1 Presentation of the research questions and the motivation for the research . ....
2.5.9 Estimation of the representativeness of the study sample .........52.
Performance, Pressures, and Politics: Motivators for Adoption of Interorganizational Information Systems Helle Zinner Henriksen Ph.D. dissertation Center for Electronic Commerce Department of Informatics Copenhagen Business School
ii
Acknowledgments “Have you considered to write a Ph.D.?” professor Niels Bjørn-Andersen asked me, at our very first meeting where we had a conversation about my possible future employment as a research assistant at the Department of Informatics at CBS. I was tempted to meet the challenge. Later on I realized that this challenge would become one of the major tests in my life so far. However, at this point in time it feels more like a triumph to have written a Ph.D. dissertation. I owe a great many thanks to all the people who helped me in the process. First of all I am indebted to the whole faculty and staff at the Department of Informatics. Everybody there has been outstanding in providing valuable help and support, especially professor Mogens Kühn Pedersen and professor Karlheinz Kautz have shown particular interest in my work. I am specially greateful to professor Niels Bjørn-Andersen who first challenged me to write the Ph.D., and in the same breath I must thank my supervisor associate professor Kim Viborg Andersen who patiently guided me through many to me unknown steps in the process. Particularly I will take this opportunity to thank Kim Viborg Andersen for his willingness to introduce me to the IS-research community. From the very beginning of the Ph.D. study Kim Viborg Andersen arranged an international advisory board. I am very indebted to the four people who kindly responded to Kim Viborg Andersens’ invitation to join the board. Professor John King, professor Kalle Lyytinen, associate professor Jens Hörlück, and professor Jan Damsgaard kindly offered help and advice. I would also like to thank professor Torben Petersen from the Department of International Economics and Management at CBS. Torben Petersen shed light on many of the darker mysteries of statistics. Torben Petersen and my husband Erik Henriksen offered valuable assistance in my struggle with the quantitative part of the dissertation. Assistant professor Dan Otzen from the Department of Accounting at CBS has also been a great source of inspiration. Dan and I have had many fruitful conversations about our
iii
dissertations and in particular about philosophy of science on our trainrides to and from Copenhagen. During the three-year period of the Ph.D. study I was given the opportunity to spend seven months at CRITO, University of California at Irvine. Professor Ken L. Kraemer kindly allowed me to join his research team at CRITO. The group of researchers and Ph.D. students at CRITO were a good source of inspiration for gaining knowledge about rigor and relevance in research. Let me also use this opportunity to express my gratitude to professor Paula Swatman, professor Kalle Lyytinen, and professor Jan Damsgaard for their critique, advice, and recommendations at the pre-defense of the dissertation in May 2001. Their feedback convinced that it would not be impossible to complete the Ph.D. – though a lot of work still had to be done. There still is one important group of people, which I owe great many thanks. Namely, all those who most willingly provided data for this study. The representatives from the two major Danish business associations and from the Danish EDI Council provided valuable support throughout the project. Let me also thank the eight companies involved in the TradeDocument Project who kindly shared their experiences from the project with me and thereby provided interesting data for the qualitative part of this study. I also appreciate that about 250 managers in the steel and machinery industry offered their time and filled out the questionnaire, which provided data for the quantitative part of this dissertation. Finally, I am, as always, indebted to my dear husband who patiently lived with me, listened to me, and lovingly supported me during the ups and downs in the process.
Helle Zinner Henriksen, March 2002
iv
Table of contents 1 Presentation of the research questions and the motivation for the research ........................................................................................................ 1 1.1 Appetizer............................................................................................. 1 1.2 Presentation of the research questions ................................................ 3 1.3 Reasons for studying motivation for adoption in a Danish context ... 8 Focal area of the dissertation ................................................................... 10 1.5 Structure of the dissertation .............................................................. 13 2
Empirical design and research method............................................ 17 2.1 Introduction....................................................................................... 17 2.2 Empirical design ............................................................................... 18 2.2.1 Flyvbjerg: A source of inspiration for the research design........ 21 2.3 Assessing case studies in IS research ............................................... 25 2.3.1 The case study method............................................................... 26 2.3.2 Methodological considerations .................................................. 29 2.4 Collection of qualitative data ............................................................ 35 2.4.1 Sources of biases in interviews .................................................. 38 2.5 Quantitative data-collection .............................................................. 41 2.5.1 Introduction ................................................................................ 41 2.5.2 Design of the study survey instrument ...................................... 42 2.5.3 Collection of data ....................................................................... 45 2.5.4 Validation of the coded data ...................................................... 46 2.5.5 Response rate ............................................................................. 47 2.5.6 Distribution of adopters and non-adopters................................. 48 2.5.7 Definition of adopters, planners, and non-adopters in the present study .................................................................................................... 50 2.5.8 Non-response bias with respect to adoption of EDI .................. 51 2.5.9 Estimation of the representativeness of the study sample ......... 52 2.5.10 Selecting methods for data analysis........................................... 53 2.5.11 Critique of the quantitative data-collection ............................... 56 2.6 Research strategy .............................................................................. 57 2.7 Summing up Chapter 2 ..................................................................... 59 v
3
The Danish business environment from an IT perspective............ 61 3.1 Industry structure .............................................................................. 61 3.2 The Danish steel and machinery industry......................................... 63 3.3 Use of EDI in the Danish business environment .............................. 64 3.4 Background for use of EDI and e-commerce in Denmark ............... 66 3.4.1 Introduction ................................................................................ 66 3.4.2 The e-commerce action plan from 1996 .................................... 66 3.4.3 E-focus – an e-commerce agenda from 1999 ............................ 75 3.5 Assessment of the influence of the action plans............................... 77 3.5.1 Coordinated diffusion initiatives................................................ 78 3.5.2 Examples of institutional initiated diffusion projects ................ 80 3.6 One, two, three – regulate and execute adoption of EDI! ................ 84 3.7 Summing up Chapter 3 ..................................................................... 90
4 Looking into an initiative launched by the industry and trade associations................................................................................................. 91 4.1 Introduction....................................................................................... 91 4.2 The TradeDocument Project ............................................................. 93 4.2.1 The preface of the TDP.............................................................. 93 4.2.2 The project organization and participants.................................. 96 4.2.3 The work .................................................................................. 100 4.3 Interviews with the TDP participants ............................................. 108 4.3.1 The actual use of EDI in the company..................................... 108 4.3.2 Reasons for adoption................................................................ 111 4.3.2.1 Company B ......................................................................... 111 4.3.2.2 Company C ......................................................................... 113 4.3.2.3 Companies D and H ............................................................ 115 4.3.2.4 Company F.......................................................................... 116 4.3.2.5 Summary of reasons for adoption....................................... 116 4.3.3 Reasons and reflections of non-adopters ................................. 117 4.3.3.1 Company E.......................................................................... 117 4.3.3.2 Company G ......................................................................... 118 4.3.3.3 Company I........................................................................... 120 4.3.3.4 Summary of reasons and reflections of non-adopters ........ 121
vi
4.3.4 Summing up motivators for adoption of EDI among the TDP participants ......................................................................................... 121 4.3.5 Degree of EDI usage among the TDP participants.................. 124 4.3.6 Degree of adoption of the software developed as part of the TDP .................................................................................................. 128 4.3.7 The suitability of the EDI subset developed during the project .... .................................................................................................. 131 4.3.8 The companies general benefits from the project.................... 132 4.4 Discussion of the findings............................................................... 136 4.5 Assessment of the data collection procedure based on the principles presented by Klein and Myers ............................................................... 141 4.6 Summing up Chapter 4 ................................................................... 147 5
IOS and EDI...................................................................................... 151 5.1 Introduction..................................................................................... 151 5.2 Views on IT and IS ......................................................................... 152 5.3 Defining IOS in a MIS context ....................................................... 157 5.3.1 Business value of IOS .............................................................. 161 5.3.2 IOS from an organizational point of view ............................... 163 5.3.3 Applied view on IOS................................................................ 164 5.4 EDI .................................................................................................. 164 5.4.1 Research strategy ..................................................................... 165 5.4.2 Defining EDI............................................................................ 167 5.4.3 Views of EDI ........................................................................... 169 5.4.4 Review of one decade’s publications on EDI from a MIS perspective.......................................................................................... 171 5.4.5 Micro, meso, and macro levels of analysis.............................. 172 5.4.6 Adoption issues ........................................................................ 175 5.4.7 Summary of a decade’s EDI research published in the top-five MIS journals ....................................................................................... 179 5.5 Revisiting the TDP case.................................................................. 182 5.6 Summing up Chapter 5 ................................................................... 184
6
Adoption of IOS................................................................................ 185 6.1 Introduction..................................................................................... 185 6.2 Adoption and diffusion ................................................................... 189 vii
6.2.1 A definition of adoption........................................................... 189 6.2.2 Diffusion .................................................................................. 190 6.2.3 Adoption and diffusion of innovations .................................... 194 6.3 Three conceptual elements from adoption research ....................... 196 6.3.1 Main drivers for adoption ........................................................ 197 6.3.2 Approaches to adoption research ............................................. 197 6.3.3 Explanatory factors .................................................................. 199 6.4 Perceived attributes of innovations................................................. 201 6.4.1 Defining innovation ................................................................. 203 6.4.2 Classifying innovations............................................................ 204 6.4.2.1 Source ................................................................................. 205 6.4.2.2 Type .................................................................................... 206 6.4.2.3 Effect................................................................................... 208 6.4.3 Rogers’ five perceived attributes determining adoption of an innovation in relation to EDI.............................................................. 210 6.4.4 Summing up innovation and perceived attributes of innovation ... .................................................................................................. 215 6.5 The innovation-decision, communication channels, nature of social system, and extent of change agents’ promotion efforts ....................... 215 6.5.1 Communication channels......................................................... 219 6.5.2 Nature of the social system ...................................................... 221 6.5.3 Extent of change agents’ promotion efforts............................. 223 6.6 Adoption of IOS innovations.......................................................... 224 6.6.1 Recent studies focusing on IOS adoption and non-adoption... 225 6.6.2 Tornatzky and Fleischer or Rogers – or a fusion of the two models?............................................................................................... 231 6.7 Operationalization of the three contexts from previous IS research .... ......................................................................................................... 232 6.7.1 The organizational context....................................................... 232 6.7.2 The environmental context ...................................................... 234 6.7.3 The technological context ........................................................ 237 6.7.4 Explanatory power of the three contexts ................................. 238 6.8 Summing up Chapter 6 ................................................................... 241 7
Operationalization of variables motivating adoption ................... 245 7.1 Introduction..................................................................................... 245 viii
7.2 7.3 7.4 7.5
Nature of the opinion data items..................................................... 246 Organizational context .................................................................... 248 Environmental context .................................................................... 255 Technological context..................................................................... 259
8 A survey of the motivation for EDI adoption in the Danish steel and machinery industry.......................................................................... 265 8.1 Introduction..................................................................................... 265 8.2 Quantitative assessment of insights gained from the TDP ............. 267 8.2.1 Introduction .............................................................................. 267 8.2.2 Exploring the relationship between the size of the company and adoption of EDI .................................................................................. 268 8.2.3 Exploring the relationship between legal status of the company and EDI adoption................................................................................ 271 8.2.4 Exploring the relationship between manufactures and wholesalers and EDI adoption............................................................ 273 8.2.5 Multivariate analysis of adoption, company size, dependency, and sector............................................................................................ 276 8.3 Testing the Tornatzky & Fleischer model for adoption ................. 278 8.3.1 Objective of the statistical analysis of motivators for the adoption-decision ............................................................................... 279 8.3.2 Overview of procedures applied in the analysis ...................... 280 8.3.3 Measurement ............................................................................ 281 8.3.4 Reliability................................................................................. 283 8.3.5 Descriptive statistics of the fifteen items included in the analysis .................................................................................................. 284 8.3.6 Exploratory two-way analysis of opinion data items statistically related to the three adoption levels..................................................... 286 8.3.7 Exploratory multivariate analyses of opinion data items statistically related to the three adoption levels ................................. 287 8.3.8 Selection of opinion data items for logistic regression analysis.... .................................................................................................. 288 8.3.9 Defining the analysis samples and the holdout samples.......... 290 8.3.10 Logistic regression analyses for adopters, planners, and nonadopters based on the analyses samples............................................. 290 8.3.10.1 Logistic regression analysis for adopters.......................... 291 ix
8.3.10.2 Logistic regression analysis for planners.......................... 292 8.3.10.3 Logistic regression analysis for non-adopters .................. 293 8.3.11 Classification of events and non-events for adopters, planners, non-adopters ....................................................................................... 293 8.3.11.1 Definitions used in the classification tables ..................... 293 8.3.12 Validation of results from logistic regression analyses ........... 295 8.3.13 Inclusion of variables from proposition 1 to 3 in the logistic regression analyses ............................................................................. 297 8.3.14 Summary of logistic regression analyses................................. 297 8.4 Quantitative assessment of the second research question .............. 298 8.4.1 Factors motivating adoption .................................................... 300 8.4.1.1 Interpretation of motivators leading to adoption ................ 301 8.4.2 Factors motivating companies planning to adopt EDI ............ 302 8.4.2.1 Interpretation of motivators influencing planners .............. 303 8.4.3 Factors causing a non-adopting attitude towards EDI ............. 304 8.4.3.1 Interpretation of reasons for remaining a non-adopter ....... 305 8.5 Summing up Chapter 8 ................................................................... 307 9
Discussion and triangulation ........................................................... 309 9.1 Outcome of the qualitative assessment of the second research question .................................................................................................. 309 9.2 Outcome of the quantitative assessment of the second research question .................................................................................................. 313 9.3 Triangulation based on the two bearings taken in businesses in the Danish steel and machinery industry..................................................... 317 9.4 An empirically derived motivation model for IOS adoption.......... 322 9.5 Assessment of the strength of the DDMM as a theoretical contribution ............................................................................................ 326 9.6 Summing up Chapter 9 ................................................................... 328
10 Conclusion ......................................................................................... 331 10.1 Introduction.................................................................................... 331 10.2 Overview of the study.................................................................... 332 10.3 Explicit answers to the dissertation’s research questions.............. 333 10.3.1 Objective A and research questions 1a and 1b ........................ 334 10.3.2 Objective B and research question 2....................................... 335 x
10.4 Conclusions of the study................................................................ 336 10.5 Contributions from this study ........................................................ 340 10.6 Limitations ..................................................................................... 342 10.7 Recommendations for future research........................................... 344 10.8 Post Scriptum................................................................................. 345 11 References ......................................................................................... 347 12 Appendixes ........................................................................................ 367 12.1 Appendix A: Questionnaire ........................................................... 368 12.2 Appendix B: Statistical tables and figures..................................... 374 12.2.1 Tables belonging to Chapter 2 ................................................. 374 12.2.2 Tables and Figures belonging to Chapter 8 ............................. 377 12.2.2.1 Abbreviations used in analyses of motivation for adoption of EDI.............................................................................................. 377 12.2.2.2 Estimation of the odds ratio for AgreeI08 for adopters.... 404
xi
List of Figures Figure 1-1. Research questions and their objectives..................................... 6 Figure 1-2. Focal area of the dissertation.................................................... 10 Figure 1-3. Schematic overview of the dissertation.................................... 16 Figure 2-1. Research approaches used in the project.................................. 18 Figure 2-2. Distribution of percentages of adopters, non-adopters, and planners................................................................................................. 49 Figure 4-1. The project organization of the TDP........................................ 97 Figure 4-2. Trade among the participating companies ............................. 100 Figure 4-3. Activities and communication during the TDP process......... 101 Figure 4-4. Motivators and barriers for adoption of EDI among the TDP participants ......................................................................................... 122 Figure 4-5. Degree of usage and EDI transaction format for the TDP participants ......................................................................................... 125 Figure 5-1. Views of technology............................................................... 156 Figure 6-1. The organizational innovation process................................... 195 Figure 6-2. An innovation typology.......................................................... 209 Figure 6-3. The Tornatzky and Fleischer (1990) model for adoption ...... 229 Figure 8-1. Relative Risk of being a non-adopter amongst all manufacturers with respect to company size ............................................................. 275 Figure 8-2. Dependency graph illustrating EDI adoption and the three variables related to propositions 1, 2, and 3....................................... 277 Figure 8-3. Dependence graph from an exploratory, initial screening of adopters and EDI adoption motivators............................................... 399 Figure 8-4. Dependence graph from an exploratory, initial screening of planners and EDI adoption motivators............................................... 401 Figure 8-5. Dependence graph from an exploratory, initial screening of non-adopters and EDI adoption motivators ....................................... 403 Figure 9-1. The Double Domain Motivation Model for IOS adoption ... 323
xii
List of Textboxes Textbox 3-1. Greetings to the 1996-action plan, a review of the plan ....... 74 Textbox 4-1. Narration of experiences from the October meeting of the TDP..................................................................................................... 103
List of Tables Table 2-1. Seven principles for interpretive field research......................... 34 Table 2-2. Response details......................................................................... 48 Table 2-3. Test for possible non-response bias......................................... 374 Table 2-4. Contingency table showing geographical location and number of employees for all manufactures in the steel and machinery in Denmark based on figures for the year 2001 provided by the manufactures business association............................................................................ 375 Table 2-5. Contingency table showing geographical location and number of employees for the survey sample of manufactures in the steel and machinery in Denmark ....................................................................... 376 Table 2-6. Shapiro-Wilks test for normality of opinion data items ............ 55 Table 3-1. The 1996-action plan ................................................................. 70 Table 4-1. TradeDocument Project: Objectives.......................................... 95 Table 4-2. Success criteria for the TDP ...................................................... 96 Table 4-3. Main activity, number of employees and annual turnover in the participating companies in the two TDP work groups......................... 98 Table 4-4. The role in the adoption process among the participants in the TDP..................................................................................................... 123 Table 4-5. The four VIDS-measures for EDI usage ................................. 126 Table 4-6. VIDS-test of EDI users in the TDP ......................................... 126 Table 4-7. Overview of the findings from the interviews......................... 137 Table 5-1. Four epistemological views of technology.............................. 153 Table 5-2. EDI articles from the top-five MIS journals, 1991 to 2000 .... 166 Table 5-3. Classification of EDI research themes in the 1991 to 2000 topfive MIS journal review ..................................................................... 180 xiii
Table 6-1. Three diffusion perspectives.................................................... 192 Table 6-2. Variables determining the rate of adoption of innovations ..... 201 Table 6-3. MIS studies of adoption motivators related to IS.................... 227 Table 6-4. Explanatory power of the three contexts ................................. 240 Table 8-1. Definition of adoption levels ................................................... 377 Table 8-2. Definition of research constructs ............................................. 377 Table 8-3. Opinion data items and the respective questions from the survey instrument ........................................................................................... 377 Table 8-4. Adoption level versus company size ....................................... 378 Table 8-5. Adoption level: Non-adopters and planners versus company size ............................................................................................................ 379 Table 8-6. Adoption level: Non-adopters and adopters versus company size ............................................................................................................ 380 Table 8-7. Adoption level: Planners and adopters versus company size.. 381 Table 8-8. Dependence versus adoption level .......................................... 382 Table 8-9. Adoption level versus company size for dependent companies ............................................................................................................ 383 Table 8-10. Adoption level versus company size for independent companies ........................................................................................... 384 Table 8-11. Position in supply-chain versus adoption level ..................... 385 Table 8-12. Relative Risk of being a non-adopter for manufactures versus wholesalers ......................................................................................... 275 Table 8-13. Analysis of two-way tables.................................................... 386 Table 8-14. Analysis of hidden associations............................................. 386 Table 8-15. The final model...................................................................... 386 Table 8-16. Internal consistency indices of reliability for the constructs: Organizational context, environmental context, and technological context for adopters ............................................................................ 387 Table 8-17. Internal consistency indices of reliability for the constructs: Organizational context, environmental context, and technological context for planners ............................................................................ 387 Table 8-18. Internal consistency indices of reliability for the constructs: Organizational context, environmental context, and technological context for non-adopters..................................................................... 388
xiv
Table 8-19. Summary of the values of Cronbach’s coefficient alpha test for scale reliability .................................................................................. 284 Table 8-20. Marginal distribution of opinion data items on seven-point Likert-scales for adopters in percentages ........................................... 389 Table 8-21. Marginal distribution of opinion data items on seven-point Likert-scales for planners in percentages ........................................... 390 Table 8-22. Marginal distribution of opinion data items on seven-point Likert-scales for non-adopters in percentages.................................... 391 Table 8-23. Spearman correlation coefficient for adopters ...................... 392 Table 8-24. Spearman correlation coefficient for planners ...................... 393 Table 8-25. Spearman correlation coefficient for non-adopters ............... 394 Table 8-26. Summary of adoption-decision motivators based on Spearman correlations ........................................................................................ 285 Table 8-27. Adoption level versus concordance/ discordance.................. 395 Table 8-28. Listing of results from analyses of two-way tables of adopter versus different items based on Fischer’s Exact two-sided p-values. 396 Table 8-29. Listing of results from analyses of two-way tables of planner versus different items based on Fischer’s Exact two-sided p-values. 396 Table 8-30. Listing of results from analyses of two-way tables of nonadopter versus different items based on Fischer’s Exact two-sided pvalues.................................................................................................. 397 Table 8-31. Exploratory multivariate analysis for adopters...................... 398 Table 8-32. Exploratory multivariate analysis for planners...................... 400 Table 8-33. Exploratory multivariate analysis for non-adopters .............. 402 Table 8-34. Listing of items for inclusion in logistic regression analysis 289 Table 8-35. Summary of logistic regression analysis for adopters........... 291 Table 8-36. Summary of logistic regression analysis for planners........... 292 Table 8-37. Summary of logistic regression analysis for non-adopters ... 293 Table 8-38. Combinations of predicted and actual cases.......................... 294 Table 8-39. Classification table for logistic regression for adopters based on the analysis sample............................................................................. 405 Table 8-40. Classification table for logistic regression for planners based on the analysis sample............................................................................. 406 Table 8-41. Classification table for logistic regression for non-adopters based on the analysis sample.............................................................. 406
xv
Table 8-42. Predicted number of cases and actual number of cases for adopters............................................................................................... 407 Table 8-43. Predicted number of cases and actual number of cases for planners............................................................................................... 407 Table 8-44. Predicted number of cases and actual number of cases for nonadopters............................................................................................... 408 Table 8-45. Combinations of AgreeI08 and AgreeI11 versus predicted and actual number of cases for adopters ................................................... 409 Table 8-46. Combinations of AgreeI10 and AgreeI15 versus predicted and actual number of cases for planners ................................................... 410 Table 8-47. Combinations of AgreeI08, AgreeI10, and AgreeI11 versus predicted and actual number of cases for non-adopters ..................... 411 Table 8-49. Summary of logistic regression including variables from propositions 1-3 ................................................................................. 297 Table 8-50. Summary of logistic regression analysis for adopters, planners, and non-adopters showing odds ratios for significant decision variables.............................................................................................. 298 Table 9-1. Overview of motivators for EDI adoption.............................. 320
xvi
Dansk resumé I 1996 blev den nationale handlingsplan for elektronisk handel i Danmark udsendt af IT og Forskningsministeriet (1996). Formålet med denne handlingsplan var i løbet af få år at gøre det muligt for danske virksomheder og den offentlige sektor at udveksle elektroniske meddelelser i form af EDI (electronic data interchange). To af dansk erhvervslivs store interesseorganisationer, som til sammen repræsenterer cirka 7.800 industrivirksomheder og grossister, iværksatte et pilotprojekt i forlængelse af den nationale handlingsplan for elektronisk handel. Dette pilotprojekt, der blev benævnt Projekt HandelsDokumenter, havde som målsætning, at skabe de ideelle betingelser for anskaffelse og udbredelse af EDI i jern- og maskinindustrien. Projektet opnåede imidlertid ikke den forventede succes. Med empirisk udgangspunkt i den nationale handlingsplan for elektronisk handel samt i Projekt HandelsDokumenter belyser denne afhandling faktorer, som motiverer til anskaffelse og udbredelse af EDI i en branche, der er karakteriseret ved en lang række små og mindre virksomheder, ganske få store virksomheder, stor specialisering samt en høj grad af samhandel. Anskaffelse og udbredelse af teknologiske innovationer har i en årrække været et aktivt forskningsområde. En af de hyppigst anvendte forklaringsmodeller for anskaffelse og spredning af teknologiske innovationer hviler på en teori af Rogers (1995), som tager udgangspunkt i, at motivationen for anskaffelse og udbredelse af teknologiske innovationer primært skal søges i sociale processer. I de senere år er der imidlertid rejst kritik af denne og andre traditionelle teoriers evne til at forklare netop anskaffelse af interorganisatoriske, teknologiske innovationer som for eksempel EDI (Prescott og Conger, 1995; Lyytinen og Damsgaard, 2001). Et af hovedmålene med denne Ph.D. afhandling er derfor systematisk at belyse, hvilke forklaringsmodeller der anvendes i forbindelse med anskaffelse og udbredelse af teknologiske interorganisatoriske innovationer.
xvii
EDI involverer en lang række aktører i virksomhederne og forudsætter samarbejde på tværs af de involverede virksomheder. Disse forhold lader sig vanskeligt forklare alene ved hjælp af Rogers’ teori. Kritikken af Rogers’ teori er søgt imødekommet ved at afprøve Rogers’ teori i forhold til data fra jern- og maskinindustrien i Danmark. I forlængelse af denne afprøvning og analyse er litteraturen inden for området gennemgået med henblik på at undersøge, hvilke alternative modeller der tidligere er anvendt som tolkningsramme. Denne analyse resulterede i tre overordnede kategorier af motivationsfaktorer: Den organisatoriske kontekst, den omgivende kontekst og den teknologiske kontekst. Disse tre motivationsfaktorer er inspireret af Tornatzky og Fleischers (1990) model for anskaffelse og udbredelse af teknologiske innovationer i organisationer. Hvorvidt disse tre kontekster kan forklare motivationen for anskaffelse af interorganisatoriske informations systemer som for eksempel EDI, er et af afhandlingens hovedspørgsmål, og det er inspireret af disse tre kontekster. Selve studiet er drevet af praksis fremfor af teori (Zmud, 1998), og det er baseret på en eksplorativ undersøgelse af jern- og maskinindustrien i Danmark. Undersøgelsen, som tager udgangspunkt både i data fra et casestudie og en spørgeskemaundersøgelse af sektoren, inkluderer de virksomheder, som har anskaffet EDI, de virksomheder der planlægger at anskaffe EDI samt de virksomheder, der ikke har nogen planer om at anskaffe EDI. På baggrund af en triangulering af kvalitative og kvantitative data i dette studie kan det konkluderes, at motivationsfaktorerne for anskaffelse af EDI i jern- og maskinindustrien primært er drevet af faktorer relateret til den omgivende kontekst. Det vil sige, at motivationen for anskaffelse af EDI primært skal ses i relation til konkurrencemæssige forhold og i relation til pres fra samhandelspartnere. De to øvrige kontekster kan imidlertid ikke helt udelukkes som motivationsfaktorer i forhold til virksomhedernes valg eller fravalg af EDI. Baseret på analysen af kvalitative og kvantitative data kan det imidlertid konkluderes, at det er nødvendigt at medtage andre faktorer, hvis der skal
xviii
findes en meningsfuld forståelse for, hvorfor virksomhederne agerer, som de gør. Empirien peger i retning af, at de parametre, Rogers inkluderede i sin teori, giver et væsentligt tolkningsbidrag til de tre kontekster præsenteret af Tornatzky og Fleischer. På baggrund af denne observation foreslås det, at de to modeller kombineres i den ”Dobbelte Domæne Motivations Model”, der kan anvendes som et mere fyldestgørende analyseværktøj i forbindelse med en afdækning af barrierer og drivkræfter for anskaffelse og udbredelse af interorganisatoriske informations systemer, som hviler på hierarkiske samhandels strukturer.
xix
xx
1 Presentation of the research questions and the motivation for the research 1.1 Appetizer Why are some organizations adopting a technological innovation that is announced to yield both operational and strategic benefits while others hesitate, or decide not to adopt for the time being? This question is highly relevant especially in the case of interorganizational information systems (IOS) exemplified by EDI (electronic data interchange), due to the earnest attention this particular technology had received throughout the 1990s from academia, practitioners, and governments all over the world. Surprisingly few Danish organizations have adopted EDI. If their technical capabilities and their high degree of computerization are considered then the organizations’ reluctance to adopt EDI appears to be even more irrational and incomprehensible. The phenomenon of organizations lagging behind adoption of IT regardless of their opportunities to do so is well known (Harrison et al., 1997). What is missing so far are factors explaining the reason for this situation. Small companies dominate the Danish business sectors. About two third of approximately 50,000 companies within the industrial production sector have less than ten employees. National and international industry and trade associations have created a number of awareness campaigns and have focused on the development of beneficial conditions for the SMEs (small and medium sized enterprises) in order to get them to adopt new technology, especially IOS such as EDI. The aim is to help the companies reduce or eliminate work routines and to support them in a market characterized by increased competition. The technological development has 1
led to an increase in the quality and functionality of hardware and software and a corresponding decrease in cost (Harrison et al., 1997). The traditional technological barriers for organizational adoption of IOS such as EDI might therefore not play the same dominant role as it did earlier. These conditions make it highly relevant to examine explanatory factors for EDI adoption among SMEs, which traditionally have less slack resources or resources allocated to IS compared to larger companies (Lai and Guynes, 1997). Throughout the history of IS research a number of explanatory factors have been found to be prevalent in relation to adoption and diffusion of IT in and among organizations. A summary of an extensive review of the literature on IOS and EDI and adoption and diffusion of IOS and EDI has resulted in a number of possible explanatory factors and models for adoption. One of these models comprises factors related to the organizational context, the environmental context, and the technological context (Tornatzky and Fleischer, 1990). These three contexts or dimensions are the pivotal point for this dissertation on adoption and diffusion of IOS in a Danish setting. Inherent in the three contexts is for example the influence on organizations from environmental factors such as action plans promoted by governmental institutions or by industry and trade associations and competitive market forces. From an internal point of view, the motivators for adoption of innovations in organizations focus on the operational benefits and human resources. Finally, the influence from innovation attributes related to the technology aspects is comprised in the third context. The purpose of Chapter 1 is to present the research questions along with motivations for exploring the research questions. The focal area of the study is presented. This presentation includes a conceptual overview of the literatures applied in the study. The aim is to create a coherent picture of the different theoretical fields that are applied in the study of factors motivating adoption of interorganizational information systems exemplified by EDI. Chapter 1 ends with an overview of the remaining chapters in the dissertation.
2
1.2 Presentation of the research questions The objective of this Ph.D. dissertation is to examine motivators for adoption or non-adoption of IOS in a business sector dominated by small enterprises. Previous research suggests that there is need for separately examining IS adoption in small businesses due to the unique characteristics of small businesses (Thong, 1999). Unique characteristics include centralized structures where the owners and daily managers make most of the critical decisions (Mintzberg et al., 1976). Previous research has focused on the complexity of the adoption decision-making process, when technological innovations are considered for adoption in organizations. Five reasons have been found to be important for understanding the IS decision-making process: 1) Development of a strategic IS system is a long process where it is important not to fall behind competitors, 2) Potential benefits are difficult to evaluate, 3) The systems require significant resources, 4) It is difficult to evaluate whether or not the competitive advantage can be sustained, and 5) The decision-process might be quite complicated, being influenced by diverse groups in the organization (Sabherwal and King, 1995). In this project, focus is on the fifth reason, the decision-process, and especially the motivating factors leading to an adoption-decision. Instead of solely focusing on the influence from diverse groups in the organization, focus is rather on influences from the business environment including professional business associations and governmental units. Instead of depending solely on previous research, the research field has been approached in an exploratory manner allowing for any outcome, which appears to possess some degree of generalizable value. This results in a less structured analysis of adoption motivators compared to applying the more traditional and well-documented frameworks. The strength of this non-structured exploratory study is however, that it is capable of drawing attention to alternative views that might have been missed if a predefined framework had been used. However, after a preliminary analysis of data it became evident that the adoption-decision framework proposed by Tornatzky and Fleischer (1990) could be appropriate for further 3
investigation of the motivators for adoption of EDI in the Danish steel and machinery industry, which is the business sector under investigation. It is recognized that though motivators for adoption are multiple and that it might be coincidences that determine adoption or non-adoption of an innovation in an organization (Mohr, 1987) it was found appropriate to focus on a set of motivators. The next step was therefore to analyze whether or not these motivators had any explanatory power related to adoption of EDI in the Danish steel and machinery industry. Tornatzky and Fleischer (1990) suggested that adoption of an organizational innovation comprise three contexts: The organizational context, the environmental context, and the technological context. These three contexts are chosen as the overall guiding framework for exploring the motivators related to adoption or non-adoption of EDI in the Danish steel and machinery industry. The three contexts are operationalized based on previous research and empirical experience in the field. The theory of diffusion of innovations as presented by Rogers (1995) focuses on presenting the process from the point in time where the potential adopter becomes aware of the innovation and until the innovation is integrated among a group of users. Rogers recognizes that the innovationdecision process plays a vital role in the process leading to acceptance or rejection of an innovation. The explicit motivators for the organizational adoption are described by a set of variables, which can be related to most technological innovations. In relation to IOS and EDI it has been discussed whether or not these adoption determinant variables are capable of explaining adoption and diffusion of this type of complex technological artifacts (Lyytinen and Damsgaard, 2001; Prescott and Conger, 1995). It is however worth noting that IS studies on adoption and diffusion of innovations in organizations often refer to Rogers, even though it is widely recognized that the explanatory power of Rogers’ theory is found to be low, when it comes to organizational adoption and diffusion of information systems. Instead of rejecting the explanatory power of the variables presented by Rogers, an assessment of each variable is made in relation to the case study data obtained in this study.
4
Prescott and Conger (1995) concluded in their review of 70 studies on diffusion of innovations, that the studies based on interorganizational technologies have found innovation diffusion literature to be less relevant than studies on internal or intraorganizational information technologies. In a recent work by Lyytinen and Damsgaard (2001), this line of thought is pursued. Lyytinen and Damsgaard examined the appropriateness of the Rogerian diffusion theory in relation to EDI. The two authors concluded that even though the traditional diffusion of innovation theory has had a positive impact on IS research it falls short on some theoretical constructs, that address how complex networked technologies can and will diffuse. A departure from technical determinism as a focus point in relation to IS research has been advocated a decade ago (Kling, 1991). Kling argued that, “Social and economic considerations have to be included.” The issue is however, which social and economic characteristics should be included and how should they be operationalized. Previous research on EDI within MIS has to a large extend rested upon economic premises related to value-chain thinking (Krcmar et al., 1995), market driven forces such as transaction costs (Malone et al., 1987) and agency theory (Gurbaxani and Wang, 1991), whereas socio-political factors have played a minor role (Premkumar and Ramamurthy, 1995). It is however presumed, that the regulatory regime and the interorganizational conditions along with the economic incentives influence the innovationdecision, especially when it comes to adoption of complex information systems. This combination of theory of diffusion of innovations, organizational change, and management of information systems is not an unknown phenomenon (Premkumar et al., 1994; Cooper and Zmud, 1990). Regulatory regimes have also been used as a starting point for research (Klein, 1995). Based on the above-mentioned considerations and theoretical inputs three explicit research questions are formulated:
5
Figure 1-1. Research questions and their objectives
Objective A To consider possible improvements of the present adoption models used in MIS research
Research question 1a How are the explanatory variables related to motivators for EDI adoption defined in MIS research at present? Research question 1b Which models are used to explain motivation for IOS adoption at present?
Objective B To identify motivators for adoption of interorganizational information systems in a business sector dominated by small businesses
Research question 2 To which degree can the motivation to adopt interorganizational information systems, exemplified by EDI, be explained by issues related to the organizational context, the environmental context, and the technological context?
To answer research question 1a it is necessary to define what EDI is. When studying diffusion of innovations one necessary requirement is to have a clear set of concepts (Scott Poole and Van De Ven, 1989). The literature on EDI is vast and approaches to EDI are numerous. One research stream focuses on EDI from a technical point of view (Damsgaard and Truex, 2000; Horlück, 1994). Another way of looking at EDI is from an economic and socio-political perspective. The economic and socio-political perspective is exclusively the one explored in this context. MIS (Management of Information Systems) is therefore the authors’ approach to this study. The major implication of this view is that the included literature is solely focused on MIS publications and journals. The explanatory variables related to motivators for EDI adoption are defined based on this research stream. In order to answer research question 1b focus is on how traditional diffusion theory explains adoption and diffusion of technological innovations. This is based on innovation theory where the different 6
dimensions of EDI as an innovation are classified. Thereafter diffusion theory is presented. Diffusion theory illustrates the different types of adoption and diffusion frameworks that have been applied as explanations for adoption and diffusion of IT/ IS. In the discussion of the adoption and diffusion theory, it is conceptually considered how this theory addresses adoption and diffusion of interorganizational complex concepts, such as EDI. The second research question concentrates on elements that can strengthen the explanatory power of diffusion theory in relation to IOS, especially in relation to EDI. The argumentation in this dissertation is supported by empirical evidence, both qualitative and quantitative. In this study, emphasis is on the motivation leading to the innovation-decision related to adoption of EDI in businesses in the Danish steel and machinery industry. Even though it not doubted that the overall diffusion process can not be isolated from cause and consequence e.g. critical mass and regulatory regimes, and even though it is recognized that diffusion of innovations is not necessarily a continuous process that goes through clearly defined stages (Newell et al., 2000), particular attention is paid to the specific state where motivation for adoption becomes explicit in the organization. Though the stage-oriented diffusion view is rejected as a tenable approach to explain the diffusion process a compromise is made and a particular point in the adoption process, which is considered as a stage in the “stageschool of thought”, is explored. However, it is the authors’ conviction that regardless of the view of the adoption and diffusion process (stages Rogers, 1995, or episodes Newell et al., 2000) there is a certain point in time where the motivation for adoption is articulated. The research questions rest on the assumption that attributes of an innovation play an important role for the adoption motivation. Even when technical issues are excluded from the discussion EDI is still a very complex innovation that involves several actors, functions, and levels of the organization. It is therefore the authors’ claim that the theory of adoption and diffusion of innovations provides the best explanatory factors for understanding the spread of EDI among businesses. However, the
7
author will challenge the traditional diffusion theory by providing qualitative and quantitative empirical evidence to the known diffusion models in order to attain better explanatory power with respect to the adoption process of complex technological innovations.
1.3 Reasons for studying motivation for adoption in a Danish context The purpose of this section is to show why the questions raised in this dissertation are of interest to the author. Different factors and perspectives motivate the study of adoption motivators with respect to adoption of IOS in private businesses. By presenting the main incentives for the study, insights into local environmental factors, empirical implications, theoretical challenges, and personal drives are offered. First, this study is motivated by a need to understand drivers and barriers for diffusion of EDI in Denmark. Despite initiatives made by governmental units and the major industry and trade associations to promote EDI and ecommerce in Denmark the diffusion of EDI seems to be hampered by factors beyond the factors presented by the traditional diffusion theory. Therefore, the incentive for exploring the issue of adoption motivation has to be seen in a local context where an e-commerce agenda was set in Denmark in the mid 90s. Two action plans for EDI and e-commerce had been launched in Denmark during the period 1996 to 1999. Both plans aimed at promoting contemporary business methods in the Danish business community and the public sector. Even though the action plans only served as recommendations, major industry and trade associations introduced pilot projects among their members specifically focusing on diffusion of EDI. These pilot-projects led to an increased awareness of EDI among businesses, but not necessarily to adoption. This is the case with one of the pilot projects, the TradeDocument Project (TDP), which is presented in detail in Chapter 4. Along with the introduction of the first action plan for EDI, a longitudinal study of the use of EDI was initiated (Andersen et al., 1999). The study, which is still going on, shows that there is significant growth in EDI traffic in Denmark. The study does not systematically 8
include qualitative explanations of why some companies use EDI while others hesitate. The primary incentive of the present study is therefore to look broadly at the effects of the Danish action plans related to EDI and ecommerce. The second motivation for this study is based on the results from the abovementioned longitudinal study and rests on the fact that the study without doubt shows there is growth in the use of EDI in Danish businesses and governmental units. What remains unknown is what motivates this growth in EDI traffic, especially when it comes to the internal adoption motivators in organizations. The second incentive for this study is therefore to examine what initiates or hampers motivation for adoption. In order to include a wide variety of explanations, perspectives covering internal factors, external factors and characteristics related to the innovation itself are included. The third incentive for studying motivation for adoption of EDI has a theoretical offset. About fifteen years ago Kwon and Zmud (1987) published a harsh critique of IS implementation1 research. Kwon and Zmud stated that, “No core set of constructs exists”. It was argued that there was a lack of common perspectives and that most studies focused on small pieces of the MIS implementation puzzle (ibid.). From the review of IOS adoption literature primarily from the last decade of the twentieth century it has become clear that the lack of core constructs is still the reality. One objective of the study is hence to operationalize a generic adoption and diffusion model using a broad framework for conceptualizing EDI adoption.
1
It should be noted that Kwon and Zmud define the term implementation as ”an organizational effort to diffuse an appropriate information technology within a user community.” Though the dissertation is focused on adoption implementation literature should not be rejected when it comes to conceptual matters.
9
1.4 Focal area of the dissertation Figure 1-2. Focal area of the dissertation
General adoption determinants Motivators
Decision
Adoption
Specific organizational adoption contexts
The reason for focusing on motivators rather than on the innovationdecision is that organizational adoption motivators are seen as the immediate expressions of intended behavior, which is based on influences from general adoption determinants (Rogers, 1995) and the specific organizational adoption contexts (Tornatzky and Fleischer, 1990). The purpose of this focus is to be able to identify factors influencing the decision to adopt. The differentiation between the initial steps leading to adoption and the innovation-decision is not unusual. Tabak and Barr (1999) distinguished between the intention to adopt and the decision to adopt. Kurnia and Johnston (2000) conceptualized the move as “organizational action leading to adoption”. Yet other researchers have inserted links between the promoters of action and adoption. For example Rai and Yakuni (1996) and Mitropoulos and Tatum (2000) respectively labeled the link “the adoption context” and “the forces for adoption”. Harrison et al. (1997) described this particular state as “the initial adoption-decision”. The term motivation is however found to be a more appropriate term from a linguistic point of view for this “link”. The word motivation is derived from the Latin word moveo. There are two meanings of this root word moveo. 1) To initiate, to leave, and to move which was originally related to
10
a military terminology. 2) To induce and bring about which was originally related to lyrics. The word motivation therefore relates to an action, which can involve both individuals and processes. Motivation therefore reflects the process defined as, “... a set of actions that begin with the identification of a stimulus for actions and ends with the specific commitment to action” (Mintzberg et al., 1976). In summary, organizational action is viewed as somewhat purposive and intentional or externally constrained and controlled. One argument for choosing this approach rather than looking at organizational action from a process point of view involving actions and structures, which largely can be viewed as socially constructed, is that the available data allows analysis based on a certain degree of rationality and efficiency. In the presentation of general adoption determinants, the model by Rogers (1995) is applied. The generic variables determining adoption are viewed as fuel for understanding the motivation for organizational adoption of EDI. The organizational adoption context model exemplified by the organizational context, the environmental context, and the technological context are related to the adoption model described by (Tornatzky and Fleischer, 1990). The purpose of including the Tornatzky and Fleischer model is to be able to specify the explanatory variables. The Tornatzky and Fleischer model is considered complementary rather than an alternative to Rogers’ model. Interorganizational systems (IOS) such as EDI have so far had a low level of diffusion (Timmers, 1999; Andersen et al., 2000). In addition, diffusion of IOS at an organizational level has until recently received relatively little attention by researchers (Lai and Guynes, 1997). Adoption of technological innovations in organizations is, however, considered a major diffusion of innovations research stream (Lai and Guynes, 1997; Prescott and Conger, 1995). The theory of diffusion of innovations as presented by Rogers (1995) is mainly concentrated on characteristics of the adopters, the characteristics of the innovation, and the communication process. Rogers’ model has been challenged by Abrahamson (1996), who suggests that diffusion is driven by managerial fashions and fads. In addition, Tornatzky
11
and Fleischer (1990) suggest that the diffusion process is a more multifaceted process, which has to be viewed from different angles such as the environmental context, the organizational context, and the technological context. Other researchers have viewed the diffusion process from a structuration point of view (Giddens, 1984) and thus distinguished between the individualist, structuralist and interactive process as the main drivers for the diffusion process (Slappendel, 1996). Diffusion of innovation theory has proven to possess a high explanatory power in diffusion of innovations at the individual level (Lai and Guynes, 1997). A review of diffusion studies within IS has however, challenged the ability to predict diffusion of technological innovations at an organizational level (Prescott and Conger, 1995). In recent studies of organizational diffusion of e.g. ISDN (Lai and Guynes, 1997) and EDI (Lyytinen and Damsgaard, 2001) there was evidence that adoption of innovations at an organizational level is not well understood if one merely applies the parameters found in the traditional diffusion of innovations theory. Mohr (1987) attempted to explain this conflict arising from having a theory with a high degree of explanatory power that however does not fit a specific type of innovation (information technology) in a specific environment (organizations). Mohr claimed that innovation theory fails to predict adoption of new technologies. He questions whether it at all is possible to predict adoption of organizational innovations. He argues that one reason, why innovation theory does not tell us what we want to know about new technology in organizations, is our failure to pinpoint precisely what our questions are. Another reason for the inability of innovation theory to describe adoption is according to Mohr, that none provides definitive answers to the individual questions we ask. To make matters worse the innovation research according to Mohr does not leave room for intuitive insights into the diffusion process. Chau and Tam (1997) discussed the shortcomings of the diffusion theory and argued that rejection of traditional diffusion theory as presented by Rogers (1995) is due to its failure to consider differences in units of analysis, environments and technology characteristics.
12
Guided by the above-mentioned considerations two different fields of literature are applied to the present study: 1) The adoption and diffusion literature, and 2) The literature on innovations. As indicated above it appears as if the diffusion theory stated by Rogers is insufficient to explain diffusion of IOS exemplified by EDI, which undoubtedly is an innovation in the traditional sense of the word. It is nonetheless seen as a natural starting point for the analysis to take an offset in the theory presented by Rogers due to the fact that almost all diffusion studies within MIS quote Rogers. Based on works by Rogers and by Tornatzky and Fleischer, and studies from MIS, motivation for organizational adoption of EDI is examined. The second stream of literature included in this study is related to innovations. Innovation is in this context limited to the process whereby an existing innovation becomes a part of an adopter’s cognitive state and behavioral repertoire (Damanpour and Gopalakrishnan, 1998) thereby excluding all types of innovativeness, which concern development and incubative behavior. By identifying general innovation characteristics, it is expected that information systems such as EDI can be placed in a theoretical better-defined context. As is the case with diffusion the theoretical offset is taken in traditional works such as Zaltman (1973), Tornatzky and Fleischer (1990), and relevant contributions from the MIS literature operating with innovation.
1.5 Structure of the dissertation Inspired by the structure of the work by Flyvbjerg (1998) this dissertation introduces the case study at an early stage in the presentation of the study. The purpose for doing so is to create a frame of reference for the later sections where the theoretical framework is presented. The use of the case in that way illustrates the exploratory approach, which has been applied throughout the study. A consequence of the early presentation of data is that data is evaluated in relation to theory throughout the presentation of theory. In the following a brief overview of the chapters of the dissertation is given.
13
Chapter two presents the empirical design of the study. All considerations related to method and methodology are discussed in Chapter 2. The chapter is also used to present all practicalities related to the accomplishment of the case study and the survey. Chapter three is a presentation of the socio-political context for the study. The proposed research questions will be answered in a sequential manner starting with exploration of the Danish industrial environment. Governmental units, industry and trade associations and major business organizations have invested substantial resources in defining, promoting and carry through actions plans on EDI and e-commerce from the mid 1990s. Part of the empirical study for this dissertation is to be found in the line of thought stemming from the first national action plan on ecommerce. The initiatives on policy level are thus a natural starting point for the presentation of the Danish business environment from an EDI point of view. Chapter four is focused on presenting the case study. The reason for presenting the case early in the dissertation is to illustrate the exploratory approach that has guided the study. Lessons learned from the case study are thus used as a frame of reference for further work in the dissertation. The case is centered on the participants of the TradeDocument Project and their motivation for adopting or non-adopting the tools developed during the project. Data from the case study is used to give a qualitative assessment of the second research question. Chapter five explores the three terms: Technology, IOS, and EDI. The empirical work is centered on adoption of EDI, which is considered a subset of IOS. To provide an overview of research on that particular technology previous research on EDI is reviewed. In order to get to a better understanding of IOS and EDI the underlying conceptualization and rationality for technology is explored at the beginning of Chapter five. The chapter provides an answer to research question 1a.
14
Chapter six is a presentation of the theoretical foundation for the study. Adoption and innovation are explored for conceptualizing the terms. Two of the most dominant adoption models are presented and discussed in relation to the empirical input from Chapters three and four. Previous research focusing on the motivation for adoption of IOS is examined and used as guidance for further exploration of the motivation for adoption of EDI in the Danish steel and machinery industry. The chapter provides an answer to research question 1b. Chapter seven is a presentation of the operationalization of the survey items based on the Tornatzky and Fleischer (1990) adoption model presented in Chapter six. Inspired by the content of the policy statements described in Chapter three, the experiences from the case study presented in Chapter four, and theoretical insights from Chapters five and six fifteen propositions related to IOS adoption are defined. Chapter eight is a presentation of the statistical analyses performed on data from a survey of the Danish steel and machinery industry. The qualitative observations from the case study that could be quantified are tested and the significance of the fifteen propositions defined in Chapter seven is explored. Based on this analysis a quantitative assessment of the second research question is made. Chapter nine presents the final analysis and discussion of the study. The outcome of the triangulation of data and an empirically derived model for the IOS adoption motivation is presented. This model is fittingly named the Double Domain Motivation Model (DDMM). Chapter ten gives an overview of the study along with the conclusion, presentation of limitations and contributions of the study.
15
Figure 1-3. Schematic overview of the dissertation
Theme Chapter 1
Purpose
Research question and motivation for research
To present the objectives and research questions at the core of the study
Empirical design, method, and methodological issues
To describe how the research questions are answered
Presentation of the Danish business environment and IT initiatives
To present the socio-political context
Chapter 2
Chapter 3
Chapter 4
Presentation of the TDP case
Chapter 5
Conceptualisation of technology, IOS, and EDI
Providing answer to research question 1a
Presentation of the theoretical framework
Providing a theoretical framework and answering research question 1b
Chapter 6
Chapter 7
Operationalization of the theoretical framework for the survey instrument
Chapter 8
Description of survey results
Chapter 9
Triangulation
Chapter 10 Conclusion
16
Qualitative assessment of research question 2
Definition of the content related to the survey instrument Quantitative assessment of research question 2 To sum up the qualitative and quantitative answers to the second research question To outline the contributions of the study
2 Empirical design and research method 2.1 Introduction In this chapter, the empirical design of the study is presented. The first section is a quick tour through the different phases of the collection of data and an introduction to the area of investigation. After the overall presentation of the empirical design, the research methods for qualitative research are discussed. Next, methodological reflections are discussed and a framework for assessing qualitative data is presented. After the discussion of these theoretical considerations, the collection of qualitative data is detailed. The presentation of the qualitative data collection comprises considerations related to an assessment of the validity of data. The last sections in this chapter concentrate on the quantitative data collection. Apart from the description of the data collection procedure, data is discussed with respect to validation of data, response rate, and nonresponse bias. The chapter ends with a critique of the quantitative data collection method and a description of considerations related to the applied research strategy. The objective of Chapter 2 is to provide an overview of the applied methods and methodologies. By presenting empirical design, considerations related to method and methodology, and procedures followed in the collection of data in one separate chapter, it is the authors’ hope that the readers will get a better overview of the entire process. An alternative would be to present collection of data and considerations related to the particular method in the two chapters concerning the qualitative data and the quantitative data. This last presentation method would however hamper the flow of the presentation of the empirical data.
17
2.2 Empirical design As stated in Chapter 1 the overall aim of the dissertation is to explore motivators leading to adoption or non-adoption of EDI in the Danish steel and machinery industry. In the initial phases of the study, knowledge of the field of inquiry was obtained through direct observation of the field, by visiting two professional business associations, by participating in business conferences, and other business meetings. The author was introduced to a pilot project initiated by two major business associations. The two business associations represent approximately 5,800 manufactures and 2,000 wholesalers. The pilot project, the TradeDocument project (TDP) aimed at supporting adoption of EDI in the steel and machinery industry. Along with direct observation, archival data of the project was also studied. At the completion of the pilot project, eight of the nine participating companies were interviewed about the project and the outcome of the project. Based on the insights gained from eight semi-structured interviews and study of archival material a questionnaire for a quantitative assessment of adoption motivators was developed. In the spring of 2000, the selfadministered questionnaire was sent to the management of 917 manufactures and wholesalers within the steel and machinery industry in Denmark. Figure 2-1. Research approaches used in the project
Observation of field of inquiry and analysis of historical data
Interviews
Survey based on interviews
As illustrated in Figure 2-1 the field of inquiry was approached in a stepwise manner and through different means of data collection methods. A qualitative, exploratory approach was applied in the preliminary phase. The 18
direct observation yielded primary data while the study of archival data provided secondary data. Both sources were used as tools to get insight into the field. In the next phase, when the interviews were performed, a more structured process was used to collect primary data. However, the interviews were semi-structured in order to get a more multi-faceted view of the field of inquiry. Semi-structured interviews could be categorized as quantitative data even though they do necessarily not meet the rigorous demands that characterize surveys (Silverman, 1993). However, the semi-structured interview leaves room for unstructured data both in the form of speech, atmosphere, and other rich sources of information. In the actual interview situation all responders had additional remarks and comments that did not fit into the interview guide, and five of the eight informants demonstrated their administrative system, showing for example how they generated a purchase order through electronic means. They also presented the different functional units of the organization explaining how orders were filled, presenting the core product of the company etc. This insight into supplementary conditions and the social setting suggests that the interviews are sources of qualitative data. Due to the fact that the object of inquiry, the TradeDocument Project, was the same in the direct observation, in the study of historical data, and in the interviews all these sources are considered to generate data for one case study. As such it can be subject to qualitative interpretation. The final step in the data collection was quantitative. In that phase a survey was performed among the entire population of the steel and machinery organizations organized in the two major industry and trade associations in Denmark. The 917 questionnaires mailed to the management of the organizations had predefined response options and as such there was no need for any further interpretation of the individual responses. The motivation for performing the survey was to get more breadth through a larger sample and a more structured information gathering process than what is normally possible through in-depth interviews.
19
To sum up, the data collection process has been a mix of primary and secondary data and of qualitative and quantitative data. By combining qualitative and quantitative data a synergistic view of evidence can be achieved (Eisenhardt, 1989) because data leaves room for both a positivistic and an interpretivist approach. The multiple data collection methods strengthen grounding of theory by triangulation2 of evidence (ibid.). A number of researchers within the MIS field have advocated combining research methods (Mingers, 1997; Mingers and Brocklesby, 1997; Pinsonneault and Kraemer, 1993) and researchers of diffusion of innovations have stressed the importance of triangulation: ”… instead of spending time attempting to find the proper type or level of analysis, we encourage using all types of research strategies and data sources, with due attention to the limitations of each. There is a tradition of, an argument for, using data from different research methodologies in a triangulation approach. Given the complexity and messiness of innovation phenomena, it is probably a wise strategy, and one that students of the field are increasingly recommending.” (Tornatzky and Fleischer, 1990).
The triangulation approach can broadly defined as, “... the combination of methodologies in the study of the same phenomenon” (Jick, 1979; Silverman, 1993). Triangulation can be “between methods” and “withinmethod” (Jick, 1979). The “within-method” triangulation is used when multiple techniques within a given method are used to collect and interpret data. An example of “within-method” triangulation is comparison of a number of cases. The “between-methods” approach to triangulation is related to the situation where the use of multiple methods is used to examine the same dimension of a research problem. Different methods are applied in the data collection related to the present study thereby leading to “between-methods” triangulation. The triangulation can, “... capture a more complete, holistic, and contextual portrayal of the unit(s) under study.”
2
The Danish astronomer Tycho Brahe apparently conceived the idea of triangulation before the end of the 16th century. Later the term has gained a footing in military strategy and navigation. It is a technique for determining distances and angles for location of, for example, a ship's position, which is obtained from bearings of multiple points. (This description is based on information from Encyclopaedia Britannica Online and Jick, 1979).
20
(Jick, 1979).3 Though the benefits of triangulation are quite obvious, the triangulation method is rarely used in MIS research (Pinsonneault and Kraemer, 1993). In the process of reporting data from the qualitative and the quantitative studies a number of different methods have been applied. The survey, which is presented in Chapter 8, is a straightforward quantitative description where output from statistical packages is presented and explained. In the case study reported in Chapter 4, different means are used to illustrate the richness of data. In addition to ordinary description of data, a narrative is included. Finally, some displays (Miles and Huberman, 1994) are included to sum up a number of important aspects presented in the case study. 2.2.1 Flyvbjerg: A source of inspiration for the research design As mentioned in the overview of the dissertation in Chapter 1 the structure of the dissertation is strongly inspired by Flyvbjerg (1998). In this section the overall ideas of Flyvbjerg in relation to research design are presented. Next, a justification for why the Flyvbjerg approach is chosen is given. Finally, an explanation of how based on the ideas of Flyvbjerg the data presentation and the data analysis is structured. Before the research design applied by Flyvbjerg is presented it should be stressed that the content of the work of Flyvbjerg is of limited relevance for the present study of adoption and diffusion of EDI in the Danish steel and machinery industry, nor are the methodological considerations (“antiEnlightenment”) applied by Flyvbjerg of any value in the present context. However, in order to illustrate the research design applied by Flyvbjerg the content of his case study is used to describe his research design, which is found to be both inspiring and appealing.
3
There are naturally some shortcomings related to the triangulation research strategy (Jick, 1979) these include difficulties in replicating research and inadequacy in relation to focusing a not clearly defined research topic.
21
Flyvbjerg characterizes his book “Rationality and Power – Democracy in Practice” as an in-depth case study, where he explores politics, administration and planning in Aalborg, a major Danish town. He uses the Wittgensteinian approach for telling the story, which Flyvbjerg denotes, narratology. This implies according to Flyvbjerg that he, “takes the reader through the streets and alleyways of Aalborg”. The city is explored before studying maps and actual practices of politics are investigated before rules. He explicitly stresses that focus is on practices rather than discourse and theory. The Aalborg case is therefore presented in terms of events where the chronology of events are in their natural context rather than in terms codes. Flyvbjerg starts his work by stating: “… it has been my aim to present my findings in the form of a narrative that would help readers move about in the dense case material, so as to provide them with the basis to form their own judgments about the case and its implications.”
In order to provide the dense case material Flyvbjerg gives detailed accounts of those interest groups and individuals involved in a specific city-planning project in Aalborg. Throughout the book most of the data is just described and sparse commented. At certain points elements are taken out and discussed theoretically. At the very end of the book summaries and generalizations are made. Analysis is therefore kept to a bare minimum, thereby giving the reader an opportunity to make her or his own interpretations. Flyvbjerg outlines his basic methodological and theoretical assumptions briefly and he presupposes that those assumptions are familiar to the reader. He does in this way spare the reader for long theoretical explanations at the beginning of his book. Instead he adds the necessary theoretical elements while outlining the case. An example of this method of adding complicated theoretical elements is the episode where Wittgenstein is invited to share his interpretation of the situation with Flyvbjerg on a walk through the streets of Aalborg. Flyvbjerg emphasizes that the role of the narrator is limited and that the case approach is subjective: 22
“... I have demurred from the role of omniscient narrator and summarizer in favor of gradually allowing the story to unfold from the diverse, complex, and sometimes conflicting stories that the actors in the case have told me.”
That the narrator is incapable of structuring complexity is however argued by Flyvbjerg to be one of the strengths of the narrative method. The narrator can provide a version of the story, but there is room for the reader to create other interpretations from the material provided by the narrator. He uses the method in such a way that it is open for multiple interpretations, and he lets data resemble the universe of Aalborgs’ city planning in a non-structured way, which reflects that the world is necessarily not driven by rationality and coherence. Apart from that he recognizes that the complexity of a given story can not be told in a linear fashion where everything fits. One reason for this is that the researcher’s insights and overview increase over time another reason is that individuals involved in the case have different understandings of a given situation depending on their knowledge and vested interests. The way Flyvbjerg structured his work was a rewarding source of inspiration since it allows the reader to draw her or his own conclusions while reading the case. Though it moves frontiers for the way doctoral work usually is conducted it was found intriguing to try to apply some of the ideas in the present dissertation. The result is that the traditional order of presentation of the work is changed. Instead of first describing and building a theoretical frame based on previous work data from the case study is presented early in the dissertation. The aim was to minimize theoretical judgment and interpretation while presenting the case, which (cf. Figure 2-1) included observation of the field of inquiry, analysis of historical data, and interviews. It is fascinating to apply the idea of guiding the reader into the universe of the Danish adoption and diffusion efforts in relation to EDI, which were initiated by the mid 1990s. This is done through a description of action taken at the macro level, where policy statements were formulated for the purpose of initiating the process of diffusion of EDI in the Danish business environment. The reason for introducing contextual information and facts, 23
which constitute the case study, is to create a frame of reference for the later sections where the theoretical framework is presented. Two aspects are benefiting from this approach: 1) The story is presented as it was told to the author “unbiased” of theoretical considerations and definitions. An example of this is that understandings of the terms EDI and pressure as expressed by the informants and stakeholders are accepted as valid even though they do not necessarily fit more formal and academic definitions. An attempt is however made to accept that practitioners have different understandings of terms compared to researchers. 2) The multiplicity of stakeholders and events that formed the environment for adoption of EDI among the TDP participants are illustrated in an a-theoretical manner. The implication of this approach is that other than researchers can also read the case without knowledge of any theoretical assumptions and constructs. This is in line with the practice driven research strategy (Zmud, 1998) which is preferred to the theory driven research strategy (for further description of the research strategy of the Ph.D. see Section 2.6). The implication of this choice of research design is that events - rather than codes guide the first part of the dissertation. First of all it must be admitted that courage failed with regards to carry out the Flyvbjerg research design rigoursly. The consequence was that theoretical chapters and reviews of previous literature are included in the dissertation. Furthermore, the Flyvbjerg research design is not followed rigoursly in relation to the method of narratives. However the presentation of data aims at illustrating the a-theoretical and exploratory approach. The presentation of data is therefore guided by the inputs gathered in the field and from browsing through archival data. Long passages of case descriptions are outlined without any theoretical judgment. Whenever it is found appropriate parallels are made to previous research and theory. The consequence is that data sometimes from a theoretical point of view appears to be “out of context”. However, this information was part of the story and it is included to illustrate the complexity of real world settings. This is as such not unusual compared to traditional case research, which per se is characterized by richness of data and complexity (cf. Section 2.3.1). What makes the presentation unusual is the analysis of data. A
24
consequence of the early presentation of data is that data is evaluated in relation to theory throughout the presentation of theory. Combined with the selected theory data from the case is used to clarify examples in relation to theory and the reviewed literature which are presented in Chapters 5 and 6. An implication of having presented data in the beginning of the dissertation is that data analysis can take place in what might appear to be “reverse order”: Analyzing data in relation to theory instead of testing theory on data.
2.3 Assessing case studies in IS research In the following two sections, reflections related to collecting and interpreting qualitative data are presented. The purpose of these sections is to describe the different approaches to research design, data collection, and interpretation of data, and especially to clarify which research method and methodology has driven this research project. These two sections focus on two concepts: Method and methodology. In this context the term method is related to research design and how the field study can be conducted, whereas the term methodology is related to the epistemological stance of the researcher. Although method and methodology are closely related an attempt is made to discuss them separately. The epistemological stance influences the approach to interpretation of data. The research epistemologies can be classified into three types: Positivistic, interpretive, and critical studies (Orlikowski and Baroudi, 1991). Orlikowski and Baroudi define the three types as follows: The premise for positivist-type studies are the existence of a priori fixed relationships within phenomena, which are typically investigated with structured instrumentation. Studies based on the interpretive approach assume that people create and associate their own subjective and intersubjective meanings as they interact with the world around them. The aim of critical studies is to critically look at the status quo, through exposure of what are believed to be deep seated, structural contradictions within social systems, thereby transforming these alienating and restrictive social
25
conditions. Due to the nature of the present study the critical approach is not given any further consideration. 2.3.1 The case study method The qualitative data-collection methods comprise a broad variety of datatypes and techniques. The data types include written documentation, archival records, direct observation, interviews, or physical artifacts (Yin, 1994; Eisenhardt, 1989). The characteristic of qualitative research is that it is conducted through an intense and/ or prolonged contact with a field or life situation. These situations are usually reflective of the everyday life of individuals, groups, societies, or organizations (Miles and Huberman, 1994). The special feature of qualitative research methods is the depth of study compared to the breadth of study, which is one of the hallmarks of quantitative methods. Qualitative data-collection is often used in case studies. Case studies cover exploration of complex, social, empirical phenomena – either contemporary or historical – by applying multiple sources of data (Andersen, 1992). This richness of data challenges the important point in scientific research - the generalizability of achieved results to other similar settings. The case study is a research strategy, which focuses on understanding the dynamics present within single settings (Eisenhardt, 1989). The case study can be designed as action research (Argyris and Schon, 1991). In action research the researcher participate for example in systems development (e.g., Kautz et al., 2001). Another way of performing case studies is through grounded theory where the constant comparative method is applied, by examining ‘incidents’ recorded in one’s data (Glaser and Strauss, 1967). By using the grounded theory method the researcher develops the theory further4 based on observations in the field (e.g., Crook and Kumar, 1998).
4
In their early work Glasser and Strauss (1967) argue that data collection is based on theoretical sampling and that the point of theoretical sampling is to select sources of new data in ways that permit further development of theory.
26
Yin (1994), author of one of the most frequently cited books on case studies, has given explicit directions on how to perform a case study. Yin has, however, not defined what a case study is. Ragin and Becker (1992) concluded that there are several possible interpretations of the definition of a case study. In the name of simplicity a straightforward definition of a case study, taken from MIS research, is in this context used as guidance: “A case study examines a phenomenon in its natural setting, employing multiple methods of data collection to gather information from one or a few entities (people, groups, or organizations)” (Benbasat et al., 1987), p. 370.
One of the major critiques of data obtained through case studies is lack of rigor in relation to collection of and evaluation of data. Yin (1994) has recommended that a set of rules be applied while gathering qualitative data. One critique of the framework defined by Yin is that it to a large extend resembles the positivistic ideals (Klein and Myers, 1999) and hence does not fulfil the requirements of social science research, which often accumulates rich data, which might be interpreted subjectively. Lee (1989) is among those who has accepted the positivistic paradigm in case studies and tried to apply it to case research using the same positivistic measures for evaluating data. Though Lee in a convincing manner succeeds in evaluating qualitative data according to the rules of positivism, or natural science as Lee denotes it, it can be argued that for research related to social science phenomena it might be more appropriate to look for methods of evaluating qualitative data outside the realm of the tradition of natural science research. The above-mentioned article by Lee (1989) is one of two influential articles from the late 1980s on the use of case study as a research method within MIS. The two articles present some of the major challenges that confront case research namely: 1) The problem of identifying when to apply the case study method (Benbasat et al., 1987), and 2) The inherent weaknesses of the case study approach (Lee 1989). Yin (1994) has formulated some general directions on when to favor the case method and when to use other methods. Yin recommends the case study approach when the research focuses on answering “how or when questions” whereas the survey method 27
is more suitable for research driven by “who-, what-, and where-questions”. Benbasat et al. challenge the Yin model by arguing that one of the major reasons for the suitability of case research within IS is that the IS technology is relatively new and that interest has shifted to organizational rather than technical issues. The deviation from Yin (1994) in relation to the application of the case method is thus not a result of the research question per se but a consequence of the realm of inquiry. The use of case studies is therefore suitable for exploratory rather than explanatory studies. Lee maintained that case studies were as valuable as the natural science approach, which according to the positivist school of thought is superior to any other method with respect to the power of explanation, prediction, and control (Lee, 1991). In his 1989-article Lee argued that the four principles, controlled observations, controlled deductions, replicability, and generalizability, which are the hallmarks of natural science research, can be applied to IS research as well. The natural science research model can thus be applied. The question is, however, is it appropriate? For the author the answer to this question is “No!” The case study method has been applied in several IS and EDI studies. Some researchers have used a single case to analyze the impact of for example EDI (Teo et al., 1997) and the business value of EDI (Chatfield and Bjorn-Andersen, 1997; Jelassi and Figon, 1994). Other studies have made cross case studies using two or more cases. Riggins and Mukhopadhyay (1994) performed two case studies to illustrate the obstacles for initiators with respect to benefits due to limited implementation of EDI in the adopters’ organizations. Iacovou et al., (1995) studied factors that influence the EDI adoption practices of small firms through seven case studies. Finally, Chatfield and Yetton (2000) used three cases to illustrate the large variation in strategic pay-off derived from strategic EDI initiatives. Some of the studies (Jelassi and Figon, 1994; Mukhopadhyay et al., 1995) aimed at providing generalizations to similar cases whereas others described unique situations (Teo, et al., 1997) which are perhaps more informative than generalizable.
28
The purpose of this study is to determine the strength of the three contexts, organizational, environmental, and technological as explanatory variables for motivation for adoption or non-adoption of EDI. The study is therefore based on an examination of the relevance of using existing models for understanding adoption and non-adoption of EDI. Based on this analysis further development of the existing models is suggested. In the present case study of the TDP the ambition is to get an understanding of the mechanisms that influence adoption of EDI in a group of related businesses. The method for obtaining data in the case study has been exploratory rather than theoretical excluding the method of grounded theory. The authors’ active participation in the project was limited to a single meeting (see Textbox 4-1, page 103). The participation took place at a point in time where most decisions had already been taken and the participants had accepted their roles in the project, leaving little or no room for external interventions. This excludes the action research approach. Instead the strategy of triangulation was chosen. One of the weaknesses related to conducting a research project that has an offset in an exploratory search is that the theoretical guidance is absent, both in relation to the practical data-collection and in relation to focus. The “research method coin” is two sided. One side of the coin is the method of collecting data in the field. The other side is related to the basic belief of the researcher, her or his epistemological viewpoint. Method and methodology are closely interrelated though not often explicitly mentioned when research is published (Orlikowski and Baroudi, 1991). In the following the methodological lens through which the TDP is observed will be presented. 2.3.2 Methodological considerations The issue of how to perceive objects and understand cause and consequence of interplay among objects and social entities has puzzled philosophers and scientists for centuries (see for example Descartes,
29
1968).5 The objective of this section is not to give an overview or introduction to the philosophy of science but rather to argue in favor of the choice of methodological approach applied in the TDP case study. According to the positive-inductivism and the hypothetico-deductive school of logic, scientific knowledge is reliable knowledge since it is objectively proven from data resulting from direct observations and experiments (Chalmers, 1982). It is believed that if a high number of n’s are having the characteristic “X” then it can be generalized that this will be the case for all n’s. At least until the theory is proven to be wrong (Popper, 1972). In case of massive falsification and proof of alternative explanations a scientific revolution can take place (Kuhn, 1996). This view of scientific inquiry has served the natural sciences well. For social sciences it seems to be a less practicable strategy. For the social sciences one of the major problems is that the multiplicity of causal factors increase the complexity of data and the number of explanatory factors. It is in practice impossible to include a social unit, e.g. an organization, in a controlled experiment and thus isolate irrelevant or disturbing elements. There are at least two aspects in relation to the controlled experiment in the context of social science research. Firstly, it is impossible to place an organization in a vacuum and it can be difficult or impossible to decide which elements are irrelevant and disturbing. Secondly, human beings with individual preferences and their own ideas populate the universe of the organization. They are part and parcel of the organizational structures and they influence these structures. The uniqueness of each n makes it difficult to get a sufficiently large number of n’s to make generalizations in the sense of natural science,6 where causal relationships are derived from a sample to a population. An example of the obstacles social science research face in relation to similar cases is illustrated in Barley’s study of CT scanners (Barley, 1986). Barley studied the same innovation (the CT
5
Descartes lived from 1596 to 1650. The reference from 1968 is a reprint of his original works. 6 Though Lee (1989) argues the opposite, at least in relation to the strength of the model that is tested.
30
scanner), in the same situation (adoption and implementation), among the same personnel (radiologists and technical personnel) in two hospitals in the same geographical site (Massachusetts). However, he ended up with two different outcomes due to different behaviors among the humans that populated the sites that were subject to analysis.7 The list of challenges for social sciences in relation to the natural science model is long and of limited relevance in this context. Examples are problems with bias due to richness of data sources, problems with reproducibility due to the ever-changing environment etc. One way out of the dilemma of how to tie different types of data with a specific set of rules is to consider the researchers epistemological viewpoint. Walsham (1993) argues that: “… epistemology, the basis of one’s claims to knowledge, and research methods are interrelated. If one adopts a positivist epistemological stance, then statistical generalizability is the key goal. However, from an interpretive position, the validity of an extrapolation from an individual case or cases depends not on the representativeness of such cases in a statistical sense, but on the plausibility and cogency of the logical reasoning used in describing the results from the cases, and in drawing conclusions from them.” (p. 15)
According to Walsham the interpretive position is a means for understanding organizational issues related to computer-based information systems. It is: “… broadly interpretive methods of research, aimed at producing an understanding of the context of the information systems, and the process whereby the information system influences and is influenced in its context.” Walsham, 1993, p. 4. (Emphasis in original).
The structuration approach is an important underlying model in the interpretive position as presented by Walsham. One of the key ideas of structuration theory is the duality of action and structure (Giddens, 1984). Giddens argues that structure and agency should not be viewed as independent and conflicting elements but as a mutually interacting duality. 7
Barley used the study as evidence for the interplay between actors and structures as described in structuration theory (Giddens, 1984).
31
Thus social structure is viewed as being created by human agents through their actions, while the actions of humans in social contexts serve to produce, and reproduce, the social structure. The perspective of structuration theory is thus that organizational change is the joint effect of the actions of individuals interacting with institutional structures like business strategies, communication vehicles, and information systems (Pozzebon and Pinsonneault, 2001). The theory of structuration recognizes that human actions are enabled and constrained by structures, yet that these structures are the result of previous actions (Orlikowski, 1992). This dialectic interpretation of organizational change is intriguing though difficult to evaluate empirically. It is argued that Giddens has stated in several of his works that, “... structuration is not intended as a concrete research programme and that the principles are essentially procedural and do not supply concepts useful for the actual prosecution of research” (Jones, 1999). The structuration approach has none the less been applied by several IS researchers e.g. (Barley, 1986; Orlikowski, 1992; Pozzebon and Pinsonneault, 2001). Another approach to interpretivism is the hermeneutic view. Hermeneutics is traditionally an approach concerned with interpreting the meaning of texts (Knudsen, 1994; Winograd and Flores, 1987). The word hermeneutics is derived from the Greek word hermeneun, which means, “to understand”. The hermeneutic method has been applied since the Middle Ages where it for example was used for studying and interpreting the Bible. Later historians applied the hermeneutic approach. At present the social sciences including economics (Knudsen, 1994) have adopted the hermeneutic approach. The main objective of the hermeneutic methodology is to develop empathy and understanding, which leads to identification of the subjective meaning of the behavior and actions of persons (ibid.). Gadamer (1976) states that any individual is continually involved in activities of interpretation in order to understand his or her world. The hermeneutic research approach aims at reflecting this continuous involvement, which leads to an understanding of the realm of inquiry.
32
Hermeneutics has been on the IS research agenda for a number of years. At the ICIS conference in 1991 one panel was dedicated to hermeneutics. Recently Klein and Myers (1999) have formulated a set of principles for evaluating interpretive field studies in IS based on the hermeneutic principles. They outlined seven principles which can help IS researchers to understand human thought and action in social and organizational contexts. Klein and Meyers define research as interpretive if: “…it is assumed that our knowledge of reality is gained only through social constructions such as language, consciousness, shared meanings, documents, tools, and other artifacts. Interpretive research does not predefine dependent and independent variables, but focuses on the complexity of human sense making as the situation emerges.” p. 69.
