Knowledge Diffusion through “Strategic Communities”

Several years ago, Chase Manhattan Bank embarked on an ambitious plan to develop a common back-office processing system for its overseas branches.1 The bank established at headquarters a cross-functional team of information systems, finance, and operations professionals to develop the design and lay out the implementation approach. Later, the bank added branch operations managers to meet “local” requirements. These managers retained their reporting relationships within their country organization structure, yet headquarters staff managed and evaluated their work on the project team. Conflicts were inevitable. After months of frequently fractious meetings, the headquarters staff prevailed, and a small group of branch managers reluctantly agreed to support implementing the new system on a pilot basis. Because the team concentrated on reconciling differences rather than on understanding similarities, the project was extended over several years. Ultimately, the bank achieved its objective of operating on a common infrastructure, but at high political, human, and competitive costs.

When facing a similar need, a large Korean chaebol took a different approach.2 To achieve its objective of global expansion, during the past decade the company focused on building production and marketing capabilities in many diverse geographies. Developing management control and reporting systems was lower priority and was principally the responsibility of local staff in each country. Under pressure to improve the timeliness and accuracy of financial reporting, the small corporate staff asked the functional managers in each country to make additional investments in the infrastructure of their management control systems. General guidelines were communicated at occasional meetings in Korea. Although the local managers discussed the generic applicability of these guidelines, they did not have a formal mechanism to identify substantive changes and elaborate on requirements. Local managers exercised substantial latitude in system specifications and implementation approach. Thus, although the reporting systems improved, concerns remained about the reliability and consistency of information across countries.

These two efforts are examples of infrastructure transitions, complex undertakings that require the cooperation of broad swaths of organizational resources. Not only is the economic justification for such efforts frequently unclear, but the legacy infrastructure also must remain operational during the transition. In managing these transitions, most global organizations establish teams drawn from headquarters staff and line business units, as did Chase and the Korean firm. Typically, such teams operate in a matrix structure. The way they function follows one of two patterns. In what we term a “thick” matrix (e.g., Chase), team leaders from headquarters interpret corporate strategy and ensure compliance through governance processes. In a “thin” matrix (e.g., the chaebol), team members from line business units interpret corporate strategy and determine the extent to which their units should comply.

Xerox Corporation, in making the transition from a proprietary information technology (IT) infrastructure to an industry standard, took a third approach: the purposeful development of a strategic community consisting of a large group of IT professionals working at corporate headquarters and in the business units (see “More about the Transition Alliance”>). We believe that this structure is a new, important organizational phenomenon, of great use in a world of increasingly dispersed human resources where firms need to wisely leverage their intellectual capital. For Xerox, this approach produced a variety of benefits. First, the strategic community provided a means for IT professionals to manage their complex infrastructure more effectively. They were better able to provide high-quality, validated solutions to issues; to handle unstructured problems; and to deal with the never-ending new developments in hardware and software. Second, the group operated as an efficient mechanism for knowledge sharing, which filters into the business units. In either of the more traditional matrix structures, knowledge sharing may occur, but that knowledge frequently remains unused. Finally, motivation for learning and developing at an individual level seemed greater in this community structure than in other organizational forms, which has important implications for the longer-term performance of these individuals at Xerox (see Table 1).

We use the label “community” because it captures the sense of responsible, independent action that characterized this group, which, at the same time, continued to function within the standard boundaries of a large organization. Similarly, we believe that establishment of the community was “strategic,” because its members’ activities centered on a broad goal — more effective management of the global IT infrastructure — that was integral to overall business strategy. Moreover, we suggest that, in many important respects, this group differed from a task-oriented team. The strategic community established its internal work processes, and its organizational structure was fluid. Although membership was managed and defined principally by organizational function, individuals chose whether to be highly active. Motivation to participate actively was based mostly on needs: to improve organizational performance, to learn, and to sustain professional identity.

We describe the Xerox experience, arguing that the existing community was a more effective infrastructure management strategy than either the thick or the thin matrix. Because community members engaged in the same professional practice, one might think of them as a community of practice. However, communities of practice are usually considered to be voluntary groups that emerge from common work practices,3 whereas Xerox top management quite deliberately established the Alliance. By contrasting the Xerox initiative with traditional community-of-practice views, we can begin to form an understanding of how other firms might replicate Xerox’s success. We also illustrate how strategic communities create value for the sponsoring organizations, and we conclude by identifying principles that drove the success of the Xerox group. Our conclusions are based on participating in community activities during the past several years, as well as directly observing the group in conjunction with a research study.

Building the Alliance Community

At the end of 1994, Xerox was a successful, large multinational corporation with an aging IT infrastructure based on technology that Xerox had developed internally and sold to others with limited success. Although the firm’s technology was superior to other products on the market, Xerox managers also recognized that continuing down the proprietary path would limit the firm’s ability to develop new products with broader market appeal. Xerox also needed an IT infrastructure that could support corporate-wide strategic initiatives, including mobile sales and service applications and increased electronic communications with suppliers and customers. The resulting strategy was to establish a consistent, industry-standard office environment in all Xerox locations with improved corporate-wide ability to exploit new technology promptly.

To support the new IT strategy, Xerox hired a program manager, whose IT career spanned 15 years and several industries. She had consulted for large U.S. and European companies, focusing on strategic client/server infrastructure and systems development projects. Her listening and motivational skills were considered critical to the success of the largest infrastructure transition in the corporation’s history. Her challenge: To persuade 38,000 people worldwide to retire their beloved workstations, many of which had been in service since the early 1980s.

Initial ideas about how to organize to accomplish the transition ranged from what might be perceived as one-to-one confrontations with each business-unit manager to a traditional, thick-matrix governance structure, in which the corporate staff sets goals to ensure compliance through regular progress reporting. To identify the requirements for success and the major barriers to a worldwide transition, the corporate information-management planning group sponsored a roundtable discussion of representatives from the organizational units. Senior information systems (IS) managers from nearly all business units sent someone responsible for retiring the proprietary infrastructure. During the initial 2-day meeting, managers from the corporate unit discussed their objectives and reviewed the status of their program initiatives. In addition, the group engaged in facilitated brainstorming sessions. Attendees unanimously agreed to meet again in 2 months, and thereafter agreed to convene at 6-week intervals.

The thirty to forty people who attended the first few meetings represented the stakeholders in the new IT strategy — the corporate information-management function, the business units, and the outsourcing firm responsible for day-to-day operations and maintenance. Each meeting began with a clear agenda, and knowledgeable participants were present to answer questions and engage in dialogue. Action items were noted, classified, and assigned: “A” items were important to the whole group, whereas “B” items affected a subset of the business units. Each action item had a “question owner,” who would ensure follow-up by an “action owner.” The question owner reported the item’s status at subsequent meetings until they resolved the problem. A dedicated scribe recorded key messages, decisions, and things learned at each meeting. A collective, post-meeting quality check was a key source of feedback used to improve subsequent meetings. During the quality check, each participant stated high and low points of the meeting (including assessments of food and facilities) to identify ways to obtain greater value from the gatherings.

Early on, each business unit agreed to meet the corporation’s high expectations for the transition. The annual process of setting goals and objectives for the entire organization specified that the legacy infrastructure had to be retired in 3 years to allow Xerox to focus on the support and growth of a standards-based architecture. Retiring the legacy system was a key plank in the overall IT strategy (called IM2000), and top managers had committed the corporation to this goal. As a result, the infrastructure management program received exceptional visibility at senior levels. At each meeting, the group tracked the number of workstations and servers retired on a month-by-month basis. Visibly demonstrating each business unit’s progress heightened the motivation of team members to maintain a high level of performance.

During seven meetings held in the first year, topics ranged from infrastructure deployment plans to progress reviews about developing transition tools. At each meeting, the group reviewed previously identified action items. The participants started to publish a group-sponsored, company-wide newsletter for end users — an important communication device that solidified the group’s identity. A year later, Xerox’s CIO reviewed progress and provided feedback about the group’s effectiveness. The group, which began calling itself the “Transition Alliance,” defined its mission as follows:

“The IM2000 Transition Alliance advocates the requirements and concerns of end users throughout the development and deployment of the IM2000 infrastructure. The team will identify, communicate, and leverage best practices, enabling Xerox to sustain a more productive, profitable, and competitive environment.”

