The Critical Success Factors For Erp Implementation An Organisational Fit Perspective

Information & Management 40 (2002) 25±40The critical success factors for ERP implementation:an organizational ®t perspectiveKyung-Kwon Hong, Young-Gul Kim*Graduate School of Management, Korea Advanced Institute of Science and Technology (KAIST),207-43 Cheongryangri-Dong, Dongdaemun-Gu, Seoul 130-012, South KoreaReceived 20 May 2001; accepted 1 September 2001AbstractSince early 1990s, many ®rms around the world have shifted their information technology (IT) strategy from developinginformation systems in-house to purchasing application software such as enterprise resource planning (ERP) systems. ITmanagers responsible for managing their organization's ERP implementation view their ERP systems as their organizations'most strategic computing platform. However, despite such strategic importance, ERP projects report an unusually high failurerate, sometimes jeopardizing the core operations of the implementing organization. This study explores the root of such highfailure rate from an ``organizational ®t of ERP'' perspective. Based on the relevant literature, we de®ne the concept oforganizational ®t of ERP and examine its impact on ERP implementation, together with ERP implementation contingencies.The results from our ®eld survey of 34 organizations show that ERP implementation success signi®cantly depends on theorganizational ®t of ERP and certain implementation contingencies.# 2002 Elsevier Science B.V. All rights reserved.Keywords: ERP implementation; Organizational ®t; ERP adaptation; Process adaptation; Organizational resistance1. Introductionquality [28]. Although application packages have thesebene®ts over custom design of applications, packagedUnder the pressure to proactively deal with thesoftware has problems of their own: uncertainty in ac-radically changing external environment, many ®rmsquisition [14] and hidden costs in implementation [29].have changed their information system (IS) strategiesIn a survey of the IT managers responsible forby adopting application software packages rather thanmanaging their organization's ERP projects, two-in-house development [12,25]. An application pack-thirds of the respondents viewed their ERP systemsage such as enterprise resource planning (ERP) systemas their organizations' most strategic computing plat-is one solution to the information technology (IT)form [47]. Despite such perceived importance, it wasindustry's chronic problems of custom system design:reported that three quarters of the ERP projects werereduced cost, rapid implementation, and high systemjudged to be unsuccessful by the ERP implementing®rms [13]. What makes ERP implementation so*unsuccessful? Swan et al. [46] argued that the rootCorresponding author. Tel.: ‡82-2-958-3614;of such high failure rate is the difference in interestsfax: ‡82-2-958-3604.E-mail addresses: kyungkh@kgsm.kaist.ac.kr (K.-K. Hong),between customer organizations who desire uniquedomino2@unitel.co.kr (Y.-G. Kim).business solutions and ERP vendors who prefer a0378-7206/02/$ ± see front matter # 2002 Elsevier Science B.V. All rights reserved.PII: S 0 3 7 8 - 7 2 0 6 ( 0 1 ) 0 0 1 3 4 - 326K.-K. Hong, Y.-G. Kim / Information & Management 40 (2002) 25±40generic solution applicable to a broad market. Suchof technology, task, people, structure, and culture [7].con¯icting interests led us to explore an organizationalThus organizational resistance to change is identi-®t perspective of ERP implementation.®ed as a critical success factor for ERP implementa-An important criterion used in selecting an ERPtion [26,34]. In this study, we de®ne the concept ofsystem is the ERP ®t with the current businessorganizational ®t of ERP and empirically examine itsprocesses [9]. Although the ®t between ERP andimpact on ERP implementation success along with thethe organizational context is believed to be criticalmoderating roles of ERP implementation contingencyfor successful ERP implementation, few examined thevariables such as ERP adaptation, process adaptation,organizational ®t issues of ERP empirically. Soh et al.and organizational resistance.[44] suggested that the organizational ®t of ERPThis paper is organized in ®ve sections. First,might be worse in Asia, because the reference processERP related literatures are reviewed. The next sectionmodel underlying most ERP systems is in¯uenced byintroduces the research model and hypotheses.European or US industry/business practices, which areResearch methodology is then described, followeddifferent from Asian business practices.by the presentation of the results. The paper concludesThe relative invisibility of the ERP implementationwith the discussion of the research ®ndings and impli-process is also identi®ed as a major cause of ERPcations for future research and practice.implementation failures [13]. Markus and Robey [33]attributed such invisibility to the unpredictably com-plex social interaction of IT and organization. The2. Theoretical perspectivescritical challenge of ERP implementation is believedto be the mutual adaptation between the IT and user2.1. Organizational ®t of ERPenvironment [49]. Such mutual adaptation processbrings the organization's existing operating processesBecause of the multiplicity of the organizationaland the packaged software's embedded functionalitydimension, researchers studying IS contingenciesinto alignment through a combination of softwarehave typically focused on the ®t between speci®ccon®guration and organizational change [49]. Butorganizational dimension and IS [21,23]. In a reviewthere are con¯icting views on which type of adapta-of the IS contingency research, Weil and Olsontion, package adaptation or organizational adaptation,[51] found that over seventy percent of the studiesis more desirable in the different contexts.followed a model assuming that the better the ®tERP diffusion agencies including ERP vendors andamong the contingency variables, the better the per-consulting ®rms recommend strongly that ERP pro-formance. They categorized the contingency variablesjects embody the universally applicable `best practice'of interest to IS researchers into strategy, structure,and should be implemented without extensive adapta-size, environment, technology, task, and individualtion of the packaged software [1]. In contrast, somecharacteristics.academics maintain that the notion of `best practice' isAttributing the inability to realize value from ITillusory and potentially disruptive because ERP doesinvestments to lack of alignment between the businessnot provide models for every process of every industryand IT strategies, Henderson and Venkatraman [19]and most ®rms usually recon®gure or add new func-developed the `strategic alignment model', emphasiz-tionality to ERP systems for optimal use within theiring the multivariate ®t among business strategy, ITunique context [46].strategy, organizational infrastructure and processes,Besides, since ERP philosophy is process-based,and IT infrastructure and processes. With an explora-rather than function-based, ERP necessitates disrup-tory survey of small business, Marius and Ashok [30]tive organizational changes [17,49,50]. Successfulhypothesized that packaged software implementationERP implementation must be managed as a programsuccess is positively associated with the degree ofof wide-ranging organizational change initiativesvendor ®t with user organization and the degree ofrather than as a software installation effort [18]. Suchsoftware ®t with user organization, respectively.IT-driven initiatives require change of the organiza-In ERP research, Gattiker and Goodhue [10] sug-tion's socio-technical system, which is intertwinedgested that while inter-dependences among sub-unitsK.-K. Hong, Y.-G. Kim / Information & Management 40 (2002) 25±4027give rise to better ®t of ERP with global operationto the package [7]. ERP implementation also involvesneeds, differentiation among sub-units develop poorbusiness process change (BPC) and ERP adaptation to®t of ERP with local operation needs. Soh et al. [44]align the software with the business processes [20].also suggest that ERP mis®t stems from the ®rm-, orWhich direction is desirable depends on one's pointcountry-speci®c requirements that do not match theof view and various implementation contingenciescapabilities of ERP. Swan et al. [46] argue that orga-[27,40].nizational mis®ts of ERP exist due to the con¯ictingTypically, ERP vendors recommend process adap-interests of user organization and ERP vendors. Thus,tation and discourage ERP adaptation for fear ofthe concept of organizational ®t seems to be the corepotential performance and integrity degradation asresearch construct to explain the implementation suc-well as maintenance and future upgrade dif®culties.cess in diverse IT implementation contexts. In the ERPOn the other hand, the user departments of the cus-implementation context, however, the impact of orga-tomer organization would prefer ERP adaptation tonizational ®t of ERP was not empirically tested. In thisprocess adaptation which will necessitate the signi®-paper, based on Markus and Robey [32] de®nition, wecant changes in their work environment.de®ne organizational ®t of ERP as the congruencebetween the original artifact of ERP and its organiza-tional context and will examine its impact on ERP3. Research model and hypothesesimplementation success empirically.This study examines the relationship between the2.2. ERP implementation contingenciesorganizational ®t of ERP and ERP implementationsuccess. The effects of three implementation contin-Implementation of a technical innovation is viewedgencies on this relationship are explored as moderat-as a dynamic process of mutual adaptation, that is:ing variables. The research model is illustrated in``re-invention of the technology and simultaneousFig. 1 and discussed below.adaptation of the organization'' [27]. Adaptationmay address the user' procedures, assumption, knowl-3.1. ERP implementation successedge, or relationships as well as physical aspects of thetechnology [48]. In the adaptation stage of IT imple-While some ®rms announced ERP implementationmentation model, an IT application is developed,success, many others reported negative results of ERPinstalled, and maintained; organizational proceduresimplementation [5,6]. An ERP implementation failureare revised and developed; organization members aremay be fatal to a ®rm: either wasting enormous sumstrained both in the new procedures and in the new ITof money or destroying the competitive advantage ofapplication [4]. The adaptation concept is applicablethe ®rm [5,6]. To manage ERP implementation suc-not only to custom software but also to off-the-shelfcessfully, a high level ERP implementation successpackages [34]. Most studies on the implementation ofmeasure is required [39]. Markus and Tanis [34]application software packages emphasize the criticalargued that a minimum set of success metrics includesnature of the adaptation process [10,14,28].project metrics, early operational metrics, and long-From the duality perspective of technology, ERP isterm business results. In this study, we de®ne ERPa ``technological artifact bundling material and sym-implementation success from the ERP implementationbolic properties in some socially recognizable formproject perspective.(e.g. hardware, software, practice)'' [36]. The adapta-While there is a wide range of con®gurations avail-tion of ERP and organizational processes is an iterativeable in any major ERP product, ERP is frequentlyprocess entailing on-going social action that is clearlyunable to model some of the adopting ®rm's existingconstrained by both the structural properties of theprocedures [10]. A critical challenge in ERP imple-organization and the built-in properties of the ERP [50].mentation has to do with ®rst identifying gaps betweenThere are two alternative approaches to the imple-the ERP generic functionality and the speci®c organi-mentation of a packaged software: package adaptationzational requirement, and then deciding how theseto organizational needs and organizational adaptationgaps will be handled [1,44,49,50]. For the successful28K.-K. Hong, Y.-G. Kim / Information & Management 40 (2002) 25±40Fig. 1. Research model.implementation of ERP, organizational ®t of ERPadaptation, identifying eight ERP adaptation types.is important, because organizational mis®t of ERPGlass [11] categorized ERP adaptation types intorequires massive changes in the adopting organization'scustomization, extension, and modi®cation. ERP cus-business process, ERP systems, or both [37]. Besides,tomization (also called con®guration) is to choosewhile there must be the minimum organizational ®tamong the reference processes and set the parameterslevel which ERP implementation is destined to failin ERP without changing the ERP source code. ERPregardless of whatever adaptations, none of the ERPextensions utilize the ``user exit'' function for localvendors acknowledge the fact or attempts to assesscode, a specialized programming language and third-the implementation feasibility based on the organiza-party bolt-on software to ®ll the gap between ERPtional ®t of ERP. Laughlin [26] regarded matchingfunctionality and organizational requirements. In con-functionality with ERP as one of the major factorstrast, ERP modi®cation changes the ERP source code.for ERP implementation. This leads to our ®rstIn this paper, we restrict ERP adaptation to extensionhypothesis.and modi®cation because customization does notchange the basic ERP identity.Hypothesis 1. Organizational fit of ERP is positivelyIn base relation, we hypothesize that organizationalrelated to ERP implementation success.®t of ERP is positively related to ERP implementationsuccess. In the case of low ERP adaptation, we expect3.2. ERP adaptationthat organizational ®t of ERP will be more stronglyassociated with ERP implementation success, becauseTechnological adaptation is referred to as thethe low level of ERP adaptation would not affect theadjustments and changes following the installationinitial organizational ®t of ERP value signi®cantly.of a new technology in a given setting [48]. WhileBut in the case of high ERP adaptation, we expect thatERP vendors presume that ERP embodies the uni-organizational ®t of ERP will not be as stronglyversally applicable best practices, most organizationsassociated with ERP implementation success as inadapt ERP to their unique organizational contextslow ERP adaptation because the high level of ERP[46]. ERP