An alternative model for managing technological change may enable

organizations to take advantage of the evolving capabilities^ entering practices,

and unanticipated outcomes associated with the use of new technolo^es.

An Improvisational Model forChange Management: The Case of

Croupware Technologies .

Wanda J. Orlikowski • J. Debra Hofman

In her discussion of technology design, Suchmanrefers to two different approaches to open seanavigation — the European and the Trukese:

"The European navigator begins with a plan — acourse — which he has charted according to certainunivei^al principles, and he carries out his voyage by re-lating his every move to that plan. His cftort through-out his voyage is directed to remaining 'on course.' Ifunexpected events occur, he must first alter the plan,dien respond accordingly. The Trukese navigator be-gins with an objective rather than a plan. He sets offtoward the objective and responds to conditions asthey arise in an ad hoc fashion. He utilizes informa-tion provided by the wind, the waves, the tide andctirrent, the fauna, the stars, the clouds, the sound ofthe water on the side of the boat, and he steers accord-ingly. His effort is directed to doing whatever is neces-sary to reach the objective."'

Like Suchman, we too find this contrast in ap-proaches instructive and use it here to motivate ourdiscussion of managing technological change. In par-ticular, we suggest that how people think about man-aging change in organizations most often resemblesthe European approach to navigation. That is, theybelieve they need to start with a plan for the change,charted according to certain general organizationalprinciples, and that they need to relate their actions

to that plan, ensuring throughout that the change re-mains on course.

However, when we examine how change occurs inpractice, we find that it much more closely resemblesthe voyage of the Trukese. That is, people end up re-sponding to conditions as they arise, often in an adhoc fashion, doing whatever is necessary to implementchange. In a manner similar to Argyris and Schon'scontrast between espoused theories and theories-in-use, we suggest that there is a discrepancy betweenhow people think about technological change andhow they implement it.' Moreover, we surest thatthis discrepancy significantly contributes to the diffi-culties and challenges that contemporary organiza-tions face as they attempt to introduce and effectivelyimplement technology-based change.

li-aditional ways of thinking about technologicalchange have their roots in Lewins three-stage changemodel of "unfreezing," "change," and "refreczing."'According to this model, the organization prepares forchange, implements the change, and then strives to re-gain stability as soon as possible. Such a model, which

Wanda J. Orlikowski is an associate profe^MV oj utjormation technologiesat the MIT Sloan School of Mafiagement, J. Debra Hofman is a seniorindustry analyst at Benchmarking Partners, Inc. When she coauthoredthis article, she was a research associate at the Center for InfonnatiouSystems Search, MIT Sloan School

Si.OAN MANAGKMtNT Rl.VIKW/WlNTER 1997 ORLIKOWSKI & HOFMAN 11

treats change as an event to be managed during aspecified period,' may have been appropriate for orga-nizations that were relatively stable and bounded andwhose functionality was sufficiently fixed to allowfor detailed specification. Today, however, givenmore turbulent, flexible, and uncertain organization-al and environmental conditions, such a model is be-coming less appropriate — hence, the discrepancy.

This discrepancy is particularly pronounced whenthe technology being implemented is open-ended andcustomizable, as in the case of the new informationtechnologies that are known as groupware.'* Group-ware technologies provide electronic networks thatsup[̂ )ort communication, coordination, and collabora-tion through facilities such as information exchange,

Amodel for managing changewould accommodate —indeed, encourage —

ongoing and iterativeexperimentation, use, and learning.

shared repositories, discussion forums, and messag-ing. Such technologies are typically designed with anopen architecture that is adaptable by end users, al-lowing them to customize existing features and cre-ate new applications.'' Rather than automating a pre-defined sequence of operations and transactions,these technologies tend to be general-purpose toolsthat are tised in different ways across various organi-zational activities and contexts. Organizations needthe experience of using groupware technologies inparticular ways and in particular contexts to betterunderstand how they may be most useful in practice.In such a technological context, the traditional changemodel is thus particularly discrepant.

I he discrepancy is also evident when organizationsuse information technologies to attempt unprecedent-ed, complex changes such as global integration or dis-tributed knowledge management. A primary exampleis the attempt by many companies to redefine and in-tegrate global value chain activities that were previous-ly managed independently. While there is typicallysome understanding up-front of the magnitude of

such a change, the depth and complexity of the inter-actions among these activities is fully understood onlyas the changes are implemented. For many org-aniza-tions, such initiatives represent a new ball game, notonly because they haven't played the game before butbecause most of the rtiles are still evolving. In a worldwith uncertain ailes, the traditional model for devis-ing and executing a game plan is very difficult toenact. And, as recent strategy research has su^ested,plajining in such circumstances is more effective as anongoing endeavor, reflecting the changing, unfoldingenvironments with which organizations interact."

In many situations, therefore, predefining the tech-nological changes to be implemented and accuratelypredicting their organizational impact is infeasible.Hence, the models of planned change that often in-form implementation of new technologies are lessdian effective. We stiggest that what would be moreappropriate is a way of tliinking about change that re-flects the unprecedented, uncertain, open-ended, com-plex, and flexible nature of the technologies and orga-nizational initiatives involved. Such a model wouldenable organizations to systematically absorb, respondto, and even leverage unexpected events, evolvingtechnological capabilities, emerging practices, and un-anticipated outcomes. Such a model for managingchange would accommodate — indeed, encourage —ongoing and iterative experimentation, use, and learn-ing. Such a model sees change management more asan ongoing improvisation than a staged event. Herewe propose such an ;dternative model ajid describe acase study of groupware implementation in a cus-tomer support organization to illtistrate the value ofthe model in practice. We conclude by discussing theconditions imder which such an improvisational modelmay be a powcrfU way to manage the implementa-tion and use of new technologies.

An Improvisational Model for ManagingChange

The improvisational model for managing technoloei-cal change is based on research we have done on theimplementation and use of open-ended informationtechnologies. The model rests on two major assump-tions that differentiate it from traditional models ofchange: First, the changes associated with technology

12 ORUKDWSKI & HOFMAN SLOAN MANAGKMENT RKVII'W/WINTKR 1997

Figure 1 An Improvisational Model of Change Management over Time

Opportunity-BasedChange

AnticipatedChange

AnticipatedChange

Opportunity-BasedChange

implementations constitute an on-going process rather than an eventwith an end point after which theorganiziuion can expect to return toa reasonably steady state. Second, allthe technological and organizationalchanges made during the ongoingprocess cannot, by defmition, beanticipated ahead of time.

Given these assumptions, our im-provisational change model recognizes three different change occur, the deployment of new technology

EmergentChange

EmergentChange

p gtypes of change: anticipated, emergent, and opportu-nity-based. These change types are elaborations onMintzberg's distinction between deliberate andemergent strategies.^ Here, we distinguish betweenanticipated changes — changes that are plannedahead of time and occur as intended — and emer-gent changes — changes that arise spontaneouslyfrom local innovation and that are not originally an-ticipated or intended. An example of an anticipatedchange is the implementation of e-mail softwarethat accomplishes its intended aim to facilitate in-creased, quicker communication among organiza-tional members. An example of an emergent changeis the use of the e-mail network as an informalgrapevine disseminating rumors throughout an orga-nization. This use of e-mail is typically not plannedor anticipated when the network is implementedbut often emerges tacitly over time in particular or-ganizational contexts.

We fiirther differentiate these two types of changesfi-om opportunity-based chaiiges — changes that arenot anticipated ahead of time but are introduced pur-posefully and intentionally during the change processin response to an unexpected opportunity, event, orbreakdown. For example, as companies gain experi-ence with the World Wide Web, they are finding op-portunities to apply and leverage its capabilities inways that they did not anticipate or plan before theintroduction of the Web. Both anticipated and op-portunity-based changes involve deliberate action, incontrast to emei^ent changes that arise spontaneouslyand usually tacitly from peoples practices with thetechnology over time.^

The three types of change build on each other it-eratively over time (see Figure 1). While there is nopredefined sequence in which the different types of

g poften entails an initial anticipated organizationalchange associated with the installation of the newhardware and software. Over time, however, use ofthe new technology will typically involve a series ofopportunity-based, emergent, and Rirther anticipatedchanges, the order of wliich cannot be determined inadvance because the changes interact with each otherin response to outcomes, events, and conditions aris-ing through experimentation and use.

One way of thinking about this model of changeis to consider the analogy of a jazz band. While mem-bers of a jazz band, unlike members of a symphonyorchestra, do not decide in advance exactly whatnotes each is going to play, they do decide ahead oftime what musical composition will form the basis oftheir performance. Once the performance begins,each player is free to explore ajid innovate, departingfrom the original composition. Yet the performanceworks because all members are playing within thesame rhythmic structure and have a shared under-standing of the rules of this musical genre. What theyare doing is improvising — enacting an ongoing se-ries of local innovations that embellish the originalstructure, respond to spontaneous departures and un-expected opportunities, and iterate and build on eachother over time. Using our earlier terminology, thejazz musicians are engaging in anticipated, opportu-nity-based, and emergent action during the course oftheir performance to create an effective, creative re-sponse to local conditions.

Similarly, an improvisational model for managingtechnological change in organizations is not a prede-fined program of change charted by managementahead of time. lather, it recognizes that technologicalchange is an iterative series of different changes, manyunpredictable at the start, that evolve from practical

SLOAN MANAGEMI.NI-Ri;vihw/WiNTEU ORLIKOWSKI & HOKMAN 13

experience with the new technologies. Using such amodel to manage change requires a set of processesand mechanisms to recognize the different types ofchange as they occur and to respond effectively tothem. The illustrative case we present next su^eststhat when an organization is open to the capabilitiesoffered by a new technological platform and willingto embrace an improvisational change model, it canachieve innovative organizational changes.

The Case of Zeta

Zeta is one of the top fifiy software companies in theUnited States, with $100 million in revenues andabout 1,000 employees. It produces and sells a rangeof powerful software products that provide capabili-ties such as decision support, executive inforniation,and marketing analysis. Zeta is headquartered in theMidwest, with sales and client-service field officesthroughout the world.

Specialists in the customer service department(CSD) at Zeta provide technical support via tele-phone to clients, consultants, value-added resellers,Zeta client-service representatives in the field, andother Zeta employees who use the products. Thistechnical support is ofi:en quite complex. Specialiststypically devote several hotirs of lesearch to each prob-lem, often searching through reference material, at-tempting to replicate the problem, and reviewing pro-gram source code. Some incidents require interactionwith members of other departments such as qualityassurance, documentation, and product development.The CSD employs approximately fifty specialists andis headed by a director and two managers.

In 1992, the CSD purchased the Lotus Notesgroupware technology within which it developed anew incident tracking support system (ITSS) to helpit log customer calls and keep a history of progresstoward resolving the customers' problems. Followinga successfiii pilot of the new system, the CSD decid-ed to commit to the Notes platform and to deployITSS throughout its department. The acquisition ofnew technology to facilitate customer call trackingwas motivated by a number of factors. The existingtracking system was a homegrown system that hadbeen developed when the department was muchsmaller and Zetas product portfolio much narrower.

The system was not real-time, entry of calls was hap-hazard, information accuracy was a concern, andperformance was slow and unreliable. It provided lit-de assistance for reusing prior solutions and no sup-port for the management of resources in the depart-ment. The volume and complexity of calls to theCSD had increased in recent years due to the intro-duction of new products, the expanded sophistica-tion of existing products, and the extended range ofoperating platforms supported. Such shifts had madereplacement of the tracking system a priority, as theCSD managers were particularly concerned that thehomegrown system provided no ability to track calls,query the status of particular calls, understand the

"he new database of informationserved as an unexpected,

informal leaming mechanism bygiving the specialists exposure to

a wide range of problemsand solutions.

workload, balance resources, identify issues and prob-lems before they became crises, and obtain up-to-diteand accurate documentation on work in progress andwork completed. In addition, calls would occasionallybe lost, as the slips of paper on which they wererecorded would get mislaid or inadvertently thrownaway.

• Introduction of ITSS. The initial introduction ofthe new ITSS system was accompanied by anticipat-ed changes in the nature of both the specialists' andmanagers' work. In contrast to the previous system,which had been designed to capture only a brief de-scription of the problem and its final resolution,11SS was designed to allow specialists to documentevery step they took in resolving a particular inci-dent. That is, it was designed to enable the captureof the fiill history of an incident. As specialists beganto use ITSS this way, the focus of their work shiftedfrom primarily research — solving problems — toboth research and documentation — solving prob-lems and documenting work in progress.

14 ORI iKinvsKi & HOFMANSi CAN MANAGEMENT REVIEW/WINTER 1997

The ITSS database quickly began to grow as eachspecialist documented his or her resolution processin detail. While documenting calls took time, it alsosaved time by providing a rich database of informa-tion that could be searched for potential resolutions.Moreover, this new database of information servedas an unexpected, informal learning mechanism bygiving the specialists exposure to a wide range ofproblems and solutions. As one specialist noted: "Ifit is quiet, I will check on my fellow colleagues to seewhat . . . kind of calls they get, so I might learnsomething from them . . . just in case somethingmight ring a bell when someone else calls." At thesame time, however, using the ITSS database as asole source of information did pose some risk be-cause there were no guarantees of the accuracy of theinformation. To minimize this risk, the specialiststacidy developed informal quality indicators to helpthem distinguish between reliable and unreUabledata. For example, resolutions that were comprehen-sively documented, documented by certain individu-als, or verified by the customer were considered reli-able sources of information.

In addition to these changes in specialists' work,the CSD managers' use of the new system improvedtheir ability to control the department's resources.Specialists' use of ITSS to document calls providedmanners with detailed workload information, whichwas used to justify increased headcount and adjustwork schedules and shift assignments on a dynamicand as-needed basis. ITSS also supplied managerswith more accurate information on specialists' workprocess, for example, the particular steps followed toresearch and resolve a problem, the areas in whichspecialists sought advice or were stalled, and thequality of their resolutions. As managers began torely on the ITSS data to evaluate specialists' perfor-mance, they expanded the criteria they used to dothis evaluation. For example, quality of work-in-progress documentation was included as an explicitevaluation criterion, and documentation skills be-came a factor in the hiring process.• Structural Changes. As the CSD gained experi-ence with and better understood the capabilities ofthe groupware technology, the managers introduceda change in the structure of the department to fur-ther leverage these capabilities. This change had not

been planned prior to the implementation of ITSS,but the growing reliance on ITSS and an apprecia-tion of the capabilities of the groupware technologycreated an opportunity for the CSD to redistributecall loads. In particular, the CSD established "firstline" and "second line" support levels, with juniorspecialists assigned to the first line, and senior spe-cialists to the second line. The CSD created partner-ships between the less experienced junior specialistsand the more experienced senior specialists. Front-line specialists now took all incoming calls, resolvedas many as they could, and then electronically trans-ferred calls to their second-line paitners when theywere overloaded or had especially difficult calls. Inaddition to handling calls transferred to them, seniorspecialists were expected to proactively monitor theirfrondine partners' progress on calls and to provideassistance.

While this partnership idea was conceptuallysound, it regularly broke down in practice. Juniorspecialists were ofi:en reluctant to hand off calls, fear-ing that such transfers would reflect poorly on theircompetence or that they would be overloading theirmore senior partners. Senior specialists, in turn, wereusually too busy resolving complex incidents to spendmuch time monitoring their junior partners' call sta-tus or progress. In response to this unanticipatedbreakdown in the partnership idea, the CSD man-agers introduced another opportunity-based struc-tural change. They created a new intermediary roletiiat was filled by a senior specialist who mediated be-tween the first and second lines, regularly monitoredjunior specialists' call loads and work in progress, anddynamically reassigned calls as appropriate. The newintermediary role served as a buffer between the ju-nior and senior specialists, facilitating the transfer ofcalls and relieving senior specialists of the responsibili-ty to constantly monitor their fi-ontline partners. Withthese structural changes, the CSD in effect changedthe prior undifFerentiated, fixed division of labor with-in the department to a dynamic distribution of workreflecting different levels of experience, various areasof expertise, and shifting workloads. In response tothe new distribution of work, managers adjustedtheir evaluation criteria to reflect the changed re-sponsibilities and roles within the CSD.

Another change that emeiged over time was a shift

SLOAN MANAGEMENT REVIEW/WINTER 1997 ORUKOWSKI & HOFMAN 15

in the nature of collaboration within the CSD from aprimarily reactive mode to a more proactive one. Be-cause all specialists now had access to the database ofcalls in the department, they began to go througheach others' calls to see which ones they could helpwith, rather th;ui waiting to be asked if they had a so-lution to a particular problem (which is how they hadsolicited and received help in the past). This shififrom solicited to unsolicited assistance was facilitatedby the capabilities of the groupware technology, thecomplex nature of the work, existing ev;iluation crite-ria that stressed teamwork, and the long-standing co-operative and collegial culture in the CSD. Severalspecialists commented: "Everyone realizes that we allhave a certain piece of the puzzle.... I may have onecritical piece, and Jenny may have another piece. . . .If we all work separately, we're never going to get thepuzzle together. But by everybody working together,we have the entire puzzle"; "Here I don't care whograbs credit for my work. . . . This support depart-ment does well because we're a team, not becausewe're all individuals."'" Managers responded to thisshift in work practices by adjusting specialists' evalu-ation criteria to specifically consider unsolicited help.As one manager explained: "When fm looking atincidents, I'll see what help other people have of-fered, and that does give me another indication ofhow well they re working as a team."• Later Changes. After approximately one year ofusing n SS, the CSD implemented two fiirther orga-nizational changes around the groupware technology.Both had been anticipated in the initial planning forITSS, although the exact timing for their implemen-tation had been left unspecified. First, the 1 i'SSapplication was installed in three overseas support of-fices, with copies of all the ITSS databases replicatedregularly across the four support sices (United States,United Kingdom, Australia, and Europe). This pro-vided al! support specialists with a more extensiveknowledge base on which to search for possibly hclp-fiil resolutions. The use of ITSS in all the stipport of-fices Rirther allowed specialists to transfer calls acrossoffices, essentially enacting a global stipport depart-ment within Zeta.

Second, the CSD initiated and funded the devel-opment of a number of bug-tracking systems thatwere implemented within groupware and deployed

in Zeta's depanments of prodtict development, prod-uct mani^ement, and quality assurance. These bug-tracking applications were linked into ITSS and en-abled specialists to enter any bugs they had discoveredin their problem resolution activities directly into therelevant products bug-tracking system. Specialistscotild now also directly query the status of particularbugs and even change their priority if customer callsindicated that such an escalation was needed. Special-ists in particular found this change invaluable. Forthe other departments, the link with I FSS allowedusers such as product managers and developers to ac-cess the ITSS records and trace the particular inci-dents that had uncovered certain bugs or specific useproblems. Only the developers had some reservationsabout the introduction of the bug-tracking applica-tion — reservations that were associated with the se-vere time constraints under which they worked to pro-duce new releases of Zeta prtxlucts.

In addition to the improved coordination and in-tegration achieved with other departments and of-fices, the CSD also realized further opportunity-based innovations and emergent changes within itsown practices. For example, as the number of inci-dents in ITSS grew, some senior specialists began torealize that they could use the information in the sys-tem to help train newcomers. By extracting certainrecords from the ITSS databa.se, the specialists createda training database of sample problems with whichnewly hired specialists could work. Using the com-munication capabilities of the groupware technology,these senior specialists could monitor their trainees'progress through the sample database and interveneto educate when necessary. As one senior specialistnoted: "We can kind of keep up to the minute ontheir progress. . . . If they're on the wrong track, wecan intercept them and say, 'Co check this, go look atthat.' But it's not like we have to actually sit withthem and review things. Its sort of an on-line, inter-active thing." As a result of this new training mecha-nism, the time for new speciiilists to begin taking ctis-tomer calls was reduced from eight weeks to aboutfive.

Another change was related to access control. Anongoing issue for the CSD was who (if anyUxly) out-side the CSD should have access to the ITSS databasewith its customer call information and specialists' work-

16 ORIIKOWSKI & HOFMAN SLOAN MANAI;KMENT REVIEWTWINH ER 1997

Figure 2 Zeta's Improvisational Management of Change over Time

ITSS System forEntering Calls andDocumenting Process

Structural Change• Front/Back Lines• Intermediaries

ITSS ExtendedGlobally

ProactiveCollaboration

iTSS Linked toBug System

Morms aboutQuality andDocumentation

ITSS asTrainingMechanism

AccessControlProtocols

in-progress documentation. This issue was not antici-pated before the acquisition of the technology. Whilethe managers were worried about how to respond tothe increasing demand for access to ITSS as the data-base became more valuable and word about its con-tent spread throughout the company, they continuedto handle each access request as i: came up. Overtime, they tised a variety of control mechanisms rang-ing from giving limited access to some "trusted" indi-viduals, generating summary reports of selected ITSSinformation for others, and refusing any access to stillothers. As one manager explained, only after sometime did they realize that their various ad hoc re-sponses to different access requests amounted to, inessence, a set of rules and procedures about accesscontrol. By responding locally to various requests andsituations over time, an implicit access control policyfor the use of ITSS evolved and emerged.

Zeta's Change ModelAlong with the introduction of the new technologyand the development of the ITSS application, theCSD first implemented some planned organization-al changes, expanding the specialists' work ro includework-in-progress documentation and adjusting themanagers' work to take advantage ot the real-time ac-cess to workload information. (Figure 2 representsthe change model aroimd the groupware technologythat Zeta followed in its CSD.) 7"he changes were an-ticipated before introducing the new technology. Asspecialists and managers began to work in new wayswith the technology, a number of changes emerged inpractice, such as the specialists developing norms todetermine the quality and value of prior resolutions,and managers paying attention to documentationskills in hiring and evaluation decisions.

Building on these anticipated and emergent changes,the CSD introduced a set of opportunity-basedchanges, creating junior-senior specialist partnershipsto take advantage of the shared database and commu-nication capabilities of the technology and thenadding the new intermediaiy role in response to theunexpected problems with partnership and work re-assignment. The CSD did not anticipate thesechanges at the start, nor did the changes emergespontaneously in working with the new technology.Rather, the CSD conceived of and implemented thechanges in situ and in response to the opportunitiesand issues that arose as it gained experience and bet-ter understood the new technology and their particu-lar use of it. This change process around the group-ware technology continued through the second yearat Zeta when some anticipated organizational changeswere followed by both emergent and opportunity-based changes associated with unfolding events andthe learning and experience gained by using the newtechnology in practice.

Overall, what we see here is an iterative and ongo-ing series of anticipated, emergent, and opportunity-based changes that allowed Zeta to learn from practi-cal experience, respond to unexpected outcomes andcapabilities, and adapt both the technology and theorganization as appropriate. In effect, Zeta's changemodel cycles through anticipated, emergent, andopportunity-based organizational changes over time.It is a change model that explicidy recognizes the in-evitability, legitimacy, and value of ongoing learningand change in practice.

Enabling Conditions

Clearly, there were certain aspects of the Zeta organi-

SLOAN MANAGEMENT Ri;v!r.w/WiNTER 1997 ORUKOWSKI & HODMAN 17

zation that enabled it to effectively adopt an improvi-sational change model to implement and use thegroupware technology. Our research at Zeta and othercompanies suggests that at least two sets of enablingconditions are critiail: aligning key dimensions ot thechange process and dediaiting resources to provideongoing support for the change process. We considereach in turn.

Aligning Key Change DimensionsAn important inHuence on the effectiveness of anychange process is the interdependent relationshipamong three dimensions: the technology, the organi-z;itional context {including culture, structure, roles,and responsibilities), and the change model used tomanage change {see Figure 3). Ide;illy, the interactionamong these three dimensions is compatible or, at aminimum, not in opposition.

First, consider the relation ot the change modeland the technology being implemented. When the

A least two sets of enablingconditions are critical:

aligning key dimensionsof the change process and

dedicating resources to provideongoing support for the change

process.

technology has been designed to operate like a "blackbox," allowing little adaptation by users, an improvi-sational approach may not be more effective than dietraditional approach to technology implementation.Similarly, when the technology is well establishedand its impacts are reasonably well understood, a tra-ditional planned change approach may be effective.However, when the technology being implementedis new and imprecedented and, additionally, is open-ended and customizable, an improvisational modelproviding the flexibility for organizations to adaptand learn through use becomes more appropriate.Such is the c;ise, we believe, with the groupware tech-nologies available today.

Figure 3 Aligning the Change Model, the Technology,and the Organization

Change Model

Organization Technology

Second, the relation of the change model to orga-niz;uional context is also relevant. A flexible changemodel, while likely to be problematic in a rigid, con-trol-oriented, or bureaucratic culture, is well suited toan informal, cooperative culture such as the one at theCSD. In another study, we examined the MidQ) or-ganizations successful adoption and implementationof CASE {computer-aicied software engineering) toolswithin its information systems organization." WhileMidCo, a multinational chemical products companywith reventies of more than $1.5 billion, was a rela-tively traditional organization in many ways, key as-pects of its culture — a commitment to total qu;ilitymanagement, a focus on oi^nizational learning andemployee empowerment, as well ;is a long-term out-look — were particularly compatible with the impro-visational model it used to manage ongoing organiza-tional changes around the new software developmenttechnology.

Finally, there is the important relationship betweenthe technology and the organiziuional context. AcZeta, the CSD's cooperative, team-oriented culturewas compatible with the collaborative nature of thenew groupware technology. Indeed, the CSDs existingculture allowed it to take advantage of the opportunityfor improved collaboration that the groupware tech-nology afforded. Moreover, when existing roles, respon-sibilities, and evaluation criteria became less salient, theCSD managers expanded or adjusted them to reflectnew uses of the technology. Cximpare these change ef-forts to those of Alpha, a profe.ssion;il sewices firm thatintroduced the Notes groupware technology to lever-age knowledge sharing and to coordinate distributedactivities.'' While the physical deployment of group-ware grew very rapidly, anticipated benefits were real-

18 ORUKOWSKI & HOFMAN SLOAN N4ANAGEMENT REVIEW/WINTER 1997

ized much more slowly. Key to the reluctance to usegroupware for knowledge sharing was a perceived in-compatibility between the collaborative nature of thetechnology and the individualistic and competitivenature of the organization. As in many professionalservices firms, Alpha rewarded individual rather thanteam performance and promoted employees basedon "up or out" evaluation criteria. In such an envi-ronment, knowledge sharing via a global Notes net-work was seen to threaten status, distinctive compe-tence, and power. In contrast to Zeta, managers atAlpha did not adjust policies, roles, incentives, andevaluation criteria to better align their organizationwith the intended use and capabilities of the tech-nology they had invested in.

Dedicating Resources for Ongoing SupportAn ongoing change process requires dedicated supportover time to adapt both the organization and the tech-nology to changing organizational conditions, usepractices, and technologiail capabilities. Opportunity-based change, in particular, depends on the ability ofthe organization to notice and recognize opportunities,issues, breakdowns, and unexpected outcomes as tlieyarise. This reqtiires attention on the part of appropriateindividuals in the organization to track technology useover time iuid to initiate organiziuional and technologi-cal adjusmients that will mitigate or take advantage ofthe identified problems and opportunities.

At Zeta, the managers and technologists played thisrole, incorporating it into their other responsibilities.So, for example, the managers adjusted the structureof their department by introducing first-Iine/second-line partnerships to facilitate a dynamic division otlabor and then made Rirther adaptations by introduc-ing an intermediary role to overcome some unantici-pated (difficulties associated with the initial change.Similarly, the technologists working with the CSDincorporated enhancements to the ITSS system asthey realized ways to improve ease of use and accesstime. The CSDs commitment to noticing and re-sponding to appropriate changes did not end afterthe implementation of the technology. The managersclearly realized that the change process they had em-barked on with the tise of groupware was ongoing, asone manager noted: "We've had ITSS for two years.I'm surprised that the enthusiasm hasnt gone avray....

I think it's because it's been changed on a regularbasis. . . . Knowing that [the changes are going to geeimplemented] keeps you wanting to think about itand keep going."

Ongoing change in the use of groupware technolo-gy also requires ongoing adjustments to the technolo-gy itself as users learn and gain experience with thenew technology's capabilities over time. Withoutdedicated technology support to implement theseadaptations and innovations, the continued experi-mentation and learning in use central to an improvi-sational change model may be stalled or thwarted.At Zeta, a dedicated technology group supported theCSD s use of groupware and ITSS. Initially consist-ing of one developer, this group grew over time asgroupware tise expanded. Aft:er two years, the groupincluded four full-time technologists who providedtechnology support for the various systems that hadbeen deployed within Zeta via the Notes platform.The group also maintained strong ties with all theirusers through regular meetings and communica-tions. This dedicated, ongoing technical support en-sured that the technology would continue to be up-dated, adjusted, and expanded as appropriate.

The value of ongoing support to enable ongoing or-ganizational and technological change was similarlyimportant in another organization we studied, theR&D division of a large Japanese manufacturingfirm." A newly formed product development teamwithin the R&D division installed a groupw;ire tech-nology, the Usenet news system (a computer confer-encing system). Similar to the CSD at Zeta, the team'sLLse of this new technology also iterated among antici-pated, emergent, and opportunity-based changes overtime. Here, a small group of users who had previouslyused the groupware technology took on the responsi-bility to man^e and support its ongoing use for them-selves and their colleagues. They tracked technologyusage and project events as tluy unfolded, respondedas appropriate with adjustments to communicationpolicies and technology Rinctionality, and proactivelymade changes to the team's use of the conferencing sys-tem to lever^e opportunities as they arose.

Conclusion

Global, responsive, team-based, networked — these

SLOAN MANAGRMENT RKVII-.W/WINTER 1997 ORUKOWSKI & HOFMAN 19

arc the watchwords for organizations of the nineties.As managers redesign and reinvent organizations in anew image, many are turning to information tech-nologies to enable more flexible processes, greaterknowledge sharing, and global integration. At thesame time, effectively implementing the organization-al changes associated with these technologies remainsdifficult in a turbulent, complex, and uncertain envi-ronment. We believe that a significant factor con-tributing to tliese challenges is the growing discrepan-cy between the way people think about technologicalchange and the way they actually implement it.

We propose that people's assumptions about tech-nology-based change and the way it is supposed to

'he improvisational modelrecognizes that change is

typically an ongoing processmade up of opportunities and

challenges that are not necessarilypredictable at the start.

happen are b;ised on models that are no longer ap-propriate. Traditional models for managing technol-ogy-based change treat change as a sequential seriesof predefined steps that are bounded within a speci-fied time. With these models as a guide, it makessense to define — as the European navigator does —a plan of action in advance of the ch;mge and trackevents against the plan, striving throu^out the changeto remain on track. Deviations from the intendedcourse — the anticipated versus the actual — thenrequire explanation, the subtle (and sometimes not-so-subde) implication being that there has been somefailure, some inadequacy in planning, that has led tothis deviation. Indeed, many organizational mecha-nisms such as btidgeting and resource planning areb;ised on these notions. The problem is that changeas it actually occurs today more closely resembles thevoyage of the Trukese navigator, and the models andmechanisms most commonly used to think aboutand manage change do not effectively support thisexperience of change.

We have offered here an iniprovisacional changemodel as a different way of thinking about managingthe introduction and ongoing use of informationtechnologies to support the more flexible, complex,and integrated structures and processes demanded inorganizations today. In contrast to traditional modelsof technological change, this improvisational modelrecognizes that change is typically an ongoing pro-cess made up of opportunities and challenges thatare not necessarily predictable at the start. It defines aprocess that iterates among three types of change —anticipated, emergent and opportunity-based — andthat allows the organization to experiment ;uid lc<ini asit uses the technology over time. Most imponantly, itoffers a systematic approach with wliich to understandand better manage the realities of technokigy-basedchajige in today's organizations.

Becatise such a model requires a tolerance for flex-ibility and uncertainty, adopting it implies that man-ners relinquish what is often an implicit paradigm of"command and control.'" An improvisational model,however, is not anarchy, and neither is it a matter of"muddling through." We are not implying that plan-ning is unnecessary or should be abandoned. We iiresuggesting, instead, that a plan is a guide rather thana blueprint and that deviations from the plan, ratherthan being seen as a symptom of failure, are to be ex-pected and actively managed.''

Rather than predefining each step and then con-trolling events to fit the plan, management creates anenvironment that facilitates improvisation. In such anenvironment, management provides, supports, andnurtures the expectations, norms, imd resources thatguide the ongoing change process. Malone refers tosuch a style of manning as "cultivation.""' Consideragain the jazz band. While each band member is freeto improvise during the performance, the result istypically not discordant. Rather, it is harmonious be-cause each player operates within an overall frame-work, conforms to a shared set of values and norms,and has access to a known repertoire of rules and re-sources. Similarly, while many changes at Zeta's CSDwere not planned, they were compatible with theoverall objectives and intentions of the departmentsmembers, their shared norms and team orientation,and the designs and capabilities of the technology.

Effectively executing an improvisational change

20 ORUKOWSKI & HOFMAN SLOAN MANAGEMENT REVIEW/WINTER 1997

model also requires aligning the technology and theorganizational context with the change model. Suchalignment does not happen auton^atiaJly. It requiresexplicit, ongoing examination and adjustment, whereand when necessary, of the technology and the orga-nization. As such, mechanisms and resources allocat-ed to ongoing support of the change process are criti-cal. Tracking and noticing events and issues as theyluifold is a responsibility that appropriate members ofthe organization need to own. Along with the re-sponsibility, these organizatioiiiil members require theauthority, credibility, influence, and resources to im-plement the ongoing changes. Creating the environ-ment; aligning the technology, context, and changemodel; and distributing the appropriate responsibilityand resources are critically important in the effectiveuse of an improvisational model, particularly as theyrepresent a significant (and therefore challenging) de-parture from the standard practice in effect in manyorganizations.

An improvisational model of change, however,does not apply to all situations. As we have noted, itis most appropriate for open-ended, customizabletechnologies or for complex, unprecedented change.In addition, as one reviewer noted, "Jazz is not ev-eryone's 'cup of tea.'. . . Some people are incapable ofplaying jazz, much less able to listen to what theyconsider to be 'noise.'" We noted above that somecultures do not support experimentation and learn-ing. As a result, they are probably not receptive to animprovisational model and are less likely to succeedwith it. As these organizations attempt to implementnew organizational forms, however, they too mayfind an improvisational model to be a particularlyvaluable approach to managing technological changein the twenty-first century. •

ReferencesWe would like to thank the editor and revieivers for their helpfid com-ments on an earlier version of this manuscript. We gratefully appreciatethe research support of MIT's Center for Coordination Science and Centerfor Infnntwtiou Systems Research.1, Berrcman (1996, p. 347} as cited in:L. Suchman. Plans and Sittiated Actions: The Problem of Human

Machine Communication (Cambridge, En^and: Cambridge UniversityPress, 1987), p. vli.

2. C. Argyris and D.A. Schon, Organizational Learning (Reading,Massachusetts: Addison Wesley, 1978).

3. K. Lcwin, "Group Decision and Stxial Change," in E. Newcombeand R. Harley, eds., Reading in Social Psychology (New York: HenryHolt. 1952), pp. 459-473; and

T.K. Kwon and R.W. Zmud, "Unifying the Fragmented Models ofInformation Systems Implemcncition," in R.j. Boland, Jr., and R.A.Hirschheini, eds.. Critical Issues in Infomiaiion Systems Research (NewYork: John Wiley. 1987), pp. 227-251.

4. A.M. Pcttigrcw, The Awakening Gi4int {OvikivA, England: BlackweilPublishers. 1985).

5. Not all groupware [ccKiiologies are flexible and customizable (e.g.,fixed-function e-mail systems). We are interested here only in thosethat are (e.g.. Lotu.s Notes).

6. D. Dejean and S.B. Dejean, Lotus Notes at Work (New York: bitusBooks, 1991); andT.W. Malone. K.Y. Lai, and C. Fry. "Experiments with OVAL: ARadically Tiiilorable 'I'ool for Cooperative Work (Toronto. Canada:IVoceedings of the Third Conference on Computer-Supported Coop-erative Work, Novt:inbcT I ')92). pp. 289-297.

7. H. Mint7bei^, "The Fall and Rise of Strategic Planning," HarvardSWwfif flw/rtf, volume 73, Januaiy-Februaiy 1994. pp. 107-114; andR.G. McGrath and I.C. MacMillan, "Discovery-Driven Planning,"Harvard Business Review, volume 73. July-August 1995. pp. 44-54.

8. H. Mintzbei^. "Crafting Strategy," Harvard Business Review, vol-ume 65. July-August 1987. pp. 66-75.

9. W.J, Orlikowski, "Improvising Organizational Transformationover Time: A Situated Change Perspective," Information SystemsResearch, volume 7, March 1996, pp. 63-92.

10. W.J. Orlikowski. "Evolving with Notes: Oi^ani/ationd Changearound Groupware Technology" (Cambridge. Massachusetts: MITSloan School of Management, working paper 3823, 1995).

U. M.l- Gallivan. J.D. Hofman, and W.J. Oriikowski, "ImplementingRadical Change: Gradual Versus Rapid Pace" (Vancouver, BritishColumbia, Canada: Proceedings of the Fifteenth Internationd Confer-ence on Information Systems, l4-17December 1994, pp. 325-3.^9).

12. W.J. Orlikowski, "Lciming from Notes: Organizational Issues inGroupware Implementation" (Toronto, Canada: Proceedings of theThird Conference on Computer-Supported Cooperative Work,November 1992, pp. 362-369).

13. W.J. Orlikowski, J. Yates, K. Okamura, and M. Fujimoto, "ShapingPJfCtronic Communication: The Metastructuring of Technolojjy inUse," Organization Science, volume 6, July-Augusi 1995. pp. 423444.

14. S. Zuboff. In the Age of the Smart Machine (New York: Basic

Books, 1988).15. Suchman (1987).16. T.W. Malone, informal conversation, 1996.

Reprint 3821

SLOAN MANAGF-MENT RLVIKW/WINTLK 1997 ORLIKOWSKI & HOFMAN 21