Monday, November 15, 2004

Pop-tarts and beer.

Ok, now that I have your attention, I would like to talk about data management. Yesterday's New York Times discusses Wal-Mart's predictive technology, in which they mine past sales records to anticipate purchase surges in specific items. The unexpected result for an oncoming hurricane was that Pop-tarts sell at about 7 times their normal rate. After that, beer is a no-brainer. What else would you use to wash down the Pop-tarts? Calling Dave Barry...

I went through the whole article, and I think something is strangely missing. Yeah, Wal-Mart has mountains of data, and they use that information to stock products and to admonish suppliers, but at what point is this different (except in scale) from the corner grocer? Moreover, a lot of other big retailers are also collecting and mining data. Why is Wal-Mart so successful relative to them? I still don't know. However, data-based predictive approaches can be hugely effective in many settings, as David, Joanna and Marcy can all attest!

Lastly, the article mentions customer's privacy concerns. This section of the article was especially brief and vague. Since Wal-Mart apparently tracks products and shopping carts, not identified customers, it may be that concerns about linking e.g. credit history with particular profiles did not belong in this article. Other retailers with loyalty card systems are far more interested in knowing all about you. This is a very complex issue, and deserves better discussion than I can manage. I do have to say that the genie is already out of the bottle, and that information, once given out, will no longer be erased or forgotten.

No comments: