Marketers hone targeting

By Published on .

With their rich stores of information about customer behavior, demographics, purchasing activity and other valuable information, data warehouses are helping marketers better reach and target consumers, as well as improve efficiencies at their stores.

While data warehousing has been in vogue for at least a decade, new tools using video and Web data are providing marketers with even better ways to build bigger data warehouses with a host of business applications.

Data warehousing is the use of huge databases on customers, prospects and products to drive corporate strategy and tactics. Wal-Mart Stores, Bentonville, Ark., has a warehouse with more than 100 terabytes of data on NCR equipment.

That's 100 trillion bytes, equivalent to 16,000 bytes for every one of the world's 6 billion people. "The foundation for a data warehouse is still customers, sales and inventory," says Nelson Gomez, NCR account director. Wal-Mart can learn which items are moving, what different types of customers are buying products, and how often it must restock shelves through the integration of its various data-warehousing solutions.

Even nontraditional data types, such as video and graphics of store designs, can deliver even more insight when cross-referenced to the warehouse. Merchants can learn which products are examined and then replaced in shoppers' carts, for instance, or other patterns of behavior. "You can look at loss prevention," Mr. Gomez says. "You can track video of losses to find patterns, rather than just stare at a screen."

Wal-Mart is getting so much value from its in-store databases, in fact, that it has stopped most outside marketing for its Web site. Instead it's using tabs in its flyers, signs in stores, and ads on its in-store radio network heard by 100 million customers per week, says spokeswoman Cynthia Liu. "Our focus is on using our stores to tell our story," she says.

Smaller data warehouses can also yield valuable insights. Fingerhut, a mail-order unit of Federated Department Stores in Minnetonka, Minn., has a "mere" 6 terabytes in its warehouse. But 30 analysts are kept busy trying to predict whether a customer will buy from a specific catalog.

Randy Erdahl, director of business intelligence for Fingerhut, says his company mails 100 catalogs per year to lists with 6 million people on them, so maximizing yield is vital. But even more value is emerging as Fingerhut moves into e-mail marketing. "We're not just looking at transactional data but shopping behavior," he says, using "behavior [consumers] have shown on the Web site as indicators in targeting e-mails."

Fingerhut, like most merchants, is increasingly dealing with what Mr. Erdahl calls "multichannel customers" and finding that its databases are the best way to understand them.

The problem with extracting data from most Web data warehouses is that users are anonymous, and there are privacy issues. But if marketers are careful, they can use the data in ways that will protect the consumer and add value to their marketing efforts.

I-Frontier, a Philadelphia interactive-marketing agency that counts AstraZeneca and Universal Studios among its clients, has launched a backend tool that tracks user Web-site activity back to promotional efforts, using data collection.

"We can see that 100 users clicked on banner A on Yahoo!, arrived at a client's site, looked at four pages, spent a total of 4.5 minutes, then left without completing the registration form," says Jeremy Lockhorn, director of media technologies for I-Frontier.

However, combining anonymous Web data with the customer information in a store warehouse can be problematic. Bill Wise, VP of US Media Networks at DoubleClick, the ad network that also owns Abacus Direct, a direct-mail database, learned this first-hand last year. DoubleClick's move to merge the databases led to a 10-month investigation by the Federal Trade Commission.

DoubleClick was cleared in January, but along the way Mr. Wise learned what data-warehouse owners can and can't do. "We can't link personally identifiable data to serve advertisements," he says. "But we can anonymously take a look at behavioral data."

The problem is, "You need a critical mass of data" to extract the meaning. Thus, most data-warehouse analysis is done by third parties such as Information Resources Inc. in Chicago. Privacy requirements mean the output doesn't yield personally identifiable information.

Even the output of new data-warehouse products such as IRI's eMarkit yields general trends, not specific consumer data, says Dan Sherr, exec VP for IRI New Ventures.

"We use cookies [tracking tags passed to user browsers by ad servers] in combination with HTML tags and Web bugs to connect the Web site to our consumer panel," says Mr. Sherr. "We're able to determine when a household who is a member of the IRI panel has gone to one of their Web sites."

Behavioral changes in IRI panel households are then extrapolated to the general market, he says. Most retailers seem satisfied with the arrangement, says John Thompson, VP-worldwide marketing at WhiteCross Systems, a Chicago vendor of customer relationship management (CRM) services.

"One of the great shames in America is the stores have given out these loyalty cards and they've let it degenerate into pure price competition. They sell that data to IRI and Nielsen, and the retailers get no benefit in the end. That's an industry that's given up the benefits from collecting its data."

That is a mistake, says Tom Ronk, CEO of Salesmation, Newport Beach, Calif., another CRM vendor owned by e-Synergy Ltd. of London. "The lines between advertising, marketing and sales are blurring. You can advertise, market and sell at the same time, driving people directly to the transaction," he says, using a data warehouse.

Outsourcing analysis of a data warehouse can also bring this within the range of most advertisers' budgets, Mr. Ronk says. Salesmation charges between $2,000 to $5,000 per month for a service that can extract data from Web sites, phone banks or direct-mail channels and come up with business rules from it.

By registering users and disclosing the use of their data, valuable opportunities can emerge from the rules, Mr. Ronk says. "If a male between 40-50 comes to the site and makes over $100,000 per year and has looked at the golfing section before, present a pop-up or send an e-mail offering a 10% discount," he says. "So you can spiral a brand and target with precision the people most likely to buy your product."

Stores will be able to take advantage of such insights at Web-speed using new shelf price labels containing liquid crystal displays [LCDs], according to Tracy Flynn, VP-product and solution marketing with NCR. In Connecticut, he says, LCD shelf labels can already replace hand-stamped prices.

Not only do the shelf labels eliminate human error, but "if you get a price change on a 16-ounce bottle of ketchup you can also change prices on other brands and other sizes."

Most Popular
In this article: