I have something else I need to be working on so I don't have time to comment on this at length but the New York Times has a piece on statistics and marketing that worth look. I can't give a ringing endorsement (despite having written on the subject before, I'm not sure the author, Charles Duhigg has that firm of a grasp of predictive analytics), but it's still informative.
I also have a similarly mixed reaction to Felix Salmon's comments on Duhigg's piece. I'm not questioning Salmon's understanding of the principles at work here (when it comes to analysing business problems, he's about the best we've got), but in this one case, I don't think he's supported his conclusion.
Check it out and judge for yourself.
25 minutes ago