This signaled a shift in focus from simply transition to a broader and more complex scope. The Alliance was not only replacing the legacy infrastructure, but was also trying to improve and expand functionality. For example, following rapid adoption of the Internet via the World Wide Web, the firm revisited its earlier technology decisions regarding an industry-standard infrastructure. The Alliance had to develop methods for distributing upgraded versions of operating systems and other software components via a redesigned global communications network.

Even with broadened goals, the group’s perceived effectiveness did not decline. In mid-1997, the CEO of Xerox formally recognized the contribution of the Transition Alliance, identifying it as one of the top-performing teams for achieving the initial infrastructure transition on time and within budget. Xerox Business Services included material about the infrastructure transition in documentation that garnered the business unit a Malcolm Baldridge National Quality Award.4

As the Alliance efforts expanded, Xerox managers started to consider whether they could share with other business areas what had been learned about the formation and functioning of this group. On the basis of surveys and interviews of Alliance members, we make three key observations that we discuss in greater detail.

• The attitudes and behavior of Alliance members differed from those of a traditional cross-functional, cross-divisional team. The Alliance’s distinctions, which closely resembled the attitudes and behaviors of members of a community of practice, enhanced the value of the group to the organization.
• The Alliance represented a strategic knowledge management capability for Xerox and provided a means of more broadly exploiting tacit knowledge than would have been possible in a matrix structure. Members explored potential solutions to issues, learned from each other, transferred knowledge into business units (and from business units to headquarters staff), and willingly applied what they learned. Thus, this approach was a good way to develop and maintain effective infrastructure management practices.
• Although the Alliance was established to operate only in one domain — that is, the management of IT infrastructure — its management processes, encouragement of individual learning, and appreciation of a community culture represent lasting value for Xerox.

New Organizational Form

Superficially, the Xerox Transition Alliance seems similar to the teams that Chase and the Korean chae-bol set up to manage their infrastructure transitions. In all three cases, people joined the group from headquarters (selected for their specialized knowledge) and business units (selected for job function). The cohesive Xerox group effectively resolved issues and managed implementation of agreed-on solutions. We suggest that a group like the Alliance differs from a traditional team along a number of dimensions; moreover, its characteristics are similar to those of a community of practice. After forming the Alliance, corporate intervention was minimal, taking the form of sponsorship rather than integration with management processes. This explains our initial motivation for using the term “community” in referring to the Alliance. Yet, recognizing that top management deliberately established it to help achieve corporate goals, the Alliance was more than a group that meets occasionally to discuss common issues relating to a single functional or professional area. That is, the Alliance was not what most people think of as a community of practice, because it had a defined relationship to formal organizational objectives. Therefore, we add the adjective “strategic” to our term. Although the word “community” is frequently used rather loosely in reference to informal organizations, Wenger has outlined three defining “dimensions of practice” that can inform our thinking about why the Alliance was such a successful corporate initiative.5 Next, we discuss these dimensions of practice and why the Alliance’s organizational structure differed from that of a typical task-oriented team or a community of practice.

Mutual Engagement

Both task-oriented teams and communities of practice are characterized partly by the nature of their members’ relationships, which begin after people are assigned to a team or become team members by virtue of their organizational function. In a community of practice, relationships are defined not only by the way people engage with each other, but also by their interaction on a “playing field of practice.” While the shape of the field is independent of how people engage when playing on it, the field establishes an initial set of boundaries on their relationships.

Xerox encouraged rather than mandated membership in the Alliance. Many members characterized their involvement as stemming from their own initiative —supported, rather than directed, by their manager. Attendance at meetings and electronic distribution of materials was loosely controlled by the facilitators. To participate, sponsorship by a current Alliance member and an internal accounting transfer of a nominal meeting fee were required. Interestingly, when the fee was imposed after about 2 years of operation, there was no decrease in meeting attendance. Through normal turnover, members from business units and headquarters rotated. A central core of about twenty members attended 80 percent of the meetings. Depending on the agenda for the meeting, people outside of the core were invited. Some of these attended three or four meetings in a row; others just participated in one meeting. Clearly, with forty-five to fifty active members, the Alliance was much larger than the average team, yet the Alliance did not include an even larger community of about 250 systems professionals who were engaged in infrastructure management issues.