Based on these characteristics of interpretive research and hermeneutics seven principles for interpretive field research in IS are outlined. The foundation for the seven principles is the hermeneutic circle (Gadamer, 1976). The hermeneutic circle implies that the meaning of an individual text contextually depends on the moment of interpretation and the horizon brought in by the interpreter (Winograd and Flores, 1987). Klein and Myers state that the seven guidelines are principles not a set of bureaucratic mandatory rules. The primary reason for choosing the seven principles as interpretive framework of the present case study is the nature of the data available and the objective of the study. The data for the present case study does not invite normal hypothesis testing8 or drawing inferences to the population based on a representative sample. Nor is it the objective, contrary to critical research, to eliminate the causes of unwarranted alienation and domination thereby enhancing the opportunities for realizing human or social capital. Most of the data available was related to written documentation such as letters, minutes of meetings, agendas, and internal working documents. By choosing a methodology originally dedicated to the understanding of texts
8
A set of propositions was derived from observations in the case study (cf. Chapter 8). These propositions were tested on data gathered in accordance with the principles of positivistic research.
33
it is found that the hermeneutic interpretive approach is most appropriate in relation to the assessment of the qualitative data obtained in the study. Table 2-1. Seven principles for interpretive field research 1) The fundamental principle for the hermeneutic circle. This principle suggests that all human understanding is achieved by iterating between considering the interdependent meaning of parts and the whole that they form. This principle of human understanding is fundamental to all the other principles. 2) The principle of contextualization. This principle requires critical reflection on the social and historical background of the research setting, so that the intended audience can see how the current situation under investigation emerged. 3) The principle of interaction between the researchers and the subjects. This principle requires critical reflection on how the research materials (or “data”) were socially constructed through the interaction between the researchers and participants. 4) The principle of abstraction and generalization. This principle requires relating to the ideographic details revealed by the data interpretation through the application of principles one and two to theoretical, general concepts that describe the nature of human understanding and social action. 5) The principle of dialogical reasoning. This principle requires sensitivity to possible contradictions between the theoretical preconceptions guiding the research design and actual findings (“the story which the data tell”) with subsequent cycles of revision. 6) The principle of multiple interpretations. This principle requires sensitivity to possible differences in interpretations among the participants as are typically expressed in multiple narratives or stories of the same sequence of events under study. Similar to multiple witness accounts even if all tells it as they saw it. 7) The principle of suspicion. This principle requires sensitivity to possible “biases” and systematic “distortions” in the narratives collected from the participants. Source: Klein and Myers, 1999
If methodological frameworks are viewed as meta-theory then they serve as a tool for reflection and analysis of the criteria, which influence the choices the researcher makes in relation to acceptance or rejection of statements and theories (Knudsen, 1994). The inclusion of a methodological framework in a study therefore serves to explicitly present accepted assumptions. In the actual case study the hermeneutic stance as outlined by Klein and Myers provides the criteria for accepting or rejecting statements which make sense in relation to the interpretation of data. The underlying
34
assumption in hermeneutics is a conviction or belief that it is possible to understand the realm of inquiry from insights into written documentation, observation of situations, and conversations/ interviews with people involved in a given project. Following this argumentation on the significance of methodological frameworks in research projects the interpretive dimension is related more to an analysis of which elements guided the researcher to her conclusions than to analysis of data. It is at this point the seven principles outlined by Klein and Myers prove their value. By advising the researcher to observe key elements in the hermeneutic view the researcher is forced to consider how she came to understand the realm of inquiry. The analysis of data is on the other hand made by applying relevant theory concerning the particular realm of inquiry and the explicit research question(s). The particular case research was concerned with adoption and diffusion issues and appropriate theory from this research tradition was therefore used in the analysis. However, methodological framework influences the way data is analyzed. The practical implication of the use of the Klein and Myers principles is that context is given a high priority. The fundamental principle of the hermeneutic circle implies that the researcher moves from texts or other sources of information to the world and then back to the available information. The move between the parts and the whole facilitates an interpretation of the situation. After this discussion of research methods and methodologies the datacollection of qualitative and quantitative data is presented in detail.
2.4 Collection of qualitative data Step one “observation of field of inquiry and study of historical data” (cf. Figure 2-1, page 18) was the least structured process. Direct observation included participating in formal and informal meetings in the professional business associations, visiting organizations, and attending a work meeting in the TDP. Minutes of meetings distributed to participants were prepared 35
from formal meetings in the professional business associations. Additionally, notes for personal use were taken. Data from informal meetings such as seminars and workshops arranged by the Danish EDI Council and the business associations was not filed in a structured way. After each visit the log of official statements and personal impressions was updated. Notes were taken during the TDP work meeting. Afterwards those notes were written up in form of a narrative (cf. Textbox 4-1, page 103). The study of archival data was based on meeting agendas and minutes of meetings, data accessible at public sources such as the Internet, letters, and internal working documents. The meeting agendas and minutes of meetings were made available via an extranet established for the participants of the TDP. The project initiators provided a password to the extranet. This made it possible continuously to keep track of the developments in the project reported to the project site on the extranet. Letters and internal working documents were made available by the project initiators. The written documentation was as far as possible examined in chronological order. In order to gain the best possible understanding, all vague statements in the written documentation were discussed with the project initiators both during the formal meetings and by phone. Apart from that the project was discussed with the TDP participants whenever possible. This resulted in establishing a close interaction both with the project initiators and with the participants. During August and September 1999 representatives from eight of the nine companies participating in the TDP were interviewed. One company declined to participate in the interview. The reason being given was that the participants from that company found that the project and especially the outcome of the project had been unsatisfactorily. The interviews lasted a minimum of one hour. The longest interview lasted two hours. The interviews were taped and afterwards transcribed by an independent third party. After the first transcription the interviewer listened to the tapes checking for possible errors in the transcription. This led to minor retranscriptions of portions of some interviews. Thereafter the transcripts were returned to the informants for their approval.
36
The main participant from each of the eight companies was interviewed. A common characteristic of the informants was that they had played an active role in the TDP in the sense that they had participated throughout the project.9 There was variation in the organizational status among the responders. Six responders were in charge of accounting departments and two were responsible for IT-departments. Whether the informant came from an accounting department or from an IT department was solely a consequence of which person had been designated as being the key person representing the company in the project. About one week in advance of the planned interview the responders were via a postal letter informed of the four main questions that were to be discussed during the interview: - Does your company make use of EDI? - Does the company use or intend to use the software developed during the project? - Does the company find that the EDIFACT subset developed during the project is suitable for the company? - What is the company’s general benefit from participating in the project? The theoretical motivation for the first question was driven by the measures defined by Massetti and Zmud (1996) in relation to the adopters of EDI. According to Massetti and Zmud (ibid.) usage can be addressed in four dimensions.10 The relevance of this question in relation to the non-adopters should be seen as a starting point for supplementary questions concerning incentives and barriers for EDI adoption. The first question was hence not directly related to the TDP but rather focused on a general discussion of EDI adoption or non-adoption in the organization.
9
Some of the informants had the role of an innovation champion in the organization (Grover, 1993; Schon, 1963). Others just represented the company in the project, but they had no special interests in the IT policy of their organization or they had no influence on the overall IT policy in their organization. 10 For a description and clarification of the four dimensions see Table 4-5 (page 126).
37
The next three questions were directly related to the TDP. From a theoretical point of view they were related to the considerations presented in Chapter 3 concerning the effects of creating optimal conditions for adoption and diffusion of an IOS. Policy issues motivated the questions. Bearing in mind that it had already become apparent in the interaction with the business associations and the participants that the project had not been a solid success it was found relevant to investigate the defined objectives. This could lead to a better understanding of the tensions and obstacles the participants had experienced during the TDP. From a practical point of view the tree questions reflected the objectives of the TDP (cf. Table 4.1, page 95) in the sense that they dealt with the software developed during the project, the subset developed during the project, and finally the creation of awareness. The responders knew the interviewer from prior meetings in their TDP work groups and informal arrangements organized by the project initiators and all were aware that the organizers of the TDP sponsored the interviewer. The TDP participants were also aware that the interviewer met and discussed the project with the project initiators. The project initiators and the interviewer decided prior to the interviews that the responders would remain anonymous in the reporting of findings. 2.4.1 Sources of biases in interviews Independent of the method of data-collection there are a number of sources of biases,11 which have to be considered. During personal interviews some of these biases are related to the relationship and interaction between the interviewer and the informant while others are related to personal and institutional factors which influence the responder. Andersen (1990) lists a number of such bias-producing situations: - The responder misunderstands the questions formulated by the interviewer.
11
Bias here refers to a systematic distortion of data opposed to a random distortion of a measure as a result of a sampling procedure (Osterlind, 1983).
38
- The responder has forgotten the facts or she remembers the situation differently. - The responder has a defensive attitude. - The responder gives the answer that she expects the interviewer wants. - The responder wants to impress the interviewer. - The responder prefers to answer positively rather than negatively. There is no reason to assume that the different sources of bias listed above have been avoided in the present study. For clarification a few comments are given to each of the listed biases in order to clarify the author’s interpretation of the situation. One way of reducing biases related to misunderstandings is to apply a semi-structured interview guide, which leaves room for more informal talk and discussion of the topic under investigation. The four main questions, which were mentioned above, were closely related to the major objectives of the TDP. All informants answered the questions in a way that made sense. The second source of bias is more critical. The description of the meetings in the project leaves the impression that a number of opposing interests were at play. Facts were thus ambiguous. Additionally, as depicted in Figure 4-3 (page 101) at least four different types of activity and communication were taking place at the same time. The triangulation of different sources of information is therefore used for uncovering possible problem areas. The author did not observe the third source of bias, the defensive attitude of the responder. The participants were generally eager to convey their interpretation of the situation. Two of the informants told the author that they, when they received the four questions prior to the interview, had been somewhat worried. Therefore, they had contacted one of the project initiators to make sure that they without undue consequences could narrate
39
the story as they saw it.12 13 The fourth source of bias is probably the most serious in this particular case. The researcher was not looked upon as being independent since the participants had experienced her interaction with the project initiators. As described in Textbox 4-1 (page 103) the author was introduced to the participants by the project initiators and her presence was defended by one of the project initiators. The participants were not quite sure about the position of the author. However, as mentioned above the informants were eager to narrate their interpretation of the project. However, some of the informants were somewhat hesitant about voicing their opinions in strong terms. The fifth source of bias in relation to collection of data is related to the situation where the informant wants to impress the interviewer. The first question that was discussed with the informants was related to whether or not the organization had adopted EDI. It can not be refuted that the adopters simplified the adoption and implementation process, nor can it be denied that the degree of usage was presented in a slightly more attractive light. It is however the impression that possible exaggerations related to adoption were of minor importance regarding motivators for EDI adoption among the TDP participants. The sixth source of bias, the responders’ preference for positive rather than negative answers, was not observed (cf. Table 4-7, page 137, which summarizes the outcome of the TDP). The above-mentioned reflections on biases are no guarantees against biases nor do they guarantee the validity of data, but they help and guide the researcher during the process of collecting and describing data. They can
12
The project initiator told me about this incident prior to the interview. He told me that he had encouraged the informants to give their interpretation – positive or negative - of the project and he had promised that critique would have no negative consequences. Additionally, the project initiator emphasized that the informants were secured anonymity. (This anonymity can of course be questioned since only eight companies with individual characteristics were represented). 13 The situation illustrates a conflict between loyalty towards the business association, which had initiated and sponsored the project on one side and the frustration related to the achievements of the project on the other side. This conflict was revealed in all interviews though the loyalty towards the business association varied amongst the different companies.
40
also, as previously mentioned, help the researcher to clarify her or his interpretation of the situation.
2.5 Quantitative data-collection 2.5.1 Introduction Gathering of quantitative data is usually related to survey research. Survey research can be classified as field studies with a quantitative orientation (Kerlinger and Howard, 2000). The survey research involves, “... gathering information for scientific purposes from a sample of a population using standardized instruments or protocols” (Kraemer, 1991). The quantitative survey method has three distinct characteristics (Pinsonneault and Kraemer, 1993): Firstly, to produce quantitative descriptions of some aspects of the studied population either leading to uncovering of relationships between variables or projecting findings descriptively to a predefined population. Secondly, it is a method based on structured and predefined questions. Thirdly, information is generally collected from a fraction, a sample, of the study population; however, it is collected in such a way that results from the sample hopefully can be generalized to the study population. Ten years ago Kraemer and Dutton stated that, “Survey research is both the most widely used and most widely questioned method in the MIS field (Kraemer and Dutton, 1991). Orlikowski and Baroudi (1991) reported that 49 percent of their research analysis sample were surveys, whereas the share of case studies was 14 percent. In a recent analysis of research in IS (Claver et al., 2000) these figures remained almost unchanged. The ratio of survey research and case studies were in 1996/ 97 respectively 52 percent and 18 percent. Comparison of these two studies should be done with due care to the different sources of information and different definitions applied. However, these studies can be used as an indicator of the distribution of research methods applied in IS research. Although the survey research method is often used it is somewhat problematic. A number of disadvantages related to the survey research method have been reported (Kerlinger and Howard, 2000). These include: 41
1) The scope of information is often emphasized at the expense of depth. The reason being that survey information does not normally penetrate very deeply below the surface. 2) Another weakness is that the survey situation can temporarily lift the respondent out of his or her own social context, which may invalidate the results of the survey. In relation to organizational studies problems particularly arise in relation to survey research. The first problem is related to selection of the respondent. Maaloee (1996) calls attention to the problem of finding a single person in the organization that can give a reliable answer to the broad variety of questions often included in the survey instrument. Another aspect is that answering questionnaires may not have a high priority in the organization (ibid.). The questionnaire is often sent to the managing director. She or he may however, pass the questionnaire on to a subordinate who may not be able to provide the correct answer to the questions that were targeted to a specific person in a specific job function. The answers may therefore be invalid. These abovementioned shortcomings are some of the drawbacks of survey research. 2.5.2 Design of the study survey instrument In order to get a more comprehensive and broader view of the factors that determine adoption of EDI and B2B e-commerce (business-to-business electronic commerce)14 it was decided to make a survey of the Danish steel and machinery sector. Due to the unique qualities of each company it was also decided not to make a sample from the entire population of the steel and machinery sector organized in the two industry and trade associations. However, the total number of companies in the steel and machinery sector also played a considerable role. Finally, it was found of great importance to include both users and non-users of EDI and B2B e-commerce in the survey. Since the business associations had no knowledge about the distribution of the two groups it was found appropriate to include the entire sector.
14
The motivation for including e-commerce was that in 1999, when the questionnaire was developed, e-commerce was considered to be an inevitable innovation in the business community. It was at the same time found that it would be wasteful not to include e-commerce in a questionnaire for a sector where little was known about usage of EDI and e-commerce.
42
The items for the survey were defined by a group of key policy makers from the two industry and trade associations in collaboration with the author. The policy makers were familiar with EDI and B2B e-commerce from their involvement with the national action plans (cf. Section 3.4) and from interacting with EDI and e-commerce users in their respective member firms. By involving practitioners in the development of the questionnaire the research items tended to be exploratory and practicedriven rather than theory-driven due to the practical approach to this research topic (Zmud, 1998). After prolonged discussions during a number of meetings a questionnaire was developed and made ready for pilot testing among a number of randomly selected firms in the steel and machinery industry. Before mailing the final questionnaire the design and content was discussed with an expert in development of questionnaires.15 These discussions led to several adjustments of the visual appearance of the questionnaire and to re-formulations and adjustments to the layout of a number of items/ questions. Like other similar studies (Johnston and Gregor, 2000) it was found necessary to analyze the actions of the individual firm even though the business sector was the focus of the analysis. Furthermore, even at this level it was sensible to choose a specific responder among the individuals in the organization in order to be able to interpret and explain activities at the specific industrial level. Due to the position of the two professional business associations the responders chosen were management representatives. By including both adopters and non-adopters among the responders it was expected that a more multifaceted picture of motivators for adoption could be revealed. The strategy of including both adopters and non-adopters have been applied in previous IS research e.g. (Lai and Guynes, 1997; Crum et al., 1996; Cox and Ghoneim, 1996) the reason being, that it can be argued, that it is as appropriate to consider rejection of an innovation as a form of action similar to adoption (Abrahamson, 1996).
15
Debbie Dunckle at CRITO, University of California at Irvine kindly shared her expertise with development of questionnaires and gave valuable advice.
43
The questionnaire was divided into five parts. For a thorough presentation of the questionnaire see Appendix A (page 368) where the questionnaire is printed. The first part consisted of a set of general questions such as ownership and core product of the company. Section two was aimed at companies that had not adopted e-commerce or EDI. The non-adopters were asked about questions such as number of customers and suppliers, use of ERP-systems, and considerations in relation to B2B e-commerce and EDI. The non-adopters were also asked whether they were planning to adopt B2B e-commerce or EDI. The third section was targeted to the companies that had adopted B2B e-commerce. Here the potential responders were asked to answer questions on how e-commerce was used in the organization, the incentives for adopting e-commerce, years of experience with e-commerce and the perceived benefits of the use of ecommerce. The forth section was aimed at the EDI adopters. The EDI adopters were asked similar questions to those asked in section three. A few items were added concerning the number of business partners with whom EDI messages are exchanged. Finally, a closing section was dedicated to general questions to all responders. Here questions included electronic document exchange performance in general and the possible interest in obtaining further information about e-commerce and EDI. The questionnaire consisted of 34 main questions and a total 144 items including sub-questions. The business associations provided information on the responders’ geographical location and number of employees from their member databases. These items were later added to the database after dataentry of the questionnaires. Most questions were close-type questions and those questions relating to opinion/ perceptual matters were presented on seven-point Likert-type scales. As recommended in previous research (Grover, 1993) definitions of EDI and e-commerce were included in the questionnaire to ensure that the responders had a common understanding of the phenomena under investigation. Those definitions closely resembled the definitions presented in Sections 3.4.2 and 3.4.3, which had been exposed to the responders through media or business press.
44
In the following sections the data-collection procedure and response rate are detailed. Thereafter, an analysis of non-response bias and representativeness are presented. These sections are followed by a critique of the data-collection procedure. Finally, considerations in relation to the choice of statistical methods are presented. 2.5.3 Collection of data The method selected for the collection of survey data was a postal questionnaire. A pilot survey was conducted by randomly choosing 15 companies from the steel and machinery sector. The nine responses that resulted from this pilot survey led to some changes and improvements of the questionnaire based on the opinions expressed by responders of the pilot survey. The two industry and trade associations represent the management of the manufactures and wholesalers in the steel and machinery industry. The questionnaire was as a natural consequence mailed to the managerial unit in the organizations involved. The package sent to responders contained three items: One covering letter, one questionnaire, and a prepaid reply envelope. The covering letter explained the purpose of the survey and asked the responder to return the completed questionnaire within three weeks in the prepaid reply envelope or to be returned by fax within the same timeframe. The responders were guaranteed confidentiality and anonymity. The package was mailed by the two industry and trade associations and the cover letter carried the letterhead of the respective association. It was expected that the questionnaire using this method would be perceived as more authoritative than usual questionnaires received in organizations, and that this procedure accordingly would result in a better response-rate. Apart from sending the package the associations were also responsible for receiving the completed questionnaires from the responders. Upon receipt of the completed questionnaires the company that had responded was ticked off against a mailing list and the date of receipt was recorded. It would therefore be possible to differentiate and compare early and late responders within this single mailing.
45
2.5.4 Validation of the coded data The returned questionnaires were coded in a data-entry system developed using EPI-INFO 6.0,16 a shareware product. The strategy for data entry focused on three major objectives and aimed at achieving the highest possible data quality. Firstly, to reduce typing errors all data entry values were during data-entry chosen from pre-defined menus. Secondly, predefined logical structures in the questionnaire automatically coded data and made sure that the typist jumped to the proper fields in the questionnaire. Thirdly, logical tests were performed during data-entry to make sure that the data entered logically made sense. “Most truly revolutionary results from data analyses are based on data entry errors.” (Cody and Smith, 1997). To avoid this situation careful control procedures were applied to test for data-entry errors. After coding the questionnaires the coded values were validated using a special module in EPI-INFO. This module allows for re-entry of data and it provides an automatic check of previous entered data with the new entries. All the questionnaires were kept in six large binders. The following data validating procedure was used. All forty-eight questionnaires in binder number one were validated one hundred percent using the above-mentioned method. Since the percentage of typing errors in binder number one based on single characters were less than 0.1 percent every fifth questionnaire in binders number two, three, and four were validated. Since the error-rate was about the same as in binder number one every tenth questionnaire in binders number five and six were validated. Based on validation of data in binders number five and six it was estimated that the error-rate was similar to the error rate in binder number one. The chosen strategy for data-entry thus resulted in a satisfactory data quality. Finally, one extra check was performed. 25 questionnaires where controlled in detail by two persons. One person was reading the coded values from a printout from the dataentry system while another person was checking the read-out value against 16
EpiInfo version 6.0 is developed by the staff at The Division of Surveillance and Epidimiology Program Office at Centers for Disease Control and Prevention, Atlanta, Georgia, USA.
46
the questionnaire.17 The error-level was negligible. By performing this validation procedure it was found that the coded data reflected the completed questionnaires in a most satisfactory way. 2.5.5 Response rate The questionnaire was sent to the management of 917 manufactures and wholesalers in the steel and machinery sector in Denmark. It is part of the internal policy of the two involved associations not to burden their members with too much mail. Therefore, a second mailing was not allowed. A total of 252 responses were received, out of which 247 were retained for the purposes of the study. Out of the five non-useable questionnaires one was completely blank, one was a duplicate first received as a fax and later as the printed questionnaire originally mailed to the responder. The last three questionnaires that were excluded from analysis were only sporadically filled out and could not be included in the sample. The 247 valid returned questionnaires equal a response rate of 27.4 percent. The response rate is approximately at the same level as similar studies related to the adoption of IOS. For example 25.7 percent (Chau, 2001) and 27 percent (Masters et al., 1992). Other studies focusing on both adopters and non-adopters of information systems have generally reached response rates that are somewhat lower. For example 18.4 percent (Lai and Guynes, 1997), 19 percent (Crum et al., 1996), whereas Cox and Ghoneim (1996) obtained a response rate of 28 percent. Compared to other studies, which have included both adopters and non-adopters, the response rate is found to be satisfactory, especially when the collection procedure with only a single mailing is taken into consideration. There are 665 manufactures and 252 wholesalers out of the 917 members of the two major business associations within the steel and machinery industry. The response rate was somewhat higher among the wholesalers than among the manufactures. One possible and very likely explanation is that due to a technical irregularity in the numbering of the questions in the questionnaires, some 17
Thanks to kind assistance from my husband it was possible to perform this demanding but intellectual little inspiring task.
47
of the manufactures received a questionnaire where the numbering of the sequence of the questions were more or less random. Since there were a number of conditional jumps in the structure of the questionnaire the responders that had received the irregular questionnaires could not follow the conditional jumps indicated in the instructions. This mishap resulted from combining two Microsoft Word functions: Auto-numbering and Mail merge. Due to a high degree of automatisation of mail handling in the business association this numbering error was, unfortunately not found until the questionnaires were returned. At this point in time it was too late to correct this mishap. Approximately 60 percent of the returned questionnaires from the manufactures had errors in the numbering sequence of the questions. After comparing the irregular questionnaires with the regular questionnaires from the manufactures, it was decided to keep the irregular questionnaires for further analysis, since there was not found to be any statistical significant differences between the regular and the irregular questionnaires with regard to the opinion data items included in this research.
Table 2-2. Response details Questionnaires mailed out Questionnaires received Unusable questionnaires Useable questionnaires Response rate in percent
Manufactures Wholesalers 665 252 174 78 4 1 170 77 26.2 30.6
2.5.6 Distribution of adopters and non-adopters Of the 247 valid questionnaires, 77 companies had adopted e-commerce and/ or EDI. The 3:1 ratio of non-adopters to adopters is unusual compared to similar studies, which include EDI adopters and non-adopters. Crum et al. (1996) reported data from 83 adopters and 98 non-adopters in their survey and Cox and Ghoneim (1996) reported data from 75 users and 10 non-users. One likely explanation could be non-response bias, which will be addressed below. Another more likely explanation is national differences. The two above-mentioned surveys were performed in the US
48
among large firms, which generally have a larger ratio of adopters/ nonadopters. In the present survey 16 percent18 of the participants had adopted EDI (cf. Figure 2-2). This number matches statistics from a contemporary survey performed by Statistics Denmark (Ministry of Information Technology and Research 2001), where it was found that 15 percent of all Danish businesses had adopted EDI. It is therefore likely that national differences and the size of the companies included in the survey can explain the different ratios of adopters/ non-adopters found in previous studies compared to this study. Figure 2-2. Distribution of percentages of adopters, non-adopters, and planners
12% Adopters of EDI
15%
Adopters of e-commerce
4%
Adopters of both EDI and e-commerce Planners
47%
Non-adopters
22%
The adopters are divided into three distinct groups: Adopters of EDI (12 percent), adopters of e-commerce (15 percent), and adopters of both EDI and e-commerce (4 percent). The non-adopters were divided into two categories. Those that plan to adopt EDI and/ or e-commerce in the near future and those that so far have rejected the idea of adoption. Twenty-two percent of the responders are planners and 47 percent of the responders are non-adopters. Compared to Lai and Guynes (1997) a different strategy was 18
The responders that indicated that they have adopted both e-commerce and EDI are included in the sixteen percent (12% + 4%).
49
followed in the categorization. Lai and Guynes classified the responders in relation to the adoption-decision into three groups: Adopters, non-adopters, and “unknowns”. Lai and Guynes treated the responders that answered “unknown” as being organizations still considering adoption of the innovation. They defined non-adopters as those responders that had evaluated but rejected adoption of the technology. Grover (1993) divided the responders into three groups: Non-adopters, companies that had decided to adopt but had not yet implemented the innovation, and companies that currently use the innovation. The two last groups were considered as adopters since it was found that these two groups did not indicate significant differences. The non-adopters were subdivided into three groups: 1) No action or consideration, 2) Some discussion but the idea of implementation was rejected, and 3) Some consideration but no decision taken. No significant differences in responses were found in the three subgroups of non-adopters. The study of the Danish steel and machinery industry applies a categorization which resembles Grover’s (1993). 2.5.7 Definition of adopters, planners, and non-adopters in the present study As mentioned in Section 2.5.5 there were 247 questionnaires available for further analysis. Out of these 247 questionnaires 209 questionnaires were related to EDI. These 209 cases are referred to as “the EDI sample” and they form the basis for the quantitative analysis of motivators for EDI adoption in the Danish steel and machinery industry. The EDI sample consisting of 209 responders was divided into three categories: Adopters, planners, and non-adopters. Among the 209 responders there were respectively 115 non-adopters (55.0 %), 55 planners (18.7 %), and 39 adopters (26.3%). The responders that indicated that they used EDI or both EDI and e-commerce were defined as adopters. Planners were those responders that indicated that they considered adoption of EDI. The last group, the non-adopters, was those that neither indicated that they used EDI nor considered adoption of EDI. In those cases where the
50
questions related to considering adoption of EDI were left completely unanswered the responders were defined as non-adopters.19 The non-adopters are considered to be potential adopters. This group is still considering adoption whereas the planners are viewed, as a group of responders that have made a decision to adopt, but have not yet made the physical arrangements necessary for adoption. The division between the two categories, planners and non-adopters is a mental construct (Rogers, 1995) and what distinguishes planners from adopters is having or not having the innovation (Tornatzky and Fleischer, 1990). Since it is not certain that a change in actual behavior will manifest (Harrison et al., 1997; Kerlinger and Howard, 2000) it can be argued whether or not the uncertainties regarding this categorization should be accepted. It has however been found that the information that separates non-adopters from planners is valuable for gaining a better understanding of the driving forces related to motivation for adoption. 2.5.8 Non-response bias with respect to adoption of EDI There is always a risk of non-response bias with any mail questionnaire with a response rate less than 100 percent. The ratio of adopters to nonadopters and their responses might be different in the population compared to the sample. One method that is recommended in the literature is to look at possible changes in the sample response with respect to key items over time (Masters et al., 1992). The theory is, that compared to early responders late responders behave more like non-responders (Lambert and Harrington, 1990). Accordingly a test for non-response bias for adopters was performed. The received questionnaires were divided into three groups. Those that were returned immediately (1 to 5 days), those that were returned after approximately one week (6 to 9 days), and finally those that 19
Non-adoption was a situation, which was not carefully considered in the design of the questionnaire. Do adopters that for example tested EDI five or ten years ago consider themselves to be adopters or non-adopters? We did not elaborate further on the classification of non-adopter, planner, and adopter than to ask whether or not the company did not use, planned to use EDI, or used EDI. Inherent in this categorization is bias due to different understandings of use or no use. The underlying conviction was however, that EDI is subject to very little trialability (cf. Section 6.4.3).
51
were returned after nine days (10 to 27 days). This final group was considered late responders. Test for possible non-response bias (Table 2-3, page 374 20) suggests that the two variables Adopters21 and Response-time in days are independent (p value = 0.56). The rate of Adopters answering YES (32.6, 25.6, and 25.8 percent) is approximately the same for the three categories of Responsetime in days. This suggests that the sample with respect to adoption of EDI and e-commerce is a good estimate of the use of EDI and e-commerce in the Danish steel and machinery industry, given the theory of non-responses as outlined by Lambert and Harrington (1990). Another method for assessing non-response bias is to compare the study sample to other available data on EDI adoption (Thong, 1999). As shown in Section 2.5.6 the ratio of adopters to non-adopters in the Danish steel and machinery sector matches contemporary surveys of a broad sample of industries reported by Statistics Denmark in year 2000 (Ministry of Information Technology and Research, 2001). This indicates that the study sample with respect to the level of adoption resembles the entire population. 2.5.9 Estimation of the representativeness of the study sample In order to estimate the representativeness of the study the major business association representing manufactures in Denmark provided additional data. The supplied data included all manufactures from the steel and machinery industry.22 The data provided consisted of zip codes that could be used to geographically locate the companies and sizes of the companies measured as number of employees classified in six categories.23 The objective was to estimate whether or not the study sample resembled the
20
Data and statistical runs can be found in Appendix B (page 374). All variables used in the statistical analyses are emphasized in Italics in the text. 22 Similar data could not be obtained from the wholesalers in the sector. 23 The categories were designed to match those defined by Statistics Denmark. The categories are as follows: 1 to 5, 6 to 9, 10 to 19, 20 to 49, 50 to 99, 100+ employees. 21
52
study population given geographical location and number of employees. However, since it was not possible to identify the responders from the study population the relevant statistical tests that could indicate whether or not our study sample differed from the study population given geographical location should be interpreted with great caution. The study population is shown in Table 2-4 (page 375). The zip codes are divided into three categories of geographical locations: Sjælland, Lolland-Falster, Fyn og Øerne, and Jylland. In the contingency table showing geographical location and number of employees for all manufactures in the steel and machinery in Denmark based on figures for the year 2001 provided by the manufactures business association (Table 2-4, page 375) the chi-square value is 8.4768 with a corresponding p-value of 0.5824. This suggests that geographical location (Zip_code) and number of employees (Employees) are independent variables. In the contingency table (Table 2-5, page 376) for the study sample Zip_code by Employees the two categories 50 to 99 and 100+ employees have been collapsed since too many cells had expected counts less than five. The chi-square value for Table 2-5 is 6.2545, with a corresponding p-value of 0.3953 indicating that Zip_code and Employees are independent variables. It should however, be noted that the study sample is included in Table 2-4 (page 375). The test statistics from the two tables suggest that the study sample with respect to geographic location and number of employees is an unbiased sample and that the study sample regarding these two variables is an unbiased representation of the study population. 2.5.10 Selecting methods for data analysis A number of similar studies have applied various multivariate parametric data analysis methods. These methods include factor analytic techniques such as principal component analysis, which has been applied in IS research by (Chau, 2001; Premkumar and Ramamurthy, 1995) and factor analysis, which has been applied by (Chau and Tam, 1997; Grover, 1993). The strengths of applying factor analytic techniques are their usefulness for summarization of items and data reduction. In summarizing data items, the
53
aim of factor analysis is to uncover “latent” underlying dimensions that hopefully describe data in a smaller number of concepts compared to the original variables. Data reduction can for example be achieved by calculating scores for each underlying dimension and substituting these derived scores for the original variables (Hair et al., 1998). The factor analytic approach does not set any a priori constraints on the estimation of components or the number of components to be extracted and they are therefore suitable for research in topics which are relatively new and unexplored. Both the principal component method and the method of exploratory factor analysis are based on a matrix of Pearson correlation coefficients and data should therefore satisfy the assumptions for these statistical methods (Hatcher and Stepanski, 1994). The underlying requirements are the following five assumptions (Hatcher, 1994): - Interval-level measurement. All analyzed variables should be assessed on an interval or ratio level of measurement. - Random sampling. Each subject will contribute one score on each observed variable. These sets of scores should represent a random sample drawn from the population of interest. - Linearity. The relationship between all observed variables should be linear. - Normal distributions. Each observed variable should be normally distributed. - Bivariate normal distribution. Each pair of observed variables should display a bivariate normal distribution. All five assumptions should be fulfilled in order to apply the factor analytic techniques. Therefore, all relevant Likert-scale items from the survey of the steel and machinery industry were tested for normality. Following the recommendations of Hair et al. (1998) both graphical plots and statistical tests were performed to assess a possible departure from normality of the distributions of the data. The graphical plots of most of the items under investigation did not appear to follow the bell-shaped normal distribution
54
curve. The results of the Shapiro-Wilks test, one of the most common statistical tests used for estimating the deviation from the normal distribution (ibid.) showed p-values ranging from 0.0979 to < 0.0001. Table 2-6. Shapiro-Wilks test for normality of opinion data items Shapiro-Wilks p-value range p-value < 0.0001 < 0.0010 p-value >= 0.0001 < 0.0100 p-value >= 0.0010 < 0.1000 p-value >= 0.0100
Number of items 14 9 15 17
Table 2-6 shows the number of items falling in four different ShapiroWilks p-value ranges. About one third of the items have p-values in the range from 1 to 10 percent. And the remaining two third are all statistically highly significant. Therefore, it follows that the distributions of the opinion data items generally in a statistical sense differ significantly from a normal distribution. It is possible that some transformations of data could have overcome the non-normal distribution of data. However, it can be argued that Likertscales are merely manifestations of ordered categories (Siegel and Castellan, 1988) and therefore the requirement of at least an interval-scale for the Pearson correlation coefficient is not met. Based on the abovementioned considerations it was found prudent to focus on non-parametric methods of analysis. For analysis of multivariate categorical data the DIGRAM statistical package24 was used. Among other things DIGRAM can be used for discrete graphical modeling such as: Analysis of high-dimensional contingency tables by chain graph models, exact conditional tests for conditional independence, analysis of ordinal categorical data, and analysis of independence graphs. The outputs are shown e.g. in Table 8-31 (page 398) and Figure 8-3 (page 399). These outputs are results of initial, exploratory 24
DIGRAM is a software-package developed by Associate Professor S. Kreiner, Department of Biostatistics at University of Copenhagen. DIGRAM is based on the statistical principles presented by Whittaker (1990).
55
screenings performed in DIGRAM. The SAS System version 8.01 was used for the remainder of the statistical analysis. 2.5.11 Critique of the quantitative data-collection The most serious critique of the data-collection is the bias related to the responders in the survey. Using only a single responder in the organization may limit the explanatory capacity due to a possible inflation of the correlation between causes and effects (Premkumar and Ramamurthy, 1995; Harrison et al., 1997). The optimal situation would have been to have at least two responders from each organization expounding their views on adoption or non-adoption of EDI (Maaloee, 1996). The study utilizes the responses from the managers as being representative inputs to the constructs developed on the Likert scales. As demonstrated by Grover (1993) the problem in this type of research is, that even though managers are capable of making accurate judgements regarding a variety of organizational concepts over-reporting or under-reporting of a certain phenomenon may occur as a result of the responders’ personal role in the adoption issues. Another issue related to the responder is the choice of responder. It is important to determine not only exactly what the unit of analysis is, but also who is the agent acting on behalf of the unit of analysis of interest (Pinsonneault and Kraemer, 1993). But due to the size of the firms involved in the survey it would in many cases be difficult or impossible to find other people with opinions and knowledge of the issue. A third issue concerning critique of the data-collection is related to the mix of ex ante and ex post assessments from the responders. A subsequent collection of criterion data is always a problem. It is connected with uncertainties whether or not those companies that plan to adopt actually later on adopt or whether or not the non-adopters remain non-adopters (Harrison et al., 1997). That is however one of the inherent weaknesses in the factor approach to adoption. The factor approach to adoption searches for explanatory variables related to adoption, whereas the process approach
56
analyses the whole adoption process including actions and structures of the involved individuals (Langley, 1999; Markus and Robey, 1988).
2.6 Research strategy The driving force in research can be either practice driven or theory driven (Zmud, 1998). This study was accomplished in close collaboration with the professional business associations, the Danish EDI Council, and the researcher. The associations introduced the TDP leading to assessment of the outcome of the project and further to the elaboration of the survey instrument. The research project has therefore been practice driven rather than theory driven. Four issues characterize the nature of practice-driven research (ibid.): - The topic or phenomenon to be studied is defined by the sponsors not the researcher. - The research does not have, at its initiation, a specific research outcome. - The research is framed by the current understanding of a phenomenon rather than by a well-defined research model. - The research team is expected to propose and direct the research design. These four issues fit the design and fulfillment of the study since the curiosity of the business associations related to the motivation for adoption or non-adoption in the TDP to a large extent guided the research. The business associations had deep insight into the national action plan for EDI (see Section 3.4.2), but the reasons for non-adoption of EDI were unknown to the practitioners. Though the practice-driven research is not undertaken to confirm a prior research model the practice-driven research presents three major benefits to the IS research community (Zmud, 1998). First, the topics researched are extremely relevant to practice. Second, the projects themselves are executed in an objective, rigorous manner. Third, knowledge and insights 57
from the domains of both practice and academia jointly influence the findings. The benefits of practice-driven research are found to bridge the ultimate goals for this study. The goals are to make a theoretical contribution to the existing theory and at the same time to be able to convey useful recommendations to practitioners at the business association level. It is inevitably found that businesses in the future will be confronted with technological innovations that in many ways resemble EDI. It is therefore the author’s opinion that it is very likely that professional business associations in the future will have to initiate projects similar to the TDP in order to support their members. Insights gained from this study may therefore be a valuable tool for the professional business associations when they design future projects and campaigns that aim at supporting potential adopters in their decision process. In relation to the other goal, which was to make a theoretical contribution to the existing theory the opinions differ. Theory has been defined as, “... an ordered set of assertions about a generic behavior or structure assumed to hold throughout a significantly broad range of specific instances.” (Weick, 1989). One indication of a strong theory is that it, “… delves into underlying processes so as to understand the systematic reasons for a particular occurrence or non-occurrence.” (Sutton and Staw, 1995). It has been suggested that a complete theory must contain four essential elements: 1) A description of which elements logically should be considered as part of the explanation of the social or individual phenomena of interest, 2) A description of the relationship between the elements, 3) An explanation of the underlying psychological economic, or social dynamics that justify the selection of elements and the proposed causal relationships, and 4) A description of the range of the theory (Whetten, 1989). Even though it is not the aim of this dissertation to build a complete new theory these four questions are valuable guidelines when considering improvements of existing theory: Improvements which are based on the findings from the qualitative and the quantitative study in the Danish steel and machinery industry.
58
2.7 Summing up Chapter 2 As previously stated the triangulation approach is applied in this study. Two bearings are taken to assess the motivators leading to adoption or nonadoption in the Danish steel and machinery industry. The first bearing is related to an exploratory and qualitative assessment of the motivators leading to adoption or non-adoption amongst a group of organizations, which have been exposed to awareness concerning EDI, and which have been offered a low cost entry to EDI. The second bearing in the triangulation is a quantitative assessment of a larger population of the steel and machinery industry. Different means for evaluating data are chosen. The case study is subject to hermeneutic interpretation, whereas the survey follows the rules of positivistic research. Different means for datacollection and assessment of data may have been more appropriate. A number of studies have for example made the qualitative and the qualitative data-collection in the opposite order (Jick, 1979). Suggesting that a survey is an initial screen of the field and a good basis for in-depth case studies. An advantage of doing qualitative field studies before a survey is that triangulation can lead to a prominent role of qualitative evidence (ibid.). There is however, also a more pragmatic version of the order of the applied research methods: The order of the two bearings does in many ways reflect the learning process both of the researcher and of the business associations that were involved in the project.
59
60
3 The Danish business environment from an IT perspective The purpose of this chapter is to provide an insight into the structure of the Danish business and the IT/ IS initiatives that have taken place to promote the diffusion of IOS in the Danish business environment. The immediate aim is to present the context for the study of adoption and diffusion of EDI in the Danish steel and machinery industry. Apart from describing the industry structure per se a few details concerning the Danish steel and machinery industry are given. The next step in the presentation of the Danish business environment is to present the general trends in the use of IT/ IS. Focus is on the official number of users of EDI published from 1996 and onwards by the national statistics bureau, Statistics Denmark. Next, the two action plans for e-commerce that have been launched in Denmark are presented and parallels are drawn to other similar initiatives from abroad. After the presentation of the two Danish action plans for e-commerce and the coordinated diffusion initiatives from abroad the effect of regulation as a means for supporting adoption and diffusion will be discussed.
3.1 Industry structure To provide a frame of reference a very short presentation of the structure of the Danish industry is given in the following. Compared to other countries the overall figures of the Danish economy are modest. In the year 2000 the GDP was 151.3 billion USD, that is about 0.6 percent of the total in OECD. The US for comparison contributed 37.1 percent or 9,926.6 billion USD (OECD, 2001). Denmark has, similar to a number of other western economies, experienced an overall shift in the industry structure during the last fifty years. The share of primary and secondary industries has 61
decreased leaving a large share of the workforce employed in tertiary industry. In 1998, 72 percent of the workforce was employed in tertiary industry. Especially the public sector employs a large share of the workforce. Thirty percent of the total workforce or more than 800,000 people were employed in the public sector in 1998. Denmark has thus moved from having agriculture as the primary activity to being an industrial nation, and further on to become a ‘service nation’ (Abildgren and Nielsen, 2000). The Danish industry is characterized by manufacturing of highly specialized products, high wages, and a high employment rate (OECD 2001). Denmark ranks as number five among the OECD-countries if wealth is measured as an average of the country’s production per capita (Erhvervsministeriet, 1999). Denmark has up through the 1990s improved its position among the richest OECD-countries. One explanation for this progress is that Denmark has experienced better market conditions than the rest of Europe. Denmark has for example witnessed a growth in employment during the 1990s. Small businesses are the backbone of the economy in many countries. And they are very important in the economic development process of a country. The share of small and medium sized enterprises (SMEs) is high in the Danish business environment. In 1998 approximately fifty percent of the registered Danish companies had an annual turnover less than EUR 70,000 (Statistics Denmark, 2000a). There are almost 300,000 Danish companies. Of these approximately two third are associated with the service industry, the rest are about equally distributed among farming and industrial production. There are almost 49,000 companies in industrial production. Of these companies the majority (35,708) are very small employing 1 to 9 persons. In 1999, the sectorial contributions to gross value from agriculture, industry and services were respectively 2.7 percent, 25.5 percent, and 71.9 percent (comparable figures from the US were respectively 1.7, 26.1, and 72.2).
62
3.2 The Danish steel and machinery industry Twenty-four percent of the workforce was employed in secondary industry in 1998. Of these seven percent were employed in the steel and metal industry25 that is the largest sector among the secondary industries closely followed by the building and construction industry, which employs about six percent of the total workforce (Statistics Denmark, 2000b). While the general trend has been a decrease in number of employees in the secondary industries the steel and metal industry among a few others has encountered a growth in number of people employed. The steel and metal industry has traditionally produced for the national market due to the high transportation costs of the manufactured goods. The sector has however during the last years established subsidiaries abroad and the turnover in the sector has increased. In 1997 the globalization26 rate for the steel and metal industry was 52 percent whereas the rate was 57 percent for the industry in general. The share of exported goods in the steel and metal industry was 35 percent and for the industry in general 45 percent (Erhvervsministeriet, 1999). This indicates that the steel and machinery is catching up with respect to globalization and export of goods. Grunfoss and Danfoss, two of the largest companies in the steel and machinery industry in Denmark, are examples of companies in the sector that export a high share of their production and which have established subsidiaries abroad. The companies included in the sample from the Danish steel and machinery industry, which has been explored in depth in this dissertation are mainly focused on production, repairs, trade, and service within the following industries: Blacksmithing, machinery, electronics, and sanitation. Most of the businesses in the sector are dealing with highly specialized products. 25
Statistics Denmark uses the category steel and metal industry instead of the steel and machinery industry as defined by the industry and trade associations. The steel and machinery industry covers a larger sample of organizations than the steel and metal industry and the numbers reported in the sector are thus conservative estimates. 26 The Danish Ministry of Trade and Industry has developed a measure for the globalization rate. By summing the value of export of goods and services and the activities in subsidiaries located abroad the globalization rate is operationalized.
63
One thing which is characteristically for the companies in the Danish steel and machinery industry is that they are often run by their owner and a high number of the companies in the sector are using the expertise of other companies either as subcontractors or in a direct collaboration.
3.3 Use of EDI in the Danish business environment The growth in use of IT in the Danish industry and trade sectors has been steady from the mid 1990s when the issue caught attention due to awareness campaigns launched by governmental units and professional business associations. From 1996 and onwards the Ministry of Information Technology and Research has made yearly publications (except in 1998) describing the use of IT in Danish businesses. In the following a summary of the use of EDI is presented based on these publications. The objective of presenting five years of statistics published by the Ministry of Information Technology and Research is to illustrate the degree of IT and EDI use during that period. The publications vary in content. 1996 was the only year where opinion data was obtained from responders. Afterwards data is reported on mere numbers of users. The selection of responders is not described and it is hence difficult to interpret the large differences in the number of users over the time period. In 1996 when the first publication on the use of IT in Danish businesses was released (Ministry of Information Technology and Research, 1996a) it was stated that in 1995 thirty-one percent of all companies with more than five employees used EDI. The explanation for the scope of diffusion was according to this first IT survey, that a few central actors were willing to initiate the process. 58 percent of the users said they used EDI due to a request from a business partner. The major barriers for further diffusion were according to this survey: Practical and economic considerations. Strategic issues were on the other hand not viewed as a major incentive. In 1996 the number of EDI users in the Danish business community had risen to thirty-three percent (Ministry of Information Technology and Research, 1997).
64
The Ministry of Information Technology and Research did not in 1998 publish any statistics related to the use of EDI. However, in 1999 when the next set of figures was published related to EDI use in 1998 the number of reported users had increased to 56 percent. Of these 32 percent used the EDIFACT standard27 for their EDI communication (Ministry of Information Technology and Research, 1999). The statistics covering the use of EDI in 1999 (Ministry of Information Technology and Research, 2000) showed that 36 percent of the Danish companies used EDI. The latest figures related to EDI use in the Danish business community report that fifteen percent of the companies with more than five employees are using EDI (Ministry of Information Technology and Research, 2001). The decrease in the number of users compared to the previous year is explained by a change in the sample. Contrary to the previous year the category of small companies (5 to 9 employees) was included in the statistics for year 2000. The decrease in the number of users during the period 1995 to 2000 somewhat challenges the findings by Andersen et al. (2000). Andersen et al. found an ongoing growth in the exchange of EDI messages based on scope of messages sent via seven major vendor companies in Denmark. There was a substantial growth in number of EDI messages exchanged ranging from an annual growth of 34 percent to 46 percent. Though the same study revealed that the growth in bytes transmitted during the same period increased by the same rates, indicating that larger messages are exchanged, there is no reason to believe that the number of EDI users have been reduced during the period 1995 to 2000. Especially when considering the increase in assigned EAN-numbers28 (Andersen et al., 2000). There are however, good reasons to believe that the latest figures from Statistics
27
EDIFACT is an acronym for EDI for Administration, Commerce and Trade. The United Nations has defined the EDIFACT standard. UN/EDIFACT includes a whole range of guidelines and directories including syntax rules, a data element directory which is a subset of the United Nations Data Element Directory, code lists, standard data segments directory, and United Nations standard message types (Horlück 1994). 28 The number of assigned EAN location numbers reflects the number of companies that are able to send and receive EDI messages.
65
Denmark are a good representation of the current EDI use in Denmark (cf. the survey results reported in Section 2.5.8).
3.4 Background for use of EDI and e-commerce in Denmark 3.4.1 Introduction The purpose of this section is to sketch the efforts done to communicate the importance of EDI and B2B e-commerce to private businesses and the public sector. The aim is not to present the Danish e-commerce initiatives launched by governmental units and professional business associations in detail, but rather to illustrate which elements were considered as important and who the players were that took part in the communication process. Ecommerce has throughout the 90s increasingly been labeled as a major business imperative (Choi et al., 1997; Kalakota and Whinston, 1997; 1996). Several efforts have been made to promote the ideas and to support businesses that wanted to adopt the innovation. OECD and EU have allocated resources for marketing the innovation at an international level and several national initiatives have also pursued the idea of diffusion of ecommerce (See Section 3.5.2. for examples). 3.4.2 The e-commerce action plan from 1996 In 1996 the Danish Ministry of Research and Information Technology, the Ministry of Business and Industry, the major industry and trade associations and public institutions proposed an action plan for electronic commerce29 (Ministry of Research and Information Technology, 1996). The action plan will in the following be referred to as the 1996-action plan. The 1996-action plan was an untraditional joint initiative between governmental units and players from the private and public sector. It was acknowledged in the preface to the 1996-action plan that the success of the plan depended on the active involvement of the Danish businesses and public institutions.
66
The forerunners for the 1996-action plan were two policy statements prepared for the Danish Parliament. The first statement “From Vision to Action – Info-Society 2000” (Ministry of Information Technology and Research, 1995), which was published in 1995 by the government, was used as a lever to create awareness of the great significance of the information revolution. It was stated that this movement towards the InfoSociety was a public movement involving everybody. In a true Danish spirit it was argued that if N.S.F. Grundtvig30 had been alive he would probably be connected to the Internet. This rather strong symbol illustrates the expectations of the move towards the information society, which was seen as the means for everybody to be involved and to get educated. The strategy should thus be based on a Danish model. Market forces should not, similar to initiatives in Italy (Kumar et al, 1998) or Hong Kong (Damsgaard and Lyytinen, 1998) be allowed to be the only forces determining the development. Though the statement primarily focused on building a strong infrastructure the responsibility of the public sector to engage in the movement was stressed. Additionally, the role of Danish companies was by no means underestimated. It was thus acknowledged that the great opportunities created by the information society implied several challenges for Danish companies. The policy statement expressed that: “Danish companies must not only effectively introduce new technology for rationalization purposes, it certainly also means that they must be able to transform new technology into new products to respond to special customer requirements.” (Ministry of Information Technology and Research, 1995).
It was suggested that companies adopted EDI which, “... will result in considerable rationalization gains and a closer interplay between organizations.” Internally in the organization it was expected that the 29
Though the action plan was called an action plan for e-commerce it focused on EDI. It makes sense in a historical perspective whereas the lapse might be less appropriate from a linguistic point of view. 30 N.S.F. Grundtvig (1783 to 1872) was a famous Danish poet, statesman, and church father. He was the driving force behind the establishment of the first Danish folk highschools where national poetry and history formed an essential part of the instruction. The folk highschools have played a significant role in the education of the rural youth until the 1950s when the governmental education system took over.
67
increased use of IT would contribute to the personal development of the employees since the individual employee would be able to communicate more easily and obtain the information they needed for improved performance of work routines. At that point in time the 1996-action plan had not been formulated. But it was announced that the Ministry of Research and Information Technology, the Ministry of Business and Industry, the Danish EDI Council, and relevant industry and trade associations were about to launch a campaign to further the use of EDI and e-mail among businesses and public administration. In 1996 the next IT policy statement was published. The statement was named “The Info-Society for All - the Danish Model” (Ministry of Information Technology and Research, 1996b). In the 1996 statement it was announced that: “…, this new technology presents a number of opportunities and problems, which demand political consideration and action. A cohesive, aggressive strategy for how we wish to further the developments in Denmark is necessary.”
Included in this “aggressive strategy” was the 1996-action plan for electronic commerce. It was found that, “... the importance of a fast, effective and consistent implementation of e.g. EDI could hardly be overestimated.” First and foremost it was claimed that technological landmarks such as EDI could give Denmark an international lead along with improvement in efficiency of working procedures and development of new products and production processes. The goal of the 1996-action plan was thus to provide the necessary conditions for companies, the public sector, and not least, the consumers to reap the gains resulting from EDI. The 1996-action plan acknowledged that due to the growing globalization of commerce it was found to be essential for Denmark to be able to follow the trend of doing business electronically across the borders. The parties involved in the action plan knew that EDI as such was not a novelty but due to the diffusion of IT in the Danish society and the decrease of software and hardware prices a fertile environment for diffusion of EDI was likely to appear. It was recognized that most Danish 68
companies possessed a sufficient level of experience and know-how regarding IT to be able to adopt and implement EDI. In 1996 when the action plan was launched fifty percent of the Danish companies exchanged data via either telephone technologies or network technologies. Especially exchange of electronic messages to financial institutions had a high diffusion rate among Danish companies (Horlück, 1996). The relatively new Internet31 was already adopted by one out of five companies and it was reported that thirty percent of all businesses planned to adopt Internet in the near future (Ministry of Information Technology and Research, 1996a). It seemed as if the time was ripe for a coordinated effort to spread electronic communications from a few sectors to all industry and trade sectors in Denmark.32 The action plan was, based on this broad agenda, launched to the business environment and the public sector. In the foreword to the action plan it was stated: “The plan is to provide dynamism and accelerate growth. This will be achieved through the public sector joining forces with a large number of commercial organizations to create joint solutions. Thereby we will avoid a state in which everyone waits for everyone else, or in which the approaches chosen are not coherent.”
In order to create the necessary dynamics and consistency seven initiatives were formulated. The initiatives primarily aimed at supporting the diffusion of EDIFACT based EDI communication, thereby supporting electronic data transactions between private companies and the public sector. The expectations towards the ability of the public sector in relation to development and implementation were high. It was found that the public sector had to take the lead and show an example.
31
It is a truth with modifications that the Internet was relatively new in 1996. The fundaments for the Internet was build in mid 60s with the launch of ARPANET. Until 1993 with the release of Mosaic, the first multimedia browser, the Internet (ARPANET) however mainly consisted of research universities and military contractors (Kalakota and Whinston, 1996). 32 The financial sectors and the grocery sector were amongst those, which already had applied electronic communication to their business processes.
69
Table 3-1. The 1996-action plan Initiative Policy consensus 1. Establishment of EDI No later than 1998, the EDIFACT standard must be standards in all sectors established in all industries and sectors, for all commercial documents of significance, such as orders, invoices, payment messages, transport notes and registration of real property. The goal is to ensure the availability of a vital prerequisite for companies’ options to participate in the electronic marketplace, within trade, manufacturing, transport, finance, etc. 2. EDI for public procurement contracts
Through forthcoming EU framework agreements, the public sector will include its suppliers’ ability to participate in fully electronically based document interchange as an integral part of its tendering conditions, no later than 1998.
3. Handling EDI in public-sector financial systems
By the end of 1998, public-sector financial systems will be able to handle all relevant commercial documents in EDIFACT format.
4. EDIFACT-based interchange of administrative information with the public sector
In order to ease the administrative burden on companies, the opportunity must be created before the end of 1998 for companies operating in areas in which serviceable standards exist, to undertake EDIFACT-based electronic reporting to the public sector. And initiatives will be aimed at areas in which there is a need for new standards.
5. Development of EDI software
A number of initiatives are being aimed at software developers. These initiatives are intended to promote the development of a range of EDI software products destined for the market. The price and functionality of these products must satisfy the needs of all types of companies, regardless of an individual company’s level of ambition concerning the use of EDI.
6. Legislation on digital New legislation on digital signatures will prepare the way signatures and electronic for ensuring that the use of electronic communications is just as secure and clear cut as the use of conventional documents communications on paper. 7. Danish EDI Council as initiator and coordinator
The Danish EDI Council will assume a central initiating and cross-sectorial role in the implementation of the action plan.
Source: Adapted from the Danish national EDI action plan (Ministry of Research and Information Technology, 1996)
70
It was acknowledged in the introduction to the 1996-action plan that the project was ambitious and that it required involvement from several business sectors and institutions. The timeframe for the adoption and implementation of EDI was short. It was thus expected that the use of EDIFACT-based communication could be tested and diffused among all relevant sectors by year 2000. As initiatives one to four in Table 3-1 show the expectation was that EDI could be implemented by year 1998. The means to meet this end was to introduce awareness campaigns arranged by the Danish EDI Council. Another way to fulfil the initiatives was to make the necessary arrangements to be able to execute public procurement via EDI. It was however the adopters that carried the main responsibility for acting according to the recommendations in the action plan. This strategy was in accordance with the IT policy statement from 1996 where it was stated that: “One decisive feature of “the Danish model” is that, without grandiose plans, but precisely through dialogue and effective action, we are in a position to implement the necessary infrastructure quickly and to remove the barriers to it.”
This attitude was explained by former experiences with the long Danish tradition of the establishment of the cooperative dairies and abattoirs, which successfully took place at the end of the 19th century.33 The expectations related to the industry and trade associations and the individual organizations were thus clearly expressed. Though the individual initiatives were to be carried out by the industry and trade associations in concert with the individual organizations the Danish EDI Council was
33
The co-operative movement led to a social and economic lift for a large number of small and often impoverished farmers. The establishment of co-operative dairies and abattoirs improved the quality and quantity of the production, which benefited the Danish export of agricultural products to for example Great Britain, which at that time was engaged in industrialization. The co-operative movement is from a political point of view seen as an important factor for the development of the Danish parliamentary democracy.
71
appointed as the coordinating unit. In addition the involved ministries provided some financial resources.34 Though the Danish EDI Council was appointed to monitor the implementation of the 1996-action plan, it had however, not any fiscal authority towards the private or the public sector. The active role played by the Danish EDI Council mainly resulted in support of several projects and provision of information and publications on EDI to the business community.35 In their efforts to create awareness of EDI in the Danish business community the Danish EDI Council provided a definition of EDI in order to support a shared understanding of the term. The Danish EDI Council used the following definition of EDI: The term EDI is defined as the exchange of structured, electronic messages. This exchange is conducted with a minimum of human interaction. A requirement for defining an electronic exchange, as EDI is, that messages are exchanged based on a beforehand agreed standard. This format can be an individual proprietary standard or an international standard e.g. EDIFACT. 36
This definition includes those elements that are outlined in definitions in academic publications e.g. Hansen and Hill (1989) and Pfeiffer (1992) (cf. Section 5.4.2). In this context the most interesting issue is that the Danish business and administration environment was exposed to this definition. In relation to the 1996-action plan the Danish EDI Council has broadened the definition by including proprietary standards. The action plan on the other hand favors the EDIFACT standard and encourages organizations to adopt this standard.
34
DKK 24.6 millions were granted by the two involved ministries. DKK 18 million were earmarked for awareness initiatives to be carried out by The Danish EDI Council and DKK 6.6 million were made available for projects related to standardization issues. The Danish EDI Council would administer the money. 35 A complete list of activities and projects supported by the council can be found at www.edi.dk/ 36 This definition was inserted in the survey, which is described in detail in Chapter 8, since it was expected that this definition was familiar to the responders.
72
The two IT policy statements were especially focused on building a telecommunications infrastructure to support the Internet thereby gaining the benefits and opportunities connected to that medium of transportation of information. The EDI definition from the Danish EDI Council does not go any further than stating that concerning EDI, there is an electronic exchange of messages connected to EDI whereas the means of transportation was found less relevant. In the same manner the degree of organizational integration is subject to individual interpretation by only mentioning a minimum of human interaction as the ultimate goal. In 1998 the Danish EDI Council evaluated the outcome of the 1996-action plan.37 It was acknowledged that though a number of initiatives had taken place within different sectors including development of low-cost EDI software, diffusion of EDI had not happened at the pace envisioned in the 1996-action plan. Based on information obtained from the business associations initiative one (establishment of EDI standards) was close to being fulfilled in sectors like mortgage, shipping, and insurance whereas e.g. industry and trade sectors were progressing somewhat slower. It was however, expected that standardization would be completed by 1999. It was realized that especially the SMEs had not implemented EDI as expected. In relation to the standardization issue it was realized by the Danish EDI Council that the XML-standard (eXtended Markup Language) seemed to be a promising alternative to EDIFACT. It was also suggested that the Internet could be used as an attractive medium for transportation of EDI messages. It would be cost effective and uncomplicated. The fifth initiative in the 1996-action plan (development of EDI software) has according to the assessment performed by the Danish EDI Council had a slow start. One problem was that the different needs of the companies had not been considered from the very start. Development of EDI software had thus not considered price and functionality in relation to the individual organizations’ needs and their level of ambitions for EDI usage. These
37
This information is based on an internal working paper “Midtvejs-status for Den Nationale EDI-Handlingsplan” [“The mid-way evaluation of the National EDI Action Plan”] provided by The EDI Council.
73
deficiencies had however been mended by the initiation of pilot-projects e.g. the TDP (see Chapter 4). In the overall evaluation document of the 1996-action plan from the Danish EDI Council38 it was concluded that the 1996-action plan had been a success. One of the reasons for the success was found in the wide support for EDI from business associations. The criterion for success was, according to the managing director of the Danish EDI Council, to provide the necessary opportunities and tools for EDI for business and administration. It was not seen as an objective in itself to get as many organizations to use EDI as soon as possible but rather to create the opportunities for EDI. Textbox 3-1. Greetings to the 1996-action plan, a review of the plan “Catch the Ball” The following text is a report from a feature article in a business magazine, which 39 focuses on the steel and machinery industry. “The 1996-action plan is a fantastic opportunity for the Danish business community” with these words the managing 40 director Mr. Agner Mark praises the emergence of the EDI action plan. Mr. Agner Mark reviews the seven initiatives of the action plan and argues that the action plan is ambitious because it is the first of its kind in the world. Due to is specificity it is rather a users manual than a declaration of intent. By adopting EDI as prescribed in the action plan it is possible to gain efficiency, reduce cost on business routines, and give a number of Danish businesses a competitive lead. “The consequence of the governmental initiative is that we build a solid infrastructure securing Danish businesses optimal conditions for improved service and collaboration with suppliers and other business partners. At the same time the administrative burdens are reduced and Danish businesses get a good position in the global competition because we get a lead in the utilization of the technology.” Mr. Agner Mark continues by saying “ Electronic communication is not enough in itself. Data has to be transferred automatically from one IT-system to another. It is the 38
This information is based on an internal working paper “Slutrapport for Den Nationale EDI-Handlingsplan” [“Final report on the National EDI Action Plan”] provided by The EDI Council. 39 The article is from the magazine “Jern og Maskinindustrien” [“The steel and machinery industry”], vol. 26 (1996), issue 22, pp. 26-27. The magazine is the leading channel for information concerning the steel and machinery industry. An independent private publishing house publishes the magazine. 40 The managing director of Dan Net Ltd. Agner N. Mark writes the feature article. Dan Net Ltd. was, and still is, one of the largest VANs providers in Denmark.
74
prerequisite for fast and low-cost transfer of large volumes of data where errors are avoided. EDI (EDIFACT) is the solution - or rather it is the technological prerequisite.” The final statement from Mr. Agner Mark is: “ The action plan implies a number of good initiatives, it is obvious that the business community should catch the ball – it has already been thrown into the field. The businesses can’t afford not to.”
The feature article is an illustration of how EDI and the 1996-action plan was presented in publications read by a broad audience in the business community. 3.4.3 E-focus – an e-commerce agenda from 1999 The e-commerce agenda from 1999 (Ministry of Research and Information Technology and Ministry of Trade and Industry, 1999), "Focus on ecommerce", was the next step in supporting diffusion of e-commerce among Danish businesses, the public sector, and private households. The initiative was, similar to the 1996-action plan, an untraditional joint initiative launched by the Danish government and a number of important commercial and interest associations. Focus on e-commerce thus involves more than just the government. It comprises a broad partnership across the nation. Its action plan is continually being adjusted, but consists initially of seven points for action. The specific stages of the project are carried out by a number of work groups. “Focus on e-commerce” aims at achieving fast, flexible and efficient action. In recognition of the fast changing technological environment of ecommerce the main players of the action plan have taken new means into use for adoption of the action plan. “The option of adopting a traditional action plan has been abandoned because it would probably mean that the content and issues involved would become irrelevant or need to be modified before the plan could be put into effect.”
“Focus on e-commerce” is organized to meet the conditions imposed by the information society and the Internet. The aim is to initiate concrete activities, which can promote the development of electronic commerce in Denmark. The seven points for actions are:
75
1. Awareness and sharing of experiences. 2. Reliable and secure basic functions on the Net. 3. Establishing an effective infrastructure for electronic commerce. 4. Contractual and regulatory frameworks. 5. Research and training. 6. Electronic commerce in the public sector. 7. Social aspects of e-commerce. In order to meet the multiplicity of e-commerce the action plan focuses on the above-mentioned seven points of action. There are however, not any specific goals that have to be achieved within a timeframe in contrast to the 1996-action plan. The involved partners in “Focus on e-commerce” have recognized that ecommerce is broader and more multifaceted than the exchange of EDI messages. Contrary to the 1996-action plan the e-commerce agenda from 1999 introduces a dynamic definition of e-commerce where all possible actors, business processes, and technologies are included: The concepts electronic commerce and e-commerce are used synonymously in Focus on e-commerce, and should be understood in a broad sense. The general understanding can be expanded in the following three dimensions (Ministry of Research and Information Technology and Ministry of Trade and Industry, 1999): Application of technology – e-commerce includes the use of a wide range of technologies, both EDI (Electronic Data Interchange) and the Internet as well as whatever new technologies may appear – for example in the form of set-top-boxes for Internet commerce via TV screens and mobile terminals such as hand-held computers and mobile telephones. The trading parties – The concept of e-commerce covers trade between companies, trade between companies and consumers, and trade with public institutions. Components of the trading process – The whole of the trading process is included, from marketing to entering into agreements and payment, as well as after-sales service. Electronic marketing and service are expected to offer exceptionally good opportunities, so the actual occurrence of money transactions is not considered essential.
In line with the international trends the Danish business environment adapted to the developments that took place during the period from the 76
time the 1996-action plan was designed and launched and until “Focus on e-commerce” was introduced. The demanding EDIFACT standard was replaced by “whatever new technologies may appear” and the scope of trading parties was expanded to include consumers. E-commerce was setting the agenda rather than EDI.
3.5 Assessment of the influence of the action plans Even though the goals of the 1996-action plan, which by many means were ambitious (Henriksen and Andersen, 1999), were not fulfilled it served as a tool for creating awareness especially among SMEs. As such it served as a communication tool. According to the diffusion theory presented by Rogers (1995) adoption of innovations is largely a communication process (Kautz and Larsen, 2000). The large companies had already adopted EDI (Andersen et al., 2000) but the SMEs hesitated due to lack of the necessary knowledge of the innovation. It is a fact that there was growth in the EDI traffic (measured as number of messages and number of bytes exchanged) from 1996 and onwards (ibid.) but a large EDI landslide did not happen. However, due to the involvement of the industry and trade associations in the formulation of the 1996-action plan several EDI projects were initiated. Due to the well-defined goals of the 1996-action plan the assessment of the 1996-action plan is less complex compared to the 1999-plan. In the midway evaluation of the 1996-action plan from 1998 which was produced by the Danish EDI Council it was stated, that several efforts had been made to meet the recommendations, but that the goals so far had not been reached. The writings of Lai and Guynes (1997), Gregor and Johnston (2001) Johnston and Gregor (2000), King et al., (1994) uncover a discrepancy in relation to the effects of coordinated efforts made by industry and trade associations or governmental units in relation to adoption and the actual diffusion of IT. It has been stated that institutional (especially governmental units) interventions (Lai and Guynes, 1997), and change in the remote environment (e.g. regulation) and industry structures (Gregor and Johnston, 2001) are among the most powerful causes for IOS adoption. Deliberate institutional interventions or refraining from interventions are 77
thus found to play a vital role in technology diffusion (King et al., 1994). The opposite point of view has however also been stated recently by (Johnston and Gregor, 2000) who argue that, “... deliberative coordinated action by an industry as a whole, or units purporting to represent such a group position, may be severely limited in effectiveness”. In the next section a number of governmental and private initiated projects are presented to illustrate the outcome of coordinated efforts aimed at supporting adoption and diffusion of IOS in business communities. 3.5.1 Coordinated diffusion initiatives Efforts to diffuse IOS in business communities have been initiated through several initiatives similar to the 1996-action plan. The efforts have primarily focused on building electronic infrastructures and to provide information about the opportunities resulting from linking to electronic networks. The purpose of this section is to describe the objectives, means and especially the reported outcomes of these initiatives. The main focus point for the current section is to examine how coordinated adoption and diffusion efforts influence adoption of IOS at the organizational level. Different parties can address the building of electronic infrastructures, which are a necessity for IOS diffusion (Damsgaard and Lyytinen, 1998). The most well known are the large governmental initiatives such as the 1996-action plan, the TEDIS project launched by the European Commission, and the Singaporean TradeNet. Industry segments and private associations might however, play an important role too (ibid.). An example of a project initiated by industry segments includes the TDP (Henriksen and Andersen, 1999; Henriksen, 2000), which will be presented in details in Chapter 4. The establishment of an electronic infrastructure based on initiatives from private associations can be illustrated by the French Minitel project (Hill, 1997) or the initiative in Hong Kong (Damsgaard and Lyytinen, 1998). The SPRINT project is an example of a project where all three parties (governmental units, industry segments and private associations) were involved (Kumar et al., 1998).
78
The major reason for choosing the above-mentioned projects is to present projects, which similar to the 1996-action plan, aimed at building an infrastructure for a larger group of participants. Newer initiatives could have been chosen e.g. some of the more recent projects launched by the European Commission.41 This idea was however abandoned. First of all, because that would be a research project in itself and secondly, because no official assessments of these initiatives have yet taken place. Instead, older studies, which often have been used as references are included. Some of the advantages of including initiatives, which are contemporary to the 1996action plan, are that the projects met with similar technical obstacles and the projects have had the necessary time to be absorbed in the business community hence making it possible to evaluate the projects. The projects presented in the following have been subject to examination by researchers and the description of the projects is based on academic publications. One overall challenge in relation to assessment of the outcome of coordinated initiatives is the level of analysis when measuring the outcome. One way to evaluate the success or failure of a project is based on the projects’ rate of adoption among potential users. As the following presentation of coordinated efforts shows this might be a less fruitful path since some or all adopters at a later state might reject the innovation e.g. (Kumar et al., 1998). To denote adoption rate as the dependent variable for success thus has to be seen in relation to time (DeLone and McLean, 1992). Another common way of assessing the outcome of the coordinated initiatives has been to look at the degree of use within the individual organizations that are within the scope of the initiative (Andersen et al., 2000). This approach might however, show more about the individual motives for adoption; not the actual success of the coordinated efforts for diffusion of EDI. The following section on evaluations of the coordinated initiatives does reflect this trend. Except from the SPRINT project the evaluations of the projects have taken an offset in the micro level of analysis. 41
These include MESSENGER (Manufacturing and Engineering Service Status Electronically Negotiated for Greatly Enhanced Response in the European Commission's IST Programme. (www.europa.eu.int/-ISPO/ecommerce).
79
3.5.2 Examples of institutional initiated diffusion projects In 1988 the European Commission launched the TEDIS programme, which is an acronym for Trade Electronic Data Interchange Systems. The general focus of the industrial policy of the European Commission is to facilitate economic integration within Europe as a precondition for political integration as well as an overall strengthening of European industries within the global competition (Klein, 1995). According to Klein, it is widely accepted that the guidelines for the industrial policy acknowledge the inevitable influence of government on industry. The ideal for public policies is therefore to develop and maintain clear and predictable conditions for private activities. A closer look at the conditions outlined in the TEDIS programme highlights the following features of the two stages of the programme. The first stage of the programme was launched in 1988. The primary goal for stage one was to develop the underlying conditions necessary for a successful implementation and swift diffusion of EDI (ibid.). Development of industry standards took place in stage one. These standards include the Odette standard for motor vehicle design, Edifice for electronics and dataprocessing, and EAN-EDI for retail, trade, and distribution (European Council, 1991). The objective of the first stage of the TEDIS programme was thus to build an electronic infrastructure based on common industry standards. Especially the Odette standard for the automotive industry has served as a useful tool for the EDI diffusion within that sector (Horlück, 1994; Tuunainen, 1998).42 In 1991 the second stage of the programme was announced. “The objective of the project, is to ensure that EDI systems are set up within the Community in the most efficient manner possible” (European Council, 1991). According to the Council Decision a list of eight objectives have been specified including: Standardization of EDI messages towards the 42
It is however, difficult to assess whether the general readiness for EDI adoption in the European automotive sector was due to the EDI developments in the US automotive industry (Srinivasan et al., 1994; Mukhopadhyay et al., 1995), or whether the development of the European Odette standard motivated adoption of EDI in this particular sector.
80
EDIFACT standard, development of telecommunications networks, analysis of the impact on company management, and information campaigns. Objectives that in many ways resemble the objectives specified in the succeeding Danish 1996-action plan. The outcome of the TEDIS programme was evaluated based on fourteen case studies of the use of EDI in the European countries (Krcmar et al., 1995). The fourteen case studies investigated the potential benefits and the wider social consequences of the use of EDI in the European industry. The case studies especially focused on the impact on company management and the influence of adoption in relation to the particular organizations. In the overall assessment of the fourteen case studies of selected businesses in eight European countries43 it was found, that eleven of the participating businesses evaluated their business relations with their trading partners as improved due to usage of EDI (ibid.). At the interorganizational level it was found that EDI had a wide impact on the competitive arena of the industry as well as on the value-chain. None of the cases indicated that the electronic infrastructure for EDI was inadequate (Bjorn-Andersen and Krcmar, 1995). One interpretation could be that the TEDIS had succeeded in building a satisfactory electronic infrastructure. Another less optimistic assessment could be that the institutional development of the necessary infrastructure played a marginal role in diffusion of EDI, since the companies were already willing to change behavior and adopt EDI. In accordance with this latter interpretation the TEDIS project provided general support rather than individual help. The majority of the studies assessed the value of EDI from a strategic point of view. The penetration of EDI to a larger number of business partners was thus of less importance. In a number of cases the number of EDI links established were limited e.g. (Bjorn-Andersen and Nygaard-Andersen, 1995a; Bjorn-Andersen and Nygaard-Andersen, 1995b). It can therefore be questioned whether all businesses in the European community benefited 43
Belgium, The Netherlands, France, Italy, Germany, Denmark, Switzerland, and Spain.
81
from the electronic infrastructure or whether the selected companies, which in general were large companies or public agencies, maybe would have invested in an electronic infrastructure anyway in order to reap the strategic benefits. A common feature of the fourteen businesses investigated is that all of them were initiators of EDI connections with their trading partners. As the review of the EDI literature in Chapter 5 shows the incentive for EDI adoption is more attractive for initiators than for followers due to the higher pay-off for initiators. The fourteen case studies resemble the findings from a number of studies that evaluated the outcome of the 1996-action plan (Andersen et al., 2000). Andersen et al. assessed the adoption rate using four measures44 for utilization of EDI at the micro level. The study, which focused on the use of EDI in private businesses and public agencies, concluded that the overall number of EDI links established is limited. The number of links is generally limited to a few links to preferred trade partners. An exception is the MedCom II project, which is a joint project among health-care agencies such as hospitals, laboratories, and physicians and as such is a mix of public and private players. An interpretation of the two multiple-case evaluations is that the major benefits on the organizational level realized from the TEDIS programme and the 1996-action plan are to be found in strategic gains whereas the operational gains are minimal or non-existent. It is acknowledged that it is difficult or maybe even impossible to evaluate the benefits of EDI implementation in the organization based on a cost-benefit analysis (BjornAndersen and Krcmar, 1995). It can therefore be questioned whether the objective of the plans in relation to providing an electronic infrastructure has paid off as expected when only relatively few (and powerful) corporations have made use of the facilities. The Singaporean TradeNet is, contrary to the TEDIS project and the Danish 1996-action plan, launched as a mandatory initiative from a 44
The four measures, volume, integration, depth, and scope are presented in detail in Chapter 4, where the measures are used as an evaluation tool for the TDP case study.
