37
Accting., Mgmt. & Info. Tech. 9 (1999) 223–259 www.elsevier.com/locate/amit Software stories: three cultural perspectives on the organizational practices of software development Line Dube ´ a,* , Daniel Robey 1,b a E ´ cole des Hautes E ´ tudes Commerciales, 3000, Chemin de la Co ˆte-Ste-Catherine, Montre ´al, Que ´bec, H3T 2A7, Canada b Department of Computer Information Systems, Georgia State University, P.O. Box 4015, Atlanta, GA 30302-4015, USA Abstract Postindustrial organizations have come to depend upon the steady production and modifi- cation of software products to meet their competitive needs. This study reports insights into software development practices that were revealed through a cultural interpretation of organiza- tional stories told by members of SWC, a company engaged in software development. Through interviews with 38 members of SWC, 83 stories were extracted and analyzed to identify their main themes. By grouping these content themes, we produced nine broader cultural themes that represented the organization’s cultural context. Two management practices applied in SWC—development team organization and outsourcing—were subjected to an analysis in which cultural themes were interpreted from each of three perspectives proposed by Martin, J. [(1992) Cultures in Organisations; Three Perspectives. New York: Oxford University Press]: integration, differentiation, and fragmentation. The interpretation provides a rich reading of SWC’s cultural context. Despite management attempts to develop a unified culture based on collaboration and communication among devel- opment groups, the team approach to software development was problematic. Imposing team- work upon groups that manifested distinct subcultural differences disturbed the work life of group members, and the change was only partially successful. SWC’s management also sought survival and tighter strategic focus through an outsourcing arrangement. However, our interpretation identified significant difficulties created by the partnership between two organiza- tions with very different cultures. The presence of the outsourcing partner also brought greater uncertainty and ambiguity because work priorities and practices were subject to constant * Corresponding author. Tel.: + 1-514-340-6765; fax: + 1-514-340-6132. E-mail addresses: [email protected] (L. Dube ´), [email protected] (D. Robey). 1 Tel.: + 1-404-651-2086; fax: + 1-404-651-3842. 0959-8022/00/$ - see front matter 2000 Elsevier Science Ltd. All rights reserved. PII:S0959-8022(99)00010-7

Culture- Software Stories- Three Cultural Perspectives on the Organi

Embed Size (px)

Citation preview

Page 1: Culture- Software Stories- Three Cultural Perspectives on the Organi

Accting., Mgmt. & Info. Tech. 9 (1999) 223–259www.elsevier.com/locate/amit

Software stories: three cultural perspectives onthe organizational practices of software

development

Line Dubea,*, Daniel Robey1,b

a Ecole des Hautes E´ tudes Commerciales, 3000, Chemin de la Coˆte-Ste-Catherine, Montre´al, Quebec,H3T 2A7, Canada

b Department of Computer Information Systems, Georgia State University, P.O. Box 4015, Atlanta, GA30302-4015, USA

Abstract

Postindustrial organizations have come to depend upon the steady production and modifi-cation of software products to meet their competitive needs. This study reports insights intosoftware development practices that were revealed through a cultural interpretation of organiza-tional stories told by members of SWC, a company engaged in software development. Throughinterviews with 38 members of SWC, 83 stories were extracted and analyzed to identify theirmain themes. By grouping these content themes, we produced nine broader cultural themesthat represented the organization’s cultural context. Two management practices applied inSWC—development team organization and outsourcing—were subjected to an analysis inwhich cultural themes were interpreted from each of three perspectives proposed by Martin,J. [(1992)Cultures in Organisations; Three Perspectives. New York: Oxford University Press]:integration, differentiation, and fragmentation.

The interpretation provides a rich reading of SWC’s cultural context. Despite managementattempts to develop a unified culture based on collaboration and communication among devel-opment groups, the team approach to software development was problematic. Imposing team-work upon groups that manifested distinct subcultural differences disturbed the work life ofgroup members, and the change was only partially successful. SWC’s management also soughtsurvival and tighter strategic focus through an outsourcing arrangement. However, ourinterpretation identified significant difficulties created by the partnership between two organiza-tions with very different cultures. The presence of the outsourcing partner also brought greateruncertainty and ambiguity because work priorities and practices were subject to constant

* Corresponding author. Tel.:+1-514-340-6765; fax:+1-514-340-6132.E-mail addresses:[email protected] (L. Dube´), [email protected] (D. Robey).

1 Tel.: +1-404-651-2086; fax:+1-404-651-3842.

0959-8022/00/$ - see front matter 2000 Elsevier Science Ltd. All rights reserved.PII: S0959 -8022(99)00010-7

Page 2: Culture- Software Stories- Three Cultural Perspectives on the Organi

224 L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

renegotiation. Members from both organizations dealt with contradictions between their pre-vious norms, values and work practices and those required by the new relationship.

Overall, our analysis demonstrates the importance of understanding the cultural foundationof management practices used in software development. These practices evoke interpretationsfrom members of a culture, who collectively redefine what might have been intended. A cul-tural analysis may prepare management to move more gradually or to introduce specialapproaches to managing change. 2000 Elsevier Science Ltd. All rights reserved.

Keywords:Organizational culture; Stories; Software development; Outsourcing; Development team

The outsourcing contract between SWC and Outsourcing Co.2 had produced manychanges in the lives of SWC employees, who now found themselves working closelywith Outsourcing Co. As one of SWC’s programmers spoke privately to one of theresearchers, she told the following story about Outsourcing Co.’s strict dress code.

As a matter of fact, [Peter] and I went to a programming class that was given byOutsourcing Co., and normally they give it to their own employees, but theyallowed us to attend this class. And most classes that you go to out of town,you’re allowed to dress casually. Well, I came into class the second day wearinga nice pair of gold slacks and a silk blouse, and one of the Outsourcing Co.higher-ups went over to our instructor and said, “Who is that person? Send herhome. Tell her to come back in a dress.” Because they don’t allow women towear slacks! So, if I had been an actual Outsourcing Co. person instead of acustomer, they would have sent me all the way back home, no matter where thatmight have been, 3,000 miles or whatever, to show back up in the proper attire.So, I think that Outsourcing Co. puts the emphasis on the wrong thing.(Programmer, 25)3.

While ostensibly an unimportant story about an inconsequential event, this brieftale is symbolic of more fundamental difficulties that SWC experienced with itsoutsourcing partner. When interpreted along with other stories told by otheremployees, a rich picture of the organization’s culture may emerge and we may gaingreater insights into outsourcing and other organizational practices affecting softwaredevelopment. To the extent that outsourcing, for example, creates the resentmentmanifest in the story above, it may impede rather than enhance the production ofuseful software products and reduce the quality of working life of system pro-fessionals. Because software development has assumed growing importance as aneconomic activity, it is important to understand the cultural implications of differentapproaches to organizing and managing software development.

2 SWC and Outsourcing Co. are pseudonyms. All names of companies and people in this report are dis-guised.

3 Stories and excerpts are followed by the role of the teller and a serial number assigned to identifythe story.

Page 3: Culture- Software Stories- Three Cultural Perspectives on the Organi

225L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

An organizational story is defined as a connected discourse about a unifiedsequence of events that appear to be drawn from an oral history of an organization’spast (Martin, 1982). Stories contain specific details about people (usually organiza-tional members), actions, times and places. Since the early 1980s, organizationalstories have been treated as important and pervasive components of an organization’sculture (Martin, 1982; Martin, Feldman, Hatch & Sitkin, 1983). They are importantfor cultural members because they help make sense of critical events that shape theculture of an organization and therefore the context in which organizational eventsoccur and are interpreted (Dandridge, Mitroff & Joyce, 1980; Schwartzman, 1993).The morals embedded in organizational stories provide members with normativeguides for their behavior and communicate the values that make an organizationunique (Martin, 1982; Martin et al., 1983; Wilkins, 1984). Organizational storiesserve as precedent for individual assumptions, decisions, and actions by communicat-ing cultural beliefs and values, indicating acceptable behavior and attitudes, and byproviding examples of general themes or ideas (Boje, 1991; Wilkins, 1984). AsSchank (1990) observed, “People remember what happens to them, and they tellother people what they remember. People learn from what happens to them, and theyguide their future actions accordingly” (p. 1). Stories can also offer an explanation oforganizational successes and failures (Martin et al., 1983; Boje, 1995).

Stories are useful in research on organizational culture because they contain mem-bers’ interpretations of organizational history and policies (Martin, 1982). At thecore of the concept of organizational culture are the expressive, subjective, and non-rational qualities of organizational experience (Smircich, 1983). The power of a cul-tural analysis resides in the contextual interpretations of observable behaviors andartifacts so that patterns of deeply held assumptions and underlying values arerevealed (Schein, 1985; Smircich, 1983). Cultural analysis acknowledges that vari-able social meanings may emerge from a given context. When members interpretthe meanings of stories and other symbols of culture, “. . . their perceptions, memor-ies, beliefs, experiences, and values will vary, so interpretations will differ even ofthe same phenomenon. The patterns or configurations of these interpretations, andthe ways they are enacted, constitute culture” (Martin, 1992, p. 3).

Cultural analysis pushes researchers beyond formal organizational charts and pro-cedures to investigate the organization’s fundamental norms, values, and assump-tions. The study of symbols acknowledges the ubiquity of symbolic activities inorganizations and draws attention to the observable artifacts produced by and usedwithin a culture (Morgan, Frost & Pondy, 1983). The study of organizational cultureand symbolism offers a holistic view of the organization, and is less concerned withquantification, prediction, and determinism than with qualitative, appreciative, andcontextual understanding of organizations (Louis, 1983; Knights & Willmott, 1987).Researchers who treat organizational stories as symbols of culture are often wellpositioned to interpret a culture’s implications for organizational practice.

Software development, the focus of the present study, may be conceived as theorganizational process whereby information systems are produced. As the transitionto a worldwide, postindustrial economy continues, software production assumes amore significant economic profile. Each year in the US, more than $250 billion are

Page 4: Culture- Software Stories- Three Cultural Perspectives on the Organi

226 L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

spent on application development in approximately 175,000 projects (Johnson, 1995;The Standish Group, 1995). For leading edge companies engaged in electronic com-merce within virtual markets, software applications have become a core economicactivity. For many other firms, more rapid development of software applicationshas assumed a competitive importance not imagined just a few years ago. Thus,postindustrial organizations have come to depend upon the steady production andmodification of software products to meet their competitive needs.

Organizational arrangements for software development continue to evolve. Tra-ditional data-processing shops, which originated in the 1960s to provide mainframe-based computing services to client departments, have been transformed into smallerand more responsive units using flexible project structures and other team-baseddesigns. Significantly, organizations have also reduced their commitment todeveloping software applications by contracting with external providers for theseand related services. Growing at a rate of 14.4 percent annually (Bruno, 1995), theUS market for systems-related outsourcing is projected to reach somewhere between$26 and $46 billion in a few years (Bruno, 1995; Mulqueen, 1996). Worldwide,recent estimates indicate the outsourcing market to be $164 billion (Bobrow, 1995),and between 20 and 50 percent of the largest US firms use some form of technologyoutsourcing (McFarlan & Nolan, 1995; Verity, 1996). Outsourcing has attracted con-siderable research attention (e.g., Huff, 1991; Loh & Venkatraman, 1992a, b; Lac-ity & Hirschheim 1993, 1996; Teng, Cheon & Grover, 1995; Hiu, Saunders & Gebelt,1997). Outsourced software development places distinct new pressures on organiza-tions because it creates partnerships between two or more organizations with differenthistories, cultures and ways of conducting business (Lacity & Hirschheim, 1996).

This study focuses upon organizational stories as verbal symbols that reveal theorganizational culture of a single software development company that underwentmajor changes in its software development practices. The research question is:Whatinsights into the management of software development practices are revealed througha cultural interpretation of organizational stories?To answer this question, weinvestigated 83 stories told by 38 organizational members of SWC and analyzedtheir contents to produce nine major cultural themes. We interpreted these themesusing three perspectives on organizational culture articulated by Martin (1992): inte-gration, differentiation, and fragmentation. Our interpretations yielded a culturalunderstanding of two contemporary practices used at SWC: team organization andoutsourcing. The results reveal the problematic nature of these practices while show-ing how organizational members coped with their disruptive effects.

Fig. 1 provides an overview of our approach. We continue by describing the theor-etical framework used to guide our investigation: Martin’s multiple perspectives forthe interpretation of organizational culture.

1. Three perspectives on organizational culture

In our approach to cultural analysis, the researcher must first gain access to mani-fest symbols and activities and then interpret the meaning of those symbols within

Page 5: Culture- Software Stories- Three Cultural Perspectives on the Organi

227L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

Fig. 1. Research design.

the organization’s context. Whether acknowledged by the researcher or not, interpret-ation is subject to biases inherent in the researcher’s own preconceptions about cul-ture. Ott (1989), for example, found 38 different definitions of organizational culturein the literature, each of which would lead a researcher to focus on different aspectsof culture and to draw different conclusions from field observations. Martin, withMeyerson (Martin, 1992; Meyerson & Martin, 1987; Martin & Meyerson, 1988),addressed this problem by articulating three distinct perspectives for interpreting cul-tures:integration, differentiation,andfragmentation. These perspectives differ prim-arily in their treatment of ambiguity in an organization. Martin (1992) uses theseperspectives as alternative interpretive frameworks that may be applied in collectingand analyzing cultural data. The objective is not to assimilate the three perspectives

Page 6: Culture- Software Stories- Three Cultural Perspectives on the Organi

228 L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

or to determine which one best describes a culture, but rather to use each of themin turn to understand a culture more completely.

The integration perspectiveviews culture as a set of values and assumptions thatare shared by all members of an organization. Integration portrays people unitedaround cultural themes, such as valuing innovations or rewarding hard work, fromwhich members draw the strength and guidance to compete with other organizations.Under the integration perspective, culture functions to remove ambiguity by focusingmembers’ attention on unifying values and assumptions. This is a functionalist per-spective on culture. During the 1980s, popular books reflected the integration per-spective, and the management of corporate cultures through the manipulation of sym-bols and signs became a significant pursuit (Deal & Kennedy, 1982; Peters &Waterman, 1982; Schein, 1985).

Thedifferentiation perspectiveis skeptical of the integration perspective’s assump-tion of organization-wide consensus. From the differentiation perspective, interpret-ations of themes, practices, and assumptions are often inconsistent, and consensusexists primarily within subcultural boundaries. Differentiation directs researchers tolook beyond apparent unity and harmony toward the inconsistencies and contradic-tions among diverse subcultures. The differentiation perspective acknowledges ambi-guity directly by treating events and social relationships as actions that are vestedwith multiple and divergent meanings (Young, 1989). It enriches the interpretation oforganizational culture by introducing the prospect of conflicting relationships amonginterest groups within the organization.

The fragmentation perspectiveseeks to escape the simplicity, order, and pre-dictability imposed by the first two perspectives and considers ambiguity, com-plexity, multiplicity, and flux to be the essence of organizational culture. Contraryto the differentiation perspective, the fragmentation perspective does not assume thatgroup identities form stable subcultures. Rather, consensus, dissent, and confusioncoexist, making it too difficult to draw cultural and subcultural boundaries. Members’views are dynamic and may change from moment to moment as the importance andsalience of issues shift, as tasks and people change, and as new information becomesavailable. Different systems of meaning coexist, leading to irreconcilable interpret-ations being simultaneously entertained and accepted. Under fragmentation, ambi-guity prevents agreement on shared values, and neither organization-wide consensusnor differentiation of subcultures is possible, as advocated by the integration anddifferentiation perspectives, respectively.

The present study seeks insights available from all three cultural perspectives byincorporating multiple interpretations as part of our research method. By doing so,we provide a more comprehensive analysis of software development activities thancurrently exists in the information systems literature. Although several authors haveadvocated the use of organizational culture as a theoretical foundation for researchin information systems (e.g., Scholz, 1990; Hirschheim & Newman, 1991; Walsham1991, 1993; Robey & Azevedo, 1994; Avison & Myers, 1995; Morieux & Suther-land, 1988; Jones, 1991), relatively few empirical studies have adopted a culturalperspective on software development. None of these has employed Martin’s multipleperspectives (first published in 1987 (see Meyerson & Martin, 1987)) to interpret

upali
Highlight
upali
Highlight
Page 7: Culture- Software Stories- Three Cultural Perspectives on the Organi

229L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

their data (see Kaarst-Brown & Robey, 1999; Kendall, Buffington & Kendall, 1987;Kling & Iacono, 1989; Romm, Pliskin, Weber & Lee, 1991; Pliskin, Romm, Lee &Weber, 1993).4

2. Method

2.1. Basic assumptions

We designed this research as an interpretive case study that generated qualitativedata for analysis and interpretation. We consider interpretive research to capture anontological assumption that organizational culture is socially constructed, that is,open to various interpretations both by actors and by researchers (Orlikowski &Baroudi, 1991; Walsham, 1996). This assumption is wholly consistent with contem-porary thinking about organizational cultures; indeed, the interest in interpretiveresearch in organizations is partly attributable to the emergence of organizationalculture as a theory (Smircich, 1983). Although organizations have material propertiesand consist of real actions, socially constructed reality consists of social meaningsthat cannot be defined objectively or precisely. Accordingly, our interpretive method-ology was designed to draw subjective understandings from the stories told byorganizational members. This epistemological approach is appropriate to the natureof the problem and the research question we sought to address (Walsham, 1996).

Moreover, we acknowledge that, as researchers, we provide our own interpret-ations of organizational realities and that we are not value-free (Orlikowski & Bar-oudi, 1991). We initiated the generation of knowledge and cooperated with organiza-tional members in producing it (Deetz, 1996). Thus, it is appropriate to “confess”our biases, to the extent that we are aware of them. Our value perspective is coloredby our employment in North American business schools, teaching courses on infor-mation systems. Our interest is to generate knowledge that contributes to improve-ment in the software development process. From our perspective, improvement is notconfined to “provable” refinements in development methods but rather encompasses awide range of social phenomena: motivation and satisfaction of systems pro-fessionals, conflicts among groups and individuals involved in the development pro-cess, organization and governance of software development, and so on. We are con-cerned about improvements for all participants in the process: professionals, businesspartners, customers, and managers. We are idealistic to the degree that we believethat satisfaction of multiple stakeholders’ interests is not impossible, and we hopethat the knowledge generated from this research helps to make software developmentless problematic.

4 In addition to the studies mentioned here, IS researchers have produced a growing number of inter-pretive studies of information systems development and implementation (e.g., Orlikowski, 1991; Wal-sham, 1993). Typically, these studies examine the subjective meanings surrounding software developmentpractices. However, because these and other studies do not explicitly adopt a cultural perspective onsoftware development, they are excluded from our review of the literature.

Page 8: Culture- Software Stories- Three Cultural Perspectives on the Organi

230 L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

Because qualitative research is relatively new in the information systems field, itis subject to frequent misconceptions. We address potential misconceptions aboutthe present study by clarifying what it isnot. First, it is not exploratory researchdesigned to develop theory that can be tested later with positivist methodology. Wal-sham (1996) regards the rhetoric of exploration to be a weak justification for inter-pretive research, and we conduct our research within an established theoretical frame-work that does not require further grounding. From this argument, it should be clearthat our research is also not an exercise in grounded theory (Strauss & Corbin, 1990).Second, although our focus is upon organizational culture, our study is not an eth-nography. Although controversy about what constitutes “real” ethnography continuesin organization science and anthropology (Avison & Myers, 1995; Creswell, 1997),we agree with Barley (1990, p. 232) that ethnography requires “extended periods offield work,” more than the four months used to collect our data. Nonetheless, ourmethods do allow a writing of culture through “the method of observer as participant,spending extended periods of time during work and breaks observing work and inter-actions and trying to enter the subjective worlds of organization members” (Prasad,1993, p. 1407). Finally, our stance is not critical or postmodern. While we seek toimprove the practice of software development, we do not seek to emancipate anyparticular participant or champion any particular cause. Neither do we view theorganizational practices surrounding software development as inherently chaotic, aspostmodernism would imply (Deetz, 1996).

2.2. Research strategy

Our research strategy is that of an interpretive case study. Yin (1994) describesthe distinguishing characteristic of case study research as an attempt to examine “acontemporary phenomenon in its real-life context, especially when the boundariesbetween phenomenon and context are not clearly evident” (p. 59). This strategy isappropriate in studies of organizational culture because it is usually meaningless totreat culture as an influence that is separate from other organizational phenomenabeing studied. We do not assume that culture determines or causes particularapproaches to managing software development. Rather, managing software develop-ment is an inextricable part of an organization’s culture. It reflects cultural assump-tions and values, and cannot stand apart from them. Case studies allow an under-standing of complex phenomena in their natural settings and require that researchersimmerse themselves in the culture so that they can get close to members’ own subjec-tive understandings.

2.3. Research site

With its 952 employees, SWC builds and maintains software products for its cli-ents in the global industry of travel services. Fig. 2 traces some significant eventsin the company’s history. SWC was originally the information systems departmentof a large corporation (here called Parent) but became an independent subsidiaryapproximately ten years prior to the research. Since then, SWC was acquired by

Page 9: Culture- Software Stories- Three Cultural Perspectives on the Organi

231L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

Fig. 2. SWC’s history.

another parent company (here called New Owner), experienced the bankruptcy ofNew Owner, and sought protection for itself under bankruptcy law. To assure itslong term survival, SWC outsourced a major part of its activities to a corporatepartner (here called Outsourcing Co.), a contract that included the transfer of about2,000 employees from SWC to Outsourcing Co. Following this transition, SWCundertook a massive business process reengineering effort. This brief yet eventfulhistory allowed SWC and its employees to experience many of the difficulties facingNorth American corporations during the 1980s and 1990s: changes in ownership,financial instability, reorganizations, strategic partnerships, and massive organiza-tional change.

Access to SWC was negotiated with the Vice-President of its TechnologyDivision, who was interested in the research for the insights it would provide intothe management of software development at SWC. This division included about 160people and was divided into three departments—one focusing on PC products, oneon mainframe products, and one on the administration of contracts with externalparties. Fig. 3 shows the configuration of these departments. The TechnologyDivision was charged with the technical aspects of the development of the software(design, programming, testing, and technical implementation). The researchers hadno consulting contract with SWC or any of its employees, and it was agreed thatall identifying information would be concealed in subsequent research reports.

Fig. 3. SWC’s organizational chart. Note: For the sake of readability, the chart shows only the maingroups directly involved in software development before June 1994. The Project Management Office wasadded during the course of the study. The number in each box (x/y) gives the number of people interviewed(x) and the total number of people in a group (y).

Page 10: Culture- Software Stories- Three Cultural Perspectives on the Organi

232 L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

2.4. Data collection

Studying cultural phenomena requires one to get immersed “in the flood of aliencultures in order to grasp the direction of the stream and to feel the temperature ofthe water” (Czarniawska-Joerges, 1992, p. 197). Accordingly, the first author helddiscussions with several directors prior to scheduling interviews in order to appreciatethe business and organizational context of SWC. During the four months (fromMarch to June 1994) in which all subsequent interviews were conducted, this authorused time between interviews at the research site to wander around, have lunch there,read memos and bulletins, and work in an assigned office. This allowed theresearcher to appreciate the professional jargon, the many abbreviations and acro-nyms, and the local meanings of particular words. It was also an ideal way to observeorganizational members in a variety of situations (staff meetings, planning sessions,informal encounters between co-workers, coffee breaks) and to experience first handsome of the events mentioned during interviews. By maintaining a presence in theresearch site, the researcher also developed the trust needed to investigate more sensi-tive areas in subsequent interviews.

Thirty-eight semi-structured interviews were conducted with members of SWC’sTechnology Division. Fig. 3 indicates the number of people in each group with whomscheduled interviews were conducted. Only one person was interviewed twice, atthe very beginning and the very end of data collection. Interviews typically lastedbetween 30 and 45 minutes. Respondents were chosen using two nonprobabilitysampling techniques (Bernard, 1988).Quota samplingwas used to assure that allsubpopulations of interest were represented. For each working group, respondentswere selected who represented different genders, ethnic groups, formal positions, andlevels of experience with the organization. The second technique used wassnowballsamplingin which respondents were asked to nominate other individuals to be inter-viewed about a certain topic. Snowball sampling was helpful in identifying the sub-cultures, informal groups, and social networks emphasized by the differentiation per-spective.

The interviews were designed to gather stories about particular events in the workcontext of software development. Individuals were encouraged to talk about a broadrange of topics (e.g., their current work, early experiences with the organization, andso on). Respondents were not asked specifically to “tell stories” about their experi-ences, but, as noted by Schwartzman (1993), many chose the story as a form ofdiscourse to illustrate their points. The method of gathering data in story form viainterviews differs from the practices of gathering published stories (Martin et al.,1983) and observing stories performed in situ (Boje, 1991). However, none of thesemethods is inherently superior or closer to the “truth” of organizational culture thanother methods. All narrative accounts contain interpretations provided by the story-teller, and organizational participants relate stories on numerous occasions: in situperformances, “off stage” to each other, and to outsiders such as friends, family, andresearchers. All interviews were tape-recorded and transcribed for analysis.

In addition, direct observations about organizational members’ habits, behaviorsand physical environment were recorded in field notes taken by the researcher. Docu-

Page 11: Culture- Software Stories- Three Cultural Perspectives on the Organi

233L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

ments such as the organizational chart, list of employees, statements of corporatemission and objectives, written analyses of competitive history, and promotionalvideos were also gathered. These notes and materials were used to confirmimpressions gained from the interviews. For example, stories describing the survivalinstinct of the organization were confirmed by the aggressive image of SWC con-veyed in marketing videos produced by the organization. Our use of confirmingevidence was not an attempt to validate the accuracy of the stories, but rather to aidin their understanding and context.

2.5. Data analysis

Despite their references to real events, stories were treated asinterpretationspro-vided by the respondents. In each of the steps described below, stories were treatedas symbols of a socially constructed culture rather than objective properties of thatculture. Our approach was to search for similarities in the interpretations providedby a large number of participants, i.e., interpretations shared by participants ratherthan idiosyncratic meanings ascribed by individuals. In all steps of the analysis, weapplied our own interpretations about the degree of similarity and differences amongstories. Thus, although our method consists of orderly steps taken to reduce the rawdata to a few themes, the approach is essentially interpretive rather than positivist.No special analytical tools or software programs were used to aid this analysis.

The data analysis method adopted in this study was derived from Martin (1992,p. 38). The first step was examining the transcripts of the 38 interviews to identifyinstances of stories. A story was defined as a segment of an interview that was aconnected discourse about a sequence of past events that occurred in the organizationor that affected it. This definition included events involving both organizational mem-bers and other people related to the organization in some way.5 Stories containedspecific descriptions of events, actors and context. The second step was assigning acontent theme to each of the 83 stories that met the specified criteria. The contenttheme identified the primary message in the story, the point that the storyteller wasmaking. Thirty-five content themes emerged from the data and were not specified apriori. The third step was reducing the 35 content themes to fewer categoriesreflecting more fundamental cultural themes. The content themes were examined forlogical similarities, and nine “grand” themes were defined. For example, contentthemes about how “changes of ownership had happened and were still possible,”how “the future was uncertain,” how “the business environment was unstable,” and“problems with keeping the employees informed”, among others, comprised thegrand theme ofuncertainty. The grand themes represent dimensions of SWC’sorganizational culture.

5 A general statement without specific actors or events would not be considered a story. For example,the statement “before this, we used to react by assigning blame to the analyst when a bug was found”is not a story because no specific actor is named and no specific occurrence of a bug is described. Storiesrelating events that happened when the respondent was working for another company (former employer)were also ignored.

Page 12: Culture- Software Stories- Three Cultural Perspectives on the Organi

234 L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

Fourth, two managerial practices were drawn from the analysis of story content:SWC’s attempted use of team organization in software development and SWC’srelationship with Outsourcing Co. Although the broad objective of the research wasto study software development practices, these specific practices were not selectedbefore the study began. Rather, they emerged as the fieldwork progressed. Teamorganization and outsourcing were the subjects of many stories and were clearlyimportant and broad enough in scope to warrant the application of a cultural analysis.Because these two practices are frequently encountered in other software develop-ment organizations, the potential contribution of the study to the practice of infor-mation systems management is enhanced. The fifth and final step in the analysis wasto apply the three perspectives of Martin (1992) on organizational culture as inter-pretive lenses for understanding the two software development practices.

3. Results

Table 1 lists the nine grand cultural themes and shows their relation to the 35content themes. The nine grand themes express characteristics, values, beliefs, pri-orities, norms, and emotional responses to commonly experienced events. The cul-tural themes arehumanist image, pride, survival instinct, cultural differenceswith outsourcing partner, uncertainty, constant changes, challenging work, dif-ficulties in working together, and resistance to change.

Fig. 4 portrays the relationships among the 83 stories used in the analysis and thenine grand themes. Although stories were not directly related to grand themes in the

Fig. 4. Themes and stories.

Page 13: Culture- Software Stories- Three Cultural Perspectives on the Organi

235L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

analysis,6 the mapping of stories to the grand themes in Fig. 4 illustrates how somethemes have many stories associated with them while other themes are connectedwith only a few stories. Grand themes that are supported by many stories representcultural themes that were more widely expressed in the organization. Each of thegrand themes is described below along with story excerpts and references to otherstories.7

3.1. Humanist image

This grand theme is related to the care for individuals demonstrated by employeesand members of SWC’s management, including Parent’s president. Many storiesexpressed gratitude for management’s response to personal tragedies (e.g., highmedical bills, weather catastrophes) by directly helping the affected employees and/orletting employees use company time to solve their personal problems. Stories alsorevealed how management assisted employees who were transferred to the outsourc-ing partner. Despite difficult times (layoffs, outsourcing, pay cuts, high uncertainty),SWC retained a humanist image in employees’ minds. This grand theme is illustratedby the following story:

We had an unfortunate case here where one of our employees was being stalkedby an old boyfriend and she was shot outside and another person was killed. OurPresident flew in to come to the one person’s funeral and to visit the other personin the hospital. Now to me that means so much because it takes a lot for a Presi-dent of the company, in my mind, to take the time to do that and to show thatthey care enough that you are important. And they do a lot of that. The Presidentis approachable here and that is not common in most companies. He will comedown and say hello to me and ask me how I am doing or tell me that I did agood job on something. (Analyst, 13).

Despite many difficulties, SWC managed to maintain a caring image among itsemployees. Despite many layoffs over the years (including the transfer of 2,000employees to Outsourcing Co. and massive layoffs later), SWC was able to maintaina “non-firing” employer image among its employees (Story 54). SWC employeesoften spoke about other employees as friends and members of their families, andabout the company as a nice place to go in the morning.

3.2. Pride

The grand theme of pride encompasses stories about people’s feelings ofaccomplishment and affiliation with SWC and Parent. Many employees expressed

6 Stories were used to generate content themes, and content themes were organized into grand themes.7 For the sake of readability, only excerpts of stories are reported hereafter. These have been cleaned

up; hesitations and repetitions have been removed, and missing words have been added. For a completereading of every story, the reader is directed to Dube´ (1995).

Page 14: Culture- Software Stories- Three Cultural Perspectives on the Organi

236 L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

Table 1Relationship between Grand Themes and Content Themes

GRAND THEMES Content themes

HUMANIST IMAGE(related to the care for individuals demonstrated —difficult split inside the divided organizationby employees and members of management, (created by the major outsourcing and the transferincluding the parent company’s president) of employees to Outsourcing Co.)

PRIDE(related to the ability of SWC to compete against —going far with little resourceslarger companies in their industry despite their —fond memoriesmodest beginnings and difficult conditions)

SURVIVAL INSTINCT a

(related to SWC’s ability to survive despite thefailure of its parent companies and competitionfrom larger companies)

CULTURE DIFFERENCES(relates to the differences in culture between SWC —culture clashand the outsourcing partner. Because so many —Outsourcing Co. does not keep its promisesformer SWC employees now worked for the —Outsourcing Co. as a member of cold corporatepartner, performing outsourced functions for their Americaformer employer, differences in culture were seen —not understanding their business environmentmore clearly as individuals seemed to change their —formalization increasepersonalities to fit the expectations of the partner. —negotiations processThe partner was seen as very different from the —recognize in one occasion as caring humansurviving SWC) beings

—outsourcing impacts

UNCERTAINTY(related to SWC’s historical context, this theme —changes of ownershipreflects the frequent changes in ownership, —uncertain futurefinancial conditions, and competitive environment) —unstable business environment

—unexpected external demands—unstable working environment—“don’t worry” attitude—frustration—problems with keeping the employees informed

CONSTANT CHANGES(related to the experience of so many radical —reengineering process (success and difficulties)changes over a short period of time) —new working procedures

—resentment(continued on next page)

Page 15: Culture- Software Stories- Three Cultural Perspectives on the Organi

237L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

Table 1 (continued)

GRAND THEMES Content themes

CHALLENGING WORK(related to the need to provide competitive —exciting and varied workproducts that reflect the frequent changes in the —stresstravel service industry. Challenging work includes —actualizationreferences to stressful working conditions, the —lack of resourcesneed for frequent training, and frustrations over —training issuesinadequate resources)

DIFFICULTIES IN WORKING TOGETHER(relates to the divisions between subunits in SWC —lack of communicationand the traditional functional structure that —improved communicationseparates different stages of product development. —priorities are differentDifficulties included lack of communication and —dependence/independencethe inadequacies of other mechanisms designed to —lack of understandingpromote teamwork) —management differences

RESISTANCE TO CHANGEa

(related to ambiguity of dealing with constantchanges)

a It can be noted that the two grand themes, resistance to change and survival instinct, have no relatedcontent themes. The reason is that those two grand themes were initially two content themes. Becausethese content themes were repeated through several stories (see Fig. 4) and, at the same time, not easilyrelated to any other content themes, they became grand themes.

great pride in working for SWC and in participating in its development over theyears. One source of this pride was the modest origins of the organization and thedifficult conditions faced in its early years, as illustrated in the following story:

We started in a little trailer with five people in the early 1980s, and we are aforce to contend with. We are only 950 employees and we are up against peoplelike [their main competitor]. To give you an example, when I had my old job, itwas me; at [our main competitor], they had ten people doing what I did. Early1980s, like 1981 or 1982. So, as a company, we have accomplished a lot with alot smaller budget and a lot smaller work force than any of the other competitors.So, we have a lot to be proud of. I believe we also market more positively andhonestly. I do believe that our system is better. (Analyst, 30).

This feeling of pride seemed even stronger among the older employees who onceworked for Parent, whose disappearance had not been easy for them to accept; afew of them expressed fond memories of their earlier days working for Parent:

Parent was a wonderful company to work for. It always was. We always had thenewest toys in town, so it was easy to stay there. Newest computers and newestsomething. There was always something interesting happening. Parent going

Page 16: Culture- Software Stories- Three Cultural Perspectives on the Organi

238 L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

away—there was a lot grief involved, the people who’d worked there for a longtime. You had everything you could—no one will ever have that good of workingconditions again. (Manager, 76).

At different levels, most employees seemed very proud to be SWC employees. Whilemost were very candid about its weaknesses, most expressed an emotional link withthe company.

3.3. Survival instinct

Many of the content themes deal with obstacles to corporate success and difficult-ies for employees of SWC. Beset by bankruptcies and other uncertainty, respondentsmanaged to feel proud of SWC’s accomplishments and its humanist surroundings.Indeed, the theme of surviving against the odds comprises a major theme in itself.The following two stories are among the many that told how the organization wasable to survive and succeed despite adversity.

What happens is that we did not know where we were going to be without Parent.Who was going to provide the capital that we needed to grow? So, we had nocapital. Parent was going under. All we had was liabilities. So, the only thing wehad was the organization. (Manager, 73).

However, we were successful. We were very successful in the service that weprovided to [our clients], and we don’t have the market share that [our competitor]has, but at least people thought that when Parent went under, that we’d disappear,and it hasn’t been like that. (Manager, 42).

Survival instinct was also apparent in stories about the decision to transfer animportant part of SWC’s operations to an outsourcing partner:

New Owner was going bankrupt, and they took from SWC, because SWC wasmaking money during all this time, whereas New Owner was losing money likecrazy. We even had wage cutbacks and everything else, even though this particularcompany was making money. So, at the time, for us to survive, I think we hadto sell part of ourselves to Outsourcing Co. This was necessary at the time.(Programmer, 22).

Despite (or because of) all that had happened, SWC maintained its survivalinstinct. While everything was crumbling around them, they managed to stay alive.Their aggressive strategy included an expansion of their previous computer systems,strategic alliances with outside partners, and a reengineering project to transformtheir internal operations. The survival instinct theme reflects the belief that employeesat SWC intended to remain a competitive force in their market.

Page 17: Culture- Software Stories- Three Cultural Perspectives on the Organi

239L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

3.4. Cultural differences

This grand theme is related to the differences in culture between SWC and Out-sourcing Co., and it encompasses content themes such as “Outsourcing Co. is partof cold corporate America”, “it doesn’t keep its promises”, “it doesn’t understandour business”, “it increased formalization”, and others. Outsourcing Co. was seen asvery different from SWC. Because so many former SWC employees were transferredto Outsourcing Co., performing outsourced functions for their former employer, dif-ferences in culture were seen more clearly as individuals seemed to change theirpersonalities to fit the expectations of Outsourcing Co. Stories contained many detailsabout relationships between SWC and Outsourcing Co. These were expressed asclashes between the cultures of the two companies, especially the extreme differencesbetween Outsourcing Co.’s perceived organizational philosophy and that of SWC.Some stories, such as Story 25 presented at the beginning of this paper, related thedifficulty that some SWC people had adapting to their new work environment.Others, such as the following story, focused on changed relationships among groupsthat formerly worked together within SWC:

. . . before they were SWC, and we were all working together. When they wentto Outsourcing Co. and we stayed at SWC, [because of] the philosophy of thecompany over there, you can feel some friction between the same people we usedto work with all the time. It used to be very collaborative and not anymore. Eventhough we get things solved, you can feel the friction. I believe it’s OutsourcingCo.’s management. Now they’re all quarreling themselves. They want everythingin writing, they want this and this . . . If you say something, they won’t take yourword, they want it in writing. That’s the way it is now. I don’t know why; itwasn’t like that before. (Analyst, 26).

Even though Outsourcing Co. ensured SWC’s survival, it seemed to play the roleof the “bad guy”. Outsourcing Co. was a big organization and was often portrayedas a representative of “cold corporate America” or “a bottom line company”(Analyst, 5).

To tell you the truth, I wanted to go there. But I’m glad I didn’t. Because of thephilosophy they have, the management style, the restrictions they impose on theirpeople. I think it’s like an old philosophy; they’re very strict. I don’t think theytreat [their employees] like professionals. (Analyst, 28).

Other stories refer to Outsourcing Co.’s restrictions against personal photographs ondesks and a strict dress code that specified “on a man’s shirt how far apart yourstripes can be” (Programmer, 24).

Outsourcing Co. was also criticized for its lack of understanding of SWC’s busi-ness environment. This situation, according to SWC employees, caused friction andmore work for them.

Page 18: Culture- Software Stories- Three Cultural Perspectives on the Organi

240 L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

I was told that Outsourcing Co. told a client that they had to move a telephoneline and they wanted to cut the line over during the middle of the afternoon,which shuts down my operations to that [client]. That is not what I want. [It] isnot smart from a customer-service point of view. So, there are issues that theydon’t always think of the implications to us before they are doing something.There are degrees of inconvenience. This happens to be major. We are talkingabout big money. (Analyst, 17).

These stories show that a close outsourcing partnership entails a long period ofadaptation and negotiation. Even after three years, cultural differences emerged asa strong theme, supported by 18 individual stories. Criticism of the outsourcing part-ner for its cultural differences was one of SWC’s most obvious cultural themes.

3.5. Uncertainty

This grand theme reflects content themes such as “changes of ownership”, “uncer-tain future”, “unstable business environment”, “communication problems betweenmanagement and employees”. All reflected the great uncertainty felt by SWCemployees. The historical context of the organization largely explains the uncertaintythat was present at SWC. Some important destabilizing events were the changes ofownership that SWC faced over the years as reported in the following story:

Back in ’91, we used to be a part of Parent. The way it happened is when NewOwner bought Parent, they separated SWC and created a computer division. Thecomputer operation was separated into its own entity. And then we decided tooutsource. We gave our host computer mainframe operation to Outsourcing Co.They’re responsible for operating the big CPUs, the big computers; they’reresponsible for the telecommunications. (Manager, 40).

The unstable business environment of SWC and the repeated bankruptcies of itsparent companies, along with its own protection request in 1991, created a climateof uncertainty and precariousness surrounding the future of the organization.Employees felt that they did not know what was going on at the higher corporatelevels, and rumors of new takeover attempts were frequent.

Right now, who knows what’s going to happen with SWC? And if SWC andOutsourcing Co. and [a third company] come up with that deal, I have no idea.I just know that something is going on. I don’t know what . . . There is somedeal . . . (Programmer, 79).

Even positive news coming from management was interpreted skeptically, asthough the goal was to place the organization in a more favorable light in order tocommand a better price on the market. Previous experiences had reduced employees’trust in management, as the following story illustrates:

Page 19: Culture- Software Stories- Three Cultural Perspectives on the Organi

241L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

Parent had never been stable for the past 10, 20 years. But when I started in ’87,pretty much all we would hear—there’s a reorganization, there’s a layoff, there’sall kind of troubles. So, in our mind, we know something is going to happen.Then Parent went to bankruptcy and there were strikes going on. All the[employees]; they went on strike. All kinds of stuff. So we know that we are infinancial trouble, at that time. And, actually, the company didn’t let us know, isnot telling us—it’s in the newspaper. And in the meantime, we know that peopleare bidding for SWC. (Programmer, 37).

Such stories showed how uncertainty was a normal condition at SWC. Facing anuncertain future and having difficulties reacting to outside demands, SWC was unableto control its own destiny. This caused SWC’s employees to feel powerless anddependent upon the actions of external parties.

3.6. Constant changes

SWC experienced many radical changes over a short period of time and the grandtheme,constant changes,emerged from a series of stories about the successes anddifficulties of the reengineering process, the working procedures that were frequentlychanged and the resentment and frustration that were caused by these changes. Oneof these changes involved radical redesign of the company’s core business processes.Just out of bankruptcy and with 2,000 fewer employees, SWC’s management neededto rethink its internal operations and become more efficient. For more than two yearsit had engaged in business process reengineering under the guidance of OutsourcingCo. reactions to reengineering varied. Many people seemed to be uninterested anduninformed about the efforts to reengineer SWC while others were told that theywere “saved” from the reengineering effort. For those who participated, the reengin-eering process seemed to be quite difficult:

Some things were done strangely. They identified all the workloads and so forth,and made a series of recommendations. The way they do the reengineering, usu-ally, at least the way Outsourcing Co. did it, it’s very interesting, and you becomeostracized if you point out any practical imperfections in these things. You’reencouraged to participate, but only positively. (Manager, 85).

Despite many procedures to try to keep the employees informed, such as a hotline and electronic mail messages, changes surprised more than a few employees.

Reengineering hasn’t hit me personally yet, or directly. But one thing that—it hadto be attributed to reengineering—is that I come in one day and I find out aboutthis change in one group. It was known as “integration”. I don’t even know whatit’s known as anymore, but I come in one day and someone says, “Oh. This groupis no longer under [that director] and is now under [another director]”, and I donot really know who this director is, someone on the third floor. And I said, “Well,

Page 20: Culture- Software Stories- Three Cultural Perspectives on the Organi

242 L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

when does it take effect?” and they said, “immediately”. I couldn’t believe it.(Analyst, 67).

What is most surprising is that this analyst did not know that this change wasunrelated to the recommendations of the reengineering committee. Top managementinitiated it. This is evidence that management was not sharing information and thatthe employees did not really know what was going on.

3.7. Challenging work

This grand theme expresses the challenge felt by members responding to the needto provide competitive products to suit the frequently changing markets of the travelservice industry. It was reflected in content themes such as “exciting and variedwork”, “stressful situations”, “lack of resources”, “training issues”, and so on. Oneparticular story illustrates how exciting and varied the work was:

I’ve been lucky in being exposed to several previous systems and several environ-ments since I’ve been working here. The first product that we were working onwas based on the OS/2 platform. Then there was this group in [US State] thatdeveloped a lot of applications and a fairly large system based on OS/2 and C.SWC said, “Well, we’re going to close that center”. So, they asked me if I wantedto go to [US State] and study the system so that I can bring it over here. I said,“Yes, I will go”. It was a great experience, a lot of learning. (Programmer, 80).

Challenging work also means stressful working conditions, the need for frequenttraining, and frustrations over inadequate resources. The following story reflectssome of these concerns:

When we had the last big fare war last July, none of the computer companiesanticipated the hits to the system that happened. We were working all sorts ofhours because you would take the resources from one computer to take care ofsomething else and that made something else not work. We put in what we calledthe war room. It was literally a war room. So, what you try to do is let me takesomething off-line that will give you some more resources. So, you sit here andhave to decide what program to pull. Then you have to run around and ask “fromyour area, I want to pull these three programs, is that going to cause you a prob-lem?” So, you negotiate that. So, it becomes bedlam. We thrive on that though.We must thrive on the stress otherwise we wouldn’t still be here. (Analyst, 4).

Because of SWC’s need to remain on the cutting edge of technological develop-ments, employees needed to stay up-to-date. This was particularly true of the PCproducts department where knowledge requirements changed quickly. Expectationsof new employees were high. People were hired with a master’s degree (if possible),received a minimal introduction to their jobs, and were expected to be smart andresourceful enough to learn on the spot. Any factor that eased the challenge, such

Page 21: Culture- Software Stories- Three Cultural Perspectives on the Organi

243L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

as training (one formal class per person per year) or resource acquisitions (newcomputers) became very important. External demands also placed pressure oninternal activities and increased the challenge of some jobs. Finally, as the nexttheme illustrates, the difficulties that groups experienced in working together alsoincreased the challenge of working at SWC.

3.8. Difficulties in working together

SWC suffered from poor communication and collaboration among groups. Storiesabout the difficulties in working together referred to the lack of communication andthe inadequacies of mechanisms designed to promote teamwork. These issues werereflected in content themes such as “lack of communication”, “improved communi-cation”, “dependence among groups”, “difference in priorities”, “management differ-ences”, and others. Many factors were cited to explain this phenomenon: the sequen-tially organized software development process, the diversity of hardware platformsand software tools, inadequate knowledge levels in other groups, workload, and theorganization structure that kept groups from knowing what each other were doing.

Two major changes at SWC were often mentioned in the stories contributing tothis theme. First, management tried to change from a sequential software develop-ment process to one in which contributing parties worked together as teams. Histori-cally, software products had been handed off from one subunit to the next along witha complete transfer of responsibility. This practice resulted in a very long productdevelopment cycle and, consequently, slow response to market changes. Managementintroduced a different way to develop software so that, instead of the old, sequentialprocess, subunits would work simultaneously on software products. Second, manage-ment tried to integrate the work of groups working on different hardware platforms.Again, in the past groups working on different platforms did not have to worktogether. But with a higher integration of systems requested by clients, historicallyseparate subunits were required to work more closely together. The following storyis a good example of the difficulties imposed by these new arrangements:

Well, just recently a project was loaded on-line and [another development group]didn’t know anything about it and it was corporate accounts . . . programs likethat completely blew them out of the water. So, it can cause big problems. Onetime we add a little space . . . again it was [the same group]. When they wentto test it they didn’t know . . . the document was different than what we wereactually getting. So, when they coded. . . they got something back with an extraspace in it and they couldn’t end any records. A little space sounds little but ifsomeone is editing this entry, not to have a space and all of a sudden one is there,it can cause big problems. It is hard sometimes because we don’t really knowwhat they check for. We are learning as we run into problems. (Analyst, 8).

Such difficulties in intergroup communication within the same division were diffi-cult to manage, but communication between divisions was also problematic. Onefrequently mentioned cause was the large number of contributors that needed to be

Page 22: Culture- Software Stories- Three Cultural Perspectives on the Organi

244 L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

involved before a project could be completed. In one story, the frustrated respondentsaid, “We are not in control of our destiny; we have to interact with four thousandother departments” (Analyst, 56). Even product groups that were more technologi-cally independent than others still had to interact with other divisions, such as docu-mentation and customer services, to get a product out the door (Programmer, 36).The multitude of people involved made it difficult to get adequate feedback andinput at the right time. Because of its specialized organizational structure, closerworking relationships were hard to achieve. Management hoped that the new “teamapproach” to software development would address the problem of collaboration.

In addition, limited resources increased the level of competition in the organiza-tion, which had a direct impact on the willingness of groups to communicate andcollaborate with each other. As one respondent said, “When you don’t have money,stupid things become important like the size of your office, whether you have awindow or not, whether you have a decent parking spot or not, what your title is andall of those things that shouldn’t make a difference” (Analyst, 49). Such perceiveddifferences in the treatment of employees across groups were a source of frictionamong groups (Manager, 16).

3.9. Resistance to change

In many of the stories, resistance to change is expressed directly as a means ofcoping with the ambiguity of constant changes. Management was concerned aboutemployee resistance to the reengineering project and other initiatives. Toward theend of the reengineering project, a consultant was hired to evaluate the organization’sreadiness for and attitude toward change (Manager, 43). Much of the concern aboutchange also was related to New Owner’s influence, as the following story indicates:

I think people would be willing to change as long as things don’t get crazy withNew Owner. A couple of years ago, New Owner came out of bankruptcy, andpeople had their salaries frozen. Not frozen, but they had to take a reduction, andthere’s no equity. Our president came last week, and things are looking very upand very positive, and we’re going to get all the reinstatements of money we lostand for the increases. I think, if that stays up, then it’s fine. Then people wouldbe willing to change, but I think if that doesn’t, if that’s a really short-term thingand then all of a sudden the rug is pulled back under you, then I think it’s goingto be difficult to keep the change and be really positive. (Manager, 43).

According to this manager, uncertainty about the future of the organization wasan important factor in the capacity of people to accept changes inside the organiza-tion. While the future seemed brighter, past events were still remembered, and manypeople expressed caution with respect to impending change.

Page 23: Culture- Software Stories- Three Cultural Perspectives on the Organi

245L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

4. Cultural perspectives on software development practices

In the following analysis, we seek to understand the implications of the culturalthemes described above for two specific software development practices at SWC:team organization and outsourcing. These practices were prominent in SWC, andthey are common practices for many other organizations (Niederman, Brancheau &Wetherbe, 1991; Brancheau, Janz & Wetherbe, 1996). Thus, our interpretation ispotentially relevant beyond SWC’s particular culture, which, although unique inmany respects, may also be representative of contemporary software developmentin other respects.

In making the connection between SWC’s organizational culture and these twosoftware development practices, we used the three cultural perspectives describedby Martin (1992): integration, differentiation, and fragmentation. We apply these toeach practice successively in order to enrich and expand the interpretation. Fig. 5suggests that the analysis assumes mutual associations among the cultural themes,software practices, and interpretive perspectives. Practices are not caused by culture,or vice versa. Rather, organizational culture and practice are mutually implicated,and the interpretive perspectives are applied to gain insight into the cultural impli-cations of practice and the practical implications of culture. A description of eachpractice is provided first, followed by the successive application of each perspective,with references to specific cultural themes indicated by bold type. Table 2 summar-izes the analysis, noting the contribution of each cultural perspective.

4.1. Software development team organization

4.1.1. PracticeLike many organizations engaged in software development, SWC faced problems

of cost overruns and delays. Lengthy system development projects decreased its

Fig. 5. Relationships among themes, practices, and perspectives.

Page 24: Culture- Software Stories- Three Cultural Perspectives on the Organi

246 L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259T

able

2IS

activ

ities

bype

rspe

ctiv

e

INT

EG

RA

TIO

ND

IFF

ER

EN

TIA

TIO

NF

RA

GM

EN

TA

TIO

N

Tea

mO

rgan

izat

ion

The

chan

geis

anat

tem

ptat

inte

grat

ing

unde

rN

ewte

amor

gani

zatio

nch

alle

nges

the

actu

alC

hang

esin

crea

seun

cert

aint

y.N

ewpe

ople

aco

mm

ongo

also

me

dive

rgen

tfo

rces

inth

ecu

lture

amon

ggr

oups

base

d,in

this

case

,on

havi

ngto

wor

kto

geth

erw

illha

veto

orga

niza

tion

(mak

ing

grou

psha

ving

diffe

rent

pow

eran

dco

mpe

titio

nra

ther

than

cons

tant

lyne

gotia

tepr

iorit

ies,

way

sto

doan

dpr

iorit

ies

and

way

sof

doin

gth

ings

wor

kan

dco

llabo

ratio

n.F

orth

isre

ason

,it

isea

syto

goab

out

thin

gs,

valu

esan

dno

rms.

colla

bora

teto

geth

er).

pred

ict

impl

emen

tatio

ndi

fficu

lties

.M

embe

rsw

illne

edto

deal

with

cont

radi

ctio

nsar

isin

gfr

omdi

ffere

nces

betw

een

thei

rpr

ojec

tte

am’s

cultu

rean

dth

eir

func

tiona

lun

it’s

one.

Key

them

es:p

ride,

surv

ival

inst

inct

,K

eyth

emes

:resi

stan

ceto

chan

ge,

diffi

culti

esK

eyth

emes

:unc

erta

inty

,co

nsta

ntch

ange

shu

man

ist

imag

e,ch

alle

ngin

gw

ork

inw

orki

ngto

geth

er

Out

sour

cing

Out

sour

cing

may

help

the

orga

niza

tion

topu

tO

utso

urci

ngbr

ings

toge

ther

two

grou

psT

heim

plic

atio

nsof

vend

ors

inda

ilyac

tiviti

esto

geth

eral

lits

stre

ngth

san

dus

eth

emto

havi

ngdi

ffere

ntvi

sion

s,pr

iorit

ies,

valu

esan

dco

ntrib

ute

toin

crea

seth

eun

cert

aint

yby

focu

san

dac

hiev

ea

com

mon

goal

.H

avin

ga

norm

s.T

his

expl

ains

som

edi

fficu

lties

inth

ead

ding

ane

wpl

ayer

inth

edy

nam

icfo

rces

clea

rgr

oup

toop

pose

tom

ayal

sohe

lpcr

eate

rela

tions

hip.

Thi

spe

rspe

ctiv

eal

sohi

ghlig

hts

cons

titut

ing

the

cultu

reof

the

orga

niza

tion.

ase

nse

ofun

ityin

side

the

orga

niza

tion.

the

need

toge

tto

know

anev

entu

alpa

rtne

rbe

fore

sign

ing

alo

ng-t

erm

cont

ract

.A

sit

wou

ldha

ppen

for

new

peop

lew

orki

ngto

geth

erin

side

the

orga

niza

tion,

anou

tsou

rcin

gw

illbr

ing

anad

ditio

nal

dim

ensi

onto

team

wor

king

and

prio

ritie

s,w

ays

todo

and

goab

out

thin

gs,

valu

esan

dno

rms

will

have

tobe

nego

tiate

d.T

hese

nego

tiate

dva

lues

,no

rms,

proc

edur

esm

ayen

dup

bein

gin

cont

radi

ctio

nw

ithS

WC

’son

esan

dor

gani

zatio

nal

mem

bers

will

have

tode

alw

ithth

ose

cont

radi

ctio

ns.

Key

them

es:h

uman

ist

imag

e,cu

ltura

lK

eyth

eme:

cultu

rald

iffer

ence

sK

eyth

eme:

unce

rtai

nty

diffe

renc

es,

prid

e,su

rviv

alin

stin

ct

Page 25: Culture- Software Stories- Three Cultural Perspectives on the Organi

247L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

ability to respond rapidly to a changing market in a very competitive environment.Some of these problems could be traced to SWC’s organizational structure, whichwas designed to support the traditional, sequential model for software development:the system development life cycle (SDLC). Under this structure, software projectswere handed off from one department to another until completion, and little senseof ownership developed over the project’s life. No work was performed in parallelbecause each department’s contribution was dependent upon the completion of priorsteps by other departments. The SDLC treated new products a bit like a hot potatothat people catch and throw quickly to avoid being burned. As one programmer said:“The habit seems to be: ‘I covered my area. I did my job, and it’s your fault ifsomething doesn’t work’”.

SWC’s first attempt to solve the problems associated with the SDLC was to createone large development team composed of all the people involved in every phase ofa given project. This did not work because it tied up too many people in unproductivemeetings that bored and frustrated almost everybody. Precious time was lost ondetails only of interest to one or two people. The Technology Division then decidedto develop a new approach that created two separate teams: a core team and anextended team. The core team included the project leader and a representative fromeach department working on the product, typically five or six people. The extendedteam included additional people from those departments whose expertise and workwere required. In theory, this design would eliminate the large, unproductive meetingand would allow groups to work in parallel, thereby decreasing project completiontime.

