Bullseye Analytics

By now, most people in the analytics industry have seen the infamous New York Times article detailing how eerily accurate…

By now, most people in the analytics industry have seen the infamous New York Times article detailing how eerily accurate Target’s inferred statistical models can be. But believe it or not, the amazing part isn’t so much that such statistics can be gathered, but rather how.

Think about it—Target gathers data from credit cards, email addresses, and phone numbers to associate a unique “Guest ID” with each customer. There’s also web log data associated with your email address any time you open one of Target’s emails, visit their website, or call customer service. Already, at least four systems are in play: data from credit cards, web visits, the phone center, and email campaigns. Associating all that data with a single unique ID must be quite the challenge—but then again, the results are pretty stellar.

(If you want to add a degree of difficulty, think of how many customers have more than one credit card, or how many families have at least one credit card that changes hands. )

Will predictive analytics forever be known Target’s game? If you’ve been in a Gap, Urban Outfitters, or Apple store recently (just to name a few) you might have opted to have your receipt emailed to you, rather than printed—bingo, that’s credit card-to-email association. To me, it’s also evidence that at least a few companies are looking for data-driven boosts in sales revenues.