The Data Game
Sophisticated marketing wizards can track just about everything a consumer does
In John Grisham’s novel The Pelican Brief, a beautiful law student is running for her life because she knows too much about a political conspiracy. Played by Julia Roberts in the movie of the same name, she crisscrosses the Mississippi delta with a gang of mobsters and the FBI at her heels. But she has trouble shaking them off; the bad guys can follow her wherever she goes, as long as goes, as as she keeps using her credit cards. To cover her tracks, she has to put away the plastic and stick to cash. Sound far-fetched? Except for the presence of spies and mobsters, it isn’t. In real life, the people chasing Grisham’s heroine would be sophisticated marketing wizards, armed with the best computer technology money can buy, intent on gathering every tidbit of information she inadvertently scatters along the electronic trail.
Their goal would be to find out what she does all day—when she gets up, whether she turns on the television or radio, which channels she flips through, what she eats for breakfast. Does she shower or bathe? Use deodorant or antiperspirant? Roll-on or spray? Does she linger in front of the mirror, or dash out the door? What route does she take to school or to work? If she picks up a soft drink on the way, where does she buy it, and does she purchase anything else at the same time? Heaven forbid that she should have an extramarital affair or order racy underwear via the Internet, because if she does, it’s a safe bet the marketing team will soon know that, too.
Big Brother really is watching—only he isn’t the omniscient, all-controlling state portrayed in George Orwell’s chilling political fantasy 1984. Instead, he has strolled right up the front walk in the guise of friendly corporations offering a plethora of new products and services that ostensibly make life easier or cheaper for consumers. Credit and debit cards, loyalty programs and the Internet—all such perks and conveniences allow companies to keep tabs on just about everything customers do. It is not espionage or intrigue; it is commerce. Meanwhile, governments are grappling with the question of what, if anything, should be done to regulate this private-sector eavesdropping. This fall, for instance, Ottawa plans to introduce legislation that would extend its current privacy legislation, but it is not clear yet how far it will go. In the United States, Vice-President Al Gore is proposing an electronic bill of rights that would not only limit how much information is being collected on an individual, but would force companies to reveal how it is being used.
The minutia of daily life—what people eat, wear, watch, ride in, play with and think about—is quickly becoming one of the most sought-after commodities in the industrialized world. It is no coincidence that the computer experts who are making it possible have dubbed this business “data-mining.” For the corporations that collect, store and analyze it, the information represents a rich vein of raw material that, properly managed, has the potential to make some of the world’s wealthiest organizations a great deal wealthier. The companies that are leading the way in data-mining hope to accomplish this in two ways. The first is to invent increasingly smarter software programs that can sift through huge amounts of information and uncover patterns of consumer behavior that most marketing professionals would never even think to ask about. The best known example is the so-called beer and diapers correlation, which came to light after a random bit of data-mining by an executive at NCR Corp. revealed that men who pop out to buy diapers in the evening are disproportionately likely to pick up beer on the way home. Across North America, major retailers now dream of discovering their own unique correlations, much as astronomers long to stumble across an uncharted comet. In this case, the result is not a contribution to the store of human knowledge, but a rearrangement of store shelves, with a consequent contribution to the bottom line.
The second, and more controversial, goal of data-mining is what IBM executives call “mass customization.” Translated into everyday language, it means using a variety of sources of information to create dossiers on individuals that will enable IBM’s corporate customers, and a growing number of other companies, to determine whether a particular person is a likely customer and what, exactly, needs to be done to encourage that person’s patronage. In the words of Norbert Dawalibi, IBM North America’s vice-president of networking sales: ‘The target market is each individual.”
Naturally, all of this clever marketing requires a great deal of sophisticated technology, which is where IBM, NCR and a host of other computer companies come in. After several false starts comparable to the setbacks experienced by personal computer manufacturers in the late 1970s, the market for data-mining software is exploding. According to International Data Corp., a market research firm based in Framingham, Mass., companies worldwide spent $16 billion on data-mining technology and expertise in 1997. The figure is expected to reach $21 billion this year and as much as $26 billion in 1999. Experts say the quantity of data contained in computerized warehouses is doubling every 18 months. This week, IBM will announce that it is setting up a separate company to develop and market products that will help clients manage their customer relationships—another euphemism for electronic snooping. IBM expects the entire “business intelligence” market to keep increasing at a rate of 35 per cent a year.
The extent to which big retailers and manufacturers are capable of monitoring their customers’ behavior would probably surprise most consumers. Take WalMart. Don’t be fooled by the folksy greeters; the Bentonville, Ark.-based discount chain actually controls one of the largest data-collection systems anywhere in the world. Tom Watson, managing editor of Computing Canada, a weekly newspaper, calls the company “the poster child of the data-mining industry.” Aside from low prices, the retailer prides itself on its ability to reconstruct customer behavior from cash register tapes. It claims to be able to trace the exact route each customer has taken through each store, based on what he or she has purchased.
Major corporations see themselves as having absolutely no choice but to get into data-mining
The trend goes beyond the major multinationals. At the Torontobased Speedy Muffler King chain of auto repair shops, the goal is to increase revenue by digitally searching the company’s own data and vehicle ownership records. They then contact people whose cars might be due for a new muffler, a set of shocks or a brake job.
The good news, for people who value their privacy, is that crunching all this data is far from a simple task. Not every company is successful. The T. Eaton Co., for example, poured upwards of $2 million into a data warehouse earlier this decade, only to have the project fall apart during the retailer’s 1997 brush with bankruptcy.
The bad news? Eaton’s is looking for a way to get back in the game, because it, like every other major corporation, is viewed as having absolutely no choice. “Lots of energy is going into trying to figure out how to do this, says one company insider. “Even Wal-Mart is still working on it. When somebody figures it out—boy oh boy, it will be power beyond belief.”
Even some people who make their living in the data-mining business say the potential to pry into people’s lives is frightening. “It gets a little Big Brotherish,” says John Shoesmith, research manager at International Data Corp. (Canada) Ltd. Other executives say they refuse to respond to requests for personal information because they know how the business operates. They worry that every detail they cough up will find its way into some personal dossier. Jonathan Tice, marketing director for Netscape Canada, for example, will not subscribe to magazines that sell their mailing lists. A lawyer at one of Canada’s largest retail chains insists that his own firm would never misuse the information it collects, but he worries about how some other companies might respond. “They can sell this information to whomever they like, and without my consent.” Ron Fryer, an NCR marketing specialist in Denver, does not like it when he is asked for personal information, and jokes about how he gives retailers inconsistent details to mess up their databases.
Interestingly, the most frequent assertion made by people employed in this business is that they would never sign up for an Air Miles card. Says Shoesmith: “Do people even wonder why they have these cards? Do they actually think it’s all about giving away free plane trips?”
Run by a Toronto-based company called Loyalty Management Group Canada Inc., the Air Miles program is the brainchild of president and CEO Craig Underwood, a 41year-old West Virginia native, and two partners. In 1991, while working as management consultants for Bain & Co. in Boston, they became intrigued by a British customer loyalty program that used coupons to reward people for shopping at participating retailers; consumers who collected enough coupons could trade them in for free flights. “In a way, the coupons were just like Green Stamps,” Underwood said. He liked the idea, but decided it would be far more effective if there was a way of keeping an electronic record of every purchase. He and his friends promptly acquired the Canadian rights to the British program and launched Air Miles in Canada in 1992, handing out plastic cards to every member and installing digital scanners beside the cash registers of participating retailers.
Today, more than 7.2 million households belong to the Air Miles program, accounting for 56 per cent of all households.
The 134 corporate sponsors include the Bank of Montreal,
Shell Canada, the Bay, Blockbuster Video, several food chains including Dominion, A&P and Safeway, and, in Ontario, the provincial liquor control board’s retail outlets. (They also include Maclean Hunter, publisher of this magazine.)
On one level, Air Miles is a conventional rewards program that allows people to “earn” free air travel and other incentives while shopping for gas, groceries and other items. Corporate sponsors pay a set amount up front for a designated number of air miles; the Loyalty Group then buys plane tickets in bulk from the airlines, and earns a markup on the travel rewards.
The key to the program, however, is the use of the Air Miles card to research consumer preferences and spending habits. One source of data is the sign-up form for new members, which, in addition to requiring the person’s name and address, asks for details about age, family size and total household income. According to a case study of the program by a postgraduate student at the Harvard Business School, this information has proved to be a treasure trove for Air Miles sponsors, most of whom had never had access to such personal data.
On top of that, the Loyalty Group collects more information every time a customer makes a purchase. Through its consulting arm, Underwood’s company sorts and packages the data on behalf of its corporate sponsors—helping them decide, for example, where to locate new stores and how to maximize sales to each customer. It also encourages them to set up so-called cross-promotions as a means of collecting and sharing information with other members. Sponsors can pay extra for a separate peek at one another’s data: Holt Renfrew, for example, might want the names of Bank of Montreal gold MasterCard holders for a special promotion. Although the Loyalty Group guards information like an old-fashioned prospector sitting on top of a gold find, it does not take a great leap of imagination to conclude that if Holt Renfrew and the Bank of Montreal are sharing personal customer data, then so are other Air Miles members. Anything Blockbuster Video knows about an individual’s viewing preferences, the local liquor outlet can know, too—and vice versa. The possibilities are endless. “The whole company is really a data miner, a marketing outsourcer,” says Shoesmith. “That’s what they do.” Adds Underwood: “We operate on the principle that the more you know about a customer, the better you can service that customer.”