adaptation increases the feature-functionadaptation would have reduced the gap between ERP®t between ERP and the adopting organization, whichand organization, mitigating the effect of organiza-is likely to result in lower resistance, reduced trainingtional ®t of ERP. Thus, when the level of ERP adapta-needs, less organizational adaptation [2]. It is alsotion is low, we expect a stronger the base relationshipargued that re-invention or adaptation of an innovationand expect the opposite when the level of ERP adapta-might reduce mistakes resulting from the implemen-tion is high.tation uncertainty [40].Brehm et al. [3] contend that ERP implementationHypothesis 2. There is an interaction effect of thesuccess depends on the type and extent of ERPlevel of ERP adaptation on the relationship betweenK.-K. Hong, Y.-G. Kim / Information & Management 40 (2002) 25±4029organizational fit of ERP and ERP implementationimplementation. In a political perspective, Markussuccess.[31] explains resistance to IT implementation in termsof power distribution mis®t of IS. She noted that the3.3. Process adaptationpolitical perspective appears to be primarily applic-able for cross-functional IS. In an MRP study, CooperERP implementation may cause radical organiza-and Zmud [4] suggested that organizational resistancetional changes that need to be carefully managedand lack of MRP understanding had more explanatory[2,17,18]. Unlike the design and development of apower of inhibiting MRP infusion within its workcustom software, implementation of a packaged soft-environment than the task ®t of MRP.ware requires that its implementing organization adaptSince the organization and process changes inducedsome of its organizational processes to ®t the basicby ERP implementation force involuntary changesbusiness practices that are embedded in such applica-and frequently lead to different power and resourcetion packages [8,28]. When ERP implementationallocations, ERP implementation usually triggers ainvolves adapting the existing business processes todiverse group of overt and covert opponents within thethe standard business process of ERP, other organiza-organization. Based on this realization, we expect thattional components (e.g. organizational structure, mea-organizational resistance will moderate the relation-surement compensation, organizational culture,ship between organizational ®t of ERP and ERPtraining, etc.) and their interactions must also beimplementation success. When the level of organiza-changed together [17].tional resistance is low, we expect the stronger baseThe adaptation of organizational process in BPCrelationship and expect the opposite when the level ofliterature emphasizes the need to take account of theERP adaptation is high.management of organizational change. Grover et al.[15] empirically found that change management isHypothesis 4. There is an interaction effect of thethe most critical source of BPC implementation. Sinceorganizational resistance on the relationship betweenprocess adaptation induces the same effect on theorganizational fit of ERP and ERP implementationrelationship between organizational ®t of ERP andsuccess.implementation success as ERP adaptation, based onthe same reasoning for hypothesis two, when the levelof process adaptation is low, we expect a stronger4. Research methodrelationship and expect the opposite when the level ofprocess adaptation is high.4.1. Sample and data collectionHypothesis 3. There is an interaction effect of theThe target of this study was the organization thatlevel of process adaptation on the relationship betweenhas implemented ERP. We used the key informantorganizational fit of ERP and ERP implementationmethod for collecting information on a social settingsuccess.by interviewing (or surveying) a selected numberof participants. Fifty ®rms were identi®ed from the3.4. Organizational resistancelists provided by the ERP vendors. We contacted theERP project managers in charge of ERP implementa-ERP implementation will affect most of the com-tion in each ®rm. About 350 survey questionnairespany's business functions and in¯uence users directly.were sent to the ERP project manager of each ®rm,Resistance to a change stems from change in thewho forwarded our questionnaires to his/her projectjob content and uncertainty of the new system [22].team members in charge of individual process. In total,Any ERP project team will face a certain level ofone hundred and six questionnaires were collectedorganizational resistance due to its disruptive changefrom 34 ®rms. One case was dropped due to incom-[26]. Previous IS research, based on different theore-plete data entry. Two missing values, one item fortical perspectives, has made a substantial progressimplementation success and another item for ERPin understanding the resistance introduced by ISadaptation, were identi®ed and replaced with the30K.-K. Hong, Y.-G. Kim / Information & Management 40 (2002) 25±40average value of the rest of the items. On the average,4.4. Organizational ®t of ERPthree or four persons per ®rm participated in oursurvey (Appendix A).ERP technological artifacts are basically character-ized as integration of data and process within and4.2. Measurement developmentacross functions of an organization [6,24,34]. In atraditional software application perspective, Soh et al.We ®rst conducted a series of in-depth interviews[44] examined organizational ®t of ERP in terms ofwith various ERP project managers and members todata, process, and output. Considering that user inter-examine the external validity of our research model.face plays an increasingly critical role for the mission-We then developed the questionnaire items based oncritical enterprise systems such as ERP, we operatio-the literature and the ®eld visits, as well as thenalized the organizational ®t of ERP construct in termscomments gathered from the interviews. The surveyof data, process, and user interface ®t of ERP before orquestionnaire was revised through four rounds ofat the initial implementation period.line-by-line discussion with the help of businessprocess experts with signi®cant ERP project experi-4.5. Implementation contingenciesence. To pretest the reliability and validity of theinstruments, we conducted a pilot study, using the 32We identi®ed the type of ERP adaptation based onquestionnaire responses from six ®rms. Some itemsthe industry expert interviews and developed the ERPwere revised on the basis of the pilot result. Theadaptation measurement. We con®rmed that the ERPoperational de®nitions of the variables used areadaptation instrument was applicable to different ERPsummarized in Table 1.systems through a series of the project team's reviewsand pretests. Process adaptation and organizational4.3. ERP implementation successresistance measurements were developed and evalu-ated based on the relevant literature and validatedERP implementation success, the dependent vari-through a series of in-depth discussion with differentable in this study, is different from the traditional ISERP project teams.success measure in that it is not evaluated by thegeneral users but by the project team members. Imple-4.6. Characteristics of respondent ®rmsmentation project success is frequently de®ned in termsof the achievement of some predetermined goals,Respondent ®rms consist of 25 manufacturing ®rmswhich normally include multiple parameters such asand nine non-manufacturing ®rms, including servicetime, cost, function [34]. In this study, we measuredand distribution ®rms. Firms were using four ERPERP implementation success in terms of the perceivedproducts: SAP R/3 (14), UniERP (13), Oracle ERP (6),deviation from the expected project goals such asBann BPCS (1). Table 2 shows pro®les of the respon-cost overrun, schedule overrun, system performancedent ®rms in terms of annual revenue and number ofde®cit, and failure to achieve the expected bene®ts.total employees.Table 1Operational de®nitions of variablesVariableOperational definitionImplementation successaThe degree of deviation from project goal in terms of expected cost, time, system performance and benefitsOrganizational fit of ERPThe degree of alignment between ERP model and organization needs in terms of data, process and user interfaceERP adaptationThe extent of efforts and time spending in ERP alteration to align with organizational process needs except forERP customizationProcess adaptationThe extent of efforts and time spending in process change to align with ERPOrganizational resistanceThe strength of negative organizational response to ERP implementationa Reverse scale.K.-K. Hong, Y.-G. Kim / Information & Management 40 (2002) 25±4031Table 24.10. Construct validityPro®le of respondent ®rmsRangeFrequency PercentageConstruct validity is established by relating a mea-suring instrument to a general theoretical frameworkAnnual revenuein order to determine whether the instrument is tied toMore than $1 billion and above617.6$100 million to below $1 billion1441.2the concepts and theoretical assumption they are$10 million to below $100 million1132.4employing. In order to obtain evidence of the constructLess than $10 million38.8validity of an instrument, a researcher must make useTotal34100of both a convergent validity and a discriminantvalidity [35].Number of employeesFor convergent validity, we evaluated the item-to-10,000 and above25.9total correlation, based on the correlation of each item1000 to below 10,000823.5100 to below 10002161.8to the sum of the remaining items. This approachLess than 10038.8assumes that the total score is valid and thus the extentto which the item correlates with the total score isTotal34100indicative of convergent validity for the item. Table 3shows