Monday, June 6, 2011

Yeah, n = 1, but based on this sample, I can't see Groupon doing any targeting at all

After finishing the last post, I got to thinking about the emails I'd gotten from Groupon and I decided to go through them and try some back-of-the-envelope reverse engineering on their targeting methods. Obviously, I just had one account to look at but that's often enough to let me make an educated guess at things like what sort of data they might be looking at.

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:

$125 for 24-Karat-Gold Specialty Facial and Chocolate Foot Scrub at Chocolat Day Spa in Arcadia ($260 Value)

See Deal
Value $260
Discount 52%
You Save $135
Chocolat Day Spa on Groupon

The Company


107 S 1st Ave.
Arcadia, California 91006

Get Directions

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).


  1. This is my biggest qualm with Groupon is that I'm male yet the only tailored ads I get are for women. I think once in the past 6 months I've received an ad for a "male" product; the rest are all neutral (auto detailing) or "female" (massages, facials, shoes).

    I don't recall what information you gather when you sign up, but I assume one was "gender" and another "age" so this shouldn't be happening.

    Anyways, feel free to bump your n up to 2!

  2. I believe the majority of their customers are female. I strongly suspect they look at the aggregate numbers and send out a standard offer to everyone.

  3. If you're interested, we're maniacally focused on targeting - to the point where we flip the model and become a personal shopper for our customers on the internet. I found your blog through Felix's posts - really interesting stuff. Thanks!

  4. Erik,

    Just to be clear, lots of companies are developing smart, sophisticated approaches to marketing. What's notable about Groupon is how far behind the industry standard it seems to be.