MINING MONEY FOR POLITICIANS
If mote database toothpaste marketing and cars, can be why used not to use proit to sell politicians? In fact, it is already happening. Back in the 1970s, U.S. political consultants pioneered the art of using direct mail to influence voters in federal and state-level elections. Today, many of those same experts are using powerful computers and fancy software to identify potential supporters and increase a candidate’s chances of winning in a tight race.
Veteran U.S. political consultant Edward Grefe doesn’t mince words about the importance of information technology to
political campaigners. "The right data means votes. The right data means money,” he recently told readers of Campaigns & Elections, a Wash ington-based monthly magazine. By analyzing a wide range of demographic and lifestyle information on voters in a particular district, he said, a campaign “can completely tailor its message to certain constituents with multiple variables." Those variables include race, gen der, income, education, buying habits and dozens of other characteristics, any
one of which might provide a clue as to how an individual is likely to vote. In addition to helping campaign staff target specific kinds of voters, computers are being used to identify potential donors. At least 33 U.S. consulting companies now offer database management services for political campaigns; of those, 15 also offer assistance with fund-raising. One such firm, Aristotle Publishing, boasts that its fundraising software not only helps to identify potential new supporters in designated tax brackets, but even helps to track down opponents’ big-spending contributors. The product’s name: Fat Cats. K.N.
Yet even the amount of information flowing into the Air Miles database pales by comparison with what is being gathered by Interac, a co-operative owned by financial institutions that processes electronic banking transactions and purchases made with debit cards. Interac expects to handle nearly 1.5 billion transactions this year, at a rate of 42 a second. For 1997, it knows the number of outlets at which bank cards were used (330,000) and it knows the percentage of transactions that involved debit cards (22 per cent) compared with credit cards (17 per cent) and cash (50 per cent).
In other words, nearly 40 per cent of the consumer transactions in Canada last year left an electronic trail. Interac purchases alone accounted for $1.1 billion worth of beer, wine and liquor, $2 billion worth of prescription drugs and sundries, and $3.8 billion in spending at gas stations. The most recent breakdown of Interac purchases does not specify how many couples used debit cards to buy their marriage licence at City Hall, but according to the co-operative, a growing number of Canadians are doing that, too.
So far, most of the information gathered by Interac goes to waste, from a marketing perspective. The fact remains, however, that Canada’s banks are sitting on a motherlode of searchable data: who went shopping, at what time of day, where and how often, in onequarter of the retail transactions in Canada last year.
Anyone who wonders whether this information will eventually be put to use need only glance across the border to the United States, where several powerful financial institutions are at the forefront of the data-mining revolution. San Francisco-based Wells Fargo & Co. and MBNA Corp. of Wilmington, Del., one of the world’s largest issuer of credit cards, have developed software programs that electronically scan the financial profiles of hundreds of thousands of individuals and small businesses in order to pinpoint potential customers. When Canadian bankers talk about the need for huge investments in new technology, those are the sorts of services they are talking about. Indeed, the electronic revolution is one of the driving forces behind the recent wave of megabank mergers in Canada and the United States. As big as they are, Canadian banks say they are going to have to pool their resources to afford the elaborate computer systems required to identify and woo key customers.
For some banking customers, the trend could herald a significant improvement in service. Several large American financial institutions, for example, are already using computer software to assign a rating—from one to five, say—to each of their millions of customers, based on the potential value of that person to the bank. The rating. which pops up on a computer screen whenever a teller or other bank employee keys in that customer’s name or account number, helps to determine the level of service that individual receives. Top-ranked customers, those with liquid assets and healthy personal cash flow, would merit first-class treatment, which might include overlooking the odd bounced cheque. Customers relegated to the bottom of the heap, on the other hand, would receive a more perfunctory level of service. The rationale is that banks actually lose money providing services to some of their less prosperous clients; hence, it would be better for the bottom line if those people took their business elsewhere.