the correlations for each of research variables4.7. Instrument validationwhose item-to-total correlation score was greater than0.4. Discriminant validity was checked by factorInstrument validation is a prior and primary processanalysis. In Table 3, discriminant validity was con-in empirical research. We followed Straub's [45]®rmed when items for each variables loaded ontoguidelines to validate the instruments used in oursingle factors with loadings of greater than 0.5 [16].research.Since each variable was measured by the multi-itemconstructs, we conducted factor analysis to check the4.8. Content validityunidimensionality of the items. Table 4 provides thefactor pattern matrix that shows the loadings of eachContent validity means how representative anditem on independent and dependent variables. Fivecomprehensive the items were in creating the scale.factors emerged with no-cross construct loadingsThe content validity of the instruments was estab-above 0.5, indicating good discriminant validity.lished through the adoption of the relevant construct inThe instrument also demonstrated convergent validitythe literature, a series of reviews with the help ofwith factor loadings exceeding 0.5 for each construct.business process experts with deep ERP knowledge,These results con®rm that each of these construct isand a pilot pretest.unidimensional and factorially distinct and that allitems used to operationalize a particular construct is4.9. Reliabilityloaded onto a single factor.Reliability is the accuracy or precision of a measur-ing instrument, that is the extent to which the respon-5. Resultsdent can answer the same or approximately the samequestions the same way each time [45]. The internalThe correlation matrix between variables is pre-consistency reliability was assessed by calculatingsented in Table 5. A common concern of any regressionCronbach's alpha values. The reliability results ofanalysis is the multi-collinearity that may exist amongthe constructs are summarized in the ®fth columnthe independent variables [16]. In our model, multi-of Table 3. The internal consistency (Cronbach's alph)collinearity is not a problem since we have only oneof the construct ranged from 0.75 (for implementationindependent variable. Correlations among the pro-success) to 0.93 (for process adaptation). Given theposed moderating variables do exist but would notexploratory nature of the study, the result seemspose a serious problem since their values are less thanacceptable.0.5 [16] and each of them will be analyzed separately.32K.-K.Hong,Table 3YSummary of reliability and validity of the measurement.-G.KimMeasureItemsMeanS.D.ReliabilityConvergent validityDiscriminant validity(Cronbach's alpha)(correlation of item with total score-item)(factor loading on single factors)/InformationImplementation success (R)a44.061.600.750.451, 0.613, 0.590, 0.5510.655, 0.799, 0.802, 0.773Organizational fit of ERP110.830.798, 0.806, 0.759, 0.7460.890, 0.896, 0.865, 0.857Process fit43.491.320.900.620, 0.714, 0.595, 0.6660.790, 0.854, 0.768, 0.823&Data fit43.951.260.820.515, 0.637, 0.5480.600, 0.737, 0.637ManagementUser fit33.711.330.740.741, 0.826, 0.814, 0.767, 0.624, 0.6030.822, 0.890, 0.881, 0.836, 0.724, 0.702ERP adaptation64.571.460.890.845, 0.842, 0.825, 0.830, 0.7750.904, 0.903, 0.890, 0.894, 0.854Process adaptation54.631.360.930.682, 0.804, 0.797, 0.748, 0.6030.798, 0.892, 0.889, 0.844, 0.726Organizational resistance54.331.500.8940a Reverse score.(2002)25±40K.-K. Hong, Y.-G. Kim / Information & Management 40 (2002) 25±4033Table 4The result of factor analysisaScale itemsFactors12345ORGFIT10.745À0.255À0.0030.0700.187ORGFIT20.752À0.3470.0120.0100.187ORGFIT30.744À0.2500.018À0.0480.166ORGFIT40.756À0.217À0.0150.0270.083ORGFIT50.7140.019À0.2030.0590.090ORGFIT60.7540.063À0.037À0.1560.079ORGFIT70.629À0.278À0.055À0.0910.137ORGFIT80.681À0.196À0.0610.0560.265ORGFIT90.546À0.2280.048À0.258À0.294ORGFIT100.585À0.0390.050À0.446À0.175ORGFIT110.7350.052À0.060À0.2480.061ERPADPT2À0.1890.6690.2700.066À0.160ERPADPT3À0.1720.8060.220À0.066À0.177ERPADPT4À0.2050.7880.2560.058À0.237ERPADPT5À0.1350.8460.1140.106À0.169ERPADPT6À0.1860.7370.0710.057À0.087ERPADPT7À0.1620.6430.3610.1180.100PROCADPT1À0.1130.2600.8380.128À0.104PROCADPT2À0.0810.1650.8480.203À0.117PROCADPT30.1260.1800.8710.145À0.069PROCADPT4À0.0810.1390.8470.212À0.027PROCADPT5À0.0630.2620.7800.1890.038ORGRST10.0790.0830.2170.745À0.199ORGRST2À0.0620.0150.3140.786À0.191ORGRST3À0.168À0.0960.3210.792À0.035ORGRST4À0.192À0.0150.2300.800À0.068ORGRST50.0110.230À0.0300.7780.023PJTSUC10.115À0.059À0.124À0.2200.635PJTSUC20.089À0.235À0.1990.0400.781PJTSUC30.346À0.3570.049À0.0770.648PJTSUC40.323À0.2160.184À0.3350.555Eigenvalues5.914.374.293.782.34Percentage of variance explained19.114.113.912.27.5Cumulative percentage19.133.2475966.7a Extraction method: principal component analysis; rotation