Based on the selected article Repairing ERP, answer the following questions:

1. What is the basic theme of the article? Try to state it in just one or two paragraphs.
2. Did the article present a good support base? Theoretical framework? Explain.
3. What additional questions are suggested by the articles conclusions?

INTRODUCTION

Enterprise resource planning (ERP) has proved to be the most popular business software of the 20th century (Robey, Ross, & Boudreau, 2002). Firms have imple- mented ERP because they anticipate that it will enable strategically important organi-

THE JOURNAL OF APPLIED BEHAVIORAL SCIENCE, Vol. 42 No. 1, March 2006 40-57 DOI: 10.1177/0021886305284290

2006 NTL Institute

40

Wagner, Newell / REPAIRING ERP 41

zational change (Sawyer & Southwick, 2002). However, many organizations that implement ERP find that achieving such technologically enabled change is problem- atic (J. E. Scott & Vessey, 2002). Indeed, today ERP is associated with high levels of project failure (ERP Survey Results Point to Need, 2001) and user resistance (Aladwani, 2002; S. V. Scott & Wagner, 2003). However, research that claims ERP as wholly problematic may concentrate too heavily on presenting snapshot reports on the divisive aspects of the software. Thus, it is reported that 20% of projects are shut down prior to installation (Cooke, Gelman, & Peterson, 2001). This means that 80% of implementing organizations are committed to making the software work and demon- strates that conceptualizations of project failure and user resistance might change over time. Yet few studies have looked at how these failures are turned around and work- ing information systems are created. To explore this, we draw on the concept of social ordering.

Successful organizational change associated with the use of ERP involves the reso- lution of conflicts and the acceptance of a new social order. The concept of social ordering is central to understanding the adjudication of disputes because it provides the foundation for understanding how individuals come together to secure advan- tages for all participants (Fuller, 1978). Distinguishing the forms of social ordering lies in considering the goals of those who seek a mutually advantageous position. Organization by common aims or reciprocity represent opposing ways in which prob- lems are solved and goals achieved with the former being based on participants having the same goal and the latter on their wanting different things (Fuller, 1978). Based on this distinction, we identify controversies encountered during the ERP project and use a narrative methodology to see the reparation of these issues and the forward move- ment of the project. Narratives present accounts of the world from a particular perspec- tive and in a sequential manner to explain departures from the norm and so illuminate how the teller makes sense of the world and their relation to others (Bruner, 1990; Czarniawska, 1997). This study thus illustrates that even problematic ERP projects can be made to work within an organization.

Our research is conducted in a large U.S. university (Ivy) that partnered with Vision Corporation, an international ERP vendor seeking entrance into the higher education market. The university context is a valuable analytical focus for this study because its structure is designed to meet the needs of multiple stakeholders. Histori- cally, a plethora of functional information systems (IS) existed throughout the univer- sity with limited standardization of that data, which required aggregation and consoli- dation for university-wide reporting. This context creates challenges for the adoption of an integrated ERP (Swartz & Orgill, 2001; Wagner & Newell, 2004).

We would like to thank the special issue editors and reviewers for their helpful comments on an earlier ver- sion of this article. Thanks to Jim Henderson of Cornell Law School for inspiring this article.

Erica L. Wagner is an assistant professor of information systems at Cornell Universitys School of Hotel Administration. She conducts research into the sociopolitical dimensions of technology.

Sue Newell is the Cammarata Professor of Management at Bentley College. She is also a visiting professor at Royal Holloway, University of London.

42 THE JOURNAL OF APPLIED BEHAVIORAL SCIENCE March 2006

The research question is therefore as follows: How is social order produced to cre- ate a workable ERP system? In preparation for our analysis, we consider coordinated action during information technology (IT) enabled change initiatives as well as the nature of social ordering. This is followed by a description of the research methodol- ogy, where we discuss a narrative perspective that helped us analyze how one organi- zation used stories to repair damaging setbacks in its ERP implementation.

COORDINATED ACTION

ERP packages are undeniably popular in organizations (Davenport, 2000) because of their ability to streamline and integrate business operations across an organizations value chain. Enterprise systems provide both a depth of information by function and also a breadth of information horizontally across the value chain because they are organized in an integrated manner around a central data repository. This provides a powerful business system infrastructure for organizations. However, accomplishing organizational integration requires changes that affect an unprecedented amount of people across the enterprise. In particular, the standardization of work practices across functional areas of the business often leads to resistance from stakeholder groups. Hence the challenge of making ERP work across the organization for multiple constituencies.

In exploring the challenge of organizational change associated with ERP imple- mentation, it is simplistic to believe that IT drives or even enables such change (Robey & Boudreau, 1999). Rather, we must recognize that there is a complex rela- tionship of reciprocal causality between IT and organization, with the outcomes emer- gent and difficult to predict in advance (Orlikowski, 2000). In particular, in relation to ERP, emergence is inevitable because the change that is required for the organization to benefit from the integrative potential of the software creates a condition of interde- pendence (Pfeffer & Salancik, 1978); the change will only be effective if parties in all affected departments embrace the initiative. It is therefore important that those involved coordinate their thoughts, actions, practices, and goals so as to be attentive to the interdependencies of the community (Boland, Tenkasi, & Teeni, 1994). This is because individuals do not create change; rather, it is through coordinated action that change takes place (Boland & Tenkasi, 1995; Brunsson, 1985). This need for coordi- nation poses problems in a university setting given their fiefdom-like structure where information systems have tended to develop organically to support the values of aca- demic freedom and scientific separateness (Allen & Kern, 2001; Cornford, 2000; Pollock, 1999) rather than coordination and control, which is the central value behind ERP systems.

Coordinated action is accessible by studying various controversial moments over time with an emphasis on how such coordinated outcomes are (are not) achieved (Boland & OLeary, 1991; Bruner, 1990). It is argued that woven into individual narra- tives of change are connections and politics that highlight the basis for coordination (Boland & Schultze, 1996; Bruner, 1990). We consider the basis of such coordination from the perspective of producing social order.

Wagner, Newell / REPAIRING ERP 43

THE PRODUCTION OF SOCIAL ORDER

Organizing by common aims or reciprocity are opposing ways in which people come together to seek a mutually advantageous position (Fuller, 1978). In an organiza- tional context it might be assumed that the best way to achieve change is through estab- lishing common aims with narratives that focus on achieving buy-in to a common vision. However, although organizations are built on common aims, it is dangerous to assume that all parties will want the same thing over time and that if different goals emerge it is just a matter of educating the dissenters so that they can see the truth (Fuller, 1978, p. 361). Moreover, the more stakeholders are affected by a change initia- tive, the more challenging it is to create and maintain common interests due to the dif- fuse nature of the effort (Latour, 1999) and due to differences in goals and beliefs about how best to approach and solve a particular problem (Pfeffer, 1992). These differences are the inevitable result of specialization because this means there are groups with dif- ferent backgrounds and training and so different thought worlds (Dougherty, 1992). The coexistence of these disparate views is particularly relevant when considering ERP implementations as a major form of organizational change because the softwares integrated design encourages the institutionalization of a dominant perspective across the organization resulting in the silencing of all other ways of working (Wagner & Newell, 2004). As a result, ERP implementations are notorious for attempting to imbue a social order that impacts power relations in a dictatorial mannereither all agree to the common vision or they keep their dissenting views to themselves.

However, given the existence of heterogeneous perspectives in an organization, Fuller (1978) argues that it is important that one is skeptical of organizing through common aims. Instead, he argues one should consider what the other party wants and how to best negotiate with them. This suggests that in such situations of interdepen- dence and divergence of goals and beliefs, it will be necessary to use power and influ- ence to get things done (Pfeffer, 1992). If power and influence are not used there will be a paralysis, reflecting an inability to mobilize sufficient political support and resources to take action (Pfeffer, 1992, p. 4). Although it may be possible to move for- ward using hierarchical power to impose a particular decision, this will not in itself lead to successful problem solving because as Pfeffer (1992) notes, A decision by itself changes nothing (p. 19). For example, although senior managers can dictate that a particular IS be implemented, this does not necessarily translate into IS use because there are many ways in which users can resist using the IS, at least as it was intended (Boudreau & Robey, 2005).

Other means must therefore be found to mobilize action in these situations of inter- dependence, and capitalizing on the norm of reciprocity is important. That there is no such thing as a free lunch is an often quoted saying that defines the norm of reciproc- ity that implies that we are obligated to future repayments of favors, gifts, invitations, and so on received from others (Gouldner, 1960). This is different from a straightfor- ward exchange based on a market transaction because the returned favor is not explic- itly specified. Rather, the favor implies a diffuse, generalized obligation for repayment sometime in the future. This norm facilitates transactions between individuals and groups over time and is extremely important in facilitating organizational change

44 THE JOURNAL OF APPLIED BEHAVIORAL SCIENCE March 2006

(Pfeffer, 1992). In a decision-making situation it suggests that although each party must stand up for their goals and needs, they must at the same time be aware of the needs and goals of others because at some time in the future, each might want what the other party can give. Central to this form of working then is that each party involved has an understanding of what makes the other tickwhat are their motivations and goals.

We recognize that both forms of social ordering are narratively produced in any social environment, and therefore during the ERP project under study here the pres- ence of reciprocity does not negate the notion of common aims occurring within dif- ferent pockets of the project. However, we have found that our analysis is facilitated by focusing on the form of the narrative that dominated a particular controversy. We explore these forms of social ordering as a means for understanding how the goal of creating a working ERP is achieved. We next turn to the narrative perspective that allowed us to explore how social order was produced during these contentious episodes.

METHOD

An in-depth longitudinal case study was conducted between 1999 and 2000 in Ivy University. The conceptual ideas of Fuller (1978) were applied to the field data post hoc, using these ideas as a lens through which to view the world, sensitizing us to particular issues (Walsham, 1993). In doing this, we focus on controversial episodes during the Ivy implementation, examining the narratives that emerged during these periods that helped to move the stalled project forward. Researchers have begun exploring the narratives that accompany ERP projects (Hall & Grant, 2005), and such data are increasingly being recognized as a valuable tool for ensuring preferred organi- zational outcomes (Denning, 2004). In particular, narrative data are helpful when exploring controversies because individual stories of negotiation speak on behalf of a network of interests and highlight conflicting agendas at work. So when analyzing the visionary narratives of a powerful stakeholder, one would interpret his or her words as representing a particular ontological network. As such, researchers can study the extent to which visionary narratives are repeated within the narratives of diverse actors. If we agree that it is only through coordinated action that change takes place (Boland & Tenkasi, 1995), then it follows that the appropriation and retelling of a par- ticular narrative is an expression of affiliation; a connecting up to a powerful group of interests that in turn helps to perpetuate a particular network and bring about social order. It is this interpretation of language as a tool for conveying particular agendas that focuses data analysis and helps researchers have the eyes to see the sociotechnical ensemble at work.

Stories of controversy need not be seen as solely divisive; Bruner (1990) reminds us that narrative accounts of controversy can provide insight into reparation:

Wagner, Newell / REPAIRING ERP 45

It is probably the case that human beings forever suffer conflicts of interest, with attendant grudges, factions, coalitions, and shifting alliances. But what is interesting about the fractious phenomena is not how much they separate us but how much more often they are neutralized or forgiven or excused. (p. 95)

To gain access to how such factious phenomena are resolved, one must be commit- ted to following change as it unfolds over time. As Latour (personal communication, November, 2001) argues, let us study the bombardment of offers for contradictory social groups because by sitting at the controversy of group formation we can compare group making to group making. To study group making, we focus on networks of action rather than particular individuals whose status and relevance may change over time (Czarniawska, 1997). We use multiple sources of narrative data that are analyzed to illuminate the interdependence of groups and their emerging controversies. Group formation directs the researchers gaze to specific controversies where options come to be (or fail to be) designed into the artifact. Following controversies provides insight into how groups negotiate for the dominance of their preferred work practices, so cre- ating social order. During software projects, the goal is to create a sociotechnical ensemble that works for its constituents. At Ivy, this approach gave us the eyes to see the emergence of new social orders around the IT artifacts that resulted from a dynamic change process.

Fieldwork

The field researcher conducted five site visits each lasting an average of 8 weeks. Evidence from 129 narrative interviews with 53 individuals was collected, totaling nearly 200 hours of empirical data. Additional data sources included technical and official documentation, observation of weekly project meetings, informal conversa- tions, and field researcher notes. The interviews adhered to the narrative interviewing (NI) convention, which provides a time frame to help structure the interview (e.g., What has happened since we last met?) and then encourages uninterrupted storytell- ing related to issues of central importance for the interviewee at the time (Bauer, 1996). All interviews were tape recorded and transcribed. Interviews were supple- mented with the field researchers interpretation of events and description of issues not captured in the transcripts. Direct observation of user group meetings and confidential project documentation were treated as other organizational narratives.

Not only did we seek to collect multiple types of data over time, we deliberately set out to gather evidence from a variety of actors throughout the organization who would provide diverse perspectives. The fieldwork protocol began with preliminary inter- viewing of team members, and transcripts were immediately produced and analyzed in terms of actors and issues that were raised during the interview. Importantly, both allies and antagonists were interviewed as a result of this protocol as we let the stories themselves drive our interview schedule.

Once all transcripts were preliminarily coded, data analysis followed a two- pronged approach. First, we determined significant content-based themes across manuscripts. Second, we grouped emerging narratives across ontological groups over time and analyzed the ways in which these networks were interconnected, thereby pro-

46 THE JOURNAL OF APPLIED BEHAVIORAL SCIENCE March 2006

viding insight into how coordinated outcomes were/were not achieved during the ini- tiative. These narratives were organized to illuminate the interconnectivity involved in negotiating through IT-enabled change. These storylines focused on resolving an unknown future and provided insight into the concept of social order.

CASE DESCRIPTION

In the summer of 1996, Ivy began modernizing its administrative information sys- tems through a 2-year ERP implementation that was officially announced as a project to benefit all constituencies. Ivy was one of the first U.S. universities to select ERP and contracted with Vision Corporation to develop software modules related to grants and contracts. That October, the project structure was formed and functional teams were created with cobusiness and -technical leaders. The teams were comprised mostly of Ivy middle managers from central administration whose permanent positions had been backfilled for the duration of the project. Although an experienced ERP project manager had been hired, the real authority lay with the teams who communicated directly with Ivys newly appointed VP for finance and administration.

Once under way, the project team found that Visions technical experts were not on site as often as expected, leaving team members working with an incomplete software suite and having to imagine how the grants and contracts module would be integrated. Worried about the projects progress after a year of high-level theorizing, Ivy hired consultants to audit the readiness of the software for its scheduled big bang imple- mentation in October 1998. Their findings caused Ivy to modify expectations and switch to a phased implementation strategy, adopting a revised deadline of July 1999 for the fully integrated suite. This date marked the beginning of Ivys year 2000 fiscal calendar, and as such it represented the deadline for retiring legacy systems at risk of the millennium bug. Ivy met its drop dead date in that the skeleton ERP was opera- tional on the first day of the new fiscal year, but the user interface and reporting envi- ronment still required significant development. The suboptimal rollout of the ERP was complicated by user resistance to the grants and contracts design. The academic con- stituencies who had expectations of an improved working environment were unable to complete crucial administrative tasks. Faculty demanded changes in the ERPs design as well as interim support for their administrators whose workload increased dramati- cally in the ERP-enabled environment. For more than 2 years, the project team was involved in postimplementation design changes before receiving buy-in from the academic community.

In the next section, we focus on two contentious episodes that nearly caused the derailment of Ivys project and analyze the narratives that emerged in each situation in terms of how social order was produced. These episodes have been selected because they illustrate the kinds of factions that one might typically come in contact with dur- ing large software projects where multiple stakeholders are involved. The first episode adopts an internal-external perspectiveimportant in light of the growing trend for contractual relationships between client organizations and external experts such as

Wagner, Newell / REPAIRING ERP 47

software vendors. The second episode concentrates on the goals of different organiza- tional groups who share the universitys mission but also have unique aims. In reality, a project of this length and complexity involves many episodes of contention. However, given our intent to explore episodes in depth, using the rich qualitative material that has been collected, we have selected two episodes that are archetypal of the ways stakeholders solve problems in projects and so produce social order.

CASE ANALYSIS

Ivy-Vision Group Formation

The Ivy-Vision strategic partnership was created to benefit both parties through the development of a higher education ERP suite. This product would form the basis of Ivys administrative infrastructure and would help Vision enter an untapped vertical market. Through coordinated action, Ivy and Vision were expecting to achieve this common aim because alone neither had all the necessary skills to create a higher edu- cation enterprise solution. However, during the first 2 years of the project, Vision failed to become enrolled in the project to the extent that Ivy expectedbeing largely absent from the project site.

I had a sense that [Vision] didnt even staff this thing for a year . . . when I asked questions at a cocktail party like . . . How big is your staff now? [Id learn] . . . its up to two people. They would catch themselves and say No, no, no. . . . Weve got seven now. Then theyd reveal things like We assigned our first person in April 98.

Ivy felt the level of resources provided by Vision was inadequate and a misrepre- sentation of their partnership agreement. The partnership had created a condition of interdependence (Pfeffer & Salancik, 1978) between the two parties, and Ivy did not control all of the conditions necessary to create the ERP product: Without the [Vision] guys here with us we were still talking philosophies and strategies . . . and had not even set up the system and figured out the decisions that needed to be made. In the absence of Vision representatives, Ivy failed to produce action in Brunssons (1985) sense because despite their cognitive activities and attempts at problem solving they were unable to effectively collaborate with Visiona necessary player in the process. The lack of resources provided by Vision was understandable given that Ivy repre- sented one client within a small vertical market that had limited growth opportunities; and they staffed the project to reflect this. Over time, those on the Ivy team came to realize this.

The thing about [Vision] is, they have made a commitment to higher ed but its much harder than . . . they thought it was going to be because they thought we were much more like governmental than we really are. . . . They have made an investment and they continue to invest in this market but you have to wonder how long theyre going to do that. There are only about 50 institutions that com- prise the [U.S.] market.

48 THE JOURNAL OF APPLIED BEHAVIORAL SCIENCE March 2006

In contrast, the project represented a major commitment for Ivy in terms of time and resources and was a substantial capital investment that was expected to have long-term implications for their operation and governance. A project leader reflects on this situation,

The strategic development partnership [is] a risky implementation because you dont know what youre going to getits dependent on a future releaseyoure not quite sure at the last minute whether your partners going to say Sorry, I can only do five of these features. . . . Its almost like trying to fly a plane and youre not quite sure whether youre going to land in LA or Chicago. So youre constantly recharting your flight path.

It is not surprising that Ivy and Vision had different views of the situation given that the former is a nonprofit organization seeking what is best for its constituencies and the latter an international publicly traded software company operating from a market- based perspective. As Dougherty (1992) notes, backgrounds and training create dif- ferent thought worlds that will result in variable interpretations, and as such one should be skeptical of relying on common aims (Fuller, 1978) to achieve organiza- tional stability.

When the original deadline arrived, Vision was still developing one of the modules and decided to completely redesign the other. Their absence coupled with the failure to produce tangible products led Ivy to organize its teams more closely and simulta- neously reinterpret their relationship with Vision. Ivy began considering alternative ways to achieve its goal given its reliance on Visionwhat did Vision want, and how best should Ivy negotiate with them? It was clear that Vision needed discipline exper- tise with regards to university grants management and Ivy began to realize that its modernization initiative was more complex because of the partnership with Vision due to the vendors need to create a marketable product.

Were going to have to find good ways to work together because [Vision] is committed to doing a pre- award [grant] system . . . because they want to market this product to higher education institutions and pre-award is an important part of the business. This is where we find we have the most duplica- tion because we dont need a pre-award system. (Functional team member)

Ivy had a well-respected grants management office with which they were happy that administered the external funding process for faculty. However, they lacked the ability to centrally manage the way in which faculty budgeted and spent their award dollars. So although Ivy would have preferred to leave preawards out of the scope of the ERP project, they realized that they were in an interdependent relationship (Pfeffer & Salancik, 1978) with Vision who saw pre- and postaward activities as parts of the same process, both of which were necessary to develop as part of a higher education solution that would be sold to research institutions. Ivy saw the potential to leverage their grants management expertise to obligate (Gouldner, 1960) Vision for repayment to Ivy sometime in the future.

As Ivy began to reinterpret their partnership with Vision by seeking compromise, the VP decided to mediate divergent goals and beliefs by using his power and influ- ence to get things done. The VP began by exerting his power over the project and hired

Wagner, Newell / REPAIRING ERP 49

an enterprise computing expert as technical director of administrative systems whose charge was to stop waiting for Vision and ramp up the technical development in time for the Y2K deadline. In addition to refocusing the team through strong leadership, the VP and newly hired director felt that it was necessary to give Vision an ultimatum. The technical director explains,

We were in crisis in late spring of 98. The grants management piece was simply not working and [the other module] had not even been delivered. . . . Thats the first time . . . that we were going to tell [Vision] that we were going to chuck em on grantsthat we were going to come up with an alterna- tive strategy. That would have been bad news for them because thats how they want[ed] to sell this market. . . . So we went through the discipline of [asking] What would we do instead? And it was ugly . . . but we were prepared to do it. We really meant businessif they couldnt execute we were not going to install.

Power relations are key to leveraging a reciprocal relationship between involved parties. Because a win-win situation is necessary to produce this form of social ordering, one must understand the goals of those involved and must have enough power to change the course of events, to overcome resistance, and to get people to do things that they would not otherwise do (Pfeffer, 1992, p. 30). The VP was influential in achieving concessions from Vision because of his ability to mobilize Ivy resources toward alternative development activities during a crucial time when the project was experiencing paralysis. In-house technical expertise was mobilized.

I had the geniuses working for me. . . . I asked them to do the detailed homework and gave them full empowerment . . . so that we could create the development that had to be done . . . and that required taking the geniuses from other groups like the Warehouse [team] and making it happen. (Technical director of administrative systems)

The VP also made a personal visit to Vision headquarters where he indicated that there was indeed no free lunch available for the vendor but rather they had an obliga- tion to repay Ivy for its commitment to developing the higher education product. While the project team geniuses continued to develop an alternative solution for grant and labor functionality, Vision produced what they called essential functional- ity in time for Ivys fiscal year 2000, and Ivy chose not to drop Vision as a development partner. Their continued involvement with the vendor is directly related to their own interests for a robust product, which they understand will only be achieved if they are able to influence Visions behavior and development trajectory by visiting the vendor and being involved in product functionality decisions, despite the costs involved:

We sent a team of people out to California for two weeks in June to do Beta testing on the [new] ver- sions of those applications. . . . It costs a lot of money . . . [and] resource time but we think that its worth it in the long run. (Team leader)

This example of conflict between Ivy and Vision despite their strategic alliance illustrates clearly that it is a mistake to assume that common aims will consistently pro- duce social order within a software project and that problems can be resolved by appealing to individuals sense of shared goals. Rather, we see here that what looked

50 THE JOURNAL OF APPLIED BEHAVIORAL SCIENCE March 2006

like common aims between Ivy and Vision at the beginning of the project tended to break down further along the road. This is because conflicting interests were also pres- ent at the same time as the common aim, and these other interests overshadowed the shared goal. We further analyze these forms of social ordering in the next subsection, focusing on the divergent aims of internal stakeholders.

Academic Constituencies Reject ERP Design of Grants Management

During implementation, team members focused on professionalizing the univer- sitys administrative practices to ensure institutional governance and mediate financial and regulatory risk. This goal was spearheaded by the financial management (FM) team leader who persuaded project members of this agenda, and through coordinated action they purposely excluded legacy grants management practices (based on a com- mitment accounting approach) from the ERP in preference for a corporate accounting approach based on time-phased budgeting interpreted as more rigorous. The connec- tions and politics that highlight the basis for coordination (Boland & Schultze, 1996; Bruner, 1990) are represented in the FM team leaders story.

I would say that the mentality that weve had . . . for managing is primitive . . . and its old fash- ioned . . . the corporate world left it . . . years ago . . . faculty think of things fundamentally wrong. We want to move people towards a management model where were going to ask [them] to put together a time-phased budget and management plan.

The FM team leader went on to liken the legacy accounting approach to Quicken a simplistic software program for the management of personal finances.

If they dont like it, we ought to fire emand get new users! . . . Its a . . . retreat. . . . I taught karate for yearsyou know what? If youre afraid to fight, youll never fight! Got to decide to get up there and get hit . . . [were] spending millions and millions of dollars to go forward, not to duplicate what we had . . . [Ivy] needs more than a copy of Quicken for each grantwe have 4,000 grants . . . we dont do that here any more. I meanwe just dont!

The rhetoric of this story excludes the possibility of reciprocity in favor of squash- ing the old ways of working. The content and tone of his message illustrates little respect for the different university thought worlds (Dougherty, 1992).

The team leader makes a mistake however when he equates his decision about the best way to manage grants through time-phasing with the capacity to make that change happen through coordinated action (Brunsson, 1985). When the ERP was rolled out, it was met with resistance from academic administrators who were unable to inform their faculty members about the financial details of their grant and contract awards using the time-phased approach embedded in the ERP. Faculty in turn became unhappy about the ERP because they were unable to receive the answers they needed to do their jobs effectively. At this time, the project entered paralysis because the team was unable to gain political support from the academic departments. It is in this moment of controversy that all parties involved realize the lack of common aims and

Wagner, Newell / REPAIRING ERP 51

begin to consider how they might advance their particular interests. A dissenting cen- tral leadership voice shares his interpretation of this controversy,

We took an environment that wasnt complex and added a level of complexity that was a 100 fold . . . in the old world people invented shadow systems around the accounting system in order to do their jobs. . . . They understood how to take faculty information here, and central university information there and make them both accurate. . . . But with the ERP central administrators win!

Frustrated by this situation and cognizant of the increased stress levels of their staff, powerful faculty members reminded the VP of his official promise that the ERP would improve university administrative practices for the entire community. Faculty and their administrators joined forces and used their power to secure a meeting with the provost, VP, and project leadership where they gave an ultimatum threatening to build local systems and use the ERP only as a data repository unless the legacy functionality replaced the time-phased approach. Their influence is illustrated in the following narrative of a faculty administrator:

I took the message over to the [project team], and said . . . We have looked at a creative way of using the ERP for [grants] and its become clear to us that we need commitments . . . were poised . . . to cre- ate our own system but I would like to present this as a university issue and I want to know whether or not you would like to join us in this effort. Boom, boom, boom. All of a sudden it . . . happened . . . overnight. They had a working group that quickly went into designing a customized system.

As a result of these meetings, the project began to move out of paralysis and estab- lish stable patterns of organizing through an agenda of reciprocity (Gouldner, 1960). The legacy commitment accounting system remained live and new, and bolt-on tools were developed that customized the ERP system so that faculty needs were met. The expectation was that by appeasing faculty with regards to grants management, the ERP as a whole would be more likely to succeed.

DISCUSSION

It is a much cited statistic that 70% of software projects fail (Standish Group Inter- national, 2001) and two thirds of ERP implementations fall short of expectations (Cliffe, 1999). Historically, ERP research has tended to focus on these costly project breakdowns or at least the problems associated with such IT-enabled change. This focus leads to what we argue is a skewed representation of ERP in context. There is a need therefore for more research that studies workarounds and forms of articulation work that enable people to make dynamically complex systems work in practice (Orlikowski & Iacono, 2001). We have explored this process through a study of Ivy where they created a working information system from ERP software after confront- ing multiple challenges that could have derailed the project. In doing this, we demon- strate how narrative data coupled with a conceptual framework of social ordering (Fuller, 1978) allowed us to see not just the sociotechnical controversies at Ivy but also

52 THE JOURNAL OF APPLIED BEHAVIORAL SCIENCE March 2006

their resolutionor as Bruner (1990) says, how the fractious phenomena are neutralized or forgiven or excused (p. 95).

Fuller (1978) argues that organization by common aims or reciprocity represents opposing ways in which goals are achieved. The rhetoric that surrounds both intra- and interorganizational alliances and partnerships implies consensus building as a founda- tional concept, suggesting that common goals should be the preferred option in situa- tions where cooperation among different parties is necessary. For example, the Asso- ciation for Operations Management (APICS) observes that strategic alliances are formed when neither party has the expertise to achieve their goal alone and so share and develop linked and common processes to increase the performance of both com- panies (APICS, 2004) and thereby achieving a common aim. Similarly, in the litera- ture on projects involving multiple stakeholders it is suggested that the emphasis should be on creating common goals, trust, and shared understanding (Nonaka, 1994).

This focus on organizing by common aims was evident in the early narratives in the Ivy case, where the common aim among the internal stakeholders was improved administrative efficiency and the common aim for the Ivy-Vision partnership was the development of an ERP system to modernize the administrative practices of higher education institutions. None of the parties possessed the skills to create these goals alone. However, by joining forces the entities expected to be able to create the desired solution by participating in organization by common aims. Although the impetus for creating such relationships, either internally or externally, is therefore common aims, over time the extent to which consensus is truly felt in more than name alonewhere it informs the actions of those involvedwill ebb and flow. Thus, in both the controver- sial episodes we have studied, the common aim among those involved did not prevent problems arising because of conflicting interests that threatened to derail the project; there was a real danger that the Ivy-Vision partnership would be disbanded and that a system would be produced that faculty departments would not use. Our narrative methodology allowed us to examine how stability and order was achieved around the ERP despite these conflicting interests. We argue that actors use narrative to express shared meanings among community members, but as complexity increases and change involves multiple functional groups, it becomes more difficult to rely on con- sensus and common aims across narrative accounts. Rather, during complex change processes individuals and groups are forming interpretive procedures that help them evaluate and judge the interests of other parties so that they can begin to negotiate on the basis of reciprocity rather than common aims. Narrative accounts can thus be analyzed to reveal the ways in which communities repair controversy.

For example, underneath the common aim of changing administrative practice the FM developed a narrative around professionalizing practice and was able to enroll the project team around this narrative. However, this meant that they designed grant accounting functionality based on corporate budgeting techniques that were directly at odds with faculty accounting methods. It was only during the use phase of the pro- ject that faculty interests surfaced through narratives of betrayal and demands for design modifications. Although initially this opposing narrative was dismissed as amateurish, eventually it was recognized that ignoring the demands embedded in

Wagner, Newell / REPAIRING ERP 53

this narrative would not help in creating the working IS that all were seeking. The result was that the team worked with faculty to develop a hybrid IS involving postinstallation customization. This hybrid was negotiated once the decision makers involved recognized reciprocal interests and stopped trying to impose a common aim (the FMs vision). Similarly, the tension between Ivy and Vision was resolved when both parties began to give a little, recognizing and accepting as legitimate the narra- tives of the other that were initially seen as unacceptableIvy to get a functioning ERP ready for their Y2000 deadline and Vision to develop the full range of grants management functionality.

Although it is possible in such situations for one party to impose its design require- ments as the common aim through the use of hierarchical power, this approach is likely to be limited (Pfeffer, 1981). For example, as the client in the relationship, Ivy could have withdrawn from the partnership agreement with Vision. However, this would not have helped them to develop the integrated system that they sought. Similarly, the team could have insisted on sticking to the time-budgeting approach, ignoring faculty demands for commitment accounting functionality. However, the likelihood of creat- ing a stable postimplementation environment where users actually use the software as intended was small [italics added] (Boudreau & Robey, 2005), especially given the setting where the use of absolute power is antithetical to the fiefdom-like structure of universities (Allen & Kern, 2001; Cornford, 2000; Pollock, 1999). In such contexts, attempting to impose a software system on different users without consideration of their interests and demands is likely to at best create a system that will be only partially used and at worst lead to complete system failure. Instead, what is important is to give a little so that the demands of users are met even though decision makers may believe that these demands are based on old thinking. Our analysis supports this idea by showing how bringing forward some of the valued legacy practices was essential to ensure buy-in as a necessary part of change. This may not be rational from the per- spective of the project team, in the sense that the underpinning legacy thinking may be unnecessary in the changed organizational context, but it is necessary to ensure the commitment and motivation of those who need to be enticed to use the system. Such irrationality, argues Brunsson (1985), is often the key to effectively implementing decisions in organizations. Over time, change involves a process of reordering the dominant organizational landscape and cycles of activity by making decisions about what will be carried forward into the future and what will be left behind to make room for the new (Latour, 1999, p. 71). In other wordswe may have to carry the past for- ward in time if we want users to come along with us into the future.

The coordinated action required during IT-enabled change initiatives means that actors confront an intense pace of work and must negotiate multiple and conflicting perspectives, priorities, and deadlines if a sociotechnical ensemble is to be accepted and used. It is through narrative accounts that actors articulate their situated under- standing of these negotiations and how they relate legacy work practice to future oper- ating environments. Our narrative methodology has helped us to craft an analysis that illustrates this transformation by highlighting the interconnectivity of detailed negoti- ations over time.

54 THE JOURNAL OF APPLIED BEHAVIORAL SCIENCE March 2006

CONCLUSIONS AND IMPLICATIONS

Today, more software is being purchased than written from scratch. In addition, the enterprise-wide nature of software means that costs escalate rapidly. The whole-house nature of such initiatives means that process and functional owners from across the value chain must be involved as well as user representatives from different parts of the organization. Successfully implementing such software therefore requires consid- erable organizational change. The scale and scope of change that is required also means that organizations experience problems. Research to date has tended to focus on these problems. However, given that ERP projects develop momentum and it is dif- ficult to pull the plug once the projects are ramped up, our approach has been to focus on how to make troubled software projects work so that an information system will be produced that is a good enough solution for all stakeholders.