82
governmental agency, the National Computer Board. The companies involved in the export sector were forced to implement EDI for the exchange of export-documents with the public authorities. The electronic infrastructure was launched in 1989. The TradeNet is the first nation wide EDI ever implemented in the world (Teo et al., 1997). By the end of 1994 almost 99 percent of all trade declaration documents were processed through TradeNet. The Singapore experiment can at the macro level be viewed as a success in relation to diffusion of EDI in a business community, where it has achieved or surpassed its goals (Thong, 1999). The assessment however presents a different picture when it comes to the micro level (Teo et al., 1997). If the number of participants is used as a measure of success or failure of an electronic network then the French Minitel must be characterized a solid success. The Minitel initiative offered access to a wide range of information sources and allowed products to be ordered. Basically, Minitel can be considered to be a forerunner for the Internet.45 46 The Minitel service was offered by France Telecom and is as such a private initiated network. Though the Minitel project by no means can be compared to business-to-business projects, it serves as a good example of a project that diffused among the target group within a short time-span. The Minitel, which was launched in 1984, was adopted by 36 percent of the French population over the age of fifteen years by the end of 1994 (Hill, 1997). The high penetration of Minitel can be explained partly due to the low cost of the terminals. The terminals were in fact provided free of charge by French Telecom in the early years of the Minitel-service (ibid.). The Italian SPRINT project had, similar to Minitel a high penetration rate in its initial phases. Similar to the Minitel the SPRINT-service was offered to the users at a very low price. But, contrary to the Minitel project the users of
45
The same reservations in relation to the chronology of the Internet history as described in footnote 31 should be kept in mind. 46 Minitel is not an interorganizational information system. The reason for including Minitel is that it in many ways assembles the other projects included in this section. The objective of Minitel was to establish an electronic infrastructure. Minitel additionally illustrates the importance of critical mass in relation to service networks.
83
SPRINT abandoned the services offered as soon as the subsidies to the project ended (Kumar et al., 1998). The above-mentioned initiatives show that industry wide penetration of IT does not happen just by launching coordinated diffusion initiatives. What is common for the examples mentioned above is, that there has been a wide support in the initial phases of the projects and that a stable infrastructure has been established including development of common standards. It is however, questionable whether the diffusion of the technological innovation is directly related to the coordinated initiatives in relation to the non-subsidized voluntary projects such as the TEDIS project and the Danish 1996-action plan. Minitel, the TradeNet, and the SPRINT projects serve more as an illustration of how it is possible to establish and diffuse communications networks among private or business entities and are thus less relevant than the TEDIS project and the 1996-action plan in this context. These two are especially relevant in this context since they are focused on efforts to diffuse EDI in a business community based on businesses’ voluntary commitment. The organizations that were investigated in both evaluations (Bjorn-Andersen and Krcmar, 1995; Andersen et al., 2000) were large companies, which hold the position of initiators in the business community. As stated above it can be argued that the willingness in relation to investment is a major driver for adoption of technological innovations. It can thus be argued that the outcome of the coordinated initiatives is limited due to the fact that the initiators most likely would have made the investment anyway.
3.6 One, two, three – regulate and execute adoption of EDI! As mentioned in Section 3.5 there are different opinions regarding the effectiveness of regulation and interventions as means for directing action related to IT adoption. The examples presented above illustrate the mixed outcomes of projects that were initiated based on formal (e.g. TradeNet) or informal regulation (e.g. TEDIS and the 1996-action plan). None of the reported projects have led to substantial changes in behavior at the organizational level in relation to EDI adoption and implementation or to 84
industry wide diffusion. This raises the question whether regulatory initiatives can possibly result in voluntary changes in organizational behavior leading to adoption and diffusion based on political initiatives, or whether incentives for adoption are to be found in mechanisms driven by economic and strategic interests. In the following a number of considerations related to regulation are discussed. It should be emphasized that the initiatives presented above, including the 1996-action plan, in a legal sense are categorized as soft law.47 An analogy is however, made to considerations related to explicit rules and regulation in general. The significance of the preparatory works of legal acts should not be underestimated, because they can provide a clue about the intentions of the regulation and the underlying concerns and opinions. The “will of the legislator” is however, an abstract concept mainly because the legislator is not a single person (Ross, 1966). In the preparatory works of regulation three issues are of relevance (Zahle, 1989): The personal, the social, and the contents of a given set of rules. The personal issues are related to persons or interest groups who get the idea and who transform the idea into regulation. Of greater importance are however, the social issues that aim at explaining why the idea emerges and how it is possible to manifest the idea into a given set of normative expressed rules. Closely related to the social issues are the issues related to the contents of a given set of regulatory norms. The main question in relation to contents is related to why the ideas are manifested into one and not another particular set of regulatory norms. The questions to these three issues are multifaceted ranging from political factors, social and economic conditions, and psychological subjects (ibid.). The motivation for regulatory intervention therefore becomes an important issue. Legally binding laws in a Danish setting are well documented from the initial stages to the final set of rules qua minutes of parliamentary sessions etc. However, for soft law, which can be defined by a broad variety of interests groups ranging from ministries to professional business and 47
Soft law concerns rules of conduct that are on a legally non-binding level (Nielsen, 1999).
85
industry associations (Nielsen, 1999), it is more difficult to find causal factors related to the three issues: Personal, social and contents. To get the best possible insight into the concerns and opinions of the different interests groups involved in the process it is optimal to be present during the process and to have background knowledge of the driving forces behind expressed opinions. The present research project was initiated after emergence of the 1996-action plan. Those involved in the process of developing the 1996-action plan from the two trade and industry associations served as the primary informants in the research project. However, they were hesitant to reveal their influence and motivation for their contribution to the formulation of the 1996-action plan. A possible interpretation of the process could be that political considerations related to support of industry structures were pursued by members of the professional associations. Inherent in a set of rules is the means of regulatory intervention. Three types of regulatory intervention, the pedagogical, the economic, and the normative (Eckhoff, 1983) have been described as means for regulating organizational and human behavior. Though it has been stated that the Danish government did not see it as its primary role to pursue top-down steering or regulation of the EDI diffusion process (Andersen et al., 2000) the goal of the 1996-action plan was to fulfil the objectives described in the plan. Based on the three types of intervention it is in the following discussed, which one of the three interventions the initiators of the 1996action plan brought into play when they launched the action plan. Prohibition and commands characterize normative interventions. Normative interventions are means for enforcing the objective of the regulation. This type of interventions in relation to adoption of EDI leads for example to setting of standards (Andersen et al., 2000). The economic interventions are characterized by influencing what people find advantageous to do. Through the economic interventions organizations are rewarded to perform certain acts or punished for certain other types of behavior. Direct subsidy of EDI projects can be considered an example of an economic intervention (ibid.). The pedagogical intervention is
86
characterized by information campaigns initiated by governmental units and larger associations where the aim is to influence the opinion of a given group of potential adopters. Eckhoff (1983) argues that information is a must in most cases where the economic and normative regulatory interventions are brought into play, but information in itself can be used as a means of intervention. Through information it is possible to influence opinions and values thus making individuals more motivated for certain types of actions wished for by the governmental units or associations. Knowledge building hence becomes an important aspect in pedagogical intervention (Andersen et al., 2000). In the case of the 1996-action plan no normative interventions were introduced. Contrary to the TradeNet in Singapore, there was no regulatory enforcement with respect to the use of EDI. One natural consequence of the 1996-action plan could have been to force private businesses to report their financial statements to the national taxation bureau via EDI messages. This had been relevant since the 1996-action plan aimed at having the public sector as a locomotive for EDI adoption, which could have made EDI communication possible.48 Though financial resources were used as direct subsidies to support the diffusion of EDI in the Danish business community, these resources were limited and they were not available for any individual business. Instead financial support was provided to projects such as the TDP which was initiated by the two major Danish trade and industry associations (See Chapter 4). It can thus be argued that for individual businesses the economic incentives for adoption from a regulatory perspective were absent in the 1996-action plan. The pedagogical intervention was the sole strategy pursued by the initiators of the action plan. Through information campaigns by the Danish EDI Council and those associations that supported the 1996-action plan the advantages of EDI were communicated to potential adopters. This communication process thus becomes a central issue for the adoption and diffusion of EDI in the Danish business environment. 48
Research, which has looked at the adoption rate of EDI in the public sector, has however found that the adoption rate was slow (Andersen et al., 2000) thereby limiting the possible enforcement of electronic communication based on EDI.
87
One issue is how the message about the innovation is communicated. Another issue is how the innovation is presented. The message about the innovation can be communicated via mass-communications networks or through interpersonal relations (Rogers, 1995). In the 1996-action plan both types of communication channels were used. The Danish EDI Council and the professional business associations communicated the advantages of the innovation to their members via publications, newsletters and social arrangements. The second issue, how the innovation is presented, is perhaps of higher relevance in this context. Two different types of information can be conveyed: Signaling and know-how (Attewell, 1992). Signaling refers to information about the existence and potential benefits of a new innovation, whereas know-how refers to knowledge transfer in relation to the innovation. “The technical know-how is relatively immobile, and often has to be recreated by user organizations.” (ibid.). This places a heavy burden on potential adopters since they have to update and upgrade their skills and knowledge regarding EDI before adoption can take place. The first policy statement “From vision to action – Info society 2000” suggested that EDI adoption could lead to, “… considerable rationalization gains and a closer interplay between organizations”. In the 1996-action plan EDI was described as a tool that could give Denmark an, “... international lead along with improvement of efficiency of working procedures and development of new products and production processes”. Efficiency was an attribute, which was related to the innovation, whereas the traditional innovation attributes49 played a secondary role. Those statements formulated by the business community itself might however, be of even greater significance. The feature article presented in Textbox 3-1 (page 74) represents a quick two-page read-through of the initiatives in the 1996-action plan and the advantages of EDI. Mr. Agner Mark argued in favor of efficiency, cost reduction, and increased competitiveness resulting from adoption of EDI. The article left the impression that the investment
49
Rogers (1995) argues that the following five innovation attributes are of major importance for adoption of an innovation: Relative advantage, compatibility, complexity, trialability, and observability.
88
would pay off. This case indicates that the information conveyed about EDI focused on signaling rather than on know-how. The expenses related to EDI adoption were an issue that played a minor role in the policy statements. One of the seven initiatives in the 1996-action plan explicitly mentioned development of EDI software as an objective. However, the cost of hardware, integration50 of EDI software, and implementation of EDI in the organization were not mentioned in the policy statements.51 In a 1996 EDI survey among the members of The Confederation of Danish Industries the arguments in favor of EDI and the hindrances to the implementation of EDI were presented (Horlück, 1996). The large investments related to EDI adoption and the uncertainty about the actual advantages of using EDI were the two highest scoring items found as expressed barriers for EDI adoption. The third highest scoring item reported to be a barrier in the survey was related to the lack of standardization of the trade data standards. Apart from serving as a tool for creating awareness of EDI the 1996-action plan aimed at introducing the EDIFACT-standard in the Danish business community and the public sector. According to Horlück’s 1996-survey there seemed to be overlapping interest in the action plan and among the potential users in relation to standardization issues.52 The major incentives for adoption of EDI were related to issues such as quicker access to data provided by EDI, general cost reduction, automatic handling of purchase orders, invoicing etc., better quality of data, improved relations with customers and suppliers, and request from major trading partners to 50
It is generally acknowledged that organizations in order to take full advantage of EDI should integrate their internal systems with the external IOS (Riggins & Mukhopadhyay 1999). 51 A later publication from The Danish EDI Council, “Elektronisk handel og samarbejde – det betaler sig at bruge EDI” [“Electronic commerce and co-operation – it pays off to use EDI”], from 1999, elaborates on the expenses related to the EDI investments in an abstract way. The publication outlines the items that should be included in a cost-benefit analysis related to purchase of EDI. So, even though research had found that the costbenefit analysis was of limited value in relation to EDI (Bjorn-Andersen & Krcmar 1995) it was found that cost was an influential factor for EDI adoption. 52 The standardization issue in relation to EDI is however not an easy one to solve (Damsgaard & Truex 2000).
89
introduce EDI. Efficiency and the strategic role of EDI were according to the responders of major importance for the adoption of EDI. Based on the needs expressed in the 1996-survey and the message communicated about EDI in the policy statements there seemed to be some ideal agreement about which direction to take and which goals to reach.
3.7 Summing up Chapter 3 Governmental units and professional business associations did from mid 90s recognize that Danish businesses in order to stay competitive had to prepare themselves for the information society through a technological upgrade. One of the means for this technological upgrade was adoption and usage of EDI. A national action plan for electronic commerce was prepared. The action plan was ambitious in the sense that it presumed that adoption and implementation of EDI based on the EDIFACT-standard amongst private businesses and public administration would be possible within a few years. Similar initiatives have been launched in other countries involving similar technologies through the 1980s and 1990s. Some of these initiatives were mentioned briefly in order to compare the effect of these initiatives to the Danish initiative. A general impression from these projects is that their success is limited in relation to the rate of adoption and diffusion. The means and effects of intervention were discussed. It was found that the Danish initiative was driven by pedagogical regulatory intervention. Voluntary commitments and information campaigns characterize pedagogical regulatory interventions. One major issue in this context is whether or not it is possible to influence the business community to change administrative behavior and make substantial investments based on recommendations from governmental units and professional business associations. In the following chapter the ideas of bringing the 1996-action plan into practice are pursued.
90
4 Looking into an initiative launched by the industry and trade associations 4.1 Introduction The intention of the 1996-action plan was to support diffusion of EDI in the Danish administration and business community. By defining explicit recommendations in relation to adoption of EDI for the private and the public sector it was expected that adoption and use of EDI would reach a high level within a four year period (Ministry of Research and Information Technology, 1996). As previously mentioned, the action plan was formulated in concert with governmental entities, the major industry and trade associations and other interest groups. In order to follow up on the recommendations outlined in the action plan two of the major industry and trade associations in Denmark agreed to initiate a pilot project that aimed at creating ideal conditions for further adoption of EDI among small and medium sized businesses within manufacturing and wholesale. The pilot project named the TradeDocument Project (TDP) explicitly mentioned in the five objectives of the project (see Table 4-1, page 95) that it should be viewed as a contribution to the 1996-action plan. The TDP was a joint project between two of the major Danish industry and trade associations that represent approximately 7,800 small and medium sized manufactures and wholesalers situated in Denmark. Apart from the two industry and trade associations involved in the project the Danish EDI Council provided financial support. An IT-consultant was hired to manage the work in the project and to be in charge of programming a low cost EDI software and mapping an EDIFACT-subset that was to be developed
91
during the project. These players involved in planning and executing the pilot project are in the following referred to as the project initiators. The project initiators introduced the TDP in 1997. A group of managers in companies in the steel and machinery industry was invited to join the project. The purpose was presented as an opportunity to create ideal conditions for adoption and implementation of EDI among the members in the Danish steel and machine industry. The main aim of the project was to implement trade documents based on EDIFACT-standards. The means to this end was to create awareness of EDI, to provide the necessary resources, to develop a low cost EDI software, and to develop an EDI subset for the industry sector. The case study of the TDP is primarily used as an illustration of adoption and diffusion in an environment that has been massively exposed to the content of the 1996-action plan. The TDP case study thus becomes a means for examination of a group of organizations that were involved in operationalization of the 1996-action plan. As shown in Figure 1-3 (page 16) the theme for this chapter is a presentation of the case with the specific purpose of being able to make a qualitative assessment of research question 2. In the following sections data from the TDP is presented. The data reflects three stages in the project: Prior, during and after the TDP. The first stage, which is considered to be prior to the TDP, describes the overall aim of the project as presented in the project description distributed to the organizations invited to join the project. After presenting the project and after describing the project organization, the work that took place during the project is examined. The data explaining the work is mainly based on agendas and minutes of the meetings that took place among the participants over a period of 12 months. Thus far the data is only based on archival sources. Ex post data includes interviews with the participants of the TDP. Eight of the nine participating organizations are represented qua semistructured interviews. The chapter ends with an assessment of data from the case study. The sequence of the description is as close as possible to the observed chronology.
92
4.2 The TradeDocument Project In January 1998 the objectives for the TradeDocument Project (TDP) were introduced to managers in the steel and machinery industry. At that time fifteen companies agreed to participate. Three work groups were formed. In June 1998, when the practical work of the project started, only nine companies were still interested in participating and the number of workgroups were reduced to two. Instead of managers as originally intended, employees from IT- and accounting departments were involved in the project. The workgroups met on a monthly basis from June 1998 to January 1999. The project initiators and the two workgroups held a total of 24 meetings. The workgroup meetings were mainly used to discuss the EDI subset, but other issues such as legal and organizational topics in relation to EDI were also considered (Henriksen and Andersen, 1999). In the spring of 1999, the work related to the development of an EDI subset was finalized and the software was ready to be tested. The project finished by August 1999. 4.2.1 The preface of the TDP The pilot project on EDI TradeDocuments was launched at the beginning of January 1998. In an introductory letter written by the project initiators the purpose and objective for the project were outlined: “With the two major industry and trade associations as organizational backup an EDI project with the aim of defining EDI messages for trade among businesses and their suppliers primarily in the steel and machinery industry is established.” (introductory letter)
The introductory presentation of the project aimed at achieving the following goals: To create templates for common business documents such as invoices, order confirmations, credit notes and other commonly used business documents. Additionally, it was expected that the pilot project would serve as a model for other similar projects at industry level. The synergy effect of the project in the entire industry seemed to play a vital role for the success of the project:
93
“One outcome of the pilot project is that other pilot projects are initiated resulting in implementation of EDI for the support of transactions such as placing of purchase orders, order confirmation, and payments. Another outcome is that the lessons learned in the pilot projects are applied in the rest of the steel and machinery industry and adjacent industries. Based on the fact that there is wide support from a number of industry associations a high degree of diffusion is expected both among the members of the associations and among business partners in other areas. The diffusion is dependent on the degree of awareness created in and among the involved organizations, that are a part of the project.” (introductory letter)
Though not explicitly mentioned in the opening paragraph, the pilot project was meant to include more than just developing an EDI subset for the business sector. The above-cited paragraph indicates that the project initiators aimed at indirectly supporting the diffusion of EDI in and among a large scope of businesses. The diffusion process however, depended on the active involvement of the organizations participating in the pilot project. The project initiators defined the business value of the project for the involved firms and sectors in terms of rationalization gains both among the wholesalers and the manufactures due to automatisation and optimization of routines resulting from applying EDI in the organizations. Apart from these benefits it was expected that the implementation of the project would lead to substantial cost reductions due to simplification of routines among those organizations that had close relations. Finally, the business value was to be found in the development of a low cost EDI software that was affordable even for the smallest businesses. The project initiators defined five objectives that were a mix of tangible results and general EDI guidelines for the business environment in Denmark. The second objective (cf. Table 4-1, page 95) had the character of a general statement and it would be fulfilled the moment the project was launched. The objectives one and four, the development of standards and development of an EDI software were objectives where the outcome could be evaluated in relation to whether or not the products were available after completing the project. As the later assessment of the outcomes from the project will show the interpretation of whether or not the objectives, 94
number four in particular, had been achieved is ambiguous. Additionally, adoption and implementation in the individual organization is difficult to assess. The immediate evaluation of initiative three and five, the sharing of experiences and know-how and the cooperation across business sectors, is difficult to assess ex post because the measures for evaluation of those objectives are rather ephemeral and subjective, due to the nature of the innovation and due to the nature of business relations in a market economy.
Table 4-1. TradeDocument Project: Objectives # 1
2 3 4 5
Objective A thorough analysis of the existing EDI trade documents for instance the projects EAN/ HANCOM53 and VIBE54. If these projects are found useful to make the necessary changes to these documents so that they can be used as standard documents for the relevant business sectors. The EDI project is carried out to ensure that similar standard documents are developed and to minimise the cost for the enterprise that is going to use EDI. The main goal is to standardize the procedures and to reduce the total amount of work needed to implement EDI in any specific business. The project is the industry and trade associations’ contribution to the National EDI action plan. To launch pilot projects involving members of different business sectors and sharing experiences and know-how with other members of the respective business sectors. Development and quality testing of EDI software solutions in the various pilot projects. This software will be shareware and the selling price will be fixed so that even the smallest enterprises can afford to participate. This project will make every effort to cooperate with other relevant EDI projects within other business sectors for instance with the transport sector and other sectors directly involved with the current project in the Danish steel and machine industry sector.
In the preliminary project description the project initiators had defined five criteria for success, which to a large extent reflected the five objectives:
53
EAN/HANCOM is the European retailer’s common subset of EDIFACT, which is defined by the International Article Numbering Association. 54 VIBE is an electronic procurement system for the public sector that was developed under the governance of one of the involved business associations in the TDP.
95
Table 4-2. Success criteria for the TDP # 1 2 3 4 5
Success criteria To develop a number of EDI trade-documents suitable for the sector through reusing as far as possible already existing EDI documents, guidelines, standards etc. To launch a number of pilot projects which can lead to establishment of EDI cooperation between a number of organizations. To diffuse EDI to as many of the involved sectors’ members as possible. To develop a standard software – ideally as shareware. To establish relations to other EDI projects of relevance for the sector e.g. the transportation sector.
4.2.2 The project organization and participants The preliminary selection of participants for the project consisted of representatives from four industry associations, which in total represented approximately 900 wholesalers and manufacturing companies within the steel and machinery sector. The selection of businesses was primarily based on supplier/ customer relations. The involved businesses comprised supply of raw materials, semi manufactured articles, components, systems, process-equipment, process-software etc. The involved companies were located in different parts of the country to secure a wide geographical distribution. The project initiators concluded that in order to create a dynamic project group the number of participants should be restricted to 8 to 15 companies. If too many companies wanted to join the project a selection had to take place due to these restrictions. If too many companies wanted to join then a reference group would be established that would receive information on the TDP in order to be able to follow the developments in the TDP. The Danish EDI Council was invited to participate in both the steering committee and the project group – an offer that the Danish EDI Council accepted. Apart from being a part of the project the Danish EDI Council also provided financial support to the project. The grant from the Danish EDI Council was DKK 250,000. The practical work was to be done in the project groups with an independent IT-consultant as project leader. Two pilot groups were established. A condition for joining the project was that there was an established business connection among at least one of the other participants 96
in the work group, so that establishment of an EDI connection would make sense among the participants in the project groups. In each project group there was one main contractor and some subcontractors.
Figure 4-1. The project organization of the TDP
Project initiators Representatives from the two business and industry associations and the Danish EDI Council
Steering committee The project initiators An IT-consultant Representatives from: two of the participating companies two managers from the steel and machinery industry
Work group I 1 IT-consultant 1 main contractor 4 subcontractors
Work group II 1 IT-consultant 1 main contractor 3 subcontractors
The nine companies in the two work groups had a broad variety of sizes both measured in terms of number of employees and annual turnover at the time of initiation of the project. As shown in Table 4-3 below there were both larger companies and SMEs involved in the project.
97
Table 4-3. Main activity, number of employees and annual turnover in the participating companies in the two TDP work groups Main activity
Number of employees (1997)
Annual turnover in mill. USD* (1997)
Work group I Manufacturer 871 133 Wholesaler 856 326 Manufacturer 1226 260 Wholesaler 70 no data Manufacturer 15 no data Work group II Wholesaler 56 174 Manufacturer 244 236 Wholesaler 56 140 Wholesaler 92 251 Source: GreensOnline Business Relations, http://abon.greens.dk/ * Based on the exchange rate 1st of February 1999.
A brief presentation of the activities and the legal ownership of the nine companies is given below. To maintain the anonymity of the participating companies the following description of the tasks performed by each company is reduced to a generic description, where each company is labeled with the letters “A” – “I”. Company A is the first manufacturer mentioned in work group I and Company “B” the first mentioned wholesaler in work group I and so forth. Company A is one of the two main contractors in the TDP. Company A is part of an international group of companies operating within two main business areas: Marine boiler plants and related equipment and energy and environmentally friendly plants on shore. Company B sells and distributes heating and sanitary equipment, steel and other metals, tools, machinery and technical products, and also provides advisory services. At Company C the main activity is production of steel plates. Environmental problems are solved by transforming scrapped metal products into new steel plates and bar steel. Approximately 70 percent of the production is exported. Company C is part of one of the major industry groups in Denmark. The business activities in Company D are distribution and sale of industrial components and integrated systems within the fields of hydraulics, pneumatics, transmission systems, tools, filtration equipment and flow
98
technology. The company is an international industry group with branches in four Scandinavian countries. Company E, which is the smallest of the nine participating companies employing only 15 workers, is an order producing company. Company E is one of the three independent companies in the project. Company E produces and installs conveyance systems and custom-built machines and produces steel constructions. Company E also performs turn and mill tasks, sheet-metal work for example CNCcontrolled cutting and bending machines and stainless steel work. Company F is the other main contractor in the project. Company F produces water hydraulic products, including power packs, pumps, valves, motors, cylinders and accessories. The company is a subsidiary to one of the largest manufactures in Denmark. Company G is one of Europe’s leading producers of machinery for snow clearing and machines for clearing in general. Apart from that production the company makes machinery and equipment for construction. Company G is an independent company. Company H is a subsidiary of a multinational industry group. The company is mainly an outlet for the products of the parent company. Company H distributes pneumatic and hydraulic aggregates. Finally, Company I which is the third independent company among the nine participating companies. Company I is a wholesaler of hydraulic components, steel pipes and valves. Based on the information above related to the nine companies there appears to be rich opportunities for business relations in and among the two work groups. Figure 4-2 below illustrates only those business relations that were brought to light during the interviews with representatives of the participants in the TDP. A close examination of the customer database might show a different picture, since the informants most likely only remember the major customers.
99
Figure 4-2. Trade among the participating companies
Work group I
Work group II
A
I H
B C
G D
F E
In Figure 4-2 each letter represents a company. The dashed line separates the two work groups and an arrow pointing to a company illustrates that this company is a supplier to the company in the project. As illustrated in the figure business relations are more intense among some companies than others. But, there is some business relation between at least two companies in each work group. Only a single company penetrates the dashed line separating the two work groups. 4.2.3 The work In the following a description of the work undertaken in the project is presented. As mentioned above this description is based on the minutes of the meetings. The major weakness of portraying the work in the TDP qua minutes of the meetings is that the minutes of the meetings mostly reflect formal communication. The inner relations, tensions, conflicts and tacit assumptions that were present in the meetings are not directly reflected in the minutes of meetings. Minutes of meetings rather reflect assessments of past agendas, achievements, and future agendas. Another weakness is that the minutes of meetings provide a fragmented picture rather than a coherent picture of the process. One does not get insight into the interactions among the companies between the reported meetings. Figure 4-3 below illustrates
100
the actors and activities of the meetings and the activities that took place between the meetings.
Figure 4-3. Activities and communication during the TDP process
Actors and actions Project initiators Participant involvement Development of subset Softwaredevelopment
Time M1
M2
M3
M4 ... Mn
Legend:
M1
Formal articulated activity communicated through minutes of meetings Informal articulated activity not communicated collectively Formal articulated information communicated at meetings and through mail, newsletters etc. Meeting 1 (and so on)
M1 to Mn on the x-axis in Figure 4-3 illustrate meetings that took place during the TDP process. Four types of activities are presented on the y-axis illustrating actors and activities. Three different modes of activities in the TDP process are depicted in four differently labeled horizontal lines. The horizontal lines serve as an illustrative tool. The lines are not intended to exactly define time and space for communication and actions. The IT-consultant was responsible for development of the EDI software. He was also the coordinator of input from the TDP participants concerning the definition of the EDI subset. The software development was performed between the meetings and the progress was communicated at the meetings.
101
The subset definition was based on input resulting from the work meetings. Between the meetings the IT-consultant coded the input received from the participants. These two activities are well described in the minutes of meetings. The communication “informal articulated information not communicated collectively”, which is illustrated by the gray line, is never directly articulated in the minutes of meetings except in those cases where the project initiators communicate official messages to the participants either at the meetings or between the meetings. The hatched lines indicating that information communicated was selective illustrate the communication from the project initiators. There are however, good reasons to believe that the “Gray activity” influenced the TDP. An example illustrating this is presented in Textbox 4-1 (page 103). The project initiators distributed an introductory letter outlining the project in the beginning of January 1998. By June that year it was unanimously decided by the project initiators to start the TDP. It was agreed to make a job description for the software IT-consultant with special focus on issues concerning copyrights and ownership of the not yet developed EDI software solutions. A general description of the project was prepared by the project initiators including a financial budget and a time schedule. A website was established by the Danish EDI Council to publicize and promote the TDP. In July 1998 the first meeting in the two work groups took place. Two representatives, one from the IT-department and one from the accountingdepartment participated from each company. The IT-consultant suggested that the groups commenced the project by preparing EDIFACT-standards for invoices and orders. During August and September the two work groups met separately. The groups spend much time exchanging and sharing general information concerning the present status of hardware and software within the individual companies and the degree to which EDI was used. In both work groups a number of participants found it very important to focus on legal issues. Therefore, it was found prudent to make a draft for a legally binding contract for exchange of EDI messages.
102
In October 1998 a meeting was held for both work groups. Most companies involved in the TDP, the IT-consultant, one of the project initiators and his personal secretary, who did all the coordination of meetings in the two workgroups, were present. At the meeting the project initiator informed the participants that a general information letter about the project had been sent to the management of all companies involved in the TDP. At the meeting some of the members of the work groups mentioned that they considered the time schedule to be too tight. Furthermore, the work groups suggested dividing the groups into sub-groups. One sub-group was to focus mainly on technical issues and the other sub-group was to focus on legal, commercial and administrative issues. The rationale for this suggestion was a more efficient use of time and resources, especially the resources of the ITconsultant. A draft for an EDI standard contract was presented. The group concluded that no further work should be done on this contract since it is possible to use an already existing EDI standard contract based on EUstandards. Textbox 4-1. Narration of experiences from the October meeting of the TDP Personal experiences from the October meeting In October 1998 the initial negotiations between the two business associations, the Danish EDI Council and Center for Electronic Commerce at Copenhagen Business School had taken place. It was agreed that a research project investigating the usage of EDI and electronic commerce among the members of the two business associations was an interesting and promising research topic. It was hence suggested that the ongoing TDP should be followed closely. An opportunity to get insight into the project was to participate in one of the work meetings that took place in the TDP. The author was invited by the project initiators to join the October 1998 meeting as an observer. The following narration is based on the authors personal observations and experiences from the meeting. The narration, which is based on the author’s personal notes and subjective interpretations of the situation, leaves the impression that the participants from the large and small companies have different interests. It looks like the project leader hesitated intervening in conflicts and that the participants were uncertain about their goals. It was also observed that there was some resistance in the group to the idea of being a research object. The October meeting was a full day session from 9 a.m. to 4 p.m. where both work groups were invited for the purpose of making a status. The physical setting was a room on neutral grounds rented for the occasion. The room was furnished as a traditional meeting room with tables arranged in the form of a “horseshoe” and with a desk with different types of AV-equipment closing the horseshoe. The IT-consultant was appointed chairman of the meeting and he had occupied the desk with papers, binders
103
etc. He had arranged a laptop and a projector at a table inside the horseshoe. The participants seated themselves in a manner that at first glance seemed to be random. Men, women, young, elder, some casually dressed, and others in white shirts and ties mingled. The participants did not occupy any of the chairs close to the desk nor were those seats directly facing the desk taken. All the participants had brought various papers and a few had also thick binders in front of them. Fifteen people including the IT-consultant were present. The participants were small talking and exchanging papers. The activities were more intense at one side of the horseshoe. At the busy section heaps of paper were leafed through, documents were found in binders, remarks were exchanged low-voiced nodding to the IT-consultant. At the other side of the table the main activity focused on distributing coffee-cups, sugar and polite comments on the distributed agenda of the meeting. The IT-consultant had arranged his equipment nearer the “coffee-side” of the horseshoe and the participants involved him in their considerations related to the agenda. It looked like the “coffee-side” of the horseshoe was collaborative whereas the other side of the circle was more aloof. One representative from the project initiators, his secretary, and the author were seated close to the desk opposite the “coffee-side” of the horseshoe. The project initiator, who said that he could only be present for a short period of time to share some information, opened the meeting. (Here it is worth noting that the project initiator had traveled for more than three hours to get to the meeting). Before sharing the information he introduced the author. None of the participants had beforehand been informed that the author would participate in the meeting. Not all the participants were happy about the presence of an outsider. One of the participants directly said, “What is the purpose of having an academic snooping in the project?” The project initiator defended the presence of the author by saying that it could be valuable for everybody to get a neutral third party to look at the activities of the project with fresh eyes. The “coffee-side” of the circle agreed whereas the “aloof side” of the horseshoe appeared to be somewhat uncomfortable with the situation. The IT-consultant did not express any opinion. After this minor dispute the project initiator communicated the information he had come to convey. He explained that the two business associations had informed the management of the participating companies in the TDP about the purpose of the TDP. He also explained that the project initiators and the steering committee had reconsidered the scope of the EDI subset that was to be developed. No one made any comments. After this information the author was asked to present herself and her research project. Perspectives related to the role of the wholesalers in the future and the drivers and barriers for diffusion of EDI were presented and discussed. The reactions towards the author and the project were less negative after this presentation by the author. The meeting now proceeded as a regular work meeting. Focus was on the achievements related to the specifications of the EDI subset for the steel and machinery sector. Especially data-fields related to the invoice were discussed. The IT-consultant outlined what had been done and what needed to be done in relation to the development of the EDI subset. The participants, who addressed each other by their first name, discussed the input. The “aloof side” of the horseshoe was critical to the issues expressed by the IT-consultant. And after about an hour the “aloof side” stated that the project itself was too ambitious within the given time frame and that there were too many uncertainties regarding the legal status of EDI messages. It was also argued that it was waste of time to continue to work in the same manner as had been done so far. An alternative strategy
104
was suggested where the technical people would concentrate on the development of the subset and make optimal use of the expertise of the IT-consultant whereas the administrative people should be engaged in soft issues e.g. legal matters and strategies for introduction of EDI in their respective organization. The same group suggested that they should be in the technical group and they also right away appointed a person from the “coffee-side” of the horseshoe. The IT-consultant took note of the critique and said that he accordingly would change the agendas for the future meetings. The subset discussion was ousted by the legal status discussion. The “aloof side” of the horseshoe initiated the discussion. This discussion had obviously also been on the agenda on previous meetings. And one of the participants had prepared a draft for a legal EDI contract. He presented the draft, and received a few comments mainly regarding concerns related to uncertainties within the legal domain. Some of the participants from the “coffee-side” expressed their confidence in this draft since a large company, which had its own legal department, had prepared it. At this point of the meeting it became obvious that the “aloof side" represented the large companies and that there seemed to be a tacit agreement in the group that they represented the opinion leaders. One person had not joined any of the two sides of the circle. He had seated himself closer to the “coffee-side” but he had left some empty seats between himself and the rest of the group. He did not agree with the aloof group either on the idea on splitting the work group into two or in the legal draft. He suggested that the standard contract for EDI messages outlined by the EU should be used and that it was waste of time to spend more time on the legal discussion at all. Most of the participants were not aware of the existence of the standard EDI contract from the EU and the comment created some unrest. The person that had presented the proposal got annoyed by the situation and stated that his company’s legal department had looked at the EU-contract, but found it to be inadequate. Representatives from small companies and the “opponent” (who represented one of the largest companies in the TDP) argued that the EU-contract ought to be adequate since it was professionally prepared. After this discussion the atmosphere was rather tense. It was time for a lunch break. At lunch the two sides of the horseshoe split into two groups having lunch in two different rooms. The IT-consultant and the “opponent” joined the “aloof group” leaving no more space in that room for others. The rest of the participants including the author had lunch together. The atmosphere was friendly, the people exchanged EDI anecdotes with the author, they showed interest in the research project and they were interested in getting new insights related to EDI from the author. They expressed their concerns in relation to the difficulties the author might have understanding what was going on during the meeting. One or two at the table expressed that some people were very powerful and that they in their opinion were taking the project in a wrong direction. Nobody was really interested in discussing this issue and the conversation on that topic died. After lunch the IT-consultant took over. He presented the status of some of the datafields in the invoice EDI formula. Some of the administrative people discussed the appropriateness of including certain types of terms of delivery and terms of payment in the EDI subset for the steel and machinery industry. They used their practical
105
experience from their companies as reference. The involvement was half-hearted and none from the “aloof-side” paid attention to the ongoing activity. At 2:30 p.m. it was decided to close the meeting.
In November 1998 the IT-consultant distributed a new type of agenda for the meetings in the work groups. The IT-consultant had specified suggestions for the different items on the agenda. The IT-consultant mentioned that the reason for this change was that it gave the participants a better opportunity to prepare the meetings in the work group. There were two different agendas for the two subgroups within the work groups. The IT-consultant suggested that the technical work group focused their efforts on making minimum standards for the EDI documents. He suggested that the administrative group focused on general issues concerning the implementation of EDI in the respective companies. These issues included questions such as: What to do if your business partners reject the idea of doing business via EDI? Are the workers in the company informed about the implementation of EDI? Who will be responsible for the implementation? How are human resource issues handled? During the meetings held on the basis of the above-mentioned agenda the groups dedicated most of their time to discussion of the technical issues. Issues concerning the administrative matters were not discussed. They were at least not included in any of the minutes of the meetings. In December the IT-consultant distributed the new meeting agenda again. This agenda had the same structure as the previous one, and the IT-consultant suggested that the administrative issues were discussed again. However, in the work groups the participants were more interested in presenting the results reached on the technical part of EDI. And the issues concerning the administrative matters were still not discussed. The prevailing opinion in work group I was, that the project should be tailored to meet the special needs related to the business partners in the work group. While work group II maintained the original objective of the project, aiming at IT diffusion in the whole sector and in other industrial sectors. This discrepancy in perception of the objective of the project created tension in the groups and among the two groups.
106
In the beginning of January 1999 the last project meeting was held. Participants in the meeting were the representatives from the two industry and trade associations and from the Danish EDI Council. At that meeting it was agreed that further work on the project should be done by two of the three present in the meeting. It was found that what was missing was to make a general description of the project along with an invitation to a meeting for those interested in the outcome of the project. This proposed meeting never took place. Instead an external consultant was hired to write a coherent document based on documentation from the project. This 51-pages document, which is made available on the Danish EDI Council’s official web-site,55 gives a thorough introduction to EDI in relation to possible benefits of implementing EDI in the organization. The short term benefits described are related to operational benefits, for example less data errors, reduced physical paper handling, and avoidance of re-keying. The long term benefits of EDI adoption are related to strategic benefits, for example closer ties to business partners, improved service, and promoting efficiencies leading to a reduction in inventory and hence to less capital demand. A couple of successful cases are included to illustrate the operational benefits of EDI adoption. The document includes a detailed guide on how to get started with EDI including a standard contract for business partners who plan to establish an EDI connection. The document is also used as a promotional tool for distributing the EDI software that was developed during the TDP. It is possible to download a free test-version of the EDI software and the prices for the software are provided.56 At this point the commercial interests of the involved IT-consultant is evident. It becomes clear that the IT-consultant is biased in relation to EDI adoption based on the software developed in the TDP, since he has an economic interest in this very same software.
55
http://www.edi.dk/edi_4.htm
107
4.3 Interviews with the TDP participants The objective of the interviews was to evaluate the outcome of the TDP in relation to adoption of EDI and especially to assess the incentives for adoption in the participating organizations. In order to evaluate the project some evaluation criteria were formulated. These criteria resembled the original objectives formulated by the initiators of the project. The evaluation criteria were defined and formulated after studying the written documentation concerning the project, the authors’ direct observation during one of the work meetings, and after discussions with the representatives from the two industry and trade associations and the Danish EDI Council representative. A semi-structured interview protocol was prepared for the purpose of collection and evaluation of data. The semi-structured interview method was chosen, because this method allows the informants to elicit information on topics not included in a more rigidly structured format. The following presentation of data from the eight interviews is a mix of the authors’ interpretation of the obtained data and a number of quotations from the informants. Danish was the medium for all the interviews. The author has afterwards translated the quotations into English. 4.3.1 The actual use of EDI in the company The first question asked (cf. Section 2.4), was not directly related to the TDP, and therefore the EDI use in the organization is not necessarily based on the EDI software and the EDI subset developed during the TDP. The intention of the question, “What is the actual use of EDI in the company?” was to find out whether or not the TDP had led to further diffusion of EDI in the business sector and also to get insight into the general use of EDI among the involved organizations. The question also served as a general assessment of the adoption and diffusion of EDI in a particular business 56
The most basic version of EDItalk costs DKK 3,995 ( ≈ EURO 540). Apart from information on prices for the software prices for support and prices for integration to the organizations’ ERP-system is provided. The cost of integration into standard ERPsystems is DKK 4,995 (≈ EURO 670).
108
sector. Especially this question has contributed to a better understanding of the diffusion patterns within the group, because the informants shed light on their motivations for adoption or their reluctance to adopt EDI. At the time of the interviews five of the eight companies had adopted EDI. Four out of the five companies that use EDI already used EDI prior to joining the TDP. The fifth company adopted EDI during the TDP and started using EDI during the spring of 1999. The reason that this company adopted EDI was that its parent company suggested it. The adoption for this company can therefore not be considered to be a direct consequence of the TDP. The informant however, would not deny that the knowledge gained from the TDP had benefited the company in assessing and accepting the offer to adopt EDI from its parent company. The informant, an accounting manager, mentioned that he through the work on the project had gained more knowledge about EDI, and that this made him more confident that he was making the right decision. The remaining four companies that used EDI by the time of the interview had different motivations for their adoption. These motivations are described in detail in Sections 4.3.2 and 4.3.3 where incentives and barriers for adoption of EDI among the informants are reported. One striking attribute of the companies that used EDI at the time of the interview was that all the five companies were subsidiaries or part of an industry group. Some of them had been forced to implement EDI by their parent company or other partners in the industry group, while others were requested to use it. Three out of these five companies exclusively exchange EDI messages with the parent company or other partners in the industry group. It is worth noting that the two subsidiary companies have their parent company as their major supplier. In one of the industry groups, another business unit in the industry group is their biggest customer. Therefore, only two out of the five adopters used EDI with customers and suppliers other than their parent company or other units in the industry group. The two companies have a limited number of suppliers and/ or customers, with whom they actually exchange EDI messages compared to
109
the total number of suppliers and customers (cf. Figure 4-5, page 125). Since they exchange EDI messages with so few business partners they have not yet achieved any large-scale savings from EDI. Company B: ”If we include both the customers we exchange proprietary EDI and EDIFACT messages with then there are close to fifty customers – and we have approximately 25,000 customers. So if you think about it in that way then it is not that many. These customers do however represent a big share of our turnover. These fifty customers are our large customers. It is no secret that we receive almost one percent of the total number of order lines via EDI.”
According to the above statement the strategic benefits rather than the operational benefits dominate. From an operational perspective the savings realized through EDI transactions are marginal especially in the light of the resources invested in running both EDIFACT and proprietary formats. From a strategic point of view the company has as a result of EDI managed to establish close ties with their large customers. Another strategic benefit is expressed in the following statement: Company D: “We exchange EDI with only a few customers and that is very expensive for us. But we have decided not to consider expenses – we do not even calculate them. So we do not look at the economic consequences and say that the cost of handling an EDI message is DKK 100 to 150, whereas an ordinary purchase order in a paper format costs approximately DKK 40 to 50. You have to look at it in another way. We can offer a technological advantage to our customers. That gives us an innovative image.”
Similar to the above-mentioned statement from Company B, Company D reports that it does not have economic gains from EDI. However, Company D states that it was a good investment since it creates an innovative image of the company and since it is advantageous for the customers. The two companies do not exclusively use the EDIFACT-standard in their exchange of messages. Both companies have realized that in order to appeal to a broader clientele it is necessary to use proprietary standards along with the EDIFACT-standard. Both companies are wholesalers and have chosen different strategies in their efforts to include their suppliers (and customers) in an EDI partnership. One of the companies has tried to force their suppliers to adopt EDI. The other company has chosen to make it optional for their suppliers to use EDI. 110
Two of the three companies that did not use EDI at the time of the interview had expected that the TDP would have given them the necessary support to get started. They expressed some disappointment regarding the level of support from all involved parties in the TDP and in the fact that the TDP had not resulted in adoption of EDI in their companies. Company I: “If all the efforts invested in this project are not to be wasted then he [the IT-consultant] has to offer us more support and help us a bit more in order to get started. The level we are at right now is not taking us any further. There is also another thing I would like to say. I know [company xx] is not going to know this, but [company xx] invited us to join in the TDP and I had expected that they had taken the role as primus motor. But, it is not my impression that they have lived up to that role. They had already their own EDI system and they just remained passive. That disappointed me, especially because I believe they both had the expertise and the resources.”
As the above statement illustrates the non-adopters do have some frustrations both in relation to the other participants in the project and in relation to the possible outcome of the project. The informant from Company I wants and expects more support. A common characteristic of those companies that do not use EDI yet is that none of them are subsidiary companies or part of an industry group. Two of the companies are producers and the third is a wholesaler. The three non-users mentioned that they gained some but not enough knowledge during the project to feel comfortable adopting and implementing EDI. In the following sections reasons and reflections related to EDI are presented for each of the eight informants. 4.3.2 Reasons for adoption 4.3.2.1 Company B Company B is among the EDI champions in the Danish business environment. The company has used EDI for about 15 years. The incentive for adopting EDI was that the company wanted to disseminate the use of IT in the steel and machinery sector. Instead of being passive the company therefore established its own IT-department that developed EDI solutions, which they offered to other companies in the sector. Company B realized after a while, that selling EDI solutions to their customers was not ideal. As 111
the informant commented: “To sell EDI solutions to our customers was not the smartest thing to do.” Instead they sold the product and goodwill to an established IT company. Digging deeper into the incentive for developing and distributing EDI solutions to the business partners the informant replied: “First of all we realized that there was a competitive advantage at that time. Another reason was that we could give our customers a better market position with the result that we ourselves would get a better market position.”
Being one of the major wholesalers of steel components in the Danish market and thus being in a dominant market position made a situation for Company B where it had the opportunity of becoming an EDI champion (Grover, 1993; Schon, 1963). With a dominating position in the negotiation process, the EDI champion is able to convince or even force their small partners to invest in EDI capabilities (Hart and Saunders, 1998). Many authors have noted that EDI can benefit its champion but doubted that EDI adopters, who are often coerced to implement the EDI by the champion, can gain similar pay-off from their EDI investments (Lee et al., 1999). Whereas the short-term benefits for Company B were obvious in relation to market positioning and competitive advantages it was realized that the benefits accruing from introducing EDI to its business partners were marginal: “Yes it was [a competitive advantage] in the short run – but not in the long run. You can not use the advantages in the long run, because everybody can do this. Everybody can develop EDI solutions and everybody can exchange EDI messages – if they are forced to … We were modern in the beginning and we got a good image among our business partners. However, after some time our competitors were also capable of introducing EDI, and by that time the advantage was gone.”
To sum up, for Company B the incentive for EDI adoption was to maintain or perhaps gain market position. Strategic positioning of the firm due to use of EDI is in previous research recognized as an incentive for adoption (e.g., Bergeron and Raymond, 1992). The company achieved a competitive advantage in the short run, but due to the general diffusion of EDI in the business environment the competitive advantage has not been sustained in
112
relation to an innovative image. It is worth noting that the informant commented on the general developments in the IT environment by saying: “But EDI is not in fashion anymore, nowadays it’s called Internet or other exotic names.”
4.3.2.2 Company C For Company C the incentive for adoption was quite different from the case of Company B. Company C is part of one of the largest industry groups in Denmark. As a manufacturer of unique products the power from customers is insignificant because there are none or few alternative suppliers of that particular product. Even though the international market on raw steel has adopted EDI to a large extent (Henriksen, 2001) Company C has not been confronted with pressure from suppliers of raw materials for its production. Company C was encouraged to adopt EDI from one of its largest customers within its industry group. Even though it was out of the question that Company C would be impositioned by its trading partner (Iacovou et al., 1995) if they decided to reject the ‘offer’ to adopt EDI the informant from Company C admitted that the consequences of a negative response was considered. According to the informant the primary cause for adoption was: “The reason [for the adoption] was that they are one of our largest customers. We sell a unique product. It is not commodities taken directly from the shelf. The customer wanted to have the purchase information electronic so that they could avoid re-keying. They had to re-key data when they received an order-confirmation from us and when we booked their order we re-keyed data from computer generated lists. The major issue was not the cost of re-keying - the problem was generating dataerrors. Data-errors that might lead to disturbances in the production process. By entering data only once a lot of time consuming control processes were avoided.”
The major reason for adoption was thus improved operational performance. One of the things that characterize Company C is that they solely produce on demand, and that they do not have any product codes. This feature has created a number of obstacles when implementing the EDI solution with the customer, but the solution has led to a substantial reduction in wrong 113
deliveries. Apart from that it is recognized by Company C that the use of EDI has improved its business relations with its customer. The successful EDI project with this customer has been the motivating factor for Company C to suggest EDI to other of its customers: “It turned out to be beneficial [to exchange EDI messages]. One thing was that we achieved closer ties to the customer, and we did in addition gain a lot both in relation to order processing and avoidance of errors. So, we started another project, because we thought - if we can do it with one customer then why not try with some of our other customers. We have the opportunity to be first-movers… We wanted to set the standard. We already had a suitable standard so we were in a win-position.”
Company C initiated a project with one of their other major customers based on the standard that had proven to be suitable. Two years were used on negotiations and programming to get the EDI system running. Before commencing the project the customer went bankrupt. This experience has resulted in reluctance from Company C to suggest EDI to business partners due to the sunk costs in that particular project but also due to uncertainty about future developments in e-commerce. The appropriateness of staking everything on a single customer and considering the efforts as sunk can be discussed especially in relation to EDI. The company has chosen to use a proprietary EDI format, which reduces access to a broader audience. The company is however, in a unique position due to the lack of competitors in the market and their bargaining power should thus be optimal. However, the following quotation reflects considerations that transcend mere annoyance with sunk costs. “Our intention was without doubt that we would broaden the project to include all our customers. The intention was that the initial project was the first step in a chain of projects. And the chain has not slipped off – but the wheel is turning slower, partly because of our disappointment with the customer that went bankrupt and partly because we do not know if we are doing the right thing. You know e-commerce and that type of developments.”
Similar to Company B, Company C gives some attention to the ecommerce developments. In a historical context this makes sense. In August 1999 when the interviews were recorded Internet-based e-
114
commerce was one of the business imperatives in the business press and among consultants (Timmers, 1999). One interpretation could therefore be that the informants were influenced by those trends. From a broader diffusion perspective Company B and Company C are examples of organizations that are willing to set the standard being innovation champions. The significance of the innovation champion in the business community has been found to play a major role in diffusion of an innovative idea (Schon, 1963; Van De Ven, 1986). Apart from taking the role of champions and thereby taking the lead Company B and Company C are exerting some degree of power to influence the adoption process among their business partners (Hart and Saunders, 1997; 1998; Iacovou et al., 1995). Their own direct experience with EDI turned out to be a success leading to savings in relation to less re-keying and reduction in errors. This success inspired Company C to encourage other business partners to exchange EDI messages with them. Company B has not experienced direct savings, but they are in a good strategic position due to long-term experience with both EDIFACT and proprietary EDI. 4.3.2.3 Companies D and H Company D and Company H are part of international industry groups. It has been beyond discussion for both companies whether or not to adopt EDI. As expressed by the informant from Company H: “That type of initiative just come from above.”
Pressure is the best way to characterize the adoption of EDI in Company D and Company H. Whether the pressure should be denoted as persuasive or coercive power (Hart and Saunders, 1998) is a matter of interpretation. Compared to the cases of Company B and Company C it is however relevant to emphasize, that the risk of adopting EDI for Company D and Company H is much smaller because the parent company takes the risk of sunk cost in relation to development and implementation of EDI in the first case.
115
4.3.2.4 Company F The fifth company among the informants that had adopted EDI at the time of the interview is Company F. Company F adopted EDI during the TDP because of a suggestion from its parent company. The adoption-decision was related to expected savings since Company F receives forty percent of its supplies from its parent company. “The adoption was voluntary, but in relation to savings it would be wrong not to adopt EDI. We have, however, not felt ready until now.”
The readiness referred to by the informant was related to organizational readiness, financial readiness, and technological readiness. Especially the organizational readiness revealed an interesting aspect. The informant explained that one of their sales mangers did not like the idea of EDI. The sales manager felt that adoption of EDI was a threat to his position, he had therefore done whatever possible to delay the decision. Company F has already gained substantial savings from less re-keying and handling of physical documents and has no doubts in relation to the achievement of operational benefits due to EDI use. Even though it is recognized that the cost of adopting EDI has been high and that it has taken a lot of resources it is the general impression that it has been worth it. “It has been hard work to integrate EDI with our administrative systems and during the process a lot of modifications of systems set-ups have been made. It requires a lot of resources.”
Company F is an example of how operational benefits can be achieved (O’Callaghan and Turner, 1995). Even though the company has been subject to some degree of persuasion from the parent company this has not hindered substantial cost savings. There are however also strategic gains. The resulting reduction of workload in the accounting department has not led to lay-offs. Instead the resources have been reallocated to, “… more intelligent logistics tasks aimed at improving sales and customer services.” 4.3.2.5 Summary of reasons for adoption Summarizing the section on incentives for adoption leads to the following conclusions:
116
- Strategic benefits played a more important role than operational benefits in relation to the adoption-decision. - Cost was not considered to play a significant role. - The incentive for being a first-mover is the opportunity to set the standard. - An incentive for adoption is to get an innovative image. - Being a parent company or part of an industry group is positively related to the adoption of EDI. - The companies that are part of an industry group or subsidiaries have adopted EDI due to encouragement or pressure from the parent company or other companies in the industry group. 4.3.3 Reasons and reflections of non-adopters Three of the eight companies interviewed had not adopted EDI at the time of the interview. The comments below present reasons and reflections that these non-adopters expressed. Additionally, the potential motivators for adoption expressed by the non-adopters are included. 4.3.3.1 Company E Company E, which is the smallest of the participating companies in the TDP, has not yet adopted EDI. Directly asked why not the informant answered: “The reason is that none of our suppliers has demanded it… Even though we are among the smallest steel producers our goal is to have all our IT totally integrated. The pay-off is low if EDI is not fully integrated. To integrate EDI is expensive. We are not yet willing to spend that amount of money. However, if the need arises then we will make the investment.”
Company E is aware that it can be a target for pressure from suppliers. The company is at the same time knowledgeable about the challenge that adoption of EDI is in relation to integration into existing administrative systems (Massetti and Zmud, 1996). The response is mixed in the sense that the integration issue plays an important role due to lack of necessary IT-resources. Another argument is also given for the reluctance for adoption. The informant does not expect that the company can achieve a critical mass of business partners. So the informant is not convinced that 117
the investments in EDI will pay off. Finally, the company produces highly specialized and custom-made components, so they do not think that EDI messages are suitable for their company. “EDI is only attractive to us when we have several orders with many units.”
To sum up, Company E puts forward the following reasons for its status as a non-EDI user: No external pressure, cost, lack of critical mass, and finally the company does not find that its products are suitable for EDI. The barrier for Company E thus seems to be that there is no incentive in relation to improved performance resulting from adoption of EDI. It seems as if operational considerations play a more important role than strategic considerations. 4.3.3.2 Company G Company G is an order-producing company. The reason for being reluctant to EDI adoption is that it has not been possible for the accounting unit to document possible benefits resulting from usage of EDI. “We bought new IT-equipment approximately 2 years ago. I applied for financial support to purchase of EDI software on that occasion… However, it was not possible for me to document [to the management] that the use of EDI would lead to savings. We have two suppliers who want to cut prices if we exchange messages via EDI.”
Those two suppliers were ready to give a 2½ percent discount on all purchases. This was however not enough to convince the management of Company G that EDI would be a good investment. One argument is that only a few of the suppliers are using EDI and so far only mild pressure has been exerted on Company G in order to adopt EDI. “Our bolt supplier, has expressed an interest in exchanging of EDI messages – but they will of course not force us, because we can ourselves decide where we want to place our purchase orders. What they have done has been to influence us by offering discounts. That means economic incentives – and that is often the best carrot.”
The management of Company G has not been convinced that those discounts can yield a reasonable rate of return on the investment, which the
118
company expects to be approximately DKK 100,000. The informant, the accounting manager, agrees with that viewpoint. Contrary to the two other non-adopters Company G uses a standard administrative system, Concorde XAL, which can easily be integrated with a standard EDI module. The informant recognizes that purchase of the software is not enough and that adding new components to the system might create problems. None the less the account manager states that the accounting unit would be pleased to purchase and implement the EDI software, even though, he doubts whether there is a critical mass due to the general low EDI usage in the sector. Company G has, similar to Company E, the problem that it is a producer of custom-made products. Therefore, several of the components that are purchased for its production are unique. It is therefore felt that EDI is not suitable for the company. The case of Company G illustrates the ambivalence towards EDI in relation to the rate of the return of the investment. On one hand the account manager pursues the idea of adoption of EDI by applying for the necessary financial resources for purchasing standard EDI software for the administrative system, which enables integration of EDI messages. On the other hand the account manager admits that he doubts whether this investment will pay off taking into account critical mass and the number of transactions that can be executed via EDI. Another theme which is well known in relation to adoption of technological innovations is the importance of management support (Van De Ven, 1986; Rai and Yakuni, 1996; Premkumar and Ramamurthy, 1995; Ramamurthy et al., 1999, Sabherwal and King, 1995). In the case of Company G there is no management support. This lack of support has so far, for Company G, been the prime inhibitor for the adoption of EDI. Company G has the same reservations towards adoption of EDI, as Company E. Company G does not expect that the EDI investment will pay off. Furthermore, none of its business partners have so far put enough pressure on the company to adopt EDI.
119
4.3.3.3 Company I The last of the eight interviewed companies Company I is represented by the account manager of this wholesaling company. The informant states that the reason for the lack of EDI adoption is considerations related to cost. For years Company I wanted to adopt EDI, but it has found that the cost was too high especially in relation to the integration of EDI into its existing administrative systems. This integration is a perquisite for adoption of EDI. Company I gets its software developed ad hoc by an ITfirm that supports a number of businesses within the steel and machinery sector. The IT-firm has for a long time ignored EDI - to the annoyance of Company I. Even though there are a number of reservations in relation to EDI the informant has no doubts that the company in the future will make use of EDI: “I am quite sure that we will be using EDI within the next year. That is our plan.”
Directly asked if this decision was due to pressure from business partners the accounting manager replied: “No, that is not the case. Nobody has put any pressure on us – that is very “un-Danish”. However, you have to be careful not to miss the opportunity of being an attractive business partner. We have already announced for a couple of years that we wanted to adopt EDI. We can’t continue to postpone it. We feel it is the right time now especially since some of our business partners have expressed that they wanted to exchange EDI messages with us.”
One of the explanations for the low pressure from business partners could be seen in relation to the high degree of specialization of the products that Company I deals with. There are few or no alternatives for the customers. Even though Company I’s market position is good the informant still expressed anxiety in relation to loss of market position if EDI is not adopted within a short period of time. The company expects that the adoption of EDI will lead to savings. “… less re-keying etc. It would be an advantage, and I hope we can realize some savings in these processes – because we would like to see some returns on our investments.”
120
The incentive for Company I to adopt EDI is thus to maintain market position and to realize savings. The barrier for EDI adoption has been the substantial investment needed. Company I is different from the two nonadopters since the first steps towards adoption have been taken though no authoritative or financial commitment have yet been made. 4.3.3.4 Summary of reasons and reflections of non-adopters Summarizing the section on reasons and reflections of non-adopters leads to the following conclusions: - Small independent companies are not adopting EDI. - External pressures are the most important incentives for adoption. - The cost of EDI equipment is found to be too high. - The business activities are not suitable for exchanging EDI messages. - There are not enough business partners that are interested in exchanging EDI messages with small companies. - The adoption of EDI does not lead to substantial savings. - Integration into existing ERP-systems is too costly. 4.3.4 Summing up motivators for adoption of EDI among the TDP participants In Figure 4-4 below the nine companies are represented by nine squares each named with a single letter. The dashed line illustrates the division of the two work groups. Block arrows for non-adopters include motivators for adoption and barriers for adoption. For adopters and Company I, which has decided to adopt in the near future the block arrows only contains motivators. The figure illustrates the significance of the different kinds of pressures both in relation to direct pressures and indirect pressures. Company B and Company C wanted to be in a position where they could put pressure on their business partners. Especially Company B wanted to be in a position where it by developing and distributing EDI software based on a proprietary standard could set the standard. Company C took the opportunity to be a first mover in its industry segment. This lead for Company C was considered to be a chance to influence the development in
121
the sector. Company B, however, invested substantial resources in its own IT-department, whereas Company C has out-sourced all development and maintenance activities related to IT. Figure 4-4. Motivators and barriers for adoption of EDI among the TDP participants
Work group I
Work group II
no data available M: Market position gain M: Wanted to set the standard M: Wants to set the standard
A
M: Maintaining of market position M: Savings
I
B C
M: Pressure from parent company
M: Decision from parent company
H G
B: Cost B: Critical mass B: Lack of management support M: Pressure
D
B: Cost B: Critical mass M: Pressure
E
F
M: Pressure from parent company
Legend = Non-adopter = Adopter = Adoption has been decided
M = Motivator for adoption B = Barrier for adoption
Companies D, F, and H were just told by top management to adopt and implement EDI. For Company H and Company D it was a decision beyond discussion. In the case of Company F no coercive force was exerted from the parent company. The company was encouraged and the necessary resources were provided by the parent company leading to a smooth adoption and implementation process for Company F. Companies B and C were in a position where they wanted to exert pressure on their business partners. They were first-movers and they wanted to set the standard in the business community. Two of the non-adopters, Company E and Company G are in a waiting position. They will take action in case their business partners should force 122
them. The two share the opinion that the investment will most likely not pay off due to lack of critical mass. The third non-adopter, Company I has decided to adopt. The decision of Company I to adopt EDI is mainly influenced by considerations related to a possible loss of market position. Table 4-4. The role in the adoption process among the participants in the TDP Status Role Active
Passive
Adopter
Non-adopter
The incentive was to set the standard and keep or gain market position. “B”, “C” It just happened due to a decision or a pressure from the parent company. “D”, “F”, “H”
To avoid loss of market position and realize savings. “I” Nobody pressured the company. “E”, “G”
Table 4-4 illustrates the role of the TDP companies in the adoption process from a power perspective. The roles can be classified as either active or passive. The active role is characterized by an urge to influence the process in order to maintain or gain market position. As for the late mover, Company I, it is more a question of not loosing its share of the market. The passive role is characterized by an acceptance of pressure being exerted on the company. The active and passive roles resemble the labels proactive and reactive often used in the IS literature (Premkumar and Ramamurthy, 1995; Chau and Tam, 1997; Swatman and Swatman, 1992). Iacovou et al. (1995) found in their study of seven companies that the adopters could be classified in a similar way. They had the active adopters labeled as “ready adopter” and “EDI initiator” and those playing a passive role in the adoption process labeled “unprepared adopter”, “coerced adopter”, and “unmotivated adopter”. Even though the “ready adopters” are pressured to adopt EDI by trading partners they were already prepared for adoption. According to Iacovou et al. ready adopters have the necessary resources to integrate EDI with their existing computer applications and they thus have the potential of realizing benefits from the investment. Company I can be characterized as a ready adopter in the sense that they
123
have allocated the necessary resources for adoption and implementation of EDI. Company B and Company C can be characterized as EDI initiators. Especially Company B has taken the role of being EDI initiator in the business sector. EDI research has shown that adopters that play an active role in the adoption process view the perceived benefits of the investment and the network externalities in a positive way (Riggins et al., 1994; Wang and Seidmann, 1995). 4.3.5 Degree of EDI usage among the TDP participants Two of the non-adopters expressed reservation about EDI due to critical mass issues. If the motive for adoption is to achieve savings or improved performance as a consequence of EDI use (O’Callaghan and Turner, 1995) then the optimal benefit of EDI is without doubt achieved if and when the technology is used among a large number of business partners (Jones and Beatty, 1998; Premkumar et al., 1997; Iacovou et al., 1995). Based on data from the informants this critical mass is far from reached. In Figure 4-5 the nine companies are illustrated in the form of nine squares each named with a single letter. The dashed line illustrates the division of the two work groups. Block arrows are included for the adopters listing EDI format used (EDIFACT and/ or proprietary) and scope of EDI use with customers and suppliers. The dotted line between Company A and Company B illustrates an established EDI link. This is the only EDI link amongst the nine companies in the TDP. Figure 4-5 illustrates that companies C, F, and H are exclusively exchanging EDI messages with a single business partner. The number of business partners with whom EDI messages are exchanged are higher for Company B and Company D. However, the degree of EDI use for Company B and Company D is still marginal. The low utilization of the investment is not surprising compared to other surveys on EDI usage among businesses in Denmark (Andersen et al., 2000; Horlück, 1996). In a survey of the entire steel and machinery industry in Denmark made by the author in year 2000 the ratios of EDI
124
messages to paper based communication both among customers and suppliers were higher than other reported studies have shown.57
Figure 4-5. Degree of usage and EDI transaction format for the TDP participants Work group I
No data available Format: E & P Scope: Customers: 50/25,000* Suppliers: 30/6,000** Format: P Scope: One company in concern exclusively Format: E Scope: Customers: 2/5,000* Suppliers: 5 /50**
Legend: E = EDIFACT P = Proprietary format
Work group II
A
I
B
Format: E Scope: Exclusively parent company
H
C
G D
F E
= Non adopter = Adopter
Format: P Scope:Exclusively parent company
= EDI link * = EDI customers/ total number of customers * * = EDI suppliers/ total number of suppliers
A more thorough analysis of the degree of EDI usage should include more dimensions than the number of business partners that exchange EDI messages compared to the total number of business partners (Massetti and Zmud, 1996). Andersen et al., (2000) redefined the four dimensions suggested by Massetti and Zmud and based on these four dimensions operationalized four measures to a “VIDS-test”.
57
The actual statistics are reported in the publication “Elektronisk handel og EDI i praksis: Anvendelsen i jern- og maskinindustrien i år 2000” [“Electronic commerce and EDI in practice: Usage in the steel and machinery industry in year 2000”] written by H. Z. Henriksen and published by the Danish EDI Council.
125
Table 4-5. The four VIDS-measures for EDI usage Variable Definition Volume The magnitude of EDI-based traffic relative to all structured messages. Integration Back-office integration of EDI applications with other functions and document handling systems within the organization. Diversity The number of different types of EDI-messages relative to the total number of different document types. Span Number of co-operating organizations that apply EDI with the organization. Source: Andersen et al., 2000, p. 46.
By applying the four measures from the VIDS-test58 a more informative picture of the utilization of the technology in the five companies that have adopted EDI is revealed. Table 4-6. VIDS-test of EDI users in the TDP Company Volume B Customers: 1 (3) percent* Suppliers: 30 (20) percent* C Customers: 10 percent Suppliers: none
Integration Diversity Full Purchase order Order confirmation Invoice
Span Customers: 50 Suppliers: 30
1 Purchase order Order confirmation Invoice D Customers: < 1 percent None Purchase order Customers: 2 Suppliers: 50 percent Invoice Suppliers: 5 1 F Customers: none Full Purchase order Suppliers: 40 percent Order confirmation Invoice 1 H Customers: none Partial Purchase order Suppliers: 85 percent Order confirmation Invoice * The first number indicates the percentage of purchase orders sent/ received via EDI. 59 The number in parenthesis indicates the percentage of invoices sent/ received via EDI. Full
58
The test has been modified compared to the original VIDS-test with respect to quantification of each of the four measures. Instead of using qualitative data levels (low, medium, and high) data supplied by the informants are included. 59 Some of the customers find it convenient to receive the invoice electronically but they are not capable of generating an EDI order from scratch. Some of the suppliers are able to receive orders, perhaps due to pressure from customers, but they have not invested in the necessary equipment to handle EDI messages in the organization.
126
In the following sections conclusions related to data from the five TDP participants that have adopted EDI and the four dimensions in the VIDStest are presented. One should exercise caution when reading these conclusions due to the limited number of cases included in the analysis. The VIDS-analysis suggests that the coefficient of utilization of the innovation is limited. This somewhat supports conclusions related to “assimilation gaps”60 (Fichman and Kemerer, 1999) and the “productivity paradox”61 (Hitt and Brynjolfsson, 1996). From a non-economic point of view there may still be benefits that cannot be measured and interpreted as direct pay-off in relation to the resources invested. Volume is the first dimension in the VIDS-test. Companies B, D, F, and H are benefiting from their investment in EDI with their suppliers by exchanging thirty percent or more of all purchase orders through EDI. All these four companies are wholesalers. Whereas the operational benefits are relatively high in relation to suppliers they are marginal or non-existent in relation to customers. A comparison to another VIDS-analysis (Andersen et al., 2000), which included both customers and suppliers, should be done with caution. Andersen et al. assessed that the volume dimension in the wholesale segment was “medium to low”. In the case of the TDP participants the assessment is similar if both customers and suppliers are included in the assessment. Only one manufacturer, Company C, is among the EDI users in the sample. Company C’s operational benefit of EDI is estimated to be low, since EDI is only used with ten percent of its customers down-stream in the supply-chain. Integration is the second dimension in the VIDS-test. One of the companies reported that they had no integration of EDI with their ERP-systems whereas three of the companies reported that they had full integration. There does not seem to be any dependence between degree of integration 60
Assimilation gaps refer to the discrepancy between acquisition of an innovation and deployment of the innovation. 61 The productivity paradox refers to the fact that productivity growth has slowed every decade since the 1960s while investments in information technology have grown dramatically. Some take this as proof that information technology does not affect productivity.
127
and transaction volume. Company H which has an extremely high volume on the supplier side has no integration whereas Company C, which has a low transaction volume both in relation to customers and suppliers, has invested in full integration. Diversity is the third dimension in the VIDS-test. The diversity of different EDI documents is low for all five companies. It has been argued (Premkumar et al., 1994) that in order to utilize the electronic links fully it is necessary to expand the communication with business partners to include more types of documents. Only the most basic business documents are included in the EDI exchange amongst the TDP organizations. Span is the fourth dimension in the VIDS-test. Three of the five companies have the lowest possible span. They are only exchanging EDI messages with a single business partner. Compared to the total number of customers and suppliers Company B and Company D exchange messages with a marginal number of business partners. It is however, worth noting that though the span is limited the volume is relatively high. This indicates that the suppliers with whom EDI messages are exchanged are among the most important suppliers. The strategic role of EDI thus seems to play an important role for the adopters. After this detailed reporting of usage, motivations for adoption, and reasons and reflections for non-adoption of EDI among the TDP participants focus is shifted to the tools meant for promoting adoption of EDI in the Danish steel and machinery industry. The following sections focus on assessing the TDP in relation to the development of the EDI software and EDI subset for the industry sector that were formulated as objectives of the TDP. 4.3.6 Degree of adoption of the software developed as part of the TDP One of the main objectives of the TDP was to develop low-cost EDI software. The IT-consultant was responsible for this task. At the time of the
128
interview, the software was fully functional and the IT-consultant had installed it in the companies (companies G and I62) that wanted it. Four of the interviewed companies would not use this software under any circumstances. A common characteristic among those who rejected to use the software was that they already used EDI. All of them found that the software did not meet their requirements and needs. A puzzling feature surfaced during the interviews. Those who had declined to use the software did not agree on which tasks the software actually fulfilled. One of the participants, a head of an IT-department in one of the large participating companies, made the following comments regarding the developed software: Company D: “I don’t know much about it. We are not interested in it at all. Because we can not use it. We will never use it. However, if I have understood it correct then it is some kind of a “Concorde software” – that is some kind of software for a standard system. So, I don’t know anything about it… In my understanding it is a module, which you can add to your administrative system – some kind of converter. But to be honest I have not paid much attention to this part of the project, because it had absolutely no relevance for our company. We would never use such a simple software solution in our company. We have already bought the converter we need.”
A head accountant from another participating company expressed doubts about the functionality of the software: Company F: “I have only seen it demonstrated once. I have no idea how it’s going to work in practice. But, if the software depends on manual keying then its success depends on the kindness of the users. I do not see any rationale in keying data from one system to another – of course depending on the complexity and the number of transactions. But, then again where are the savings?”
The two above-mentioned statements illustrate different aspects of refusal to use the TDP-software. The informant from Company D does not express the slightest interest in which tasks the software fulfills. Even though the 62
The two companies explained that they had the software installed in order to get some experience with EDI. The two companies had decided to explore EDI with each other, they were however not interested in applying the software in their ordinary business routines.
129
informant represents one of those companies that can be characterized as advanced EDI user. However, the statement also clearly illustrates that the participant has paid an absolute minimum of attention to what has been going on with regard to the EDI software during the meetings, where he after all was in an ideal situation to understand what was going on. The informant from Company F on the other hand doubted the functionality of the software and he doubted whether this system would realize any savings. The quotation below also illustrates uncertainty and ambiguities with respect to the software and the project in general. On one hand the functionality of the software is questioned. It is against the policy of the organization to adopt software that is not integrated in existing systems, and since the TDP-software is not a ‘plug n’ play’ software the solution is not suitable for the organization. On the other hand it is acknowledged that the software provides a good opportunity for the organization to join the EDI community due to the marginal investment. Company I: ”The software has been installed in our company, and I would use it if it was required. However, as long as nobody requires it I’ll not use it. It would be a breach of our policy. Our policy is that all our software has to be integrated [with our ERP-systems]. Because we say that nothing should be keyed more than once due to risk of data errors. So using the software would first of all not lead to timesaving and it would mean more errors. But we have discussed these things many times in our work group – and I have criticized this point several times. And as long as the software can not be fully integrated with our accounting systems then it is irrelevant. However, the advantage is that it is a low cost solution and that is of high value for small companies like us. We are not capable of making large investments in computer systems. And we do not have the same needs as large companies. But the real reason why we have not used the developed software is that we do not want to re-key data. But I think that the software is okay.”
Two informants were positive regarding the use of the software. The two informants represented one company that had adopted EDI and one nonadopter. However, they used the same argument: It is a low-cost and a straightforward solution. The company that already uses EDI runs EDI in a proprietary format. Therefore, it is not an easy task to expand the use of EDI among business partners. The TDP-software is an easy solution. However, the company did not have plans to suggest new EDI partners to 130
exchange EDI messages based on the TDP-software. They would, if necessary use the software if they were requested by a business partner to start an EDI partnership. The other company, which had a positive attitude towards the software, had the following comments: Company G: “Yes, we will use it - if necessary. That does however, not mean that we will integrate it into our existing system. But, we’ll have to check it out - if it actually works. We actually do not know if it works.”
The last two informants answered: ”Yes, if EDI is absolutely necessary in the near future.” They also said that the software is low-cost and a straightforward solution. All the responses to the question concerning use of the TDP-software had one thing in common. The common complaint was that the software could not directly be integrated with their other intra organizational systems. All agreed that one of the main reasons the project had not achieved the necessary power of penetration was this lack of integration. All agreed that an implementation of the TDP EDI software would be a step backwards, since it would mean that data had to be re-keyed. 4.3.7 The suitability of the EDI subset developed during the project Six out of eight informants found that the EDI subset, which is based on the EDIFACT-standard, was suitable for their company and their business sector in general. The reason for this positive response among both users and non-users of EDI was that the work groups have had a great influence on the development of the subset. However, the companies, which do not use EDI, are reluctant to fully trust the subset since they do not know if it actually works. One of the more positive EDI users expressed the following opinion: Author: “Is the subset useful for your organization?” Company F: “Definitely, we defined it. No doubt – we could use it.”
One of the non-users commented as follows: Author: “Is the subset useful for your organization?”
131
Company E: ”I guess it would be my starting point. Now I have a good stepping stone, I am not an IT-person, so it is easier when you don’t have to start from scratch.”
The two informants who did not find the subset suitable had the following reasons. One did not expect that the subset was suitable because the company is a producer and the products being produced are custom-made. The other informant remarked that his company already uses many different subsets, and the TDP subset is simply another subset. 4.3.8 The companies general benefits from the project Benefits from a project such as the TDP can be measured in a number of different ways. Benefits can be approached from a learning perspective or as the achieved output. The feedback from the participants regarding benefits of the TDP is mixed. Especially the non-adopters view the benefits from the project from a learning perspective whereas the adopters are more focused on concrete benefits. There was a fundamental difference in the participants’ perceived benefits from the project based on the amount of time invested in the project. The predominant response from the non-adopters of EDI was that the awareness created through the project had been very useful. Their common opinion was that the project had not yet finished and that they still expected that something would happened that would help them to adopt EDI in their respective organizations. Company E: ”… I’ll say, that the project has prepared me for EDI. I think, if the situation arises where we have to use EDI, then we are ready. Because we know a lot more about it now.” … Author: “Do you find that the project has been a success?” Company E: “Well you might not call it a success – because we are still not using EDI. But I think it is our own fault, that we haven’t taken the final step yet. The reason being that we are reluctant to spend the necessary money. In our case that would be approximately DKK 50,000 (approximately 6,000 USD) for an integrated EDI solution. For us that is a lot of money – because there are also current expenses to existing ITsystems, and there are also expenses associated with EDI. … Author: ”Based on your present knowledge – would you invest time and energy in a similar project again?”
132
Company E: “No, I don’t think so. But time will show – I still hope that the project continues.” Author: ”So you don’t find that the project has had much relevance for your company?” Company G: “No it hasn’t. Well only in a visionary sense. We liked to take part in the project, because it is interesting to take part in something new. And it is interesting to see what is going to happen in the future.” Author: “So, you do after all find that you benefited from participating?” Company G: “Oh yes, I have learned a lot about EDI. I am much better prepared some time in the future to make a decision.” … Author: “Have your expectations towards the project been fulfilled?” Company G: “No, we haven’t come any further. We have reached a point where we ourselves may have to do something. I also think that almost everybody in the work group have completely misunderstood the task. The task was, in my opinion, to develop a basic system – and by the way we talked about creating an order confirmation. You already heard my opinion on management by exception – what is the use of an order confirmation? Why do you have to create documents like that? EDI is suitable for commodities not for order produced goods, such as ships where you need plenty of documentation for each component.” Author: ”What has the project meant for your organization?” Company I: “The outcome has so far been marginal, partly because we have started another project in another setting. But the TDP has promoted our organization in the steel and machinery community. If we had done it on our own we had not been able to show our willingness to use EDI. The greatest advantage has without doubt been that we can also establish connection to small companies.” Author: “What is the short term benefit?” Company I: ”None! But let’s see what happens when we meet after the summer vacation.”
The adopters on the other hand except from one informant did not have many positive remarks about the project. The adopters were not interested in the general information on EDI that the project generated, because they were the sources of that information. The adopter informant that was positive towards the project especially expressed enthusiasm about business opportunities and the process that the project had initiated: Author: ”What has the project meant for your company?” Company F: “Well, what has it meant – I think it has been interesting to take part in the project. If only we could tempt some of our customers to join, so that we could establish closer ties to them. That would be a good incentive [to continue the project]. But I feel sad about the way the project has failed – in my opinion it has been due to lack of means or
133
will from the other participants. They hesitated to spend the necessary resources. It would have been beneficial for everybody if we had agreed to break the ice and had started to communicate via EDI among all five in the group. However, I do feel that it has not only been economic considerations that have obstructed the process, technology has also played an important role.” … Author: “Do you think that the time invested in the project has been worthwhile?” Company F: “Yes, in our case it has been an good investment. And I hope that the rest of the group feels the same. I never expected that the system would be running by 1st of January 1999. I know things take time and IT is difficult to handle. And I realized, as autumn [1998] came that it was impossible to realize the goals that were defined for the project. To me it doesn’t matter if we realize the goals this year or next year. The most important thing is that the project generated ideas among the participants that relations are build, and that we get to an understanding of what EDI is. Now we can always say, “Let’s go”.
Those less positive towards the project expressed the following remarks: Author: ”Has the project fulfilled your expectations?” Company D: “No, I don’t think so. I don’t think that it lived up to our expectations. We had expected faster development of the subset. If only we had been able to use it with those involved in the project – we had to develop another subset instead. The process was simply too slow.” Author: “Did you expect a concrete outcome from the project?” Company D: “We had expected a subset that was ready within a short period of time.” Author: ”What has the project meant for your company?” Company C: “This is not a critique of the project – but our company has such features that makes it unsuitable for that project. This project is targeted towards companies with standard IT-systems. Companies that handle commodities, have many transactions etc. But, the system developed is not compatible with our IT-systems. It would demand, as many resources to integrate the developed system to our IT-systems as it would take to develop a new system from scratch. So to us the participation was driven by general interest in the topic. It does – unfortunately – not affect us in the short run. Nothing will happened unless we are met with demands [of exchanging EDI messages] from participants in the project.” Author: “How about your expectations towards the project, were they fulfilled?” Company C: “Oh yes, you could say so. We had no expectations! I was, from the very beginning aware, that the timeframe for the project was unrealistic. I said so at the first meeting. They wanted to finish the project by January 1999.”
134
Though the non-adopters found that they had gained useful information about EDI during the project they were not very happy with the actual outcome. Especially the two larger non-adopters had expected that the project would have led to adoption and implementation of EDI in their companies. Another complaint voiced by the non-adopters was, that they found that the development of the subset had been too time consuming. So the non-adopter companies tended to feel that time could have been used more effectively especially since they had not gained any operational benefits in the sense that they had not adopted and implemented EDI. Due to their limited understanding of technical issues they did not feel that they had been in a position to influence the development, neither the subset nor the software. They reported no strategic gains. The EDI adopters on the other hand found that the overall outcome was positive. They had achieved some benefits from the subset that had been developed and also because they had had the opportunity to exchange experiences related to EDI with the other participants in the project and it had been a good opportunity for networking. The adopters considered the resources such as time invested in the project as a strategic move, which could enable them to have some influence on the future developments of B2B e-commerce in their business sector. When asked if the project had been a waste of time the EDI adopters answered, “No”. A summary of the findings from the interviews, study of written documentation and direct observation is provided in Table 4-7 (page 137). As the table shows, there are patterns regarding main activity, type of corporation, the use of EDI and the perceived benefit of the TDP. The five companies that use EDI are subsidiaries or part of an industry group. And four out of the five adopters are wholesalers. Another striking feature is that the majority gives the project a positive evaluation regarding the time spent on the project although they have not realized the benefits they expected.
135
4.4 Discussion of the findings In light of the above reported findings the TDP and implicitly the operationalization of the 1996-action plan may at best be assessed as being less than successful. None of the participants were satisfied with the developed software. The non-adopters appreciated that a low-cost EDI solution was made available but they were not eager to use it due to the limited functionality of the system. The immediate outcome of the efforts invested in development of the software has been that awareness of operational benefits related to EDI use has been increased. The argument for reluctance to use the software is thus that operational benefits are absent (e.g. re-keying). The positive attitude towards the software is solely related to strategic considerations. The non-adopters and the users of proprietary EDI have a quick solution at hand in case they are met with demands from trading partners. The objective of the project was to create favorable conditions for further adoption and diffusion of EDI in the steel and machinery industry in Denmark. Does it then make sense in this context to discuss diffusion of innovations when half of the companies already had adopted EDI before the TDP started? Given the complexity of EDI, diffusion of EDI is not merely a matter of having or not having EDI in the company (Ramamurthy et al., 1999). It is also a question of the intensity by which EDI is used (Massetti and Zmud, 1996; Truman, 2000; Premkumar et al., 1994). No evidence was furnished that the TDP had led to an increased use of EDI among the established users, neither related to an increase in the number of EDI partners nor in relation to an increase in types of EDI messages. One company decided to start using EDI during the project. However, in this specific case adoption was not a result of the TDP. And one of the nonadopters had made the initial commitments towards adoption. This company wants to get started within a short time frame. The remaining two non-adopters are thus still non-adopters and the adopters have not increased their use of EDI on account of the TDP.
136
Table 4-7. Overview of the findings from the interviews Characteristics of organizations Position in Type of EDI use/ supply-chain corporation number of years*
Perception of the TDP TDPTDP software subset Adoption suitable
Wholesaler
Yes/ 15
No
No
Yes/ 2
No
No
Yes/ 8
No
Yes
Manufacturer Independent No
Yes
Yes
Good networking. High degree of influence on development process. Helpful information.
Wholesaler
Yes
Yes
Good networking.
Manufacturer Independent No
If forced
Yes
Helpful information.
Wholesaler
Subsidiary
No
Yes
Good Networking.
Wholesaler
Independent No
If forced
Yes
Gained enough information to feel ready for EDI adoption.
Industry Group
Manufacturer Industry Group Wholesaler Industry Group
Subsidiary
Yes/ 1
Yes/ 5
General benefits from the project
Critical issues
Waste of time
Good networking. High degree of influence on development process. Good networking.
No gain of new knowledge.
No
Little new knowledge gained. No gain of new knowledge.
No
Insufficient support from adopters. Little new knowledge gained. Insufficient support from adopters. Not deeply committed to the development process. Insufficient collaboration with adopters.
Neutral
* Not based on the software developed in the TDP
137
No
No Neutral No Neutral
A contrasting feature is that even though the project did not lead to an increased use of EDI among the participants a closer look at the objectives, italicized in the following numbered listing of the project, indicate that the objectives are to a certain extent fulfilled: 1. To develop subsets, which are suitable for the business-sector, based on the EDIFACT-standard. A subset was developed which most participants are actually satisfied with and interested in using. However, none of the companies have so far started using the subset. 2. To contribute to the National EDI Action Plan. The industry and trade associations and the companies have shown their good will to the 1996-action plan by initiating and participating in the TDP. It is however, difficult to evaluate this type of objective, since it is more like a declaration of intent. 3. To launch pilot projects involving members of different business sectors for sharing experiences and knowledge with other members of the respective business sectors. The TDP has led to sharing of experiences among the businesses involved in the project. However, the feedback of opinions from the participants is mixed. Adopters find that no new EDI knowledge was gained. But, at the same time they did find that time had not been wasted. The non-adopters benefited from the project since they learned a lot about EDI, but they were unsure whether or not they had spent their time most effectively. 4. To develop and quality test an EDI software solution, the software, which is shareware, has a fixed selling price so that even the smallest enterprises can afford to participate. Low-cost EDI software has been developed. However, a major weakness pointed out by the participants is the software’s lack of integration with existing systems unless substantial investments are made after purchase of the low-cost software. The participants found that even though the software is inexpensive to buy it will be expensive to put to use. Either data has to be
138
handled twice or the company must make additional investments in order to integrate the software into the company’s intra organizational systems. 5. This project will make every effort to cooperate with other relevant EDI projects within other business sectors, for instance, with the transport sector and other sectors directly involved with the current project in the Danish steel and machine industry sector. This objective, like the second, is more like a declaration of intent. No activities known to the author regarding cooperation with other EDI projects have taken place during the TDP. Although it can be argued that the ‘measurable objectives’ (development of a subset and software) for the project were fulfilled, the participants do not perceive the project as a success. This raises the question: What was not realized through the project? The project did not lead to any further diffusion of EDI. The adoption and implementation of the TDP-software and subset did never take place. None of the companies increased their use of EDI during or after the TDP. One company adopted EDI during the project but did not use the tools developed in the project. One straightforward explanation may be found in Rogers’ framework for diffusion of innovations. Referring to the stages of Rogers, adoption takes place when an innovation is put to use (cf. Figure 6-1, page 195), that is when mental exercises are transformed to behavior change (Rogers, 1995). One explanation for the lack of adoption (and diffusion63) within the businesses that joined the TDP could be, that the EDI software developed in the project is not the appropriate information technology for the companies. IT adoption in general is an organizational effort directed toward diffusion of appropriate information technology within a user community (Cooper and Zmud, 1990). The participants reported that the 63
The adopters in the project had the opportunity to increase their use of EDI based on the tools developed in TDP. They did, however, not expand their scope of business partners to include potential EDI partners from the other TDP participants.
139
software did not fulfill the needs of their companies. The software itself might be usable, but lack of integration makes it unusable for potential adopters. The decision to adopt an IT application, such as EDI, depends upon the alternatives available (ibid.). In the case of the TDP-software it is difficult to justify that adoption and implementation of the TDP software would lead to significant improvements compared to adoption and implementation of other available EDI software already on the market. Authorities both in the industry and trade associations and on company level drove the innovation-decision in the TDP. In the context of authority innovation-decisions, decisions to adopt or reject an innovation are made by relatively few individuals that possess power, status, or technical expertise in a system (Rogers, 1995). The participants in the project were those with “hands on”. They had very good insight into the work procedures that could be optimized by implementing EDI. They had the technical expertise but not necessarily the power or status to make the final decision about EDI. This could lead to the conclusion that lack of management support, which in numerous studies have been found to be of major importance for adoption and implementation success (Premkumar and Ramamurthy, 1995; Sabherwal and King, 1995; Ramamurthy et al., 1999), was one of the main problems that led to lack of success of the project. Another more complex set of explanatory factors can be found in the interorganizational conditions, which influence the adoption climate. As stated by Premkumar and Ramamurthy (1995) the adoption and implementation of IOS and in particular EDI requires cooperation and commitment of all the participating members. These members may have complex economic and business relationships among themselves that can result in a number of social, political, and economic factors influencing the adoption and implementation of IOS.
140
4.5 Assessment of the data collection procedure based on the principles presented by Klein and Myers The purpose of this section is to evaluate the data collection procedure of the case study, which was presented in this and the previous chapter. The challenge at this point is to: “… [move] from a shapeless data spaghetti toward some kind of theoretical understanding that does not betray the richness, dynamism, and complexity of the data but that is understandable and potentially useful to others.” (Langley, 1999), p. 694.
The methodological framework for evaluation of the case data was presented in Table 2-1 (page 34). The findings from the field study will be discussed in the following in relation to the methodological framework that comprises the seven principles for interpretive IS field research (Klein and Myers, 1999). No study known to the author has applied these seven principles rigoursly.64 The Klein and Myers article included three articles (Orlikowski, 1991; Walsham and Waema, 1994; Myers, 1994) as examples of how the principles had been applied in IS research. From a study of these three articles it is obvious that the seven principles are considered to be a checklist, that can be used as an analytic tool in relation to how the researcher approached the field.65 The seven principles are not considered to be a set of rigid rules. It can be argued that a detailed presentation of each of the seven principles with respect to the present case study is out of proportion in relation to the intentions of the original framework proposed by Klein and Myers. It is however, found to be a useful exercise to systematically go through each of the principles in order to evaluate the data collection from the case study. The Klein and Myers framework is therefore used as a tool for cross-checking the data collection rather than for interpreting the case. In order to get to a better understanding of the 64
A search in the Social Science Citation Index (November 2001) for articles citing the Klein and Myers article resulted in twelve hits. These twelve articles were however, using the Klein and Myers article as a reference for the appropriateness of using the interpretive approach in IS research.
141
TDP data, the seven principles for conducting and evaluating interpretive field studies in information systems will be examined in the following. It should be stressed one more time that the seven principles for interpretation should not be viewed as a set of bureaucratic rules. Klein and Myers claim that it is incumbent on authors to exercise judgement in deciding whether, how and which of the seven principles will be appropriate to apply in a given research project. 1. The fundamental principle of the hermeneutic circle. This principle, which is the foundation for the other six principles, suggests that one comes to understand a complex whole from preconceptions about the meanings of its parts and their interrelationships. According to Klein and Myers the two terms “parts and whole” should be given a broad and liberal interpretation. In this context the whole is the TDP and the adoption motivators related to EDI revealed in the TDP – or even broader the whole is the attempt to fulfil the 1996-action plan. The participants and the project initiators are seen as the parts. The exercise then consists in moving from the parts to the whole and back to the parts again in order to gain an understanding of the processes leading to adoption or non-adoption of the tools developed during the project. 2. The principle of contextualization. This principle requires a critical reflection on the social and historical background of the research setting in order to understand how the current situation under investigation emerged. The principle of contextualization borrows from the inherent assumptions of structuration (Giddens, 1984) where structures and actions are closely interrelated (Klein and Myers, 1999). The historical context of the TDP was presented in detail in Chapter 3. It was illustrated how the trade and industry associations were involved in work related to a “technological upgrading” of the Danish business environment. The significance of regulation as a means for action was also discussed. It was questioned whether or not regulation is a means for 65
It would clearly have been an anachronism if the researchers had applied the Klein and Myers framework thoroughly, but the theoretical foundations that guided Klein and Myers could have been applied more or less systematically in the three articles.
142
creating structures that efficiently influence actions. The structures created from policy statements at the governmental level did unquestionably induce the professional associations to act. Apart from taking part in the preparation of the 1996-action plan, which aimed at supporting adoption and diffusion of EDI in the Danish business environment, the professional business associations met the challenge of operationalizing the recommendations of the 1996-action plan in a project, which necessarily had to be tested in a real world setting. In a wider context EDI initiatives and other technological infrastructures aiming at improving business performance had already taken place in a number of countries (cf. Section 3.5.2). This could indicate that Denmark in order to avoid a “digital divide”66 among businesses had to put EDI on the political agenda. Accordingly, the business associations felt pressured to adopt the EDI agenda in order to look after the interests of their members. The content of the 1996-action plan was transformed to the objectives of the TDP, which thereafter were presented to the participants of the TDP. The professional associations aimed at building a structure but the participants did not act in accordance with this imaginary structure. The participants at the time of initiation of the project had received information on the importance of EDI through opinion leaders (cf. Textbox 3-1, page 74) and through multiple channels in the business community. Advantages and disadvantages of EDI had also been discussed both in scientific articles (cf. Chapter 5) and in the business press. These initiatives where however not enough to convince the participants of the value of EDI adoption.
66
The designation digital divide was originally used to describe divisions of classes in society. “In just about every country, a certain percentage of people has the best information technology that society has to offer. These people have the most powerful computers, the best telephone service and fastest Internet service, as well as a wealth of content and training relevant to their lives. There is another group of people. They are the people who for one reason or another don’t have access to the newest or best computers, the most reliable telephone service or the fastest or most convenient Internet services. The difference between these two groups of people is what has been called the "Digital Divide".” (http://www.digitaldivide.gov/)
143
3. The principle of interaction between the researchers and the subjects. This principle requires critical reflection on how the research materials - or “data” - were socially constructed through the interaction between the researchers and participants. Klein and Myers claim that facts are produced as part of the social interaction of the researcher and the participants. They argue that interpretive researchers must recognize the fact that the participants, just as well as the researcher, can be seen as interpreters as well as analysts, because participants are interpreters due to the fact that participants alter their horizons according to the view of the researcher. This will have an effect on the kind of data the researcher obtains. There were two occasions of direct interaction between the researcher and the subjects. The first time was during one of the work meetings where most of the participants were present (cf. Textbox 4-1, page 103). The second time was when the researcher met with the involved people in the project during the individual interview in the informants’ company. The researcher did interfere during one of the work group meetings by her mere presence and she presented an alternative view of EDI and the relevance of adoption for the involved companies. Some of the participants expressed that they felt that they had become subjects of investigation, since the researcher was “snooping around”. For others her presence and especially the alternative view of the importance on EDI was found to be a source of inspiration, which led to new ideas of EDI in relation to the importance of EDI in the organization. This became clear at the informal gathering at lunch, where the participants shared “EDI anecdotes” with the author. Additionally, an interest was shown in how the author viewed EDI and the benefits related to EDI. In the next interactive situation each participant individually met with the author in a semi-structured interview. The informants had received the questions in advance and had prepared their answers before the interview took place. As mentioned in Section 4.3 some of the informants had considered how to align their opinions of the questions with the expectations of the business associations. The interaction with the researcher brought focus on issues that had not prior received explicit verbal attention.
144
4. The principle of abstraction and generalization. Klein and Myers claim that it is important that theoretical abstractions and generalizations should carefully be related to the field study details as the researcher experiences them. They stress that the key point is that theory plays a crucial role in interpretive research, which distinguishes it from mere anecdotes. Data has been outlined in detail in the present chapter in order to enable the reader to follow or even to deduce the same conclusions as the researcher using the same data. This does however not necessarily mean that the reader should agree in the choice of theoretical framework, which is outlined in the two forthcoming chapters. One stream of theory, the adoption and diffusion of innovations theory, is specifically found to be appropriate as a theoretical framework for explaining the outcomes of the case study. Other studies within MIS have used different theories in order to explain the adoption and diffusion of IS in businesses and in the business community.67 The major issue is however not, whether or not one theory is more appropriate than other theories. It is rather to enable the readers to follow how the researcher arrived at her theoretical insights. One way of meeting this challenge has been to mimic the procedure presented by Flyvbjerg (1998). The strategy followed is therefore to present data in the initial stages of the dissertation and then to relate the insights gained from the case study to previous research and the theoretical elements applied along the way. 5. The principle of dialogical reasoning. This principle requires that the researcher confront her preconceptions that guided the original research design with the data that emerge from the research process. One of the major issues related to the researchers intellectual history is embedded in this principle. Given the researcher’s academic training, which is poles apart from MIS and adoption and diffusion of innovations, there were no particular preferences at the initial stages of the study. The study started as an exploratory examination of the environment and there were no articulated preconceptions. 67
Examples include game theory (Barua and Lee, 1997), BPR (Chatfield and BjornAndersen, 1997; Clark and Stoddard, 1996), transaction cost theory (Lyytinen, 1991), and value-chain thinking (Krcmar, Bjorn-Andersen & O'Callaghan 1995).
145
6. The principle of multiple interpretations. The principle of multiple interpretations requires the researcher to examine the influences that the social context has on the actions under study. Or to put it plainly: Is it possible to find different versions of the same story? The viewpoints of the various stakeholders are absent in the description of the TDP and in the preface of the TDP. As mentioned in Section 3.6 the motivation for the involvement of the industry and trade associations in the preliminary phases of the 1996-action plan is not accessible. Therefore, it is not fully known what made the business associations support adoption and diffusion of EDI among their members, nor is the rhetoric and explanations used to involve the members known. The objectives of the involvement in the EDI development communicated to the members of the industry and trade associations and especially to the participants in the TDP is however quite clear. They recommended their member organizations to adopt EDI. The participants of the TDP did not share this enthusiasm for EDI. The non-adopters were not convinced of the suitability of the EDI solution, and the adopters involved in the project were not motivated to make efforts to broaden their scope of EDI with their business partners. It therefore appears as if the professional business associations and their members are not sharing the same goal of adoption and diffusion of EDI. EDI was announced as a tool for improving efficiency and for cost reduction by the business associations, whereas their members perceive EDI as a strategic tool. Another interpretation of the project is that the industry and trade associations felt an obligation to launch the project, but they were not investing the necessary time and other resources in order to support the project. The direct investment in the project was DKK 250,000 (approximately EURO 40,000), which were obtained from a grant from the Danish EDI Council. The grant only covered the salary to the IT-consultant leaving no resources for additional management and support of the project. 7. The principle of suspicion. The seventh principle accordingly to Klein and Myers is concerned with the discovery of “false preconceptions”. This principle is different from the other principles, which are more concerned with the interpretation of
146
meanings. This principle of suspicion therefore requires the researcher to pay attention to possible “biases” and systematic “distortions” in the narratives gathered from the participants. This final point is a problematic one since it is the basic belief of the researcher that the TDP participants have given their interpretation and view of the TDP as truthfully as possible. No evidence was found that any of the involved parties deliberately changed the story. Nobody tried to hide that the project was not the success it was expected to be. There were many rich opportunities to examine the “surface problems” in the project since all participants and policy makers acknowledged that the outcome of the project was not as successful as was expected. If any systematic distortions are to be found, it is in relation to the political interaction that took place. Politics played a major role in the project in the sense that some of the involved companies wanted to set the agenda at the expense of the outcome of the project. None of the involved companies were however eager to outline the political structures that governed the interaction in the TDP.
4.6 Summing up Chapter 4 The qualitative data, which is used to examine the second research question of this dissertation, was presented in this chapter. From examination of different sources of information the outcome of an initiative launched by two business associations was presented. Throughout the process it became clear that the reasons for adoption or non-adoption were not solely to be found in the TDP itself. Development of an inexpensive EDI software, development of an industry subset for EDI, and sharing of EDI knowledge was not sufficient for adoption and diffusion of EDI in the Danish steel and machinery industry. From analysis of the individual characteristics of the involved companies it seemed as if there was a pattern concerning the size of the company, the legal status of the company, and the position of the company in the supplychain. It was found that those companies that were dependent companies in the sense that they were either part of an industry group of subsidiaries had adopted EDI, whereas companies that were independent had not yet 147
adopted EDI. The position in the supply-chain also seemed to influence adoption of EDI. The wholesalers involved in the TDP seemed to be more inclined to adopt EDI than the manufactures. This “quantification of qualitative data” does however only partially lead to an understanding of motivators for adoption of EDI amongst the TDP participants. A closer look at the individual reasons and considerations related to EDI adoption in the individual organizations revealed, that competitiveness and pressure seemed to play a central role. For those adopters that voluntarily had adopted EDI the major incentive was their wish to set the standard for EDI communication in their business sector. They also wanted to make sure that they had an innovative image and that they were technologically attractive to their customers. Competitive considerations therefore played an important role for this type of adopters. For the companies that were forced to adopt by their parent company or by other companies in their industry group, the strategic considerations played a minor role. Companies such as “C” and “F” were however not blind to the strategic opportunities their more or less voluntary adoption of EDI had created. Though they in the first case were encouraged to adopt EDI from another company in the industry group and the parent company respectively they were in a position where they wanted to pursue a more proactive strategy for including more business partners in an EDI partnership. Strategic concerns therefore became an issue for those reactive adopters. Cost of acquisition was one of the major barriers for the non-adopters in relation to EDI, but the expected absence of potential savings from EDI usage also seemed to play an important role. The incentive for nonadopters to adopt EDI was competitive threats. They were willing to adopt, if their business partners forced them to do so. Based on these observations it is concluded that competitive concerns and the power factor were the dominant forces determining EDI adoption amongst the eight interviewed companies from the TDP.
148
So far the author has had no preconceptions related to the term EDI, nor have the theoretical conceptualizations of the three contexts included in the second research question been explained. The use of the term EDI has so far reflected the TDP participants’ understanding. In the following chapter an examination of EDI as a research object is made. Specifically considerations related to benefits accruing from EDI is discussed in relation to the MIS literature.
149
150
5 IOS and EDI As outlined in Figure 1-3 (page 16) the theme of Chapter 5 is to conceptualize technology, IOS, and EDI for the purpose of providing an answer to the research question, “How are the explanatory variables related to motivators for adoption defined in MIS research at present?” After identifying the explanatory variables for EDI adoption these variables are related to the three contexts mentioned in the second research question: Organizational, environmental, and technological.
5.1 Introduction The case study presented in Chapter 4 focused on motivators for adoption and diffusion of EDI. The purpose of this chapter is to take a closer look at Interorganizational Information Systems (IOS) including EDI. The purpose and theme of the dissertation is to pinpoint the motivators for adoption of IOS. In Chapter 5 IOS will be defined and exemplified by EDI. There is broad agreement in the IOS literature that EDI is a subset of IOS (Swatman and Swatman, 1992). EDI is one of those IOS that has been intensively researched during the last two decades. Though EDI has recently been challenged by business-to-business e-commerce there are still organizations that use this technology and even more important in this particular case there are still organizations that consider adopting EDI. However, as mentioned in Chapter 1, EDI is used as an exemplification of an IOS and EDI is not necessarily the optimal solution to interorganizational coordination and communication of business information. There are strengths and weaknesses related to exploring a concept, which is subject to competing concepts such as e-commerce. Influence from ecommerce is present in some of the later articles in the EDI review presented in this chapter. E-commerce might be a more attractive topic to
151
explore from a prospective point of view, but one advantage of working with an older type of technology is, that the involved informants are more familiar with the concept in a real life setting. This reduces an imminent source of bias: Lack of knowledge of the research topic. Apart from that it should now be clear that substantial resources were invested in promotional efforts related to the diffusion of EDI in the Danish business community (cf. Chapters 3 and 4). It is hence of general interest to pursue and examine the effect of the EDI campaigns and projects in an environment where EDI is still considered to be an innovation. The present chapter starts with a presentation of different views of IS/ IT. The purpose is to clarify how IS/ IT is viewed in the present project. Besides, it is presumed that the basic assumption of IT/ IS among decisionmakers influences the expectations of e.g. EDI among adopters. A classification of the IT/ IS terms from a more philosophical point of view may therefore be helpful in providing a better understanding of adoption or non-adoption of EDI. Thereafter, IOS is defined and discussed based on reviews of existing literature. IOS is viewed from two different angles, an economic in relation to the business value of IOS and an organizational in relation to the impact on actors and routines caused by adoption and implementation of IOS. The following section is used to define and provide an overview of EDI and the myriad of approaches to EDI. After defining EDI a systematic review of ten years of EDI research articles in the top-five MIS journals is presented. The objective of the review is to provide an overview of research approaches to EDI from a MIS perspective. After identifying the dominating foci of EDI research the case study is revisited in order to discuss the features of the case in relation to the major research themes during the last decade of the twentieth century.
5.2 Views on IT and IS The fundamental perception of a given phenomenon is of major importance for the interpretation of the phenomenon. The purpose of this section is to illustrate different views related to the significance of technology in
152
organizations and among individuals. By introducing a frame of reference a tool for further interpretation of the TDP is provided. Technology and the influence of technology in particular are viewed in different ways depending on the fundamental stance of the viewer. Barley (1998) suggested a classification based on a two-by-two matrix consisting of determinism versus volunteerism on one dimension and materialism versus idealism on the other dimension. The materialistic determinists view technology and other physical artifacts as direct causes of social phenomena. The view is influenced by Marxist ideology, which considers technology as a tool for further deployment of the political and economic interests of powerful actors. The idealistic determinists consider sociotechnical developments as being driven by cultural ideologies. The materialistic voluntarists believe that technology more or less shape and determine human behavior. They argue that technology can be altered and that humans can affect the social impact of a technology by re-designing it or refusing to adopt it. The idealistic voluntarists hold the view that a technology’s effects are rooted in the beliefs and values of designers. Following this stance it is believed that the effects of the technology can be changed by changing the designer’s images of the users. Table 5-1. Four epistemological views of technology Determinism
Volunteerism
Materialism Social phenomenon causes technological change. Idealism Technological change is a manifestation of cultural ideologies. Source: Adapted from Barley, 1998
Technology causes social phenomenon. Technological change is a process controlled by humans.
These four different viewpoints of technology are useful when it comes to interpreting the role technology is given in an organizational context by the involved actors. Though the four views presented by Barley are based on technological change and are hence the driving forces for development of technology, it is a useful tool to apply when an understanding of organizational adoption of technology is pursued. It is relevant because the
153
views reflect the expectations towards technology or organizational behavior prompting the motivation for adoption. Markus and Robey (1988) examined the role of technology in organizations by looking at researchers’ conceptions of the nature and direction of causality. Theoretical considerations rather than practical organizational behavior therefore drive the views presented by Markus and Robey. The value of the Markus and Robey framework as an interpretive tool should however not be underestimated due to the role that managers give technology in organizations. Markus and Robey describe the role of technology in organizations based on three perspectives: The technological imperative, the organizational imperative, and the emergent perspective. These three perspectives are given a more normative importance in this context than originally envisaged by Markus and Robey. In the technological imperative technology is viewed as causing organizational change. According to this view, the characteristics of the technology determine actions in terms of usage and consequences of adoption. This view resembles Barley’s materialistic volunteerism where the impact on organizational behavior caused by technology is the point of departure. Markus and Robey describe technology in the technological imperative as an, “... exogenous force which determines or strongly constrains the behavior of individuals and organizations.” The premise for the technological imperative model is that technology as well as the organizational and individual factors can be measured and predicted (Orlikowski 1992). The organizational imperative is characterized by the belief that human actors are the designers of information systems, and that these systems will be able to satisfy organizational needs for information. The organizational imperative described by Markus and Robey thus resembles the idealistic voluntarism described by Barley. According to Markus and Robey the organizational imperative does, “... assume almost unlimited choice over technological options and almost unlimited control over the consequences.” The emergent perspective, which holds that the uses and consequences of information technology emerge unpredictably from complex social interactions, is rooted in structuration theory (Giddens, 1984). The third
154
perspective, the emergent perspective, can be interpreted as related to the materialistic deterministic view presented by Barley. The works of Barley (1998) and Markus and Robey (1988) are mainly concerned with identifying how technology at the meta-level is conceptualized in the organization or by individuals. The major issue is to determine whether technology influences organization or whether it is the other way around. The emergent perspective includes the role of the users of the technology. However, the impact on individuals caused by automating work processes is underplayed. Kling addressed this issue in his description of the rationalities of technology (Kling, 1980; Kling 1991). Kling described the systems rationalism and the segmented institutionalism. The systems rationalism perspective is based on a technoeconomical view whereas the segmented institutionalism perspective is centered on a socio-political approach to technology. Systems rationalism emphasizes the positive roles that computerized technologies play in social life. This view is a proponent for seeing technology from an efficient and practical view where organizational action is viewed as purposive and rational or externally constrained and controlled (Swanson, 1987). The systems rationalism builds on the neo-classical ideals in the sense that it presumes that all actors and stakeholders in an organization subscribe to the same economic goal: Maximizing the organization’s economic efficiency and effectiveness through technology (Kumar et al., 1998). The segmented institutionalism on the other hand examines the consequences of technologies in different aspects of social life. Kling exemplifies the segmented institutionalism perspective by describing how computerization of work procedures can lead to enhanced personal status, improved quality of decisions, and increased economic efficiency of specific activities. The segmented institutionalism is not technocentric and it does therefore not presume a technological imperative or economic rationality in human behavior. Instead it opens the door to the investigation of social phenomena in management of information systems (Kumar et al.,
155
1998). The perspective assumes that organizations are fora for political activity. Kumar et al. (1998) used Kling’s two rationalities to interpret their study of the Pratesian business community. Kumar et al. however found it necessary to add a third rationality of technology. Kumar et al. added to the two rationalities defined by Kling (1980; 1991), a rationality, which is beyond the culture where the two perspectives originally where defined. Kumar et al. found that in this particular context (the Prato district in Italy) it was more relevant to apply a trust and relationships rationality. The trust and relationships rationality focuses on collaboration and cooperation as the key interaction processes. The key concepts are trust, social capital, and collaborative relationship. The major difference between the ideas presented by Barley and Markus and Robey and those presented by Kling and Kumar et al. is that Kling and Kumar et al. focus on the impact of technology, whereas Barley and Markus and Robey focus on the conceptualization of technology. Barley and Markus and Robey are more concerned with the ontology of technology whereas Kling and Kumar et al. concentrate on the motivation for adoption of technology.
Figure 5-1. Views of technology Conceptualization of technology
§ Materialistic determinism * Materialistic volunteerism § Idealistic determinism * Idealistic volunteerism
Rationalities of technology
Action
§ Techno-economical * Socio-political * Relations-based
Legend: § Outcome of technology is given and is subject to governance and control * Outcome of technology emerges and is a process with a high degree of complexity
156
As illustrated in Figure 5-1 the two elements, conceptualization of technology and rationalities of technology, can be seen as having one thing in common: Action. People act according to their basic belief of technology. A management, which views technology from a materialisticvolunteeristic point of view, invests substantial resources in technological hardware and software giving lower priorities to organizational and human consequences. The motive or rationale for acting can be to improve control of work processes in the organization. In that case the rationality for technology could be labeled socio-political. An ex ante interpretation of a given project can possibly lead to different outcomes in relation to the intentions of the project and the immediate outcome. In the TDP the rationale for diffusion of EDI was presented in techno-economic terms. In the introductory letter from the business associations the main argument was that EDI could increase efficiency and optimize work processes. The ex post assessment of the project does however, point more to the relations-based rationality for EDI adoption in the Danish steel and machinery industry. The project initiators exhibited a materialistic-volunteeristic conceptualization of EDI primarily emphasizing the development of low-cost EDI software and an industry subset in the five objectives of the TDP. The participants on the other hand did not share the same convictions regarding EDI adoption.
5.3 Defining IOS in a MIS context IOS consists of a computer and communication infrastructure that permits sharing of information systems (IS) applications across company boundaries (Cash and Konsynski, 1985). IOS is thus a combination of IS and an interorganizational environment and has as a consequence essentially both technological and organizational characteristics (Johnston and Vitale, 1988). An organization’s information system is hence no longer merely a computing device for facilitating internal information processing activities, it is emerging as a crucial means of facilitating business strategies (Wang and Seidmann, 1995).
157
Before going any further two definitions will be presented. A prerequisite for defining Interorganizational Information Systems (IOS) is to define Information Systems (IS) which is a part of the IOS. One widely used definition of IS is the following: “An IS [Information System] is a formal, deliberately planned technological innovation composed by man, machine, and procedures that is introduced into an organization in response to a perceived need on the part of one or more organizational members.” (Kwon and Zmud, 1987).
An IOS is defined as: “An automated information system shared by two or more companies. An IOS is built around information technology, that is, around computer and communication technology that facilitates the creation, storage, transformation and transmission of information. An IOS differs from an internal distributed information system by allowing information to be sent across organizational boundaries.” (Johnston and Vitale, 1988).
The Johnston and Vitale definition of IOS, which is based on the widely used Cash and Konsynski (1985) description of IOS attributes, (see for example Swatman & Swatman 1992; Bergeron et al., 1991; Iacovou et al., 1995; Lee et al., 1999; Choudhury 1997) is particularly focused on the technical attributes of IOS. Other aspects which have to be included especially in relation to EDI, which is a subset of IOS (Swatman & Swatman 1992; Ramamurthy and Premkumar, 1995) are socio-political traits, which influence the collaboration across organizations (Damsgaard and Lyytinen, 2001; Damsgaard & Lyytinen 1998a; Damsgaard and Lyytinen, 1998b). These traits are among others related to: The interorganizational and networked aspects which are not centrally controlled by any authority, the linking of organizations which imply that organizational boundaries are lowered leading to increased transparency, the dependency on critical mass and, the dependency of infrastructures to secure efficient and reliable transactions between trading partners (Damsgaard and Lyytinen, 2001; Damsgaard & Lyytinen 1998a; Damsgaard and Lyytinen, 1998b). Those aspects have to be considered when studying the drivers and barriers for adoption and diffusion of IOS 158
including EDI. It may be technically possible to fulfill the requirements for an IOS outlined by Johnston and Vitale (1988). Another issue is however, whether the additional features outlined by Damsgaard and Lyytinen (1998a; 1998b; 2001) related to IOS in the form of EDI or other types of business-to-business IOS can be met, since it involves complex sociopolitical processes among the business partners. The foundation for IOS is an IT based information system.68 IS in organizations have multiple roles. From a task perspective the roles include increasing of scale efficiencies of the firms operations, processing of basic business transactions, collecting and providing information relevant to managerial decisions, maintenance of records of status, and change in the fundamental business functions within the organization (Gurbaxani and Wang, 1991). The adoption of an IS innovation has a number of consequences in relation to the roles of the employees including fashioning and incorporating new roles, responsibilities, relationships, lines of authority, control mechanisms, work processes, and work flows – in short, new organizational designs. Based on these multifaceted sources of organizational interference IS innovations can be seen as presenting themselves intrinsically as organizational forms (Swanson and Ramiller, 1997). IS in organizations include an administrative core, a functional IS core (which serves as a link to the two other cores), and a technical core. As a consequence adoption of IS has impact on both the business’ administrative and technical work processes (Swanson, 1994). Swanson argues that the overall domain of IS innovation may be mapped on two basic dimensions: The business impact and the technological and organizational feature composition. In the present focus is on the business impact rather than on the technological and organizational dimension. Though it is acknowledged that separating business impact from organizational issues is often neither easy nor appropriate. Both IS and the interorganizational environment are multifaceted entities. The interorganizational environment consists of supply-chains, trading 68
Opposed to information systems based on physical artifacts e.g. roads, railway tracks and other forms of infrastructures.
159
partners, standards organizations, industry bodies, transport companies, trade associations, software providers etc. (Kurnia and Johnston, 2000). The scope of the systems, the involvement of different organizations with differing goals and the range and nature of possible relationships between the parties involved makes the situation one of extreme complexity (Gregor and Johnston, 2001). The number of variables to consider in relation to adoption and use of IOS is therefore multifarious. The interorganizational connection can be established at different levels. The connection can be established as electronic dyads, multilateral IOS, or electronic monopolies (Choudhury, 1997; O'Callaghan and Turner, 1995). In the electronic dyad the trade partners establish individual logical links with a selected number of attractive businesses. EDI is an example of an electronic dyad (Choudhury, 1997). The electronic dyad often leads to hierarchical structures whereas the multilateral IOS assembles a marketstructure exemplified by electronic marketplaces (Malone et al., 1987). Electronic marketplaces have with the emergence of the Internet as a means for transportation of commercial information and messages become one of the central issues in business-to-business e-commerce (Kalakota and Whinston, 1996; Wigand, 1997; Zwass, 1996). The electronic marketplaces do not exclude EDI as the standard for communication (Henriksen, 2001) nor is the Internet irrelevant as a means for transportation of EDI messages (Andersen et al., 2000; Henriksen and Göersch, 1999). The focus of this dissertation does not include how linkages between companies are organized (marketplaces or hierarchical structures), nor does it include mapping the means of transportation of EDI messages. The benefits of IOS, strategic or operational, are not unconditionally achieved by introducing an IOS to a trading partner. It is widely accepted that in order to make the system successful and compensate for the initial investment, a number of partner organizations have to agree to join and to use the system (Gebauer and Buxmann, 2000; Kumar et al., 1998). The agreement or agreements are connected to different types of factors. Allen et al. (2000) suggested a typology for the maintenance factors for IOS consisting of five interrelated elements: Data and code standards, shared
160
objectives, power, trust, and understanding. As will be shown especially the last four elements play a significant role in relation to EDI from a MIS perspective. Based on the above-mentioned characteristics it should be clear that the adoption of IOS has an impact on multiple functions in the involved organizations both in relation to strategic and operational matters and in relation to employees. 5.3.1 Business value of IOS IOS can be viewed from different angles such as the business value represented by strategic versus operational uses (Gebauer and Buxmann, 2000; O’Callaghan and Turner, 1995). The organizational impact where the impact of IS in relation to employees and work procedures is the main focus (Barley, 1986; Orlikowski and Baroudi, 1991; Zuboff, 1988) is another approach to IOS. A third approach is the interorganizational impact where structures and power relations are of importance (Kling, 1980; Kumar et al., 1998). IOS have drawn considerable attention from IS researchers because of their strategic significance (Barua and Lee, 1997). Dominant themes are the linking of IT and corporate strategy (Porter and Millar, 1985) and new institutional theories, including transaction costs (Malone et al., 1987) and agency theory (Gurbaxani and Wang, 1991). Porter and Millar (1985) have described the corporate strategy concerning value-chain thinking. The authors called attention to the fact that IT could affect competition in three vital ways: It changes industry structure and thus alters the rules for competition, it creates competitive advantage by giving companies new ways to outperform their rivals, and it spawns whole new businesses. In order to take advantage of the opportunities created by IT the authors among other things recommended that senior executives determine the role of information technology in their industry structure and develop a plan for taking advantage of information technology. As the analysis of adoption processes in relation to EDI shows these recommendations make sense at first. They are, however, difficult to put into practice. Another 1985-article also outlines the strategic impact of IOS
161
(Cash and Konsynski, 1985). Cash and Konsynsky concluded, similar to Porter and Millar (1985) that IOS have the potential to radically change the balance of power in buyer-supplier relationships, provide entry and exit barriers in industry segments, and shift the competitive position of intraindustry competitors. Though the two articles are relatively old they none the less both still point at the core role of IOS from a MIS perspective. The new-institutional theories have had substantial impact on the view of IOS. Especially the transaction cost theory (Coase, 1937; Williamson, 1975) has been utilized to explain the operational and strategic benefits of IOS (Malone et al., 1987; Malone et al., 1989; Bakos, 1991; Bakos, 1997; Picot et al., 1997; Choudhury 1997). Malone et al. (1987; 1989) argued in favor of reduction of coordination costs due to the ‘electronic communication effect’ which would allow more information to be communicated in the same amount of time, and a dramatic decrease in the cost of communication. Malone et al. predicted that the electronic markets, not electronic hierarchies, would become the dominating structure as a consequence of the electronic communication effect. Clemons et al. (1993) have challenged this stance by suggesting that “a move to the middle” is more likely to take place in the future market, where benefits of the electronic communication effect can be achieved both in hierarchies as well as in markets. Gurbaxani and Wang (1991) challenged the transaction cost view by suggesting that the impact of IS on organizations should be viewed in relation to agency theory. Gurbaxani and Wang thus question the view of the actors in the organization as having shared goals benefiting the interests of the organization. One central consequence of adoption of IOS is the changes in powerrelations. Within the context of interorganizational systems, it becomes necessary to redefine what is considered to be “outside” the firm’s boundary and therefore beyond the firm’s control. For large buyers with substantial leverage over their suppliers, mandating changes to suppliers internal systems and processes is now becoming possible. If buyers possess substantial market leverage over their suppliers, they may be in a position
162
to mandate change (with or without subsidy) at the supplier’s facility (Riggins and Mukhopadhyay, 1994). 5.3.2 IOS from an organizational point of view The IOS literature is dominated by the paradigm of bounded rationality, utility maximization, and opportunistic behavior to describe the actors’ activities. These actors are however to a large extent limited to managers and other decision-makers in the organization. In relation to that particular group of people in the organization centralization versus decentralization due to adoption of IT is widely discussed (Sampler, 1996). The considerations in relation to the influence of technology on organizational work-processes and employees are less explored than the economic perspectives. Though it is more than forty years ago that academics began speculating on the impact IT would have on organizational structure (Leavitt and Whisler, 1958) MIS research has been modest in its exploration of this topic. Apart from the contributions discussed in Section 5.2 related to rationalities of IS the sources are limited when it comes to exploring how adoption of IS (and IOS in particular) influences workprocesses and employees. In her pioneering book “In the Age of the Smart Machine” Zuboff (1988) introduced the terms “automating” and “informating”. Zuboff argues that technology can be designed with different intentions, which will have different implications for workers. Automating can lead to controlling and deskilling, whereas informating can lead to empowering and upskilling of the workforce. Zuboff hence acknowledged that economic utility maximization is not always the driving force for the employees that have to operate the technology. It has recently been acknowledged that adoption of IOS involves significant changes in organization’s culture, structure, business relationships and working practices over time and space (Kurnia and Johnston, 2000). Many articles concerning the use of IOS, such as EDI, ignore or minimize the organizational or process changes required to take advantage of the technological capabilities. Some note the changes but assume that technological innovations automatically result in new 163
processes that effectively use the new capabilities (Clark and Stoddard, 1996). It has however been found, that the shift in the underlying business processes and communication patterns bring about changes in the skills of employees and, in some cases, even employee categories (Cash and Konsynski, 1985; Swanson and Ramiller, 1997). These changes might however not automatically be adapted in the organization, and may not lead to efficient uses of the new technology. A shift in the work-processes and power-relations internal in the organization is thus inevitable when adoption of IOS takes place (Kling, 1980; 1991). 5.3.3 Applied view on IOS To summarize, IOS is viewed as a complex mechanism, which includes technological as well as organizational components. It represents multiple aspects relating operational tasks and actors. It is an important strategic tool for coordination and maintenance of competitive advantages. Based on the new-institutional approaches, which often have been used as a framework for identifying the basic characteristics of IOS it is accepted that bounded rationality, utility maximization, and opportunistic behavior describe and guide the actors’ activities. In the following section the focus is on EDI, a particular type of IOS.
5.4 EDI The purpose of the following section is to present EDI research in order to illustrate how EDI has been conceptualized and researched in MIS. The section serves as a literature review of previous research on EDI from a MIS perspective. The review is divided into two parts. First, a general description and conceptualization of EDI is presented. Next, a systematic review of articles on EDI from the top-five MIS journals during a whole decade is examined. The purpose is to identify the focus of EDI research published through these selected channels. To assure coverage of the multiplicity of EDI a complementary description of EDI based on a broad variety of sources is made. The section starts with a presentation of the method chosen for selecting the papers included in the review. The next 164
section deals with a definition of EDI and a description of the major features related to EDI. Finally, the ‘systematic’ review based on articles from the top-five MIS journals is outlined. 5.4.1 Research strategy The literature on EDI is vast. Since the mid-eighties when academics started to publish on the topic there has been a continuos stream of publications in IS journals. To review the entire list of publications would be an enormous task both in relation to time and space, even if a strict set of limitations related to the focus applied in this study were made. The criteria for selecting the publications for the review has been to include those articles related to EDI that have been published in the top-five MIS journals (Hardgrave and Walstrom, 1997) from 1991 to 2000. The top-five MIS journals identified by Hardgrave and Walstrom are: MIS Quarterly, Communications of the ACM, Information Systems Research, Management Science, and Journal of Management Information Systems. The publications from the top-five MIS journals are mainly used to illustrate the trends in EDI research within MIS during the decade. In the retrieval of publications the search criteria were the keywords “Electronic Data Interchange” and “EDI”. Finally, abstracts from the journals were browsed for the two keywords. This research strategy has both advantages and drawbacks. One of the advantages is that the review is rooted in the “MIS paradigm” (if such one exists). Another advantage is that a clear set of restrictions for inclusion of articles are defined – which journals and which period of time. The drawbacks include that the ratings of the journals are based on individual preferences of some authors in this particular case Hardgrave and Walstrom. A second drawback is that the journal ratings are not definitive. MIS journal ratings are an ongoing process which can best be illustrated by visiting the IS World homepage (www.isworld.org). A third drawback is that the top-journals not necessarily reflect contemporary research and especially contemporary data. To get research published in top-end journals can be a prolonged process (Mingers, 2002). An alternative would have been to select papers from less high profile journals, conferences and workshops. This would however have been a never ending
165
task, and there would always be a danger of excluding some works, which deserved to be included in favor of less important works. Finally, it should be mentioned that top journals tend to be biased towards acceptance of positivistic studies conducted in US (Mingers, 2002; Claver et al., 2000). It is however, recognized that EDI is adopted and diffused throughout the world and that non-US journals also have contributed with excellent studies on EDI. Despite all these drawbacks it is found that the top-journals after all have a greater audience among researchers and that they thereby create a collective body of knowledge, which is shared amongst researchers throughout the world and which thereby influence researchers in their own work. A search was made in the Social Science Citation Index with a crosscheck in the ABInform database. A total of twenty-four articles appeared:
Table 5-2. EDI articles from the top-five MIS journals, 1991 to 2000 Journal MIS Quarterly
# 5
Information Systems Research Management Science
2 6
Publications Massetti and Zmud, 1996; Iacovou et al., 1995; Mukhopadhyay et al., 1995; Jelassi and Figon, 1994; Bergeron et al., 1991* Barua and Lee, 1997; Lee et al., 1999 Cachon and Fisher, 2000*; Lee et al., 2000; Chen, 1998*; Wang and Seidmann, 1995; Srinivasan et al., 1994; Riggins et al., 1994 Riggins and Rhee, 1998; Rotenberg, 1993*
Communications of 2 the ACM 9 Chatfield and Yetton, 2000; Truman, 2000; Hart & Saunders Journal of 1998; Chatfield and Bjorn-Andersen, 1997; Teo et al., 1997; Management Clark and Stoddard, 1996; Premkumar et al., 1994; Riggins Information and Mukhopadhyay, 1994; Kimbrough and Moore, 1992 Systems Note: The articles marked with an * has the term electronic data interchange or EDI in their keywords, but the text does not explicitly refer to the issue or EDI is just included in a superficial way and therefore does not illustrate EDI research as such. Those articles are therefore omitted from the review.
For a definition and general description of EDI a set of complementary articles is included in the sample because it is found that those twenty-four publications from the top-five MIS journals do not cover the entire area of EDI. 166
5.4.2 Defining EDI There are several definitions of Electronic Data Interchange. One of the first definitions that appeared in an academic article is by Hansen and Hill (1989): “Electronic Data Interchange (EDI) is the movement of business documents electronically between or within firms (including their agents or intermediaries) in a structured, machine-retrievable, data format that permits data to be transferred, without re-keying, from a business application in one location to a business application in another location.” (Hansen and Hill, 1989).
A few years later Pfeiffer (1992) listed four criteria for what constitutes EDI: -
-
At least two organizations having a business relationship, i.e. conducting joint business transactions. Data processing at both (all) organizations pertaining to a transaction are supported by independent application systems. The integrity of the data exchange between application systems of transaction partners is guaranteed by ex ante agreements concerning data coding and formatting rules, eliminating or at least reducing the need for human intervention. Data exchange between the application systems is accomplished via telecommunication links.
Later definitions do not substantially differ from the 1989 and 1992 definitions e.g. (Williams and Frolick, 2001; Raymond and Bergeron, 1996; Damsgaard, 1996; Arunachalam, 1995) in relation to the challenges that organizations face when adopting and implementing this technology. Basically it is still the same actors and the same procedures that are involved in the process of exchanging EDI messages. From a technical perspective the electronic movement of information and the means of transportation have improved in speed and quality and also in new types of formats e.g. XML. In this context the discussion of competing technologies are omitted, and the discussion of alternative means of transportation is kept to a bare minimum. In short, EDI allows business partners to make commercial transactions by sending and receiving digital documents over telecommunication networks 167
(Raymond and Bergeron, 1996). The core purpose of EDI in business-tobusiness transactions is: To transport business documents via electronic means in a format that is reusable throughout the entire organization and beyond organizational boundaries. The scenario is thus when for example a purchase order is generated the corresponding data is used both within the buyers’ organization and within the sellers’ organization in for example production, accounting, and inventory units (O’Callaghan and Turner, 1995). Besides that, data can be transmitted to external transportation agencies and governmental units such as the taxation bureau etc. (Andersen et al., 2000). EDI can thus closely integrate business procedures. Though it is a complex process, it has been claimed that this technology can yield great benefits for the companies (Kimbrough and Moore, 1992; Bergeron and Raymond, 1992). Both of the above-mentioned definitions include the important characteristic of EDI messages – the highly structured format. Depending on how sophisticated the systems are in the businesses involved and the degree of integration of administrative systems the exchange of information can be both external (e.g. purchase orders) and internal (e.g. logistics and production planning). To derive those benefits from EDI, it is necessary to have both external and internal EDI connections. Thus it has been argued, that the entire range of capabilities of EDI cannot be realized unless EDI is fully integrated with internal IS applications (Ramamurthy et al., 1999). Several studies have found that EDI gives the opportunity to secure short transaction time for messages, high data quality, and integration of data (Jones and Beatty, 1998; Cox and Ghoneim, 1996; Arunachalam, 1995). EDI is an interactive medium from which businesses only derive full benefit if there is a broad access. A critical mass is essential (Jones and Beatty, 1998; Premkumar et al., 1997; Iacovou et al., 1995). Even though universal access is optimal (Markus, 1987), universal access is not essential. However, the more users there are in the EDI community the greater the benefits (Mukhopadhyay et al., 1995) even though some authors have drawn attention to possible negative network externalities (Riggins et al., 1994; Wang and Seidmann, 1995) critical mass has generally been
168
found to be essential. Literature (e.g. Ramamurthy et al., 1999) has shown that especially small companies do not derive full benefits from EDI since they still have to maintain their traditional paper-based routines because they only exchange EDI messages with a few customers. It is thus recognized that EDI is both an inter- and intraorganizational information system, which promises a broad variety of benefits if full integration and critical mass is realized. The adoption of EDI has none the less been considered to be more complicated than adoption of any other interorganizational technology due to the uncertainties related to the increase in interdependence and a more vulnerable position with respect to competitors (Hart and Saunders, 1997). 5.4.3 Views of EDI EDI can be seen from a broad variety of perspectives. Recent research has focused on the technical challenges of EDI (Henriksen and Görsch, 2000; Jui-Lin Lu and Hwang, 2001) due to the development of new means of transportation of data, e.g. Internet (Falch, 1998) and new programming protocols, e.g. XML (Jui-Lin Lu and Hwang, 2001). With the introduction of the Internet an alternative to the traditional VANS-operators (Value Added Network Services) has been introduced (Kalakota and Whinston, 1996). According to some authors this has led to a decrease in the cost of exchanging EDI messages and especially a decrease in the cost of establishing an EDI link (Falch, 1998). This may have resulted in a more attractive environment for EDI. Another research stream has focused on the standardization issue (Damsgaard and Truex, 2000), which by some researchers has been considered as one of the major obstacles for diffusion of EDI (Kimbrough and Moore, 1992). The problem of different standards makes it hard especially for SMEs to adopt EDI because they have to adapt their internal systems to different business partners using different standards (Chen and Williams, 1998). Besides, they have to maintain their traditional paperbased routines (Ramamurthy et al., 1999). One issue worth noting in relation to lack of common standards is that research has found that the
169
perceived problem might be larger than the actual problem. Drury and Farhoomand (1996) found in a survey of the attitude towards EDI among non-adopters and adopters, that 29 percent of the responding non-adopters found that the standardization issue was a major impediment to adoption of EDI whereas 13 percent of the responding adopters shared that same opinion. Three types of formats have dominated EDI messages throughout its history. The EDIFACT format, which is developed and maintained by the UN, the ANSI x.12, which is primarily used in North America, and the proprietary formats, which to a large extend have been used as a lock-in mechanism by larger organizations towards their suppliers. Yet another research stream has focused on integration issues (Massetti and Zmud, 1996; Ramamurthy et al., 1999; Truman, 2000; Premkumar et al., 1994). Integration can be viewed as a technical issue facing the technical possibilities and challenges of integrating EDI messages in the intraorganizational information systems instead of just having computer to computer communication. Integration is found to be a prerequisite for achieving the potential operational benefits related to EDI. From an economic point of view the direct and strategic benefits of EDI have played a central role (O’Callaghan and Turner, 1995). The discussion of strategic benefits has been initiated early in the EDI research history at a very general level (Dearing, 1990). Specific issues related to the strategy discussion have included the strength of critical mass. It has for example been argued that if enough suppliers adopt EDI then it may become a competitive necessity for late joiners. They simply have to join in order not to lose important customers (Riggins and Mukhopadhyay, 1999). The economic and organizational implications of EDI will be discussed in more details in the following sections. Though technical issues and the standardization discussion are of great importance, much attention is paid to the business aspects of EDI usage including, organizational and managerial issues. The adoption of EDI brings about a certain degree of change in the organization. Actually any individual in the organization from executives to blue-collar workers can
170
be affected by adoption of EDI. Restructuring and training of personnel is therefore crucial. As the following review of ten years of publications in the top-five MIS journals reflects there is a tendency to focus on improved performance, strategic or operational, due to the use of EDI whereas the importance of organizational structure issues are underplayed. The business value of EDI is thus the dominating theme in the twenty reviewed articles from MIS journals. 5.4.4 Review of one decade’s publications on EDI from a MIS perspective Though EDI had been used extensively in the industry for several years69 the publications on EDI in the selected MIS journals were sparse during the first four years of the decade from 1991 to 2000. 1994 was the breakthrough for EDI articles in the five reviewed MIS journals. That year five articles appeared. From 1995 and afterwards the numbers have been three to five per year. The articles differ in focus such as standards (Kimbrough and Moore, 1992), network externalities (Riggins et al., 1994; Wang and Seidmann, 1995), benefits for initiators or adopters (Chatfield and Yetton, 2000), and adoption (Premkumar et al., 1994; Iacovou et al., 1995). The focus of investigation also varies in two dimensions. Most studies focus on the interorganizational context whereas other focus on the intraorganizational context (Teo et al., 1997; Chatfield and BjornAndersen, 1997; Srinivasan et al., 1994). So even though all authors agree that EDI is an IOS several articles look at the benefits – or drawbacks – of EDI from an internal point of view. The second dimension in relation to focus is related to the unit of analysis. A number of articles investigate a single organization (Srinivasan et al., 1994; Jelassi and Figon, 1994; Mukhopadhyay, et al., 1995; Chatfield and Bjorn-Andersen, 1997), others look at a business sector (Clark and Stoddard, 1996; Truman, 2000; Chatfield and Yetton, 2000), while others look at organizations broadly (Kimbrough and Moore, 1992). Another categorization of unit of analysis is made in relation to the role of organizations that are subject to investigation. Some articles focus on the 69
For example, the Chrysler Corporation implemented EDI in the mid 80s (Srinivasan et al, 1994).
171
perspective of the adopters (Premkumar et al., 1994; Massetti and Zmud, 1995) while others look at the initiators (Riggins et al., 1994; Chatfield and Yetton, 2000). A number of authors apply the view of the suppliers (Wang and Seidman, 1995; Barua and Lee, 1997; Hart and Saunders, 1998). All articles except of a few views EDI in relation to performance. The benefits, strategic or operational, are thus examined. Only a few articles focus on issues such as power (Hart and Saunders, 1998; Lee et al., 1999; Iacovou et al., 1995), technology (Riggins and Rhee, 1998), or adoption, implementation and diffusion (Premkumar et al., 1994; Iacovou et al., 1995). The twenty articles can be categorized in two broad groups. The first group focuses on the level of analysis. These articles are divided into studies that specifically target a particular organization, business sector, or community. The second group includes issues related to barriers and incentives for adoption of EDI. However, the second group of articles is not very welldefined since not all articles directly or indirectly relate to adoption. Viewed from a factor approach (See Section 6.3.2) related to adoption of EDI these studies do however present a number of factors which directly or indirectly may explain adoption of EDI. With the increased use of the Internet for commercial purposes the traditional VANS based EDI has been challenged. This had led to a technical redefinition of EDI. The redefinition is largely manifested in the replacement of proprietary communication protocols by the TCP/ IP communication protocol (Truman, 2000). One contribution to this trend is by Riggins and Rhee (1998). In their conceptual article they suggested that it was useful to understand the differing views of e-commerce that have emerged from the use of Internet for commercial transactions. 5.4.5 Micro, meso, and macro levels of analysis The micro level of analysis is characterized by focusing on a single organization and the benefits or drawbacks gained from EDI adoption and implementation. Two studies have concentrated on EDI use in the Chrysler Corporation (Srinivasan et al., 1994; Mukhopadhyay et al., 1995). Both
172
studies focused on the operational benefits of EDI. Srinivasan et al. focused on the efficiency related to accurate exchange of information among trading partners, which supported better shipment performance and thus optimized JIT (just-in-time) operations. Mukhopadhyay et al. (1995) conducted an ex post analysis of the EDI program at Chrysler and assessed the management’s goal of reducing manufacturing and logistics costs and streamlining operations for JIT. Using a framework with four components (locus of impact, magnitude of impact, production/ operation process, and exogenous or economy-wide trends that may mitigate the results) for assessing the business value of EDI, the authors concluded that the total tangible benefits of EDI resulted in more than $100 savings per vehicle produced. The study by Mukhopadhyay et al. is unique in the sense that exact numbers quantify the savings. Jelassi and Figon (1994) focused on a single organization in France. The authors examined a distributor of office supplies in relation to customer service, lead-time, management costs, and new business opportunities. By looking at the relations towards the customers it was found that the outcome of the implementation of EDI had led to both operational and strategic benefits. Chatfield and Bjorn-Andersen (1997) explored the relationship between the focal firm’s ability to exploit IOS-enabled business process change and business outcomes. Using the case of Japan Airlines the authors paid particular attention to the ways in which IOS not only contributed to improved competitiveness, but also to how IOS enabled Japan Airlines to leverage its strategically important value-chain as an engine of growth and a new source of sustainable competitive advantage. In order to stay competitive the management in Japan Airlines found it prudent to reduce paper-based transactions. This decision led to the development of EDI applications in a short period of time. EDI was mainly used for strategic purposes, with focus on value-chain logistics coordination and cost reduction. The common feature of the four micro level studies, which focus on a single organization, and the achievements of that particular organization is that all conclude that the organization has benefited from the use of EDI
173
both with respect to operational savings and strategically. The three organizations are all initiators of EDI in relation to their suppliers and customers. Several studies (Gebauer and Buxmann, 2000; Swatman and Swatman, 1992) have pointed at the difference in advantages of EDI depends on whether the organization is an initiator or a follower. Research on that issue will be presented later in this section. Right now emphasis is on the unit of analysis in relation to the level of investigation. Departing from the micro perspective of a single organization, where EDI was found to be beneficial, the meso level is the next level of analysis. Three articles (Clark and Stoddard, 1996; Truman, 2000; Chatfield and Yetton, 2000) deal with sectors at industry level. Clark and Stoddard examined the relationship between technological and process innovations in the US grocery sector. They found that technological and process innovations are interdependent and that both are needed to capture the potential benefits of EDI implementation through interorganizational process re-design. The investments in EDI could be justified on the basis of costs, but the largest pay-off for retailers and manufacturers result from the combination of process and technological innovation. The authors underlined this statement by stating that, EDI combined with process innovations can provide dramatic performance improvements and can even enable new channel structures to emerge. Whereas EDI without process innovation provides little or no benefit for firms investing in these capabilities. The benefits of EDI at the industry level were also subjects of investigation in the study by Truman (2000). Truman had two objectives for his study of EDI usage in the insurance industry. The first objective was to theoretically discuss and empirically examine the relationship between EDI’s integration with internal systems and performance outcomes. The second objective was to theoretically discuss and empirically examine the relationship between the two integration concepts, interface integration and internal integration. Based on quantitative data it was found that there was a positive association between interface integration and performance outcomes. It was also found that there was a positive association between
174
interface integration and internal integration. The third and final study at the meso level is by Chatfield and Yetton (2000). Chatfield and Yetton looked at EDI initiators in three different industry sectors. Based on these case studies the authors concluded that the benefits for initiators could only be realized if adopters use EDI strategically. Compared to the micro level of analysis the meso level of analysis is more critical to the value of EDI. At the micro level the operational benefits played an important role along with the strategic benefits. The studies at the meso level are more centered on strategic benefits. These benefits are however often found to be uncertain for followers. One study focuses on the macro level. The impact of the TradeNet initiative in Singapore is examined in a case study by Teo et al. (1997). The authors examined how the TradeNet had changed the organizational structures and assessed the strategic impact of the government initiative. Teo et al. concluded that while the traditional use of IT for organizational automation has led to marginal gains, the innovative use of IT for organizational transformation could lead to phenomenal gains. The authors hypothesized that IT could influence organizations at multiple levels with varying amounts of benefits. The four constructs identified by the authors organizational structure, business processes, business network, and business scope – were all found to be influenced by the TradeNet project. 5.4.6 Adoption issues The research topics related to adoption issues are characterized by a broad span of topics. The topics are however all found to be of explanatory power in relation to adoption and non-adoption in the sense that they represent factors that may or may not influence the decision to adopt. Only a few studies explicitly link their research to adoption and diffusion theory. Two articles (Premkumar et al., 1994; Iacovou et al., 1995) examined adoption and implementation of EDI from a traditional diffusion of innovations perspective (Rogers, 1983, 1995). Premkumar et al. (1994) examined the relationship between innovation characteristics and various
175
aspects of diffusion of EDI. Using a stage model of diffusion the authors studied four variables that were found to capture the various aspects of diffusion of EDI in organizations: Adaptation, internal diffusion, external diffusion, and implementation success. Through statistical analysis of data the authors found that there existed different determinants depending on the aspect of diffusion. In a study by Iacovou et al. (1995) organizational readiness, external pressures, and perceived benefits were found to be the major factors influencing EDI adoption practices of small firms. Iacovou et al. operationalized the above-mentioned three factors. The article concluded that the major reason for small companies becoming EDI capable is external pressure, especially from trading partners. The results from Iacovou et al.’s case study also indicated that subsidies, promotional efforts, and other influence tactics from trading partners could lead to a faster and a more integrated adoption. The power perspective leading to EDI adoption has been examined in a number of articles. The power perspective represents various aspects such as direct power (Hart and Saunders, 1998), encouragement (Lee et al. 1999), and strategic superiority (Barua and Lee, 1997). Subject for investigation was customer power and supplier dependence, and supplier commitment and supplier trust in a study by Hart and Saunders (1998). Hart and Saunders found that the diffusion of network technologies, such as EDI, ought to lead to electronic partnerships building on trust. The authors found that trust was a more influential factor in increasing the diversity of EDI use than exercised power. Lee et al. (1999) surveyed one EDI initiator – the Campbell Corporation and thirty-one of its followers. Lee et al. provided empirical evidence that the companies that were encouraged to adopt EDI by EDI initiators could realize significant productivity gains from their EDI investments. The authors found that significant levels of benefits could be realized from the EDI implementation if EDI is merged with changes of interfirm processes and policies. Barua and Lee (1997) applied a game theoretical approach to examine the incentives for suppliers to join an EDI network in their case study and survey. They found that suppliers might have to join an EDI network out of strategic necessity, due to the presence of an IT efficient
176
supplier. Another finding from the study was that large suppliers are more likely to join EDI networks due to the large risk related to loss of business relations. Chatfield and Yetton (2000) revisited the problem of pay-off for initiators of EDI networks. From their case analysis they found that for the initiator in order to realize strategic benefits, the adopters (followers) need to use EDI strategically in such ways that they also create value through interfirm joint action. Along the same line of thought Riggins and Mukhopadhyay (1994) examined the existence of interdependent benefits that can cause unique problems for initiators of IOS. By recognizing that lack of enthusiasm among trading partners that are forced to implement EDI, and therefore may not use the technology optimally will lead to inability to realize many of the originally anticipated benefits. A natural consequence, the authors argued, was, that it would be expected in the future that powerful buyers would require internal changes in the adopters’ organization in order to achieve the embedded benefits of fully-integrated EDI. During the decade reviewed Kimbrough and Moore (1992) are the only authors that have focused on a problem that appear to be one of the major adoption and diffusion obstacles EDI has faced during its long history: Standards. Kimbrough and Moore argued that advanced communication systems would change the nature and dynamics of switching costs to all contracting parties, leading to much more symbiotic relationships that extend and evolve over time. Second, they claimed that due to the requirements of substantial investment to convert existing systems, firstmover opportunities and barriers to entry might arise depending on the cost of conversion. They argued that EDI is only useful for well-defined business documents thereby ignoring a vast part of business operations. EDI was therefore found to be insufficient for the industries long-term communications needs.
177
The business value of EDI as an incentive for adoption is a general issue in a number of the reviewed articles. One article in particular focused on the economic benefits of EDI based on the premises, which EDI rest on. Focusing on possible dimensions for integration of EDI Massetti and Zmud (1996) developed four measures related to integration of EDI in organizations. Massetti and Zmud suggested that the EDI integration should be measured in four dimensions: Volume, diversity, breadth, and depth. By focusing on these four dimensions Massetti and Zmud suggest that managers can get a useful tool for measuring IS usage thereby linking EDI strategy and operations. Truman (2000) revisited the integration issue when examining the relationship between EDI integration with internal systems and performance outcomes. The two articles by Massetti and Zmud and by Truman are focused on the internal business value of EDI. Lee et al. (2000) explored an interorganizational view of the business value of EDI. They looked at the benefit of information sharing in the valuechain among a retailer and a manufacturer. Based on their econometric modeling they found that information sharing would be especially useful for improving the efficiency of the supply-chains in the high-tech industry. EDI has traditionally been viewed as creating positive network externalities due to its increased usefulness in relation to other users on the network. Network externalities have been examined in two articles (Riggins et al., 1994; Wang and Seidmann, 1995). Referring to Malone et al. (1987), Riggins et al. (1994) characterized EDI as an electronic hierarchy. Based on this framework Riggins et al. argued that negative network externalities may exist for suppliers due to the fact, that as the number of suppliers on the network increases the network externalities decreases. The suppliers found it to be more difficult to reap the economic benefits from joining the network, because they had to consider and take care of the individual needs of their customers. The authors focused on the adoption-decision in relation to the initiator of the EDI network. Based on their analytical modeling the authors concluded that there to some degree are negative or sub-optimal network externalities connected with joining an EDI network. Wang and Seidmann (1995) found similar to Riggins et al. (1994) that adoption of
178
EDI can lead to negative network externalities. Based on their econometric model Wang and Seidmann concluded that suppliers’ adoption of EDI could generate positive externalities for the buyer and negative externalities for other suppliers. 5.4.7 Summary of a decade’s EDI research published in the top-five MIS journals To summarize, the twenty articles included in the review of MIS publications on EDI are mainly focused on the improved performance that can be achieved by adopting and implementing EDI. Especially the strategic benefits play a major role. The studies presented in Section 5.4.5 on the micro, meso and macro level are concentrated on achievements in relation to both strategic and operational benefits. It is recognized that initiators of EDI gain both operational as well as strategic benefits. The operational benefits include direct savings, shortened lead time, reduced management costs, and more accurate information. The strategic benefits included new business opportunities and increased competitiveness. It was however recognized, that the strategic pay-off were less likely when initiators and adopters use EDI simply to automate existing interfirm information flows and decision processes. It follows that a firm’s unwillingness to re-engineer its internal systems is a barrier to capturing the strategic benefits of EDI (Chatfield and Yetton, 2000). It was thus recognized that mere adoption of EDI is not enough. The adoption, implementation and diffusion themes examined during the decade is concentrated on incentives and barriers. Some studies took offset in the traditional diffusion theory (Premkumar et al., 1994; Iacovou et al., 1995). These adoption studies applied a factor approach in their examination of determinants for adoption. Especially factors such as pressure (Iacovou et al., 1995), power (Hart and Saunders, 1998), encouragement (Lee et al., 1999), and strategic superiority (Barua and Lee, 1997) were found to play an important role for the adoption of EDI. Two studies (Chatfield and Yetton, 2000; Riggins and Mukhophadhyay, 1994) focused on the pay-off for the EDI initiators. Both studies found that the initiator in order to realize benefits from the EDI investments have to set certain conditions for the adopters.
179
The articles focusing on barriers for EDI adoption were concentrated on standards (Kimbrough and Moore, 1992) and network externalities (Riggins et al., 1994; Wang and Seidmann, 1995). In two studies on network externalities it was concluded that negative network externalities could emerge in relation to EDI adoption. Table 5-3. Classification of EDI research themes in the 1991 to 2000 top-five MIS journal review Focus Operational performance
Exemplified by General performance improvements Accurate exchange of information Reduction of logistics costs Information sharing Improved customer service Reduced lead-time Benefits related to EDI integration
Strategic performance
Network externalities New business opportunities Interdependent benefits Competitive advantage Value-chain logistics coordination Changes in interfirm processes and politics Trust
Pressure
Sociotechnical issues
External pressures Power Strategic superiority Regulation (TradeNet) Standards Significance of the Internet
Source Clark and Stoddard, 1996 Srinivasan et al., 1994 Mukhopadhyay et al., 1995 Chatfield and Bjorn-Andersen, 1997 Lee et al., 2000 Jelassi and Figon, 1994 Jelassi and Figon, 1994 Massetti and Zmud, 1996 Truman, 2000 Premkumar et al., 1994 Riggins et al., 1994 Wang and Seidmann, 1995 Jelassi and Figon, 1994 Chatfield and Yetton, 2000 Riggins and Mukhopadhyay, 1994 Chatfield and Bjorn-Andersen, 1997 Jelassi and Figon, 1994 Chatfield and Bjorn-Andersen, 1997 Hart and Saunders, 1998 Lee et al., 1999 Iacovou et al., 1995 Hart and Saunders, 1998 Lee et al., 1999 Barua and Lee, 1997 Teo and Tan, 1997 Kimbrough et al., 1992 Truman, 2000 Riggins and Rhee, 1998
In Table 5-3 a classification of EDI research themes of the 1991 to 2000 top-five MIS journal review is presented. Four EDI research themes are 180
found to dominate the two groups (micro, meso, and macro level of analysis and adoption issues). These four themes are: 1) Operational performance, 2) Strategic performance, 3) Pressure, and 4) Socio-technical issues. The exemplification of one or more of the four foci is related to the reported outcomes of the studies. As illustrated in Table 5-3 the majority of these studies focused on operational and strategic performance. A straightforward answer to the research question, “How are the explanatory variables related to motivators for adoption defined in MIS research at present?” is therefore that the bulk of explanatory variables related to motivators for adoption as defined in MIS research during the last decade of the twentieth century are related to issues concerning organizational performance.70 Performance can be related to operational performance as well as strategic performance. Operational performance is related to improvements in storing, transferring, and sharing of information and issues related to cost reductions due to adoption and implementation of EDI. A number of these studies have focused on the necessity of a certain degree of implementation in order to reap the benefits of EDI. Strategic performance is related to benefits (or drawbacks) related to competitive positioning in the business environment and opportunities for getting closer ties to business partners. Going a step further and focusing on the second research question which is related to the explanatory power of the three contexts, the organizational, the environmental, and the technological (Tornatzky and Fleischer, 1990), an interpretation of the research foci and a reclassification is necessary (for a description of the three contexts see Sections 6.7.1, 6.7.2, and 6.7.3). The 70
A reservation should be made in relation to a direct interpretation of the reviewed literature in relation to the research question concerning motivators leading to adoption. Most of the reviewed studies represented ex post assessments of the benefits gained from EDI adoption. Only a few studies (Riggins et al., 1994; Wang and Seidmann, 1995), which were related to theoretical modeling, were concerned with ex ante assessments of benefits or drawbacks related to EDI adoption. Two studies (Premkumar et al., 1994; Iacovou et al., 1995) explicitly included the adoption theory and thereby rigoursly pursued those considerations related to adoption and motivators for adoption. The majority of the studies did therefore not directly have adoption motivators in mind in their investigation of the field and in their reporting of findings.
181
operational performance focus is mainly related to the organizational context where operational and procedural needs are of major interest. A number of research themes were related to the organizational context. For example general performance improvements and improved customer service. The environmental context did, however, seem to dominate the research themes and resulting findings in the MIS articles. Strategic performance is related to the environmental context, which is concerned with the competitive environment organizations operate in. Included in the environmental context are also those research themes related to pressure and the socio-technical item related to regulation. From a MIS perspective technological issues play a minor role in relation to EDI research. Most of the research themes pursued in the top-five MIS review focused on operational and strategic improvements. This suggests that the conceptualization of technology and EDI generally is rooted in the materialistic-volunteeristic paradigm. The major belief in the materialisticvolunteeristic paradigm is that technology more or less shapes and determines human behavior. The benefits related to efficiency are therefore found to be within reach given the technology works as planned. The rationality for the adoption on the other hand provides evidence of the relations-based rationality (e.g. Hart and Saunders, 1998) and the technoeconomical rationality influenced the views on technology in the studies. However, the socio-political rationality is not explicitly represented in the twenty articles that cover EDI research from a MIS perspective in the last decade of the twentieth century.
5.5 Revisiting the TDP case Most of the studies in the top-five MIS review were theory-driven and they represented as such an explicit research agenda where a specific research outcome is pursued. As mentioned in Section 2.7 the empirical part of this dissertation is based on a practice-driven approach. The empirical design has been guided by an exploratory search for explanatory factors motivating adoption of EDI in the Danish steel and machinery industry. At this point it is found relevant to examine whether the findings from the 182
TDP match the research themes and outcomes from the top-five MIS review.71 The reported motivation for adoption in the TDP case was (cf. Section 4.3.2.5) mainly to be found in relation to strategic performance. The adopters of EDI stated that strategic benefits played a more important role than operational benefits in relation to motivation for adoption. Similar to the reported studies in the top-five MIS review there was found to be a significant difference in the reported motivators depending on whether the adopters were initiators or followers. The EDI followers in the TDP were not as interested in the significance of the possible improvements related to strategic performance as the initiators. The EDI followers had been subject to pressure from their parent company or other companies in their industry group. The EDI initiators on the other hand wanted to gain a competitive lead and to set the standard for EDI exchange of messages in their industry. The reasons and considerations for non-adoption (cf. Section 4.3.3.4) were to a large degree related to costs of purchasing and integrating EDI software. The absence of pressure from business partners and lack of strategic incentives were the immediate reasons for non-adoption. The characterization of the non-adopters in the TDP obviously does not match the summary of the top-five MIS review. This is not surprising when the sources are taken into consideration. As noted in footnote 70 the reviewed studies reflect ex post assessments related to EDI adopters. There is to a high degree of correspondence between TDP data and the EDI themes in MIS research in relation to the TDP adopters. That is at least the case if the assessment is based on self-reported information from the TDP participants. Factors related to company size, position in supply-chain, and legal ownership seemed to play an important role in relation to adoption or 71
This exercise may not be a fair representation of the each of the twenty articles, nor may it be scientifically correct to put twenty articles, which were generated through different means of methods and in different environments into one common pool. If the decade’s top-five MIS review on EDI research on the other hand is seen as a common body of knowledge about EDI then the summary in Table 5-3 is a generic representation of EDI themes of interest.
183
non-adoption in the TDP case. None of the articles in the top-five MIS review paid particular attention to these issues.
5.6 Summing up Chapter 5 As stated in the beginning of this chapter the primary objective was to provide an answer to research question 1a, which comprises the explanatory variables related to motivators for adoption of EDI defined in MIS research. In the pursuit of the explanatory variables it was found relevant to conceptualize IT in general in order to get to a better understanding of how technology is perceived amongst users (and researchers). The second theoretical perspective that was explored was the rationality of technology. The rationality of technology is seen as another tool for understanding what adopters may expect to gain from technology. After the identification of definitions of IOS and EDI different perspectives on EDI were presented leading to the specific MIS context of EDI. Explicit adoption themes were not prevalent in the decade’s MIS research reviewed. The review however, illustrated a number of beneficial features related to EDI that may be viewed as motivators for adoption. These motivators did not represent a set of clear defined explanatory variables. They were accordingly grouped into four categories representing: 1) Operational performance, 2) Strategic performance, 3) Pressure, and 4) Socio-technical issues. Especially the themes related to performance were found to play a prominent role in MIS research during the decade. The answer to research question, “How are the explanatory variables related to motivators for adoption defined in MIS research at present?” is therefore that the explanatory variables in the present MIS research related to motivators for EDI adoption are dominated by performance, pressure and socio-technical issues.
184
6 Adoption of IOS “What makes IOS adoption special compared to other organizational adoption: IOS projects are inherently more risky than traditional internal IT projects because there is less control due to the uncertainty of external trading-partner actions. Additionally, since interorganizational systems often have interdependent benefits, the way in which a trading partner implements a system may affect the benefits realized by the other party.” (Riggins and Mukhopadhyay, 1999).
6.1 Introduction This chapter presents the theoretical foundation for this dissertation: The motivation for adoption of IOS innovations. Three elements are inherent in this theoretical paradigm: Adoption, motivation and innovation. Each of these three elements will be discussed in the present chapter. Based on the presentation of these three theoretical elements an answer to research question 1b, “Which models are used to explain the motivation for IOS adoption at present?” is provided. In Section 6.1 the relevance of the research project is outlined. A short overview of previous research in adoption and diffusion of IT is presented and discussed in relation to the present research project and the basic assumptions guiding the research in relation to diffusion theory is outlined. The author’s underlying conviction is that the traditional adoption models per se are not interesting in this context. Instead underlying constructs in relation to innovation and motivators leading to adoption are in focus. The implications of this view is that the innovation attributes and the mechanisms that influence the adoption-decision will be discussed rather than attempting to improve the generic adoption model such as for example the one outlined by Rogers (1995) (cf. Figure 6-1, page 195). The
185
theoretical postulate is that in order to get a better understanding of motivators for adoption of IOS a closer look at the innovation attributes and motivators for adoption is more relevant than merely examining sequential adoption models per se. Much innovation-diffusion theory is based on the assumption that potential adopters voluntarily choose to adopt or not adopt an innovation based on the benefits expected from the use of the innovation (Fichman, 1992). The underlying argumentation for this view is that motivation and innovation attributes cause adoption. Innovation attributes can be abstract as well as concrete. The concrete attributes related to EDI were presented in the previous chapter when IOS and EDI were presented. The abstract attributes are described and discussed in the present chapter. Though the voluntary element in relation to adoption of IOS can be questioned it is presumed that some specific factors influence motivation for adoption. These mechanisms are related to innovation attributes along with social, political, and economic interests. This view is pursued theoretically and empirically through a presentation of one of the more dominant adoption and diffusion models outlined by Rogers (1995). Along with the presentation of this model the suitability of the model as an explanatory model in relation to data from the TDP is discussed. Based on the ongoing evaluation and discussion in this chapter of the model conceived by Rogers an alternative model for further analysis of the adoption motivators in the Danish steel and machinery is examined. The alternative adoption model used is based on the framework suggested by Tornatzky and Fleischer (1990). The implicit objective is hence to examine the suitability of the traditional adoption and diffusion theory in relation to adoption of a specific technological innovation. This is an extension of the argument given by Fichman (1992). Fichmann suggested that diffusion theories need to be tailored to the adoption context. In this particular context tailoring is done both in relation to adoption context, small businesses, and adoption content, which is IOS with special emphasis on EDI.
186
Several authors (Tornatzky and Klein, 1982; Prescott and Conger, 1995; Tornatzky and Fleischer, 1990) have already suggested that alternative models to Rogers’ adoption model are required in order to understand adoption of technological innovations in organizations. They have presented a critique of the traditional adoption theory as presented by Rogers. Recent research (Lyytinen and Damsgaard, 2001) has once again stressed the need to find alternatives to the traditional view of adoption, which considers technology adoption as dual and thus fails to recognize adoption of only selected fragments (Rai and Yakuni, 1996). None of the mentioned authors have however, operationalized the factors motivating adoption of EDI and examined these factors empirically in relation to EDI. One way of solving the problem of not having an explanatory theory is to narrow the focus to more specific innovations and contexts and thus develop a more powerful explanatory theory around the distinctive characteristics of those innovations and contexts (Fichman and Kemerer, 1994). This strategy is only conditionally accepted by the author in the sense that the innovation that is subject for investigation is an IOS. IOS technologies are (cf. Chapter 5) complex. When looking at the determining factors for adoption a number of variables have therefore to be considered. One issue that is most relevant in relation to motivators for adoption is the perception of the adoption process. Some authors have argued that the adoption process reflects rational behaviors (Cooper and Zmud, 1990) while others have suggested that adoption is determined by for example political and social constraints (Markus, 1983; Lyytinen and Damsgaard, 2001; Premkumar and Ramamurthy, 1995). Bounded rationality (Simon, 1977) is the premise for the presentation of a theoretical framework for adoption is an acceptance for the present study. This implies that the motivators for adoption can be driven by less rational behaviors such as political and social elements. The purpose of Chapter 6 is to provide an overview of adoption and diffusion models and adoption research in general. Data from the TDP is used to assess the suitability of Rogers’ (1995) theory in relation to
187
adoption of IOS. After an assessment of the general adoption theory, research related to IOS and EDI will be explored for the purpose of illustrating elements of importance in relation to adoption of IOS innovations. In the previous chapter IOS and EDI were presented in detail. In this chapter they are conceptualized in relation to existing classifications of innovations. Innovation plays a central role in relation to adoption and diffusion. Adoption and diffusion of innovations as a research stream reflects numerous facets. In order to narrow the field of interest some restrictions are needed. First of all research dealing with the development of innovations (Anderson and Tushman, 1997; Schon, 1963) and organizational innovativeness (Damanpour and Gopalakrishnan, 1998) are omitted. The entire discussion concerning innovation in relation to core competencies from a resource-based view (Prahalad and Hamel, 1990) is also omitted. A vast literature has focused on this particular view in relation to economic and regulatory regimes in relation to innovative environments (Lundvall, 1999). This research stream is also omitted. Additionally, management of innovative organizations (Van De Ven, 1986) is considered to be beyond the scope of this study. In other words focus is on the outcomes of adoption of an innovation rather than processes leading to development of innovations. When referring to innovation in this context, innovation therefore refers to an acquisition of something, which is perceived as new in the organization whereas the creativity per se being part of innovativeness in relation to the development of for example EDI software is excluded. The sequence of the chapter is as follows. The traditional adoption and diffusion theory is presented in order to conceptualize adoption and diffusion. Thereafter different approaches towards adoption and diffusion are outlined. The purpose of this exercise is mainly an academic discipline to illustrate the multiplicity of approaches to adoption and diffusion. Adoption motivators as defined by Rogers are presented next. Innovation attributes are included in the adoption motivators defined by Rogers. Due to the importance of the characteristics of the innovation the term
188
innovation is defined and explored in depth. The purpose of the examination of the concept innovation is to link adoption motivation to the content of this dissertation. Hence, the innovation typology is used as a generic conceptualization of IOS. The traditional adoption motivators defined by Rogers are discussed in relation to the TDP data. As such it is an acceptance of an invitation from Lyytinen and Damsgaard (2001) to challenge the traditional adoption and diffusion theory in relation to EDI. After discussion of pro et contra of Rogers’ theory in relation to the TDP data, models for organizational adoption are presented and discussed. The next step is to describe IS research, which has applied an adoption approach. Finally, Chapter 6 ends with a presentation of a relevant model that can embrace adoption motivation in relation to EDI. The aim of this effort is to present a model, which can be operationalized in a quantitative investigation of the steel and machinery industry with respect to factors motivating EDI adoption.
6.2 Adoption and diffusion Adoption is considered to take place in a singular unit e.g. an organization (Rogers, 1995; Tornatzky and Fleischer, 1990). When adoption takes place across a business community then diffusion of an innovation is said to happen (Rogers, 1995). Adoption is of primary interest in this context, whereas the successful adoption in the business community leading to diffusion plays a secondary role in this analysis. 6.2.1 A definition of adoption There are two dominant views on adoption. Adoption can be seen as having or not having an innovation (Tornatzky and Fleischer, 1990) or it can be seen as using the innovation versus not having it (Rogers, 1995). Adoption is according to Rogers, “… a decision to make full use of an innovation as the best course of action available and rejection is a decision not to adopt an innovation.” (ibid.). According to this view the line between adoption and use of the innovation is thin or non-existent. Rogers argues (cf. Figure 6-1, page 195), that the adoption process is the process through which a decision-making unit passes from first knowledge of an innovation, to 189
forming an attitude toward the innovation, to a decision to adopt or reject, to implementation of the new idea, and finally to confirmation of this decision. In this course of events the adoption process is considered to be merely a mental exercise until implementation takes place. In this dissertation the core understanding of the term adoption is “having versus not having” (Tornatzky and Fleischer, 1990) rather than “not having versus using” (Rogers, 1995). As a consequence this mental exercise is the absolute minimum requirement separating adoption from non-adoption. If the mental exercise has led to some form of authoritative commitment among decision-makers and the point of purchase or development is within a short timeframe (Tornatzky and Fleischer, 1990) then the line between non-adoption and adoption has been crossed. The important point is that some dividing line is crossed where the participants decide to invest resources necessary to accommodate the effort to change (Kwon and Zmud, 1987). Adoption is thus a mental or financial commitment towards the innovation or a physical acquisition of the artifact. The use of the term adoption is therefore exclusively related to the initiation stage as opposed to the implementation stage. Organizational implementation and the obstacles related to IS implementation72 are consequently not considered. 6.2.2 Diffusion “Diffusion is the process by which an innovation is communicated through certain channels over time among the members of a social system.” 72
Assessment of the implementation of complex technological innovations is a complicated task in several ways. One aspect is that implementation requires both individual and organizational learning (Attewell, 1992). These two areas are, however, entire research topics in themselves. Another aspect is that the assessment of implementation in relation to information systems such as EDI is complex. Massetti and Zmud (1996) extensively deal with this issue in their article on quantitative measurements for the use of EDI in organizations. However, a number of researchers have worked with EDI implementation (Chan and Swatman, 1998; Chan and Swatman, 2000; Cox and Ghoneim 1996; Riggins and Mukhopadhyay 1999) in relation to improvement of internal processes, metrics to measure effects etc. This literature is not included because these two stages in the diffusion process are viewed differently. The adoption and implementation literatures are raising substantially different issues. Where adoption research mainly relates to ex ante considerations the implementation research involves ex post assessment.
190
(Rogers, 1995). The distinction is therefore that while adoption refers to the decision to invest resources necessary to accommodate the change effort (Kwon and Zmud, 1987) diffusion refers to the spread over time of an innovation within the unit or to other units (Rogers, 1995). An examination of diffusion in the steel and machinery industry would have required a completely different research design where e.g. community networks and quantitative assessments of time of adoption had been the foci of interest. However, before leaving the issue of diffusion completely a few comments on the conceptualization of diffusion are made in order to define and clarify the term, which after all is closely related to adoption. The diffusion process can be illustrated in three different ways: The Sshaped curve, the exponential curve, and the gravity model (Attewell, 1992). The S-shaped curve describes diffusion over time. At first few organizations adopt, then there is a sudden take off followed by a slowing rate of adoption. Economists explain the S-shaped curve in terms of the shifting balance in supply and demand, where the sudden take off is due to substantial drop in the price of the new technology (ibid.). Though the market has experienced a drop in the price of EDI software, for example the TDP-solution, a take off in relation to EDI adoption has not manifested. Diffusion illustrated as an exponential curve is related to interactive communications media, such as EDI, where adoption becomes more and more attractive when more adopt it. Lynne Markus is exponent of this critical mass theory as a conceptual model for diffusion (Markus, 1987).73 Similar to the absence of the take-off of the S-curve, the exponential curve has also not yet manifested itself with respect to EDI adoption. The third, and perhaps most widespread conceptualization of diffusion is the gravity model also called the spatial approach. The gravity model predicts 73
Positive network externalities, which are related to the critical mass view, have however, been subject to discussion in relation to EDI (Riggins et al., 1994; Wang and Seidmann, 1995). The determining factor which leads to positive or negative network externalities is whether or not the organization that joins an EDI network can differentiate itself by joining the network. The potential benefit a supplier can achieve by joining is dependent on whether or not competing suppliers join (Riggins and Mukhopadhyay 1999).
191
diffusion based on a function of the population size of an area and the distances of that area from other centers of populations (Attewell, 1992). Diffusion is according to that view primarily a social process (Rogers, 1995). One criticism of the gravity model is that it seem to be a useful explanation for diffusion of social innovations where individuals are adopters, but it appears to be a less useful explanation for adoption and diffusion of organizational innovations (Attewell, 1992). An illustrative example of this paradox can be taken from the TDP. Though the business associations introduced EDI as an attractive organizational innovation in the business community by launching the TDP, the network of organizations was not susceptible to this initiative. The traditional understanding of the word diffusion is, “The spread of cultural elements from one area or group of people to others by contact.”74 Research on this spread of cultural elements, which in this case represents technological innovations in the form of EDI, has been approached from different angles. Researchers have suggested diffusion perspectives such as supplier-focused, user-focused, and fashion-setting (Newell et al., 2000). Table 6-1. Three diffusion perspectives Diffusion Supplier-focused User-focused focus The active role of Sociological and Diffusion users in the process perspective organizational perspectives Source: Adapted from Newell et al., 2000, pp. 243-249.
Fashion-setting Mechanisms surrounding the diffusion process
The first perspective, the supplier-focused diffusion perspective, is rooted in the diffusion framework exemplified by the work of Rogers (1995). “The supplier-focused perspective is aimed at helping central technology suppliers to promote more rapid diffusion of predefined best-practice innovations to communities of potential adopters” (Newell et al., 2000). The main focus for explanatory factors in the Rogerian diffusion framework is the characteristics of the innovation, the characteristics of the potential adopters, and the communication channels such as social 74
Cf. Merriam-Webster’s collegiate dictionary, www.webster.com/
192
networks. It thus becomes crucial to understand the attributes of the innovation and the social networks through which the ideas are communicated (Newell et al., 2000; Wolfe, 1994). In their critique of the supplier-focused diffusion model Newell et al. draw attention to the fact that the attributes of complex technologies such as EDI are not fixed and rigid. These complex technologies are to a large extent socially constructed. The attributes of complex innovations can thus be unpacked and communicated selectively by groups who have vested interests in promoting or limiting the adoption. The supplier-focused perspective underestimates the importance of the active role that users play in the adoption-diffusion process. Other researchers have looked at the supplier perspective from a push perspective (Drury and Farhoomand, 1999; Rai and Yakuni, 1996; Chau and Tam, 2000). Rai and Yakuni (1996) claimed that technology push not only influenced adoption behavior but could also shape elements of the organizational context, which, in turn, may affect adoption behavior. Drury and Farhoomand (1999) found that technologypush depending on the industry was rarely a driver for adoption. The user-focused diffusion perspective compensates for the lack of including the role of users. According to this view the understanding of adoption must involve considerations of the different communication networks, both internal and external through which individuals gain access to new ideas (Newell et al., 2000). Newell et al. exemplifies the userfocused perspective with the role of professional associations in social networks. They argue that, “The main aim of the professional association is to provide its members with access to information about the latest technological developments in a particular knowledge area.” A third perspective, which is presented by Newell et al., is the fads and fashions perspective. The fads and fashions perspective is rooted in the writings of Abrahamson (1996). Abrahamson defines management fashions as “… transitory collective beliefs that certain management techniques are at the forefront of management progress.” The fashion setters are to be found among management gurus, consultants, and mass media business publications. Newell et al. state that, “In the context of IT in particular the
193
technology suppliers and professional associations have played a key role in disseminating rhetoric about the latest best-practice tools and techniques to managers in industry.” What differentiates the fads and fashions perspective from the userperspective is that whereas the users play an active role in relation to adoption and diffusion of an introduced innovation, managers according to the fads and fashions perspective actively search for the innovations. The major difference in the three perspectives is hence that supply forces drive the user and supplier perspectives, whereas the fads and fashions perspective is driven by demand. In the TDP supply forces rather than demand drove the diffusion. The SMEs in the business environment were exposed to EDI. They did not actively seek EDI as a managerial tool. In the TDP case and in the case of EDI diffusion in Denmark in general the industry and trade associations have played a central role in communicating information about the innovation to the potential adopters. In the TDP there was however, a mix of interests in the communication network initiated by the associations. The professional associations created awareness but the necessary tools were developed and distributed by private vendors. The distinction between the user-focused and the supplier-focused perspectives are hence not clear cut in relation to the TDP. In that particular case the industry and trade associations on the one hand supported the user-perspective and the private software vendors on the other hand supported the supplier-perspective in relation to adoption and diffusion of EDI in the business community. 6.2.3 Adoption and diffusion of innovations Adoption of innovations can be conceptualized in two stages: Initiation and implementation (Zaltman et al., 1973; Rogers, 1995), (See Figure 6-1). The appropriateness - or inappropriateness - of considering different stages in relation to adoption will be discussed below. However, at this point the simplification of the process is accepted for illustrative and pedagogical reasons. The initiation stage includes all activities associated with the decision-makers’ perception, interpretation, and evaluating of the potential
194
innovation leading to a decision, while the implementation stage involves activities pertaining to the actual usage of an innovation (Rogers, 1995; Tabak and Barr, 1999). In the present study focus is on the former stage, the initiation, and in particular on the motivation leading to the adoptiondecision which separates initiation and implementation. The motivators for adoption are viewed as the initiating forces leading to the decision.
Figure 6-1. The organizational innovation process
Decision I. Initiation Stage 1 AGENDA-SETTING
II. Implementation Stage 2 MATCHING
Stage 3 REDEFINING/ RESTRUCTURING
Stage 4 CLARIFYING
Stage 5 ROUTINIZING
Source: Rogers, 1995, p. 392
Rogers (1995) views the diffusion of an innovation as a stage process where particular factors play an important role. Some of the later opponents to the stage view are Newell et al. (1998). Newell et al. argue that it is more appropriate to look at the diffusion process as a number of episodes rather than as a set of continues stages. Newell et al. suggest that episodes do not represent discrete stages. It has also been argued that even though there evidently are some diffusion stages it can not be taken for granted that they occur in the expected order (Wolfe, 1994). Especially if innovations are complex, stages tend to be “muddled and overlapping”, and the process is not simple and linear, but it is rather a complex, iterative process having many feed-back and feed-forward cycles and feed-back loops (ibid.), which may act in positive or negative manners (Kwon and Zmud, 1987). Discontinuation,75 re-invention,76 or even exnovation77 can and do appear
75
Discontinuities are “breakthrough innovations that advance by an order of magnitude the technological state-of-the-art, which characterizes an industry. They are based on new technologies whose technical limits are inherently greater than those of the previous dominant technology, along economically relevant dimensions of merit.” (Anderson & Tushman 1997).
195
especially in the information systems field, where there is a perpetual development of smarter, faster and cheaper solutions (Abrahamson, 1996; Galliers, 1999).
6.3 Three conceptual elements from adoption research In the presentation of adoption theory three elements are used to conceptualize the span of the adoption literature. These elements are found to be distinct and yet overlapping. The overlapping features illustrate the mutual inter-relatedness of the three elements. The purpose of presenting these three elements separately is to portray the complexity of the adoption literature. Besides it is found relevant to position the study in relation to the three elements in order to create an overview of the theoretical views on adoption leading to a better understanding of adoption motivators. The first conceptual element focuses on the main driver for adoption, or to use the terminology of Kling (1980) the rationality for adoption of an innovation. The second element is related to the focus of a given study. Focus can be a factor approach or a process approach. The third element related to the description of adoption is concerned with the explanatory factors for adoption. This element is mainly concerned with identifying a set of generic adoption motivators. While the first two elements are 76
Re-invention is the degree to which an innovation is changed or modified by a user in the process of its adoption and implementation (Rogers, 1995; Slappendel 1996). The adoption of an innovation is thus not necessarily a passive role of accepting a standard template of the new idea (Rogers, 1995). Re-inventions occur when the innovation is changed or modified by users in the process of its adoption and implementation. The organizations might have taken an offset in EDI when they decide to adopt an interorganizational information system. However, since the process from the initial decision to adopt an information system and to the implementation is a prolonged process (Sabherwal and King, 1995) the technological trend might have changed from one managerial trend towards another (Abrahamson, 1996; Galliers, 1999). This might for example be the case for EDI, which might be substituted by business-to-business electronic commerce. Because of the rapid change in technology, those who have not yet adopted EDI might not be seen as ‘laggards’ but as ‘innovators’ (Rogers, 1995) of the next emerging business-to-business electronic commerce technology. 77 Exnovation is the disappearance of an adopted and implemented information technology (Kwon and Zmud, 1987).
196
described and discussed briefly the third element, which is related to the explanatory factors for adoption, is presented in more depth. Especially elements concerning innovation attributes are examined in detail. 6.3.1 Main drivers for adoption Two perspectives have been identified as main drivers for adoption (Attewell, 1992): Economics and social processes. From an economic point of view adoption is primarily assessed in relation to cost and benefit considerations. Ceteris paribus, the higher the cost the lower the rate of adoption. The higher the perceived benefits the greater the rate of adoption. This view is supported by theories emphasizing the S-shaped diffusion curve and the exponential diffusion curve where price mechanisms and demand are used as explanatory variables for diffusion (cf. Section 6.2.2). Adoption explained from a social process point of view emphasizes that adoption and diffusion takes place through social processes and information flows. It is presumed that firms, which are closely connected to users of a given innovation, learn about the innovation and are early adopters. Firms on the periphery of networks are slower to adopt. The adoption theory as presented by Rogers (1995) is based on this social process view. 6.3.2 Approaches to adoption research Most diffusion studies employ either a process or a factor approach (Prescott and Conger, 1995). Broadly speaking the factor approach attempts to identify static forces which lead to adoption whereas the process approach focuses on the dynamics of adoption and implementation, examining the behavior of stake-holders over time (Cooper and Zmud, 1990). Five different research streams have been identified in a review of the IS adoption and implementation literature (Kwon and Zmud, 1987): 1) The factor research stream, 2) The mutual understanding research stream, 3) The process research stream, 4) The political research stream, and 5) The prescriptive research stream. Political processes motivating adoption were outlined in Chapter 3. The political approach to adoption and diffusion of innovations recognizes that the diverse interests of IT
197
stakeholders affect adoption efforts and that successful adoption and implementation depend on recognizing and managing this diversity (Cooper and Zmud, 1990). In the TDP the early adopters did not show any particular interest in supporting adoption of EDI with their business partners involved in the project neither with those companies that had adopted EDI nor with those that were non-adopters. The political approach will not at this point be discussed any further. Instead the factor approach and the process approach, which are found to be of more relevance in this context, are introduced in details below. A process or stage approach involves an in-depth study of the sequence of events leading to an adoption-decision within a firm. The process research stream focuses on activities related to social change rather than technical activities (Kwon and Zmud, 1987). The process approach provides explanations in terms of the sequence of events leading to an outcome (Langley, 1999). The TAM (Technology Acceptance Model) is an example of a theory working with the process approach (Davis et al., 1989; Davis, 1989). The TAM aims at explaining the determinants of computer acceptance across a broad range of end-user technologies and user populations. It is hence an explanatory theory of individual behavior. As opposed to the longitudinal approach, which is the basis process approach, the factor approach examines a cross-section of firms in an attempt to isolate significant characteristics governing adoption (Cooper and Zmud, 1990). The factor or variance research stream focuses on identifying the variables being potentially relevant for IS adoption and implementation effectiveness (Markus and Robey, 1988). The main purpose of the factor approach is to provide explanations for phenomena in terms of relationships among dependent and independent variables (Langley, 1999). On that account a variety of individual and organizational variables have been examined (Kwon and Zmud, 1987). Identifying possible key variables has so far been less successful even though previous work has supported the importance of a number of variables (Prescott and Conger, 1995).
198
The overall purpose of the study of the Danish steel and machinery industry is to identify variables that motivate organizational adoption. The study therefore employs the factor approach with special attention to incentives that motivate the adoption-decision. Though the factor approach generally has been found to be unable to capture the complex and dynamic interactions with business partners in relation to IOS adoption, the search for explanatory variables for adoption of EDI among small businesses is found to be more relevant than an examination of processes leading to adoption (Kurnia and Johnston, 2000). Kurnia and Johnston actually make an exception to their reservation towards the suitability of the factor approach in relation to IOS adoption. They argue that the factor approach is well suited to studies of IOS adoption by small industry partners. The reason is that external forces such as pressure from other more influential trading partners may determine small business partners’ actions. The decision to apply the factor approach will however not completely exclude the process-oriented approach. The distinction between these two approaches is of higher theoretical than practical value since the two types of views and the phenomena related to the approaches in practice are intertwined (Langley, 1999). In this context the underlying discussions of whether or not the process approach or the factor approach is more appropriate is less relevant due to the focus of the present study. Subject for investigation is a particular point in the adoption process. Previous and future activities related to the adoption-decision are not studied. It can thus be argued that the research represents a hybrid of the two approaches in the sense that a particular stage in the process is selected for examination of the adoption motivation across a population. 6.3.3 Explanatory factors There are, according to Rogers, two conditions that primarily determine adoption: 1) The characteristics of the adopters, and 2) The characteristics related to the innovation. Rogers argues that, “… subjective evaluations of an innovation, derived from individuals’ personal experiences and perceptions conveyed by interpersonal networks, drives the diffusion process.” These two conditions are in principle not questioned instead they
199
will be described shortly and discussed in relation to the present study. The aim of this exercise is to examine the appropriateness of the theory in relation to clarification of adoption motivators related to EDI. Re: 1) Rogers has outlined a number of features that increase the likelihood of adoptive behavior of individuals.78 These are, not considered much further. The reason being that the characteristics of the adopters are found to be less important in relation to organizational adoption, where the decision to adopt is made among a group of people rather than by a single individual. In small organizations, which have the primary focus in this study it is presumed that even though personal characteristics play an important role the explanatory factors for adoption are to be found elsewhere. Characteristics related to the innovation and environmental influences such as competitive pressures and constraints are seen as more plausible explanations for adoption than the socio-economic status, personality values, and communication behavior of the individual manager. Re: 2) The characteristics related to the innovation are a set of variables that aim at supporting the “subjective evaluations of the innovation”. These subjective evaluations are found to closely correspond with the term motivation for adoption in the way this term is used in this context. Rogers identifies five categories of variables that are determining the relative speed by which members of a social system adopt an innovation. In the following sections an assessment of the variables determining the rate of adoption is made in relation to EDI in general and to the TDP in particular. The purpose of this exercise is to analyze the suitability of Rogers’ adoption model in relation to the present project.
78
These include three categories of characteristics: Socio-economic status, personality values, and communication behavior. Socio-economic status is operationalized as for example formal education, literacy, and social mobility. Personality variables are operationalized as for example ability to deal with abstractions, ability to cope with uncertainty, favorable attitude towards science. Finally, communication behavior is operationalized as for example degree of social participation, interconnection through interpersonal networks, and exposure to communication channels.
200
Table 6-2. Variables determining the rate of adoption of innovations Variable Perceived attributes of innovations
Type of innovation-decision Communication channels Nature of the social system Extent of change agents’ promotion efforts Source: Rogers, 1995, p. 207.
Exemplified by Relative advantage Compatibility Complexity Trialability Observability Optional Collective Authority Mass media Interpersonal Norms Degree of network Interconnectedness -
The first variable in Table 6-2 “Perceived attributes of innovations” is described and discussed in the next section whereas the remaining variables are described and discussed in Section 6.5. The reason for splitting the five variables in this manner is that especially innovation needs to be explored in depth whereas the description of the four other variables are found to need a less in-depth description and discussion.
6.4 Perceived attributes of innovations The perceived attributes of innovations are related to a set of characteristics of the innovation under consideration. The attributes of innovations have been examined in several studies. Rogers has identified five predominant attributes: Relative advantage, compatibility, complexity, trialability, and observability (cf. Table 6-2) that help explaining the rate of adoption. A number of IS studies have operationalized the five attributes (Moore and Benbasat, 1991). Based on their extensive review of innovation literature Tornatzky and Klein (1982) expanded the list to thirty attributes. This list included the five attributes identified by Rogers. A number of studies have applied these five attributes in their search for explanatory factors related to adoption and diffusion of IOS innovations (Kurnia and Johnston, 2000; Premkumar and Ramamurthy, 1995; Premkumar et al., 1994). Common for 201
all these studies is the acceptance of Rogers’ perceived attributes of innovations as being valid explanations of the diffusion of various kinds of technological artifacts. Lately researchers (Prescott and Conger, 1995; Lai and Guynes, 1998; Lyytinen and Damsgaard, 2001) within IS have questioned whether the Rogerian theory is able to capture the diffusion of IOS innovations based on the innovation attributes identified by Rogers. The concept of innovation is explored before a further discussion of the five innovation attributes defined by Rogers. Mohr (1987) suggested that in order to approach adoption in an appropriate manner it is necessary to focus on innovation. This focus is crucial for obtaining intelligible, manageable information. Wolfe addressed five aspects related to innovation research: “Researchers must clearly address the following points, when studying innovation (1) which of the various streams of innovation research is relevant to a research question; (2) the stage of the innovation process upon which a study focuses; (3) the types of organizations included in the study; (4) how a study’s outcome variable is conceptualized, and (5) the attributes of the innovation being investigated (Wolfe, 1994).
Inspired by Wolfe’s five aspects of importance the following considerations in relation to this particular study will be outlined. The demarcation of the research stream of interest is as mentioned above the adoption of an innovation. From this it follows that the vast literature on development of innovations and management of the innovation processes is excluded. The stage of the innovation process is the motivation leading to the adoption-decision, keeping the reservations towards the notion of stages in mind. The types of organizations included in the study are businesses in the Danish steel and machinery industry, as presented in Chapter 3 and 4. The conceptualization of the study’s outcome variable, adoption based on the organizational, environmental, and technological context, will be presented at the end of this chapter. Chapter 5 presented the attributes of EDI, which is the innovation under investigation. In the following the specific attributes of IOS and EDI will be related to the innovation characteristics described in previous research on technological innovations.
202
6.4.1 Defining innovation In general, when considering diffusion of something it is implicitly diffusion of something new. It does not have to be new in the sense ’never seen before’, but it has to be perceived as an innovation in the particular setting where it is to be adopted and implemented. Broadly defined an innovation is an idea, practice, or object perceived as new by an individual or other unit of adoption (Rogers, 1995; Zaltman et al., 1973). It is the successful introduction into an applied situation of means or ends that are new to that situation (Mohr, 1969). It should be stressed that innovation is a perceived attribute. It means that something can be viewed as new in a given setting even if it to others elsewhere is already passé (Tornatzky and Fleischer, 1990). The distinguishing characteristic of an innovation is therefore that instead of being an external object, it is the perception of a social unit that defines its newness (Zaltman et al., 1973). Therefore, a particular practice can be considered an innovation for one organization, but not for another. Whether it objectively is new or not is not important, as long as it is considered to be new to the adopting unit (Tabak and Barr, 1999). The common characteristic of an innovation based on the abovementioned considerations is that innovation “is the perception of newness among a group of people” (Tornatzky and Fleischer, 1990; Zaltman et al., 1973). The crucial characteristic of an innovation in an organization is, that not all members of the organization may think of the adopted idea, practice or object as an innovation (Zaltman et al., 1973). An illustration of this in many ways paradoxical phenomenon of innovation was found in the TDP. During the time of the project the concept of EDI had been known for years in the business community and among most of the involved businesses, but it was still perceived as an innovation among the non-adopters in the project at least in relation to the organizational setting. The persons involved in the TDP were all familiar with EDI, but the representatives of the non-adopters acknowledged that EDI would be perceived as new in their organizations. In their evaluation of the project the adopters expressed disappointment in relation to the informative value of the project while representatives from the non-adopting organizations stated that they had received useful information. The information was the
203
same but it varied in relevance depending on the persons or units knowledge of the innovation. The TDP case also illustrates how ideas develop over time and get reinvented in new forms. Some of the informants among the users thus argued that their EDI practices over time might be superseded by e-commerce. The examples and the definitions clearly indicate that innovation is not something static. It is rather a mindset that shapes the involved people and their organizations. Tornatzky and Fleischer (1990) and Zaltman et al. (1973) have focused on innovations in an organizational context. The work of Zaltman et al. focuses on units of adoption larger than an individual, which also implies that not all members of an organization may think of the item as an innovation. One implication of this are the possibility of conflicts between those in the organization who view the object or practice in question as an innovation and tend to resist change and those who perceive it as not significantly new and advocate adoption of the product or practice (ibid.). Zaltman et al. argue that the factors that influence individual perceptions of innovation directly or indirectly also influence an organization’s perception. They put forth the following argumentation to support their claim. First, they argue that the age of the organization influences its history of experiences with potential innovation and hence influences the perception of how familiar the practice or product is. The second argument is that organizations identify themselves with other firms that have adopted the innovation. The important point here is that organizations, like individuals, have perceptions and that the nature of their perceptual processes is particularly important. 6.4.2 Classifying innovations Source, type, and effect, three dimensions of an innovation, will be discussed in the following leading to a typology of innovations. The purpose of defining the typology is twofold. One purpose is to be able to identify, how EDI is to be understood in relation to innovation theory in general. Another purpose is to identify a framework, which can be used to classify other innovations, which may posses identical innovation
204
characteristics. This would make it possible to generalize aspects related to EDI to other technological innovations that have a similar profile as EDI.79 6.4.2.1 Source Source of an innovation is related to the origin of the innovation. A distinction is here made depending on whether the innovation is invented in-house or if it is adopted from outside the organization (Zaltman et al., 1973). The distinction between invention and adoption can be viewed from three different angles. Firstly, innovation can be viewed as being synonymous with invention. Here the source is incubative (Damanpour and Gopalakrishnan, 1998) in the sense that the firm develops its own innovations. In that context innovation refers to a creative process whereby two or more existing concepts or entities are combined in some novel way to produce a configuration not previously known by the persons involved. According to Zaltman et al. this understanding of innovation is the most common. Secondly, innovation can refer to a process of adoption. In that case the source of innovation is imitative (Damanpour and Gopalakrishnan, 1998). Here the term innovation is used to describe the process whereby an existing innovation becomes a part of an adopter’s cognitive state and behavioral repertoire. Thirdly, the use of the term innovation is related to a situation where an idea, practice, or material artifact is invented or regarded as being new independent of it being adopted or not. Here the innovation is acquisitive (ibid.). In the first two cases processes are involved whereas the third category emphasizes a description of why something is novel. In the first and the third case the organization can be innovative without adopting. In the second case the individual or the organization can adopt the innovation without being inventive. Whereas invention implies bringing 79
Downs and Mohr, 1976 argue that ”… it is fruitless and premature to attempt to construct typologies of innovations to be used to generalize across a large sample of organizations or sites.” The argumentation for this radical statement is that innovation attributes are essentially interactive with features of the organization, which to a large extent base their preferences on perceived dimensions of the innovation. This caveat does however, not stop the author from developing a typology based on the existing innovation literature. The author thereby supports the argumentation against a total subjectivity of innovation characteristics expounded by Tornatzky and Klein (1982). They questioned whether or not innovation characteristics are based on complete situational specificity.
205
something new into being; innovation implies bringing something new into use (Mohr, 1969). In relation to the TDP EDI adoption of the innovation from outside the organization is the most adequate description. The companies were imitative rather than incubative or acquisitive. The organizations adopted goals, processes or policies that are new in the sense of being departures from their own tradition (Mohr, 1969). Especially in situations where the innovation is externally induced, the adoption of the innovation may create complications and be difficult to carry out. People in the organization have their own interests and perspectives, and therefore there is a tendency for different interpretations of meaning and significance of a given technology (Marcus and Weber, 1989). This may lead to complications in the further adoption process leading to implementation. In the literature review of IOS and EDI (cf. Chapter 5) most companies had adopted an existing innovation and acted according to the opportunities and limitations that were inherent in the innovation. The interpretations of the innovation were in several of the reviewed cases in concert with the objectives defined by the initiators. For example the case of the French office supplier (Jelassi and Figon, 1994) or the Danish State Railroad System (Bjorn-Andersen and Nygaard-Andersen, 1995a) where the adoption of EDI was interpreted as beneficial from both operational and strategic points of view. Objectively the outcome could have been interpreted as less beneficial if other assessment criteria had been used. In the TDP a quite unusual behavior was observed in relation to company “B” which combined the development of EDI with the adoption of EDI in the organization. “B” was thus at the same time both inventor and adopter of an innovation. But, as the IT-manager expressed it, this situation was not optimal and after some time the inventive role was abandoned. The rest of the adopters were in a situation where they adopted an existing innovation. 6.4.2.2 Type Type, the second dimension, is related to the specific characteristic of the idea, practice or object that the innovation represents. Innovation in a
206
business environment may include new production inputs, machines, processes and techniques adopted by firms for their own use (Frambach, 1993). Five main categories of innovation types, which comprise the characteristics suggested by Frambach, are identified by Zaltman et al., (1973): Product or services innovations. This type of innovations is related to creative and inventive behavior. Production-process oriented innovations. For example automated assembly lines and changes in an accounting process. The productionprocess oriented innovations involve changes in the organization’s task systems and in its physical production operations. Organizational-structure innovations. For example decentralizing decision making and instituting incentive systems. Alter the informal and formal systems of the organization. People innovations. For example programs on creative decision making and the use of social research to improve social practice. People innovations entail the replacement of old personnel and the addition of new personnel or training and altering of behavior among the old personnel. Policy innovation. For example involving major changes in the organization’s strategies for achieving its major objectives.
The authors mention that the five types of innovations are highly interrelated. The successful adoption and implementation of an innovation along one aspect depends on how well changes are made along other aspects. How the type of innovation in relation to IOS and EDI is classified depends on how the system is to be used. Three of the types are relevant in relation to IOS and EDI: The production-process oriented innovations, the organizational-structure innovations, and the policy innovation. If IOS and EDI are viewed as means for improving performance then the innovation type is a production-process oriented innovation. There is a broad agreement in the literature reviewed on EDI that improvement in performance is most likely to occur (O’Callaghan and Turner, 1995), and it can be argued that the adoption of EDI as a minimum is a productionprocess oriented innovation. The organizational-structure innovation view of IOS and EDI were presented in relation to BPR in two of the reviewed articles (Chatfield and Bjorn-Andersen, 1997; Clark and Stoddard, 1996).
207
This phenomenon is a familiar issue within MIS research (Markus, 1983). It is widely accepted that organization-structure innovation may alter the informal and formal social system in the organization. In addition it is relevant to view EDI as a policy innovation in those cases where EDI is adopted and used for strategic purposes. In the TDP the initiators of EDI used it to realize their major strategic objectives, whereas the adopters that had been forced (persuaded or coerced) to adopt EDI tended to view the innovation as a means for performance improvement. 6.4.2.3 Effect Effect is the third dimension. Effect is related to how radical the adoption of the innovation influences the organization. Innovations can have three possible patterns of effect (Chau and Tam, 1997; Zaltman et al., 1973): Continuous innovations. Innovations that have little disruptive impact on behavior patterns. The adopted item constitutes only slight alteration of a current practice. Dynamically continuous innovations. Innovations that have moderate impact on behavior patterns. Discontinuous innovations. Innovations that involve the establishment of new behavior patterns. Reorientation takes place as a consequence of discontinuous innovations, which imply fundamental changes. Old routines are eliminated and entirely new ones added.
Especially the discontinuous innovations or radical innovations have caught attention among researchers. A radical innovation, besides being new to the organization, is also radically different from current organizational practices, and thus requires significant changes in organizational processes, behaviors, and structures (Tabak and Barr, 1999). A distinction is made between two types of discontinuous innovations: Competence-enhancing and competence destroying (Tornatzky and Fleischer, 1990). The competence-destroying innovations are characterized by the fact that adoption can cause a firm’s expertise to become obsolete. The competence-enhancing innovations provide the opportunity for a firm to radically improve a product or a process. EDI can best be described as a competence-enhancing innovation, which can improve the organizations’ information flow both internally and externally.
208
As mentioned earlier the degree of EDI usage is often limited to a few business partners and the adoption of EDI can at best be labeled a dynamically continuous innovation.80 In the TDP it appeared as if the nonadopters perceived the change as more radical than the adopters did. At a first glance this should be self-evident. It is however, interesting to note that the non-adopters and the adopters perceive the necessary degree of change differently. Two of the non-adopters expressed that their organization was not ready for such a change, since they expected that adoption would lead to major changes due to the high degree of integration of business documents and the thereby following reduction in the workforce. To the adopters the change is not seen as something radical, maybe due to the elapsed time since adoption or perhaps due to the low level of implementation, which have led to limited discontinuation in the organizational routines. Figure 6-2. An innovation typology
SOURCE • internal • external
EFFECT • little disruptive impact • moderate impact • new behavior patterns
TYPE • product • production • process • people • political
80
It can be argued that EDI is a continuous innovation. If all four dimensions, volume, integration, diversity, and scope (Massetti and Zmud, 1996) are at low levels it is doubtful whether EDI adoption will lead to any change in organizational behavior, since the traditional paper-based routines have to be maintained in relation to the rest of the trading partners.
209
The three innovation dimensions are illustrated in Figure 6-2. As mentioned above it is not possible to classify EDI in an unambiguous manner. Depending on the degree of implementation the dimensions differ. Each of the three dimensions represents different innovation opportunities. Whether these innovation opportunities can be exploited depends on the degree of match between the organizational practices and needs and the innovation adopted. 6.4.3 Rogers’ five perceived attributes determining adoption of an innovation in relation to EDI The five perceived attributes of an innovation determining the rate of adoption identified by Rogers are: Relative advantage, compatibility, complexity, trialability, and observability (cf. Table 6-2, page 201). A review of the diffusion literature has identified thirty innovation attributes (Tornatzky and Klein, 1982). A later review suggested that ten of these thirty attributes have the best explanatory power (Premkumar et al., 1994). The ten attributes identified by Premkumar et al. included the five attributes originally identified by Rogers (1995). This indicates that the characteristics identified by Rogers can hardly be ignored. Moreover, is not the objective of the present study to underplay the importance of these adoption determinants. The objective is rather to emphasize, that additional elements should be included to develop a more solid theory. Relative advantage is the degree to which an innovation is perceived as being better than the idea it supersedes (Rogers, 1995). Relative advantage can be divided into tangible benefits and intangible benefits (Premkumar et al., 1994). Relative advantage is according to Rogers often expressed as economic profitability, social prestige, or other benefits. The general opinion in regard to EDI has been that it can lead to relative advantage. EDI can result in economic profitability for the adopter since a number of operational and strategic benefits can be gained (O’Callaghan and Turner, 1995). This argument was examined in the top-five MIS review on EDI in the previous chapter. It was (cf. Table 5-3, page 180) concluded that economic profitability could be related to both operational and strategic benefits. The reviewed studies were however mixed in their assessment of
210
the actual benefits gained from EDI adoption. Also the degree to which EDI was perceived as being better than the routines it supersedes was questioned especially in the studies related to the meso level of analysis. The participants in the TDP had different opinions about the relative advantage of EDI. Two of the adopters explicitly expressed that exchange of EDI messages were more costly than the ordinary paper based routines. The reason for doing this non-profitable operation was to gain social prestige (innovative image) and strategic benefits. The non-adopters had no expectations regarding economic gains from EDI adoption. To them the advantage of EDI adoption was to be found in strategic considerations. The adopters and non-adopters did recognize relative advantages of EDI. The advantages were considered to be strategic in nature rather than related to economic profitability and social prestige. Since the strategic needs had not yet been a real need for the non-adopters they had so far not found that the relative advantage from adoption was superior to the traditional paper based routines. Compatibility is the degree to which an innovation is considered to be consistent with existing values, past experiences, and the needs of the potential adopter. If the usage of EDI is compatible with the socio-cultural values, previous introduced ideas, and the company’s need for innovation then adoption is more likely. The determining factors in relation to EDI with respect to compatibility is whether or not EDI is viewed as being better than the existing manual or electronic systems, whether or not it is consistent with the needs of the potential adopter, and whether or not it is perceived as easy to use and understand. No clear patterns were found regarding organizational compatibility in the TDP. The advanced users had adopted EDI out of curiosity or due to strategic considerations. The adopters that were forced to use EDI and the non-adopters more or less regarded EDI as a strategic necessity. EDI was not really viewed as being consistent with the perceived needs of the organizations. For the adopters that were forced to adopt it was just there,
211
and for the non-adopters there had not yet been any external pressure demanding EDI adoption. Complexity is the degree to which an innovation is perceived as relatively difficult to understand and use. Complexity can be viewed from two angles: The managerial and the technical. The technical complexity is beyond the scope of this analysis. The managerial view of complexity refers to the depth of the organizations’ knowledge resources (Lai and Guynes, 1997). Lai and Guynes refer to complexity as the, “… employees’ range of knowledge, expertise, experience, and professionalism.” If the managerial view is applied then there are two types of complexity related to the understanding and the use of the innovation. One type of complexity is related to understanding how computer hardware and software should be used. Training of employees in the use of computer hardware would hardly be necessary in Danish businesses. The statistics from year 2000 reflecting IT usage in Danish companies show that 91 percent of all companies use IT for administrative tasks (Ministry of Information Technology and Research, 2001). It can however not be denied that adoption of EDI software will require training of employees. This may lead to a perception of the innovation as being relatively difficult to understand and use since adoption of EDI could eliminate tasks that have been part of the organizational routines for decades. Another aspect of the managerial view of complexity is related to the interorganizational relations, which are to be reconsidered when EDI is adopted. For example EDI contracts specifying terms for delivery and settlement are to be renegotiated with existing business partners and legal uncertainties related to the status of EDI messages may be viewed as an obstacle for understanding and using the innovation (Henriksen and Görsch, 1999). Complexity was not really a subject that was discussed in the TDP. No evidence was found in the minutes of the meetings that participants had expressed concerns related to the understanding of and the use of EDI. One explanation could be that at present most companies had already invested substantial amounts of money and effort in training employees in the use of computer hardware and software. The same attitude was predominant in the
212
ex post interviews of the TDP participants. This could lead to the conclusion that the complexity of EDI from a managerial point of view is limited. If complexity on the other hand is viewed from a technical point of view then the standardization issue dominated the matter of complexity. The standardization issue in relation to EDI is far from being solved – and probably never will be (Damsgaard and Truex, 2000). Trialability is the degree to which an innovation may be tested with some limitations. One of the examples Rogers presents to demonstrate trialability is the use of new types of seed grain among farmers. Farmers can test the innovation on a limited basis. They can for example sow part of a field with a new type of corn without making substantial efforts or substantial investments. When it comes to adoption of IOS trialability is more complex. A company has to make contractual arrangements with its business partner before EDI messages can be exchanged in a meaningful way. Next, the company has to make investments in EDI software and arrangements with transport providers before the first EDI message can be exchanged. The cost of legal advice in relation to contracting and the cost of consultant services can be considerable. Trialability on a limited basis in relation to EDI investments are therefore in practice non-existent. This however, does not exclude that companies that adopt EDI experiment with EDI on a limited basis in relation to the number of business partners with whom they exchange EDI messages. A more pessimistic interpretation of the term trialability therefore leads to the conclusion that trialability is not practical in relation to EDI since the investment is sunk cost if the project is abandoned. The intention of the TDP was that the developed EDI software should be provided at a marginal cost and that the primary means for transportation of the EDI messages should be the Internet, with which most Danish companies are connected (Ministry of Information Technology and Research, 2001). Through EDI standard contracts and a low-cost EDI software the costs of adoption should be reduced to an absolute minimum. Additionally, the project initiators expected that the involved companies through their work in the project made the necessary arrangements to
213
establish EDI connections. Even though what seem to be optimal conditions for EDI adoption were created, the involved companies were not ready to try the innovation. When trialability failed in the TDP it is difficult to imagine that trialability would be realistic in a “real world setting” where the investments and uncertainties are much greater than among the TDP participants.81 Observability is the degree to which the results of an innovation are visible to others. The adoption of EDI is as such not observable to other than those directly involved in the administrative processes affected by EDI, since no special devices are visible. Theoretically this should not be a problem in relation to EDI, which is used amongst established business partners that are expected to have the necessary knowledge of each other in relation to technological capability. However, a recent study of a company’s customers revealed that close to twenty-five percent of its key customers actually used EDI without the company being aware of it (Henriksen 2001). Viewing observability from a broader perspective the following considerations may be put forward. Denmark does not similar to US provide “EDI Yellow Pages”. Another directory of EAN (European Article Numbering) location numbers82, GEPIR83, is however accessible. This means that EDI users become visible for potential business relationships based on EDI. It is one thing that it is possible to find an EDI using company in a business directory. However, the symbolic value of being an adopter of an innovation which others can recognize right away is quite another thing. To conclude it is found that even though EDI adopters literally may be observable it has no practical relevance in relation to symbolic visibility.
81
This may be a superficial and simple interpretation of the situation. Political and economic relationships have without doubt, similar to previous research (Premkumar and Ramamurthy, 1995), also influenced the willingness to adopt EDI amongst the TDP participants. 82 Assigned EAN numbers are indicators of the number of companies that are able to send/ receive EDI(FACT) messages (Andersen et al., 2000). 83 http://gepir.ean.dk/ (Global EAN Party Information Repository) lists the assigned EAN numbers in all European companies plus companies in a number of other nonEuropean countries.
214
6.4.4 Summing up innovation and perceived attributes of innovation As shown in this presentation and evaluation of the term innovation there are multiple facets that have to be taken into consideration if a meaningful assessment is to be made in relation to IOS and EDI innovations. The first three facets that were included in the innovation typology (cf. Figure 6-2, page 209) were related to more objective and normative consequences of adoption of an innovation. Depending of the nature of the innovation the three dimensions: Source, type, and effect could lead to a more or less radical reorientation for the adopting unit. The other set of facets was related to more subjective ex ante considerations. These subjective and perceived qualities of a given innovation were addressed in the five perceived attributes outlined by Rogers. If EDI is viewed in relation to Rogers’ five attributes it becomes clear that it is necessary to add complementary elements in order to find an explanatory theory for adoption of information systems, which supports interorganizational business relations. This was illustrated by the inclusion of data from the TDP. A number of dimensions seemed to be absent in relation to Rogers’ five innovation attributes. These include political and economic factors that govern interorganizational relations. The five attributes identified by Rogers primarily viewed the innovation from an internal point of view where the potential adopter consider voluntarily adoption due to perceived needs and preferences. This attitude is however less pertinent when it comes to the adoption of an innovation, which is applied by businesses in a competitive environment where strategic alliances and politics play an important role.
6.5 The innovation-decision, communication channels, nature of social system, and extent of change agents’ promotion efforts The second variable, “Type of innovation-decision”, that Rogers suggests as determining the rate of adoption manifests when a decision-making unit engages in activities that lead to a decision to adopt or reject an innovation (Rogers, 1995). Strictly speaking, the adoption motivation is the seed of the innovation-decision. The innovation-decision determines the rate of 215
adoption and is related to a commitment to action (Mintzberg et al., 1976). The innovation-decision process is essentially an information seeking and information processing activity in which the individual is motivated to reduce uncertainty about the advantages and disadvantages of an innovation (Rogers, 1995). In order to reduce uncertainty, knowledge about the innovation is a prerequisite. The process will often require new mental activities later on or subproblems have to be solved in parallel cycles leading to new insights into phases that have already been concluded (Simon, 1977). The non-linearity of the adoption-decision (or any decision for that sake) process is well illustrated by the following quotation: “Interrupts” change the pace and direction of a decision process. “Scheduling delays” operate between the stages of decision making. “Feedback delays” are those in which decision makers are delayed in progressing to the next step while awaiting the results of the previous step. “Timing delays and speedups” are situations in which managers purposely change a decision-making schedule in order to take advantage of new circumstances. “Comprehension cycles” represents the process by which decision-makers cycle and recycle through phases in effort to gain a more accurate understanding of the situation. Finally “failure recycles” occur when an acceptable solution is not found or when available solutions are rejected.” (Tornatzky and Fleischer, 1990)
An adopter goes through three stages of knowledge about an innovation which leads to the decision-making process (Rogers, 1995). First, the adopter gains awareness-knowledge when (s)he is informed of the existence of the innovation. The, next step is to gather how-to knowledge. At this point the adopter gets some familiarity with the use of the innovation. Finally, (s)he obtains principles-knowledge, which comprises information about how to deal with the functioning principles underlying the innovation. Knowledge about an idea is often quite different from using the idea. Most individuals know about many innovations, which they have not adopted. The reason being that an individual may know about a new idea but is not regarding it as relevant to her or his situation and hence does not consider it to be potentially useful (ibid.).
216
The project initiators of the TDP apparently assumed that awarenessknowledge regarding EDI had been spread through the 1996-action plan and the general promotion of EDI in the business environment. The objective of the TDP was to move the potential adopters from awarenessknowledge to how-to knowledge and then further on to principlesknowledge. Through development of the EDI subset the TDP participants gained knowledge of the underlying principles, knowledge that must be considered beyond that of ordinary EDI users. The project thus succeeded in creating both how-to knowledge and principles-knowledge. This however, did not lead to further diffusion of EDI among the TDP participants even though the non-adopters said that they had obtained useful information about the innovation. The non-adopters in the TDP case said that no demand from business partners to use the innovation was their primary reason for being non-adopters. Decisions-making is complex and often organizations enact decisionmaking procedures to compensate for the bounded cognitive abilities of decision-makers (March and Simon, 1993). That is especially the case in relation to decisions labeled non-programmed decisions in decision theory. These types of decisions are novel, unstructured and may have far-reaching consequences. There will be no routine solutions since the problem has not been confronted earlier or because its precise nature and structure is elusive and complex. Finally, the decision can be so important that it deserves a custom-tailored treatment (Simon, 1977). Frambach (1993) pinpoints the complexity of non-programmed innovation-decisions: “The innovation adoption decision is the most complex one that an organization will be faced with, because no experience on the buying process of the particular product exists.” (p. 24)
Adoption-decisions are, as mentioned by Frambach in the above quotation, closely tied to a series of uncertainties related to the nature of the innovation. The reason being that the novelty details related to the innovation may be unknown to the adopter. This is particularly relevant in relation to organizational adoption. Organizational adoption may involve changes in organizational processes and further lead to changes in organizational structures and flow of activities (Tabak and Barr, 1999). 217
Mohr (1987) has presented the uncertainties related to organizational adoption of innovations by stating that: “It is difficult to avoid the conclusion that collective decisions about almost any aspect of innovation, but about adoption in particular, are likely to reflect much of the quality of innovation itself; that is, the decisions are likely to be non-linear, tumultuous, unpredictable, and only partially subject to rigorous planning.” (p. 238).
This quotation of Mohr could be used to conclude that a search for motivators for adoption of innovations is a sheer waste of time. None the less in the TDP certain patterns in relation to motivation for EDI adoption became visible. Adoption appeared to be dependent on the size of the organization, position in the supply-chain, and legal ownership of the organization. However, different degrees of power also seemed to play a central role (cf. Table 4-4, page 122) and issues related to strategic performance, efficiency and critical mass were also mentioned as important determinants for EDI adoption. Data from the case study suggests that some more or less predictable adoption motivators could be identified. The challenge is therefore to uncover and get an understanding of the motivation for adoption leading to the innovation-decision. Rogers claims that there exists a set of innovation-decisions on a continuum starting with optional decisions, through collective decisions, to authority decisions. 1) Optional innovation-decisions are characterized by a situation where there exists a choice whether or not to adopt an innovation. This type of decisions is made by an individual who is independent of the decisions of the other members of the system. The individual has complete control of and responsibility for the optional decision process. 2) Collective innovation-decisions. In the case of the collective innovation-decisions, the decisions to adopt or reject an innovation are made by consensus among the members of a system. Regulation is an example of collective innovation-decisions where all members of the system must conform to the systems’ decision. The individual influences the collective innovationdecision. 3) Authority innovation-decisions. In the case of the authority innovation-decisions, the decisions to adopt or reject an innovation are made by relatively few individuals in a system that possesses power, status,
218
or technical expertise. The adopting individual has no influence on the authority innovation-decision. In general the decision to adopt EDI in the organization is a nonprogrammed decision. The issue could earlier have been on the agenda, but the precise nature and structure of the decision related to EDI in the organization is elusive and complex and often requires a custom-tailored treatment due to the interorganizational nature of the technology. The innovation-decision is therefore likely to be an authority innovationdecision when it comes to adoption of EDI in a small and medium sized organization where only a few individuals possesses power, status, or technical expertise to decide the best course of action. The decision-process in the TDP was, as concluded in Section 4.4, driven by authority innovation-decisions. However, these authorities did not have the necessary authority with respect to the final adoption and diffusion decision in their respective organizations. The decision-making process in relation to adoption is complex (Tornatzky and Flesicher, 1990; Mohr, 1987). It is not a single decision, or even a simple set of decisions, but rather a highly contingent chain of decisions that iterate toward an outcome that is neither inevitable nor predictable. Tornatzky and Fleischer find it important to have a firm grasp of the concept adoption of an innovation in order to understand the innovationdecision process. As mentioned in Section 6.2, adoption is defined as the process when the organization moves from not having the innovation to having it. An authoritative commitment divides having from not having (Tornatzky and Fleischer, 1990). There are however, as just mentioned, a large number of parameters associated with the innovation-decision, which make the process complex and full of uncertainties. 6.5.1 Communication channels The third element (cf. Table 6-2, page 201), the communication process, is considered a major explanatory factor for diffusion (Kautz and Larsen, 2000). The communication about the innovation can broadly speaking be addressed through two channels: Mass media channels and interpersonal
219
channels (Rogers, 1995). By using the mass media channels the information about the innovation is spread to a large number of potential adopters via media such as TV, radio, newspapers, and the Internet. The mass media channels are ideal for communicating the innovation to those who are completely unaware of the existence of the innovation. Awarenessknowledge is usually created via this mass media channel. The interpersonal channel, which involves face-to-face exchange between two or more individuals, is preferable when the potential adopters already have some knowledge about the innovation. Rogers claims that the interpersonal channel is more important than the mass media channel with respect to the adoption-decision. He argues that the interpersonal communication channel is the most efficient, since diffusion is a social process where people depend on subjective evaluations conveyed to them from other individuals rather than depending on objective scientific studies. The individual adoption is to a large extent viewed as an imitation process, which takes place in social networks. In the TDP case the interpersonal communication channel played the dominant role. Though the trade and business associations had used the mass media channels for general EDI campaigns and thus made the participants in the TDP aware of EDI the intention was to build a strong interpersonal communication network in the two work groups. Even though the necessary structure was build in relation to establishment of physical settings for meetings and financial support to activities related to the project the communication did not lead to the expected adoption and diffusion of EDI. As mentioned in Chapter 3 the Danish companies were generally exposed to information about EDI. The companies involved in the TDP had thus been exposed to “double” mass communication channels84 along with interpersonal communication. Though the TDP participants were heavily exposed to both types of communication and thereby made knowledgeable about many aspects of the innovation, this was however not enough to make the participants adopt the innovation. 84
There were ongoing EDI campaigns at the general level including distribution of information booklets published by the Danish EDI Council. Parallel to these efforts the industry and trade associations in detail informed their members about the developments.
220
6.5.2 Nature of the social system The fourth variable, which according to Rogers determines the rate of adoption, is related to social systems (cf. Table 6-2, page 201). Social systems are defined as, “... a set of interrelated units that are engaged in joint problem-solving to accomplish a common goal.” (Rogers, 1995). The underlying assumption is that the members of the social system seek to solve a common problem in order to reach a mutual goal. There are, according to Rogers, structures in social systems. These structures can be formal or informal. The formal structure gives regularity and stability to human behavior. The bureaucratic organization of a government agency is an example of a formal structure. Informal structures on the other hand reflect who interacts with whom and under what circumstances. When viewing the TDP based on the theoretical input from Rogers in relation to the nature of the social system it becomes obvious, why adoption was less than successful. The first obstacle is the “common goal”. It can be debated whether the nine companies in the project had a common problem, which could be solved, within the scope of the TDP. That at least is true if the common problem was the very low level of diffusion of EDI in the steel and machinery industry. The means of solving that problem were different depending on the level of adoption. The companies had different needs in regard to information on EDI depending on their status as adopters or non-adopters. The adopters were facing problems in relation to implementation and diffusion, whereas the non-adopters were more concerned with adoption issues. If the common problem on the other hand was that the sector lacked an industry EDI subset then the issue of a common problem becomes meaningful. An examination of the objectives and success criteria for the TDP cf. Table 4-1 (page 95) and Table 4-2 (page 96) respectively reveal that the common problem was not clearly defined. It is however, the first mentioned problem, the low level of diffusion, that drives the project, and not the second mentioned problem, the lack of an industry EDI subset. A fact supporting this interpretation is found in the general attitude in the environment, the socio-political factors, described in Chapter 3. Here it was evident that adoption and diffusion of EDI was on the agenda. The next step, solving practicalities such as
221
developing industry subsets and general implementation were not explicitly considered. These problems were to a large extent considered as a responsibility of the individual businesses involved in adoption. Another issue that can explain the limited success of the TDP in relation to the nature of the social system is the “mutual goal”. The case study of the project discloses that the participants had different goals. A comparison of the two work groups’ interpretation of the mutual goal illustrates this discrepancy. One work group considered the objective of the project to be the achievement of stronger business relations within the work group participants, while the other work group considered the objective to be building a common infrastructure based on EDI that can benefit the entire steel and machinery industry. Further, it seems obvious that the different stakeholders had different objectives for being involved in the TDP. The larger companies, which had adopted EDI, stated that they joined the project for political reasons. They wanted to make sure that they were major participants and contributors in the development of the industry EDI subset. Additionally, they would not miss the opportunity to closely follow the developments in their industry sector in general. The smaller nonadopters on the other hand were more interested in getting the necessary knowledge to get started with EDI. Data from the case study leaves the impression that the social structure was informal rather than formal. Though there was created a formal structure around the project initiators and the steering committee (cf. Figure 4-1, page 97) regularity and stability of the formal structure appeared to be absent. At this point it should be mentioned that the influence of the steering committee on the TDP was negligible. According to the sources available the steering committee hardly functioned on a formal level. The remaining part of the formal structure, the project initiators, participated very little in the ongoing work in the two work groups. The informal structure on the other hand seemed to play an important role in the project. An evident example is the authors’ narration of a meeting (cf. Textbox 4-1, page 103). A group of participants set the agenda and influenced the work according to their preferences.
222
6.5.3 Extent of change agents’ promotion efforts The last variable “Extent of change agents’ promotion efforts” was discussed in Chapters 3 and 4. According to Rogers, “A change agent is an individual who influences clients’ innovation-decision in a direction deemed desirable by a change agency.” It can be difficult to separate the change agent from the change agency in the TDP case, especially because the change agents as professionals are representing the change agencies. It was found that the promotion efforts of EDI both in relation to the macro level (the 1996-action plan, cf. Chapter 3) and meso level (the activities of the Danish EDI Council and the business associations, cf. Chapter 4) were less successful in relation to the rate of adoption. Rogers mentions that the relationship between rate of adoption and change agents’ efforts may not be direct and linear. This issue was discussed in relation to the effect of softlaw exemplified by the 1996-action plan. It was found that, depending on the type of regulatory intervention (pedagogical, economic, or normative) the governmental regulation initiatives in general were not the most efficient change agents. At the meso level one of the promotion efforts was the TDP in the sense that the pilot project served as a tool for, “… involving members of different business sectors and sharing experiences and know-how with other members of the respective business sectors.” (cf. Table 4-1, page 95, initiative 3). The significance of professional associations has been discussed recently (Swan et al., 2000). Swan et al. highlight the importance of professional associations because they provide members with information about the latest developments in their area and provide a forum where members informally can meet and discuss the advantages and problems associated with particular technologies. Rogers argues that, “The greatest response to change agent effort occurs when opinion leaders adopt…” and a consequence of adoption among opinion leaders is that, “… the innovation will then continue to spread with little promotion effort by change agents, after a critical mass of adopters is reached.” It can be questioned whether the two business associations pursued an optimal strategy for adoption of EDI among their members in the steel and machinery industry, when they selected the participants for the TDP.
223
In the narration of one of the meetings (cf. Textbox 4-1, page 103) it became clear that the “aloof side of the table” were the opinion leaders in the project. These companies were, however, not ideal proponents for the EDI option created in the TDP. They declined to use the EDI software developed during the project (cf. Table 4-7, page 137) even though it could have extended the scope of EDI use both in the TDP and beyond the group of involved companies and hence help to establish the necessary critical mass in the business community. Additionally, it can be argued based on the VISD-test, that the opinion leaders in the TDP only had adopted EDI at a low level if the elements from the VIDS-test (cf. Table 4-6, page 126) are followed scrupulously. Though the opinion leaders in the TDP had adopted EDI their degree of use was at a low level. The limited utilization of the EDI investments influenced the attitude towards EDI in a less positive direction during discussions in the workgroup. This could weaken the argumentation in favor of EDI. This review of Rogers’ five variables determining the rate of adoption of technological innovations led to mixed results in relation to the adoption model’s ability to explain adoption of IOS and EDI. The five variables determining the rate of adoption did serve as valuable guidelines for the interpretation of the TDP data. In the following an alternative to Rogers’ adoption model is presented. An offset is taken in previous research on adoption of IOS in order to explore how other researchers have approached the problem of finding a suitable adoption theory for IOS innovations.
6.6 Adoption of IOS innovations A number of studies have looked at IOS adoption in small businesses (Iacovou et al., 1995; Premkumar and Ramamurthy, 1995) in relation to incentives (Thong, 1999) and inhibitors (Chau, 2001). Common for these studies is that they include variables related to both organizational and interorganizational issues. The use of interorganizational systems in business processes implicates cooperation and commitment from all involved members. The participants may have complex economic and business relationships, that result in a number of social, political, and 224
economic factors influencing adoption of IOS (Premkumar and Ramamurthy, 1995). These issues combined with the complexity of organizational adoption in general challenges the adoption of IOS: “Diffusion among organizations presents special challenges because, unlike individuals, they are complex human aggregates with various decision centers and are endowed with traditions, values, and procedures that impede or enhance the decision adoption process.” (Pennings, 1987)
Rogers acknowledges similar to Pennings (1987) that compared to the individual adoption process the adoption process in organizations is much more complex. This complexity is partly related to the number of individuals with different interests involved in the decision process. Group decisions involve interplay between individual roles and group dynamics. Especially the individual roles are complex because they are influenced by personal characteristics as well as by the organizational position the individual holds (Tornatzky and Fleischer, 1990). Besides, the adoption usually requires organizational adaptation, which might lead to change in organizational routines and culture (Tabak and Barr, 1999). IOS such as EDI involve substantial internal effort related to change in administrative procedures. This may lead to changes in job-tasks and a reorganization of the work force. As a consequence many organizational factors can be expected to influence the decision to adopt these strategic systems (Premkumar and Ramamurthy, 1995). The considerations related to organizational adoption therefore go beyond the mere acceptance of the innovation. 6.6.1 Recent studies focusing on IOS adoption and non-adoption A number of recent studies have examined the organizational adoption and non-adoption of IOS. Hart and Saunders (1997) studied the significance of power and trust in relation to EDI adoption. Chau (2001) explored issued related to non-adoption of EDI. A number of researchers have studied a mix of adopters and non-adopters of EDI (Lai and Guynes, 1997; Crum et al., 1996; Iacovou et al., 1995). Hart and Saunders (1997) concluded based on an in-depth case study of a firm and its EDI partners, that both power and trust are important 225
determinants for adoption of EDI. It was found that EDI adoption reflects existing power arrangements among organizations. The aspect of trust was based on the assumption that adoption of EDI is linked to uncertainties related to increases in interdependency and vulnerability. A number of studies have included both adopters and non-adopters in their analysis (Lai and Guynes, 1997; Crum et al., 1996; Iacovou et al., 1995). Lai and Guynes (1997) found that there was a considerable variability between adopters and non-adopters in relation to a number of organizational influences such as size, degree of openness, and degree of uncommitted slack resources. In recognition of the deficiencies of the explanatory power of traditional diffusion of innovations theory in relation to organizational adoption Lai and Guynes concluded that a combination of traditional diffusion theory and critical mass theory (Markus, 1987) would be ideal for understanding the adoption behavior of IOS exemplified by EDI. Most studies have focused on adoption and only a few studies have focused on non-adoption (Chau, 2001). Chau focused on the inhibitors of EDI adoption. The non-adoption was according to the analysis of a large sample of non-adopters in Hong Kong caused by limited technological capabilities and perceived unimportance of the innovation. It was also found that since there had been no pressure to adopt it had not been imperative to consider adoption. Chau therefore concluded, “The ability to adopt is more important than the benefits of the adoption in the eyes of small businesses when considering adoption of EDI.” The same conclusion was made in relation to a previous study on open systems adoption in Hong Kong (Chau and Tam, 1997). In the Chau and Tam study of adoption of an IOS the Tornatzky and Fleischer adoption model was operationalized. It was found that characteristics of the technological innovation were of greater importance than the organizational and environmental context. This conclusion might however, as pointed out by the authors, depend on the geographical context rather than on the technology itself. The argument is that companies in the third world might perceive technology differently than companies in the industrial countries in the sense that companies in third-world countries are less familiar with technology in general. Also
226
strategic issues related to technology adoption play a more central role in industrial countries than in third-world countries.85 Table 6-3. MIS studies of adoption motivators related to IS Reference Kurnia and Johnston, 2000
Study object IOS/ e-commerce
Thong, 1999
IS
Chau and Tam, 1997
Open systems
Lai and Guynes, 1997
ISDN
Premkumar and Ramamurthy, 1995 Iacovou et al., 1995
EDI
Sabherwal and King, 1995
IS
Grover and Goslar, 1993
Telecommunication technologies
Grover, 1993
IOS
EDI
Explanatory variables - External factors - Nature of technology - Capability of organization - Supply-chain/ industry structure - CEO characteristics - Perception of IS attributes - Organizational characteristics - Environmental characteristics - External environment context - Technological context - Organizational context - Contextual effects - Structural effects - Strategy effects - Interorganizational variables - Organizational variables - Perceived benefits - Organizational readiness - External pressure - External environment - The organizational context - The IS function - Environmental factors - Structural (organizational) factors - IS factors - IOS factors - Environmental factors - Policy factors - Organizational factors - Support factors
The above-mentioned studies all point to explanatory variables, adoption motivators, which go beyond those included in Rogers’ adoption and diffusion model. Though Rogers’ emphasizes the social processes related 85
Rogers (1995) discusses the cultural aspects of technology adoption. He argues that adoption and diffusion of products or ideas developed in the industrialized parts of the world often meet with what seems to be obscure obstacles when introduced in other cultures. Recent research has strongly stressed that inherently diffusion of innovations is an exponent of cultural and economic imperialism (McMaster, 2001).
227
to the nature of the social system and extent of change agents’ promotion efforts (cf. Table 6-2, page 201), these elements are more concerned with the channels of communication and the characteristics of the innovation. They are not directly related to the competitive business environmental context organizations operate in. In order to explore the determining variables for IOS adoption an extensive literature review was made. In order to broaden the scope for possible adoption motivators, studies related to IS were also included in this review. This review resulted in nine studies dealing with motivators for adoption. What is common for these nine studies focusing on adoption motivators is that most of them include environmental factors and organizational factors as explanatory variables for adoption.86 A number of studies have included technological attributes as well. Only a few studies (Premkumar and Ramamurthy, 1995; Grover, 1993) include particular IOS characteristics in their search for explanatory variables for adoption even though most of the studies are related to IOS adoption. The grouping of explanatory variables into three clusters, the environmental, the organizational, and the technological is not an unfamiliar model in adoption and diffusion literature. Tornatzky and Fleischer (1990) suggested that there are three explanatory variables that influence the process by which innovations are adopted in organizations. 1) The external environmental context, 2) The technological context, and 3) The organizational context. Tornatzky and Fleischer consider the three explanatory contexts to be interconnected cf. Figure 6-3. Tornatzky and Fleischer suggest that the following factors are inherent in the three types of explanatory contexts. The organizational context reflects items such as: Company size, formalization, quality of human resources, and informal linkages between employees. The external environmental context includes: Market conditions such as competitive market forces, market uncertainty, and
86
Though all the studies include other explanatory factors than Rogers’ adoption and diffusion theory, all studies refer to editions of Rogers’ book “Diffusion of Innovations” (1995; 1983).
228
government regulation. The technological context refers to availability and suitability of technologies of relevance for the organization. Figure 6-3. The Tornatzky and Fleischer (1990) model for adoption
Environmental context
Organizational context
Technological innovation decision making
Technological context
In order to introduce the three contexts a short presentation of the three contexts will be given. The organizational context comprises attributes related to the organization. Tornatzky and Fleischer are of the opinion that the organization provides a rich source of structures and processes that constrain or facilitate the adoption of innovations. These structures and processes can be formal or informal. The informal structures and processes include the ways an organization divides its labor force into distinct tasks and achieves coordination among them. The informal structures and processes represent naturally occurring behavioral patterns and roles that perform many of the same co-ordinating functions as the formal structures. Adoption attributes related to the organizational context include: Informal linkage and communication, quality of human resources, top management leadership behaviors, organizational size, and the amount of available internal slack resources.
