There’s a great article in The Economist’s latest Technology Quarterly, called Untangling the Social Web. It’s about how the field of network analysis looks at the behaviour of people in terms of the trails of data exhaust they leave behind, but not as individuals. Rather, it considers the behaviour of people within their social context. For example, by studying the pattern of phone calls made by telco customers, network analysis can determine who is an ‘influencer’, and hence more likely to get friends and relatives to change telcos when they do. If you can tell who the influencers are among your customers, you can offer them better deals, thus ensuring fewer lost customers.
Leaving aside the question of how we non-influential misanthropes might feel about formalizing this two-tiered structure of clients, what I find fascinating is the way it refocuses how we need to think about the boundaries of privacy. Often, we think about privacy in terms of the content of our information. In this case, though, it’s not that companies are tracking who amongst their clients is saying things like “hey I think you should by stock in company x,y,z” or searching their client base to find out who have important jobs; rather they’re tracking the pattern of calls: calls made late at night, length of calls, etc, and making inferences about the social relations that underlie those patterns.