Management defined three pilot projects to experiment with team organization.The first pilot was the final phase of a small project, which was successfully com-pleted under team organization. Two other new, and more important projects (herecalled A and B) were then identified. In Project A, the development group in chargeof the analysis and programming resisted the assignment of a project leader fromoutside the group. The present leader believed that the group already had an effectiveway of managing and organizing itself. In the SDLC approach, used previously, theline manager of the department currently responsible for the project was completelyin charge of it. When management tried to assign a new leader to the developmentgroup, the project slowed down and members resisted the new leadership. Manyexcuses were given: tight schedule, group knowledge and experience, and others,but it was apparent from the interviews that resistance was a response to politicalthreat. According to its members, this group did not want to become a pilot for theteam approach. This group’s manager would have had to relinquish his power to anoutside leader, and the group’s members felt that they would improve their chancesof survival if they were able to demonstrate high achievement with their own leader-ship. Their particular expertise and management approach differed from other groups,and they were not about to share their expertise with other groups in the organization.Because of this opposition, Project A was never realized as a pilot project. ProjectB was initiated as a pilot project, assigned an independent project leader, and laterexperienced difficulties over the course of the project.

Page 26: Culture- Software Stories- Three Cultural Perspectives on the Organi

248 L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

4.1.2. PerspectivesThe team approach to development at SWC raised numerous questions that were

conveyed through the stories told and compiled into grand themes. What was therole of the line manager? What was the role of the extended team? Who was respon-sible in case of failure? How should common ownership materialize within hetero-geneous groups? The application of theintegration perspective as an interpretivelens draws attention to unifying symbols that signify shared values, assumptions,and beliefs. At SWC, management’s objective of unification led to experiments withthe team approach, which it hoped would facilitate collaboration among groups.Team organization was an effort to integrate groups under one project leader anddirect their work toward a common goal. As one manager explained: “We neededto establish a project management group that was more independent from the organi-zation, that worked with all the different parts of the organization to take the projectsthat are currently being designed and developed into the marketplace”.

At a high level, unifying values were manifest in the grand themes ofpride,survival, andhumanist image. For example, employees expressed collective con-cern when their colleagues suffered misfortune. As one story related, “Last year, wewent through quite a few employees dying of AIDS and it got rough for some ofthem financially. Instead of exchanging gifts at Christmas, we would put $10.00 inand give them a gift certificate for food. This was always a family feeling of acompany” (Analyst, 12). Stories that told about SWC’s financial struggles also con-veyed pride in overcoming adversity and outliving Parent. “People thought that whenParent went under, that we’d disappear, and it hasn’t been like that” (Manager, 42).Moreover, stories contributing to the theme ofchallenging work emphasized theexcitement of developing innovative and competitive products for customers. As ananalyst said: “As much as I may complain sometimes about the fast pace and theconstant change, it keeps your brain alive. I don’t think I could work in a normal9 to 5 office job. You can’t get bored here”.

However, the failure of team organization to achieve integration at an operationallevel was the subject of several stories contributing to the themes ofuncertaintyandconstant change. These themes suggest a lack of unification and annoyance atthe disruption produced by experiments in organizing software teams. As the ProjectA manager said: “I went in and said: I don’t want it to be a pilot, because a pilotmeant we had all these people looking at every little step that we took and that’sgoing to take a lot of our time. So, you tell me what you want to do. Either I dedicatemy time to slowly show them what we’re doing and keep them informed and talkto them about how it works, how empowerment works, or I don’t have anybodylooking into the team, and we’re just doing it, but then it will be faster” (52). Thus,despite some areas of unification, employees at SWC exhibitedresistance to changeon the issue of team organization.

To extend the analysis, we applied thedifferentiation perspective as a secondinterpretive lens. The differentiation perspective directs attention to values that arenot shared and seeks to identify different subcultures. Applied to the practice of teamorganization, differentiation draws attention to differences among groups and to thedifficulty in creating organizational changes that overcome these differences. As a

Page 27: Culture- Software Stories- Three Cultural Perspectives on the Organi

249L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

manager said “The ones that are up front with the machetes are going to have avery hard time, but they will be making progress and they will be making thingsbetter” (36).

Resistance to changeand difficulty working with others were fundamental toSWC’s culture, making team organization a relatively ineffective experiment. Thedestinies of Projects A and B were very different because of their respective subcul-tures. While Project B was very open to experimentation, had a low-profile leader,and worked with conventional methods, Project A was developing leading-edge pro-ducts with new development methodologies, and had a strong leader. When askedabout cultural differences between SWC and Outsourcing Co., one member of ProjectA responded: “I consider that I work in a group that is very, very different from allat SWC. We have been handpicked. I think we alienate a little bit the people aroundus” (Programmer, 37). This elitist attitude was evident in the following statement ofan analyst from Project A: “I don’t think our group will be affected by the reengineer-ing project because of the performance and the projects that we have to develop;we’re top performers. I mean that we are doers and we can perform”. With such sharpdifferentiation, thedifficulties in working together persisted despite the attemptedunification through team organization. The differentiation perspective focuses on thehistorically deep divisions among departments and illuminates the tremendous dif-ficulty of implementing a new structure requiring interdepartmental teamwork.

The fragmentationperspective complicates but enriches the analysis by drawingattention to the confusion, uncertainty and ambiguity in the practice of team organi-zation. For example, in one story the respondent related that he did not know thathe was part of a pilot team, even though his name was on the current project list.As the Project B manager said:

We haven’t had anything like this, in this way, in the company before. The biggestobstacle is that there is a lot of fear and intimidation by it because it is so different,and I think a lack of understanding of what I should do to accomplish my taskI’m responsible for. I don’t know where to start. I don’t know how to get goingor get into it. The whole pilot approach is new and different for everyone.

Also, because team organization grouped together people who were unaccustomedto working together and visualizing products in the same way, the priorities, values,and norms that applied to the old departmental structure were challenged and renego-tiated. More ambiguous roles and uncertainty thus exacerbated thedifficulties ofworking together.

Fragmentation also draws attention to the paradoxes and ironies within a largergroup of cultural themes. The coexistence of contradictory themes represents the sortof ambiguity that the fragmentation perspective assumes is endemic of all cultures.We have seen how the themedifficulties in working together persisted despiteattempts to reorganize software development into teams. Yet, we have also seen howmembers of SWC expressed greatpride in their accomplishments and subscribedto a commonsurvival instinct . Ironically, this survival instinct had become animpediment to continued survival because it threatened the collaboration needed to

Page 28: Culture- Software Stories- Three Cultural Perspectives on the Organi

250 L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

make the experiments with team organization succeed. That is, while individualdepartments worked very hard to ensure their survival and achieved things they couldbe proud of, they did not work well together, which placed everyone’s survival ingreater jeopardy. Faced with the choice between cooperating and exercising self-interest, SWC’s employees had apparently chosen the latter, leaving the conceptof team organization paradoxically positioned to enforce collaboration in a culturecharacterized by asurvival instinct , uncertainty, andconstant change.

4.2. Outsourcing

4.2.1. PracticeAs mentioned previously, SWC outsourced a large part of its operations in an

attempt to save the organization. Outsourcing all of the mainframe programming,telecommunications activities, computer support and operations, and client instal-lation and technical support resolved management’s uncertainty about investing inthe technology necessary to remain competitive and grow. As part of the contract,approximately 2,000 SWC employees were transferred to Outsourcing Co. Accordingto management, outsourcing allowed the survival of SWC by freeing the organizationfrom data processing problems and cost, assuring a certain level of technologicalperformance at a fixed cost (for a period of ten years), providing a quick financialtransfusion to the organization, and allowing concentration on the main business ofdeveloping applications.

4.2.2. PerspectivesFrom anintegrationperspective, outsourcing was a unifying event undertaken to

ensure survival. We have already seen that thesurvival instinct was an importanttheme for SWC, and the outsourcing contract was clearly associated in the mindsof employees with the need for survival. As the contract administrator said: “It’sbeneficial to both companies that we do not have to worry that we do not havemoney to buy the latest computers and so forth. Could we afford to operate our ownsystem? Maybe yes, maybe no. So, the decision was made to outsource”. By remov-ing various activities that were unrelated to the core business, outsourcing reducedthe amount of differentiation among subunits and laid the ground for a more unifiedculture focused on product design and marketing. The new relationship with Out-sourcing Co. also provided SWC employees with a common enemy, which furtherstrengthened internal unity. The theme ofhumanist image is also associated withthe events surrounding outsourcing. Because treating the employees fairly was avalue widely shared in the organization (in conformity with the integrationperspective), some employees were crushed to see their friends transferred to Out-sourcing Co. without any further discussion. As one analyst said: “Here you havepeople who are used to working in our environment and then you move them to avery structured, strict environment and some of the people didn’t make it”. Thethemes ofculture differences, humanist image, and pride support such an inte-grationist interpretation of outsourcing.

To conclude that the outsourcing partnership was simply a unifying event in

Page 29: Culture- Software Stories- Three Cultural Perspectives on the Organi

251L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

SWC’s culture, however, would be a limited analysis. Again, applying the perspec-tives of differentiation and fragmentation successively adds insight to the culturalinterpretation. Thedifferentiationperspective focuses directly upon the cultural dif-ferences between SWC and Outsourcing Co. As one SWC analyst described: “Weare not blue suits and white shirts kind of people. Then you bring in a group likeOutsourcing Co., that is extremely straight and narrow, that just sees down one path,and the blue suits and the white shirts, and you mix that group. It causes someinteresting fall-out”. While outsourcing had many positive features, such as unifyingagainst a common enemy and enabling survival, it also presented many difficultiesbecause of the sharp differences between the cultures of the two organizations. Thestory told at the very beginning of this paper is a good example of the many storiesillustrating this difference; because Outsourcing Co. placed so much emphasis onappearances and the “bottom line”, SWC employees regarded Outsourcing Co. as asuperficial company. The theme ofculture differencesbest reflects these differences.

Although SWC was now serviced by former employees who had moved to Out-sourcing Co., respondents were clear about the influence of Outsourcing Co.’s cul-tural norms on these people. As told in story 26, although relationships used to bequite informal and simple with other SWC departments, relationships had becomemore complicated and less collaborative after these departments were moved to Out-sourcing Co. While this seems quite natural from an outsider’s viewpoint, it seemedto necessitate a long period of adaptation for SWC’s employees. The study wasconducted three years after the outsourcing contract was struck, but similar storieswere told every time the topic of outsourcing was raised. As many stories relate,Outsourcing Co.’s organizational culture was more formal, hierarchical, and rule-oriented; SWC’s culture was much less formal, and employees were oriented tofulfilling customer needs rather than complying with arbitrary rules and procedures.Because outsourcing was a formal contract, Outsourcing Co. insisted on gettingeverything in writing with proper authorization and controls, but this was difficultfor many SWC employees who saw such rules to be obstacles to the quick problemsolving needed to deal with constant changes and unstable environment. As oneanalyst said: “The bottom line was to get it fixed. With Outsourcing Co., the paper-work has to come first”.

SWC employees frequently questioned Outsourcing Co.’s intentions and disagreedwith their business priorities. As a senior analyst said: “Outsourcing Co. is defini-tively a bottom line company. Does this make me money or doesn’t it make memoney? If it doesn’t make money, then get rid of it. I don’t see them as a peoplecompany where I see us as a people company”. Another programmer (30) said:“Outsourcing Co. is like big brother is watching you. If I step on you to get ahead,I’m going to step on you, and I don’t care. I might step twice on you, you know.That’s the kind of attitudes they have”. The differentiation perspective focuses onthese cultural differences in a business partnership and aids in understanding thedifficulties in managing an outsourcing relationship.

The fragmentationperspective draws attention to theuncertainty created by out-sourcing. Outsourcing Co. was seen as an unpredictable player, and that amplifiedthe difficulty of producing software products for customers. Employees felt out of

Page 30: Culture- Software Stories- Three Cultural Perspectives on the Organi

252 L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

control and dependent upon an outside force that neither understood their businessnor the way things were done at SWC. For example, Outsourcing Co. could assigna new account manager to SWC or replace other employees assigned to work onSWC’s projects. As one manager expressed: “I’m not in control of an operationthat’s dealing directly with the customer”.

The cultural differences exacerbated the general uncertainty prevailing at SWC.Teams now containing people from Outsourcing Co. had to negotiate work normswith people coming from a different entity while they continued to have problemsobtaining collaboration from internal groups. As one analyst noted: “We have toremember that we are a customer and not another entity of Outsourcing Co. and thatthey are not an entity of us. We were all one company at one time and now it’s likesplit, but it’s a lot of the same people. So, we have to remember sometimes that weare different companies trying to support each other”.

As an illustration of the degree of uncertainty and mistrust, SWC created a specialteam just to monitor performance levels and to administer the contract with Outsourc-ing Co. Because this group constantly had to intervene in conflicts between SWCgroups and Outsourcing Co., it was sometimes seen as being tied to OutsourcingCo. and was not there to defend SWC’s interests, which increased the uncertaintyaround the management of this relationship.

The fragmentation perspective also helps to identify contradictions betweenthemes. For example, because SWC’ssurvival instinct and pride were so strong,one might conclude thatcultural differences between SWC and Outsourcing Co.could be overcome, or that SWC would regard Outsourcing Co. as a savior. Despitethe importance of these themes in the organization, the cultural differences remainedan important theme and were extensively discussed. The fragmentation perspectivetakes these contrary expressions as signs of general ambiguity, uncertainty, and con-fusion. But these interpretations exist along side of those provided by the differen-tiation and integration perspectives. Each perspective reveals different interpretationsabout SWC’s culture, and all three are necessary to produce a fuller interpretation.

We have interpreted both software development practices from multiple culturalperspectives. We have also shown how excerpts from specific stories enrich theinterpretation from each perspective, leaving no doubt that multiple interpretation isa more complicated process, but one that is richer and more rewarding. In the follow-ing section, we draw implications of our cultural analysis for both practice andresearch.

5. Implications

5.1. Implications for practice

In contrast to strictly rational perspectives on systems development, the primarygoal of this study was to generate insights into the management of software develop-ment activities through cultural analysis of organizational stories. In this way, weseek to advance the understanding of how software development is actually conduc-

Page 31: Culture- Software Stories- Three Cultural Perspectives on the Organi

253L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

ted, and how management practices are interpreted in a cultural context. This studycontributes to a stream of research on the social context of information system devel-opment (Robey & Markus, 1984; Hirschheim & Newman, 1991; Robey & Newman,1996). Like any social artifact, management practices such as outsourcing and teamorganization are subject to interpretation and assigned cultural meaning by parti-cipants. The salience and importance of these meanings are clear from their emerg-ence in the stories told by organizational members.

The primary implication of our analysis is that practices commonly undertaken toimprove the production of software evoke interpretations from members of a culture,who collectively redefine what might have been intended or anticipated. This con-firms what many studies of organizational culture have historically revealed: thatsocial arrangements for work are not easy to design or control. When interpreted byparticipants in the culture, such practices may even be rendered ineffective or prob-lematic. Managers attempting to reorganize software production must acknowledgethe social meanings of such practices. Not only should managers anticipate conflict-ing views from subcultural groups, but they should also consider the consequencesof cultural ambiguity. Where changes are frequent and contradictory, the wisdomthat motivated the change may be masked by participants’ generalized ambiguity.By supplementing the integrationist view of culture with the differentiation and frag-mentation views, we gained insight into these cultural implications of organiza-tional practice.

While academics may understand the complexity and ambiguity associated withchanging organizations, managers’ actions do not always demonstrate sensitivity tothe complexity of organizational cultures. Indeed, SWC’s assumptions about teamsand outsourcing seemed to be rather naı¨ve. Team organization represented a funda-mental shift from prior practice, and little was done to manage the change processaside from running ill-fated pilot projects. Likewise, the outsourcing contractbetween two very different partners was rife with conflict, apparently not anticipatedby management. In all fairness, managers probably understand organizational cul-tures more intimately than academic researchers, but their roles require responsesthat demonstrate order, mastery and control. A deliberate and comprehensive culturalanalysis, incorporating the multiple perspectives considered here, may reveal whymanagers’ attempts to impose order often become problematic and why more gradualchange or special approaches to managing change may be more effective.

The early years of organizational culture studies were characterized by unrealisticexpectations that cultures could be managed through the manipulation of symbolsto produce unification around key values signifying a proper course of action. Suchis the legacy of the integration perspective. At best, the integration perspective warnsthat members may resist cultural change unless it is managed skillfully. The differen-tiation perspective more directly confronts the deep internal divides and diagnoses“resistance” as conflict among subcultures that have their own values and methods ofworking. Moreover, the fragmentation perspective completely dismisses the idealisticnotion of cultural harmony and draws attention to the signs of confusion in an organi-zation. The fragmentation perspective reveals the ambiguity associated with plannedchange. These more contemporary analyses view changes in software development

Page 32: Culture- Software Stories- Three Cultural Perspectives on the Organi

254 L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

practices as more problematic than simply overcoming cultural resistance. Storiesexpressing frustrations with constant change and difficulties in working togetherillustrate that a team approach to software development must overcome a great dealin order to be implemented successfully. Stories at SWC referred to people “losingtheir minds”, “feeling threatened”, and other such strong emotions. Anticipatingresistance, SWC’s management tried to implement the team approach to softwaredevelopment incrementally, going from pilot to pilot and adjusting the new develop-ment process. Instead of imposing their vision on the Project A team that resistedtheir new approach, managers adopted roles as observers. They tried to see how theycould integrate the more successful practices into the software development processand apply them to future projects. They learned that, to facilitate change, roles andresponsibilities needed to be clarified. They also learned that reward systems, nowin the hands of each department head and based on individuals’ contributions tothe department, would need to match the new desired values of collaboration andcommunication. The evaluation and reward systems would eventually need to beplaced in the hands of the project managers and used to measure the contributionsof individuals to project and team goals.

The friction that existed between SWC and Outsourcing Co. is evidence of culturaldifferences of great magnitude. Stories brimmed with resentment toward the newbusiness partner. How could Outsourcing Co. possibly understand SWC’s business,where business volume might triple in a single month? As one manager quipped, itwasn’t like selling Pepsi-Colas. The day the outsourcing company became animportant business partner, the management of this day-to-day relationship becamea constant source of uncertainty and friction for SWC’s employees. The culturaldifferences between two organizations precipitated constant negotiation of how busi-ness was conducted. To implement a new partnership in software development, man-agers and participants must acknowledge that organizational change necessitates cul-tural change. Even minor changes can affect the fragile relationship between twoorganizations. When new partnerships are formed, a subtle process of negotiation isinitiated that may redefine the culture of both partners. After an initial period ofpeaceful work, a group may be transformed into a state of chaos where norms, values,and work practices become subject to continuous negotiation. To help alleviate thissituation, Outsourcing Co. and SWC tried to identify people to serve as stable inter-faces to assure continuity throughout projects between the two organizations.

Because of the cultural implications of changes in software development practices,it is important for managers to understand the cultural foundations of practice. Avaluable method for obtaining cultural insights is listening to stories. Stories focuson events that are meaningful to organizational members and that have helped toshape an organization. Stories are also an important source of information for anyoneinside or outside an organization. Managers need to be attentive to the recurringmessages in stories and the broader cultural themes that they support. This studyhas shown that stories told by people at all levels of an organization are sources ofinformation that are useful for understanding the cultural context of software devel-opment practices.

Page 33: Culture- Software Stories- Three Cultural Perspectives on the Organi

255L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

5.2. Implications for research

This research sought to obtain insights into organizational culture by examiningorganizational stories. While this study examines stories told to the researcher inprivate interviews, other strategies have been used to elicit stories and discover mean-ings: asking specifically for stories during an interview (e.g., Wilkins, 1978), consul-tation of published sources (e.g., Martin et al., 1983) and observation of story per-formance (e.g., Boje, 1991). A study could be designed so that it uses all of thesesources. Research is definitely enriched and broadened by employing multiple metho-dological approaches, and further innovation in story collection should be encour-aged.

While this study looks at the main themes expressed through stories as a symbolof organizational culture, other approaches to analysis could also be used. Forexample, stories could be analyzed through formal deconstruction or through a for-mal narrative analysis undertaken to understand a story’s structure and form, its levelof language, and its rhetorical strategy. Either of these approaches could contributeto a better understanding of how people make sense of the flow of their organiza-tional experiences.

In any research effort, the researcher’s own presence is bound to affect the datagathered. By asking SWC employees to talk about their experiences into aresearcher’s microphone, the researcher creates demand characteristics affecting thedata provided by respondents (Rosenthal & Rosnow, 1991). However, organizationalstories are often told outside of organizational contexts: to friends and family as wellas researchers. We assume that the stories told during interviews remain culturalsymbols, even though the cultural setting for storytelling (the interview) does notexist naturally in the organization. Despite the artificial setting, respondents werevisibly involved in many of the stories they told. Rather than recounting factualdetails in a dull manner, storytellers interpreted the details to express an opinion oremotion directly related to their personal lives.

Finally, by focusing on orally conveyed stories, this study largely neglects othersymbolic activities or cultural manifestations in the organization. A more thoroughinvestigation could corroborate the interpretations of multiple symbols, includingpublished documents, architecture and work products. While such artifacts were usedto aid the researchers’ understanding of organizational context, they were excludedfrom formal analysis and interpretation.

6. Conclusion

Software development is an activity with increasing economic significance, andany inquiry into its management and organization is potentially useful. As the busi-ness context for software development proceeds on its revolutionary course, man-agers and professionals will continue to be confronted by the challenges of managingchange. Pressures for higher-performing products, and more frequent and fasterreleases, will increase the need for effective management response. As with all

Page 34: Culture- Software Stories- Three Cultural Perspectives on the Organi

256 L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

organized social activity, cultural understandings will emerge and play a significantrole in making sense of the changing workplace. The symbols of organizationalculture—stories and other artifacts—will provide valuable clues for managers andresearchers as they seek to effect and understand changes in software development.

This study illustrates the advantages of using multiple perspectives to understandorganizational culture and practices. As recommended by Martin (1992), the inte-gration, differentiation, and fragmentation perspectives each offer somethingadditional, and no single perspective enables a full interpretation. While emphasizingthe important shared values of pride and the survival instinct, the integration perspec-tive overlooks the organizational divisions that threaten survival. The differentiationperspective highlights the organization’s lack of cohesion and enables a deeperunderstanding of the problems of implementing team-based designs and workingwith an outsourcing partner. The fragmentation perspective acknowledges the roleof uncertainty and enables the analysis of contradictory themes. Clearly, sustainingSWC’s theme of humanist image amid the uncertainties and difficulties present inother themes is an extraordinary bit of cultural sensemaking. The fragmentation per-spective adds such insights to the cultural analyst’s other means of understandingorganizations. Contradictory themes are more clearly understood, as relevant aspectsof an organization’s culture, rather than ignored or explained away.

While the different organizational culture perspectives have been shown to beuseful in analyzing activities such as reorganization and outsourcing, they may verywell be extended to other information system activities. For example, the differen-tiation perspective can provide a framework for analyzing the reactions of differentgroups to the implementation of a particular system. The fragmentation perspectivemay help in understanding the creation of development teams composed of membersfrom diverse departments. This perspective would help highlight and focus on thenegotiations required for these people to reconcile their cultural differences. In futureresearch, we urge researchers not to embrace blindly the “managerialist” view oforganizational culture, but to question the appropriateness of the managerialist per-spective and to consider the richness of alternatives. A broader view of organizationalculture improves our understanding of complex social phenomena such as softwaredevelopment activities.

This study provides a cultural analysis of software development in one organiza-tion that has withstood many of the difficulties experienced by other companies in the1990s: outsourcing, radical organizational change, business process reengineering,multiple bankruptcies, and changes in ownership. How can the consequences of theseconditions on software development be understood? Our answer to this question isthat there are multiple ways of interpreting the culture of software development. Butrather than choosing among alternative perspectives, we advocate using more thanone. The research challenge is not to select the best cultural perspective for under-standing software development, but rather to understand in greater depth. Using mul-tiple perspectives supports cultural interpretations not attainable from the applicationof a single perspective. The culture of software development is simultaneously inte-grated, differentiated, and fragmented if researchers choose to see it that way. Asour analysis of SWC has demonstrated, cultural interpretations matter—to

Page 35: Culture- Software Stories- Three Cultural Perspectives on the Organi

257L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

employees, clients, business partners, managers, and others—and those interpret-ations can be enriched by a multi-faceted strategy.

Acknowledgements

We acknowledge the financial support of the Social Sciences and HumanitiesResearch Council of Canada (SSHRC), which partly funded the study. We alsoacknowledge the suggestions of Leigh Jin, Michelle Kaarst-Brown, and Allen Leeon previous drafts.

References

Avison, D. E., & Myers, M. (1995). Information systems and anthropology: an anthropological perspectiveon IT and organizational culture.Information Technology and People, 8, 43–56.

Barley, S. R. (1990). Images of imaging: notes on doing longitudinal field work.Organization Science,1 (3), 220–247.

Bernard, H. R. (1988).Research methods in cultural anthropology.Newbury Park, CA: Sage.Bobrow, E. (1995). Service outsourcing trend continues.Midrange Systems, 8 (5), 10.Boje, D. M. (1991). The storytelling organization: a study of story performance in an office-supply firm.

Administrative Science Quarterly, 36, 106–126.Boje, D. M. (1995). Stories of the storytelling organization: a postmodern analysis of Disney as “Tamara-

Land”. Academy of Management Journal, 38 (4), 997–1035.Brancheau, J. C., Janz, B. D., & Wetherbe, J. C. (1996). Key issues in information systems management:

1994-95 SIM Delphi results.MIS Quarterly, 20 (2), 225–242.Bruno, C. (1995). Outsourcing mania.Network World, 12 (51), 1.Creswell, J. W. (1997).Qualitative inquiry and research design.Thousand Oaks, CA: Sage.Czarniawska-Joerges, B. (1992).Exploring complex organizations: a cultural perspective.Newbury Park,

CA: Sage.Dandridge, T. C., Mitroff, I., & Joyce, W. F. (1980). Organizational symbolism: a topic to expand organi-

zational analysis.Academy of Management Review, 5 (1), 77–82.Deal, T. E., & Kennedy, A. A. (1982).Corporate cultures: the rites and rituals of corporate life.Reading,

MA: Addison-Wesley.Deetz, S. (1996). Describing differences in approaches to organization science: rethinking Burrell and

Morgan and their legacy.Organization Science, 7 (2), 191–207.Dube, L. (1995). The role of stories in understanding the cultural context surrounding information system

practices. Unpublished Doctoral Dissertation, Florida International University.Feldman, S. P. (1990). Stories as cultural creativity: on the relation between symbolism and politics in

organizational change.Human Relations, 43 (9), 809–828.Hirschheim, R., & Newman, M. (1991). Symbolism and information system development: myth, metaphor

and magic.Information Systems Research, 2 (1), 29–62.Hiu, Q., Saunders, C. S., & Gebelt, M. (1997). Diffusion of information systems outsourcing: a reevalu-

ation of influence sources.Information Systems Research, 8, 288–301.Huff, S. (1991). Outsourcing of information services.Business Quarterly, 55 (4), 62–65.Johnson, J. (1995). Chaos: the dollar drain of IT project failures.Application Development Trends, 2 (1),

41–47.Jones, M. R. (1991). The impact of organizations on computers.SIGCHI Bulletin, 23 (1), 66–67.Kaarst-Brown, M. L., & Robey, D. (1999). More on myth, magic and metaphor: cultural insights into

the management of information technology in organizations.Information Technology and People, 12,192–217.

Page 36: Culture- Software Stories- Three Cultural Perspectives on the Organi

258 L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

Kendall, K. E., Buffington, J. R., & Kendall, J. E. (1987). The relationship of organizational subculturesto DSS user satisfaction.Human Systems Management, 7, 31–39.

Kling, R., & Iacono, S. (1989). The institutional character of computerized information systems.Office:Technology and People, 5 (1), 7–28.

Knights, D., & Willmott, H. C. (1987). Organizational culture as management strategy: a critique andillustration from the financial services industry.International Studies of Management and Organiza-tions, 17 (3), 40–63.

Lacity, M. C., & Hirschheim, R. (1993).Information systems outsourcing: myths, metaphors and realities.New York: John Wiley.

Lacity, M. C., & Hirschheim, R. (1996).Beyond the information systems outsourcing bandwagon.NewYork: John Wiley.

Loh, L., & Venkatraman, N. (1992a). Diffusion of information technology outsourcing: influence sourcesand the Kodak effect.Information Systems Research, 3 (4), 334–358.

Loh, L., & Venkatraman, N. (1992b). Determinants of information technology outsourcing: a cross-sec-tional analysis.Journal of Management Information Systems, 8, 7–24.

Louis, M. R. (1983). Organizations as culture-bearing milieux. In L. R. Pondy, P. J. Frost, G. Morgan, &T. C. Dandridge,Organizational symbolism(pp. 39–54). Greenwich, CT: JAI Press.

Martin, J. (1982). Stories and scripts in organizational settings. In A. H. Hastorf, & A. M. Isen,Cognitivesocial psychology(pp. 255–305). New York: Elsevier.

Martin, J. (1992).Cultures in organizations; three perspectives.New York: Oxford University Press.Martin, J., Feldman, M. S., Hatch, M. J., & Sitkin, S. B. (1983). The uniqueness paradox in organizational

stories.Administrative Science Quarterly, 28, 438–453.Martin, J., & Meyerson, D. (1988). Organizational cultures and the denial, channeling and acknowledg-

ment of ambiguity. In L. R. Pondy, R. J. Boland Jr, & H. Thomas,Managing ambiguity and change(pp. 93–125). New York: John Wiley.

McFarlan, F. W., & Nolan, R. L. (1995). How to manage an IT outsourcing alliance.Sloan ManagementReview, 36 (2), 9–23.

Meyerson, D., & Martin, J. (1987). Cultural change: an integration of three different views.Journal ofManagement Studies, 24 (6), 623–647.

Morgan, G., Frost, P. J., & Pondy, L. R. (1983). Organizational symbolism. In L. R. Pondy, P. J. Frost,G. Morgan, & T. C. Dandridge,Organizational symbolism(pp. 3–35). Greenwich, CT: JAI Press.

Morieux, Y. V. H., & Sutherland, E. (1988). The interaction between the use of information technologyand organizational culture.Behaviour and Information Technology, 7 (2), 205–213.

Mulqueen, J.T. (1996). Industry trend toward outsourcing continues to grow strong.CommunicationsWeek, April 29, 1996. Issue: 608, p. 75.

Niederman, F., Brancheau, J. C., & Wetherbe, J. C. (1991). Information systems management issues forthe 1990s.MIS Quarterly, 15 (4), 474–495.

Orlikowski, W. J. (1991). Integrated information environment or matrix of control? The contradictoryimplications of information technology.Accounting, Management and Information Technologies, 1,9–42.

Orlikowski, W. J., & Baroudi, J. (1991). Studying information technology in organizations: researchapproaches and assumptions.Information Systems Research, 2, 1–28.

Ott, J. S. (1989).The organizational culture perspective.Pacific Grove, CA: Brooks/Cole.Peters, T. J., & Waterman, R. H. (1982).In search of excellence.New York: Harper and Row.Pliskin, N., Romm, T., Lee, A. S., & Weber, Y. (1993). Presumed versus actual organizational culture:

managerial implications for implementation of information systems.The Computer Journal, 36 (2),143–152.

Prasad, P. (1993). Symbolic processes in the implementation of technological change: a symbolic interac-tionist study of work computerization.Academy of Management Journal, 36, 1400–1429.

Robey, D., & Azevedo, A. (1994). Cultural analysis of the organizational consequences of informationtechnology.Accounting, Management, and Information Technology, 4, 23–37.

Robey, D., & Newman, M. (1996). Sequential patterns in information systems development: an applicationof a social process model.ACM Transactions on Information Systems, 14 (1), 30–63.

Robey, D., & Markus, M. L. (1984). Rituals in information system design.MIS Quarterly, 8, 5–15.

Page 37: Culture- Software Stories- Three Cultural Perspectives on the Organi

259L. Dube, D. Robey / Accting., Mgmt. & Info. Tech. 9 (1999) 223–259

Romm, T., Pliskin, N., Weber, Y., & Lee, A. S. (1991). Identifying organizational culture clash in MISimplementation: when is it worth the effort?Information and Management, 21, 99–109.

Rosenthal, R., & Rosnow, R. L. (1991).Essentials of behavioral research: methods and data analysis.New York: McGraw-Hill.

Schank, R. C. (1990).Tell me a story.New York: Charles Scribner’s Sons.Schein, E. H. (1985).Organizational culture and leadership.San Francisco: Jossey-Bass.Scholz, C. (1990). The symbolic value of computerized information systems. In P. Gagliardi,Symbols

and artifacts: views of the corporate landscape(pp. 233–254). Berlin: de Gruyter.Schwartzman, H. B. (1993).Ethnography in organizations.Newbury Park, CA: Sage.Smircich, L. (1983). Concepts of culture and organizational analysis.Administrative Science Quarterly,

28, 339–358.Strauss, A., & Corbin, J. (1990).Basics of qualitative research: grounded theory procedures and tech-

niques.Newbury Park, CA: Sage.Teng, J. T. C., Cheon, M. J., & Grover, V. (1995). Decisions to outsource information systems functions:

testing a strategy-theoretic discrepancy model.Decision Sciences, 26 (1), 75–103.The Standish Group (1995). Chaos. http://www.standishgroup.com/chaos.html. September 23, 1996.Verity, J.W. (1996) Let’s order out for technology.Business Week, May 13, 1996. Issue: 3475, p. 47.Walsham, G. (1991). Organizational metaphors and information systems research.European Journal of

Information Systems, 1 (2), 83–94.Walsham, G. (1993).Interpreting information systems in organizations.New York: John Wiley.Walsham, G. (1996). The emergence of interpretivism in IS research.Information Systems Research, 4,

376–394.Wilkins, A.L. (1978). Organizational stories as an expression of management philosophy: implications

for social control in organizations. Unpublished doctoral dissertation. Stanford University.Wilkins, A. L. (1983). Organizational stories as symbols which control the organization. In L. R. Pondy,

P. J. Frost, G. Morgan, & T. C. Dandridge,Organizational symbolism(pp. 81–92). Greenwich, CT:JAI Press.

Wilkins, A. L. (1984). The creation of company cultures: the role of stories and human resource systems.Human Resource Management, 23 (1), 41–60.

Yin, R. K. (1994).Case Study Research.(2nd ed.). Newbury Park, CA: Sage.Young, E. (1989). On the naming of the rose: interests and multiple meanings as elements of organiza-

tional culture.Organization Studies, 10 (2), 187–206.