229
The environmental context is the arena in which a firm conducts its business. Included in this context are the organization’s competitors, its industry, its access to resources supplied by others, and its dealings with government. All these elements can influence the degree to which an organization views the need for adoption of innovations. Tornatzky and Fleischer suggest that two aspects of the external environment are key determinants of innovative activity: The competitive characteristics of its industry and the existence of a relevant technology support infrastructure. Adoption attributes related to the environmental context concerning the competitive characteristics include: Intensity of competition, customersupplier relations, market uncertainty or volatility, and industry life cycle. The attributes listed in relation to technology support infrastructures include: Labor costs, the skills of available labor force, and access to suppliers of technology related services. It is worth noticing that the environment is not solely considered to be external forces beyond the control of the firm. Some firms have the ability to shape their environments through their influence on actions of customers and suppliers and through dictates due to their competitive superiority. The technological context comprises both the internal and external technologies relevant to the firm. This includes current practices and equipment internal to the firm as well as the pool of available technologies external to the firm. Decisions to adopt a technology depend on what is available, as well as how the available technology fits the firm’s current technology. Tornatzky and Fleischer call attention to the fact that not all innovations are relevant to all industries. Contrary to the organizational context and the environmental context Tornatzky and Fleischer do not mention any particular attributes of the technological context, which facilitate or hamper the adoption of technological innovations. The presentation of the technological context does however reveal, that it has close connections with traditional innovation attributes. One feature, which for example is stressed, is the degree of radicality of the innovation.
230
6.6.2 Tornatzky and Fleischer or Rogers – or a fusion of the two models? Though the adoption and diffusion theory as presented by Rogers includes explanatory variables related to the three contexts described by Tornatzky and Fleischer (1990). Rogers’ theory does not encompass environmental characteristics such as market forces (Kwon and Zmud, 1987; Tornatzky and Fleischer, 1990) and the uncertainty and vulnerability related to the increased transparency, which potentially is present in relation to adoption of IOS (Hart and Saunders, 1998). The variables determining the rate of adoption (cf. Table 6-2, page 201) presented by Rogers included “Nature of the social system” and “Extent of change agents’ promotion efforts”. These two variables are to a certain extent related to the environmental context. They are however, more closely related to the socio-political processes in the external environment. Economic and political forces related to business processes such as competitive strategies are not explicitly included in the model. Also, Rogers does not consider market uncertainties in the two variables related to the environmental context. The technological context is included in Rogers’ framework by the variables related to “The perceived attributes of innovations”. These attributes were cf. Section 6.4.3 found to cover EDI in a generic, but unsatisfactory manner. This could be explained by one of the salient critiques of Rogers’ framework when this framework is used on adoption and diffusion of complex technological innovations such as IOS. Rogers’ framework focuses on adoption and diffusion of general practices and mass-produced items rather than complex interorganizational information technologies (Chau and Tam, 1997). The third context included in the Tornatzky and Fleischer model for adoption is the organizational context. Rogers does not specifically target his model for organizational adoption. A liberal interpretation of Rogers’ framework could, however, justify that the variable “Nature of the social system” also could be said to encompass organizational cultures and norms. “Type of innovation-decision” can straight away be related to the organizational context.
231
However, instead of attempting to fuse the two different models into one model, the models should be considered to complement each other. Rogers’ framework is concerned with the social processes nourishing the motivation for adoption, and Tornatzky and Fleischers’ model for adoption broadens the scope by including other possible motivators for adoption. In the following three sections a conceptualization of the three contexts will be presented. These conceptualizations are based on the nine studies resulting from the literature review focusing on adoption of IOS. The purpose of the description of the variables determining adoption from previous research is to present and analyze how the three contexts have been considered in previous research.
6.7 Operationalization of the three contexts from previous IS research 6.7.1 The organizational context The organizational context is related to the characteristics of an organization (Tornatzky and Fleischer, 1990). The organizational characteristics influencing the motivation for adoption of IOS include the traditional organizational factors such as: Centralization, formalization, and size (Tornatzky and Fleischer, 1990; Grover and Goslar, 1993; Grover, 1993; Chau and Tam, 1997; Lai and Guynes, 1997; Sabherwal and King, 1995). Other organizational factors that motivate adoption related to IS innovations have been added in the nine studies listed in Table 6-3 (page 227). These factors are divided into two broad categories: 1) Operational and procedural needs. 2) Human resources. Operational and procedural needs comprise: Task complexity (Chau and Tam, 1997), satisfaction with existing systems (Chau and Tam, 1997), IS planning behaviors (Grover, 1993; Kurnia and Johnston, 2000), IS infrastructure (Grover, 1993; Kurnia and Johnston, 2000; Premkumar and Ramamurthy, 1995), technological resources in the firm (Iacovou et al., 1995), information intensity (Thong, 1999), slack resources (Lai and Guynes, 1997), and internal need (Premkumar and Ramamurthy, 1995).
232
Human resources include: Adequate education (Kurnia and Johnston, 2000), employees’ IS knowledge (Thong, 1999), employees’ positive attitudes toward change (Lai and Guynes, 1997), top management support (Kurnia and Johnston, 2000, Grover, 1993; Premkumar and Ramamurthy, 1995), innovation champions (Premkumar and Ramamurthy, 1995), and organizational compatibility (Premkumar and Ramamurthy, 1995). The operational and procedural needs are concerned with existing systems, needs for improvements of systems, and the resources available for further adoption of IS innovations. In relation to task complexity it is relevant to consider the type of task and the type of innovation. If the task complexity is high this will assist an organization to adopt the innovation (Chau and Tam, 1997). When dealing with the innovation typology (cf. Figure 6-2, page 209) three dimensions: Source, type, and effect were identified as being useful for classification of a technological innovation. Especially, type is relevant in this context. For EDI it was found that it as a minimum should be classified as a production-process oriented innovation. It is however hard to ignore that EDI is related to the organizational-structure innovations and the policy innovations characteristics. To keep things simple the organizational-structure and policy aspects are omitted in this context. When relating EDI to the production-process oriented innovation, those tasks, which EDI can support, should be considered. EDI is an administrative system, which in its most simple form can support information flows to and from the accounting department in the organization. After the introduction of ERP-systems in most businesses87 data related to business information such as balance sheets and inventory control is already digitized. This leads to an improved IS infrastructure, which has been found to be one of the predictors for adoption within the organizational context (Grover, 1993). The digitalization has however not led to lower task complexity. This is the case mainly because increased and improved information leads to higher task complexity for the involved
87
Two third of the EDI non-adopters included in the survey (which is reported in Chapter 8) of the steel and machinery industry indicated that the company already had an ERP-system.
233
employees. A complex IT infrastructure does according to Chau and Tam (1997) motivate organizations to further adoption of innovations. The category of human resources related to motivators for adoption of IOS are concerned with the willingness to change and the technological capabilities of the employees. Small companies are to a large extent viewed as laggards in relation to adoption of technologies for administrative tasks. This can be due to the fact that small businesses are often lacking specialized IS knowledge and technical skills (Thong, 1999; Iacovou et al., 1995). The factors related to the employees’ technical capabilities become an important parameter when assessing the motivation for EDI adoption in small businesses, since the capabilities are crucial for successful adoption and implementation. It takes time and expertise to incorporate complex technologies in organizations, because adoption of a complex technology is not a single event it is rather a process of accumulation of knowledge (Chau and Tam, 1997). That is especially true regarding EDI where the innovation is: 1) Fragile and not always operates as expected, 2) Is difficult to try in a meaningful way, and 3) Is “unpacked” in the sense that adopters cannot treat the technology as a black box, but must acquire broad tacit knowledge and procedural know-how to use the technological innovation effectively (Attewell, 1992). 6.7.2 The environmental context The environmental context is the arena in which the organization conducts its business (Tornatzky and Fleischer, 1990). Tornatzky and Fleischer included external factors such as relationships with the industry, competitors, regulations, and the government. External factors such as competitors and regulation are external to an organization and present constraints and opportunities for organizations. The environmental context in relation to regulations and government was presented in Chapter 3. It was demonstrated that the regulatory environment concerning the IT policy in Denmark was driven by a mix of public and private interests and that the driving forces were a combination of market forces and regulative recommendations. None of the reviewed articles has included regulatory factors as parameters for IOS adoption.
234
Similar to Tornatzky and Fleischer, Kurnia and Johnston (2000) call attention to the fact that companies are neither total victims of their environment nor in total control of their environment, rather they operate in the environment through interactions with other organizations. It is therefore better to refer to the interorganizational environment rather than solely referring to environment in relation to motivators for IOS adoption. As a consequence the following list of factors related to the environmental context used in previous studies include both those explicitly related to environmental factors and those interorganizational factors that are related to interactions amongst organizations. Similar to organizational factors the environmental factors can be divided into two categories: 1) Market driven factors. 2) Interorganizational relations. Market driven factors comprises: Market uncertainty (Grover and Goslar, 1993; Chau and Tam, 1997), unpredictable demand (Kurnia and Johnston, 2000), competition (Kurnia and Johnston, 2000; Thong, 1999; Grover, 1993), and maturity of an industry (Grover, 1993). Interorganizational relations comprises: Pressure from trading partners (Premkumar and Ramamurthy, 1995; Kurnia and Johnston, 2000; Iacovou et al., 1995; Grover, 1993), trust between trading partners, transaction climate (Premkumar and Ramamurthy, 1995; Kurnia and Johnston, 2000), dependence (Premkumar and Ramamurthy, 1995), and vertical integration (Grover, 1993). Market driven factors are related to environmental contingencies, and these factors are to a large extent beyond the control of the companies operating in the market. Under conditions of relatively undifferentiated and stable environments, organizations are able to handle information-processing requirements without sophisticated information technology (Grover and Goslar, 1993). On the other hand when an organization faces a market situation with uncertainty and with intense competition, then such a situation tend to stimulate rapid adoption and diffusion of an innovation (Thong, 1999; Grover, 1993; Goslar and Grover, 1993; Chau and Tam, 1997). In relation to IOS the competitive market forces are countered
235
through strategic alliances between trading partners for the purpose of maintaining a competitive edge in an increasingly volatile market environment (Chau and Tam, 1997). Another source of market uncertainty is globalization of businesses, which is related to demand uncertainty and logistic challenges (ibid.). No evidence was found that the Danish steel and machinery industry compared to other industries was facing extraordinary market uncertainties. Though the globalization rate was estimated to be high in the Danish steel and metal industry (cf. Section 3.2) none of the companies in the TDP expressed particular concerns related to market uncertainty and intense market competition. Interorganizational relations refer to relations with a firm’s existing business partners. Interorganizational relations such as trading partner considerations are important in relation to EDI. Hart and Saunders (1997) found in their study of EDI adoption between organizations that several power and trust factors had impact on electronic relationships between organizations. In the reviewed articles related to factors influencing motivation for adoption these two aspects were also present. Interorganizational relations were related to external pressure. Pressure can be related to peer pressure (Premkumar and Ramamurthy, 1995), competitive pressure, and imposition of trading partners (Iacovou et al., 1995). The external pressures to adopt IOS comprise a continuum of means ranging from recommendations to threats of business loss (ibid.). Peer pressure is pressure from industry associations, which forces firms to adopt e.g. EDI. Premkumar and Ramamurthy (1995) illustrate peer pressure by an example, which in many ways resembles the intentions of the TDP. They describe how industry associations in the US grocery and automotive industry have taken the lead in developing EDI standards thereby creating an environment with significant incentives and peer pressure for firms to use EDI as the standard mode for transaction communication. The Danish business associations have, however, so far been less than successful in their promotion of an EDI subset for the steel and machinery industry compared to the US associations. Competitive pressure refers to the level of the firm’s technological capabilities and the capabilities of its competitor (Iacovou et al., 1995). In the case of competitive pressure organizations
236
may adopt EDI despite any real internal need for EDI (Premkumar and Ramamurthy, 1995). Imposition of trading partners refers to a situation where customers require their suppliers to use EDI in their transactions. Another aspect related to the interorganizational relations is concerned with the trust and transactions climate. It might be possible to force business partners to establish an EDI connection. Trust is crucial for successful adoption. “A high level of cooperation and trust is required among participating firms because automation of the inter-firm purchase and sales transactions removes most of the manual control systems and paper trail that currently exists to ensure the accuracy and integrity of various transactions.” (Premkumar and Ramamurthy, 1995). Another issue is whether or not the business partners are capable of gaining benefits from the innovation. 6.7.3 The technological context The technological context comprises all the technologies available to the organization. The main focus of the technological context is how technology characteristics influence adoption (Tornatzky and Fleischer, 1990). Most of the nine studies have been strongly inspired by Rogers’ innovation attributes (relative advantage, compatibility, complexity, trialability, and observability) when operationalising the technological context. Only three (Chau and Tam, 1997; Grover and Goslar, 1993; Iacovou et al., 1995) of the nine reviewed studies did not include any of the innovation attributes in their operationalization of technological adoption motivators. Five additional factors were explored in relation to technological factors influencing motivation for adoption. These five factors can be classified into two categories: 1) Cost-benefit considerations. 2) Technological sophistication. The factors related to cost-benefit considerations were: Perceived benefits from technology use (Iacovou et al., 1995; Chau and Tam, 1997), perceived obstacles and risks (Chau and Tam, 1997; Kurnia and Johnston, 2000), and degree of switching costs (Kurnia and Johnston, 2000). The two factors related to technological sophistication were: IS maturity (Grover and
237
Goslar, 1993; Sabherwal and King, 1995) and interface standards (Chau and Tam, 1997). The technological context can be related to the cost-benefit trade-off of adopting a particular innovation. The perceived benefits of the innovation in relation to an organization’s specific setting play an important role in relation to the technological context (Chau and Tam, 1997). The perceived benefits are related to direct benefits such as operational savings related to internal efficiency and indirect benefits which refer to the impact of IOS on business processes and relationships (Iacovou et al., 1995). The technological sophistication factor comprises perceived importance of compliance to standards, interoperability, and interconnectivity (Chau and Tam, 1997). To some organizations proprietary systems are optimal while others aim at increasing compatibility and flexibility of the IT infrastructure and a (global) standard for example EDIFACT is preferred. 6.7.4 Explanatory power of the three contexts The nine studies revealed similarities in their operationalization of factors influencing motivation for adoption (cf. Table 6-3, page 227). An obvious conclusion is that the studies have found that one or the other of the three contexts provided the best explanatory power in relation to motivation for adoption. Six of the nine studies explicitly included the three Tornatzky and Fleischer contexts in their operationalization of motivators for adoption. A brief summary of the findings from the six studies is presented below in order to examine the explanatory power of the three contexts outlined by Tornatzky and Fleischer (1990). Grover (1993) found based on discriminant analysis of 225 responses from senior IS executives in US firms that environmental factors were the weakest predictors for adoption. The factors related to the organizational context, top management support and championship had the strongest influence on motivation for adoption of IOS.
238
In the Grover and Goslar (1993) study, which was based on multiple regression analysis of 165 responses from directors of IS departments, it was found that environmental factors (environmental uncertainty) had strong explanatory power (p-value < 0.001) in relation to motivators leading to adoption of telecommunications technologies. Organizational factors and IS factors on the other hand were not found to influence initiation and adoption of telecommunications technologies. Sabherwal and King (1995) used 85 responses from senior IS executives in their cluster analysis of IS adoption patterns. Base on their analysis Sabherwal and King concluded that none of the three contexts should be considered universally applicable. They found that explanatory power is dependent on the specific circumstances. The Chau and Tam (1997) study, which investigated adoption of open systems, used 89 responses from senior executives in corporate IT functions in Hong Kong. Chau and Tam used logistic regression analysis to test their research hypotheses related to the three contexts. No relationship between market uncertainty and adoption was found so there was no explanatory power of the environmental context. The study found no relationship between the organizational context and the factors comprising the context. The explanatory power of perceived benefits of adopting an open system were found to be insignificant, perceived barriers on the other hand were found to be significant. This indicates that the technological context holds the best explanatory power in this particular study. Thong (1999) explored the motivation for adoption of IS. The study, which concerned small businesses in Singapore, included 294 responses from CEOs. Based on discriminant analysis it was found that the environmental context was not significantly related to the decision to adopt IS. The technological context (relative advantage, compatibility, and complexity of IS) was significantly associated with adoption. The Kurnia and Johnston (2000) study, which was a case study of five leading Australian manufacturers and retailers, concluded that the factor
239
approach used for studying adoption of technologies in an organizational context was inadequate. It was argued to be inadequate due to its focus on a single organization, a single epoch, and non-inclusion of the interorganizational context. In their case studies they found that the trajectory of each organization adopting e-commerce is unique. Therefore, the factors needed to explain adoption of IOS innovations are not as straightforward as imagined in the traditional adoption and diffusion literature. Table 6-4. Explanatory power of the three contexts Organizational context Grover, 1993
Environmental context Grover and Goslar, 1993
Technological context Chau and Tam, 1997 Thong, 1999
Adoption depends on specific circumstances Sabherwal and King, 1995 Kurnia and Johnston, 2000
The results of the six studies, which included the three contexts suggested by Tornatzky and Fleischer (1990), appear to be ambiguous. The explanatory factors related to adoption motivation might depend on the technological and cultural context. The two Asian studies (Thong, 1999; Chau and Tam, 1997) found that the technological context yielded the best explanatory power. The two US studies on the other hand were ambiguous. They concluded that the organizational context (Grover, 1993) or the environmental context (Grover and Goslar, 1993) was the determining factor for adoption. The observed differences could be a result of cultural, legal, and economic differences or they might be related to the conclusions found by Kurnia and Johnston (2000). However, instead of solely accepting the process approach as suggested by Kurnia and Johnston (2000), the factor approach is still used in the search for explanatory factors for adoption motivators in a Danish context. Keeping the caveat of Kurnia and Johnston in mind along with the knowledge that organizations that have a high degree of IT knowledge integrate technology, people, and organizational factors in order to achieve business goals (Crook and Kumar, 1998) the factor approach will be pursued in the quantitative search for motivators for adoption in the Danish steel and machinery industry.
240
6.8 Summing up Chapter 6 In this chapter the central terms for the dissertation were defined and discussed. It was argued that the best understanding and definition of adoption/ non-adoption was related to having or not having an innovation as suggested by Tornatzky and Fleischer. Adoption was viewed as the (organizational) acquisition of an innovation opposed to diffusion, which was the spread of the innovation throughout a (business) community. The term innovation was explored in relation to three dimensions: Scope, type, and effect. The objective of this exercise was to be able to classify IOS and EDI according to the three dimensions thereby getting a better understanding of IOS and EDI in relation to in vast and complex term innovation. Recalling Figure 1-3 (page 16) it was stated that the theme of the present chapter was a presentation of the theoretical framework of the study and that the purpose of the chapter was to provide a framework for assessing data. Two models for adoption of innovations were presented for the purpose of exploring the motivation for EDI adoption. As illustrated in Figure 1-2 (page 10) the focal area for the dissertation is to find motivators for EDI adoption. The means to that end was to explore general adoption determinants and specific organizational adoption contexts. Rogers’ adoption and diffusion model focuses on general adoption determinants whereas Tornatzky and Fleischer’ model focuses on the specific organizational adoption contexts. Throughout the presentation of Rogers’ model data from the TDP was used to exemplify the model. In order to explore how similar studies had pursued an understanding of adoption of IOS a number of studies were reviewed. It was found that even though most of the studies were inspired by Rogers’ model for adoption a research strategy resembling the Tornatzky and Fleischer model for adoption was used in most of these studies. It was found that Rogers’ model which comprises five variables determining the rate of adoption of innovations was a useful tool for understanding the qualitative data from the TDP. One reason being that 241
Rogers’ framework is concerned with the social processes nourishing the motivation for adoption. From the assessment of each of the five variables a number of explanations for the limited success of the TDP were identified. Issues related to type of innovation-decision, nature of the social system, and the extent of change agents’ promotion efforts illustrated that an understanding of the limited success of the TDP most likely were to be found in the participants’ relatively limited influence on the decisionmaking process in their respective organizations. The nature of the social system was also interpreted to be a useful explanation for the limited success of the TDP. This assessment was primarily based on the observation that there were no common problem and no mutual goal, which could be addressed within the scope of the TDP. Finally, the extent of change agents’ promotion efforts appeared to have been unsuccessful with respect to the target group. The explanatory power of the perceived attributes of innovations was less unambiguous. The Tornatzky and Fleischer model was considered as a means for broadening the scope by inclusion of other explanatory factors motivating adoption of IOS. From the review of IS studies focusing on adoption it was found that the Tornatzky and Fleischer model for organizational adoption of technological innovations was often used. The Tornatzky and Fleischer model is therefore included and applied in the design of the survey instrument used for the quantitative assessment of the Danish steel and machinery industry. With respect to research question 1b, “Which models are used to explain the motivation for IOS adoption at present?” (cf. Figure 1-1, page 6) it is now possible to provide an answer. From the review of IOS adoption it was found that the operationalized items associated with IOS adoption were directly or indirectly inspired by Tornatzky and Fleischers’ model for organizational adoption of technological innovations. Most studies did however rest on assumptions from Rogers’ model for adoption and diffusion of innovations. Most of the reviewed studies did for example conceptualize technological attributes in relation to the innovation attributes defined by Rogers. The answer to research question 1b, “Which
242
models are used to explain the motivation for IOS adoption at present?” is therefore, that the models used to explain motivation for IOS adoption at present rest on the ideas from the frameworks provided by Tornatzky and Fleischer (1990) and Rogers (1995).
243
s
244
7 Operationalization of variables motivating adoption 7.1 Introduction The findings from the case study (cf. Chapter 4) were presented in a report and communicated to the two business associations and the Danish EDI Council. The three representatives from these associations did not express any surprise about the results. They had throughout the TDP been aware of the obstacles that the participants had faced in relation to collaboration and adoption of EDI with respect to the TDP. The representatives were however, interested in getting a better understanding of what influences the innovation-decision in a particular business sector, where awareness of the innovation had been created, where access to a low-cost EDI software had been provided, and where an EDI subset had been developed for this particular industry. From discussions with representatives in the business associations, from experiences gained from the TDP case, and from previous research fifteen opinion items were defined. These fifteen items were expected to capture a broad range of issues explaining motivators and hindrances for adoption. As argued in Chapter 2 practice-driven research does not mean that theoretical considerations are excluded from the research. It is actually the researchers responsibility to add suitable theory to the agenda set by practitioners (Zmud, 1998). The Tornatzky and Fleischer (1990) model for adoption of technological innovations in organizations, which was presented and discussed in Chapter 6, has been found to be a suitable overall guiding framework for classification of these fifteen items. Tornatzky and Fleischer suggested that three contexts influence the decision to adopt (cf. Figure 6-3, page 227): 1) Organizational context, i.e. the characteristics of an organization, 2) Environmental context, i.e. the
245
arena in which an organization conducts its business, and 3) Technological context, i.e. the technologies available to an organization. These three contexts were, similar to a number of other IS studies (cf. Table 6-3, page 227) operationalized in a survey instrument. The purpose of this operationalization was to be able to statistically analyze the different types of motivators for adoption within the three contexts.
7.2 Nature of the opinion data items The operationalized adoption-decision variables in relation to EDI are mainly related to secondary innovation attributes (Downs and Mohr, 1976). Downs and Mohr distinguished between primary and secondary innovation attributes. Primary attributes are viewed as inherent in the innovation and invariant across settings and organizations e.g. size and cost, which can be measured fairly objectively, whereas secondary attributes were defined as perceptually based on subjective characteristics for example complexity and relative advantage. The perception of secondary attributes is assumed to be influenced by characteristics of both the particular setting as well as actors involved in adoption of a particular innovation. It is therefore recognized that the measures applied are subjective in the sense that they are perceived and interpreted in the mind of the responder. It is however, as stated by Tornatzky and Klein (1982), difficult to maintain a separation between primary and secondary attributes since a primary attribute, which is assumed to be objective, in most cases is relative to the potential adopter. Cost for example is relative to the financial resources of the adopting organization. The cost might appear low to one organization and exorbitant to another. Anyway, an attempt is made to separate the primary and secondary attributes. The survey focuses solely on secondary attributes. In the context of the present study secondary innovation attributes are referred to as opinion data, which is found to be a more expressive label. This chapter is concerned with the operationalization of opinion data. Similar to Moore and Benbasat (1991) the immediate focus is to develop scales for measuring the opinion data related to EDI adoption. It is however the intention that the scales developed should be as general as possible, 246
thereby being applicable to a wide range of IOS. Consequently, items solely applicable to EDI adoption were not included in the survey instrument. This chosen strategy has both advantages and drawbacks. One of the drawbacks is that EDI compared to other IOS innovations has been considered to be constrained by for example rigid standards (Damsgaard and Truex, 2000). To make a thorough analysis of motivators and considerations related to EDI adoption it could therefore be argued that issues related to standards should be included. Another issue, which is specifically related to EDI compared to other IOS, is the dependency on critical mass in order to gain substantial organizational savings (See for example the VIDS-test in Table 4-5, page 126). It is however the author’ conviction that the advantages of having a more general IOS survey instrument are greater than the drawbacks. One of the advantages of having a generic survey instrument, which can be used for examination of a wide variety of IOS, is the prospect of being able to repeat the survey in relation to other technological innovations. Another argument proposed by Moore and Benbasat (1991) which makes the strategy of using general scales attractive is, that the study of secondary or perceived attributes of innovations is helpful for formulating a more general theory. In Sections 7.3, 7.4, and 7.5 the operationalization of opinion data items related to the organizational context, the environmental context, and the technological context is presented. The propositions in each of the three contexts are labeled with a letter, O, E, or T indicating whether the proposition is related to the Organizational context, the Environmental context, or the Technological context. The presentation of the underlying assumptions of the opinion data rests on three sources of information. First of all, the insights obtained from the interaction with the representatives from the business associations and the Danish EDI Council. The lessons learned from the interaction with the policy makers in the associations in most cases reflect the official policy statements and the 1996-action plan which were presented in Section 3.4. References are therefore often made to these official documents along with more postulating assertions expressed by the policy makers. Secondly, the
247
experiences from the TDP are used as basis for the various propositions. Finally, the theoretical insights gained from the review of the top-five MIS research themes (cf. Section 5.4.4) and in particular studies comprising motivators for IOS adoption (cf. Section 6.6.1) are included. The major distinction between the operationalization of the organizational context and the environmental context is whether the opinion data item is within or beyond the control of the organization. Control should in this context be understood as the organizations’ (perceived) capability of influencing the outcome caused by a potential adoption. That is the case for propositions O2, O3, and O4, which are related to improvement of work environment or the organizations’ capability of calculating the potential needs, benefits, or readiness, propositions O1, O5, and O6. The opinion data items related to the environmental context are on the other hand enabled/ constrained by interaction with business partners, propositions E1 and E2, or directly influenced by relations to business partners/ different degrees of power exerted towards the organization, propositions E3, E4, and E5. The opinion data items related to the environmental context therefore reflect the organizations’ external business relations whereas the opinion data items related to the organizational context are embedded in the internal affairs of the organization.
7.3 Organizational context Three themes are included in the opinion data items related to the organizational context. First an item related to operational performance or more specifically reasonability of EDI adoption is explored. This issue is operationalized in relation to possible savings. In the studies reviewed the theme of reasonability of IT adoption was a dominant theme. In order to move beyond this theme two other types of opinion data items were included in the examination of motivators for adoption of EDI related to the organizational context: 1) Human resources and work environment, 2) The organizations’ perceived capability of EDI adoption.
248
Opportunities for savings and profits due to EDI adoption were some of the major issues in the 1996-action plan for EDI (Ministry of Research and Information Technology, 1996). EDI was considered as a means for transmitting large quantities of data fast, inexpensively and without error. In the foreword to the 1996-action plan it was stated that the adoption of EDI could lead to, “… optimum conditions for improved service towards their customers, greater internal efficiency, optimum conditions for cooperation with external suppliers and – above all – make them ready to compete in the global, electronic marketplace.” It was therefore natural to explore how the companies that had been exposed to the 1996-action plan perceived the profitability of EDI adoption. The adopters in the TDP had different opinions regarding the economic benefits of EDI. The companies that were merely exchanging EDI messages with their parent company or another company in their industry group stated that savings had been realized. Recalling the VIDS-test (cf. Table 4-6, page 126) two of these companies reported that the volume of EDI messages compared to the total number of messages exchanged with suppliers was respectively 40 percent and 85 percent. Based on the VIDStest volume dimension it is understandable that the adopters find adoption of EDI to be profitable. On the other hand, the two adopters that used EDI with a limited number of suppliers and customers said that economic savings were absent in relation to EDI. The non-adopters expressed that they did not expect that adoption of EDI would lead to any direct savings. Generally profitability has been found to be a motivator for adoption (Attewell, 1992). One rationale for adopting EDI is an expectation of increased efficiency due to improvement of intraorganizational and interorganizational routines (Timmers, 1999). Improved operational performance was found to be one of the prevalent themes in the review of the articles from the top-five MIS-journals (cf. Table 5-3, page 180). Among the operational performance themes were: General performance improvements (Clark and Stoddard, 1996), accurate exchange of business information (Srinivasan et al., 1994), and benefits related to integration of EDI (Premkumar et al., 1994; Massetti and Zmud, 1996; Truman, 2000).
249
Direct savings are rarely reported in EDI studies (Cox and Ghoneim, 1996; O’Callaghan and Turner, 1995). An exception is the study by Mukhopadhyay et al., (1995) who found that the Chrysler Corporation using EDI gained over $ 100 per vehicle manufactured. Indirect savings have also been explored. These savings can be related to reduction in the workforce due to less re-keying and a decreased need for manual storing of documents, lower inventory costs, and shortened duration of transactions (O’Callaghan and Turner, 1995). The following proposition is made to examine whether the possible savings related to EDI have influenced a company’s motivation for EDI adoption: Proposition O1: Prospects of future savings motivate EDI adoption. EDI can lead to streamlining of business activities. One challenge is to identify and maintain levels of personal expertise in business operations (Swatman and Swatman, 1992). The two policy statements preparing the Danish citizens and businesses for the “information revolution” from 1995 and 1996 (cf. Section 3.4.2) were both focused on the possible benefits for employees resulting from introducing IT in the organizations. In the policystatement “From vision to Action – Info Society 2000” (Ministry of Information Technology and Research, 1995) IT was considered as the means for everybody to get involved and educated. It was expected that IT would contribute to the personal development of the employees since the individual employee would be able to communicate more easily thereby obtaining the information needed for improved performance of work routines. The next policy-statement “The Info-Society for All – the Danish Model” (Ministry of Information Technology and Research, 1996b) stated that adoption of EDI could lead to, “… improvement in efficiency of working procedures and development of new production processes.” These considerations related to work environment and human resources prompted an examination of how these issues influence motivation for EDI adoption. During the interviews none of the TDP participants paid particular attention to issues like work environment and human relations. Only one of the interviewed companies stated that adoption of EDI had freed some
250
manpower from the accounting department. It had however, not led to any lay-off. The employees had instead been posted to other job functions where they worked with improving customer relations. From studying the minutes of the work group meetings it was revealed that the IT-consultant had put human resource and employee related issues on the agenda a number of times. The participants had however, been more interested in defining the industry EDI subset. Tornatzky and Fleischer (1990) focus on human resources in relation to the organizational environment. According to Tornatzky and Fleischer the human resources are related to the quality of human resources. The review of IS studies concerning the organizational context (cf. Table 6-3, page 227 and Section 6.7.1) included issues such as adequate education (Kurnia and Johnston, 2000) and employees’ IS knowledge (Thong, 1999). The EDI literature has to a limited extent focused on issues related to work environment and human resources. Swatman and Swatman (1992) point out that adoption of EDI may lead to organizational restructuring involving staff retraining due to changing staff functions. The above-mentioned TDP example is an illustration of changes in staff functions. Especially training of all relevant employees has been found to be one of the major determinants for SMEs gaining benefits from EDI adoption (Raymond and Bergeron, 1996). O’Callaghan and Turner (1995) explicitly mention that, “Companies may be able to operate with fewer personnel, or reassign them to more productive activities.” In this context the “more productive activities” are of major interest. The savings in manpower due to elimination of re-keying and manual reconciliation, reduced time spent correcting errors, and by freeing professional personnel from administrative tasks can lead to these “more productive activities”. In order to investigate the importance of work environment and human resources in relation to motivation for EDI adoption three propositions were formulated. These three propositions are closely interrelated. They do, however, illustrate slightly different approaches to work environment and human resources. Proposition O2 is related to, whether or not adoption of EDI will create better work conditions for employees leading to more independent
251
job functions for the employees. O’Callaghan and Turner (1995) characterized this situation as “Freeing professional from administrative tasks.” Proposition O2: The notion that EDI will create a better work environment is a motivator for EDI adoption. Proposition O3 is related to whether or not adoption of EDI will lead to training and education of employees. Retraining due to changes in staff functions as described by Swatman and Swatman (1992) is the underlying assumption for this proposition. Proposition O3: The notion that EDI will benefit the development and utilization of human resources is a motivator for EDI adoption. The theoretical inspiration for Proposition O4 can be found in the oftenclaimed benefit of EDI related to the elimination of redundant re-keying and elimination of manual reconciliation (O’Callaghan and Turner, 1995; Arunachalam, 1995). One improvement of the work environment is seen in relation to the elimination of trivial work and routines. Proposition O4: The notion that EDI will eliminate trivial work is a motivator for EDI adoption. The representatives from the professional business associations had often heard the excuse “Our business activities are not suitable for EDI.” The non-adopters in the TDP expressed the same view. The stated reason was that their business activities were focused on custom-made products and therefore not suitable for EDI. Even though EDI messages in most cases are related to administrative standard procedures such as purchase orders, invoices, and payment notices both non-adopters and adopters within manufacturing in the TDP argued that their products were not really suitable for EDI messages. They argued, that were not buying and selling batches of raw materials and products.
252
Most EDI studies have focused on commodities and standardized products such as aircraft parts (Choudhury et al., 1998), hospital supplies (Steinfield et al., 1995), and office supplies (Jelassi and Figon, 1994). Research has especially shown broad implementation of EDI in the automotive industry (Tuunainen, 1998; Mukhopadhyay et al., 1995), and in the grocery sector (Andersen et al., 2000). Commodities and standardized products characterize both these sectors. Even though EDI is useful for exchanging business information regardless of whether the item is a commodity or something highly specific, the EDI literature and practice have so far mainly concentrated on commodities. To investigate whether or not the type of business activities influences the motivation for EDI adoption the following proposition is formulated: Proposition O5: The decision-makers’ awareness that the company’s business activities are well suited for EDI is a motivator for EDI adoption. The sixth proposition related to the organizational context deals with organizational readiness for EDI adoption. The two policy statements from 1995 and 1996 were meant as IT and EDI icebreakers for private businesses and public institutions. As discussed in Section 3.6 the policy statements and the 1996-action plan for EDI were seen rather as a pedagogical regulatory intervention than as an economic or normative intervention. As mentioned in Section 3.6, the pedagogical intervention is characterized by information campaigns initiated by governmental units and larger associations where the aim is to influence the opinion of a given group of potential adopters. The underlying assumption is therefore, that the addressees of the information campaigns are expected to perceive themselves as being more prepared for e.g. EDI adoption. Rogers (1995) stresses the significance of the nature of the social systems and the extent of change agents’ promotion efforts as important variables determining the rate of adoption. These two elements are interpreted to play a central role in relation to the effect of the pedagogical intervention because the “ideological” support from opinion leaders adds value to the content of the pedagogical intervention (See for example Textbox 3-1, page 74). The value of this social process should therefore not be underestimated in this
253
context. The test whether or not the pedagogical intervention and social processes have had any influence on the motivation to adopt EDI the organizational readiness was tested in the following Proposition O6. The participants in the TDP were subject to pedagogical intervention from their business associations who directly stated that one of the objectives of the TDP was to create and share knowledge of EDI (cf. Table 4-1, page 95). The non-adopters expressed that they had received useful information from the project, which made them more confident in making future decisions on EDI. The informant who represented the company that adopted EDI during the project expressed the same opinion. The adopters on the other hand demonstrated that their benefit in relation to increased knowledge of EDI was rather limited. One of the “organizational context studies” (cf. Table 6-3, page 227 and Section 6.7.1) directly referred to organizational readiness for EDI (Iacovou et al., 1995). Iacovou et al. related organizational readiness to the level of financial and technological resources. In the top-five MIS review different approaches to organizational readiness were presented. Lai and Guynes (1997) referred to employee’s positive attitude to organizational change. Premkumar and Ramamurthy (1995) included the influence of innovation champions and top management support. One aspect which has been seen as a parameter for organizational readiness for EDI adoption is related to whether the adopters are EDI initiators or followers (Premkumar and Ramamurthy, 1995; Swatman and Swatman, 1992). The companies that are persuaded or directly forced to adopt EDI are less prepared for EDI and they as a consequence do not reap the full benefits of EDI immediately - if ever. In order to investigate the importance of organizational readiness in relation to motivation for EDI adoption the following proposition was formulated: Proposition O6: The notion that companies considering themselves to be well prepared for EDI are more likely to adopt EDI.
254
7.4 Environmental context Two themes are included in the examination of the environmental context. The themes are similar to the reviewed studies in Section 6.7.2 focused on the two themes: Market driven forces and interorganizational forces. The market forces are primarily related to competitive issues whereas the interorganizational forces are related to different degrees of pressure. The five opinion data items related to market driven forces and interorganizational forces are presented in detail in the following. As described in the previous section related to the organizational context one aspect of the content of the policy statements from 1995 and 1996 was to prepare the organizations for the upcoming technological shift. Aspects such as competition and technology were however inherent elements in the communicated information from the involved ministries. Considerations related to the environmental context can therefore not be excluded in an evaluation of the impact the statements may have had on the common EDI knowledge in the business community. The 1995-policy statement explicitly mentioned that adoption of EDI could lead to closer interplay between organizations thereby strengthening their competitive position in the market. A similar message was given by one of the opinion leaders that greeted the arrival of the 1996-action plan (cf. Textbox 3-1, page 74). Here it should be stressed that one of the primary objectives of the business associations is to support their members in relation to competitiveness. The policy makers from the two business associations were as a consequence very interested in the importance of EDI as a means for the organization’s perceived improved competitiveness. Strategic considerations related to EDI adoption were evident in the TDP. This was the case for both the non-adopters and for those adopters which were initiators of EDI. The non-adopters stated that the competitive needs for EDI had not yet been a dominant force. The non-adopter that was in the process of becoming an adopter had however realized, that EDI adoption was a must in order to prevent loss of market position. The adopters that had taken the position of EDI initiators in the industry had primarily made 255
the move due to strategic considerations. They wanted to make sure that they were able to meet demands for EDI from business partners thereby being attractive business partners. Recalling that the environmental context is the arena in which the organization conducts its business (Tornatzky and Fleischer, 1990) the improvement of strategic performance is an important issue in relation to IOS adoption. Three of the IOS studies, which included the environmental context in their survey instrument, focused on competition (Kurnia and Johnston, 2000; Thong, 1999; Grover, 1993). Competitive need was found to be an important variable influencing adoption of IOS (Grover, 1993). In the top-five MIS review strategic performance directly related to competitive advantage was an often-visited theme. A number of studies found that adoption of EDI could lead to competitive advantages such as improved competitiveness (Chatfield and Bjorn-Andersen, 1997), new business opportunities (Jelassi and Figon, 1994), and changes in interfirm processes and politics (Lee et al., 1999). To investigate whether competitiveness influences the motivation for EDI adoption Proposition E1 was formulated. Proposition E1 is directly related to possible improved competitiveness due to EDI adoption. The proposition rests on some of the “traditional” assumptions related to EDI (Pfeiffer, 1992; Dearing, 1990; Arunachalam, 1995; O'Callaghan and Turner, 1995). These includes a possible increase in the company’s competitiveness due to e.g. improved customer service, shorter lead times, and more timely information about transaction status. Proposition E1: The prospect of improving the company’s competitiveness motivates EDI adoption. A different view of competitiveness, which has received some attention in research, is the argument brought forward by two of the adopters in the TDP. They argued that they saw adoption of EDI as a means for becoming an attractive business partner thereby getting the opportunity of increasing their market share. Proposition E2 is related to the strategic alliances
256
between business partners for the purpose of maintaining a competitive edge (Chau and Tam, 1997). The perspective represents the opposite view to the one given by Iacovou et al., (1995) who explored risk of discontinuation of partnerships in case of non-adoption. Proposition E2: The prospect of increasing the company’s market share is a motivator for EDI adoption. The items related to interorganizational forces contain issues related to power and pressure. The pressure and power perspective as a driver for IS adoption refers to the obligation of a firm to adopt an innovation in order to keep a customer or supplier content (Hart and Saunders, 1998). Tornatzky and Fleischer (1990) stated that the external environment should not be seen only as external forces beyond control of the firm, but it should also be remembered that some firms have the ability to shape their environments. External forces are examined in this context. The two policy statements and the 1996-action plan did not directly elaborate on perspectives related to pressure. Though the 1996 policy statement explicitly mentioned that Denmark should pursue an aggressive strategy for IT adoption no normative or economic regulatory arrangements were made. As mentioned earlier the pedagogical strategy was pursued instead. The strategy involved inclusion of strong national symbols such as N.S.F. Grundtvig and the cooperative movement in farming to stress the importance of action. No pressure was however, placed on individual organizations. The TDP was also designed as a project where organizations could join on a voluntarily basis. And even more important the project initiators did not include any conditions in the project description related to EDI adoption amongst the participants in the project. The business associations did therefore not consider peer pressure (Premkumar and Ramamurthy, 1995). From interaction with the representatives from the business associations and from the Danish EDI Council it became clear that they found that pressure related to EDI adoption was a very “un-Danish” phenomenon.
257
However, during the interviews it became evident that pressure most likely was one of the primary motivators for EDI adoption in Danish companies (cf. Table 4-4, page 123). Three out of the eight interviewed companies had adopted EDI due to a direct order from their parent company or the industry group they belonged to.88 The remaining two adopters were in a position where they were able to put pressure on their customers and suppliers. One of the companies had (with some success) tried the strategy of exerting direct pressure on their business partners on order to make them adopt EDI. Amongst the non-adopters two of the companies explicitly expressed that the reason for their status as non-adopters was that none of their business partners had put pressure on them in relation to exchange of EDI messages. The third non-adopter, which had decided to adopt EDI in the near future, said that pressure was not the reason for their adoption of EDI. The informant actually emphasized that she considered it to be “very un-Danish”. This attitude may however, be the exception that proves the rule of pressure as being an important motivator for adoption. As detailed in Section 6.7.2 issues related to pressure was an oftenmentioned theme in the operationalization of the environmental context in IOS studies. Pressure was also a theme in a number of the articles reviewed from the top-five MIS journals (cf. Table 5-3, page 180). Hart and Saunders (1998) explored the different ways of exerting power in relation to business partners. They distinguished between persuasive and coercive power. Iacovou et al. (1995) distinguished between competitive pressures and imposition by trading partners. Bergeron and Raymond (1992) included the benefits from strategic repositioning of the firm due to implementation of EDI in their survey. In order to investigate the importance of pressure as a motivation for EDI adoption the three propositions were formulated. Pressure related to imposition of business partners (Iacovou et al., 1995) was operationalized in Proposition E3 which is related to having some information that EDI is being used amongst business partners. 88
The relationship between legal ownership and EDI adoption is separately tested in Proposition 2 in Section 8.2.3.
258
Proposition E3: The knowledge that several business partners already use EDI is a motivator for EDI adoption. Increase in the degree of pressure compared to Proposition E3 is outlined in Proposition E4. Proposition E4 is related to the situation where the company is subject to persuasive power (Hart and Saunders, 1998). The company is not directly forced to adopt EDI but business partners may take steps such as informing about EDI benefits and offering assistance in relation to the adoption and implementation process (ibid.). Broadly speaking the recommendations could also originate from business associations and other opinion leader that influence the companies. Proposition E4: The fact that EDI has been recommended by others is a motivator for EDI adoption. The third proposition is related to direct pressure from business partners. The pressure can take different dimensions ranging from promises to threats (Iacovou et al., 1995). Promises include rewards such as rebates due to EDI usage and threats include negative sanctions such as suspension of the partnership. Proposition E5: The fact that the company is put under pressure to use EDI is a motivator for adoption of EDI.
7.5 Technological context One inevitable problem in relation to an examination of the three contexts in one questionnaire is the wide scope of the three contexts. As described in Section 2.5.1 it is difficult or nearly impossible to find a single person in an organization that can provide answers to all questions related to the organization. In order to avoid a situation where the manager, who had received the questionnaire, was unable to answer issues related to the technological context, the questions were formulated in general terms focusing on the managerial aspects of technology. Two categories of items
259
related to the technological context were included: Attributes related to the EDI solution and organizational innovativeness. The attributes related to the EDI solution are focused on the technical level of EDI and the cost of EDI. In the 1996-action plan the general technological developments and the decrease in hardware and software prices were seen as an advantage in relation to the diffusion of EDI in Danish businesses. The general attitude as expressed by the representatives from the two business associations and the representative from the Danish EDI Council was, that the technical obstacles related to the initial problems with EDI had almost been overcome. One issue that played an important role was the new means for transportation – the Internet – that had been introduced in most businesses. The Internet represented an alternative to VANS and private lines, and it was found to reduce the complexity and in particular the cost of establishing an EDI link. The TDP case inspired the inclusion of attributes related to the technical level of EDI solutions and the cost of EDI solutions. One of the objectives in the TDP was to operationalize the 1996-action plan by means of developing an inexpensive EDI software which could easily be adopted in the organization for the benefit of the members of the two industry and trade associations involved in the project. The experiences from the study of the TDP did, however at first reveal disturbing conclusions in relation to the importance of the cost of EDI solutions and the level of the technical solution. The TDP-software was offered to the participants at a marginal cost. However, only two of the participants were interested in adopting the TDP-software. The general attitude towards the software was that it was not sufficiently developed. In order to be attractive the software should at least be capable of being integrated with the organizations’ administrative systems. Since that was not the case the participants were not interested in adopting the software due to the additional costs of establishing the necessary integration. This indicated that cost is not the only technological attribute that matters in relation to EDI adoption. The technical level of the innovation also seems to play a major role.
260
Research (Andersen et al., 2000) has similar to the policy makers from the business associations found that the internet represents an alternative to VANS and private lines, and that it can reduce the complexity and in particular the cost of establishing an EDI link (Falch, 1998). Another aspect related to technical complexity is related to standards for EDI communication. This issue related to standards is not yet solved (Damsgaard and Truex, 2000). Other things being equal, the more costly an innovation is, the less likely it is to be adopted, but once it is adopted and adapted, the large investment may strongly motivate implementation the innovation (Cooper and Zmud, 1990). As for the cost of technical solutions, it can be seen as the direct price of purchase or in relation to both direct costs and expenses resulting from education and training of employees (Raymond and Bergeron, 1996). In order to examine the influence and importance of the technical level and the cost of the innovation as motivators for EDI adoption two propositions were formulated. Proposition T1 is related to how managers perceive the importance of the technical level of EDI. Instead of specifically investigating issues related to standards, means for transportation, or prospect of integration this opinion item is kept in very general terms. Proposition T1: A satisfactory technical level of IT solutions is a motivator for EDI adoption. Proposition T2 is directly related to price. A theoretical reference to T2 is the relative advantage of EDI adoption (Rogers, 1995) or cost-benefit considerations. Proposition T2: A satisfactory price level of IT solutions is a motivator for EDI adoption. The following propositions, T3 and T4, can be seen as an indirect investigation of the effect of the information campaigns on EDI launched by the Ministry of Information Technology and Research and the efforts
261
done by the business associations and the Danish EDI Council to inform the business community of EDI. The policy statements and the 1996-action plan can be seen, as reminders to Danish businesses that they in order to avoid the “Global digital divide” technologically had to upgrade themselves. The major concern in the business associations was that their members were not paying enough attention to the innovation and therefore could be left behind in the international markets. The spirit in the statements were that EDI (and e-commerce) was something new and interesting and that companies should get involved in order not to be left behind the technological developments. In the 1996-action plan it was acknowledged that EDI might not be the most modern technology but that it still represented technological advances for the adopters. The participants in the TDP were divided with respect to the issue of the importance of the technological developments. The more advanced users of EDI opined that EDI might be out-competed by B2B e-commerce in the near future. However, they did not show much concern on that account. This reaction can be interpreted as a form of technological maturity in the sense that the adopters already had experience with and confidence in the interorganizational technologies. The less advanced EDI users did not pay attention to competing innovations. They used EDI because their parent company or the industry group had told them to do so. The non-users were more focused on EDI which they had become familiar with from their participation in the TDP. EDI still represented something new and would therefore be considered an innovation in their organizations. None of the reviewed studies concerning EDI or adoption of IOS specifically included issues related to the threat of “technological marginalization” due to reluctance to adopt a technological innovation. The reviews were rich in examples of “economic marginalization” in the sense that non-adoption for example could lead to weakened competitive advantages. The adoption and diffusion theory on the other hand is rich in examples related to the issue of technological marginalization. The S-curve often used to explain diffusion of technology (Attewell, 1992) conceptualizes the technological marginalization in the sense that it
262
postulate that adopters can be divided into five groups: Innovators, early adopters, early majority, late majority, and laggards (Rogers, 1995). The last opinion data items related to the technological context can from a theoretical point of view be related to the issue of managerial fashion innovation (Abrahamson, 1996; Newell et al., 2000; Galliers, 1999). As shown in Section 6.2.2 the fads and fashion perspective is related to the situation where managers actively search for innovations that can upgrade their businesses technologically (Newell et al., 2000). In order to test whether or not this type of pursuit for technological innovations is present in the Danish steel and machinery industry issues related to this were included in the survey. Therefore, in order to investigate the significance of innovativeness as a motivation for EDI adoption two propositions were formulated. The first perspective is related to the situation where the potential adopter is in a neutral position towards the innovation per se. However, mere knowledge that not having the innovation might exclude the company from being up-front might serve as a motivator for adoption. The threat of being a laggard (Rogers, 1995) with respect to adoption is formulated in Proposition T3. Proposition T3: A feeling of being left behind with respect to EDI is a motivator for EDI adoption. Proposition T4 is directly related to the fads and fashion phenomenon presented by Abrahamson (1996). It should be noted that though researchers do not perceive EDI as new and interesting this might not be the case for practitioners. As discussed in Section 6.4.1 innovation is seen as a relative term, which is conditioned by the perception of the potential adopter. Proposition T4: The notion that EDI is new and interesting is a motivator for EDI adoption.
263
After this presentation of the operationalization of the three contexts that were used to systematize the empirical and theoretical input a statistically based exploratory search for explanatory factors was carried out. The purpose was to uncover which factors influenced motivation for EDI adoption for each of the fifteen items just presented. This search was carried out separately for adopters, planners, and non-adopters. The results from this statistical exploratory search are presented in the next chapter.
264
8 A survey of the motivation for EDI adoption in the Danish steel and machinery industry 8.1 Introduction So far empirical input has focused on qualitative data from the TDP case study. At this point a change in focus is made. This new orientation is chosen in order to triangulate data in search for motivators leading to adoption of EDI. Focus is still on companies in the steel and machinery industry, but a quantitative assessment is made of a larger sample of the same business sector. The steel and machinery industry in Denmark is contrary to the sample in the TDP dominated by a large number of small companies (cf. Table 2-4, page 375, which presents the distribution of size among manufactures in the entire business sector). This is a new perspective. Now focus is on small and micro-sized companies89 instead of large companies and SMEs. This new focus is a result of the sources of data that were made available. Another change is the fact, that the companies included in the survey sample have not to the same degree as the TDP participants received information about EDI. The survey sample has in trade journals received some information from their business associations and from general EDI campaigns. This means that the study sample in the survey compared to the TDP to a much larger degree resembles “the average company” in relation to knowledge about EDI.
89
The definition of SMEs includes three types of enterprises: Medium sized enterprises, small sized enterprises and micro-enterprises. The definitions are according to the European Commission based on three parameters: Maximum number of employees, maximum turnover, and maximum balance-sheet total (http://europa.eu.int/ISPO /ecommerce/sme/definition.html). Micro-enterprises have a maximum of 10 employees, small enterprises have a maximum of 50 employees, and medium sized enterprises employ a maximum of 250 people.
265
However, due to various reportings of results from the TDP, which has been communicated in the trade journals from the business associations, there is an assumption – perhaps unfounded – that the companies in the study sample are well informed about EDI.90 Despite the differences in relation to company sizes and the level of general knowledge of EDI an attempt is made to statistically test some of the findings from the TDP. In the following sections three propositions inspired by insights from the TDP are tested statistically based on data obtained from the survey. Thereafter, the strategy for testing the propositions presented in Chapter 7 is outlined. Finally, the chapter ends with an analysis of data where the main purpose is to uncover which of the fifteen propositions that provide the best explanatory power with respect to motivation for adoption or non-adoption of EDI amongst adopters, planners, and non-adopters. The following strategy for reporting data and results is chosen. Output from the statistical packages, SAS and the DIGRAM, are to be found in an appendix chapter, Chapter 12/ Appendix B. Most tables are found in Chapter 12 and references are provided in the present chapter. Appendix B in Chapter 12 also includes an overview of abbreviations used in the statistical analysis of the motivation for adoption of EDI. All variable names included in a given analysis are written in Italics. It was decided to be both practical and appropriate to present a number of statistics under each of the presented tables. These statistics are all based on the Pearson chi-square. There are several reasons for this decision. Presenting this family of statistics will make it easier for the reader to compare a given statistic with a statistic from a similar table reported in other studies. Furthermore, the reader can easily see what will happen if assumptions regarding scales are questioned. If for instance both table variables can be assumed to lie on an ordinal scale then the Mantel90
The percentages of adopters, planners, and non-adopters in the present survey, that indicated they were interested in receiving more information about EDI were respectively about 35, 60, and 40 percent.
266
Haenszel chi-square statistic would be appropriate statistic. If the reader on the other hand is of the opinion, that the scales are mere nominal then the value for the chi-square statistic is available. Format and layout considerations are another reason. It is the authors’ experience that presenting information in a standardized format generally makes it easier to find and read the presented information.
8.2 Quantitative assessment of insights gained from the TDP 8.2.1 Introduction Data from the interviews indicated that the legal status of the organization influenced adoption of EDI. As previously mentioned all the users of EDI in the TDP were either part of an industry group or a subsidiary. Company size also seemed to play a role for adoption. In the TDP it appeared that there was a tendency for larger companies compared to smaller companies to be adopters of EDI. Finally, data from the interviews indicated that wholesalers compared to manufactures were more motivated for adopting EDI. These three explanations for adoption were related to primary innovation attributes (Downs and Mohr, 1976) that could be measured and calculated by looking at legal status, company size, and organizational activities. These three motivations for EDI adoption are formulated explicitly in propositions 1 to 3. Proposition 1: The larger the size of the company the more likely the company is to adopt EDI. Proposition 2: Being part of an industry group or a subsidiary is positively related to EDI adoption. Proposition 3: Being a wholesaler is positively related to EDI adoption.
267
8.2.2 Exploring the relationship between the size of the company and adoption of EDI A number of studies have shown that the size of a company91 positively influences adoption of innovations (Attewell, 1992; Thong, 1999). The most common explanation for this relationship between size and adoption of innovations is the higher level of resources due to the size of the firm. Additionally, it has been argued that larger firms may be involved in a greater variety of production activities, and they are thus more likely to find any given innovation applicable to their operations (Tornatzky and Fleischer, 1990). The EDI figures from Statistics Denmark for the year 2000 (Ministry of Information Technology and Research, 2001) concluded that about fifty percent of companies with one hundred or more employees had adopted EDI. In comparison the average percentage of adopters amongst all businesses in Denmark was fifteen percent (ibid.).92 One study (Iacovou et al., 1995) especially focusing on EDI adoption in small firms found that there was no relation between size of the company and EDI adoption. This contradicts the observations from the TDP. In the study of the TDP it was observed that the size of the company in some way influenced the decision to adopt or not adopt EDI. The larger companies in the TDP had all adopted EDI, whereas the smaller companies involved in
91
Company size is measured as the number of employees. A number of studies have used the number of employees as a measure of company size (Thong 1999; Crag and King, 1993). Other studies have calculated company size based on annual turnover (Premkumar and Ramamurthy, 1995; Ramamurthy et al., 1999), whereas some studies have applied both number of employees and revenue as indicator for company size (Grover, 1993). The use of number of employees to represent company size is however, often used as measurement in studies of small businesses (Thong, 1999). The choice of means of measurement in this study was primarily based on environmental constraints. The industry and trade associations, which provided additional data to the survey, did not have data concerning annual turnover or other size parameters available except number of employees. Alternatively, a question concerning annual turnover could have been included as a question in the self-administered questionnaire. It was, however, expected that most responders would consider this type of information confidential. It was hence decided to base size on number of employees, which can be cross checked in publicly available information in Danish business directories such as KRAK and Greens Business information. 92 Attention should be paid to concentration of small businesses in Denmark (cf. Section 3.1).
268
the pilot-project were non-adopters.93 As mentioned in the introduction to this section the following proposition was formulated: Proposition 1: The larger the size of the company the more likely the company is to adopt EDI. The size of the company was divided into six categories.94 The variable Company size is not based on self-reported data from the responders. Instead the business associations calculated the figures from a retrieval of their member-database. To estimate the strength of the association between Company size and Adoption level a three-by-six table Adoption/ Planning/ Non-adoption by Size was analyzed. Data on adoption level and company size Table 8-4 (page 378) shows that the variables Company size measured as number of employees and Adoption level in a statistical sense are not independent (Chi-square p-value = 0.0015). The ratio of EDI adopters in larger firms (50+) is approximately one-third (13/ 37) compared to the non-adopters, which is about one tenth (11/ 108) given the same company size. The ratio of planners given company size equal to 50+ (15/ 50) is more like the ratio of adopters (13/ 37) given the same company size. In the 50+ company size the ratio for non-adopters is much less (11/ 108). More than forty percent of the nonadopting companies have less than ten employees whereas the percentage of the smallest companies (1 to 9 employees) that have adopted EDI or which are planning to adopt EDI is about sixteen percent. Three sets of two-by-three tables, Non-adopter/ Planner by Company size (Table 8-5. Adoption level: Non-adopters and planners versus company 93
The company size in the sample included in the TDP and in the survey of the Danish steel and machinery industry varies considerably. About 99 percent of the companies included in the survey are micro-enterprises (1 to 9 employees) or SMEs. In the TDP six out of nine companies had more than 250 employees and they are therefore not considered as SMEs. In this sense the TDP-sample is not representative of the Danish steel and machinery industry. 94 The applied categorization resembles the categories used by the business associations and Statistics Denmark. This categorization reflects the large share of small businesses in Denmark: 1 to 5, 6 to 9, 10 to 19, 20 to 49, 50 to 99, and 100+ employees.
269
size, page 379), Non-adopter/ Adopter by Company size (Table 8-6. Adoption level: Non-adopters and adopters versus company size, page 380), and Planner/ Adopter by Company size (Table 8-7. Adoption level: Planners and adopters versus company size, page 381) were analyzed. The objective of the analysis was to examine the relationship between adoption level and company size. In this analysis both variables are assumed to lie on an ordinal scale, therefore the Mantel-Haenszel Chi-square test is appropriate. The analysis of the two tables Non-adopter/ Planner by Company size (Mantel-Haenszel Chi-Square = 8.7258; p-value = 0.0031) and Non-adopter/ Adopter by Company size (Mantel-Haenszel Chi-Square = 7.9560; p-value = 0.0048) revealed strong dependency between company size and adoption level. However, the analysis of Planner/ Adopter by Company size (Mantel-Haenszel Chi-Square = 0.0216; p-value = 0.8830) suggested that the two variables are independent. Proposition 1 stated that, “The larger the size of the company the more likely the company is to adopt EDI.” Based on the results presented above proposition 1 can not be rejected. Additionally, it is found that the ratios of planners and adopters given the various company sizes are very similar. Originally the focus was on adopters and non-adopters. However, about one quarter of the respondents indicated that they planned to adopt EDI in the near future. It can be argued that merely indicating an intention to change behavior does not necessarily lead to an actual change of behavior (Harrison et al., 1997). The argument for including data for planners is that refinement of the categorization adopter/ non-adopter can yield more useful information. The actual finding related to adoption status can support the applied definition of adoption (cf. Section 6.2.3), where arguments were given in favor of the Tornatzky and Fleischer (1990) definition. Though Tornatzky and Fleischer distinguish between having and not having an innovation. The term “having” can according to the authors be interpreted broadly to included those situations where a decision to adopt has been made. The Tornatzky and Fleischer definition in its broadest interpretation seems to be
270
useful in regard to adoption and planning to adopt with respect to company size. 8.2.3 Exploring the relationship between legal status of the company and EDI adoption None of the studies included in the review of adoption studies focusing on IS has explored the relationship between adoption and the legal status of the company. Broadly speaking the term legal status can be divided into two major groups: 1) Independent companies, which are organized, as individual owned units or as joint stock companies. 2) Dependent companies, which are either subsidiaries or part of national or international industry groups. The dependent companies have compared to the independent companies less right of self-determination and they usually have close trading connections to the parent company or other companies in the industry group. It can hence be argued that the option to reject the innovation is non-existent if top management has decided that EDI is to be adopted in the corporation. If this is the case then the determinants for adoption must be considered to be pressure rather than to be based on free choice. A number of studies have focused on the significance of pressure in relation to adoption of EDI (Hart and Saunders, 1997; Premkumar and Ramamurthy, 1995; Iacovou et al., 1995). However, these studies have not considered the legal status of the involved companies. EDI adoption among the dependent companies was the norm in the TDP case. All the five companies that had adopted EDI were dependent companies. Some of the informants directly expressed that the adoption of EDI was a result of a request from the parent company or another company in the industry group. Based on this observation the following proposition was formulated: Proposition 2: Being a part of an industry group or a subsidiary is positively related to adoption of EDI. To analyze a possible dependency between legal status and adoption the chi-square statistics was calculated on a two-by-three table, Dependence by Adoption/ Planning/ Non-adoption. 271
The table on dependence and adoption level (Table 8-8, page 382) shows that the chi-square value equals 11.5884 and the corresponding p-value is 0.0030. Therefore, it is reasonable to assume that the two variables Dependence and Adoption are dependent. The percentage of adopters among dependent companies is 52.94 percent, which is very similar to the percentage of dependent planners, 54.72 percent. However, amongst the non-adopters 70.09 percent belong to the group of independent companies, and 29.91 percent to dependent companies. Proposition 1 stated that size of the company was related to the rate of adoption of EDI. As mentioned above it was concluded that proposition 1 could not be rejected. With respect to proposition 2 a question can be raised: Is this dependency between adoption and size of a company independent of the legal status of the company? Or to be more specific: Is the influence of the legal status more powerful than the size with respect to adoption of EDI? To test whether adoption is independent of company size given legal status two three-by-three tables were analyzed, respectively including dependent and independent companies. The test statistics for adoption level and company size for dependent companies (Table 8-9, page 383) suggests that for dependent companies the adoption level and size of the company are independent (Mantel-Haenszel Chi-Square = 2.6363; p-value = 0.1044). This suggests that the parent company or the industry group decide whether or not EDI is to be adopted regardless of the size of the company. For adoption level and company size for independent companies (Table 810, page 384) the Mantel-Haenszel chi-square statistics is equal to 4.1429 and the corresponding p-value is 0.0418. This is statistically significant though not highly significant. The table shows that the larger the size of the company the smaller the ratio of non-adopters to the total number of independent companies. The percentage of non-adopters with 1 to 19 employees is 71.83 percent compared to 4.23 percent for companies with 50+ employees. This could indicate that the smaller and independent companies have their own reasons for not adopting EDI. One reason could be that the smaller companies expect that adoption of EDI will not be 272
profitable. As detailed in Chapter 5 the general belief is that especially EDI initiators compared to followers derive benefits from EDI. Studies on EDI (Krcmar et al., 1995; Srinivasan et al., 1994) generally show that larger firms are EDI initiators whereas small companies are followers. Another reason could be limited resources allocated for IT-investments. Yet, another reason could be that they are not large suppliers and therefore not subject to pressure from their customers. An examination of Table 8-9 for dependent companies and Table 8-10 for independent companies shows respectively that 16.67 percent and 46.36 percent of the non-adopters are found amongst the smallest companies. This suggests that there are relatively fewer non-adopters among dependent companies. Proposition 2 stated: Being a part of an industry group or a subsidiary is positively related to adoption of EDI. Examination of Table 8-8 suggests that proposition 2 can not be rejected. It was found in Table 8-9 that there is no dependence between company size and adoption level. This suggests that size is not a determining factor for adoption among dependent companies. Dependent companies are adopting EDI regardless of company size whereas only larger independent companies adopt EDI (cf. Table 8-10, page 384). This suggests that the lager independent companies find that there is a rationale – operational or strategic - for adopting EDI. For dependent companies, small or large, the adoption is less voluntary. 8.2.4 Exploring the relationship between manufactures and wholesalers and EDI adoption Throughout the 1990s the e-commerce literature has generally predicted the disappearing role of the intermediary. A phenomenon, which has been labeled disintermediation (Zwass 1996; Wigand 1997; Choudhury, 1997). Disintermediation has been predicted as a result of improved means of communication (Malone et al., 1987). It has been suggested that electronic markets may be able to replace the search function performed by intermediaries in transactions characterized by high technological
273
uncertainty and high market variability (Choudhury, 1997).95 However, so far conceptual considerations rather than empirical evidence have supported the prophecies of disintermediation in relation to business-tobusiness connections. The threat of being ousted could be seen as an incentive for wholesalers to adopt tools such as EDI, that promises improved performance. A representative from the wholesaler’s business association mentioned during our conversations that he often heard from wholesaler’s business association members that they feared, “In the future to be reduced to function as a data-base”. This threat was according to the representative one of the major incentives for wholesalers to adopt EDI because the wholesalers anticipated that EDI was likely to improve customer service thereby make the wholesalers valuable in the supply-chain. No studies known to the author have compared manufactures and wholesalers and none has explored the relationship between EDI adoption and the position in the supply-chain.96 The sample size of the TDP comprised nine companies. Of these nine companies six companies had adopted EDI. Four of these six companies were wholesalers. Based on this observation the following proposition was formulated. Proposition 3: Being a wholesaler is positively related to EDI adoption. To analyze a possible dependency between position in the supply-chain and adoption a two-by-three table: Organizational status by Adoption/ 95
Choudhury (1997) has however a number of reservations related to the issue of disintermediation. Firstly, the possibility of improvement of the competitive position for wholesalers due to enhanced service inherent in an electronic market developed by the wholesaler may be an argument against disintermediation. Secondly, the prospect of establishment of an electronic monopoly direct with the manufacturer for products with uncertainty of demand and high variability among manufacturers may justify the existence of wholesalers. 96 Kurnia and Johnston (2000) addressed the issue in relation to Efficient Consumer Response (ECR) adoption. Based on case studies they concluded that industry structures influence adoption of ECR. They did, however, not discuss the likelihood of adoption based on position in the supply-chain.
274
Planning/ Non-adoption was analyzed. Organizational status is based on the membership of the business association of the responder. Wholesalers and manufactures had received the questionnaire from their respective business associations. In Table 8-11 (page 385) the chi-square statistics on position in supplychain and adoption level is equal to 27.4495 and the corresponding p-value is 0.0001. This is statistically highly significant indicating a strong dependency between position in the value-chain and level of EDI adoption. The table shows that approximately two-third (67.38 percent) of the manufactures are non-adopters. In comparison the share of wholesalers that are non-adopters is less than one-third (29.41 percent). Almost twice as many wholesalers, 26.47 percent compared to manufactures, 14.89 have adopted EDI. A complementary analysis was done to estimate the Relative Risks (RR)97 of non-adoption among manufactures given company size. The purpose of this RR analysis was for manufactures to estimate the probability of being a non-adopter. Figure 8-1. Relative Risk of being a non-adopter amongst all manufacturers with respect to company size 4,5 4,0 3,5 3,0 2,5 2,0 1,5 1,0 0,5 1-5
6-9
10-19
97
20-49
50+
Relative Risk is an association measure which estimates the probability of an occurrence of a given incident for an exposed group compared to a non-exposed group (Cody and Smith 1997). The calculated RRs shown in Figure 8-1 and Table 8-12 are based on the assumption that manufactures are exposed to non-adoption.
275
The estimated RRs for manufactures are illustrated in Figure 8-1 The RRs for all company sizes are greater than 1 and all the RRs are closer to the lower 95 percent confidence limits. This indicates that the chance of being a non-adopter is considerably larger for manufacturers than for wholesalers. This is especially true for larger companies. For manufactures with fifty or more employees the RR is 4.1176.
Table 8-12. Relative Risk of being a non-adopter for manufactures versus wholesalers Company size
1-5
6-9
10-19
20-49
50+
Relative Risk 95% Confidence Limits, lower 95% Confidence Limits, upper
1.9167 0.4782
1.0200 0.6821
1.5238 0.6611
1.4286 0.8628
4.1176 0.6425
7.6829
1.5254
3.5125
2.3653
26.3885
Proposition 3, which stated that being a wholesaler is positive related to EDI adoption, can therefore not be rejected based on the findings in the cross-tabulation in Table and the above-mentioned relative risk analysis. 8.2.5 Multivariate analysis of adoption, company size, dependency, and sector It is, as previously mentioned, commonly accepted that company size is positively related to adoption of innovations. As demonstrated in the analysis of proposition 1 it was found that company size based on the number of employees was positively related to EDI adoption amongst the businesses in the Danish steel and machinery industry. In the analysis of propositions 2 and 3 it was found that company size was dependent on variables related to legal status of the company and position in the valuechain. Therefore, a multivariate analysis of the four variables Adoption (yes/ no), Employees/ Company size (six levels), Dependency (yes/ no), and Sector (manufacturer/ wholesaler) was performed with the aim of examining possible hidden relationships between the four variables. The two-way analysis shown in Table 8-13 (page 386) illustrates the pairwise marginal dependencies between variables using a 5 percent level of
276
significance. The search for possible hidden associations is presented in Table 8-14 (page 386). A hidden association is an association where two variables are marginally independent, but are conditionally dependent given a third variable marginally associated with each of the variables of interest. Table 8-14 shows that there were no hidden associations between the variables of interest. The final model in Table 8-15 (page 386) which makes allowance for hidden interactions and conditional independencies is in this particular case identical to Table 8-13 and Table 8-14. The four-way analysis, as illustrated in Figure 8-2, suggests that the size of the company is the primary factor for EDI adoption in the Danish steel and machinery industry. Figure 8-2. Dependency graph illustrating EDI adoption and the three variables related to propositions 1, 2, and 3
SEC
ADO
DEP
EMP
The two rectangles in Figure 8-2 containing respectively Sector (SEC), Depend (DEP), Employees (EMP), and Adoption (ADO) is a graphical illustration of a one-way causal structure between the variables in the right and the left rectangle. Dependency between two variables is illustrated with a connecting line. If there is no line between two variables the respective variables are independent given the rest of the variables. The dependency graph shows that the only causal connection is between company size and adoption or non-adoption (ADO). Furthermore, the dependence graph illustrates that Adoption and Sector are independent given the variables Depend and Employees. Additionally, the dependency
277
graph shows that Adoption and Depend are independent given Sector and Employees. To sum up, the three propositions examined in this section could not be rejected. The multivariate analysis suggests that the only causal relationship is between company size and adoption. There is no direct interaction between the variables related to position in supply-chain, legal status and adoption. This once again suggests that company size is a major determinant for adoption of technological innovations like EDI. The three propositions examined above have all been dealing with primary attributes, hard facts, that is attributes of the organizations, which are based on clearly defined characteristics such as number of employees and legal ownership. The following section presents analysis performed on opinion data gathered from self-administered questionnaires distributed to businesses in the Danish steel and machinery industry.
8.3 Testing the Tornatzky & Fleischer model for adoption A plausible interpretation of the attitudes and opinion data of the TDP case is that it was not only technical obstacles that prevented the companies from adopting EDI. Though it was not possible directly to integrate the developed EDI software with the companies’ administrative systems the non-users of EDI could still try out the TDP-software if they were eager to do so. Another insight gained from the interviews indicated that pressure played some role in the steel and machinery industry in Denmark. The nonadopters expressed willingness to adopt EDI should they be forced to do so by their business partners. However, they had so far not experienced pressure neither persuasive nor coercive, even though some of their major business partners were already using EDI. The adopters that were encouraged to adopt EDI had experienced pressure from their parent company or other partners in their industry group whereas an overall environmental - persuasive or coercive - pressure had been absent according to the informants. These issues related to pressure were among others operationalized resulting in the fifteen propositions presented in 278
Chapter 7. The fifteen opinion data items were as described in Chapter 7 classified according to the Tornatzky and Fleischer (1990) model for organizational adoption of innovations. As previously mentioned the present research is practice-driven rather than theory-driven (Zmud, 1998). In a search for the most suitable theoretical model that would fit the opinion data items described, the Tornatzky and Fleischer model for adoption-decisions was considered to be the best option. As mentioned in Chapter 7 the strength of the Tornatzky and Fleischer model is that it allows for comprehensive operationalization of the three explanatory contexts: Organizational context, environmental context, and technological context. 8.3.1 Objective of the statistical analysis of motivators for the adoptiondecision The main objective of the statistical analysis is to reveal decision motivators and patterns per se; that is to say independent of the relationships analyzed in propositions 1 to 3. Therefore, the following two main objectives are: 1. To uncover the patent priorities of the responders with respect to adoption of EDI. 2. To estimate the relative strengths of the responder’s patent priorities. The patent priorities of the responders refer to motivators related to the three different contexts: The organizational, the environmental, and the technological. In this respect patent priority refers to the manifest, face value expressions of the responders. This is a different approach than searching for latent factors or structures as normally done in factor analytic techniques. The objective of estimating the relative strengths of the patent priorities is to clearly identify which items the responders give more importance. Odds ratios are used to measure the relative strengths of the items analyzed. Odds ratios are the comparison of the probability of an event to the probability of the event not happening (Hair et al., 1998).
279
8.3.2 Overview of procedures applied in the analysis The presentation of the procedure of the analysis starts with a description of measures. Next, Cronbach coefficient alpha is estimated. Cronbach coefficient alpha is used as a measure of reliability in relation to the operationalization of the constructs from the Tornatzky and Fleischer model (1990). The objective of these reliability estimations is to investigate the strength of the three groupings of items related to organizational context, environmental context, and technological context given adoption level. The next step in the analysis is a calculation of Spearman rank correlation coefficients. These are calculated pair-wise for all the fifteen items involved in the analysis. Three tables are included reflecting the three levels of adoption: Adopters, planners, and non-adopters. The purpose of these tables is to include the descriptive statistics of data and to present the correlation between all items. Additionally, it is a useful tool for examination of possible correlations between the opinion data items within the three contexts of the adoption-decision motivators. The next two steps in the quantitative analysis of data are exploratory searches for items related to the three adoption levels. Two steps are used to identify the factors motivating or de-motivating adoption of EDI. Fischer’s Exact two-sided test on two-by-two tables is used to identify those items that are strongly related to each one of the three levels of adoption. Fischer’s Exact test is chosen because many cell-counts in the two-by-two tables are rather small. Two-sided tests are chosen since there is no implied direction for the alternative hypothesis concerning the relationship between the item analyzed and the respective adoption level. Data is then analyzed using graphical models in DIGRAM. This is done, because it is found prudent to analyze the relationships between all the items taken together and the respective levels of adoption. Contrary to the two-way analysis using Fischer’s Exact test the analysis performed in DIGRAM is a multivariate analysis technique. One of the strengths of graphical modeling is the opportunity to analyze causal structures. In this particular case the causal structure between the items and a given level of adoption. The purpose of Fischer’s Exact test and DIGRAM is to find items that show statistical significant relationships to specific levels of adoption.