Our research demonstrates that achieving such a workable solution does not mean that consensus must be sought or achieved in all instances as is articulated in the princi- ple of common aims and as is promoted in much of the academic literature on partner- ships and cross-functional project work. Rather, the important thing appears to be coordinating action that will allow goals to be achieved, even if this involves compro- mise along the way. Such compromise depends on reciprocity to produce social order and does not expect a shared aim as long as a solution can be negotiated. More research exploring when and how compromise encourages the norm of reciprocity would be valuable to understanding the production of a new social order around IT implementation.

So what are the practical implications of our analysis? First, it suggests that IT- enabled organizational change involves partnerships and projects that in their incep- tion are built on common aims. Such change initiatives will involve numerous contro- versial moments during which time opposing narratives emerge indicating divergent opinions about how to achieve the common aims. How these controversies are dealt with is likely to be critical in relation to whether a working system is produced. In this respect, our study indicates that it is important for managers to be aware of the emer- gence of opposing narratives and not to dismiss or ignore them, however irrational they may appear from their particular perspective. In Ivy for example, the project implementation team was able to initially marginalize the dissenting narratives about the problems of time-phased budgeting and design a system that ignored the concerns being expressed. However, in this case the dissenting narrative came from a powerful group in the organization, and it became clear that it was not going to be possible to ignore them, at least if they wanted the ERP to be fully used. Faculty in a university set- ting may be different in this respect from employees in other kinds of organizations. Indeed, the departmental administrators had been unable to make their voices heard when they were expressing similar concerns, and it was only the involvement of the faculty that finally led to functionality being added as a customization. In other organi- zational contexts, it may be easier to continue to ignore or dismiss the dissenting narra- tive(s). This is likely to lead to a situation where the ERP is implemented but remains under- or unused, explaining perhaps why many ERP implementations do not meet expectations and often involve substantial employee workarounds.

Wagner, Newell / REPAIRING ERP 55

The savvy manager will thus study the intra- and interorganizational dynamics throughout a software project actively looking for narratives that indicate emerging differences. Seeking out the formation of a group that connects up the interests of mul- tiple actors indicates that other groups are being dismantled as a result of this particular enrollment. For example, at Ivy the signing of the development contract with Vision represents the strengthening of one network and a weakening of interests related to a multivendor approach or the upgrading of legacy systems. For managers what is rele- vant is how this decision enables and constrains certain future scenarios both in terms of organizational transformation and IT design. We argue that it is this act of compar- ing that is most valuable for understanding the future trajectory of technological change because it directs ones gaze to specific controversies where options come to be (or fail to be) designed into the artifact. When the geniuses began working on an in- house alternative to the Vision modules we see the formation of another group that is working in opposition to the Vision programmers. It is at this place of controversy that we are positioned to see its resolution. It is during these times that ability to mobilize political resources is crucial. However, although mobilizing political resources at these times is important, it is often not the case that actually using these resources to impose ones solutions on the opposition is helpful. Even during the Ivy-Vision partnership controversy, although resources were mobilized (using geniuses to develop a home-grown solution), this was more to put pressure on Vision to increase resources on the project than to actually abandon the partnership in favor of develop- ing the system alone. Thus, what appears to be important at these points of controversy is to recognize the motivations and needs of the other party(s) and be willing to com- promise on the deliverables and negotiate with the partner in a way that helps create a good enough solution for both parties. In other words, using reciprocity as the basis for the production of social order, although not ideal, is far better than stagnating and developing resentments about the changed rules of the game (we were on the same page and now we are not) or imposing a solution that will later not be fully used.

A final implication of this principle of reciprocity is that in some organizations, as in Ivy, select customizations will be necessary if the ERP is to be made to work because these valued legacy practices cannot be met through configuration alone. This goes against all the advice about avoiding customization to ensure the benefits of upgrades and so on (Alshawi, Themistocleous, & Almadani, 2004). However, our findings indicate that customization to a system is sometimes necessary to achieving a working IS. In fact, if done strategically, customization can ease tensions and sustain employees commitment and motivation through the complex transitional phases that are part of any organizational change initiative.

In conclusion, in every case of success and failure there are stories of negotiation of actors maneuvering through change and orderit is in these details that one can learn how to integrate multiple perspectives while allowing for different motivating goals. Ivy provides us with a compelling story of IT-enabled change telling us that grand narratives exist as credos about how to workmandates from visionary leaders. But also there exist stories of action that can move one closer to achieving a mutually advantageous position. In conclusion, a goal of this study has been to challenge the normative idea of ERP as uncontrollable and instead show that such software is being

56 THE JOURNAL OF APPLIED BEHAVIORAL SCIENCE March 2006

made to work within organizations. In part, this story remains largely untold because the studies we are doing fail to show the resilience of individuals as they work through problematic situations.