method: varimax with Kaiser normalization.The results from the normal probability plot andsuccess was signi®cant (r ˆ 0:51, P < 0:002), support-Kolmogorov±Smirnov test indicated no violationing the Hypothesis 1. In simple regression of organiza-of normality for the regression model (statistic ˆtional ®t of ERP on implementation success, the value0:54±0.72, P > 0:200).of R2 and adjusted R2 was 0.26 and 0.24, respectively.It indicates that 24% of the implementation success5.1. Testing the base relationshipvariance is explained by the organizational ®t of ERP.Correlation analysis was used for testing the base5.2. Testing the contingency relationshipsrelationship between organizational ®t of ERP and ERPimplementation success. The base relation betweenFor analysis of the moderator variables, we will useorganizational ®t of ERP and ERP implementationthe typology of moderator variables and the method34K.-K. Hong, Y.-G. Kim / Information & Management 40 (2002) 25±40Table 5criterion variables, implying a signi®cant interactionCorrelations matrix between variablesbetween the moderator and predictor variables.(1)(2)(3)(4)Cell 3 is referred to as quasi-moderator, which notonly interacts with the predictor variable but is aImplementation success (1)predictor variable itself. Cell 4 is referred to as pureOrganizational fit of ERP (2)0.51**ERP adaptation (3)À0.49** À0.58**moderator, which interacts with predictor variablesProcess adaptation (4)À0.22 À0.18 0.36*while having a negligible correlation with the criterionOrganizational resistance (5)À0.46** À0.39* 0.42*0.42*itself.* P < 0:05.To test the three contingency variables, we followed** P < 0:01.the four-step procedure for identifying moderatorvariables [42] (described in Appendix B). The testfor identifying moderator variables developed byresults of the three moderating effects are summari-Sharma et al. [42]. Fig. 2 depicts this typology, whichzed in Table 6. While the ®rst line in each pair ofhas two dimensions: whether the target variable isregression models of the table shows the resultsrelated to a criterion (dependent) variable and whetherof the regression run without interaction, the resultsthe target variable interacts with the predictor (inde-of the regression run with the interaction term are inpendent) variable.the second line.The target variable in cell 1 is referred to as anIn the ®rst pair of regression models, it is found thatintervening, exogenous, antecedent, suppressor, orthe interaction term was signi®cant at the level of 0.01.predictor variable depending on its other character-In the correlation matrix of Table 5, ERP adaptationistics. Variables in cells 2±4 are referred to as mod-was negatively correlated with both organizational ®terator variables. The target variable in cell 2 is referredof ERP (À0.58) and ERP implementation successto as homologizer, which in¯uences the strength of the(À0.49) at the level of signi®cance 0.01. These tworelationship between the predictor and criterion vari-facts suggest that ERP adaptation is a quasi-moderatorables across homogeneous subgroups. In contrast,of the base relationship between organizational ®tmoderator variables in cells 3 and 4 in¯uence theof ERP and implementation success, supporting theform of the relationship between the predictor andHypothesis 2.Fig. 2. Typology of moderator variables.Table 6The test results of the moderated regression effectsRegression modelab (interaction)p (interaction)R2p (model)DR2ORGFIT0.260.002ORGFIT ‡ ERPADPT ‡ interactionÀ0.580.0060.470.0000.21ORGFIT0.260.002ORGFIT ‡ PROADPT ‡ interactionÀ0.500.0050.440.0000.18ORGFIT0.260.002ORGFIT ‡ ORGRST ‡ interactionÀ0.250.2430.370.0030.11a ORGFIT, organizational ®t of ERP; ERPADPT, ERP adaptation; PROADPT, process adaptation; ORGRST, organizational resistance.Document Outline

• The critical success factors for ERP implementation: an organizational fit perspective
  • Introduction
  • Theoretical perspectives
    • Organizational fit of ERP
    • ERP implementation contingencies
  • Research model and hypotheses
    • ERP implementation success
    • ERP adaptation
    • Process adaptation
    • Organizational resistance
  • Research method
    • Sample and data collection
    • Measurement development
    • ERP implementation success
    • Organizational fit of ERP
    • Implementation contingencies
    • Characteristics of respondent firms
    • Instrument validation
    • Content validity
    • Reliability
    • Construct validity
  • Results
    • Testing the base relationship
    • Testing the contingency relationships
    • Interpreting the moderating effects
  • Discussion
    • Organizational fit of ERP
    • ERP adaptation and process adaptation
    • Organizational resistance
  • Limitation and conclusions
    • Survey instrument
      • Framework for identifying moderator variables
  • References