Friday, October 03, 2008

Social Analytics

How should we frame the question to be answered?

In Simplexity, Jeffrey Luger's book about studying the continuum between the simple and complex he uses numerous examples of social behavior - traffic jams, evacuations, standing ovations, and other events. A point is made that the level at which you study these events dictates whether or not you can see the forest from the trees. Murray Gell-Mann's term for this is 'plectics' - which I take to be the understanding the granularity at which a problem should be approached.

Web 2.0 analytics, as reported in B2B poses similar problems. All those micro-events create a false sense of precision. If we can measure seconds of this, clicks of that, loads of something else, then we must be able to nail down this thing called engagement. Intent is still an abstract concept, even with tons of data about behavior.

Social media analytics seems to pose similar problems. There is a lot of attention focused on identifying influencers or an individual's web of connections. But since every decision is likely to involve a completely different network - my search for a camera involved a completely different social graph than the search for funding - maybe some attention should be on a more macro level.

Influence, like engagement, happens - no argument about that. But rather than who links to whom, shouldn't we be thinking more about how it impacts a business rather than a sale? What can we learn from social media data if we don't know the specific individual?

No comments: