Transcription

Back to "Search By Author"Abstract NumberTitle of the paperName of Conference:::002-0040ERP implementations and cultural influences: a case study2nd World Conference on POM, Cancun , MexicoAuthor/sDr. James F. O'KaneMarco RoeberNewcastle Business SchoolUniversity of NorthumbriaNewcastle Upon Tyne NE1 8STTel 0191-227-3839Fax [email protected]

Back to "Search By Author"AbstractUser acceptance and support of a new Enterprise Resource Planning (ERP) system atthe early implementation stages are critical key success factors, which have to betargeted with appropriate change management. This becomes even more important inan Asian context where national and organisational culture with a different value andbelief system, resulting in different management styles, might not harmonize withWestern business culture embedded in the predefined standard business processes ofexisting ERP packages. This research describes and critically evaluates research intonational and organizational culture and the influence of different national cultures onthe implementation and reengineering process of ERP packages in an Asian context.Using a case study, realized through a quantitative survey testing five of Martinsons’sand Davison’s propositions (2000) in a Korean sample company, confirmed theexpected results from the literature review that culture has an impact on theimplementation process and that employee empowerment is an unavoidableconsequence of an ERP implementation.IntroductionEnterprise Resource Planning (ERP) systems have been adopted by the majority ofFortune 500 companies and an estimated figure of 33,000 companies worldwide areusing ERP software. Consequently, worldwide over 100 ERP vendors are nowtargeting small and midsize companies (Mabert et al. 2001).

Back to "Search By Author"On the basis of an AMR Research Report, Kennerly (2001) reports that the ERPmarket, including consulting, is expected to grow to 66.6 billion by 2003. Themarket leader SAP AG, headquartered in Walldorf, Germany, increased its total saleson average by 25 percent per year over the last five years, reaching total sales of 7.3billion in 2001 and capturing 32 percent of the total ERP market (SAP Online Press).However, ERP implementations have not always been a similar success story forsome of the ERP vendor’s customers. For example, FoxMeyer Drug partially blamedthe failed ERP implementation as their reason for filling bankruptcy in 1996 (Chen2001). Nevertheless, the ERP vendors cannot fully be made responsible for suchimplementation project failure as several parties such as consultants, training andimplementation teams, and other factors tend to have a more significant influence onan ERP introduction and thus can make or break the project ultimately. ERPimplementation projects can cost a company depending on its size between 1.5percent and 6 percent of annual revenues and thus project failure could deal a severeblow to a company [Mabert et al. 2001].Recently the attention of researchers, concerning success factors of an ERPimplementation, has shifted from the hard facts to the more soft facts, addressinghuman and cultural factors when trying to explain why ERP implementations fail.Some firms such as Guinness, the brewing division of the Diageo food and drinksgroup, have addressed the human issues and cultural change in their businessprogrammes and the Guinness Integrated Business Programme (IGB) is a clearexample of this. [Financial Times; Mar 1, 2000; online archive].

Back to "Search By Author"Yet, very little large scale research concerning the influence of organisational andnational culture on the implementation of ERP software packages has beencompleted. Therefore, this research attempts to further investigate this topic by meansof studying the results of a SAP R/3 implementation in a Korean subsidiary of aGerman pneumatic manufacturer and the effect that national and internal companyculture affects on the success of the project. The aim of this research was to determinewhat impact national cultures – particularly Korean culture – have on theimplementation process of enterprise systems (ES) like SAP R/3. The objectives ofthe research were to:Discuss and compare the research on national cultures and highlightagreements and differences and possible criticism.Test some of the propositions of Martinsons and Davison in one Koreancompany.Deduce from the before mentioned objectives impacts on the ChangeManagement when implementing ERP software.Recommend ways to align the company culture to the implementationstrategy to encourage loyalty to the system after implementation.Primary research was carried out by testing some of the propositions of Martinsonsand Davison (2000) in one Korean company.Enterprise Resource Planning (ERP)