280
Logistic regression analysis is the next step in the search for patent priorities. The independent, explanatory items for logistic regression analysis are the items that have been identified either by the two-way tables using Fischer’s Exact tests and/ or through the exploratory analysis performed in DIGRAM. The objective is to identify the opinion data items that have the best explanatory power in relation to adoption, planning, or non-adoption of EDI. Logistic regression is a useful method for finding the best fitting and most parsimonious model to describe a relationship between a binary dependent outcome (adoption (yes/ no), planning (yes/ no), or non-adoption (yes/ no)) and a set of independent explanatory variables (Hosmer and Lemeshow, 1989). Following the recommendations of Hair (1998) logistic regression analysis is performed on an analysis sample and the resulting logistic regression model is then validated on a holdout sample. A measure of validity of the logistic regression model from the analysis sample is obtained by predicting the number of cases in the holdout sample given the explanatory variables from the holdout sample, and then comparing these predicted cases to the actual number of cases in the holdout sample. The purpose of this comparison is to validate the logistic regression model. The final step in the statistical analysis of the opinion data is to examine whether the three explanatory variables from proposition 1 to 3 will in any way modify the results of the logistic regression analysis. This final step is important. If the results of the logistic regression are not changed by the inclusion of the variables from proposition 1 to 3, then these results from the logistic regression analyses are more likely to be of general interest. Then the items per se - independently of company size, legal status and position in supply-chain – may be generalized to other industries and other types of innovations. 8.3.3 Measurement Similar to other adoption studies (Moore and Benbasat, 1991) multi-item indicators were used for most of the research questions. All items related to the opinion data items concerning the motivation for adoption were measured using a seven-point Likert scale ranging from “fully agree” to
281
“strongly disagree”. Due to the limited number of responders it was necessary to reduce the seven-point scales to binary scales. These binary scales were constructed to reflect agreement and disagreement with the adoption item in question. The Likert scale points 1 to 3 defined agreements whereas the points 4 to 7 defined disagreements. Point four on the seven-point Likert scale in question is considered to indicate a neutral response. It was decided to include the neutral response with points 5, 6, and 7, because a response value of 1, 2, and 3 reflects definite agreement with the item, whereas response values 4 to 7 indicate neutral to nonagreement. An alternative definition of agreement versus disagreement was considered. Exclusion of the neutral value, the fourth point on the seven point Likertscale, could be an option. The opinion data items were analyzed and it was found that this strategy would seriously reduce the number of cases available for further analysis. Therefore, this idea was rejected. Another strategy, which was considered, was to randomly assign the neutral responses to either the agreement group or the disagreement group. This strategy was feasible, however, the strategy of dividing the seven-point Likert scale in two groups: Level 1 to 3 indicating agreement and level 4 to 7 indicating disagreement had two important advantages. Firstly, there would be no reduction in the number of cases for analysis. Secondly, and perhaps even more important this method would provide a more conservative estimate of the agreement/ disagreement ratio. Based on the results from the analyses of propositions 1 to 3 it is found prudent to differentiate between adopters, planners, and non-adopters in the further analysis. Adopters and planners were asked the very same questions. For non-adopters all questions were phrased negatively.98 In the subsequent analysis of data all opinion data item values were “reversed” for non-adopters.
98
An example of negative phrasing is: “The technical solutions have not reached a satisfactory technical level” versus the positive phrasing for adopters: “The technical solutions have reached a satisfactory technical level”. (See questions 16 and 23 in the questionnaire in Appendix A, page 368).
282
8.3.4 Reliability Generally, reliability concerns the extent to which an experiment, test or any measuring procedure yields the same results on repeated trials (Carmines and Zeller, 1979). In research involving questionnaire data the internal consistency indices of reliability are useful (Hatcher, 1994). One criticism of adoption and diffusion research is that testing the reliability of measurement instruments has only been done sporadically (Moore and Benbasat, 1991). One of the most widely-used indices of internal consistency reliability is Cronbach’s coefficient alpha (Carmines and Zeller, 1979; Cronbach, 1951). The Cronbach’s coefficient alpha ranges from 0 to 1. It is a measure of reliability not a statistical test (Carmines and Zeller, 1979). Generally, one of the first tasks when conducting questionnaire research is to assess the reliability of the constructs. If the scales are not reliable there is no point in performing additional analyses. An application of Cronbach’s coefficient alpha is based on certain scale-assumptions concerning the items involved, since different measures of variances are involved in the calculation of Cronbach’s coefficient alpha (Hatcher, 1994). The use of Cronbach’s coefficient alpha is therefore a violation of the underlying assumption in regard to scales put forward by the author. A compromise was however made, due to the importance of assessing the reliability of the three constructs. The statistical analysis in all other situations in this study is based on the conviction that Likert-scales in general can be considered to be no more than ordinal scales (cf. Section 2.5.10). To get a feel for the measure of reliability the estimation of Cronbach’s coefficient alpha has none the less been included. The Cronbach’s coefficient alpha shown in Tables 8-16, page 387; 8-17, page 387; 8-18, page 388 should be viewed and interpreted keeping the above-mentioned reservations in mind. Nine Cronbach’s coefficient alphas were calculated. An analysis of Cronbach’s coefficient alphas was performed based on the operationalization of the three constructs: Organizational context, environmental context, and technological context given adoption status. Three tables (Table 8-16. Internal consistency indices of reliability for the
283
constructs: Organizational context, environmental context, and technological context for adopters, page 387; Table 8-17. Internal consistency indices of reliability for the constructs: Organizational context, environmental context, and technological context for planners, page 387; and; Table 8-18. Internal consistency indices of reliability for the constructs: Organizational context, environmental context, and technological context for non-adopters, page 388) were used to calculate Cronbach’s coefficient alpha. Table 8-19. Summary of the values of Cronbach’s coefficient alpha test for scale reliability Construct Organizational context Environmental context Technological context
Adopters
Planners
0.76 0.70 0.52
Non-adopters 0.72 0.82 0.74 0.82 0.34 0.49
Cronbach’s coefficient alpha was used to assess scale reliability. For organizational context, environmental context, and technological context the Cronbach’s coefficient alphas ranged respectively from 0.72 to 0.82, 0.70 to 0.82, and 0.34 to 0.52. Generally, the lower acceptable limit for summed scales is considered to be 0.70 (Nunnally, 1978). Keeping the above-mentioned reservations in mind regarding the scales, the constructs for organizational context and environmental context are of an acceptable reliability level independent of adoption status. On the other hand, the operationalization of technological context is below the generally acceptable reliability level independent of adoption status. 8.3.5 Descriptive statistics of the fifteen items included in the analysis The marginal distributions of all opinion data items motivating adoption of EDI in the Danish steel and machinery industry are presented in the following three tables: 8-20, page 389; 8-21, page 390; 8-22, page 391. These tables list the percentage distributions of all the individual opinion data items together with all these opinion data items’ total number of cases (column N). The mode, the most frequent value, for each opinion data item is shown in boldface. The marginal distributions of the opinion data items presented in these tables demonstrate a marked skewness in a great number
284
of these items. Several items (e.g. ITEMs 3 and 4 in Table 8-20) show positive skewness since many responders strongly agree with the corresponding propositions. The Spearman correlation coefficient for adopters, planners, and nonadopters are presented in three tables (Table 8-23. Spearman correlation coefficient for adopters, page 392; Table 8-24. Spearman correlation coefficient for planners, page 393; and Table 8-25. Spearman correlation coefficient for non-adopters, page 394). All correlation values greater than 0.500 and with a significant level less than 1 percent are shown in boldface in these tables. Table 8-26. Summary of adoption-decision motivators based on Spearman correlations Discordant Total Concordant Adoption sum TT OO EE Sum Level Adopters 2 4 2 8 3 11 Planners 1 2 1 4 1 5 Non-adopters 1 5 7 13 9 22 Totals 4 11 10 25 13 38 Legend: TT = Correlation of technological context items OO = Correlation of organizational context items EE = Correlation of environmental context items
Table 8-26 lists the number of concordances and discordances for adopters, planners, and non-adopters. The concordances are subdivided into three levels reflecting the correlation within the three theoretical constructs for the adoption-decision. For example, among adopters there are eight concordant Spearman Rank correlations. Among these eight there are two correlations related to the technological attributes, four related to the organizational context and two related to the environmental context. Three of the correlations were between different constructs (discordant) for example ITEM08 and ITEM01 representing respectively the organizational context and the technological context. The ratio of concordances to discordances based on the total numbers in Table 8-26 is 25/ 38, or 65,8 percent. Even though this estimation is independent of the level of adoption it suggests that two out of three items 285
within the three constructs are significantly correlated with Spearman Rank correlations greater than 0.500. Based solely on the Spearman Rank correlations the three adoption motivation constructs appear to be meaningful and useful. A test for independence of the variables (Table 8-27. Adoption level versus concordance/ discordance , page 395), Adoption level and Concordance/ discordance, was performed. The table probability of Fischer’s Exact test was 0.0758 and the corresponding p-value was 0.7012. This suggests that Adoption level and Concordance/ discordance are independent. So far the fifteen opinion data items related to adoption motivators have been examined to assess whether these items in a meaningful way can be subdivided into the three theory driven constructs: Organizational context, environmental context, and technological context. Based on the Cronbach’s coefficient alpha the organizational and the environmental contexts appeared to be reliable. This was not the case the technological context. The purpose of following sections is to uncover the items with the best explanatory power in relation to the three levels of adoption. 8.3.6 Exploratory two-way analysis of opinion data items statistically related to the three adoption levels Fischer’s Exact tests were performed on all the original opinion data items transformed to binary variables (cf. Section 8.3.3) related to the three adoption levels: Adopters, planners, and non-adopters. As mentioned in Section 8.3.1 the objective of the Fischer’s Exact test was to identify those items that were related to each of the three levels of adoption. Table 8-28, page 396; Table 8-29, page 396; and Table 8-30, page 397 illustrate respectively: The item number (ITEM #), adoption-decision motivator (TYPE), and a third column labeled Fischer’s Exact. The column labeled Fischer’s Exact has two sub-columns presenting the p-value of each pairwise test and a graphical illustration of the level of the statistical significance.
286
Based on the results from Fischer’s Exact test in Table 8-28 (page 396) it was found that six items are strongly related to adoption. Four of the six items are organizational context items, one is a technological context item, and one is an environmental context item. Table 8-29 (page 396) shows that five items are related to companies that are considering adopting EDI. Of these five items two are related to the technological context and three items are related to the environmental context. Table 8-30 (page 397) shows that nine items are related to the responders classified as non-adopters. Out of these nine items, four items are related to the organizational context, three are related to the environmental context, and two items are related to the technological context. The relationships examined in Tables 8-28; 8-29; and 8-30 for the three adoption levels are all based on analysis of two-by-two tables. These relationships are all two-dimensional. However, to uncover influences from multi-way dimensions and to identify possible causal structures these same opinion data items were included in a multivariable analysis based on graphical models. 8.3.7 Exploratory multivariate analyses of opinion data items statistically related to the three adoption levels The main objective of the exploratory multivariate analyses was to analyze the relationships between opinion data and the levels of adoption. This was done, because it was found relevant to analyze the relationships between all items taken together versus the three levels of adoption. Three tables are included to report the outcome of the exploratory multivariate analysis: Table 8-31. Exploratory multivariate analysis for adopters, page 398; Table 8-32. Exploratory multivariate analysis for planners, page 400, and Table 8-33. Exploratory multivariate analysis for non-adopters, page 402. The graphical models as used here are based on the assumption that the opinion data items have a causal influence on the adoption level; not the other way around. This strategy can be criticized since the assessment of
287
the opinion data item depends on the adoption level. It has been suggested that adopters have a more positive perception of an innovation than nonadopters (Moore and Benbasat, 1991) leading to an inherent bias. Generally, adopters evaluate opinion data items retrospectively, where as non-adopters provide answers to the very same items prospectively. In contrast to this planners are at best providing contemporary answers. In this respect planners are not biased by their own personal experiences of EDI, which may be the case for adopters. Furthermore, planners are responding to questions that they most likely perceive as being relevant. This may not be the case for non-adopters, since they may not even consider EDI a relevant option. In order to illustrate the relationships between opinion data and the levels of adoption three graphs are included in Appendix B: Figure 8-3. Dependence graph from an exploratory, initial screening of adopters and EDI adoption motivators, page 399, Figure 8-4. Dependence graph from an explanatory, initial screening of planners and EDI adoption motivators, page 401, and Figure 8-5. Dependence graph from an exploratory, initial screening of non-adopters and EDI adoption motivators, page 403. The exploratory multivariate analysis suggests that the only opinion data items that have a causal relationship with adoption were items: C, E, J, and K99 (cf. Figure 8-3, page 399). For planners (cf. Figure 8-4, page 401) the causal relationships were opinion data items: J, N, and O. Finally, for nonadopters (cf. Figure 8-5, page 403) the opinion data items were: C, D, J, K, and O. 8.3.8 Selection of opinion data items for logistic regression analysis The opinion data items identified using Fischer Exact test on two-by-two tables and the causal relationships identified using multi-way analysis performed in DIGRAM are summarized in Table 8-34.
288
Table 8-34. Listing of items for inclusion in logistic regression analysis Item/ Name
Proposition
Fischer’s exact ADO
PLA
DIGRAM NON
ADO
PLA
Items for inclusion NON
ADO
PLA
NON
1/A T1 * * 2/B T2 ** ** + + 3/C O1 *** *** ← + + ← 4/D O5 ** *** + + ← 5/E O2 + ← 6/F O3 7/G O4 ** *** + + 8/H O6 *** *** + + 9/I E1 *** *** + + + + + 10/J E2 *** *** ← ← ← 11/K E3 *** *** ← + + ← 12/L E5 13/M T3 *** + + 14/N T4 ← 15/O E4 ** + + ← ← Legend: * = p ChiSq
Point Estimate
95% Wald Confidence Limits
- 2.2777 - 1.1409
0.0011 0.0147
0.102
0.016
0.639
- 1.4475
0.0026
0.055
0.008
0.363
- 1.9272
0.0015
0.021
0.002
0.229
Response profile Ordered value Non-adopters 1 Yes 2 No
Total frequency 27 45
Hosmer and Lemeshow Goodness-of-Fit Test Chi-Square DF PR > ChiSq 1.5645 4 0.8152
Since the Hosmer and Lemeshow Goodness-of-Fit Test p-value (0.8152) is greater than 5 percent this supports the fit of the model. 8.3.11 Classification of events and non-events for adopters, planners, nonadopters 8.3.11.1 Definitions used in the classification tables The following classification tables (Table 8-39, page 405; Table 8-40, page 406; Table 8-41, page 406) present the sensitivity, specificity, false positive rate, and false negative rate for various levels of probability of the event in 293
question. The objective of these four measures is to illustrate the predictive ability of the model given a specific probability level. Table 8-38. Combinations of predicted and actual cases Actual Yes A True positive No C False negative A+C Adapted from: Foldspang et al., 1992. Predicted Yes
No B False positive D True negative B+D
A+B C+D N
Sensitivity is defined as: A/ (A+C), that is the number of cases of true positive divided by the sum of the number of cases of true positive and false negative. Specificity is defined as: D/ (D+B), that is the number of cases of true negative divided by the sum of the number of cases of true negative and false positive. An optimal and ideal model would have a sensitivity equal to 1 (100 percent) and a specificity equal to 1 (100 percent) since there would be no cases of false negative and no cases of false positive (Foldspang et al., 1992). The closer the sensitivity and specificity is to 1 the better is the predictive ability of the model. Whether the focus should be on sensitivity or specificity or both depends on the actual situation and its consequences. Also the level of sensitivity and specificity must necessarily depend on the actual situation and its consequences. For example, to estimate the sensitivity for a responder being an adopter with a probability level greater than 30 percent the classification table (Table 8-39, page 405) shows that the corresponding sensitivity is 84.2 percent. This means that the model predicts correctly in 84.2 percent of the cases with responders stating they are adopters. The probability levels can for example for adopters be found in Table 8-42 (Table 8-42. Predicted number of cases and actual number of cases for adopters, page 407). In order to concretize the information contained in the classification tables (cf. Table 8-39, page 405; Table 8-40, page 406; Table 8-41, page 406)
294
some comments related to the values of false positive and false negative are presented in the following. The false positive rate is the ratio of the wrongly predicted positive cases divided by the total number of positive predicted cases. An example, for adopters indicating ‘yes’ to adoption this means that it is the ratio of the models’ wrongly predicted cases divided by the total number of predicted cases, “Non-adopters are wrongly being classified as adopters by the model”. Respectively the false negative rate it is the ratio of the wrongly predicted negative cases divided by the total number of negative predicted cases, “Adopters are wrongly being classified as non-adopters by the model”. To illustrate the strength of the classification models for adopters, planners, and non-adopters a probability-level of 30 percent is chosen. For adopters the false positive rate is relatively high, namely 52.9 percent. The corresponding false negative rate is low, only 7.0 percent (cf. Table 8-39, page 405). This shows that the model at this particular probability level only predicts correctly for adopters (= ‘yes’) in about half the cases. At this probability level the model only has a 7 percent false negative rate. This indicates the models’ predictive ability at this particular probability level better identifies the adopters that actually are saying ‘no’ to adoption of EDI. For planners and non-adopters the false positive rates are respectively 29.6 and 22.6 percent indicating that the respective models’ predictive ability is much better for planners and non-adopters at this probability level. The classification table for adopters shows that the model at all probability levels have a relatively high number of false positive ratio compared to the classification tables for planners and non-adopters suggesting that the models for planners and non-adopters have better fits. 8.3.12 Validation of results from logistic regression analyses As outlined in the overview of procedures used in the statistical analyses the models resulting from the three logistic regression analyses will be validated. The models are based on the previously defined analysis samples. In order to estimate the validity of the models, a measure of
295
validity of the logistic regression model from the analysis sample is obtained by predicting the number of cases in the holdout sample given the explanatory variables from the holdout sample, and then comparing the predicted number of cases with the actual number of cases in the holdout sample. The purpose is to validate the logistic regression model (Hair et al., 1998). The predicted number of adopters and the actual number of adopters (cf. Table 8-42, page 407) were analyzed using Fischer’s Exact Test. For the variable Adopter equal to “no” 103 the table probability (P) was 0.0133 with a corresponding probability of 0.6374 for obtaining values less than or equal to (P) (cf. Table 8-45, page 409). This suggests that the differences between predicted number of cases and the actual number of cases are not statistically significant. This demonstrates that the logistic regression analysis model for Adoption in this respect is valid. Table 8-43 (page 407) was analyzed using Fischer’s Exact Test. For the variable Planner equal “no” the table probability (P) was 0.0194 with a corresponding probability of 0.8839 for obtaining values less than or equal to (P) (cf. Table 8-46, page 410). This suggests that the differences between predicted number of cases and the actual number of cases for planners are not statistically significant. This demonstrates that the logistic regression analysis model for companies planning to adopt EDI is valid. The predicted number of Non-adopters and the actual number of Nonadopters were (cf. Table 8-44, page 408) analyzed using Fischer’s Exact Test. For the variable Non-adopter equal “no” the table probability (P) was 2.217E-04 while the corresponding probability of 0.9021 for obtaining values less than or equal to (P) (cf. Table 8-47, page 411). This suggests that the differences between predicted number of cases and actual number of cases for non-adopters are not statistically significant. This demonstrates the validity of the logistic regression analysis model for non-adopters of EDI. 103
The reason for the decision to test “no” for adopters, planners, and non-adopters instead of “yes” was that the “no”-tables included more cases.
296
8.3.13 Inclusion of variables from proposition 1 to 3 in the logistic regression analyses Table 8-49. Summary of logistic regression including variables from propositions 1 to 3 Proposition variables Depend Employee Organisa
ADO ChiSq score 0.9787 1.8904 0.2842
Pr > ChiSq 0.3325 0.3886 0.5940
PLA ChiSq score 0.0030 1.9903 0.5126
Pr > ChiSq 0.9561 0.3697 0.4740
NON ChiSq score 1.6376 3.0561 0.0282
Pr > ChiSq 0.2006 0.6913 0.8667
The variables concerning company size, position in value-chain, and legal ownership did not influence the outcome of the logistic regression analysis in relation to explanatory factors concerning EDI adoption (cf. Table 8-49). Previous research focusing on small firm adoption of IT has found that variables not directly related to the motivation for adoption e.g. size added nothing to the understanding of the decision variables related to motivation (Harrison et al., 1997). This was also the case for the present study. 8.3.14 Summary of logistic regression analyses In the summary table below the opinion data items have been replaced by their respective propositions outlined in Chapter 7. For example opinion data item AgreeI08, “The company felt well prepared for adopting EDI” is replaced by O6, “The notion that companies considering themselves to be well prepared for EDI are more likely to adopt EDI.” The reason for this change in reporting of results from the analysis of motivators for adoption of EDI is to clarify which of the contexts and thereby which propositions have been accepted in the analyses. Propositions related to the technological context were not found to influence the motivation for adoption for any of the responders regardless of the level of adoption. One reason could be that the opinion data items focusing on technology were not well-defined. As shown in the reliability test using Cronbach’s coefficient alpha the construct of the technological context was not well-defined for any of the three adoption levels.
297
Table 8-50. Summary of logistic regression analysis for adopters, planners, and nonadopters showing odds ratios for significant decision variables Code O6
E2 E3 E4
Proposition The notion that companies considering themselves to be well prepared for EDI are more likely to adopt EDI The prospect of increasing the company’s market share is a motivator for EDI adoption The knowledge that several business partners already use EDI is a motivator for EDI adoption The fact that EDI has been recommended by others is a motivator for EDI adoption
Context ORG
ADO PLA 3.292
ENV ENV ENV
11.019 10.744
NON 0.102
0.055 0.021
0.087
For adopters and non-adopters factors related to the organizational context and environmental context were found to explain motivation for EDI adoption or non-adoption. For planners opinion data related to the environmental context were found to explain the motivation for EDI adoption. In the following section a closer look at the significance of each of the explanatory opinion data items in relation to the adoption level is presented.
8.4 Quantitative assessment of the second research question In Section 8.3 it was stated that the main objective of the quantitative assessment of data from the businesses in the Danish steel and machinery industry was two-fold: 1. To uncover the patent priorities of the responders with respect to adoption of EDI. 2. To estimate the relative strengths of the responder’s patent priorities. Fifteen propositions were defined in Chapter 7. The propositions were based on opinion data items related to a number of statements, which based on practical and theoretical sources were found to influence motivation for EDI adoption. These fifteen statements were analyzed using Fischers’ Exact test and an exploratory analysis performed in DIGRAM. These two 298
analyses led to a selection of a number of patent priorities (cf. Table 8-34, page 289). These priorities were included in logistic regression analyses. This resulted in a substantial reduction of the number of explanatory variables or patent priorities. As a result of uncovering the patent priorities or explanatory variables it is now possible to provide an answer the second research question, “To which degree can the motivation to adopt interorganizational information systems, exemplified by EDI, be explained by issues related to the organizational context, the environmental context, and the technological context?” Analyses of two-way tables indicated that for adopters of EDI the propositions T1, T2, T4, O2, O3, E1, E2, E4, and E5 provided in a statistical sense no information about the states, “yes” or “no”, of the dependent variable Adopter (cf. Table 8-35, page 291). For planners of EDI the propositions O1, O2, O3, O4, O5, O6, E3, E5, T3, and T4 (cf. Table 836, page 292) provided no information about the states, “yes” or “no”, of the dependent variable Planner. Similarly, for non-adopters the propositions with no information about the states, “yes” or “no”, of the dependent variable Non-adopter were O2, O3, E4, E5, T3, and T4 (cf. Table 8-37, page 293). Even though all the statistically significant binary variables (cf. Table 8-34, page 289) from the variable-set {AgreeI01: AgreeI15}, which were derived from the original opinion data items, were included in the respective logistic regressions analyses. The results indicated that only four opinion data items out of the total number of fifteen opinion data items provided information on the states of adopters, planners, and non-adopters. These four opinion data items alone could be considered to be motivators for either adoption of EDI, planning to adopt EDI, or remaining as a nonadopter of EDI. It was surprising that these four opinion data items alone could explain the outcome of the three levels of adoption, and that the other eleven items in this respect provided no additional information. It is remarkable that none of the opinion data items from the technological context were of any importance compared to opinion data items from the
299
two other contexts in the regression analyses. One explanation could be that most of the responders were managers. Managers or managerial staff are naturally more concerned with business issues than with concrete technical issues and the necessity of acquiring new innovations. From a theoretical point of view it can be argued that managers are not exposed to technical complexities and problems, as for example illustrated in the review of EDI in the top-five MIS journals were technical obstacles and complexities related to EDI were seldom mentioned. In the following three sections the patent priorities in relation to the fifteen propositions are outlined for adopters, planners, and non-adopters respectively. Based on the results from the logistic regression analysis for adopters it is possible to estimate the relative strength of the propositions. 8.4.1 Factors motivating adoption Two factors were found to motivate EDI adoption. These two factors’ corresponding propositions O6, “The notion that companies considering themselves to be well prepared for EDI are more likely to adopt EDI” and E3, “The knowledge that several business partners already use EDI is a motivator for EDI adoption” could not be rejected. If a given responder answers that both statements are of minor importance (points 4 to 7 on the original seven points Likert-scales) then the probability of the responder being an adopter is about 22 percent. If a responder on the other hand answers that both statements are of major importance (points 1 to 3 on the original seven points Likert-scales) then this responder’s probability of being an adopter is about 91 percent (cf. Table 8-42, page 407). A responder answering “yes-yes” to O6 and E3 is more than four times as likely to become an adopter than a responder answering “no-no” to O6 and E3. If a responder indicates that O6 is of major importance and that E3 is of minor importance then the corresponding probability of being an adopter is about 49 percent. Finally, if a given responder answers that E3 is of major importance and O6 of minor importance then this responders probability of being an EDI adopter
300
is approximately 76 percent. So the amount of information provided by the two constructs, O6 and E3, is quite considerable. 8.4.1.1 Interpretation of motivators leading to adoption In the statistical analyses of the two-way tables of the opinion data items, corresponding to the organizational context, versus adoption, the opinion data item corresponding to O6 was the most significant. Similarly, the most significant of the environmental context opinion data items was E3. Therefore, one could argue that proposition O6, “The notion that companies considering themselves to be well prepared for EDI are more likely to adopt EDI“ from a managerial point of view comprises all the organizational context opinion items. That the company states, that it is well prepared for EDI adoption implies that the remaining organizational context items in some way are contained within proposition O6. Another interpretation of the outcome of the logistic regression for adopters could find support in the nature of the social system (Rogers, 1995). If the prevailing attitude in the social system is, that EDI adoption is the norm, then companies are likely to perceive themselves to be ready for adoption. The common problem, which according to Rogers, is one of the characteristics of the nature of the social system, is then related to efficiency. The mutual goal therefore is to improve efficiency through EDI adoption thereby creating interorganizational efficiency and network externalities. Next, the importance of organizational readiness could be a result of the influence from change agents’ promotional efforts, which through campaigns have informed about the innovation. The importance of proposition O6 is according to this interpretation influenced by social processes and communicated information about the innovation. If this interpretation is accepted, then the knowledge that several business partners already use EDI, proposition E3, is supporting the notion of social process attitude towards adoption even more strongly. Amongst the environmental context opinion data items proposition E3,” The knowledge that several business partners already use EDI is a motivator for EDI adoption” appears to be the most important statement.
301
The awareness that business partners already use EDI induces the potential adopters to perceive adoption to be the norm. Another interpretation of the importance of proposition E3 for adopters can be supported by the exponential diffusion curve (Attewell, 1992). Adoption according to this view becomes more and more attractive the more people adopt the innovation. This is especially the case when interorganizational attributes are related to the innovation. Here critical mass is important for benefits to accrue from the investment (Markus, 1987). One aspect, which is important to consider, when interpreting the preferences indicated by the adopters, is that their responses reflect an ex post evaluation. The two propositions, O6 and E3, that were found to be statistical significant in the logistic regression analysis are more abstract and less specific and therefore of a more general nature than to the rest of the propositions comprising the organizational context and the environmental context. Instead of specifically replying that the motivation for adoption was for example related to improvement of work environment or competitiveness the motivation is expressed in more general terms. 8.4.2 Factors motivating companies planning to adopt EDI Two factors were found to motivate companies considering adopting EDI. So their corresponding propositions E2, “The prospect of increasing the company’s market share is a motivator for EDI adoption” and E4, “The fact that EDI has been recommended by others is a motivator for EDI adoption” could not be rejected. If a given responder answers that both statements are of minor importance (points 4 to 7 on the original seven points Likert-scales) then the probability of this responder being a planner is about 18 percent. If the responder on the other hand answers that both statements are of major importance (points 1 to 3 on the original seven points Likert-scales) then the probability of being a planner is about 17 percent (cf. Table 8-43, page 407). If a responder indicates that E2 is of major importance and that E4 is of minor importance then the corresponding probability of being a planner is about 70 percent. Finally, if a given responder answers that E2 is of
302
minor importance and E4 of major importance then this responder’s probability of being an EDI planner is approximately 2 percent. The largest ratio for planner is a situation where a responder answers “yes-no” relative to a responder answering “no-yes” to E2 and E4. This ratio is about 35 (70.36/ 1.83). 8.4.2.1 Interpretation of motivators influencing planners For planners the factors motivating EDI adoption are solely related to the environmental context. The two opinion data items identifying companies that plan to adopt EDI were related to the prospect of increasing the company’s market share and recommendations from others. Here it should be noted that planners do not consider recommendations from others to be of any importance (odds ratio = 0.087). This would indicate that recommendations from other businesses and from business associations could not be considered to be motivators for those companies that plan to adopt EDI. This appears to be in contrast to the variables determining adoption defined by Rogers. As mentioned in relation to adopters the variables related to the nature of the social systems and change agents’ promotion efforts were used as a suitable framework for understanding why these particular propositions were of relevance for adopters. One interpretation is that rationality rather than social processes drive the motivation for EDI adoption amongst the responders that indicated that they plan to adopt EDI. One reason could be that planners compared to adopters are indicating contemporary adoption preferences contrary to the adopters who expressed an ex post evaluation of their motivation for adoption. Planners in contrast to adopters indicate less abstract and more concrete motivation preferences. The ratio of proposition E2 to E4 was 35. This suggests that planners independent of recommendations from change agents and norms in the social system consider adoption of EDI to improve the organizations’ strategic performance, thereby leading to increased market shares. The same incentive was expressed by some of the TDP participants. They explained that one of their motivations for adopting EDI had been to
303
become attractive business partners. Expectations related to savings and improved efficiency did not play an important role. The information on EDI communicated from change agents has to a large extent been focused on savings and efficiency. As mentioned in Chapter 3, EDI was in the policy statements presented as a means for improving business performance. The TDP initiators also emphasized the strategic and operational benefits accruing from EDI. However, recommendations are not given any importance by planners, when they state their preferences in relation to EDI adoption. Their attitude seem to be to adopt EDI regardless of how EDI is presented. The same pattern was observed in relation to the one planner in the TDP. This responder expressed that the company wanted to adopt EDI to avoid loosing business opportunities regardless of investments involved and the less than flattering light EDI in the TDP setting was presented in. 8.4.3 Factors causing a non-adopting attitude towards EDI Three factors were found to cause a non-adopting attitude towards EDI. Their corresponding propositions O6, “The notion that companies considering themselves to be well prepared for EDI are more likely to adopt EDI” E2, “The prospect of increasing the company’s market share is a motivator for EDI adoption”, and E3, “The knowledge that several business partners already use EDI is a motivator for EDI adoption” could not be rejected.104 If a given responder answers that the three statements are of minor importance (points 4 to 7 on the original seven points Likert-scales) then the probability of this responder being a non-adopter is about 9 percent. If the responder on the other hand answers that all three statements are of major importance (points 1 to 3 on the original seven points Likert-scales) then the probability of being a non-adopter is about 0 percent (cf. Table 844, page 408). If a responder indicates that O6 is of major importance and that E2 and E3 are of minor importance then the corresponding probability of being a non-adopter is about 1 percent. If a responder indicates that E2 is 104
In this respect it is important to remember that the questionnaire items for nonadopters were all phrased negatively (cf. Footnote 98, page 282).
304
of major importance and that O6 and E3 are of minor importance then the corresponding probability of being a non-adopter is about 0.2 percent. If a responder indicates that E3 is of major importance and that O6 and E2 are of minor importance then the corresponding probability of being a nonadopter is about 0.6 percent. For non-adopters the range of probabilities is very narrow from 0 percent to about 9 percent. Only one of the predicted probabilities was greater than one percent. This indicates that none of the opinion data items were of great significance. In other words, the non-adopters did not agree with any of the statements included in the questionnaire. The three propositions of importance for non-adopters, which were identified in the logistic regression analysis, do however provide the best profile of non-adopters of EDI in the Danish steel and machinery industry. 8.4.3.1 Interpretation of reasons for remaining a non-adopter The environmental context seemed to be the dominant explanatory factor for responders remaining as non-adopters. One opinion data item related to the organizational context was however also found to be a significant explanatory factor for non-adopters. Proposition O6, which concerned organizational readiness for adoption was found to be of major importance for adopters. Non-adopters on the other hand stated that they did not consider organizational readiness to be of any importance with respect to EDI adoption (odds ratio = 0.102). A similar pattern was found in relation to proposition E3, the awareness that several business partners already use EDI. This opinion data item was of major importance for adopters, whereas it had no relevance for the non-adopters at all (odds ratio = 0.021). However, there is some logical explanation for this inconsistency of preferences amongst the two levels of adoption – what makes good sense for adopters and planners, does not appear to make sense for non-adopters. This confirms the strength of the explanatory variables, O6 and E3, determining adoption/ non-adoption, and the explanatory variable E2, which was found to be of major importance for planners but of no relevance for non-adopters. Common for all opinion data items for nonadopters was that they did not agree with any of these statements. The
305
usefulness of the results from the logistic regression analysis is therefore that it enables prediction of the probability of a given responder being a non-adopter based on her or his response profile. A rational explanation for the situation, that non-adopters do not find EDI attractive at all, might be found in the attributes of the non-adopting companies. It was found in the analysis of propositions 1 and 3 that especially small companies and independent companies were amongst the non-adopters (cf. Sections 0 and 8.2.4). These companies have limited power to initiate an EDI partnership and they are most likely allotted the role of an EDI follower. It is generally found that followers do not derive the same benefits of EDI as initiators (Swatman and Swatman, 1992). Operational and strategic gains from EDI adoption for small companies might therefore be limited. This is also the case in relation to the five innovation attributes defined by Rogers. The relative advantage of EDI for small adopters is limited in relation to the efforts required to set-up an EDI solution with a few business partners. In this context it should be noted that the two non-adopters in the TDP expressed, that they expected a future EDI solution to be fully integrated in the administrative systems of the organization. If the non-adopters in the survey have the same expectation as the two companies in the TDP then it demands considerable resources to adopt and implement EDI in the organization. It is therefore reasonable to conclude that these companies do not consider EDI adoption to be an attractive option. It was argued that one possible reason for adopters indicating that organizational readiness was a motivator for EDI adoption was to be found in the nature of the social system and the change agents’ promotion efforts. Organizational readiness was according to the responders of no relevance for non-adopters. An interpretation of this outcome is that non-adopters did not consider themselves to be addressees of the EDI campaigns launched by change agents. Pedagogical intervention is therefore of limited value for companies that postpone or reject adoption of EDI. Additionally, the norms of the social system, which they perceive themselves to belong to, may not attach any value to EDI.
306
The two opinion data items concerning the environmental context, which resulted from the logistic regression analysis for non-adopters, were related to a possible increase of the company’s market-share due to EDI adoption and the awareness of that several business partners were using EDI. Proposition E3 was considered as the mildest form of pressure leading to EDI adoption among the fifteen opinion data items defined in Chapter 7. This external “community” pressure did not influence non-adopters. A rational interpretation could be that non-adopters could not foresee that they would reach a critical mass of business partners using EDI. An interpretation guided by social processes could be that non-adopters simply do not identify themselves with EDI adopters. Therefore, there is no basis for an imitation process. With respect to proposition E2 it could be argued that if the non-adopters had found that EDI adoption in any way would influence their market share in a positive way, then they would probably already have adopted EDI. To sum up, it looks like the non-adopters think that they can do fine without this innovation. And they do therefore not agree with or show any sign of enthusiasm with respect to any of the defined motivators for EDI adoption.
8.5 Summing up Chapter 8 The procedure of “quantifying qualitative findings” led to the conclusion that propositions 1 to 3 could not be rejected. First of all, it was found that company size positively influenced EDI adoption. Secondly, it was found that legal dependency to a parent company or an industry group positively influenced EDI adoption. Thirdly, it was found that position in the supplychain influenced EDI adoption. It was found that wholesalers compared to manufactures were more likely to adopt EDI. In the multivariate analysis it was found that the only causal relationship between the three factors and adoption was found in relation to company size. The three explanatory factors could not be rejected to influence adoption of EDI. However, when they were included in a logistic regression together with opinion data items 307
they were not found to have any statistical significant influence as motivators for adoption. In the next step of the quantitative analysis of motivators for adoption of EDI in the Danish steel industry the individual strength of fifteen opinion data items were assessed. It was found that only four of the fifteen opinion data items could be considered to be explanatory factors for the three levels of adoption: Adopters, planners, and non-adopters. Among these four factors one was related to the organizational context whereas three were related to the environmental context. The significance of propositions O6 and E3 could not be rejected in relation to the motivation for EDI adoption for adopters. For planners propositions E2 and E4 could not be rejected. Finally, three propositions were found to influence non-adoption: Propositions O6, E2, and E3. The theme of this chapter was to present the survey results and the purpose was to provide a quantitative assessment of the second research question. It was found that the motivation for adoption of EDI was related to the organizational context and the environmental context. From the managers’ point of view, the technological context on the other hand was not found to have any explanatory significance for the motivation to adopt EDI among the businesses in the Danish steel and machinery industry.
308
9 Discussion and triangulation The objective of this chapter is to present and triangulate the results from the qualitative assessment and the quantitative assessment of the adoption motivators in the Danish steel and machinery industry. Triangulation was as stated in Chapter 2 the research method applied in the present study. The “between methods” triangulation was applied in a search for motivators for EDI adoption in the Danish steel and machinery industry. The two bearings taken for the triangulation were: A case study of the TDP and a survey of the steel and machinery industry. Triangulation will be performed after presenting the findings from the two bearings. After this triangulation the implications of the findings are discussed, and finally the “Double Domain Motivation Model “ (DDMM) is presented.
9.1 Outcome of the qualitative assessment of the second research question Political agendas amongst the TDP participants seemed to play an important role for adoption and diffusion of EDI. Some participants in the project wanted to influence the overall developments in the business sector, but they were not particularly interested in supporting the adoption process. Others were in positions in their organizations where they had little or no influence on the decision processes related to adoption of EDI. Political reasons were the immediate cause for initiation of the pilot project. Adoption explanations in the political research stream comprises, “… irrational or inconsistent adoption behaviors, and outcomes can be understood only when all of the consequences of IS adoption of all the stakeholders are considered” (Kwon and Zmud, 1987). There were multiple stakeholders in the TDP. The professional business associations supported the 1996-action plan, and they were motivated to launch a pilot project that
309
met the objectives defined in the 1996-action plan. The consequences of EDI adoption for the stakeholders are however limited in the sense that adoption of EDI amongst the professional business associations’ members does not influence the relations and work routines directly in the business associations. However, the significance of the project should not be underestimated in relation to the business associations, which might feel some sort of commitment towards the implementation of the 1996-action plan. A successful outcome of the TDP would therefore give the business associations a good reputation in relation to EDI diffusion. In a broader perspective a good reputation in the Danish political environment would mean the ability to set the agenda in relation to future regulatory initiatives. This after all is one of the core tasks of business associations. Politics also played a role at the micro level. Some of the participants where particularly interested in joining the project – not because they needed to get started with EDI – rather they wanted to influence the process of defining an industry subset and to monitor the general developments of EDI in this business sector. In the overall evaluation of the benefits of the TDP the adopters expressed that they found that their time and efforts had not been wasted (cf. Table 4-7, page 137). They stated that they had the opportunity to influence the process and that the project had been useful for networking. The non-adopters mentioned that the prospect of improved performance was a motivator for adoption. This is in line with the advantages communicated by the governmental units and professional business associations involved in the promotional efforts of diffusing knowledge of EDI in the Danish business environment. The non-adopters reckoned that the software developed during the project would not improve performance in relation to operational efficiency, and there was therefore no direct incentive for EDI adoption based on the tools made available in the TDP. The non-adopter that had decided to adopt EDI in near future105 had acknowledged the strategic significance of EDI and this company expected 105
This company should strictly speaking be labeled a planner if the classification applied in the survey is also applied in the case study.
310
EDI to improve strategic performance. The adopters were mainly concerned with the strategic significance of EDI in relation to their business performance. They were willing to make the necessary investments in order to be able to exchange EDI messages with their core business partners (cf. Figure 4-5, page 125). The adopters had however realized that the improved operational performance was marginal (cf. Table 4-6, page 126). Pressure did play a role both in relation to adopters and non-adopters (cf. Table 4-4, page 123). Some of the adopters had pressured their business partners while other had been subject to pressure from their parent company or by another company in their industry group. The pressure ranged from encouragement to direct order, which can be characterized as coercive power in relation to EDI adoption (Hart and Saunders, 1998; Iacovou et al., 1995). The absence of pressure on the non-adopters had a negative effect on EDI adoption. Due to lack of requests for EDI adoption from their business partners, EDI had so far not been introduced in the organization. The objective of the second research question was to identify which factors motivate IOS adoption in a business sector dominated by small businesses.106 Based on the self-reported data from the TDP participants the answer to this research question is that the environmental context played the dominant role for adoption and non-adoption. The motivators for becoming EDI adopters were strategic considerations and pressure. The proactive adopters in the TDP expressed that they were interested in influencing the EDI developments in the Danish business environment and that they wanted to be attractive business partners. At the same time they wanted to be able to set the EDI agenda in relation to the business partners that could be forced to exchange EDI messages with them. This strategy seems to have failed so far, considering the low level of EDI usage (cf. the VIDS-test Table 4-6, page 126). The reactive adopters on the other hand were influenced by the environmental context in the sense that they were 106
It should be remembered that not all of the businesses involved in the TDP were small businesses (cf. Table 4-3, page 98).
311
told to adopt EDI by their parent company or by other businesses in their industry group. The planner in the TDP stated, that strategic considerations had driven its commitment towards EDI adoption, and that no pressure had been exerted. The company wanted to maintain its position in the market. The reason for the reluctance to adopt EDI had been a mix of a lack of organizational readiness and considerations related to the technological context, explicitly the cost of acquisition and integration to existing administrative systems in the organization. For the remaining two non-adopters absence of pressure and strategic necessity were the major reasons for their status as non-adopters. However, the two non-adopters did not deny that issues related to the technological context were of relevance for their adoption status. Cost of purchase and especially the cost of integration of EDI software were mentioned as a reservation for EDI adoption. During the interviews the non-adopters and the planner expressed some reservation with respect to the EDI technology. They said that they had to examine and test the software to see if it actually worked. This indicates that the uncertainties related to the technical level of EDI solutions played a role for the non-adopters. By comparing data from the eight companies involved in the TDP it was observed that factors related to the organizational context appeared to have influenced the adoption level of the participants. These factors included company size, legal status, and position in the supply-chain. However, as the quantitative assessment of these factors, company size, legal status, and position in the supply-chain, later revealed, these three factors however lost their statistical significance when applied to a larger sample of businesses and when combined with the opinion data items related to the three contexts (cf. Table 8-49, page 297). Though the three propositions (Proposition 1, Proposition 2, and Proposition 3) related to the three factors in a statistical sense could not individually be rejected (cf. Sections 8.2.2, 8.2.3, and 8.2.4), they had in a statistical sense no impact on the selection of the explanatory variables in the regression analyses.
312
Based on the qualitative assessment of information from the eight informants in the case study, it was found that the primary motivator for EDI adoption among the adopters was to be found in the environmental context. For non-adopters and the one planner influences from the environmental and the technological context were the explanatory factors preventing adoption. Organizational context variables were represented by company size, position in supply-chain, and legal ownership. It was however found, that the explanations for adoption and diffusion of EDI amongst the involved businesses were influenced by factors other than those explicitly defined in the three contexts outlined by Tornatzky and Fleischer (1990). Based on these insights from the case study the answer to the second research question, “To which degree can the motivation to adopt IOS, exemplified by EDI, be explained by issues related to the organizational context, the environmental context, and the technological context?” is, that the three contexts only to a certain extent can explain the motivation for adoption of EDI in the Danish steel and machinery industry.
9.2 Outcome of the quantitative assessment of the second research question The three contexts, the organizational, the environmental, and the technological were assessed quantitatively through operationalization of fifteen opinion data items related to the three contexts. In relation to the strengths of the operationalization of the three contexts, it was found that the Cronbach’s’ coefficient alpha was at a satisfactory level for the organizational context and for the environmental context for all three adoption levels. On the other hand the technological context was below the recommended minimum level for all three adoption levels. This indicated that the technological context was not well-defined. The next step in the investigation of the second research question was an attempt to isolate the factors that provided the best explanatory power in relation to motivation for EDI adoption amongst the responders. This process resulted in a total four opinion data items with statistical significance. These opinion data 313
items were related to the organizational and the environmental contexts not to the technological context. It was found that the motivators for adoption of EDI were to be found in relation to organizational readiness, the prospect of increasing the company’s market share, and the knowledge that several business partners are using EDI. The fourth opinion data item, recommendations from others, was statistical significant. This opinion data item influenced planners negatively in relation to the adoption process. The group of responders in the survey, which had indicated that they did not plan to adopt EDI, the non-adopters, did not agree with any of the defined opinion data items related to motivators for EDI adoption. This lent support to the value of the strength of the explanatory variables in the sense that those factors, which were determinants for adoption or for planning to adopt, were perceived as irrelevant by the non-adopters. An interpretation of the reason why these particular motivators, resulting from the logistic regression analysis, were of importance, led to the conclusion that especially the nature of the social system and change agents’ promotion efforts were very useful for understanding the responders’ preferences. It may be surprising that the technological context did not influence the adoption decision for any of the three groups of responders. However, compared to previous studies that applied the Tornatzky and Fleischer structure (cf. Table 6-3, page 227) it is not surprising. Table 6-4 (page 240) outlined which of the three contexts, the organizational, the environmental, and the technological, that were found to hold the best explanatory power. Two (Chau and Tam, 1997; Thong, 1999) out of the six studies included in Table 6-4 concluded that the technological context items were the determining factors for adoption. These two studies were conducted in an Asian context and the authors noted that technology might play a different role in this cultural context compared to a Western context. In the following some arguments are offered related to the operationalization of the technological context. This is done in order to examine whether or not the right items were included in the technological context.
314
It should be recalled that the case study and the survey were practice driven rather than theory driven (cf. Section 2.6) and that the items included were results of discussions with practitioners from the involved business associations. This however does not reduce the researcher’s responsibility. It is the researcher’s responsibility to apply the appropriate theoretical models and adapt items derived from interaction with practitioners in the theoretical model (Zmud, 1998). As mentioned in the presentation of the technological context (Section 7.4) the technological attributes under investigation were related to “managerial aspects of technology”. This created certain limitations to a survey of the technological capabilities of the company, since this group of responders do not often readily have that type of knowledge at hand. Control variables on for example use of technology in the organization e.g., use of computers, use of intranet or extranet, use of ERP-systems etc. could have been included. However, in the year 2000 when the survey was conducted it was known that 98 percent of all companies in the Danish industry used computers in their daily work and that 92 percent had an intranet (Ministry of Information Technology and Research, 2000), and it would be a search for redundant information to explore general technological capabilities in the surveyed organizations.107 At the same time it was acknowledged that it would be unrealistic to expect managers to be able to provide information on specific technological capabilities. The research items related to technological attributes were therefore focused on managerial issues related to technology. The studies from the literature review on adoption and diffusion on IS (cf. Table 6-3, page 227) which had operationalized the technological context or technological attributes in general had almost all operationalized the technological context by using the innovation attributes outlined by Rogers (1995). One or more of the five innovation attributes, relative advantage, compatibility, complexity, trialability, and observability, were used to characterize the technological context by six out of the nine studies (Kurnia 107
The survey included a question concerning non-adopters’ availability of ERPsystems in the organization.
315
and Johnston, 2000; Thong, 1999; Lai and Guynes, 1997; Premkumar and Ramamurthy, 1995; Sabherwal and King, 1995; Grover, 1993). These characteristics have not been found to be good identifiers of technological attributes (Lyytinen and Damsgaard, 2002). As outlined in Section 6.7.3. the nine studies also included other attributes designed to explore the technological context. These were categorized as cost benefit considerations and technological sophistication in the presentation of the attributes. Inclusion of attributes related to “technological sophistication” could have improved the reliability of the survey in relation to the technological context. The technological sophistication factor comprised perceived importance of compliance to standards, interoperability, and interconnectivity, which were included by Chau and Tam (1997) in their survey. These measures could have been used as objective measures to assess the technological capability of the studied firms. It was however expected that managers would have limited knowledge of these issues. A consequence could have been to expand the scope of responders to include e.g. people from the IT-department. That would however have created some new obstacles for example related to available responders. Most of the companies included in the survey were, as illustrated in Table 12-3 (page 376), very small. It was very unlikely that the companies would have a separate IT-department. It was therefore decided to re-engineer the technological context and include opinion data items, which could in a meaningful way be answered by managers in small companies. Returning to the second research question, “To which degree can the motivation to adopt IOS, exemplified by EDI, be explained by issues related to the organizational context, the environmental context, and the technological context?”, it is more difficult to perform an assessment of whether or not there could be other contexts or elements, which could explain motivators for adoption in relation to the quantitative evaluation of the three contexts. The reason being that only these three specific contexts were examined in the survey. The interpretation of the survey results did however leave the impression that even though the opinion data items were a set of predefined response options, the interaction of the items found in the logistic regression analyses lifted the responses out of the rigid
316
classifications defined by the three contexts. This was especially clear in relation to adopters (cf. Section 8.4.1.1). In the interpretation of the two propositions O6 and E3 it was found, that social processes played an important role in relation to why particularly these two opinion data items influenced the companies to adopt EDI. Social processes are not explicitly related to any of the three contexts in the Tornatzky and Fleischer model. It is therefore argued that the three contexts only partially explain the factors motivating adoption of IOS. Based on data from the survey the answer to the second research question, “To which degree can the motivation to adopt IOS, exemplified by EDI, be explained by issues related to the organizational context, the environmental context, and the technological context?”, is therefore, that the three contexts can identify elements, which are important motivation factors for adoption. But, they do not explain everything related to the interactions of the items explaining a given type of adoption.
9.3 Triangulation based on the two bearings taken in businesses in the Danish steel and machinery industry “It is a delicate exercise to decide whether or not results have converged.” (Jick, 1979), since the researcher using the triangulation method is likely to rely even more on a “feel” of the situation. Another aspect, which is more methodological in nature, is whether or not all components of a given triangulation can have equal weights (ibid.). Recalling the discussion of the strength and validity of different methods in Chapter 2 it was argued that data independent of the means of collection is a reliable source as long as the prescribed method is followed rigorously. The explanatory power of the findings from the quantitative study and the qualitative study are therefore juxtaposed. Jick argues that confidence in the results grows considerably, when there is convergence in the results based on triangulation. However, when divergent results emerge, alternative, and most likely more complex explanations are generated.
317
It could be argued that an explanatory theory for adoption should be capable of explaining both motivators for adoption and non-adoption. One finding – perhaps the most important – in this study is that this was actually the case. The findings related to adopters and planners could in reverse order be applied to non-adopters. The adopters in the case study were motivated to adopt EDI due to some sort of pressure and/ or due to expectations of improved strategic performance. The non-adopters on the other hand had not been subject to pressure nor did they find that their nonadoption of EDI had led to strategic losses. The same “reverse order” patters were even more evident in the survey where Propositions O6, E2, E3, and E4 determined the response profile of adopters and planners and O6, E2, and E3 characterized the response profile of non-adopters (cf. Table 8-50, page 298).108 It can be questioned whether this study rests on true triangulation. The insights derived from the case study of the TDP were applied in the design of the survey. The bearing taken in the survey was therefore not unbiased with respect to the design of the survey instrument since it depended on knowledge gained from the first bearing. One argument in favor of the claim that a true and unbiased triangulation has been made is that the “between methods” approach to triangulation was applied in the current study. This approach consists in using multiple methods to examine the same dimension of a research problem (Jick, 1979). The survey inherited data from the case study. However, the same dimension of the research problem – to identify motivators for IOS adoption - was addressed through different methods, and the outcome of the two different types of examination of the research question could in principle have turned out differently. There was partial convergence in the results from the triangulation. As illustrated in Table 9-1 (page 320) the motivators for adoption of EDI were 108
This at least is the case at the gross level where factors explaining motivators for adoption are identified. As illustrated in the interpretation of the quantitative results (cf. Section 8.4.1.1) the interpretation of the factors related to each of the three levels of adoption provides a rich and detailed picture, which can hardly be put in “reversed order” for non-adopters.
318
mainly to be found within the environmental context both in the case study and in the survey. Motivators for adoption of EDI were however also found within the organizational context and the technological context especially in relation to the informants in the case study. Technological context issues were evident in the case study amongst the non-adopters. One issue, which is interesting in relation to the technological context, is that the well-informed non-adopters had reservations with respect to this technology. This was not the case for the responders in the survey. Although about forty percent of the non-adopters indicated that they were interested in getting more information about EDI, they did not express any concerns about technological issues. One explanation for this phenomenon may be that the non-adopters in the TDP developed an understanding of the complexity of EDI, and therefore expressed a “higher level” of uncertainty. On the other hand, the nonadopters included in the survey got their information about EDI from the more general EDI campaigns, which were less critical to possible obstacles related to EDI adoption. Another interpretation of this phenomenon could be that the non-adopters in the case study in a sense were closer to become adopters than the non-adopters in the survey. The TDP non-adopters have started to consider EDI adoption and have on that account gathered a more varied information about this technology. The technological context issues were according to the responders only of relevance for the non-adopters in the case study. This could suggest that Harrison et al. (1997) are on the right track, when they argue that the traditional technological barriers for organizational adoption of IOS might not play the same dominant role as it did earlier. The reason being that the technological development has led to an increase in quality and functionality, and a decrease in cost of hardware and software.
319
Table 9-1. Overview of motivators for EDI adoption Qualitative assessment of motivators for EDI adoption Adoption level
Organizational context
Environmental context
Adopters
- Innovation champions - Innovative image
- Active pressure - Subjects to pressure - Strategic performance - Attractive business partner
Planner
Non-adopters
- Company size - Legal status - Position in supply chain - The company’s activities are not suitable for EDI - Lack of organizational readiness - No management support - No operational benefits
Technological context
- Strategic performance
- Cost of software
- No pressure - No strategic loss - No critical mass
- Cost of software - Cost of integration - Technological uncertainty
Quantitative assessment of motivators for EDI adoption Adoption level
Organizational context
Environmental context
Adopters
- Organizational readiness
- Several business partners use EDI - The prospect of increasing market share - Recommendations are not motivating adoption
Planners
Non-adopters
- No organizational readiness
- No prospect of increasing market share - No business partners use EDI
320
Technological context
A closer look at the environmental and organizational context motivators reveals that there is only some degree of convergence with respect to these particular items across the two research methods. The environmental context included two different themes in the case study and the survey: Pressure and strategic performance. Pressure was a dominant explanatory factor for the informants in the TDP, whereas the responders in the questionnaire indicated a more subtle form of pressure. For adopters in the survey the mere knowledge that business partners are using EDI is motivating EDI adoption. For the TDP participants direct pressure is a motivating factor. Strategic performance was relevant for all levels of adoption among the TDP informants. That same item was only relevant for the planners and the non-adopters in the survey. In the survey only a single item was found to be important within the organizational context. Adopters indicated that organizational readiness had been a motivator for adoption of EDI. The non-adopters on the other hand indicated that they did not possess the necessary organizational readiness for EDI adoption. Lack of organizational readiness was also a reason given for non-adoption among the two non-adopters in the TDP. The explanatory factors related to the organizational context were generally more diverse in the TDP. The triangulation of the bearings gave a fairly good indication of the position of the vessel loaded with explanatory factors for motivation of EDI adoption. However, details concerning the contexts differed depending on whether data was obtained from the case study or from the survey. One implication of the triangulation is that the three contexts may be good pointers indicating the domain, where explanatory factors are to be found, whereas a more detailed understanding of the factors within the contexts are to be found through a qualitative evaluation of data. The “feel” for the situation (Jick, 1979) becomes important.
321
9.4 An empirically derived motivation model for IOS adoption One of the aims of this study was as stated in the first objective (cf. Objective A, Figure 1-1, page 6), “ To consider possible improvements of the present adoption models used in MIS research.” It was stated in Section 2.6 that the goals were to make a theoretical contribution to the existing theory and at the same time to be able to provide useful recommendations for practitioners at the business association level. In order to meet the first goal previous research was explored to see how other studies had examined the issue of adoption of IOS. Two models were generally found to be used as references for adoption of IOS. These two models were more or less uncritically applied to data from the present study. At this point the models are evaluated. The purpose is to consider possible improvements of the models leading to refinements of the presently used adoption theory. The two adoption models that were used as theoretical framework for interpreting data on motivation for EDI adoption were expanded in the very design of the study. Both frameworks are designed to answer why individuals and organizations are adopting an innovation. In this study the frameworks were used to explain, why some organizations adopted EDI, what made others consider adopting, and why some did not find the prospect of EDI adoption attractive. The Tornatzky and Fleischer model for adoption of innovations in organizations was found to be a valuable tool for operationalising factors that might explain the motivation for adoption of IOS. These contexts were however difficult to interpret and understand without applying the characteristics outlined by Rogers (1995). Whereas the Tornatzky and Fleischer framework could be used to identify the major factors influencing motivation for adoption, Rogers’ framework was useful for gaining an understanding of the underlying preferences for a particular context. The advantages and drawbacks of these two models inspired the development of a new model, which is blend of the two well-known adoption models. It is expected that this new model can help researchers identify and understand the motivators leading to adoption of IOS in a MIS setting. 322
Figure 9-1 illustrates this new model, the Double Domain Motivation Model (DDMM), for IOS adoption. This model comprises two interacting domains inspired by Tornatzky and Fleischer’s (1990) and Rogers’ (1995) frameworks. One domain is related to specific organizational adoption contexts (Tornatzky and Fleischer, 1990) and the other domain is related to general adoption determinants (Rogers, 1995). Figure 9-1. The Double Domain Motivation Model for IOS adoption Specific organizational adoption contexts domain • Organizational context • Environmental context • Technological context - operational level - content
Adoption Motivation
Decision Nonadoption
General adoption determinants domain • Perceived attributes of innovations • Type of innovation-decision • Communication channels • Nature of the social system • Extent of change agents’ promotion efforts - interpretive level - context
The “Specific organizational adoption contexts domain” comprises measurable characteristics of the three contexts included in the Tornatzky and Fleischer framework. These measurable characteristics can be both tangible and perceived. The tangible, measurable characteristics are the primary innovation attributes, which by Downs and Mohr (1976) were described as invariant across settings and organizations. In the survey these tangible measurable characteristics were company size, position in the 323
supply-chain, and legal status of the company. The perceived measurable characteristics are those Downs and Mohr labeled secondary innovation attributes. They are characterized by being subjective and influenced by the particular situation, by the innovation, and by the responder. In the survey the perceived measurable characteristics were represented by the fifteen opinion data items. These opinion data items included different degrees of pressure and issues related to human resources. The measurable characteristics provide the content for the organizational context, the environmental context, and the technological context. Data for further interpretation thereby becomes available. The “Specific organizational adoption contexts domain” therefore comprises characteristics related to the operational level where content is established. The “General adoption determinants domain” provides a set of predefined elements related to the innovation and especially to the social processes associated with adoption. The general adoption determinants domain provides a foundation for an interpretation of the measurable characteristics provided by the “Specific organizational adoption contexts domain”. The understanding of, why organizations adopt - or do not adopt - a technological innovation is in this way placed in a particular context.109 It is a context where social processes play the primary role. These social processes are related to political, organizational, and environmental structures. The social processes are related to the perception of the technological innovation associated with the five dimensions: Relative advantage, compatibility, complexity, trialability, and observability. The social processes are also related to the types and means of receiving information about the innovation, the social networks and norms, the position of the decision-maker, and the influences from change agents. In the present study the “General adoption determinants domain” were used to interpret the TDP data. The determinants gave meaning to data – a 109
It is a paradox that the “Specific organizational adoption context” provides content whereas the “General adoption determinants” provide context. This is however considered to be merely a linguistic coincidence rather than a practical problem.
324
context was established. The content for the organizational context was for example provided by size of the company, economic capability, and organizational readiness for change. Content for the environmental context was for example provided by degree of pressure from business partners and competitive forces, and content for the technological context by for example EDI software and EDI integration. By applying the five variables, Rogers found were determinants for the rate of adoption of innovations, it became evident why motivation for EDI adoption and diffusion in the TDP was absent. The very same pattern was found in data from the survey where the nature of the social system and change agents’ promotion efforts were useful for understanding why certain opinion data items were found to be important for adopters, planners, and non-adopters. As illustrated in Figure 9-1 the DDMM is not considered to be a static model. The two arrows in the model indicate that the two domains, general adoption determinants and specific organizational adoption contexts, interact with each other. Interpretation of the content leads to a better understanding of the context, which again provides a deeper and richer insight into the content, which ultimately gives room for a contextual understanding of the motivators leading to adoption. A criticism of the “General adoption determinants domain” in relation to its ability to describe IOS adoption is that the set of predefined elements are unable to capture the environmental and technological characteristics connected to IOS (Lyytinen and Damsgaard, 2001; Prescott and Conger, 1995). This shortcoming is however to a large extent overcome by the addition of the three organizational adoption contexts. As demonstrated in the review of the studies focusing on IOS adoption the researchers operationalized a given technology based on its own premises in the technological context. The interorganizational relations were captured in the environmental context. Nothing prevented institutional structures and key players (elements which by Lyytinen and Damsgaard found to be lacking in the diffusion and innovation theory) from being included in the environmental context.
325
One of the major challenges in applying the DDMM is therefore how to operationalize the correct items within the three contexts. Here correct items means that the research design embraces the characteristics of a given innovation in the organizational, environmental, and technological contexts. A consequence of this point of view is that identification of a core set of constructs is not the ultimate goal. Kwon and Zmud (1987) regretted that no core set of constructs existed in MIS research. However, it is questionable whether it is an optimal solution to have a core set of constructs in relation to research in IOS innovations. What seems to be more appropriate is to have a generic model to which situational content can be added, depending on the particular developments of a given technology. One way of assessing whether or not the attributes of the technological innovation resemble those explored in previous research is to consider the innovation typology presented in Chapter 6 (cf. Figure 6-2, page 209).
9.5 Assessment of the strength of the DDMM as a theoretical contribution Returning to Whetten (1989) and the four guidelines for theoretical contributions, which were presented in Section 2.6, the four “building blocks” will be discussed to demonstrate the strength of the theoretical contribution. Additionally, the purpose is to clarify the theoretical contribution of the DDMM. Whereas the first and the second building block are used to describe the theoretical contribution, the third building block is concerned with explaining the model. The purpose of the fourth building block is to outline the limitations of the theoretical contribution. The first element: A description of which elements logically should be considered as part of the explanation of the social or individual phenomena of interest. The elements of the DDMM for IOS adoption motivators are related to the three contexts outlined by Tornatzky and Fleischer (1990) and the five variables determining the adoption rate of technological innovations (Rogers, 1995). By merging these concepts a tool, which is both descriptive and interpretive in nature, is developed. This tool can 326
identify a broad range of factors motivating adoption of complex technological innovations. By including elements, which are related to organizational, environmental, and technological attributes, there are opportunities for including explanatory factors related to, for example, human resources, economics, market forces, and IOS attributes. The second domain added to the model, the five variables determining the adoption rate, which focus on social processes facilitate an interpretation and understanding of the underlying motivators. The DDMM will in this way be a comprehensive, synergetic model for interpreting adoption motivators. The second element: A description of the relationship between the elements. This building block is used to explain causality between the elements in the model. This is often done by graphically depicting the elements (Whetten, 1989). The graphical illustration of the DDMM shows the causal relationship between the domains. As mentioned the model is considered to be dynamic. In the DDMM there is a flow of data and information crystallizing in a better and more profound understanding of the adoption situation under investigation. Content from the three contexts provides the basis for understanding data thereby contextualizing content. When content is contextualized motivators for adoption of IOS get identified, leading to a decision of adoption or non-adoption. The third element: An explanation of the underlying psychological, economic, or social dynamics that justify the selection of elements and the proposed causal relationships. This building block is concerned with the underlying assumptions associated with the model and the logic underlying the model. The underlying assumptions of the model are that it is not enough merely to conceptualize explanatory factors. It is also necessary to understand the interactions between the concepts in order to understand the motivators leading to adoption. By including an interpretive level in the model a tool for understanding the conceptual preferences is provided. The fourth element: A description of the range of the theory. The logical limit for the DDMM is, that it is suitable as long as other attributes, than those already included in the three contexts, do not in a better way provide
327
a causal and natural content, which determines the motivators for adoption. A positive definition of the range of the theory may be more difficult to demarcate due to the inherent range of IOS. The IOS, EDI, which was evaluated using the DDMM, focused on the exchange of standardized business information in a beforehand-negotiated format, which are the attributes traditionally associated with EDI. Electronic hierarchies dominate this type of IOS (Malone et al., 1987). This leads to the fact that the environmental context plays an important role, since the benefits of the investments also depend on the decisions and loyalty of business partners (Gebauer and Buxmann, 2000). Trust and pressure therefore become vital elements in determining motivators for adoption (Hart and Saunders, 1997; 1998). That may however not be the case in relation to IOS, which are driven by electronic markets. The motivators for adopting the concept of electronic marketplaces (Bakos, 1991; 1997) may therefore differ from EDI adoption, since the notion of for example pressure and number of business partners involved may have only marginal influence in relation to the perceived benefits of the adoption. The five variables determining the rate of adoption may also be a less reliable tool for interpreting IOS related to electronic markets, since the social system is a more vague concept in electronic markets compared to electronic hierarchies. To sum up, the empirically derived model for IOS adoption motivation, the DDMM, is found to be a valuable, useful, and comprehensive alternative to the adoption models used in MIS research at present. That is the case mainly because the model allows for inclusion of variables related to the three contexts, which covers a broad range of aspects, which MIS research focuses on in relation to IOS adoption. By including an interpretive level to the three contexts the DDMM provides a standardized set of elements to evaluate the underlying social processes and preferences related to adoption motivation.
9.6 Summing up Chapter 9 Based on triangulation data from the case study and the survey of the Danish steel and machinery industry indicated that the environmental 328
context was of major importance for the motivation leading to adoption of EDI. It was however found that factors related to the organizational context and the technological context could not be excluded when factors motivating EDI adoption were to be identified. Additionally, factors, which were beyond the three explanatory contexts defined by Tornatzky and Fleischer (1990), were interpreted to be of some importance in relation to EDI adoption in the Danish steel and machinery industry. The answer to the second research question, “To which degree can the motivation to adopt IOS, exemplified by EDI, be explained by issues related to the organizational context, the environmental context, and the technological context?” is therefore that the three contexts only to a certain extent can explain the motivation for EDI adoption. The second outcome of Chapter 9 was the presentation of the Double Domain Motivation Model (DDMM). It was stated in Objective A, that the first objective of this study was, “To consider possible improvements of the present adoption models used in MIS research.” In Chapter 9 it was argued that those two most used adoption models in IS research when combined constituted a coherent framework. The major advantage of the DDMM, based on the insights gained from the study of the Danish steel and machinery industry, is that the Tornatzky and Fleischer (1990) model appears to be a valuable instrument for operationalizing factors determining motivation for adoption, whereas Rogers’ (1995) framework is helpful for interpreting the causal structures leading to motivation for adoption. Content, from Tornatzky and Fleischer’s model, and context, from Rogers’ model, are in this way included in one single model.
329
330
10 Conclusion 10.1 Introduction One crucial question remains: “What if any is the prospective value of reporting findings from a project that rests on a technology that some consider to be outdated?” One answer could be, that not all Danish businesses - or for that matter foreign businesses (e.g. Williams and Frolick, 2001) - find EDI outdated. Even if some businesses find EDI outdated, they still invest in this technology (Andersen et al., 2000). An answer of perhaps more permanent value is that the findings may be generalizable to similar IOS innovations. The academic value of the present study is therefore, that the findings hopefully can be used for assessing the motivation for adoption of similar IOS and technological innovations in other business sectors and environments. From a normative point of view the present study may also be valuable. The study may help furnish the industry and trade associations with recommendations for future campaigns and projects that aim at supporting their members in adoption processes related to IOS innovations. The objective of this final chapter of the dissertation is to clarify the findings reported in the previous chapters and to explicitly relate the research questions to their respective objectives. The chapter starts with a brief overview of the study. Next, explicit answers to the research questions are provided. Thereafter, the findings of the study are presented and put into perspective. The next section lists contributions of the study. Then follows a discussion of the limitations of the study, and recommendations for future research are addressed. The chapter – and the dissertation - closes with a few somewhat philosophical comments.
331
10.2 Overview of the study This study is a longitudinal evaluation of the outcome of the national 1996action plan for e-commerce. The reported data originated from more than half a decade starting with the political 1995-statements, which gave a high priority to the use of information technology in businesses and administration. The 1995-statement stressed the need for formulating an action plan for EDI and e-commerce. The next step was the launch of the 1996-action plan for EDI and e-commerce. This action plan was operationalized in the TDP (TradeDocument Project) by the two major Danish business associations. The TDP which aimed at supporting the adoption and diffusion of EDI included nine companies, manufactures and wholesalers, in the Danish steel and machinery industry. As a part of this analysis of the impact of co-ordinated adoption and diffusion projects the outcome of TDP was evaluated during the autumn of 1999. Based on the knowledge gained from the TDP, from previous research on IS adoption, and from EDI research in the 1990s a survey instrument was developed to evaluate the status of the use of EDI (and e-commerce) in a business sector, which had been exposed to the 1996-action plan and the TDP. This study, which was practice driven in nature (Zmud, 1998), started as an exploratory search for motivators for EDI adoption. The exploratory approach led to changes in focus of the fieldwork throughout the study. The first step, examination of the TDP, revealed that a number of factors had influenced adoption and non-adoption amongst the participating companies. The industry and trade associations, which initiated the project, expected that the TDP provided the optimal conditions for adoption and diffusion of EDI in the Danish steel and machinery industry and especially amongst the companies involved in the project. However, the interviews with the involved companies indicated that some of the participants had their own agendas. They were more interested in networking and exerting influence. The non-adopters were interested in obtaining knowledge of EDI. However, they remained non-adopters due to organizational constraints.
332
After examining the influence of the co-ordinated adoption and diffusion initiatives, it was found prudent to focus more intensely on motivators for adoption. Based on the general attitude from governmental units and the industry and trade associations, specific knowledge from the TDP, and data from previous research, fifteen opinion data items covering a range of issues related to the organizational context, the environmental context, and the technological context were defined. These items were tested in a survey of the Danish steel and machinery industry.
10.3 Explicit answers to the dissertation’s research questions The research questions were presented in Chapter 1 and two objectives for the research questions were outlined in connection with the research questions (cf. Figure 1-1, page 6). The first objective of the study, objective A, was: - To consider possible improvements of the present adoption models used in MIS research. The corresponding research questions, 1a and 1b, related to this objective were: - How are the explanatory variables related to motivators for EDI adoption defined in MIS research at present? (question 1a) - Which models are used to explain motivation for IOS adoption at present? (question 1b) The second objective of the study, objective B, was: - To identify motivators for adoption of IOS in a business sector dominated by small businesses. The corresponding research question related to this objective was: - To which degree can the motivation to adopt IOS, exemplified by EDI, be explained by issues related to the organizational context, the environmental context, and the technological context? (question 2) In the following two sections the research questions corresponding to the two objectives of the study are answered explicitly. 333
10.3.1 Objective A and research questions 1a and 1b On order to investigate possible improvements of the presently used adoption models in IS research it was found prudent first to investigate how at present the explanatory variables related to motivators for EDI adoption were defined in MIS research. Based on a review of a decades EDI publications from the top-five MIS journals the explanatory variables related to motivators for EDI adoption in MIS research were identified. It was found that the prevalent research themes during the last decade of the twentieth century were concerned with elements related to operational and strategic performance, pressure, and socio-technical issues (cf Table 5-3, page 180). Only a few studies specifically focused on adoption and diffusion. It was however found that an implicit assumption in the reviewed studies was, that the identified elements supported or hampered the adoption and diffusion of EDI in the business community. The answer to research question 1a is therefore, that the explanatory variables related to motivators for EDI adoption based on the review of the top-five MIS journals primarily were related to operational and strategic performance. The significance of pressure was another theme, which appeared in a few of the reviewed articles. Sociotechnical issues appeared, but were less explored. Two adoption models were found to dominate the IS research in the review of studies related to adoption of IOS (Rogers, 1995; Tornatzky and Fleischer, 1990). The framework of Rogers was explicitly mentioned in most of the studies examined, whereas the Tornatzky and Fleischer model for adoption of innovations in organizations was not referred to. It was none the less clear that the elements outlined by Tornatzky and Fleischer were present in most of the IS adoption studies used to exemplify the body of IS adoption research. The answer to research question 1b is therefore, that at present two adoption models, Rogers (1995) and Tornatzky and Fleischer (1990) are used to explain IOS adoption in a MIS context. These two models were applied to qualitative and quantitative data from the present study. This led to the conclusion that the two models should be
334
considered to be complementary rather than alternatives. Based on this insight an empirically model for IOS adoption motivation, the DDMM, was presented. The DDMM suggests that the Tornatzky and Fleischer model is useful for operationalization of research items within the three contexts, thereby providing content to a given study of motivators for IOS adoption. The adoption model outlined by Rogers is simultaneously used to contextualize data through an interpretation guided by the five variables determining the rate of adoption. 10.3.2 Objective B and research question 2 Objective B of this study was: “To identify motivators for adoption of IOS in a business sector dominated by small businesses.” Three signposts indicating possible motivators for IOS adoption were included in research question 2 which stated, “To which degree can the motivation to adopt IOS, exemplified by EDI, be explained by issues related to the organizational context, the environmental context, and the technological context?” Based on a study of qualitative and quantitative data, and a triangulation of these two bearings it was found that the motivators influencing adoption of IOS in a business sector dominated by small businesses only to some degree can be explained by these three contexts. It was found that the environmental context had the best explanatory power both in relation to the qualitative and the quantitative data and that the two other contexts also influenced motivation for adoption among the companies in the TDP. However, elements related to social processes, which go beyond the scope of the three contexts, had to be included in order to identify the underlying motivators influencing adoption of IOS. The motivation for adoption of IOS, exemplified by EDI, could therefore in this setting only to a certain degree be explained by considering the organizational context, the environmental context, and the technological context.
335
10.4 Conclusions of the study The outcome of the TDP is, similar to the projects reported in Section 3.5.2, difficult to assess within a short time horizon. It is tempting to conclude that the effects of political initiatives are somewhat limited considering the amount of resources invested. Additionally, it is reasonable to conclude that the “organizing vision” may have failed.110 The companies involved in the TDP which had adopted EDI did obviously not present the innovation in such a way that it convinced the potential adopters of the advantages of EDI. It could be argued that the adopters involved in the TDP were not the best opinion leaders (Rogers, 1995) to support adoption and diffusion of EDI within and beyond the TDP. A closer examination of the adopters in the TDP indicated that they were not fully convinced about the benefits of the innovation. That is at least the case if the low levels of adoption and diffusion are used as parameters. As the analysis of the data from the Danish steel and machinery showed the low level of adoption and diffusion of EDI amongst the participants was due to lack of pressure from business partners and lack of organizational readiness. These two factors were found to influence the non-adopters taking action, and at the same time they were for adopters found to motivate adoption. That was the conclusion both from the qualitative analysis and the quantitative analysis of the businesses in the Danish steel and machinery industry. The technological aspects that were included in the case study and in the survey were not found to influence the motivation for adoption. In the survey the technological context attributes were not found to have any significance in relation to the responders’ patent priorities. For the TDP
110
The organizing vision represents the efforts of the members of a community to make sense of an innovation as an organizational opportunity (Swanson and Ramiller, 1997). Swanson and Ramiller define the organizing vision as “a focal community idea for the application of information technology in organizations.” The organizing vision is build around experiences gained from organizational adopters that are communicated to prospective organizational adopters.
