costly replication of effort and improve per- formance on critical tasks throughout an organization.

M any executives have taken steps to enhance knowledgeworker productivity. In the 1990s, organizations focusedheavily on capturing and sharing lessons and re-usable workproducts to avoid costly replication of effort and improve per- formance on critical tasks throughout an organization.1 These efforts resulted in substantial databases of knowledge assets. However, these databases did not become as central to knowledge worker effectiveness as many managers and IT vendors had hoped.2

A part of the problem can be traced to assumptions of knowledge work. Knowledge workers must often solve complex, ill-defined problems with short time horizons. These efforts entail more than simply finding an answer in a database: they require defining the relevant dimensions of a problem, crafting a solution that is feasible and appropriate for the situation, and convincing others of the correctness of a proposed course of action.3 Given this dynamic problem- solving process, it is no surprise that databases did not supplant people as a key source of information. Instead, informal networks continue to be critical to knowledge transfer and to the diffusion of innovations and ideas.4

Appreciation of this central role of informal networks5 has led to what many call the second (or third) wave of knowledge management—a movement starting in the late 1990s that has emphasized technical and organizational ini- tiatives to promote collaboration.6 Collaborative technologies have grown to account for nearly one-fifth of corporate spending on software,7 with the market for real-time collaboration tools estimated to be close to $6 billion in 2005.8

Many executives have begun to identify and support networks of employees doing similar work to better leverage expertise and best practices throughout their enterprises. Commonly called Communities of Practice (CoPs),9 managers at leading companies—such as IBM, Accenture, Procter & Gamble, Hewlett-

Using Social Network Analysis to Improve Communities of Practice

Rob Cross Tim Laseter Andrew Parker Guillermo Velasquez

CALIFORNIA MANAGEMENT REVIEW VOL. 49, NO. 1 FALL 200632

Packard, Xerox, and BP—are increasingly supporting these lateral networks and deriving substantial organizational performance and innovation benefits.10

While communities of practice have become increasingly prevalent, the question remains as to whether these efforts yield productive collaborations and business value or simply consume excess money and time, as happened in the first wave of knowledge management. Social network analysis (SNA) can help address this question by allowing executives to visualize the myriad relationships either facilitating or impeding community effectiveness.11 Over the past four years, we have engaged in a research program applying network analysis to 15 communities of practice. In this process, we found that SNA can be particularly helpful in moving a community from an ad hoc, informal group to a value-pro- ducing network by focusing on five points:

▪ improving information flow and knowledge-reuse;

▪ developing an ability to sense and respond to key problems or opportunities;

▪ driving planned and emergent innovation;

▪ nurturing value-creating interactions; and

▪ engaging employees through community efforts.

Assessing Communities of Practice with Social Network Analysis

The term “communities of practice” came into being around fifteen years ago as part of a social theory of learning in practice.12 Drawing upon the ethno- graphic research of Xerox service technicians,13 scholars argued against a sep- aration of knowledge and practice and in the process demonstrated the strong tendency for people to learn how to do their work and be functioning members of an organization from peers facing similar conditions. Etienne Wenger, often credited with coining the term in 1991, defines CoPs as “groups of people who share a passion for something that they know how to do, and who interact reg- ularly in order to learn how to do it better.”14

The communities in our research shared two features that made them particularly amenable to assessment and improvement with network analysis. First, unlike divisions or teams, CoPs are almost always voluntary groups with no reporting structures or accountabilities. Leaders attempting to develop com- munities often have no direct control over employees contributing discretionary effort and so must look for subtle means to exert influence and promote effec- tiveness. Second, CoPs are usually composed of people who are geographically dispersed as well as housed in different divisions or business units. Whereas a divisional leader at least knows who is on her payroll and where the person resides in the formal organizational chart, community leaders do not have this luxury and quite often have to work through the network just to find the mem- bers of a potential or existing community. SNA provides an x-ray of the inner workings of these groups that helps community leaders locate the high-leverage

Using Social Network Analysis to Improve Communities of Practice

CALIFORNIA MANAGEMENT REVIEW VOL. 49, NO. 1 FALL 2006 33

intervention opportunities. Working through opinion leaders, building aware- ness of expertise in the CoP, and ensuring vibrant connectivity at key fragmen- tation points are but a few of the low-cost and subtle means of promoting CoP effectiveness that SNA helps identify.

Consider Halliburton, one of the world’s largest providers of products and services to the petroleum and energy industries. An industry leader in the knowledge management realm, Halliburton has regularly employed SNA in many of its efforts to systematically build 19 communities of practice across a variety of business disciplines and technical services. Halliburton did not imple- ment these communities in an ad hoc fashion: senior management demanded more than loosely defined or difficult to measure objectives such as “improved collaboration” or “better knowledge sharing.” Rather, the community initiatives had to show measurable results directly linked to financial performance. By applying targeted interventions based on SNA assessments, Halliburton has been able to do just that across a number of communities. As an example, a global community of practice within a critical business unit produced the following measurable results in one year:

▪ lowered customer dissatisfaction by 24 percent,

▪ reduced cost of poor quality by 66 percent,

▪ increased new product revenue by 22 percent, and

▪ improved operational productivity by 10 percent.

Employees in this community design, manufacture, and install equipment enabling production of hydrocarbons from newly completed oil and gas wells. Although initial planning for the completion of a well addresses the most impor- tant design considerations, the final design is highly dependent on the opera- tional parameters of the well. This means that a completion may go through a large number of changes depending on how the drilling of the well develops as well as on the various reservoirs it may cross, expected production, and local logistics. Because of this dynamic environment, all those involved must collabo- rate closely to avoid errors in hand offs from one group to the next. Through its

community investments, Halliburton cre- ated a global, collaborative environment that helped mobilize expertise to solve problems at an individual well and also benefited drilling around the world as others avoided costly mistakes. For exam- ple, at one point a member of the comple- tions community experienced a specific problem with a deep-water well in West Africa. Through both virtual forums and direct interaction with key network partici- pants, the community found a solution to the problem and then propagated it with such speed that three other similar comple- tions to be performed within the next 24

Using Social Network Analysis to Improve Communities of Practice