data basics; Holding All the Cards

American Demographics, Feb, 2000 by Nancy Shepherdson

Swipe and save. For 46 percent of America's households, that has become a weekly routine. Frequent-shopper card programs, good for a wealth of discounts at the checkout, have convinced us to share some of the most intimate details about our lives: our grocery purchases. Eventually, we forget to worry about all of that data piling up somewhere, revealing our preferences in everything from hygiene products to junk food. As far as most people can tell, few grocers seem to be doing much with the information anyway.

But that may change as data-mining companies roll out new tools to help predict and mold consumer behavior through beefed-up loyalty programs. Coming soon to a mailbox (or e-mail inbox) near you: recipes, free merchandise, opinions surveys, alerts about products you buy, and other goodies. Coming soon, that is, if you prove to be the kind of customer that grocers and food manufacturers want to attract.

Until now, frequent- shopper programs have not been very good at identifying the most attractive customers for this red-carpet treatment. In fact, many programs have never evolved much beyond delivering cents-off discounts to encourage everyone who applies for a card to share purchase data at the checkout. The appeal of those discounts is so popular, though, that the average household participating in any frequent- shopper program has 3.2 cards, according to figures released last year by the Retail Advertising and Marketing Association International in Chicago. And there's the rub: A grocer's best frequent shoppers are likely to be someone else's frequent shoppers, too.

A recent study by ACNielsen shows that only 10 percent of an average grocery chain's customers can be considered "top loyals" - heavy spenders who are primarily loyal to the chain. Another 20 percent of shoppers, called "top switchers" by ACNielsen, are also heavy spenders, but tend to spread their purchases among several chains. Having a frequent-shopper card does not significantly affect the purchase decisions of these shoppers.

"Frequent-shopper cards are essentially devices for measuring loyalty, not for building it," says Glenn Hausfater, managing partner of Partners in Loyalty Marketing, a database marketing company in Chicago. Hausfater, who reported on the ACNielsen study at a recent workshop in Chicago sponsored by the Advertising Research Foundation, feels that much of the potential for increasing the payback from frequent-shopper programs lies in convincing these top switchers to become more loyal. He calls it "identifying the low-hanging fruit." Doing it means mining the mountains of data collected at checkout more

effectively.Some chains, mainly smaller ones w ith less data to manipulate, already use their databases to market more effectively to customers. Since 1997, Giant Eagle, a chain of 200 stores in Ohio, Pennsylvania, and West Virginia, has tailored its Taste of Value promotional mailer based on customer purchase patterns gathered through its Advantage card program. With as many as seven inserts, the mailer contains recipes, product information, and special sales offers. Giant Eagle, however, does not mine the data to ensure that the mailing only goes to its best customers. Everyone with an Advantage card receives the promotion, whether they shop at Giant Eagle with their card once a week or once every three months.

Has the Taste of Value program increased customer loyalty for Giant Eagle? Dwayne Sigler, Giant Eagle's vice president of marketing, says his database is currently unable to provide that information. But the promotion has been so well received that some customers call if they don't receive a mailer every six weeks or so. Manufacturers participate eagerly, too, paying hard cash for access to targeted Giant Eagle mailboxes. "For them," says Sigler, "it's cheaper than doing a free-standing insert coupon to everyone in the world."

Unfortunately for many grocery chains and food manufacturers, even saving big money on couponing is not good enough anymore. "One of the things we've learned is that if retailers are just using [frequent-shopper cards] as a cents-off device, sales are not increased," says Mike Duffy, director of planning and syndicated services for Kraft Foods. "But when used to understand when, where, and why customers want things, [frequent-shopper data] will increase sales."

Kraft Foods is participating in several pilot programs to test a more powerful way to track frequent-shopper behavior. This method combines a chain's frequent-shopper data with statistical modeling that predicts shopper behavior elsewhere. Heavy hitters such as IBM, Information Resources Inc., PricewaterhouseCoopers, and ACNielsen are among the technology providers experimenting with this kind of data mining. Ultimately, the technique promises to provide information never before known about the habits of heavy spenders. At the same time, it is also likely to be quite expensive. Marketing consultant Hausfater estimates the investment in in-depth loyalty analytics for retailers can run from $20,000 to $150,000; for manufacturers, from $40,000 to $200,000. That may be a hefty price tag for retailers, but probably not as costly as extracting the same data and estimates from their databases by themselves, since their core competencies lie elsewhere.