I am at the KMWorld conference in Santa Clara.

This presentation, "Social Network Analysis (SNA): Understanding Organizations and Getting Results", is by Eric Lesser of IBM, and Patti Anklam of Hutchinson Associates.

Patti is up: Mentioning the Robert Cross presentation later, and recommending seeing him speak. (I was thinking of going to that too.)

SNA (in this context) is about looking at the pathways of how knowledge flows in an organization, and how things get done in organizations. SNA is about finding out patterns, first step is to create a social network knowledge map.

If other patterns or collaboration is desired: create interventions to create, reinforce, or change the patterns to improve the knowledge flow.

Identify network to map. Create survey questions along certain dimensions of a relationship. Basic questions like: "How much do you need to talk with each person to do your job?".

Survey dimensions: know-about, information, communication, trust, problem-solving, decision-making, sense-making. Data points: distance (degrees of separation), betweeness, density, centrality, brokers, gatekeepers, outliers.

Get data that can help you do benchmarks to determine what kinds of things you want to do to intervene.

Management team case: 85-person global group within 3,000-person organization. Strong support from VP of HR. Need good HR championing to get credibility for interventions.

Goal: get more synergy across three product lines. Survey: I frequently or very frequently receive information from this person that I need to do my job. Created network map. Color coded by product line, and by frequencies.

Next step: density analysis: how well the organization is connecting. Density within product lines and accounts. Showed product lines and account junctures with a very low densities of connections. Numbers were catalyst for action.

Hired liaisons between groups with low density of connections. Some areas reorganized. Created knowledge networking initiatives: sharing strategies across groups, and more face-to-face among groups that had never met that way.

Also, personal actions. Some people realized how they had become bottlenecks in communications, and removed themselves to unblock the flow.

SNA doesn't give the answers, but it helps ask the questions. It is a diagnostic tool. SNA methodology uses a complexity model: detect patterns, make interventions, see what new emerges. You can't predict the outcome, but you can reinforce positive patterns and alter the negative ones.

Surveys, even by asking the question, can help bring awareness to people not knowing each other well enough.

Eric is up: SNA used to understand communities of practice. Organizations realizing that communities are the cornerstone of KM strategies.

Big question is often, why should we invest in helping communities of practice, and how do we decide what to do. SNA can be a way to find good actions and measure their results.

Communities of practice: voluntary around knowledge sharing or particular interest of passion.

SNA: identify key roles people play within networks. Boundary spanners: connect different groups. Isolate: people not well connected, but valuable. Brokers: help connect different groups. Gatekeepers: may prevent connections.

Case study: pharmaceutical company made up of 4-5 merged companies. Interviewed scientists about 1) knowledge needs assessment. Then, 2) knowledge network analysis. Then 3) benchmarking and solutions design.

Survey about within lab site, across different lab sites in company, with labs outside of company. Most interactions were within site, second most were outside the company. Least was between sites within the company.

Merger was supposed to leverage synergies in research, but each lab still working very independently, with little connections with each other. Dense connections in each lab, but weak and few connections across labs.

Also, did analysis of connections to outside academics. Lots of redundant connections outside: people were going outside before going to others in the company. So, this justified the need for a community intervention, because the company's research network was not optimized: people were not connecting.

Survey: why is this happening? A: because we don't know who these people are. We don't know what these people know. Objective: making connections to other people!

First solution tried, a portal. Content, but also expertise location. Areas of expertise and connecting people with the same areas of expertise: more important than simply collecting content.

Need to really find out what people actually need. What is the primary concern of the people you want to draw together into a community.

SNA can also help identify the key people who can play an important role in creating the community. Who are the nodes / connectors (he mentions Malcolm Gladwell's The Tipping Point). Identify connectors, and bring them in early in the community development, because they will be able to attract others into that process.

Communities of practice can work as a diagnostic technique for finding needs of community and also key people.

Preparing for a SNA. Group size: need minimum of 25-30, maximum of 200. Identify potential barriers right at start: functional, geographic, hierarchical. Have some hypothesis.

Survey length: keep completion less than 20 minutes. Look at 5-7 key issues.

Clear sponsor communication: study objectives, process, what will be shared at the conclusion of study.

Conduction a SNA, priorities. Response rate: need near or at 100% to achieve most accurate results. Rigorous follow-up. Completing the surveys: web based vs paper based. In person vs remote submission.

Use of pre-filled data and drop down boxes to minimize completion time.

Analyzing and communication results of SNA. Leverage both pictures and analytic data to match information needs to different audiences. Combine with interview data: social network diagrams can highlight issues, but do not provide complete diagnosis. A person with a lot of connections may be knowledge barrier vs knowledge facilitator.

Review results with study sponsor before broader audience to prevent surprises and address political issues. Provide feedback to study participants with associated actions for improvement.

Q & A.

Other solutions for scientists: face to face meetings, expertise location, community building exercises.

Making networks transparent in organization has a lot of affect. Survey alone has a certain affect, generally useful and not misleading. Look for trends, not exact numbers on survey.

Q about remapping networks over time.

I ask a Q about the Law of 150 (that Gladwell talks about in The Tipping Point) and how they have seen that in communities of practice. A: yes, research and experience has shown 150-200 as optimal size, and larger groups tend to subdivide into smaller groups.

Q about SNA graph tools. Mentions shareware / freeware things that you can find on the web. Krebbs tool.

SNA has looked at about 50 networks in formal study, but only 6-10 cases so far of using these to form communities of practice.

People are looking at email patterns as one way to analyze network connections. Email is less intrusive than surveys, but don't get qualitative data. Social network retrieval is a new genre of technology.

Email traffic is interesting, especially around email overload. Senior managers get 150+ emails a day, based on studies. Interesting to consider cultural and organizational issues that would change that.

Q on organizational size and SNA. A: SNA can work on any organizational size, it is just a matter of what is the specific group you want to look at.

Also, SNA can be combined with other disciplines like knowledge architecture or project metrics.

People look at document sharing, because it is easy, but it is not necessarily a true measure of knowledge sharing. SNA can give you another viewpoint.

Q on evolution of communities (mergers, morphing) and possible uses of SNA in that context.

Q about results and benefits of communities: innovation, productivity.