Back to "Search By Author"The roots of ERP systems go back to material requirement planning (MRP) systems,which were developed in the 1960s and 1970s, making it possible for the first time toplan production requirements based on future demand rather than on past data(Kennerly 2001). In contrast to MRP, whose functionality was limited to productionand manufacturing, ERP is cross- and multifunctional, delivering real time data fromonly one central database. It is integrated, thus linking all departments throughout theentire organization and modular meaning that any combination of modules can beused jointly together. Even the modules from different vendors, referred to as best ofbreed implementations, like those from the so called “big five” SAP, Oracle,PeopleSoft, JDE, Baan, which control approximately 70 percent of the ERP market –can be used jointly together (Light, 2001; Mabert 2001). However, the best of breedapproach embraces the danger that the modules of the different vendors were notdesigned to work together.The two most important benefits cited in the literature for implementing an ERP / ESsystem are data integration and standardization of former diversified informationtechnology (IT) systems (Kennerly 2001;Light 2001; Mabert 2001; Scott 2000). ERP/ ES enables an enterprise to have one single set of data, in one central database ratherthan inconsistent data in a multiple set of various function based databases. Inaddition, standardization lowers overhead costs owing to economics of scale intraining and documentation and reduced transactions costs (Hammer 1999).However, with the benefits emerge also the related pitfalls particularly withstandardization. This means an enterprise can only to a limited define its processeswithin the best practice processes, as these can be predefined from the vendor in an

Back to "Search By Author"ERP package. An alternative would be an “information federalism approach”, asreferred to by Davenport (2000a) which means an enterprise identifies all processeswhich need to be universal in its worldwide subsidiaries and apart from those eachsubsidiary can have its autonomously defined processes. However, obviously such anapproach results in corresponding higher costs and anticipated longer project periods.ERP in AsiaKaye and Little (1996) claim that Asia in general has been relatively slow in adoptingfacets of office automation due to the fact of their use of non-Roman characters. Inaddition, diffusion of ERP software in the world’s second biggest economy Japan isconsidered as modest and the whole of Asia is capturing a mere nine percent of theworld’s ERP market. The latest figures for 2001 from the world’s largest ERP vendorSAP show a similar picture with sales of 841 Mil. in Asia accounting for 15 percentof total sales (SAP Online Press). One reason for the low acceptance and diffusion inJapan is the fact that most of Japan’s manufacturing activities were moved toneighbouring countries where ERP was and is not persuasive (Huang and Palvia2001).Martinsons and Chong (1999) state, based on a previous study from Chong, thatmainly human factors contributed to an ERP implementation failure rate of 60% inMalaysia. In the literature it is also reported that difficulties in Singapore, which iscommonly considered as comparatively Western, were attributed to the Asian contextand the Western biased ERP software. However, although there is a relative largediffusion gap in Asia in comparison to Western countries particular to America, nowAsian countries are beginning to recognize the benefits of ERP systems. (Soh 2000)

Back to "Search By Author"In addition, after the Asian economic crises many Asian countries have lessenedforeign direct investment regulations, which resulted in a higher merger andacquisition activity from mainly Western firms into these regions (Mody and Negishi2001). A large number of these firms have already implemented ERP software and arelikely to extend it into their newly acquired Asian subsidiaries. It is clearly evidentthat most of the literature reporting from the Asian ERP market stress in some waythat the Asian context needs particular attention during the implementation period.Research on National Culture and Cultural Dimensions: Critical Issues ofHofstede’s and Trompenaars’s ResultsHofstede was the first scholar who analysed national culture on an extensive basis.His studies included about 117,000 survey responses in 1969 and 1973 in 66countries. His analysis resulted in four dimensions: power distance index (PDI),which describes the acceptance of unequal power distribution, individualism index(IDV), which indicates loose or close ties between individuals of a society,masculinity index (MAS), which encompasses masculine values like assertivenessand achievement and more feminine values like quality of life and family and the finaldimension: uncertainty avoidance index (UAI), which describes the extent to whichunknown situations threaten members of a culture. Later he added a fifth dimensionnamed Confucian Dynamism (CVS), which evaluates a long-term orientation againsta short-term orientation in life [Hofstede 1994].Hofstede findings enabled him to rate each observed country relatively on an indexscale from 0 to 100 for each dimension. Although, Hofstede’s findings were derived

