There was just one problem. In order to reverse engineer a model there has to be a model and I couldn't a trace of one. There was no indication that Groupon was using any information about me when they sent me an email. Here's an amusing but not unrepresentative example:
Of course, it's possible that I might give something like that as a gift (though I honestly don't think I know any women who would want a chocolate foot scrub -- the whole thing sounds disturbingly like a front for a fetish website), but when you have offers that the email recipient would never consider buying as anything but a gift, You can't really call it targeted marketing.
For further evidence that Groupon is not that sophisticated (or even that serious) about targeted marketing, check out the following:
The scary word here is "New." Did it really take over two years to add a feature that should have been fully operational the day the first website launched? Has Groupon actually neglected to gather rudimentary data on tens of millions of customers?
This might not be as bad as it looks (I don't see how it could be worse). Perhaps there's a level of sophistication that I'm missing. Unfortunately, as mentioned before, people who are betting on Groupon are assuming that the company will successfully pull off some extraordinarily difficult tricks with data. The apparent inability to manage simple tasks like gathering customer level demographics does not make these bets look all that smart.
p.s. Felix Salmon seems to mean something completely different by 'targeting' (though it's possible his definition is broad enough to include what I'm talking about here).