The core members were essential to maintenance of the Alliance. The core was more open than a typical team structure, allowing time for new members to become full participants. One senior manager, who had taken over for an original Alliance member, recognized that it took time to engage with the other members of the group. He remarked: “The opportunity to do networking will be important to me once I get settled into my job and understand what is on my plate.” Thus, relationships within the Alliance also had a temporal quality. In this respect, the contrast between the Alliance and a task-oriented team is subtle. Superficially, one important differentiating factor appears to be life span. However, even though most teams in organizations cease to exist after accomplishing their goal, this is certainly not always the case. For example, just as the life span of the Alliance was initially indeterminate, self-managing teams in a production setting and top-management decision-making teams may be relatively permanent features of the organizational landscape. Expecting that relationships would persist, core Alliance members were motivated to invest time in getting to know new members, which in turn led to developing mutual trust. This durable Alliance, as an organizational entity, provided the playing field of practice for its members.

Alliance members who participated in our survey believed that almost two-thirds of the group’s value was derived from face-to-face networking at the regular meetings. One Alliance member who had an especially expensive and arduous journey attended every other meeting and participated by audioconference when he could not attend in person. The importance of maintaining personal relationships in this way also distinguishes the Alliance from other high-performing teams, for which research indicates that physical proximity is not critical.6 Although face-to-face meetings are not a requisite means of interaction for a community of practice, most communities do work this way.

Authority relationships in a typical task-oriented team are organizationally determined. Although the Alliance resembled a large, self-managed team in some respects, there were two clear differentiating characteristics. First, the members of the Alliance did not select the facilitators. This was, of course, consistent with the original objective of establishing the Alliance: namely, that it was a corporate initiative guided by the headquarters-based unit responsible for infrastructure strategy and management (i.e., the Technology and Strategy Infrastructure Group; see Figure 1 for information on key organizational relationships). Second, Alliance members clearly distinguished between the facilitators and what we term “knowledge leaders.” Facilitators promoted and advanced the discussion. Knowledge leaders transferred their experience — what they had learned on their own or in other group settings — and shared their insights with the group as a whole. Knowledge leaders emerged on an issue-by-issue basis. Our analysis of e-mail files showed little consistency in whom the Alliance members gravitated toward when looking for answers. They tended to ask those Alliance members who had demonstrated knowledge of an issue in the past. They also acknowledged a small group of people as “knowing who knows.” This subgroup often was able to point a questioner to someone with an answer.

These knowledge leaders were not leaders because of their organizational status or the size of their functional group. They assumed the role of knowledge leader on the basis of their ability to communicate knowledge of a particular field. This role also exists in a self-managed team. However, within the Alliance, the action-item process separated the knowledge leadership role from the facilitation role. The clear demarcation of responsibility helped Alliance members benefit from both the leadership skills of the facilitators and the knowledge available in the group.

Shared Communication Repertoire

Other important distinguishing characteristics of the Alliance as an organizational form relate to communication patterns and work processes. Most teams must satisfy well-defined requirements to communicate information to other parts of the organization. For example, teams may be required to submit regular status reports or to conduct briefing sessions for other units. Communication patterns both within the Alliance and from the Alliance to other parts of the Xerox organization were ad hoc and diffuse (see Figure 1). There was no formal requirement that the Alliance report on its activities. In fact, the first attempt to develop a formal process for informing senior managers of issues that the Alliance considered important was not successful. In lieu of standardized reporting processes, Alliance members used many approaches to inform colleagues of plans and issues, ranging from memos prepared after each meeting to hallway conversations. This resulted in an anomalous situation in which the formal organization that established the Alliance recognizes its value and provides support for its activities, but simultaneously did not incorporate the Alliance within its formal reporting structure and normal organizational communications patterns.

Within the Alliance, the communication repertoire was built upon the leadership training programs required for all Xerox employees. There was, as one Alliance member put it, “a vocabulary that everybody shares. The leadership programs teach people to listen.” This person suggested that as a result of these programs, the Alliance was able “to distill consensus out of what appears to be a chaotic discussion.”