Back to "Search By Author"from an IBM questionnaire which originally was designed and intended to measurethe success of a worldwide organizational culture programme, his results were to alarge degree confirmed by similar research, conducted by Trompenaars with a similarsize, comprising in the raw set some 50,000 cases of 100 different countries(Trompenaars 2000). As Trompenaars used a different questionnaire and method toanalyse the results, they can only be compared on the basis of a general tendencytowards mutual characteristics.Trompenaars’s seven cultural dimensions are partly different to Hofstede’s fivedimensions but nevertheless overlap to some extent. Krumbholz and Maiden (2001)concluded when comparing Hofstede’s and Trompenaars’s dimensions, as shown inTable 1, that the overlapping of some dimensions with several others suggest that thedefinitions are too general or unclear. Also Adler (1997) identified six similar andpartly overlapping cultural dimensions to those referred to above and states “culturaldescriptions always refer to the norm or stereotype; they never refer to the behaviourof all people in the culture, nor do they predict the behaviour of any particularperson.” This has never been the intention of Hofstede or Trompenaars’s researchwhich is only meant to show the average or the most predictable behaviour of amember of a certain culture in the sense of a “central tendency” as each nation canshow an extensive individual variation (Hofstede 1997; Trompenaars 2000).While Hofstede’s conclusion and findings have been widely recognized, they havebeen especially recently severely criticized to the same extent. Holden (2002)considers Hofstede’s findings as outdated given that the data was gathered some 30years ago and that the world since has changed significantly, particularly politically;

Back to "Search By Author"however, he acknowledges his achievement. As this point was critiqued before,Hofstede (1998) claims to have found centuries-old roots, which are solid and only toa very limited extent subject to change over time. McSweeny (2002), sharplycriticising Hofstede’s analysis methods, concludes that all five assumptions on whichHofstede based his findings are all “flawed”.In addition, Hofstede’s assumptions are for some countries based on relatively smallsamples such as Pakistan with a mere 37 or Singapore with 58 respondents. Moreover,the responses for the analysis were largely limited to marketing and sales employeesand hence can hardly be considered as representative for a nation (McSweeney 2002;Smith and Bond 1997). Nevertheless, Hofstede (1994) himself states that “ the four(dimensions) together account for 49 percent of the country differences in the data,just about half.”Korean Culture and Management PracticesAs the research described in this paper was conducted in the Korean subsidiary of aGerman pneumatic components and equipment manufacturer a brief introduction toKorean culture and management practices is considered necessary at this point.Korean culture is influenced and shaped by various practiced religions andphilosophies such as Confucianism, Buddhism, and Christianity which now representsupwards of 20 percent of the Korean society. Confucian values are still prevailing inbusiness practices and central to this is harmony, thus even disagreement will behidden behind a smile (Frank, 2001).

Back to "Search By Author"The Korean workforce is generally widely committed to their company and untilrecently it was a life time commitment, which was unsettled by the Asian crisis.Morden and Bowles (1998) and Floyd (1999) report that Korean business generallytend to be run in a hierarchical, authoritarian and centralised manner. This approachreflects partially the relatively high scores on Hofstede’s PDI and UAI. Thus teammeetings tend to be dominated by vertical communication with the leader rather thaninter or horizontal team communication, which clearly shows the influence of cultureon teamwork behaviour. In addition, the inputs of team members are not encouragedas this would imply disrespect to the team leader. Since the Korean culture is a highcontext culture, meaning that moods, expressions, the situation are part of thecommunication process which is far less direct than in Western cultures, face to facecommunication is generally preferred (Hong 1997).Cultural Impacts on Business Process Reengineering (BPR) and ERPImplementationsFour out of the five leading ERP vendors are from the United States and one is fromGermany. Both countries score relatively low on Hofstede’s PDI with 40 and 35,respectively and are relatively high on the IDV particularly the United States with 91and Germany with 67 (Hofstede 1994). However, national culture influences andshapes the organisational culture, thus the so called “best practice processes”integrated in the ERP standard software packages reflect, to some extent, anindividualistic and low power distance culture environment than the related primarilyEuropean and U.S. industry practices (Soh, Kien and Tay-Yap 2000). Additionally, assuch software packages are relatively inflexible and only to a limited degreemodifiable to existing business processes, the reengineering upon which the new

Back to "Search By Author"resulting work practices are based, may cause various culture clashes. Moreover, evenestablished Western companies are struggling with their reengineering and businesssoftware implementation approaches. Countries from South East Asia for instance thathave a different value and belief system inevitably need substantially longer time forthe adaptation and acceptance of such a major organisational change. McKenna(1995) concludes that the critical point of reengineering - particularly in overseassubsidiaries - which has to be addressed beforehand is the extent to which such anapproach will result in the decline of an existing value system which currentlyprevails the business world. Huang and Palvia (2001) state:“Japanese organizations emphasize employee loyalty and provide all meansto retain employees. BPR before implementing ERP violates this belief andrestricts the use of ERP.”Additionally, ERP software packages certainly affect the distribution of power andauthority, which is far deeper rooted in high PDI cultures than in low PDI cultures(Martinsons and Chong 1999). Likewise Hofstede considers PDI as one of the moreintricate dimensions concerning change efforts.Until recently little research was conducted concerning cultural impacts on theimplementation of ERP packages unless on the basis of occasional case studies.Nonetheless, Krumbholz and Maiden (2001) report of the involvement of SAP in anongoing research which focuses on the development of a model to predict the impactof national and organizational culture on the ERP implementation. Prior case studybased research has been conducted sometimes using two different subsidiaries, which

Back to "Search By Author"implemented an ERP package within a European context, but the authorsacknowledges that although cultural differences even in a converging Europe do exist,such approaches are less straightforward as here the cultural problems seem to bemarginal and managerial approaches tend to be largely similar, which makes anadaptation of ERP software relatively easy in comparison to an Asian context.Despite the fact that, they conducted their study in two subsidiaries of a multinationalEuropean supplier of pharmaceuticals and laboratory equipment in the UnitedKingdom and Sweden, Krumbholz and Maiden (2001) indicate that cultural clashesbecame evident. Thus transferring such a result to an Asian context, it is highlyprobable that severe cultural clashes can be expected if cultural considerations are notaddressed in an appropriate manner.Straub, Keil and Brenner (1997) used in a broader study, encompassing the USA,Switzerland and Japan, the technology acceptance model (TAM), which explains thebehaviour pattern of computer usage to test the acceptance of IT. In any case, theresults tend to suggest that the TAM is not valid in Japan due to cultural differencesthus for instance resulting in less use of e-mail and a preference of face to facecommunication.In a comprehensive literature review Nah, Lau and Kuang (2001) found 11 criticalfactors influencing the implementation of Enterprise Systems and precedingreengineering. Change management and cultural factors were addressed as secondmost important with seven out of ten references. The most important factor was ERPteamwork and composition which is influenced by national culture, too. This will be

Back to "Search By Author"outlined on the example of Korean business culture in the subsequent section as theauthor’s case study is a Korean company. Although Mabert (2001) sees the failurestories of ERP implementations as vastly exaggerated, Davenport (2000a; 1998), andHammer and Champy (2001) report from several major failures partly caused by notaddressing the organizational culture and the human factor during the implementationand reengineering process.Martinsons and Davison used secondary research and developed 12 propositionsconcerning the influence of culture on ERP implementations, based on extensiveliterature review, which can be found in Table 2. Currently, Martinsons and Davisonhave not yet developed a tool for testing their propositions and are not aware if theyhave been used for any further research [Davison 2001].Critical Review of Martinsons’s and Davison’s WorkAs the author’s research into the subject deepened, some critical points concerningMartinsons’s research came to light. When comparing the propositions 8a and 8b,which refer to the influence of MAS, with propositions 11 and 12, referring to CVS, astrong overlapping tendency between the four propositions could be discerned.Additionally, Confucian, Hofstede’s fifth dimension, was incorporated unchangedinto their work. This suggests that there might have already been some overlappingtendencies which had been neglected by Hofstede. Another reasonable explanationmight be Confucian and Masculinity tend to have similar influences on ERPimplementations. However, this raises the question: why did Martinsons and Davisonnot combine the four propositions?

