The Data Game
Plenty of credit card providers popped up in the early '80s, but all the cards looked roughly the same: 18 percent interest, a 25-day grace period to pay off new charges, an annual fee of about $25. For card providers like Capital One Bank, the good cardholders, who faithfully ponied up their payments, were effectively underwriting the bad customers, whose defaults totaled about 7 percent of each dollar the company loaned.
But in 1988, top executives at Capital One had an epiphany: Their business could be about more than just extending credit. "[We realized] that the credit card business was not just a finance activity but an information-based activity about customers," says Jim Donehey, senior vice president and CIO of Capital One. The company started maintaining transaction records and searching various market segments to identify "perfect customers"--those who use the cards and carry a balance but pay bills faithfully, says Donehey. The strategy was simple: Offer better terms to consumers in those segments, draw in better customers and thus decrease the percentage of bad loans. "By focusing on the people who had good credit, we could waive the annual fee and increase our overall margin while actually reducing the cost to the customer," says Donehey.
Today, Capital One offers more than 3,000 different credit card products customized to the needs of 8.2 million customers.
Prior to its enlightenment, Capital One--like many other companies in many other industries--held little regard for leftover customer and transactional records. Consultant Stan Davis has dubbed that data "information exhaust" because reams of such information escape every corporation, a supposedly useless byproduct of the information age. But more and more companies are learning that information about customers past, present or potential is a potent source of value.
Donehey and others say the CIO must play a key role in turning dross into gold, data into dollars. To do so, the CIO has to understand the company's business, the data itself and its many possible uses in the marketplace. That understanding puts the IS executive in a position to make a direct impact on the corporate bottom line--particularly gratifying because that's not the reputation IS carries. "IT folks don't tend to be the sources of inspiration in crafting new initiatives [that turn customer information into products]," says Randall H. Russell, a senior manager at the Ernst & Young LLP Center for Business Innovation in Boston, who has written extensively about the value of customer data.
Opportunities Inside, Outside
Uses of information exhaust vary but can be lumped into a few general areas. Most companies keep their data to themselves and use it to customize their products or to target their marketing efforts more accurately. Some other brave souls are making their data available in one form or another to outsiders.
Capital One's story is an example of mass customization, tailoring products to meet the exact needs and promises of each customer. Mass customization might be defined as finding the right product for your customer, for example, tweaking interest rates and credit lines to please a customer and still make money. As Russell says, a company pursuing a mass-customization strategy breaks down its products to the smallest possible units that can be mass produced and then cultivates its expertise in putting those pieces together in the right combination to meet the individual needs of the customer.
Targeted marketing is, in some respects, the converse of mass customization: finding the right customer for an existing product or service. Taken to the extreme, it means one-to-one marketing, knowing each individual customer's preferences and presenting him with only those offers he is most likely to pursue. London-based online supermarket chain Tesco PLC, for example, uses preferences indicated by members of its frequent-shopper card program to inform its coupon mailings. For example, a vegetarian has no use for "meat vouchers" and conceivably might be offended if she received them in the mail; Tesco's database dictates that the coupon envelope mailed to a vegetarian won't include such vouchers, says company Press Officer Danielle Byrne.
Supermarkets are collecting many terabytes of customer-purchasing data through affinity memberships. Tesco has 9.5 million members who shop once or twice a week on average, yielding "massive, massive amounts of data," says Byrne. It seems a bit naive to think grocers are exerting such efforts only for the sake of minor tweaks to their direct mail campaigns. But even at that, it's clear the business that can target potential buyers more accurately, lowering marketing costs while simultaneously increasing returns, has a leg (and an arm) up on the competition. "You can't market unless you understand who your customers are," says Aaron Zornes, a Burlingame, Calif.-based executive vice president and director at consultancy Meta Group Inc.
Information exhaust needn't be limited to internal uses. MasterCard International Inc. is pursuing a different path: making data available to its member banks in order to add value to MasterCard's product offerings and win member loyalty. The company does not offer credit cards; its customers are banks that provide the cards under the MasterCard brand. MasterCard's data warehouse helps its member banks identify the behaviors of various consumer segments. (For example: "Dual-income urban dwellers spend heavily on entertainment, meals and travel.") The banks in turn use the data for marketing purposes, often by marrying it to data from other sources, says Anne Grim, senior vice president of global information services for MasterCard in Purchase, N.Y. MasterCard's member banks typically have the same data already in their possession but don't store it in a way that is easy to access and understand, Grim says. In many cases, MasterCard must work closely with external customers to make sure customers are able to get the maximum value from data. That assistance gives the banks a boost and in turn makes MasterCard a more attractive business partner.
Similarly, providing individual customers with a better view into their own purchasing habits can be an enticement to buy. A simple example is Federal Express's package-tracking information database. "The whole idea of letting the external world access your internal database is customer service," says Zornes.
Perhaps the most obvious way to turn data into money, though, is simply to sell it. According to the "infomediaries," companies that deal solely in bartering information, the appetite for data of all sorts is growing dramatically. "We're seeing a much wider set of industries using our data," says Don Miller, vice president of technology development and operations for Experian Information Solutions Inc. (formerly the credit reporting arm of TRW) in Orange, Calif. Experian has plenty to work with: The company's largest database has about 8 billion rows and records as many as 60 million updates each day.
As more businesses and consumers become buyers of data, the field widens for more sellers. Merck-Medco Managed Care LLC. is a health-care company cashing in on the demand for data. The Montvale, N.J., company started as a claims processor for the insurance industry. However, the claims records have yielded trend data that pharmaceutical companies, for instance, now purchase to examine the effectiveness of prescribed treatments. Quest Informatics is another example from the health-care industry. The Rutherford, N.J., company collects and analyzes clinical laboratory data that comes from a variety of sources including its sister company, Cambridge, Mass.-based Quest Diagnostics Inc., which does lab work for physicians nationwide. Quest Informatics produces "actionable" clinical and strategic information for purchase while still maintaining patient confidentiality, according to Brian Griffin, manager of product operations for the company.
Mind Your Own Business
Initiatives turning data into products must originate with a solid understanding of corporate goals. If ever there were an area that demands that IS department be in tune with business objectives, this is it. The job requires more than throwing data into a warehouse and doing a little data mining, say the CIOs who already are exploring the possibilities of information exhaust. "The CIO is changing from a technology role to a technology-business integrator," says MasterCard's Grim.
Grim says the IS organization that builds a data warehouse, tosses the keys to the marketing group and washes its collective hands is simply wrongheaded; much of the potential value of the data will be lost. People on the marketing side of the fence typically feel removed from the process of capturing customer data. The situation requires the IS group to form partnerships within the company and get all hands involved. "If the IT people have a foot in the business camp and understand what the business objectives are, they can play a leading role in demonstrating how information can be used," she says.
Capital One's Donehey says part of his role as CIO is to expose business people continually to new technologies. Knowing the possibilities offered by state-of-the-art computing, the CIO can foster a what-if mind-set in other business areas.
"Once you've got their imagination cooking, you've made tremendous strides," says Donehey. That's quite a reversal of fortune for the IS function, which in the days of long, inflexible project development cycles was more likely to kill such thinking than inspire it.
Know the Data
At the same time, there are some skills that IS alone brings to the party. The CIO has an intimate understanding of the data that is at the company's disposal. Often that data derives its value from being correlated with information from another area of the business or an outside source. If that's being done in a data warehouse, typically there are challenges associated with modeling the data correctly and flexibly enough. "You can never underestimate the magnitude of effort involved in integrating disparate data sources," says Larry McAferty, CIO at Source Informatics, a health-care information provider in Phoenix. Source Informatics is among the so-called infomediaries whose business is based entirely on data: They take data in from outside sources; process, package and analyze it; and either sell it back to the original source or find new markets for the information.
Data modeling and analysis are core competencies for infomediaries. While the average corporate IS shop may have to play catch-up, most are making progress. "In general, legacy systems are getting better, and order-entry shops are using better controls [to make sure data is accurate]," says Ken Rapp, president of DynaMark, a 300-employee data service bureau in Arden Hills, Minn., that processes data primarily for financial companies.
Sometimes customer data doesn't take on significance unless it is tracked over long periods of time. Trends are valuable. For example, Capital One tracks variables that can indicate how likely a customer is to default. "You have to do it over a long period of time; you have to keep a lot of information for a long time to find the trends you are looking for," Donehey says.
Dirty data is another issue. Customers move, change names, pay with different means. A simple typographical error by an order taker can create inconsistencies that are hard to spot and correct. Software tools from companies such as Vality Technology can help straighten out messy data; CIOs can build skills into the IS organization for managing those issues or they can choose to outsource modeling and cleaning chores. "Data contamination is not an insurmountable problem," says Manish Acharya, director of marketing for warehouse vendor MicroStrategy Inc. in Vienna, Va. "With this kind of money at stake, people will invest what's necessary to get their data clean."
CIOs also need to watch out for a few yawning pitfalls that await the unwary company that offers up customer data for sale at the individual level. Customer privacy is one issue, particularly in sensitive industries such as health care and finance (see "Up Close and Personal," CIO, May 15, 1997). Another is the risk of letting the corporate customer list fall into the wrong hands. Companies that have been in the list-brokering business for years have developed standard practices for managing list usage, according to Don Hinman, vice president of Acxiom. Those practices include examining each potential customer's intended use of the data and seeding the lists with "decoy names" that will indicate whether the list is being used in the agreed manner.
Each company must weigh the risks against the potential rewards. By all accounts, you have to start somewhere. "If customer information is as important as we think it is, a CIO not running a customer service application may have a problem," says Rayport. "Most companies are going to have a hard time figuring out what consumers want and how to customize products for them if they don't figure out how to collect and exploit the value of that information," he says.
Staff Writer Derek Slater can be reached at firstname.lastname@example.org.