Friday 17 September 2010

Social Network Analysis


One of the themes to come out of many of the presentations and discussions was that successful collaboration projects recognise that 80% of the effort should be put into the people issues. The technology is only part of the story. One way of getting a perspective of what issues might exist is to do a social network analysis.  Look at current relationships and connections between people, identifying nodes and information flows. If you look at how work gets done you can identify influential people, groups who work together, and loners who don't communicate. It can then be interesting to compare with the organisational structure. In the above hypothetical example there's an individual easily identified who seems key, but would not be easily identifiable from the organisational chart. Without him, the "blue" team would be totally disconnected from the rest of the group. The Head of the Dept (red) is also fairly disconnected.

So, could doing such an analysis on our own departments/teams tell us anything we didn't know? Could we use it to better understand how we work, and where there might be issues?  The presentation included some examples of social network analyses of projects and their interaction with stakeholders, which did illustrate  how integrated the members of the network are, how well information is being shared across the network and how influential certain project participants are. It seemed to provide a more useful depiction than a project organisation structure diagram since it can identify critical relationships in projects.

Definitely food for thought.

3 comments:

Drew Mackie said...

Yes SNA can be really useful in recognizing that the real working relationships in an organisation can form patterns of influence and information flow that may be (and usually are) quite different from the official organogram. Things get really interesting when you compare network centrality with the skills and resources attached to the nodes. Often, this shows highly central nodes that are poorly skilled and / or resourced, or the concentration of skills and resources in peripheral nodes.

We have been involved in several such analyses recently (one in Sheffield for the "Big Society up North"). The biggest problem is getting senior management, who are used to thinking statistically, to take SNA seriously.

pj said...

wonder if we could use gephi to look at this - been looking for a reason to use it

Arthur Clune said...

@pj (and @cloggingchris) see

http://pppeoplepppowered.blogspot.com/2010/03/my-very-first-hyperforest.html

and

http://pppeoplepppowered.blogspot.com/2010/08/neo4j-python-crawler-open-calais-gephi.html

JISC project at University of York