Tuesday's KMPro Chicago meeting was a great panel discussion on social networks and social network analysis with Stefan Lafloer of InsightKnowledge, Valdis Krebs, Konstantin Guericke of LinkedIn, and Shannon Clark of JigZaw and MeshForum. We had people participating at two gathering points in Chicago and several of the panelists on the phone from remote locations.
The session started with a brief demonstration of network analysis from Stefan, who showed us NetVis, an open source tool to play with network visualization. This can obviously be used for social network analysis, but it can also be used to analyze any network for which one has data, much like Valdis' InFlow network mapping software.
We then jumped right into the discussion with the question of "what is social network analysis or social networks." Being the SNA expert, Valdis reminded us that SNA gives us a snapshot in time that helps an organization see what is happening and then put forward plans to effect new behaviors. Ideally, one follows the change with another round of SNA to test the effects of the change. Valdis later mentioned that it is quite difficult to make effective change in social networks because they are so deeply ingrained in the way people trust one another. Trust is something that doesn't simply happen without a long pattern of behavior to draw from. Just because you claim a connection with someone in a social networking tool doesn't mean that there is a strong trust relationship there. Social network analysis attempts to look beyond these claims to actual behavior to gauge the interaction levels (trust) between people.
On the trust theme, Stefan harked back to some of the research he had presented to KMPro in the past: how trust impacts knowledge management. Clearly, in the sense of networks being collections of people, the strength of the ties between them has a lot to do with trust. The stronger the trust connection between people, the more likely they are to rely on one another for help and information, even though they may not be the de facto expert.
As the other participants indicated, connections in a network are not always personal. Shannon, in particular, was interested in how networks operate in the general sense. A social network is a collection of nodes (people) connected by some action between them. But a network of companies or airports or a supply chain could also be described in a similar way: nodes connected by transactions between those nodes. In this sense, the analysis can become more complex and more abstract. Valdis described some work in which he is combining the people aspect with a knowledge aspect in a company with over 20 main products. The network could be viewed strictly along the lines of who-talks-to-whom on a given product, but it could also be sliced along the lines of all the products to look for how those products are connected to one another: who are the mavens that transfer knowledge across product lines?
From the social networking tool (LinkedIn) perspective, Konstantin was interested in the question of how and why people participate. For LinkedIn, they explicitly look for people who want to share business connections. And they have found in surveys that people tend to share connections and pass along requests fastest amongst their most trusted connections. Many others have commented on the value of claiming a connection between X and Y. However, at some point the strong links are helpful to the network members. Wouldn't it be interesting to rate the person to whom you are connected as well as the strength and type of connection? In other words: what is the nature of the transaction(s) between X and Y? How do you know each other?
Someone brought up the question of whether there was a "best" network design. There isn't, which is obvious after a few moments' reflection. There cannot be a best design because every group has different needs and drivers. Even within the same company, there are different forces at work that create many different possibilities for the "best" way to connect people. Look at the number of organizational designs companies have tried over the years. Who is to say which is best overall? That said, some people still claim to like certain types of networks over others. I suspect this has more to do with personal preference and experience, rather than true optimal design.
What about other business applications, besides internal SNA? We talked about the use of network analysis to improve one's marketing. Pharmaceutical companies look to opinion leaders in every market and try to convince them to prescribe a given product because there is strong evidence that the opinion leaders heavily influence their networks. There are similar efforts happening in many industries, even the fashion color marketing world.
There was plenty of other discussion, but this should give you the gist. There will be no August KMPro meeting, except maybe for a visit to Chicago's movies in Grant Park as a social outing. Upcoming topics/speakers at KMPro Chicago include Tom Hoglund, Clare Hart, and a discussion of blogging for business. Check the KMPro Chicago blog for updates.