How Visa Predicts Divorce
By scrutinizing your purchases, credit companies try to figure out if your life is about to change—so they’ll know what to sell you.
If you ever doubted the power of the credit card companies, consider this: Visa, the world’s largest credit card network, can predict how likely you are to get a divorce. There’s no consumer-protection legislation for that.
Why would Visa care that your marriage is on the rocks? Yale Law School Professor Ian Ayres, who included the Visa example in his book Super Crunchers, says “credit card companies don't really care about divorce in and of itself—they care whether you're going to pay your card off." And because people who are going through a divorce are more likely to miss payments, your domestic troubles are of great interest to a company that thrives on risk management. Exactly how the credit industry does it—through sophisticated data-mining techniques—is a closely guarded secret. (Visa did not respond to a request for comment. UPDATE: After this story ran, the company responded with the following statement. “Visa does not track or monitor cardholder marital status, nor does it offer any service or product that predicts a potential divorce. These claims are false and any media outlets or authors citing that Visa has such capabilities are inaccurate and wrong.”)
The mobile social network Loopt or its competitors could conceivably predict with 90 percent accuracy where an individual will be tomorrow.
Predicting people’s behavior is becoming big business—and increasingly feasible in an era defined by accessible information. Data crunching by Canadian Tire, for instance, recently enabled the retailer's credit card business to create psychological profiles of its cardholders that were built upon alarmingly precise correlations. Their findings: Cardholders who purchased carbon-monoxide detectors, premium birdseed, and felt pads for the bottoms of their chair legs rarely missed a payment. On the other hand, those who bought cheap motor oil and visited a Montreal pool bar called "Sharx" were a higher risk. "If you show us what you buy, we can tell you who you are, maybe even better than you know yourself," a former Canadian Tire exec said.
Credit card companies have also used predictive modeling to answer questions such as, has this cardholder recently moved? "There's a whole market out there that has tried to predict whether someone has just moved, and to be first with offers," says Bob Grossman, director of the Laboratory for Advanced Computing at the University of Illinois at Chicago. "Those kinds of things tend to be pretty high value." If a credit card issuer can quickly determine that a cardholder has moved, then the issuer's marketing partners—a home refurb business, for instance—can be the first to swoop in.
Last year, American Express began offering select cardholders $300 simply to close their accounts and walk away—individuals who the company clearly felt were too much of a risk to keep on its books. And the factors that go into such a calculation have become considerably more sophisticated than the simple matter of whether cardholders have paid their bills on time.
The credit card industry is just an early adopter of a number-crunching game that’s increasingly transforming businesses from airlines to gambling. "Thirty years ago, loan officers used to look you in the eye and tell you whether you were the right kind of person to trust for a loan. That was a really inaccurate approach. Just using FICO scores did a much better job," Ayres says. "Credit card companies started using a similar approach in deciding whether to issue and how to price their card. It's getting to be a more nuanced statistical game."
Other industries have bolstered their bottom lines by predicting how consumers will behave, according to Super Crunchers. UPS predicts when customers are at risk of fleeing to one of its competitors, and then tries to prevent the loss with a telephone call from a salesperson. And with its “Total Rewards” card, Harrah’s casinos track everything that players win and lose, in real time, and then analyze their demographic information to calculate their “pain point”—the maximum amount of money they’re likely to be willing to lose and still come back to the casino in the future. Players who get too close to their pain point are likely to be offered a free dinner that gets them off the casino floor.
The statistical guessing game is also becoming one that consumers can play. For example, the New York-based startup Hunch offers personalized recommendations after users answer a series of questions that give the site a sense of their tastes. Do you live in the suburbs? Do you like bumper cars? Are you more likely to spoon or be spooned? Out of this examination, Hunch generates a “taste profile” for each of its users.
Hunch then looks for statistical correlations between the information that all of its users provide, revealing fascinating links between people’s seemingly unrelated preferences. For instance, Hunch has revealed that people who enjoy dancing are more apt to want to buy a Mac, that people who like The Count on Sesame Street tend to support legalizing marijuana, that pug owners are often fans of The Shawshank Redemption, and that users who prefer aisle seats on planes "spend more money on other people than themselves."
Through “machine learning,” the Hunch algorithm is developing a sense of what individuals with a certain taste profile will prefer—a sense that is being improved with each new user of the Web site. This knowledge then allows the system to make predictions of what an individual user might like: a movie soundtrack, a cat name, a restaurant in Los Angeles. Kelly Ford, the startup's vice president of marketing, notes that while the credit card companies rely on a small set of inputs to make predictions, Hunch's questions collect "nearly unlimited aspects of who you are and how and what you think."
As new sets of data are collected about our lives, that data will contain a new set of predictions about us, waiting to be mined. The question will be how much control we have over that process. At the South by Southwest Interactive conference in March, Sam Altman, chief executive of the mobile social network Loopt, said that by using the available data, Loopt or its competitors could conceivably predict with 90 percent accuracy where an individual will be tomorrow.
He said hedge funds have contacted Loopt to try to purchase its data set so that they can forecast how much traffic a particular store will get. The startup declined. It doesn’t take much predictive prowess to see that these issues will become major matters of contention in the years to come.
Correction: The headline of this article originally referenced MasterCard, not Visa.
Nicholas Ciarelli is the former publisher of Think Secret, an Apple news Web site. He currently works on the product team at The Daily Beast.