Mostbehavioral studies of the social andinformationprocessingarchitectures of computerbased information systems (CBIS) focus on the early stages of their life cycles, their conception, design, adoption, and organizational implementation. Because of this focus on earlyperiodsinthelifecycleofCBIS, wearemostawareoftheorganizationaldramasinthese early stages. This article examinesone aspectofthe laterstages ina CBIS' life cycle: theways in which organizational actors select and implement enhancements to "existing" CBIS. The implementation of computer-based information systems is not mechanical If implementing a new technologyenhancedaspects of worklife equally forallparticipants, allmight agree on implementation schedules, strategies, and use of the CBIS. However, CBIS implementations are often accompanied by disagreements and delays, and sometimes failures. There are two streams of research on information systems implementation: procedural analyses of implementation and studies of the political dimensions of implementation. This study builds on the political stream of information systems studies and organizational studies. The political studies of information systems implementation have traditionally focused on the early stages of implementation Readers mightassume thatonce a CBIS has been successfully implemented political elements are less central to its routine administration. We show how political campaigns can continue throughout the life of a CBIS and are equally critical in maintaining or shifting the balance of powen This paper explains the nature of political mobilization in some detail The primary data are drawn from an extensive case study of a medium sized manufacturing firm (PRINTCO) which operates a complex computerized inventory control system shared by several departments. We also show how the organization of a CBIS does not simply evolve; rather, key actors shape the developmental trajectory of the organization of computing. (A developmental trajectory for a system is a sequence of social and technical configurations through which it has developed and a sequence of future configurations) Usually there is a range of variation in future sequences, and different groups may prefer different developmental trajectories. Inordertogaincontroloverthedevelopmentaltrajectory, keyactorsattempttocreatesetsof procedures and beliefs about the computing arrangements which other participants will accept as legitimate. These actors develop long-term strategies to mobilize support fortheir own preferences and to block the emergence of conflicting preferences. Participants in different work groups have different computing preferences which derive from their own lines of work. Subunits within an organization push for information system architectures and data access in forms which secure their own interests and enable control or significant influence over others.

At any given time the infrastructure for providing computing services is structured, but not for everyone's simultaneous convenience. (Infrastructure refers toresources and procedures which support the efficient use of some focal computing resource). It includes access to programmers, terminals, or appropriate computer- based reports; provision of training in system usage or programming languages; or participation in decision- making about appropriate procedures and priorities in system development Those groups which are successful indevelopingandmaintainingfavorablearrangementsamplifyexistingstructuresincrementalls Because computing resources are insufficient to meet all actor's preferences simultaneosly, dominant coalitions can build powerby guidingthe development of a system to their own advantage and limiting other groups Theincremental developmentofa CBIS takes place overaperiod ofyearsratherthanweeks. The specific directioninwhichpowerfulactors guidea CBIS hasimportantconsequences for others inthe organizationinthatsome groups willbe betterservedthanothers. Overtimethe organization of the system becomes a"taken- for-granted" way of managing and coordinating workactivities. Asfinancialandideological commitments areincurred,itmaybecometoo costly to radically alter its developmental trajectory. We identifytwo key aspects of campaigns for computing: (1) a structural dimension, and (2) an ideological dimension. The structural dimension refers to the standardized arrangements for providing computing services (e. g., the infrastructure) and the process by which they become woven into organizational life and institutionalized. The ideological dimension focuses attention on the articulation of a world view which takes on meaning in the social worldofan organization.Key actors seeklegitimacy fortheircampaigns by convincingothers that their world view makes sense. In the paper we describe selected strategies key actors use to mobilize support for their preferences and to quiet opposition. The struggles for contzol might be "perverse" if power and control were the only issues. At PRINTCO we found that a coalition of manufacturing managers coupled their initiatives for control with material policies that could improve organizational efficienci However, efficiency was not the only issue since these manufacturing managers did not engage in many varied experiments to improve material efficiencies. They engaged in a relatively narrow array of strategies largely tied to their computerized inventory control system. The language of efficiency was both "real" and an acceptable rationale for leveraging organizational influence. Manyorganizationshavenowlivedwithseveralsuccessivegenerationsofthe"same"kindof CBIS. By common standards of successful implementation, the computerized inventory control system atPRINTCO was adequately implemented. The company was also successfulin growing rapidly and reaping good profits. The computerized inventory control system was heavilyused for sixyears, centraltothe operations ofthe organization, andthe subjectof tremendous continuing commitment However, its operation was not smooth, and its enhancement was the subject of several failed initiatives, including a major software conversion. It did not "evolve." A coalition of manufacturing managers tried to move the system, and the associated computing environment along a particular developmental trajectory. Their efforts were embodied in a series of campaigns which gave continuing life to the computerized inventory control system. This coalition dominated the local computing environment, but had neither perfectinformationnottotal control Some oftheir campaign strategies failed Inone period, they released substantial control over computing resources through a local "micro-revolutioni" but rapidly regained control when they appreciated its scale. The paper explains how key actors developed a variety of structural and ideological strategies to mobilize support for the arrangements they preferred and to quiet opposition. These structural arrangements became institutionalized; they were taken for granted and fit together in a mutually reinforcing complex. CBIS live and develop through the energies of their promoters rather than "evolve" through a life of their own.