Forgetting the privacy issues (formidable as they are), let’s ask a few business questions about the value of hyper-targeted media. Is it really valuable? Certainly not on an individual basis.
Statisticians will tell you that, with the appropriate sample size and mathematical tools, it is relatively easy to predict what a population will do. However, it is absolutely impossible to predict what any individual will do. As you well know, when it comes to your personal decisions, past performance is rarely an indicator of future performance. So, we can predict that tonight in Manhattan, 5,437 dinners will be served at restaurants that feature Mexican food, but we can’t predict who will eat them.
Even if you could predict, with a reasonable margin of accuracy, who would frequent these culinary establishments this evening, would it help you market to them? What would you do differently? Would your advanced knowledge of this particular behavior enable you to extract a greater share of wallet from this hyper-targeted audience? Probably not.
He raises an interesting point. The data wonks are asserting more control than ever, using new media to prove that ROI can be improved by reaching their targets more precisely. But all the hyper-targeting in the world can’t overcome the notion that people are emotional, irrational, and make purchasing decisions for all sorts of reasons that we can’t predict.