Work processes developed within the Alliance supplement the common corporate vocabulary. As described earlier, handling action items, creating meeting agendas, and developing other processes were evidence of the self-directed nature of the group and, moreover, provided a context for communication. These characteristics are similar to those Brown identified when describing a community of practice: It has “a shape and membership that emerges in the process of activity, as opposed to being created to carry out a task.”7

In summary, the Alliance exercised substantial group discretion not only in the scope of Alliance issues, but also in how the group defined its own operating environment — one in which the formal organization did not explicitly impose either work flow or controls. The level and frequency of engagement with each other built trust, which in turn reinforced the openness necessary to share fully. Unlike in the thick matrix of Chase, interaction was not required; Alliance members on the periphery gained even if they did not interact. When the knowledge flow was unidirectional, the Alliance was analogous to the environment of Honda’s “brainstorming camps.”8 In such a setting, social interaction provides the contextual grease that allows practitioners of the same craft to smoothly engage their intellectual gears.

Joint Enterprise

The work of the Alliance was determined in part by the organization and in part by its members. During the first 2 years, however, the balance shifted toward what the members judged to be value-adding tasks and issues. In this shift, we see similarities in the way a community of practice negotiates the focus of its activities, and we see differences from the characteristics of a task-oriented team.

Most teams are predominantly oriented toward achieving a specific goal and, as noted above, most teams are dissolved after achieving their goal. Initially, the Alliance was also oriented toward a concrete goal — replacing the old, proprietary infrastructure — but as that was being achieved, the goal was supplanted and substantially expanded. The nature of work in the Alliance changed from a focus on current issues to a prospective focus, dealing more with infrastructure management practice. At first, one important new objective — global rollout of a new operating system — was similar in purpose and scope to the original goal. The members quickly expanded this objective to encompass other technologies outside of the operating system domain, such as a new e-mail service, remote access capability, and Web application features.

Thus, the work orientation of the Alliance became centered on the general practice of technology management as well as on the accomplishment of specific tasks. This became a key superordinate goal around which a shared vision and unified identity developed. Hackman uses the term “spiral” to describe how groups move up (or down) the performance curve from events that define their identity.9 In the case of the Alliance, the performance implications of the upward spiral from the defining event — achieving its original goal — were primarily visible through the broader role that the Alliance played in infrastructure management issues. For a task-oriented team, the performance implications are usually measured only in operational output.

Alliance members determined their tasks and activities; they were not governed by a formal corporate structure. Previously hidden issues became explicit, perhaps because one-to-one discussions (business unit to corporate) were replaced by a forum in which people felt less constrained and more willing to express doubts, generate new ideas, and critique old ones. As one Alliance member put it: “Working with corporate staff used to be a series of ups and downs (mostly downs). The way we work has now changed. Instead of one-on-one discussions, our work is more of a joint effort that boosts everyone’s performance.” In seeking explanations for this change, we found that improved performance was an outcome of the nature of learning and knowledge processing within the Alliance and among its stakeholders.

New Knowledge Management Capability

We suggest that the Alliance represented a distinctive and informal, yet corporate-sanctioned, organizational entity. Nevertheless, one obvious question is: “So what?” To build on the description of the strategic community as a noteworthy organizational phenomenon, we identify the ways that it creates value. Specifically, we believe that the Alliance represented a new knowledge management capability for Xerox. This capability is replicable, provided that a firm consider the contextual characteristics of mutual engagement, shared communication repertoire, and joint enterprise. We characterize this capability as a way to manage knowledge because many of the outcomes of the Alliance revolved around learning. There is equally important evidence of more specific ways of adding value, as we discuss next. (For a summary of added values, see Table 2.)

Over time, learning became an objective equal in importance to the information sharing engaged in by the Alliance members. In a sense, this was another superordinate goal of the group. One member said: “I gather information about programs and the methodologies that are being used to deploy development modules.” At the end of a meeting, another member said publicly: “I am pleasantly shocked at the amount of information I obtained and how much I learned, even though not all of it is immediately applicable to my current job.”

The lack of emphasis on positional authority elicited behaviors in Alliance members that were conducive to deep, relevant learning. However, one behavior that is particularly hard to teach is reflection. To overcome this, at the end of each day of the Alliance’s customary 2-day meeting a small group of Alliance members discussed what they had learned that day. They distributed a one-page summary to Alliance members and to others in the larger infrastructure management community. At the meeting’s end, leaders supplemented this loosely structured reflective process by polling Alliance members about what they learned during the company’s usual end-of-meeting quality-check queries. Reflection is the mechanism that transforms tacit knowledge into explicit knowledge, not only for the attendees, but also for the larger Xerox community of interest.10 The result is learning — individually and organizationally.

In practice, the openness of the forum and the different objectives of group members led to more opportunistic learning than occurs in a task-oriented team. Such learning becomes applicable beyond the immediate task, which in turn spurs innovation. When deploying new technology in a complex infrastructure environment, getting the technology “in the door” takes an extraordinary effort. The Alliance was consistently successful at uncovering operational implications — those not identified or those of greater impact than the designers originally believed.

Moreover, the diverse makeup of this large group inherently meant there was less opportunity for group-think behavior to develop. In fact, the creative potential of the Alliance’s heterogeneous membership was frequently evident. For example, representatives from corporate headquarters, the business units, and the design group disagreed with a supposedly carefully designed approach to software upgrades. Although at first defensive, after discussion the designers acknowledged that the requested modifications to the original plan would substantially improve the upgrade process. Collaborative, frank interaction at the Alliance meetings helped to validate knowledge beyond the initial output of the design team.

We also examined how learning became diffused throughout the organization. Typical of many teams, the Alliance was composed of people from different organizational units and disciplines. However, we believe that Alliance membership was more conducive to relatively honest opinion sharing. As described previously, people were encouraged to participate on a “need to know” basis, which resulted in deeper internalization of learning and better leveraging of skills. Furthermore, Alliance members, themselves, selectively facilitated effective diffusion of learning across organizational boundaries.

The application of learning differs in strategic communities and task-oriented teams. Most teams represent only a small organizational component, and they tend to be inwardly focused; typically, they are oriented toward completing a task. Wider dispersion of lessons learned greatly depends on a team’s reputation, which may be organizationally limited. In contrast, implementation tactics and management practices that received the Alliance’s “stamp of approval” carried considerable authority within Xerox. Alliance participants recognized that the sooner the Alliance was involved in an issue, the better the chance of avoiding a “replan.” As a result, organizational units demonstrated greater willingness to modify plans when challenged. The authority and influence of the Alliance seemed to extend beyond its boundaries and reduced additional review and decision making in the business units, which were blindsided less frequently.

Almost all participants that we surveyed valued their membership in the Alliance and shared a sense of pride in belonging to this community. Sharing hundreds of years of collective knowledge fostered an environment that facilitated learning transfer and solved complex deployment and transition problems. Nonaka and Takeuchi use the term “community of interaction” to define the context in which knowledge creation occurs.11 In the Alliance’s strategic community, knowledge creation focused on what the members of the group practiced, namely, development and management of technology infrastructure. The members shared what they knew and learned through interaction.

Principles of Effective Strategic Communities

Early in the life of the Alliance, some business-unit members expressed skepticism in the form of a “we versus they” attitude. For example, one business-unit manager described the Alliance’s objective as “rollout of their [i.e., headquarters’] programs into my space.” Gradually, the Alliance developed, refined, and adopted processes such as agenda building and action-item management. These processes did not work perfectly, but they provided members with visible evidence of the commonality of their interests. In effect, interaction through adherence to these group norms stimulated and sustained the sense of community. Unlike emergent communities of practice, the Alliance did not form spontaneously or fortuitously around issues of common interest. Nonetheless, the processes and attitudes necessary to achieve the high performance and mutual learning characteristic of a strategic community had to develop over time despite initial corporate sponsorship. At the end of our study, members of the CIO’s leadership team were making regular presentations at Alliance meetings, and business units were allocating budget monies to ensure that their managers could continue to participate.

Although a strategic community may focus on relatively short-term operational needs, it provides lasting value through learning, innovation, and knowledge transfer. The Alliance demonstrated that cultivating a strategic community is an organizational capability that maximizes the effective application of intellectual assets. Firms that want to compete in the knowledge economy must exploit the potential of strategic communities and sponsor their development.

How can other companies replicate this Xerox experience? Quinn suggests that effective organizations must structure their enterprises around intellect.12 He uses the term “intellectual web” to describe phenomena similar to those that we observed in the Alliance. A company can create an intellectual web if the following key factors are present: regular interaction, mutual interest, recognition of the intrinsic value of learning, and incentives for sharing. Other research reinforces the importance of these factors.13

We have identified six key principles that were critical to the success of the Transition Alliance at Xerox (see Table 3).

1. Design an interaction format that promotes openness and allows for serendipity. In the Alliance, the facilitators managed a flexible agenda built around a few standard meeting components. Within this format, the role of the facilitators was clearly distinguished from the role of the group’s knowledge leaders. In addition, the format provided significant time for social interaction during the meetings. Although electronic interaction (e.g., an e-mail distribution list) was useful in maintaining community-mindedness, we observed that technology alone was insufficient for effective community development. In fact, the Alliance members — all of whom were completely comfortable with a variety of interaction technologies — strongly believed in the value of face-to-face meetings. They told us that this was the primary means of shaping the deeper knowledge of colleagues and mutual trust that are precursors to effective group work with digital technologies.
2. Build upon a common organizational culture. Every Xerox employee must participate in the company’s “Leadership through Quality” training program. A primary objective of this program is to develop sensitivity to the group dynamics issues that may limit productivity. Employees learn interaction and facilitation skills, but, more importantly, they internalize the use of a common vocabulary that reinforces a strong sense of organizational culture. Appropriate application of acronyms such as TTM (time to market), XMM (Xerox management model), and XBA (Xerox business architecture) provides a common ground for effective communication and development of mutual understanding.14 The result is a climate of social activism, in which the potential for counter-productive interaction is mitigated by following well-established organizational norms and processes. In this environment, the natural tenacity and determination of Alliance leaders produced successful outcomes from group work. Just as the leader of a jazz ensemble improvises around a theme familiar to everyone in the group, the leaders of the Alliance exercised these traits to develop and maintain effective infrastructure management practices.
3. Demonstrate the existence of mutual interests after initial success at resolving issues and achieving corporate goals. Group facilitators do this consciously by using and refining the processes that the community develops. In effect, these processes stimulate the emergence of “mutual influence systems” that mask differences in hierarchical status among the members.15 These processes also balance local (i.e., at the individual and business-unit level) and global (i.e., at the corporate level) value. Combining flat structure and mutual influence processes results in high commitment to achieving goals, even in an organization strongly oriented toward control structures. Alliance members acknowledged their mutual interests and were willing to reconcile differences, recognizing that working together reduces uncertainty. After initial success, the community members grew confident that the group could get the work done. Broadly accepting that they have mutual interests and that everyone values sharing experiences led to learning.
4. Leverage those aspects of the organizational culture that respect the value of collective learning. Xerox not only values learning, but also appreciates the value of collective learning. Thus, Alliance members benefited from learning on an individual basis in a natural environment, in addition to learning from direct interaction with practicing colleagues. Rather than relying solely on a centralized “push” of information, a strategic community forms and shares knowledge by “pulling” individual members into an environment in which they learn from each other. Broadly based participation and engagement emerged as the dominant learning modes for Alliance members.

5. Embed knowledge-sharing practices into the work processes of the group. Xerox introduced this policy early in the life of the Alliance by asking business-unit managers to regularly present talks on their initiatives. Later, Alliance leaders introduced an agenda-building process — a means of polling members about their knowledge requirements. During the end-of-meeting quality check, they began to ask members what they had learned, which highlighted and reinforced valuable knowledge. In addition, a small group convened after each meeting to reflect on and record the key messages for the larger community. By incorporating knowledge sharing in group work, reflecting on acquired learning became a natural behavioral characteristic of its members.
6. Establish an environment in which knowledge sharing is based on processes and cultural norms defined by the community rather than other parts of the organization. This is not an idealized view. In the Xerox case, the organization granted the community unusual autonomy to operate outside the formal control structure. This included, for example, the freedom not to report to the organization on the activities of the community. Here, the facilitator’s role was to encourage openness and commitment to the group rather than to the organization. Over time, the community established a “zone of safety” for candidness. The facilitators also actively and regularly solicited feedback from the group to ensure that its work evolved in synchrony with its sponsoring organization.

Conclusion

We believe that developing strategic communities is of substantial value, and we offer some principles to guide achieving this. Although this article focuses on effectively managing technology infrastructure, we suggest that the principles outlined apply more broadly to the effective creation and use of organizational knowledge. By definition, infrastructure affects everyone, therefore effectively managing and using it can only be achieved by a broad transfer of knowledge. Strategic communities are an important way to attain this goal.