Laube, 43, Cardlytics’ president and COO, and Grimes, 51, its CEO, have since helped pioneer a data-driven advertising niche called merchant-funded rewards. It targets people based on what they buy, not who they are. “If you know where and how someone is spending money, you know lots of things about them without having to know their personally identifying information,” Laube says.

I have found that the transactions of customers is the most important predictor of future behavior in all data I have studied. While the demo, geo and psychographics of customers is very interesting data, to maximize the revenue from known customers is to really get to know their transactions.

While I don't know how good these algorithms are, the theory is solid. I happen to be a Bank Of America customer and the deals I receive don't seem to be any better than say my Rapid Rewards dining offers which don't seem to know my eating habits whatsoever. If these can be perfected, I think it's something that would get me to use my card more often instead of using my AMEX. I'll be watching because this is very intriguing.