Back to "Search By Author"Research ApproachOnly recently, research has increasingly focused on the cultural implications on ERPimplementations and mainly in a European or American context due to the higherdiffusion of such systems there. There has been relatively little primary research todate on the effects of culture on ERP implementations in an Asian context.Nevertheless, from the little existing research there is strong evidence that particularlyin an Asian context cultural considerations might be the most important criticalsuccess factor. Therefore this case study aims at investigating in cultural implicationson an ERP implementation in a Korean context, which to the author’s knowledge todate has not been previously investigated. The authors used five propositions of theresearch approach from Martinsons and Davison [2000], as a starting-point for theirown research.Selection of five PropositionsInitially the author’s intention was to test all 12 propositions of Martinsons andDavison (2000), however, in consequence of the implementation method, which willbe described briefly below, not all 12 propositions are valid for this case study. Forinstance, Proposition Three suggests that the initiative for undertaking an ERPimplementation comes from the subsidiary itself rather than from the headquarter ofthe company. Moreover, the implementation method which was led by a Germanimplementation team puts further limitations to the validation of all propositions. Forinstance, there were no process teams involved and thus in that case Proposition Six aand Six b were not considered.

Back to "Search By Author"In addition, a pre-test showed clearly that a draft questionnaire, comprising 45statements, was too time consuming, thus two out of three test candidates did notcomplete the test questionnaire. As a result, the questionnaire was reduced, testingonly three dimensions among which PDI and UAI, which are considered by Hofstede(1994) as the most difficult in change management approaches, were included. Thusthe final questionnaire comprised 26 statements, attempting to test five propositions,namely One, Four, Five, Eight and Nine.The case study company’s implementation methodologyIn order to better understand the limitations of the case study and why the authorsdecided to test only five propositions a brief introduction to the sample company’simplementation methodology is given.The structure of the sample company’s worldwide subsidiaries tend to be similar, asthey mainly function as sales offices, however, they often have an integratedproduction for mainly country specific applications and for reasons of local costadvantages. Therefore, a basic SAP R/3 version, which tends to match 95 percent ofthe subsidiaries’ processes was developed in the German headquarters, thus savingthem from having to do a process analysis in each subsidiary and making the roll-outrelatively straight forward. The remaining five percent were adjusted to thesubsidiaries’ unique characteristics during a pre-training phase. Often the essentialreengineering and reorganisation is left to the subsidiaries’ responsibility. Obviously,this approach is combined with advantages and disadvantages which are not relevantto this research.

Back to "Search By Author"Research MethodologyFor the primary research the authors decided to use a quantitative approach in light ofthe language barriers, which tend to be less significant in a quantitative than in aqualitative approach. As the sample size, namely 35 respondents, was relatively small,a quantitative approach would be less time consuming and easier to participate in forthe respondents, especially since English was not their native language. Moreover, theevaluation of the responses, when limited to the given statements, is less complicatedin quantitative approach. The option of a translation of the statements in Korean wasalso considered, but was discarded due to the potential loss of control of the translatedcontent. This could have been overcome by a back translation from a different source,however this source was not available and it would have meant extensive coordinationbetween the two sources. Also the IBM questionnaires, which Hofstede used for hisevaluations, were in English.The authors were aware of the effect that a survey might have on the frame of mind inwhich the employees formulate their responses to the questions. A survey might implyimpending change and the employees might have wanted to respond in the light ofanticipated change. However, since the author’s survey was conducted during theimplementation of SAP R/3 and after the organisational transformation accompanyingit, there would be less room for speculation and much more clarity on their view ofthe situation they currently found themselves in. This implementation experienceadded more credibility to their responses and increase the validity of the overallresults.

Back to "Search By Author"Additionally, the authors contacted Davison (2001) who suggested interviews as thebest approach to test the propositions. However, this was not possible due to thedistance and financial limitations and the non availability of video conferencingsystems in the Korean firm.Questionnaire DesignEach of the five selected propositions was tested with four to six quantitativestatements using a seven-category Likert scale, which ranged from strongly agree tostrongly disagree. Some of the statements were derived from a similar questionnaireused by Krumbholz and Maiden (2001), which attempted to test the cultural influenceon ERP implementations in a European context. An example would be StatementsSeven and Nine, which were intended to test Proposition Four. In addition, somestatements were designed based on the conclusions of Martinsons and Davison, forinstance Statement One and Two to test Proposition One. Some propositionspurposely contained two relatively similar statements which were meant to test theattention paid and the clarity of the questions thus similar rating for those wouldsupport this. An example is in Proposition Four Statement Seven and Statement Tenfor which similar answers are expected. A copy of the questionnaire can be found inthe Appendix A.Pre-test and PilotsThe questionnaire was pre-tested with five Chinese colleagues, whose educationalbackground was engineering, similar to most of the sample company’s employees insales and marketing. In addition, their average English knowledge was considered ashaving a similar level to the average level in sales and marketing in the Korean firm.

Back to "Search By Author"The following interview with one pre-tester revealed that some questions neededadditional explanation and although there was the risk of leading the respondent in acertain direction, it was considered as necessary to rephrase the question so that itwould be more clearer and understandable. For example, in Statement One the term“group ware” was better described by giving examples. After the revision the draftwas sent to the sample company where an employee, who was not included in thesample approved and checked the questionnaire for clarity. As a result the followingquestion-statement: “Team work is used to get work done rather than hierarchy”,contained in Proposition Five, was deleted from the questionnaire as it was considereddifficult to understand.Hypotheses and Sub HypothesesThe organisational culture in the Korean sample company, which is predominantlyinfluenced and to a certain extent shaped by national culture has an impact on theBPR and ERP implementation process and thus needs to be specifically addressed inthe change management process to ensure a smooth ERP implementation and broaduser acceptance. If one of the five selected proposition of Martinsons and Davison isconfirmed by the quantitative questionnaire survey then it can be said that theorganisational culture in the Korean sample company has an influence on the ERPimplementation. Proposition one and nine are considered as valid in the samplecompany context if the mean score of the statements is below 3.5. In contrastproposition four ,five, and eight are considered as valid if the mean score is above 3.5.

Back to "Search By Author"Results and Data AnalysisSample and survey realisationOne of the authors had worked half a year in Korea during the preparation phase of aSAP R/3 implementation and came to know most of the employees of the samplecompany personally, he considered the sales and marketing department as mostappropriate, as the employees’ English knowledge appeared to be sufficient foranswering the questionnaire. Hofstede also mainly used the answers of sales andmarketing staff for his evaluation and analysis but there was no intention on the partof the authors to imitate Hofstede. The decision to use the sales and marketingdepartment however, was solely based on their English language knowledge.The marketing and sales department comprises 35 employees, who receivedpersonally addressed cover letters, which were all signed personally by the authors toemphasize the importance and thus to increase the response rate. In addition, 10 nonpersonal cover letters were added to the parcel, containing the questionnaires, whichwas sent to the sample company in Korea, for additional distribution in case a lowresponse rate had been achieved in the two selected departments. After the Koreancompany approved the final version of the questionnaire 45 questionnaires werecouriered to Korea where an employee distributed them in the sales and mark

Enterprise Resource Planning (ERP) systems have been adopted by the majority of Fortune 500 companies and an estimated figure of 33,000 companies worldwide are using ERP software. Consequently, worldwide over 100 ERP vendors are now ta