336
participants both the non-adopters and the planner stated that the cost of software and cost of integration of software was an inhibitor for adoption of EDI. These aspects were however of minor importance compared to the operational and strategic significance of EDI. The non-adopters indicated that if the need arised – that is if business partners encouraged them to adopt EDI – then they would make the necessary investment. It should be emphasized, that the informants were managers. It might therefore be a simplification of the real situation to conclude that technological attributes are of no relevance in relation to motivation for adoption of EDI. It is however fair from this study to conclude that managers do not pay much attention to technological attributes when considering EDI adoption. One reason could be the decrease in cost of hardware and software (Harrison et al., 1997). Viewed in a broader perspective it is deemed fair to conclude that the diffusion efforts from the 1996-action plan only had a limited impact on the Danish steel and machinery industry – and the Danish industry sectors in general – where about fifteen percent of the companies have adopted EDI. One explanation for this low level of adoption and diffusion of EDI in the Danish business environment, which can be supported by other findings (Mansfield et al., 1977; Romeo, 1977), is the relatively weak competitive environment in the Danish steel and machinery industry. Based on an examination of the US steel industry, Mansfield et al. found evidence indicating that intense competition stimulated rapid spread of an innovation. Mansfield et al. found that the rates of diffusion were higher in industries not dominated by a few large firms. Romeo (1977) reached the same conclusion. He found that the rate of diffusion was more rapid in industries with more firms of equal size than in industries dominated by few producers. Table 2-4 (page 375) illustrates that the Danish industry segment is characterized by a few large firms (100+ employees) and a high number of small firms. By solely looking at the percentage of adopters and the distribution of small and larger businesses it is therefore plausible to presume that the low level of adoption and diffusion in the Danish steel and machinery industry is due to moderate competition in the business environment. From the qualitative and quantitative assessment it however
337
became evident that other factors influenced the motivation for adoption of EDI among the companies involved in the analysis. The motivators were primarily found in the environmental context (cf. Table 9-1, page 320) where critical mass, pressure, and strategic performance were identified to play an important role for the motivation leading to adoption. Apart from that organizational readiness and innovative image was found to influence the motivation for adoption. It is generally assumed that the purpose of adoption of an innovation is to improve the effectiveness or performance of the adopting organization (Damanpour and Goplakrishnan, 1998). This assumption was also found to dominate EDI research from a MIS perspective. Ten years of EDI research in the top-five MIS journals showed that especially improved performance was the dominant research theme in these studies. It was also concluded in most of the studies, that adopters of EDI had achieved improved operational or strategic performance. It could therefore be tempting to consider whether or not the businesses in the Danish steel and machinery industry perceived EDI as being a less valuable innovation, since this industry sector did not find it attractive to adopt EDI. It is a fact that the bandwagon effect111 has not yet occurred in the Danish steel and machinery industry, where only about fifteen percent of the businesses in the business sector have adopted EDI. Data from the case study indicated that the nonadopters found EDI to suffer from shortcomings especially in relation to cost of purchase and cost of integration,112 which makes it less attractive to adopt EDI. It could therefore be argued that the non-adopters are aware of
111
Bandwagons are diffusion processes where adopters choose an innovation not because of its technical properties, but because of the sheer numbers of adoptions that have already taken place (Abrahamson and Rosenkopf, 1993). 112 It should be stressed that the reservations towards EDI amongst this particular group of non-adopters should be seen in the light of the TDP EDI software. That particular EDI software was perceived as less attractive due to its stand-alone nature.
338
possible pro-innovation biases113 and over-adoption114 in relation to EDI. In the TDP case this was the conclusion. In the TDP it was found that adoption might not be the best course of action (Cooper and Zmud, 1990). In the survey technological barriers were however not found to be important for non-adopters. They were more concerned with organizational readiness and environmental factors related to possible prospects of increasing market shares. Absense of pressures had so far not made these companies adopt EDI. The essential conclusion from the study is that policy makers at the macro and the meso levels focus on certain benefits assumed to accrue from EDI and IOS, whereas businesses are driven by other perceived benefits related to business relations. The policy statements and the 1996-action plan suggested that adoption and diffusion of EDI could lead to considerable rationalization benefits, closer interplay between organizations, and personal development of employees. Adoption would according to these statements lead to improvement in operational and strategic performance. This main message is similar to the overall presentation of EDI research from the 1990s as communicated in the top-five MIS journals. The business associations involved in TDP tuned their project to the statements communicated in the 1996 action plan and they operationalized the “spirit” of the action plan in the objectives of the TDP. The study of the involved companies in the TDP indicated that companies perceived the opportunities of EDI differently. Politics and strategic alliances were on the agenda rather than improved operational performance. Stategic performance played a role both for adopters and non-adopters, but pressure and absense 113
The pro-innovation bias refers to the presumption that an innovation is always to be considered the best course of action. However, the observation that a substantial number of potential adopters decide not to adopt the innovation in a particular case may not be attributable to these individuals themselves. It may well be the case that the system is at fault for not providing an innovation more appropriate to the individual’s needs, and so the individual may be well justified in rejecting the innovation (Frambach, 1993; Kwon and Zmud, 1997). 114 Over-adoption refers to the phenomenon where adoption of an innovation is done by an individual even if experts recommend that (s)he should reject the innovation. “Certain individuals have such a penchant for anything new that they occasionally appear to be suckers for change. They adopt when they shouldn’t.” (Rogers, 1995), p. 215.
339
of pressure seemed to be one of the main motivators for whether or not companies adopted EDI. The above-mentioned considerations and conclusions lead to the following four statements, which summarizes the study of factors motivating adoption of IOS in the Danish steel and machinery industry: - The benefits of EDI communicated by policy makers, and the benefits of EDI perceived by businesses are not the same. - Paedagogical intervention in the form of soft law has a limited impact on motivation for adoption of an IOS among businesses. - Pressure and organizational readiness are the primary motivating factors for adoption of IOS in the Danish steel and machinery industry. - When viewed from a managerial perspective technological attributes are not given a high priority in the Danish steel and machinery industry.
10.5 Contributions from this study The unique contribution from this study is first of all the scope of examination. Both adopters and non-adopters in a single industrial sector, which had received massive support and information about the innovation from industry and trade associations, are included in the study. Chau and Tam (1997) made a similar study of adoption and non-adoption of IOS. They did however look at several sectors. The second feature, which makes this study unique, is the operationalization of the Tornatzky and Fleischer (1990) diffusion model from an IOS perspective. Based on knowledge gained from the field study of a pilot project initiated by two major Danish industry and trade associations and previous research on EDI, the three adoption contexts, the organizational context, the environmental context, and the technological context, are operationalized and tested on quantitative data. The third aspect that makes this study unique compared to previous research is that focus is on a particular stage in the adoption process: The motivation for the adoption-decision. The study is thus based on the factor 340
approach towards adoption and diffusion; not on the process approach that focuses on the entire process from initiation to implementation (Langley, 1999; Prescott and Conger, 1995; Markus and Robey, 1988). The present work can be viewed as being complementary to previous adoption studies, since it provides new insights into organizational adoption of IOS. Even though IS research focusing on EDI adoption has widely used the factor approach no unequivocal factors explaining EDI adoption have so far been found.115 Previous research has either mainly been driven by case studies (Iacovou et al., 1995; Kurnia and Johnston, 2000) or by surveys (Grover and Goslar, 1993; Premkumar and Ramamurthy, 1995; Thong, 1999) rooted in previous research items. This study was designed to combine both an exploratory case study and a survey. The survey instrument was developed based on the insights gained from an exploratory case study, from close interaction with policy makers in industry and trade associations, and finally from an extensive study of the available literature on explanatory factors influencing motivation for adoption of IOS. Based on the analysis and discussion in Chapter 9 it was however found, that it may not be appropriate to have a set of fixed explanatory factors, which can be applied to any type of IOS. The study is focused on isolating explanatory factors related to motivation or reservation for adoption. Researchers have especially in studies related to quantitative assessment of adoption motivators designed their studies and their respective analysis with the purpose of finding latent, underlying structures explaining adoption. This study aimed at uncovering the patent priorities of the responders. Consequently, the study focused on key explanatory factors for the adoption motivation rather than grouping sets of items that may or may not meaningfully explain adoption. By identifying the patent priorities of the responders the knowledge gained from these priorities also becomes relevant to and useful for practitioners. At the same
115
It should be noted that a general critique of IS research (in relation to survey research) is that it so far has been unable to yield a cumulative body of knowledge. It has been found to be a-theoretical, and ill suited for addressing the subtle dynamics of IT in complex social settings (Kraemer & Dutton 1991).
341
time this knowledge is relevant to the professional business associations, where future campaigns targeted at businesses can be designed in accordance with the motivators and reservations for adoption that are at present found to be dominant in the business sector. The knowledge may also be relevant at the organizational level, where managers considering adoption of IOS evaluate their strategy in relation to the organizations’ business partners. Besides, the above-mentioned characteristics the theoretical contribution of this study from an academic point of view is found in the empirically derived model for IOS adoption motivators, the DDMM. Finally, the study is a demonstration of the usefulness of triangulation in relation to investigations of motivators for adoption in MIS research. For practitioners in professional business associations the contribution of this study is to be found in relation to understanding how their members perceive information campaigns and pilot-projects. This research project provides a better understanding of the preferences of adopters, planners, and non-adopters of IOS. Results from the study can therefore help to target future diffusion campaigns in order to reach potential adopters of IOS
10.6 Limitations In many ways this study is retrospective. The study more or less uncritically accepted the assumed EDI benefits reported in previous research. The study ignored that this particular technology, EDI, by some is viewed as old-fashioned. Furthermore, the study ignored the fact that EDI has been met with a fair amount of critique throughout its history due to a number of shortcomings, such as lack of a common standard, high costs, and absence of critical mass (Henriksen and Görsch, 1999). However, it is inherent in the design of the study to explore the value of EDI from the users’ perspective giving academic judgment a second priority. Therefore, the study was designed to be practice driven (Zmud, 1998), and the study applied the method demonstrated by Flyvbjerg (1994). The implications of 342
these choices are that it was attempted to avoid preconceptions about the field of inquiry. Three types of limitations are recognized in this study: - The inherent limitations of an exploratory study. - The critical time dimension related to an exploration of adoption motivators. - Limitations purposely made in the study. The limitations purposely made in the study are related to the theoretical focus of the study, the technology under investigation, and the research subjects. The research subjects all belong to the same industry segment. The small businesses in the Danish steel and machinery industry were chosen. The reason being that this sector generally is viewed as laggards in relation to adoption of new technologies, which aims at supporting administrative routines. The implications of this choice is that the conclusion of the study may reflect the opinions of more conservative responders’ attitude to EDI and generally to technological innovations. In order to maintain focus the analysis is limited to the pre-adoption phase, which excludes issues related to implementation and further evaluation of success or failure of the implementation. Therefore, the business value of EDI is not discussed. Another constraint purposely applied to the study is related to the technology under investigation. EDI was chosen. EDI may not any longer be in fashion in the research community. There are however good reasons for studying EDI in this particular context. Firstly, if impacts of initiatives such as the 1996-action plan and the TradeDocument Project are to be evaluated, then the technology, which was subject to promotion, is a natural starting point for this research. Secondly, it is the author’s belief that EDI is just one technology in a series of managerial “fashions and fads”. Therefore, when viewing EDI from a managerial point of view, conclusions from this study have a potential value for coming generations of IOS. The limitations associated with the time dimension are related to the problems of ex post and ex ante evaluations when including both adopters
343
and non-adopters in the same study. Ex post evaluations of motivations for adoption involves looking back in time to seek possible causes and relationships of the effects of the independent variables. The responders are in a situation where they are evaluating an innovation, which they themselves decided to adopt. This fact combined with their actual experiences with the innovation in question may lead to biased answers. Ex ante assessments of motivations for adoption are challenging the responders in different ways. Prospective assessments of an innovation that might not even be relevant to the responder might not receive the most attentive assessments from the responder. The inherent limitations related to an exploratory study include issues related to operationalization of research items. The operationalization is based on a-theoretical considerations in this study rather than using items from previous research. Another limitation related to the exploratory research strategy is that the study can easily become a demonstration of the researchers learning process. No attempts are made to hide that insights were gained during the research process and that focus shifted throughout the project.
10.7 Recommendations for future research. Continuing the line of thought outlined in the previous section on limitations of the study, future research should aim at compensating for some of these limitations. Implementation issues and thereby an assessment of the benefits of EDI adoption would be a natural starting point for future research. This perspective was not explored in-depth in the present study. Therefore, approaches related to the degree of adoption and implementation among adopting organizations could be targets for future research. It was recognized that the adopters in the TDP in many ways were utilizing their EDI solution in a rather limited way (cf. the VIDS-test for the EDI users in TDP, page 126). It is not easy to understand why users do not make better use of their investment.
344
Next, focus should be broadened to include other types of IOS, both those that aim at supporting hierarchial structures and those IOS that support market structures. Finally, the DDMM should be tested in different settings and with different IOS technologies.
10.8 Post Scriptum The different bearings taken to locate the motivators for adoption of IOS leave the vessel in an ocean full of possibilities. The vessel is loaded with a set of explanatory factors for adoption and non-adoption particularly for the Danish steel and machinery industry, and it seems as if cruising has just begun. It must be admitted that sometimes the cruise was a navigation between Schylla and Charybdis. Though the monsters encountered during my work on the dissertation did not have six pair of feet and six heads, nor were capable of swallowing the waters, there were however many other forces tempting me to flee or to run the vessel aground. The technique of triangulation, which cf. Footnote 2 (page 20) apparently was developed by Tycho Brahe, helped the author to see the multiplicity of motivators for adoption of EDI. Though the bearings indicated that the explanatory factors motivating EDI adoption in a business sector dominated by small businesses were to be found mainly within the environmental context, there were not found to be any Chinese walls to the two other contexts. The bearings did not exclude that explanatory factors beyond the three contexts could be explanatory factors for adoption and non-adoption. It is my hope that the empirically derived model for motivators, the DDMM, for adoption of IOS does not prove to be as wrong as the theory Tycho Brahe was spokesman for. Tycho Brahe expressed the conviction, that the Earth was the center of the Universe.116 As is well known from our classes in philosophy of science there were good reasons for having this
345
belief at that particular time. However, today there are no excuses, neither political nor religious, for not publishing adoption of innovation models, which some may consider to be controversial. Though my castle was not Uranienborg117 the optimal conditions for studying and researching the Universe of MIS and adoption of innovations were provided by the Department of Informatics at Copenhagen Business School. Kind and helpful colleagues never hesitated to encourage and give advice in the process, and rich opportunities were given to travel and visit other research environments and conferences during the Ph.D.-study. I am grateful to those who gave me the privilege of getting all these experiences and opportunities.
116
Tycho Brahe published his theories in 1588, approximately forty years after Copernicus had published his work, where Copernicus stated that the Sun was the center of the Universe and the Earth a planet among others. 117 Tycho Brahe built his observatory, Uranienborg, under the auspices of the Danish king. Uranienborg was at that time one of the finest observatories in Europe.
346
11 References Abildgren, K. & Nielsen, S.C. (2000). Dansk Erhvervsliv. Handelshøjskolens Forlag, Copenhagen. Abrahamson, E. (1996). Management Fashion. The Academy of Management Review 21, 254-285. Abrahamson, E. & Rosenkopf, L. (1993). Institutional and Competitive Bandwagons: Using Mathematical Modeling as a Tool to Explore Innovation Diffusion. Academy of Management Review 18, 487-517. Allen, D., Colligan, D., Finnie, A., & Kern, T. (2000). Trust, Power and Interorganizational Information Systems: The Case of the Electronic Trading Community TransLease. Information Systems Journal, 10, 2140. Andersen, I. (1990). Valg af organisationssociologiske metoder -et kombinationsperspektiv. Samfundslitteratur, Copenhagen. Andersen, I. (1992). On the Art of Doing Field Studies. Handelshøjskolens Forlag, Copenhagen. Andersen, K.V., Bjorn-Andersen, N., & Juul, N.C. (1999). Electronic Commerce and Document Interchange: EDI Application in the Private and Public Sector. Http://www.edi.dk/edi_5.htm. Andersen, K.V., Juul, N.C., Henriksen, H.Z., Bjorn-Andersen, N., & Bunker, D. (2000). Business-to-Business E-commerce, Enterprises Facing a Turbulent World. DJØF Publishers, Copenhagen. Anderson, P. & Tushman, M.L. (1997). Managing Through Cycles of Technological Change. In: Managing Strategic Innovation and Change, M. L.Tushman and P. Anderson (eds), 45-52. Oxford University Press, New York. Argyris, C. & Schon, D.S. (1991). Participatory Action Research and Action Science Compared. In: Participatory Action Research, W.F.Whyte (eds), Sage, Newbury Park. CA, USA. Arunachalam, V. (1995). EDI: An Analysis of Adoption, Uses, Benefits and Barriers. Journal of Systems Management, 60-64.
347
Attewell, P. (1992). Technology Diffusion and Organizational Learning: The Case of Business Computing. Organization Science, 1-19. Bakos, J.Y. (1991). A Strategic Analysis of Electronic Marketplaces. MIS Quarterly, 295-310. Bakos, J.Y. (1997). Reducing Buyer Search Costs: Implications for Electronic Marketplaces. Management Science, 1676-1692. Barley, S.R. (1986). Technology as an Occasion for Structuring: Evidence from Observations of CT Scanners and the Social Order of Radiology Departments. Administrative Science Quarterly, 78-108. Barley, S.R. (1998). What Can We Learn from the History of Technology? Journal of Engineering and Technology Management, 237-255. Barua, A. & Lee, B. (1997). An Economic Analysis of the Introduction of an Electronic Data Interchange System. Information Systems Research, 398-422. Benbasat, I., Goldstein, D.K., & Mead, M. (1987). The Case Research Strategy in Studies of Information Systems. MIS Quarterly, 369-386. Bergeron, F., Bateau, C., & Raymond, L. (1991). Identification of Strategic Information Systems Opportunities: Applying and Comparing Two Methodologies. MIS Quarterly, 89-103. Bergeron, F. & Raymond, L. (1992). The Advantages of Electronic Data Interchange. DATABASE, 19-31. Bjorn-Andersen, N. & Krcmar, H. (1995). Looking Back - A Cross Analysis of 14 EDI Cases. In: EDI in Europe: How it Works in Practice, H. Krcmar, N. Bjørn-Andersen, and R. O'Callaghan (eds.), 299-326. John Wiley and Sons. Bjorn-Andersen, N. & Nygaard-Andersen, S. (1995a). EDIfragt - The Case of the Danish State Railroad's EDI System. In: EDI in Europe: How it Works in Practice, H. Krcmar, N. Bjørn-Andersen, and R. O'Callaghan (eds.), 131-146. John Wiley & Sons, New York. Bjorn-Andersen, N. & Nygaard-Andersen, S. (1995b). IndustriLink - The Case of an Order-entry System in Denmark. In: EDI in Europe: How it Works in Practice, H. Krcmar, N. Bjørn-Andersen, and R. O'Callaghan (eds.), 185-204. John Wiley & Sons, New York. Cachon, G.P. & Fisher, M. (2000). Supply Chain Inventory Management and the Value of Shared Information. Management Science, 1032-1048.
348
Carmines, E.G. & Zeller, R.A. (1979). Reliability and Validity Assessment. SAGE Publications, Beverly Hills. Cash, J.I. & Konsynski, B.R. (1985). IS Redraws Competitive Boundaries. Harvard Business Review, 134-142. Chalmers, A.F. (1982). What is this Thing Called Science? University of Queensland Press. Chan, C. & Swatman, P.M.C. (1998). EDI Implementation: A Broader Perspective. Proceedings of the 11th International Bled Electronic Commerce Conference. Chan, C. & Swatman, P.M.C. (2000). From EDI to Internet Commerce: The BHP Steel Experience. Internet Research: Electronic Networking Applications and Policy, 72-82. Chatfield, A.T. & Bjorn-Andersen, N. (1997). The Impact of IOS-enabled Business Process Change on Business Outcomes: Transformation of the Value Chain of Japan Airlines. Journal of Management Information Systems, 13-40. Chatfield, A.T. & Yetton, P. (2000). Strategic Payoff from EDI as a Function of EDI Embeddedness. Journal of Management Information Systems, 195-224. Chau, P.Y.K. (2001). Inhibitors to EDI Adoption in Small Businesses: An Empirical Investigation. Journal of Electronic Commerce Research, 119. Chau, P.Y.K. & Tam, K.Y. (1997). Factors Affecting the Adoption of Open Systems: An Exploratory Study. MIS Quarterly, 1-21. Chau, P.Y.K. & Tam, K.Y. (2000). Organizational Adoption of Open Systems: A 'Technology-push, need-pull' Perspective. Information & Management, 229-239. Chen, F. (1998). Echelon Reorder Points, Installation Reorder Points, and the Value of Centralized Demand Information. Management Science, 221-234. Chen, J.C. & Williams, B.C. (1998). The Impact of Electronic Data Interchange (EDI) on SME's: Summary of Eight British Case Studies. Journal of Small Business Management.
349
Choi, S.Y., Stahl, D.O., & Whinston, A.B. (1997). The Economics of Electronic Commerce. Macmillan Technical Publishing, Indianapolis, IN. Choudhury, V. (1997). Strategic Choices in the Development of Interorganizational Information Systems. Information Systems Research, 1-24. Choudhury, V., Hartzel, K.S., & Konsynski, B.R. (1998) Uses and Consequences of Electronic Markets: An Empirical Investigation in the Aircraft Parts Industry. MIS Quarterly, 471-507. Clark, T. & Stoddard, D.B. (1996). Interorganizational Business Process Redesign: Merging Technological and Process Innovation. Journal of Management Information Systems, 9-28. Claver, E., González, R., & Llopis, J. (2000). An Analysis of Research in Information Systems. Information & Management, 181-195. Clemons, E.K., Reddi, S.P., & Row, M.C. (1993). The Impact of Information Technology on the Organization of Economic Activity: The "move to the middle" Hypothesis. Journal of Management Information Systems, 9-35. Coase, R.H. (1937). The Nature of the Firm. Economica, 386-405. Cody, R.P. & Smith, J.K. (1997). Applied Statistics and the SAS Programming Language. Prentice-Hall, Inc. Cooper, R.B. & Zmud, R.W. (1990). Information Technology Implementation Research: A Technological Diffusion Approach. Management Science, 123-139. Cox, B. & Ghoneim, S. (1996). Drivers and Barriers to Adopting EDI: A Sector Analysis of UK Industry. European Journal of Information Systems, 24-33. Crag, P.B. & King, M. (1993). Small-Firm Computing: Motivators and Inhibitors. MIS Quarterly, 47-60. Cronbach, L.J. (1951). Coefficient Alpha and the Internal Structure of Tests. Psychometrika, 297-334. Crook, C.W. & Kumar, R.L. (1998). Electronic Data Interchange: A Multiindustry Investigation Using Grounded Theory. Information & Management, 75-89.
350
Crum, M., Premkumar, G., & Ramamurthy, K. (1996). An Assessment of Motor Carrier Adoption, Use, and Satisfaction with EDI. Transportation Journal, 44. Damanpour, F. & Gopalakrishnan, S. (1998). Theories of Organizational Structure and Innovation Adoption: The Role of Environmental Change. Journal of Engineering and Technology Management, 1-24. Damsgaard, J. (1996). Diffusion of Electronic Data Interchange: An Institutional and Organizational Analysis of Alternative Diffusion Patterns. Ph.D. Thesis, R-96-2041, Department of Computer Science, Aalborg University. Damsgaard, J. & Lyytinen, K. (1998a). Contours of Diffusion of Electronic Data Interchange in Finland: Overcoming Technological Barriers and Collaborating to make it Happen. Journal of Strategic Information Systems, 275-297. Damsgaard, J. & Lyytinen, K. (1998b). International Trade and the Speed of Light: Building an Electronic Trading Infrastructure in Denmark, Finalnd, and Hong Kong. in Proceedings of the IFIP WG 8.2 and 8.6 joint working conference on Information Systems: Current Issues and Future Changes.Helsinki, Finland, 10-13.December, 1998. Damsgaard, J. & Lyytinen, K. (2001). The Role of Intermediating Institutions in the Diffusion of Electronic Data Interchange (EDI): How Industry Associations Intervened in Denmark, Finland, and Hong Kong. The Information Society, 195-210. Damsgaard, J. & Truex, D. (2000). Binary Trading Relations and the Limits of EDI Standards: The Procrustean Bed of Standards. European Journal of Information Systems, 173-188. Davis, F.D. (1989). Perceived Usefulness, Perceived Ease of Use, and User Acceptance of Information Technology. MIS Quarterly, 319-339. Davis, F.D., Bagozzi, R.P., & Warshaw, P.R. (1989). User Acceptance of Computer Technology: A Comparison of Two Theoretical Models. Management Science, 982-1003. Dearing, B. (1990). The Strategic Benefits of EDI. The Journal of Business Strategy, 4-6.
351
DeLone, H.W. & McLean, R.E. (1992). Information Systems Success: The Quest for the Dependent Variable. Information Systems Research, 6095. Descartes, R. (1968). The Meditations and Selections from the Principles. Downs, G.W. & Mohr, L.B. (1976). Conceptual Issues in the Study of Innovation. Administrative Science Quarterly, 700-714. Drury, D.H. & Farhoomand, A. (1996). Innovation Adoption of EDI. Information Resources Management Journal, 5-13. Drury, D.H. & Farhoomand, A. (1999). Information Technology Push/Pull Reactions. The Journal of Systems and Software, 3-10. Eckhoff, T. (1983). Statens styringsmuligheter. Særlig i ressurs- og miljøspørsmål. TANUM - NORLI, Oslo. Eisenhardt, K.M. (1989). Building Theories from Case Study Research. Academy of Management Review, 532-550. Erhvervsministeriet (1999). Erhvervsredegørelse 1999. European Council. (1991). Council Decision 91/385/EEC of 22 July 1991. Falch, M. (1998). Internet: The Future Infrastructure for EDI? In: Commercialisation of the Internet, A. Henten and K.E. Skouby (eds.), 37-53. Center for Tele-Information Technical University of Denmark, Lyngby. Fichman, R.G. (1992). Information technology Diffusion: A Review of Empirical Research. Proceedings at the Thirteenth International Conference on Information Systems, Dallas, 195-206. Fichman, R.G. & Kemerer, C.F. (1994). Toward a Theory of the Adoption and Diffusion of Software Process Innovations. In: Diffusion, Transfer and Implementation of Information Technology, L.Levine (ed.), 23-30. Elsevier Science. Fichman, R.G. & Kemerer, C.F. (1999). The Illusory Diffusion of Innovation: An Examination of Assimilation Gaps. Information Systems Research, 255-275. Flyvbjerg, B. (1998). Rationality and Power - Democracy in Practice. The University of Chicago Press, Chicago. Foldspang, A., Juul, S., Olsen, J., & Sabroe, S. (1992) Epidemiologi sygdom og befolkning. Munksgaard, København.
352
Frambach, R.T. (1993). An Integrated Model of Organizational Adoption and Diffusion of Innovations. European Journal of Marketing, 22-41. Gadamer, H.G. (1976). Philosophical Hermeneutics. University of California Press, Berkeley. Galliers, R. (1999). Towards the Integration of E-business, Knowledge Management and Policy Considerations within an Information Systems Strategy Framework. Journal of Strategic Information Systems, 229234. Gebauer, J. & Buxmann, P. (2000). Assessing the Value of Interorganizational Systems to Support Business Transactions. International Journal of Electronic Commerce, 61-82. Giddens, A. (1984). The Constitution of Society - Outline on the Theory of Structuration. Polity Press, Cambridge. Glaser, B.G. & Strauss, A.L. (1967). The Discovery of Grounded Theory: Strategies for Qualitative Research. Aldine de Gruyter, New York. Gregor, S. & Johnston, R. B. (2001). Theory of Interorganizational Systems: Industry Structure and Processes of Change. In: Proceeidngs of the 34th Hawaii International Conference on Systems Sciences. Grover, V. (1993). An Empirically Derived Model for the Adoption of Customer-based Interorganizational Systems. Decision Sciences, 603640. Grover, V. & Goslar, M.D. (1993). The Initiation, Adoption, and Implementation of Telecommunications Technologies in U.S. Organizations. Journal of Management Information Systems, 141. Gurbaxani, V. & Wang, S. (1991). The Impact of Information Systems on Organizations and Markets. Communications of the ACM, 59-73. Hair, J.F., Anderson, R.E., Tatham, R.L., & Black, W.C. (1998). Multivariate Data Analysis. Prentice-Hall International, Inc., London. Hansen, J.V. & Hill, N.C. (1989). Control and Audit of Electronic Data Interchange. MIS Quarterly, 403-413. Hardgrave, B. & Walstrom, K. (1997). Forums for MIS Scholars. Communications of the ACM, 119-124. Harrison, D.A., Mykytyn, P.P., & Riemenschneider, C.K. (1997). Executive Decisions About Adoption of Information Technology in
353
Small Business: Theory and Empirical Tests. Information Systems Research, 171-195. Hart, P. & Saunders, C. (1997). Power and Trust: Critical Factors in the Adoption and Use of Electronic Data Interchange. Organization Science, 23-42. Hart, P. & Saunders, C. (1998). Emerging Electronic Partnerships: Antecedents and Dimensions of EDI Use from the Supplier's Perspective. Journal of Management Information Systems, 87-111. Hatcher, L. (1994). A Step-by-Step Approach to Using the SAS System for Factor Analysis and Structural Equation Modeling. SAS Institute Inc., Cary. Hatcher, L. & Stepanski, E. (1994). A Step-by-Step Approach to Using the SAS System for Univariate and Multivariate Statistics. SAS Institute Inc., Cary. Henriksen, H.Z. (2001). In Search of an Efficient EDIcebreaker: Use of Electronic Marketplaces for the Diffusion of EDI among SMEs. In; Diffusing Software Product and Process Innovations.Proceedings at IFIP TC8 W.G 8.6, fourth working conference, April 7-10 2001, Banff, Canada. 277-291. Henriksen, Z.H. (2000). Can Coordinated IT Projects Lead to Diffusion? A Casestudy of an EDI Project. In: Proceedings at the 23th IRIS conference. Henriksen, Z.H. & Andersen,K.V. (1999). Internetbased EDIFACT-trade Documents for Danish SMEs: An Analysis of the Pre-implementation Phase. In: Proceedings at the 12th Bled electronic commerce conference. Henriksen, Z.H. & Görsch, D. (1999). The Challenges of EDI . In: "Distributed Multimedia - Electronic Commerce and Converging Media", P. Carstensen and M. Falch (eds.), Center for Tele Information, Danish Technical University, Copenhagen. Hill, R. (1997). Electronic Commerce, the World Wide Web, Minitel, and EDI. The Information Society, 33-41. Hitt, L.M. & Brynjolfsson, E. (1996). Productivity, Business Profitability, and Consumer Surplus: Three Different Measures of Information Technology Value. MIS Quarterly, 121-142.
354
Horlück, J. (1994). The Pragmatics of Electronic Data Interchange: The Use of EDI will Unify Business Procedures. International Journal of Information Management, 330-343. Horlück, J. (1996). EDI and Danish Companies. Confederation of Danish Industries. Hosmer, D.W. & Lemeshow, S. (1989). Applied Logistic Regression. John Wiley and Sons, New York. Iacovou, C.L., Benbasat, I., & Dexter, A.S. (1995). Electronic Data Interchange and Small Organizations: Adoption and Impact of Technology. MIS Quarterly, 465-485. Jelassi, T. & Figon, O. (1994). Competing through EDI at Brun Passot: Achievements in France. MIS Quarterly, 337. Jick, T.D. (1979). Mixing Qualitative and Quantitative Methods: Triangulation in Action. Administrative Science Quarterly, 602-611. Johnston, H.R. & Vitale, M.R. (1988). Creating Competitive Advantage with Interorganizational Systems. MIS Quarterly, 153-165. Johnston, R.B. & Gregor, S. (2000). A Theory of Industry-level Activity for Understanding the Adoption of Interorganizational Systems. European Journal of Information Systems, 243-251. Jones, M. (1999). Structuration Theory. In: Rethinking Management Information Systems: An Interdisciplinary Perspective, W.L. Currie and B. Galliers (eds.), 103-135. Oxford University Press, Oxford. Jones, M.C. & Beatty, R.C. (1998). Towards the Development of Measures of Perceived Benefits and Compatibility of EDI: A Comparative Assessment of Competing First Order Factor Models. European Journal of Information Systems. Jui-Lin Lu, E. & Hwang, R. (2001). A Distributed EDI Model. The Journal of Systems and Software, 1-7. Kalakota, R. & Whinston, A.B. (1996). Frontiers of Electronic Commerce. Addison-Wesley Publishing Company, Inc. Kalakota, R. & Whinston, A.B. (1997). Electronic Commerce, A Managers Guide. Addison-Wesley Publishing Company Inc. Kautz, K., Hansen, H.W., & Thaysen, K. (2001). Understanding and Changing Software Organizations: An Exploration of four Perspectives on Software Process Improvement. In: Diffusing Software Product and
355
Process Innovations. Proceedings at IFIP TC8 W.G 8.6, fourth working conference, April 7-10 2001, Banff, Canada. 87-109. Kautz, K. & Larsen, E.Å. (2000). Diffusion Theory and Practice: Disseminating Quality Management and Software Improvement Innovations. Information, Technology & People, 11-26. Kerlinger, F.N. & Howard, B.L. (2000). Foundations of Behavioral Research. Harcourt College Publishers, Fort Worth. Kimbrough, S.O. & Moore, S.A. (1992). Message Management Systems: Concepts, Motivations, and Strategic Effects. Journal of Management Information Systems, 29. King, J.L., Gurbaxani, V., Kraemer, K.L., McFarlan, F.W., Raman, K.S., & Yap, C.S. (1994). Institutional Factors in Information Technology Innovation. Information Systems Research, 139-169. Klein, H.K. & Myers, M.D. (1999). A Set of Principles for Conducting and Evaluating Interpretive Field Studies in Information Systems. MIS Quarterly, 67-93. Klein, S. (1995). The Impact of Public Policy on the Diffusion and Implementation of EDI: An Evaluation of the TEDIS Programme. Information Economics and Policy, 147-181. Kling, R. (1980). Social Analysis of Computing: Theoretical Perspectives in Recent Empirical research. ACM Computing Surveys, 61-110. Kling, R. (1991). Excerpts from "Social Analyses of Computing: Theoretical Perspectives in recent Empirical Research". In: Computerization and Controversy: Value conflicts and Social Choises, C. Dunlop and R. Kling (eds.), 150-166. Academic Press Inc., Boston. Knudsen, C. (1994). Økonomisk metodologi, videnskabsteori og forklaringstyper. Jurist- of Økonomforbundets Forlag, København. Kraemer, K.L. (1991). The Information Systems Research Challenge: Survey Reserach Methods. Harvard Business School, Boston. Kraemer, K.L. & Dutton, W.H. (1991). Survey Research in the Study of Management Information Systems. In: The Information Systems Research Challenge: Survey Research Methods, K.L. Kraemer (ed.), 326. Harvard Business School, Boston. Krcmar, H., Bjorn-Andersen, N., & O'Callaghan, R. (1995). EDI in Europe: How it Works in Practice. John Wiley & Sons, New York.
356
Kreiner, S. (2001). Introduction to DIGRAM and SCD. Internal working paper from Department of Biostatistics, University of Copenhagen. Kuhn, T.S. (1996). The Structure of Scientific Revolutions. The University of Chicago Press, Chicago. Kumar, K., Dissel, H.G.v., & Bielli, P. (1998). The Merchant of Prato Revisited: Toward a Third Rationality of Information Systems. MIS Quarterly, 199-225. Kurnia, S. & Johnston, R.B. (2000). The Need of a Processual View of Inter-organizational Systems Adoption. Journal of Strategic Information Systems, 295-319. Kwon, T.H. & Zmud, R.W. (1987). Unifying the Fragmented Models of Information Systems Implementation. In: Critical Issues in Information Systems Research, R.J.Boland and R.A.Hirschheim (eds.), 227-251. John Wiley & Sons Ltd. Lai, V.S. & Guynes, J.L. (1997). An assessment of the influence of organizational characteristics on information technology adoption decision: A discriminative approach. IEEE Transactions on Engineering Management, 146-157. Lambert, D.M. & Harrington, T.C. (1990). Measuring Nonresponse Bias in Customer Service Mail Surveys. Journal of Business Logistics, 5-25. Langley, A. (1999). Strategies for Theorizing from Process Data. Academy of Management Review, 691-710. Leavitt, H.J. & Whisler, T.L. (1958). Management in the 1980s'. Harvard Business Review, 41-48. Lee, A.S. (1989). A Scientific Methodology for MIS Case Studies. MIS Quarterly, 33-51. Lee, A.S. (1991). Integrating Positivist and Interpretive Apporaches to Organizational Research. Organization Science, 342-365. Lee, H., So, K.C., & Tang, C.S. (2000). The Value of Information Sharing in a Two-Level Supply Chain. Management Science, 626-643. Lee, H.G., Clark, T., & Tam, K.Y. (1999). Research Report: Can EDI Benefit Adopters? Information Systems Research, 186-195. Lundvall, B.A. (1999). National Business Systems and National Systems of Innovation. International Studies of Management and Organization, 6077.
357
Lyytinen, K. (1991). Penetration of Information Technology in Organizations: A Comparative Study Using Stage Models and Transaction Costs. Scandinavian Journal of Information Systems, 87109. Lyytinen, K. & Damsgaard, J. (2001). What's Wrong with the Diffusion of Innovation Theory? In: Diffusing Software Product and Process Innovations.Proceedings at IFIP TC8 W.G 8.6, fourth working conference, April 7-10 2001, Banff, Canada. 173-190. Maaloee, E. (1996). Case-Studier. Af og om mennesker i organisationer. Akademisk Forlag, København. Malone, T.W., Yates, J., & Benjamin, R.I. (1987). Electronic Markets and Electronic Hierarchies. Communications of the ACM, 484-497. Malone, T.W., Yates, J., & Benjamin, R.I. (1989). The Logic of Electronic Markets. Harvard Business Review, 166-172. Mansfield, E., Rapoport, J., Romeo, A., Villiani, E., Wagner, S., & Husic, F. (1977). The Production and Application of New Industrial Technology. Norton, New York. March, J. & Simon, H. (1993). Organizations. Blackwell Science, Cambridge, Massachusetts. Marcus, A.A. & Weber, M.J. (1989). Externally-Induced Innovation. In: Research on the Management of Innovation: The Minnesota Studies, A.H. Van de Ven, H.L. Angle, and M. Scott Poole (eds.), 537-559. Harper & row, New York. Markus, L.M. (1983). Power, Politics, and MIS Implementation. Communications of the ACM, 430-445. Markus, L.M. (1987). Toward a "Critical Mass" Theory of Interactive Media. Universal Access, Interdependence and Diffusion. Communication Research, 491-511. Markus, L.M. & Robey, D. (1988). Information Technology and Organizational Change: Causal Structure in Theory and Research. Management Science, 583-598. Massetti, B. & Zmud, R.W. (1996). Measuring the Extent of EDI Usage in Complex Organizations: Strategies and Illustrative Examples. MIS Quarterly, 331-345.
358
Masters, J.M., Allenby, G.M., LaLonde, B.J., & Maltz, A. (1992). On the Adoption of DRP. Journal of Business Logistics, 47. McMaster, T. (2001). The Illusion of Diffusion in Information Systems Research. In: Diffusing Software Product and Process Innovations. Proceedings at IFIP TC8 W.G 8.6, fourth working conference, April 710 2001, Banff, Canada. 67-85. Miles, M.B. & Huberman, A.B. (1994). Qualitative Data Analysis. SAGE Publications, London. Mingers,J. (2002) The Long and Winding Road: Getting Papers Published in Top Journals. Communications of the Association for Information Systems, 330-339. Mingers, J. (1997). Combining Research Methods in Information Systems: Multi-paradigm Methodology. In: Proceedings at the 5th European Conference on Information Systems, June 19-21, Cork, Ireland 760-776. Mingers, J. & Brocklesby, J. (1997). Multimethodology: Towards a Framework for Mixing Methodologies. OMEGA, International Journal of Management Science, 489-509. Ministry of Information Technology and Research. (1995). From Vision to Action - Info-Society 2000. www.fsk.dk/publ. Ministry of Research and Information Technology (1996). Electronic Commerce in Denmark – A National EDI Action Plan. URL http://www.fsk.dk/publ/. Ministry of Information Technology and Research. (1996a). Danske virksomheders brug af IT 1996. www.fsk.dk/publ. Ministry of Information Technology and Research. (1996b). The InfoSociety for All - the Danish Model. www.fsk.dk/publ. Ministry of Information Technology and Research. (1997). Danske virksomheders brug af IT 1997. www.fsk.dk/publ. Ministry of Information Technology and Research. (1999). Danske virksomheders brug af IT 1998. www.fsk.dk/publ. Ministry of Information Technology and Research. (2000). Danske virksomheders brug af IT 1999. www.fsk.dk/publ. Ministry of Information Technology and Research. (2001). Danske virksomheders brug af IT 2000. www.fsk.dk/publ.
359
Ministry of Research and Information Technology & Ministry of Trade and Industry. (1999). Focus on E-commerce. http://www.elektroniskhandel. dk/focuson.htm. Mintzberg, H., Raisinghani, D., & Théorêt, A. (1976). The Structure of "Unstructured" Decision Processes. Administrative Science Quarterly, 246-275. Mitropoulos, P. & Tatum, C.B. (2000). Forces Driving Adoption of New Information Technologies. Journal of Construction Engineering and Management - ASCE, 340-348. Mohr, L.B. (1969). Determinants of Innovation in Organizations. The American Political Science Review, 111-126. Mohr, L.B. (1987). Innovation Theory: An Assessment from the Vantage Point of the New Electronic Technology in Organizations. In: New Technology as Organizational Innovation, Johannes M.Pennings and Arend Buitendam (eds.), 13-31. Ballinger Publishing Company, Cambridge, Massachusetts. Moore, G.C. & Benbasat, I. (1991). Development of an Instrument to Measure the Perceptions of Adopting an Information Technology Innovation. Information Systems Research, 192-222. Mukhopadhyay, T., Kekre, S., & Kalathur, S. (1995). Business Value of Information Technology: A Study of Electronic Data Interchange. MIS Quarterly. Myers, M.D. (1994). The Disaster for Everyone to See: An Interpretive Analysis of a Failed IS Project. Accounting, Management & Information Technology, 185-201. Newell, S., Swan, J., & Galliers, R. (1998). The Diffusion of Business Reengineering: A Conceptual Model. In: Proceedings of the 6th European Conference on Information Systems, Aix-en-Provence, France. Newell, S., Swan, J., & Galliers, R. (2000). A Knowledge-focused Perspective on the Diffusion and Adoption of Complex Information Technologies: The BPR Example. Information Systems Journal, 239259. Nielsen, R. (1999). Retskilderne. Jurist- og Økonomforbundets Forlag, København.
360
Nunnally, J.C. (1978). Psycometric Theory. McGraw-Hill, New York. O'Callaghan, R. & Turner, J.A. (1995). Electronic Data Interchange Concepts and Issues. In: EDI in Europe: How it Works in Practice, H. Kcmar, N. Bjorn-Andersen, and R. O'Callaghan (eds.), 1-19. John Wiley & Sons Ltd. OECD. (2001). OECD in Figures 2001: Statistics on the Member Countries. Orlikowski, W.J. (1991). Integrated Information Environment or Matrix of Control? The Contradictory Implications of Information Technology. Accounting, Management & Information Technology, 9-42. Orlikowski, W.J. (1992). The Duality of Technology: Rethinking the Concept of Technology in Organizations. Organization Science, 398427. Orlikowski, W.J. & Baroudi, J.J. (1991). Studying Information Technology in Organizations: Research Approaches and Assumptions. Information Systems Research, 1-28. Osterlind, S.J. (1983). Test Item Bias. SAGE Publications, Newbury Park. Pennings, J.M. (1987). On the Nature of Technology as Organizational Innovation. In: New Technology as Organizational Innovation, J.M. Pennings and A. Buitendam (eds.), 3-12. Ballinger Publishing Company, Cambridge, Massachusetts. Pfeiffer, H.K.C. (1992). The Diffusion of Electronic Data Interchange. Physica-Verlag, Heidelberg. Picot, A., Bortenlänger, C., & Röhrl, H. (1997). Organization of Electronic Markets: Contributions from the New Institutional Economics. The Information Society 107-123. Pinsonneault, A. & Kraemer, K.L. (1993). Survey Research Methodology in Management Information Systems: An Assessment. Journal of Management Information Systems, 75. Popper, K. (1972). Objective Knowledge. Oxford University Press, London. Porter, M.E. & Millar, V.E. (1985). How Information Gives You Competitive Advantage. Harvard Business Review, 149-160.
361
Pozzebon, M. & Pinsonneault, A. (2001). Structuration Theory in the IS Field: An Assessment of Research Strategies. 9th European conference on Information Systems. Bled, Slovenia, June 27-29, 2001, 205-217. Prahalad, C.K. & Hamel, G. (1990). The Core Competence of the Corporation. Harvard Business Review, 79-90. Premkumar, G., Ramamurthy, K., & Crum, M. (1997) Determinants of EDI Adoption in the Transportation Industry. European Journal of Information Systems, 107-121. Premkumar, G. & Ramamurthy, K. (1995). The Role of Interorganizational and Organizational Factors on the Decision Mode for Adoption of Interorganizational Systems. Decision Sciences, 303-336. Premkumar, G., Ramamurthy, K., & Nilakanta. (1994). Implementation of Electronic Data Interchange: An Innovation Diffusion Perspective. Journal of Management Information Systems, 157-177. Prescott, M.B. & Conger, S.A. (1995). Information Technology Innovations: A Classification by IT Locus of Impact and Research Approach. DATA BASE Advances, 20-41. Ragin, C.C. & Becker, H.S. (1992). What is a Case? Exploring the Foundations of Social Theory. Cambridge University Press, Cambridge. Rai, A. & Yakuni, R.P. (1996). A Structural Model for CASE Adoption Behavior. Journal of Management Information Systems, 205-234. Ramamurthy, K., Premkumar, G., & Crum, M. (1999). Organizational and Interorganizational Determinants of EDI Diffusion and Organizational Performance: A Causal Model. Journal of Organizational Computing and Electronic Commerce, 253-285. Raymond, L. & Bergeron, F. (1996). EDI Success in Small and MediumSized Enterprises: A Field Study. Journal of Organizational Computing and Electronic Commerce, 161-172. Riggins, F.J., Kriebel, C.H., & Mukhopadhyay, T. (1994). The Growth of Interorganizational Systems in the Presence of Network Externalities. Management Science, 984-998. Riggins, F.J. & Mukhopadhyay, T. (1994). Interdependent Benefits from Interorganizational Systems: Opportunities for Business Partner Reengineering. Journal of Management Information Systems, 37.
362
Riggins, F.J. & Mukhopadhyay, T. (1999). Overcoming EDI Adoption and Implementation Risks. International Journal of Electronic Commerce, 103-123. Riggins, F.J. & Rhee, H.S. (1998). Toward a Unified View of Electronic Commerce. Communications of the ACM, 88-95. Rogers, E.M. (1995). Diffusion of Innovations. Fourth edition. The Free Press, New York. Rogers, E.M. (1983). Diffusion of Innovations. Third edition. The Free Press, New York. Romeo, A. (1977). The Rate of Imitation of a Capital-Embodied Process Innovation. Economics, 63-69. Ross, A. (1966). Om ret og retfærdighed. En indførelse i den analytiske retsfilosofi. Nyt Nordisk Forlag , København. Rotenberg, M. (1993). Communications Privacy: Implications for Network Design. Communications of the ACM, 61-69. Sabherwal, R. & King, W.R. (1995) An Empirical Taxonomy of the Decision-Making Processes Concerning Strategic Applicatins of Information Systems. Journal of Management Information Systems, 177-214. Sampler, J.L. (1996). Exploring the Relationship Between Information Technology and Organizational Science. In: Information Management The Organizational Dimension, M.J. Earl (ed.), 5-22. Oxford University Press. Schon, D.S. (1963). Champions for Radical New Inventions. Harvard Business Review, 77-86. Scott Poole, M. & Van De Ven, A.H. (1989). Toward a General Theory of Innovation Processes. In: Research on the Management of Innovation: The Minnesota Studies, A.H. Van de Ven, H.L. Angle, and M. Scott Poole (eds.), 637-662. Harper & Row, New York. Siegel, S. & Castellan, N.J. (1988). Nonparametric Statistics for the Behavioral Sciences. McGraw Hill, Boston. Silverman, D. (1993). Interpreting Qualitative Data. Methods for Analysing Talk, Text and Interaction. SAGE Publications, London.
363
Simon, H. (1977). The Process of Management Decision. In: The New Science of Management Decision, 39-81. Prentice-Hall, Englewood Cliffs, New Jersey. Slappendel, C. (1996). Perspectives on Innovation in Organisations. Organisation Studies, 107-129. Srinivasan, K., Kekre, S., & Mukhopadhyay, T. (1994). Impact of Electronic Data Interchange Technology on JIT Shipments. Management Science, 1291-1304. Statistics Denmark. (2000a). Nationalregnskabsstatistik. Statistics Denmark. (2000b). Statistisk tiårsoversigt 2000. Steinfield, C., Kraut, R., & Plummer, A. (1995) The Impact Of Interorganizational Networks On Buyer-Seller Relationships. Journal of Computer-Mediated Communication. Sutton, R.I. & Staw, B.M. (1995). What Theory is Not. Administrative Science Quarterly, 371-384. Swan, J., Newell, S., & Robertson, M. (2000). The Diffusion, Design and Social Shaping of Production Management Information Systems in Europe. Information, Technology & People, 27-45. Swanson, E.B. (1987). Information Systems in Organization Theory: A Review. In: Critical Issues in Information Systems Research, R.J.Boland and R.A.Hirschheim (eds.), 181-204. John Wiley & Sons Ltd. Swanson, E.B. (1994). Information Systems Innovation Among Organizations. Management Science, 1069-1092. Swanson, E.B. & Ramiller, N.C. (1997). The Organizing Vision in Information Systems Innovation. Organization Science, 458-474. Swatman, P.M.C. & Swatman, P.A. (1992). EDI System Integration: A Definition and Literature Survey. The Information Society, 169-205. Tabak, F. & Barr, S.H. (1999). Propensity to Adopt Technological Innovations: The Impact of Personal Characteristics and Organizational Context. Journal of Engineering and Technology Management, 247270. Teo, H.H., Tan, B.C.Y., & Wei, K.K. (1997). Organizational Transformation Using Electronic Data Interchange: The Case of TradeNet in Singapore. Journal of Management Information Systems, 139-165.
364
Thong, J.Y.L. (1999). An Integrated Model for Information Systems Adoption in Small Businesses. Journal of Management Information Systems, 187-214. Timmers, P. (1999). Electronic Commerce: Strategies And Models For Business-To-Business Trading. Wiley and Sons, New York. Tornatzky L.G. & Fleischer M. (1990). The Process of Technological Innovation. Lexington Books. Tornatzky, L.G. & Klein, K.J. (1982) Innovation Characteristics and Innovation Adoption-Implementation: A Meta-Analysis of Findings. IEEE Transactions on Engineering Management, 28-45. Truman, G.E. (2000). Integration in Electronic Exchange Environments. Journal of Management Information Systems, 209-244. Tuunainen, V.K. (1998). Opportuneties of Effective Integration of EDI for Small Businesses in the Automotive Industry. Information & Management, 361-375. Van De Ven, A.H. (1986). Central Problems in the Management of Innovation. Management Science, 590-607. Walsham, G. (1993). Interpreting Information Systems in Organizations. John Wiley & Sons, Chichester. Walsham, G. & Waema, T. (1994). Information Systems Strategy and Implementation: A Case Study of a Building Society. ACM Transactions on Information Systems, 150-173. Wang, E.T.G. & Seidmann, A. (1995). Electronic Data Interchange: Competitive Externalities and Strategic Implementation Policies. Management Science, 401-418. Weick, K.E. (1989). Theory Construction as Disciplined Imagination. Academy of Management Review, 516-531. Whetten, D.A. (1989). What Constitutes a Theoretical Contribution? Academy of Management Review, 490-495. Whittaker, J. (1990). Graphical Models in Applied Multivariate Statistics. John Wiley & Sons Ltd., Chichester. Wigand, R.T. (1997). Electronic Commerce: Definition, Theory, and Context. The Information Society, 1-16.
365
Williams, M.L. & Frolick, M.N. (2001). The Evolution of EDI for Competitive Advantage: The FEDEX Case. Information Systems Management, 47-53. Williamson, O.E. (1975). Markets and Hierarchies. Free Press, New York. Winograd, T. & Flores, F. (1987). Understanding Computers and Cognition. Addison-Wesley Publishing Company, New York. Wolfe, R.A. (1994). Organizational Innovation: Review, Critique and Suggested Research Directions. Journal of Management Studies, 405431. Yin, R.K. (1994) Case Study Research: Design and Methods. Sage Publishers. Zahle, H. (1989). Institutioner og regulering. Christian Ejlers Forlag, København. Zaltman, G., Duncan, R., & Holbek, J. (1973). Innovations and Organizations. John Wiley & Sons, New York. Zmud R.W. (1998). Conducting and Publishing Practice-Driven Research. In: Information Systems: Current Issues and Future Changes.Proceedings of the IFIP WG 8.2 and 8.3 joint working Conference on Information Systems. Helsinki, Finland, 10-13, December. 21-33. Zuboff, S. (1988). In the Age of the Smart Machine: the Future of Work and Power. Basic books. Zwass, V. (1996). Electronic Commerce: Structures and Issues. International Journal of Electronic Commerce, 3-23.
366
12 Appendixes
367
12.1 Appendix A: Questionnaire Common questions for all participating companies: 1 Which trade union is the company member of? (mark 1 box only)
xxxx
yyyy 2 Is the company? (mark 1 box only)
a subsidiary company (go to 3)
part of an industry group (go to 4)
a parent company (go to 5)
an independent enterprise (go to 5) 3 Is the parent company? (mark 1 box only)
Danish
International Please go to 5. 4 Is the industry group? (check 1 box only)
100% Danish
International 5 Which type of products does the company deal with/ produce? (mark 1 box only)
mainly batch production, e.g. ball bearings
mainly unique specimen, e.g. a custom-made machine
both batch production and unique specimen 6 How does the company view electronic commerce and EDI? (mark all that apply)
as a technical solution
as a business concept
as a tool for communication
as a fashion phenomenon
other, please specify: ________________________________________________________ The remainder of the questionnaire is divided into four sections: Section 1 is to be answered by companies that do not use EDI or electronic commerce. (Questions 7-17) Section 2 is to be answered by companies that use electronic commerce. (Questions 18-21) Section 3 is to be answered by companies that use EDI. (Questions 22-29a) Section 4 is concluding questions to all companies. (Questions 30-33)
Section 1: Questions to companies that do not use EDI or electronic commerce: 7 How many customers does the company approximately have? Number: ________ 8 How many suppliers does the company approximately have? Number: ________ 9 Does the company use any administrative systems, e.g. SAP or Navision? (check 1 box only)
yes
no
don’t know
368
Definition of electronic commerce: The term electronic commerce (e-commerce) is in this survey defined to mean electronic exchange of documents, goods or services. But, it can also be digital exchange of goods and services e.g. software and information from databases. Regarding electronic commerce there are no requirements for a specific structure for messages. And, electronic commerce can be carried out without any previous agreements on which messages are to be exchanged and how. 10 Has the company considered to adopt e-commerce? (mark 1 box only)
yes
no (if no go to 13) 11 Is e-commerce a part of the company’ business strategy? (mark 1 box only)
yes
no
don’t know 12 What is the main purpose of using e-commerce? (please feel free to mark more boxes)
marketing of goods and services
sale of goods and services
purchase of goods and services
service e.g. product-manuals and update of product-catalogues
other, please specify: ________________________________________________________ 13 What prevents the company from adopting e-commerce? (mark one box on a scale from 1 to 7, where 1 corresponds to great importance and 7 corresponds to no importance) 1 2 3 4 5 6 7 Technical solutions are not sufficient 1 2 3 4 5 6 7 Technical solutions are too expensive 1 2 3 4 5 6 7 It is not expected that gains will be achieved 1 2 3 4 5 6 7 Security 1 2 3 4 5 6 7 Competition does not demand it 1 2 3 4 5 6 7 The company’ activities are not suitable for e-commerce 1 2 3 4 5 6 7 The company does not feel ready for e-commerce 1 2 3 4 5 6 7 The employees are not ready for change 1 2 3 4 5 6 7 We have not considered the idea 1 2 3 4 5 6 7 We are doing well without e-commerce 1 2 3 4 5 6 7 It is not worth trying 1 2 3 4 5 6 7 None business-partners have demanded e-commerce 1 2 3 4 5 6 7 Nobody has recommended e-commerce 1 2 3 4 5 6 7 Other, specify: Definition of EDI: The term EDI is in this survey defined to mean exchange of structured, electronic messages. This exchange is conducted with a minimum of human interaction. A requirement for defining an electronic exchange as EDI is that messages are exchanged in a beforehand agreed standard. This format can be an individual proprietary standard or an international standard e.g. EDIFACT. 14 Has the company considered to adopt EDI? (mark 1 box only)
yes
no (if no go to 16)
369
15 Which facts are of greatest importance for the company to adopt EDI? (mark one box on a scale from 1 to 7, where 1 corresponds to great importance and 7 corresponds to no importance) 1 2 3 4 5 6 7 The technical solutions had reached a satisfactory technical level 1 2 3 4 5 6 7 The technical solutions had reached a satisfactory price level 1 2 3 4 5 6 7 Possibilities for savings 1 2 3 4 5 6 7 The activities of the company were well suited for using EDI 1 2 3 4 5 6 7 To provide the employees with a better working-environment 1 2 3 4 5 6 7 The company anticipated personal development prospects for employees 1 2 3 4 5 6 7 Possibilities for change of work-routines 1 2 3 4 5 6 7 The company felt well prepared for adopting EDI 1 2 3 4 5 6 7 Competitiveness 1 2 3 4 5 6 7 Opportunities for increasing the company’s market share 1 2 3 4 5 6 7 Several business partners were using EDI 1 2 3 4 5 6 7 The company was forced to adopt EDI 1 2 3 4 5 6 7 We did not reckon that we could do without EDI 1 2 3 4 5 6 7 It was new and exiting 1 2 3 4 5 6 7 Others had recommended EDI 1 2 3 4 5 6 7 Other, please specify: 16 What prevents the company from adopting EDI? (mark one box on a scale from 1 to 7, where 1 corresponds to great importance and 7 corresponds to no importance) 1 2 3 4 5 6 7 The technical solutions have not reached a satisfactory technical level 1 2 3 4 5 6 7 The technical solutions have not reached a satisfactory price level 1 2 3 4 5 6 7 No possibilities for savings 1 2 3 4 5 6 7 The activities of the company are not well suited for using EDI 1 2 3 4 5 6 7 EDI does not provide the employees with a better working-environment 1 2 3 4 5 6 7 The company does not anticipate personal development prospects for employees 1 2 3 4 5 6 7 No possibilities for change of work-routines 1 2 3 4 5 6 7 The company does not feel well prepared for adopting EDI 1 2 3 4 5 6 7 No competitiveness 1 2 3 4 5 6 7 No opportunities for increasing the company’s market share 1 2 3 4 5 6 7 None business partners are using EDI 1 2 3 4 5 6 7 The company has not been forced to adopt EDI 1 2 3 4 5 6 7 We are doing well without EDI 1 2 3 4 5 6 7 It is not new and exiting 1 2 3 4 5 6 7 Nobody have recommended EDI 1 2 3 4 5 6 7 Other, please specify: 17 Have you been encouraged to adopt EDI? (please feel free to mark more than one box)
yes, from customers
yes, from suppliers
yes, from other business partners, e.g. financial institutions or transportation companies
don’t know End of section 1. Please go to question 30.
370
Section 2: Questions to companies that use electronic commerce: Definition of electronic commerce: The term electronic commerce (e-commerce) is in this survey defined to mean electronic exchange of documents, goods or services. But, it can also be digital exchange of goods and services e.g. software and information from databases. Regarding electronic commerce there are no requirements for a specific structure for messages. And, electronic commerce can be carried out without any previous agreements on which messages are to be exchanged and how. 18 Which year did the company start to use e-commerce? Year:______ 19 How does the company use e-commerce? (please don’t hesitate to mark more boxes)
marketing of goods and services
sale of goods and services
purchase of goods and services
service e.g. product-manuals and update of product-catalogues
other, please specify: ________________________________________________________ 20 Why did the company choose to adopt e-commerce? (mark one box on a scale from 1 to 7, where 1 corresponds to great importance and 7 corresponds to no importance) 1 2 3 4 5 6 7 Technical solutions were sufficient 1 2 3 4 5 6 7 Technical solutions had reached an affordable level 1 2 3 4 5 6 7 It was expected that gains would be achieved 1 2 3 4 5 6 7 Security had reached a acceptable level 1 2 3 4 5 6 7 Competition demanded it 1 2 3 4 5 6 7 The company’ activities are suitable for e-commerce 1 2 3 4 5 6 7 The company felt ready for e-commerce 1 2 3 4 5 6 7 The employees were ready for the change 1 2 3 4 5 6 7 It gave new work-challenges for the employees 1 2 3 4 5 6 7 It was interesting to be upfront 1 2 3 4 5 6 7 It was worth trying 1 2 3 4 5 6 7 Our business-partners demanded us to use e-commerce 1 2 3 4 5 6 7 We had e-commerce recommended 1 2 3 4 5 6 7 Other, specify: 21 How do you perceive the benefits of e-commerce for the company? (mark one box on a scale from 1 to 7, where 1corresponds to I do fully agree and 7 corresponds to I disagree, please mark the options) 1 2 3 4 5 6 7 The company has gained direct savings 1 2 3 4 5 6 7 Better service towards customers and suppliers have been gained 1 2 3 4 5 6 7 Work-routines have been eased 1 2 3 4 5 6 7 e-commerce has reduced employees’ routine work 1 2 3 4 5 6 7 The company has gained more new customers 1 2 3 4 5 6 7 The company has gained a better market position 1 2 3 4 5 6 7 The relations towards businesspartners have become closer 1 2 3 4 5 6 7 e-commerce has supported new ways of doing business 1 2 3 4 5 6 7 The image of the company has become more innovative End of section 2. If the company uses EDI please continue with question 22, otherwise please go to question 30.
371
Section 3: Questions to companies that use EDI: Definition EDI: The term EDI is in this survey defined to mean exchange of structured, electronic messages. This exchange is conducted with a minimum of human interaction. A requirement for defining an electronic exchange as EDI is that messages are exchanged in a beforehand agreed standard. This format can be an individual proprietary standard or an international standard e.g. EDIFACT. 22 Which year did the company start to use EDI? Year:______ 23 What was the motivation for adopting EDI? (mark one box on a scale from 1 to 7, where 1 corresponds to great importance and 7 corresponds to no importance) 1 2 3 4 5 6 7 The technical solutions had reached a satisfactory technical level 1 2 3 4 5 6 7 The technical solutions had reached a satisfactory price level 1 2 3 4 5 6 7 Possibilities for savings 1 2 3 4 5 6 7 The activities of the company were well suited for using EDI 1 2 3 4 5 6 7 To provide the employees with a better working-environment 1 2 3 4 5 6 7 The company anticipated personal development prospects for employees 1 2 3 4 5 6 7 Possibilities for change of work-routines 1 2 3 4 5 6 7 The company felt well prepared for adopting EDI 1 2 3 4 5 6 7 Competitiveness 1 2 3 4 5 6 7 Opportunities for increasing the company’s market share 1 2 3 4 5 6 7 Several business partners were using EDI 1 2 3 4 5 6 7 The company was forced to adopt EDI 1 2 3 4 5 6 7 We did not reckon that we could do without EDI 1 2 3 4 5 6 7 It was new and exiting 1 2 3 4 5 6 7 Others had recommended EDI 1 2 3 4 5 6 7 Other, please specify: 24 How do you perceive the benefits for the company of using EDI? (mark one box on a scale from 1 to 7, where 1corresponds to I do fully agree and 7 corresponds to I disagree, please mark the options) 1 2 3 4 5 6 7 The company has gained direct savings 1 2 3 4 5 6 7 Better service towards customers and suppliers have been gained 1 2 3 4 5 6 7 Work-routines have been eased 1 2 3 4 5 6 7 EDI has reduced employees’ routine work 1 2 3 4 5 6 7 The company has gained a better market position 1 2 3 4 5 6 7 The relations towards businesspartners have become closer 1 2 3 4 5 6 7 EDI has supported new ways of doing business 1 2 3 4 5 6 7 The image of the company has become more innovative 25 How many customers does the company use EDI with at the moment? Number: ______ 26 How many customers does the company have in total? Number:______ 27 How many suppliers does the company use EDI with at the moment? Number:______ 28 How many suppliers does the company have in total? Number:______
372
Questions 29 and 29a are only to be answered by subsidiary companies or companies part of industry groups. (Otherwise please go to 30) 29 Does the parent company or the industry group use EDI? (mark 1 box only)
yes
no
don’t know 29a. The company’ use of EDI is a result of? (mark 1 box only)
an invitation from the parent company/ the industry group
a recommendation from the parent company/ the industry group
a direct order from the parent company/ the industry group
don’t know
Section 4: Concluding questions to all companies: 30 Does the company exchange electronic data with? (please feel free to mark more boxes)
financial institutions
public authorities, e.g. tax, VAT or salary information
transportation companies
others, please specify: 31 Has the company participated in information seminars on EDI and/ or e-commerce?
yes
no
don’t know 32 Is the company member of the Danish EDI-council?
yes
no
don’t know 33 Is the company interested in participating in information seminars on EDI and/ or ecommerce?
yes
no Thank you very much for your help. Please return the questionnaire to xxx no later than xxx Use the enclosed envelope to: xxx Attention.: xxx xxx xxx Or telefax to: xxx. Attention: xxx
373
12.2 Appendix B: Statistical tables and figures118 12.2.1 Tables belonging to Chapter 2 Table 2-3. Test for possible non-response bias Frequency Percent Row Pct Col Pct ADOPTERS YES
NO
Total
RESPONSE-TIME IN DAYS 1-5
6-9
10-27
Total
30
11
17
58
51.7%
19.0%
29.3%
28.9%
32.6%
25.6%
25.8%
62
32
49
143
43.4%
22.4%
34.3%
71.1%
67.4%
74.4%
74.2%
92
43
66
45.8%
21.4%
32.8%
201
Chi square = 1.16 Degrees of freedom = 2 p value = 0.55868195
118
Additional statistical runs can be found on the following web-site: http://www.cbs.dk/staff/hzh/Publications.htm
374
Table 2-4. Contingency table showing geographical location and number of employees for all manufactures in the steel and machinery in Denmark based on figures for the year 2001 provided by the manufactures business association Zip_code
Employee
Frequency ‚ Percent ‚ Row Pct ‚ Col Pct ‚1-5 ‚5-9 ‚10-19 ‚20-49 ‚50-99 ‚100+ ‚ Total ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆ Sjælland,Lolland ‚ 39 ‚ 31 ‚ 37 ‚ 40 ‚ 10 ‚ 7 ‚ 164 -Falster ‚ 5.64 ‚ 4.49 ‚ 5.35 ‚ 5.79 ‚ 1.45 ‚ 1.01 ‚ 23.73 ‚ 23.78 ‚ 18.90 ‚ 22.56 ‚ 24.39 ‚ 6.10 ‚ 4.27 ‚ ‚ 24.68 ‚ 26.05 ‚ 21.76 ‚ 25.00 ‚ 20.41 ‚ 20.00 ‚ ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆ Fyn og Øerne ‚ 18 ‚ 22 ‚ 23 ‚ 15 ‚ 9 ‚ 4 ‚ 91 ‚ 2.60 ‚ 3.18 ‚ 3.33 ‚ 2.17 ‚ 1.30 ‚ 0.58 ‚ 13.17 ‚ 19.78 ‚ 24.18 ‚ 25.27 ‚ 16.48 ‚ 9.89 ‚ 4.40 ‚ ‚ 11.39 ‚ 18.49 ‚ 13.53 ‚ 9.38 ‚ 18.37 ‚ 11.43 ‚ ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆ Jylland ‚ 101 ‚ 66 ‚ 110 ‚ 105 ‚ 30 ‚ 24 ‚ 436 ‚ 14.62 ‚ 9.55 ‚ 15.92 ‚ 15.20 ‚ 4.34 ‚ 3.47 ‚ 63.10 ‚ 23.17 ‚ 15.14 ‚ 25.23 ‚ 24.08 ‚ 6.88 ‚ 5.50 ‚ ‚ 63.92 ‚ 55.46 ‚ 64.71 ‚ 65.63 ‚ 61.22 ‚ 68.57 ‚ ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆ Total 158 119 170 160 49 35 691 22.87 17.22 24.60 23.15 7.09 5.07 100.00
Statistics for Table of zip_code by employee Statistic DF Value Prob ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ Chi-Square 10 8.4768 0.5824 Likelihood Ratio Chi-Square 10 8.3575 0.5940 Mantel-Haenszel Chi-Square 1 0.5689 0.4507 Phi Coefficient 0.1108 Contingency Coefficient 0.1101 Cramer's V 0.0783
375
Table 2-5. Contingency table showing geographical location and number of employees for the survey sample of manufactures in the steel and machinery in Denmark ZIP_CODE
EMPLOYEE
Frequency ‚ Percent ‚ Row Pct ‚ Col Pct ‚1-9 ‚10-19 ‚20-49 ‚50+ ‚ Total ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆ Sjælland,Lolland ‚ 15 ‚ 8 ‚ 19 ‚ 5 ‚ 47 -Falster ‚ 9.09 ‚ 4.85 ‚ 11.52 ‚ 3.03 ‚ 28.48 ‚ 31.91 ‚ 17.02 ‚ 40.43 ‚ 10.64 ‚ ‚ 27.78 ‚ 24.24 ‚ 38.78 ‚ 17.24 ‚ ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆ Fyen og Øerne ‚ 6 ‚ 4 ‚ 4 ‚ 6 ‚ 20 ‚ 3.64 ‚ 2.42 ‚ 2.42 ‚ 3.64 ‚ 12.12 ‚ 30.00 ‚ 20.00 ‚ 20.00 ‚ 30.00 ‚ ‚ 11.11 ‚ 12.12 ‚ 8.16 ‚ 20.69 ‚ ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆ Jylland ‚ 33 ‚ 21 ‚ 26 ‚ 18 ‚ 98 ‚ 20.00 ‚ 12.73 ‚ 15.76 ‚ 10.91 ‚ 59.39 ‚ 33.67 ‚ 21.43 ‚ 26.53 ‚ 18.37 ‚ ‚ 61.11 ‚ 63.64 ‚ 53.06 ‚ 62.07 ‚ ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆ Total 54 33 49 29 165 32.73 20.00 29.70 17.58 100.00 Frequency Missing = 5 Statistics for Table of POSTNUM by EMPLOYEE Statistic DF Value Prob ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ Chi-Square 6 6.2545 0.3953 Likelihood Ratio Chi-Square 6 6.0763 0.4147 Mantel-Haenszel Chi-Square 1 0.0004 0.9834 Phi Coefficient 0.1947 Contingency Coefficient 0.1911 Cramer's V 0.1377 Effective Sample Size = 165
376
12.2.2 Tables and Figures belonging to Chapter 8 12.2.2.1 Abbreviations used in analyses of motivation for adoption of EDI Table 8-1. Definition of adoption levels Adoption level ADO PLA NON
Adoption level defined Adopter Planner Non-adopter
Table 8-2. Definition of research constructs Construct ORG TEC ENV
Construct defined Organizational context Technological context Environmental context
Table 8-3. Opinion data items and the respective questions from the survey instrument Item 1/ A 2/ B 3/ C 4/ D 5/ E 6/ F
Proposition T1 T2 O1 O5 O2 O3
7/ G 8/ H 9/ I 10/ J 11/ K 12/ L 13/ M 14/ N 15/ O
O4 O6 E1 E2 E3 E5 T3 T4 E4
Definition of opinion data item questions The technical solutions had reached a satisfactory technical level The technical solutions had reached a satisfactory price level Possibilities for savings The activities of the company were well suited for using EDI To provide the employees with a better working-environment The company anticipated personal development prospects for employees Possibilities for change of work-routines The company felt well prepared for adopting EDI Competitiveness Opportunities for increasing the company’s market share Several business partners were using EDI The company was forced to adopt EDI We did not reckon that we could do without EDI It was new and exiting Others had recommended EDI
377
Table 8-4. Adoption level versus company size EDI_STAT by EMPLOYEE Frequency ‚ Percent ‚ Row Pct ‚ Col Pct ‚1-5 ‚6-9 ‚10-19 ‚20-49 ‚50-99 ‚100+ ‚ Total ƒƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆ non-adopters ‚ 24 ‚ 22 ‚ 19 ‚ 32 ‚ 9 ‚ 2 ‚ 108 ‚ 12.31 ‚ 11.28 ‚ 9.74 ‚ 16.41 ‚ 4.62 ‚ 1.03 ‚ 55.38 ‚ 22.22 ‚ 20.37 ‚ 17.59 ‚ 29.63 ‚ 8.33 ‚ 1.85 ‚ ‚ 82.76 ‚ 70.97 ‚ 50.00 ‚ 55.17 ‚ 36.00 ‚ 14.29 ‚ ƒƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆ planners ‚ 3 ‚ 5 ‚ 11 ‚ 16 ‚ 10 ‚ 5 ‚ 50 ‚ 1.54 ‚ 2.56 ‚ 5.64 ‚ 8.21 ‚ 5.13 ‚ 2.56 ‚ 25.64 ‚ 6.00 ‚ 10.00 ‚ 22.00 ‚ 32.00 ‚ 20.00 ‚ 10.00 ‚ ‚ 10.34 ‚ 16.13 ‚ 28.95 ‚ 27.59 ‚ 40.00 ‚ 35.71 ‚ ƒƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆ adopters ‚ 2 ‚ 4 ‚ 8 ‚ 10 ‚ 6 ‚ 7 ‚ 37 ‚ 1.03 ‚ 2.05 ‚ 4.10 ‚ 5.13 ‚ 3.08 ‚ 3.59 ‚ 18.97 ‚ 5.41 ‚ 10.81 ‚ 21.62 ‚ 27.03 ‚ 16.22 ‚ 18.92 ‚ ‚ 6.90 ‚ 12.90 ‚ 21.05 ‚ 17.24 ‚ 24.00 ‚ 50.00 ‚ ƒƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒˆ Total 29 31 38 58 25 14 195 14.87 15.90 19.49 29.74 12.82 7.18 100.00 Frequency Missing = 52 Statistics for Table of EDI_STAT by EMPLOYEE Statistic DF Value Prob ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ Chi-Square 10 28.4452 0.0015 Likelihood Ratio Chi-Square 10 29.0127 0.0012 Mantel-Haenszel Chi-Square 1 20.0441 |r| under H0: Rho=0
Adopters
ITEM01
TEC
TEC
ORG
ORG
ORG
ORG
ORG
ORG
ENV
ENV
ENV
ENV
TEC
TEC
ENV
ITEM01
ITEM02
ITEM03
ITEM04
ITEM05
ITEM06
ITEM07
ITEM08
ITEM09
ITEM10
ITEM11
ITEM12
ITEM13
ITEM14
ITEM15
ITEM02
1.00000 0.564***
1.00000
ITEM03
0.343*
0.325
ITEM04
0.433*
0.106
1.00000 0.511**
1.00000
ITEM05
0.341*
0.222
0.454**
0.229
ITEM06
0.258
0.308
ITEM07
-0.048
0.273
0.188
0.203
1.00000
ITEM08
0.325 0.588***
0.322 0.536**
0.363*
0.503**
0.416*
0.312
0.108
0.480**
1.00000
ITEM09
-0.346*
0.053
0.083
0.086
-0.047
0.048
-0.093
-0.132
ITEM10
-0.325
-0.098
0.172
0.158
-0.003
-0.005
0.032
1.00000 -0.016 0.884***
ITEM11
-0.173
-0.294
-0.105
-0.191
-0.289
-0.378*
0.104
-0.243
ITEM12
-0.465**
-0.130
-0.109
-0.313
-0.091
0.055
-0.141
-0.354*
0.118 0.510**
ITEM13
0.082
0.094
0.296
-0.151
-0.197
0.293
0.204
ITEM14
0.344*
-0.091 0.518**
0.031
0.161
0.374*
-0.249
ITEM15
0.194
0.082
0.008
0.129
0.293
0.325 0.507**
0.109
1.00000 0.164 0.697***
1.00000
0.187
1.00000 0.491**
0.387*
0.473** 0.519**
0.293
0.210
1.00000
0.231
0.288
0.275
-0.199
0.092
0.018
1.00000
0.107
0.045
0.051
-0.234
0.212
0.013
0.427*
Legend: TEC = Technological context, ORG = Organizational context, ENV = Environmental context * p
