Sonics Beat Bulls? Ice Cream Saved? Thank a Data Mine
PITTSBURGH — Every time you buy something, it yields clues marketers can use. Major companies have known this for years. Retailers tell your age by the clothes you buy; phone companies figure out what services to sell based on what you already pay for. Now, small companies are jumping into the analysis game.
Using computers to mine the riches of "data warehouses," small- and mid-size companies are figuring out everything from how to sell potato salad to which substitute to send in a professional basketball game. The result: better business decisions, no matter what your business is.
"In the last 12 to 18 months, this application has really taken off," says Van Symons, a data-warehousing executive for IBM's AS/400 computers. "Sometimes the smaller company is able to implement this much more quickly" than larger companies. For example:
*When its freezers failed in July 1995, a small British ice-cream manufacturer was staring at a major disaster. But because of its data warehouse, it was able to analyze quickly the sales to its retailers, discount the merchandise, and place its entire inventory before the stock melted.
*The Orval Kent Food Company makes 34 kinds of potato salad as well as other prepared salads. A year ago, it installed a data warehouse to carefully analyze which salads were selling where. The system not only allowed the Wheeling, Ill., company to consolidate 350 sales reports into 50, it also cut the reporting time from several weeks to one. "I call it one version of the truth," says Dennis Lynch, director of information services for Orval Kent.
*Even professional basketball teams have gotten into the act. Using an IBM software program called "Advanced Scout" during last year's playoffs, the Seattle Supersonics reportedly made an interesting discovery. When its reserve player Frank Brickowski guarded the Chicago Bulls' Dennis Rodman, the team outscored the Bulls. In the next two games, the Supersonics gave the reserve player more playing time and the team won its next two games before falling to the champions.
That's the point of a data warehouse. It allows groups to collect and sift information in new and more flexible ways, thereby fostering smarter decisions.
In a recent survey of 62 organizations that had built data warehouses, Stephen Graham of International Data Corp. found that an average warehouse cost $2.3 million and paid for itself in less than two years. (Those findings exclude 17 groups that saw either phenomenal returns or disastrous losses after building a data warehouse.)
Currently, long-distance companies use data-mining to identify types of customers. With high-value customers, who have more than one phone line and use extra features such as call-waiting, a telephone company would ensure that the customers get red-carpet treatment every time they call, are offered incentives to stay with the company (such as a frequent-caller program), and are nudged to consider purchasing even more extra calling features.
Now, the technology has gotten cheap enough - $250,000 and up - that smaller companies can begin competing on equal footing. "There's a huge opportunity for mid-size and small businesses," says consultant Ronald Loback of Sage Communications Corp. in Scottsdale, Ariz.
One concern is privacy: Tracking individuals' buying habits is growing more sophisticated. Even data warehouse advocates say that the industry must address this issue.