There is more potential to eBay’s acquisition last week of Hunch, the product recommendation engine, than might be assumed at first glance.

I’ve read with interest about the acquisition, and find fascinating the notion of data driving, well, everything, particularly shopping. (And it is interesting not just because today is said to be CyberMonday.) It’s also wildly scary.

EBay explains its attraction to Hunch as follows:

Mark Carges, the chief technology officer at eBay, said he hoped that Hunch would help eBay with one crucial blind spot in its shopping platform: suggesting items to buy or bid on.

“All retailers want to suggest items to the people visiting their site,” Mr. Carges said. “That works with a finite category of goods. We have a lot of unique inventory at eBay, from coins to auto parts.”

EBay currently has 200 million items for sale across 50,000 categories, and nearly 100 million registered buyers and sellers using the site. Put another way, that’s 9 petabytes of data. Hunch’s software could help eBay massage that data to improve the recommendations people see.

“If Hunch knows that people who like one fashion brand also usually like another brand, that kind of learning could be exported over to make suggestions on eBay,” said Chris Dixon, who co-founded Hunch in 2009.

Hunch makes a lot of sense for eBay on its own, but what happens when it synthesizes data from PayPal, which of course is owned by eBay? And what about credit score data in general? At some point will not that data get incorporated into recommendations?

Let’s start with the PayPal data. Clearly, PayPal can tell Hunch algorithms how much money an eBay user is transacting, which is a reflection — at least in part — of the consumer’s buying power. At the least, PayPal can inform Hunch when to offer a Swatch watch as opposed to a Patek Philippe. And I’m sure PayPal data can generate more refined “suggestions” than that. 

Credit score data goes beyond what PayPal can provide, because it not just offers a window into the money flows of a consumer, but also that consumer’s financial behavior to the point where a weaker consumer should not only miss the Patek Philippe promotion, but probably should be prevented from buying such a watch from the get-go.

All this points to a kind of class- or income-based marketing that goes well beyond historical segmentation practices. If I make $50,000 per year and you make $100,000 per year, for example, your online experience will be much different than my online experience. And the more substantially financial information is leveraged, the more precise will be those alternate online and mobile experiences. I have no idea what will be the socio-political ramifications of this.