An ‘electronic bill of rights’ could curb privacy invasion
Similar technology is finding its way into bank branches around the world. This summer, NCR and Union Bank of Norway unveiled a project to convert 350 UBN locations in that country to more sophisticated electronic operations. Customers swipe their bank cards the moment they walk through the door, causing the bank’s computer to spit out a numbered ticket and informs tellers who is waiting. Simultaneously, a complete financial profile of that person appears on the tellers’ computer screens. While the customer is waiting for service, advertisements geared to his or her assets, liabilities and financial requirements appear on video monitors strategically positioned throughout the branch. In this dawning age of mass customization, no tool is more valuable than the Internet. Almost every major retailer in North America is now using, or attempting to use, the Net to generate new business and, in the process, acquire more knowledge about its customers. Every time someone orders a book from the Seattlebased online bookstore Amazon.com, that information goes into the customer’s personal file, adding to the store of data about that person’s reading preferences and interests. Increasingly, Web site operators are using such information to direct specific kinds of advertising at people who seem mostly likely to be receptive to those messages. Anyone who used the Internet search service AltaVista last week and typed in the word “toys” would have immediately seen an animated ad for Toys “R” Us; the word “automobiles” summoned up an ad for one of several car models.
In the computer business, those are what are known as “real-time” messages—ads that respond on the fly to whatever information is available about the user’s preferences. “They don’t just want to know what I’ve bought,” says Netscape’s Tice, one of many companies that is striving to become an expert in this emerging business. “They want to catch me before I buy something.
As a consumer, I may appreciate that. And then again, as a consumer, I may not.” Netscape’s own In-Box Direct service, for example, allows Web users to subscribe to any of 140 free online magazines. Each time one is selected, the company collects more information about that person, data that can be used to customize the online advertising he or she sees.
In the advertising industry, executives now dream of applying similar technology to television. In a few years, says Peter Swain, chairman and CEO of Toronto-based Media Buying Services Ltd., it should be possible for cable companies to monitor exactly which programs each of their customers is watching.
That information could then be correlated with household demographics—such as age and income—so that advertisers could direct specific commercials at each member of the family. Swain, whose company is the biggest purchaser of advertising time and space in Canada, calls it “addressable advertising,” and readily acknowledges that many people would consider it an invasion of privacy. In his view, however, it makes sense both for his clients and for consumers. “I personally wouldn’t regard it as an intrusion,” Swain says,
“because the alternative is that I’m seeing ads all the time that aren’t of the slightest use to me. If advertisers are able to communicate with us more effectively, that to me is good news.”
Swain has a point, but the real problem is that no one knows how far all this corporate snooping will go. The data-gathering and data-mining business has exploded so quickly that governments have had little opportunity to take stock of the situation, let alone regulate it. “Data-mining represents a major challenge to privacy because the companies who practise data-mining cannot predict what uses the resulting information will have,” says Ann Cavoukian, Ontario’s information and privacy commissioner. In a report issued last January, Cavoukian warned that data-mining “may be the most fundamental issue that privacy advocates face in the next decade.”
Despite such concerns, there are few legal restrictions on companies in Canada that collect and analyze consumer information. The federal privacy act states that information gathered by one agency should not be disclosed to other parties without the subject’s consent, and should not be used in ways that are inconsistent with the purpose for which it was originally collected. But that law, which has been on the books since 1983, applies only to the federal government itself. Among the provinces, only Quebec requires the private sector to adhere to statutes safeguarding a citizen’s right to privacy. Meanwhile, the Canadian Direct Marketing Association has developed a set of guidelines in the hope that it can persuade Ottawa to allow it to become a self-regulating trade along the lines of the medical or legal professions.
The association’s efforts may fail, however. This fall, the federal Liberals plan to introduce a bill that would broaden the privacy act to include federally regulated industries, including banking, telecommunications and transportation. Federal privacy commissioner Bruce Phillips, a former television journalist and press aide to prime minister Brian Mulroney, says the proposed legislation represents an important step forward because those sectors “are among the most massive managers of personal information in our society.” Nevertheless, Phillips wants Ottawa to go even further, by extending privacy law coverage to all commercial activity. “Anything that is as fundamental to decent society as people’s right to personal autonomy and dignity cannot be left to the whim of the marketplace,” he says. “It has to be embedded in law.”
The alternative, of course, is that consumers will have to learn how to look out for themselves. Providing false information about oneself is one option; another is to refuse to do business with companies unless they publicly pledge not to share customer data with other organizations. And if those tactics aren’t sufficient, there is always the prospect of a technological solution. Stephen Emmott, director of NCR’s Knowledge Lab think-tank in London believes that consumers who shop online in the future will use special software programs to create “shopping identities” that serve as electronic disguises. They would then have the ability to provide their full identities to companies they trust while sharing only small fragments of information with others.
Ultimately, Emmott says, the only way consumers can completely protect themselves is to recognize the value of their own information to any company with which they do business. Like merchants, they must view personal information as a commodity, and begin asking any corporation that seeks it to compensate them accordingly—be it with cash, coupons, discounted merchandise or extra loyalty points. If they want it badly enough, make ’em pay.
With ROSS LAVER, MICHAEL MacLEAN and JOHN SCHOFIELD in Toronto