Sometimes the results from data analysis are just plain absurd. Web logs show that one of the most common search phrases entered at Computerworld’s Web site one week was “Ryder trucks.” Nobody knows why.
More often, data mining yields unexpected nuggets of information that open the company’s eyes to new markets, new ways of reaching customers and new ways of doing business.
Samsung Electronics America Inc. in Irvine, Calif., for example, analyzes reports from 10,000 resellers to identify “lost deals,” that is, orders that went to competitors. In one business segment – computer monitors sold to the health care industry – Samsung found that 40 per cent of the lost deals went to one competitor and represented 80 percent of the total lost revenue.
Knowing those chilling facts, Samsung is working more closely with hardware integrators in the health care field to win more of those orders, says Helman Lukito, a Samsung marketing manager. But until the company collected and analyzed the data from its reseller extranet, which is powered by software from San Francisco-based Allegis Corp., Samsung was “going into the market blind,” Lukito says.
Sometimes a company likes what it sees when it analyzes the data. Catalog company Lillian Vernon Corp. in Rye, N.Y., has always been strong with female shoppers but had trouble attracting male customers. Using Web analytics from Fireclick Inc. in Los Altos, Calif., Lillian Vernon discovered that men – who might not flip through a Lillian Vernon paper catalog – were happy to shop at Lillian Vernon’s Web site. Since that discovery, the company has placed products that appeal to men more prominently on its Web site, a company spokesman says.
An important lesson from data mining is that stereotypes are often wrong. For example, an image of motorcycle owners as Hells Angels fades away when you learn that they usually rank within the highest income bracket in their neighborhoods, says Jordan Modell, senior vice president of database marketing at Wunderman in New York, a division of Young & Rubicam Inc.
“What you thought previously might not be true,” Modell says. “That’s what makes data mining fun, but there’s a purpose to it, too,” because it means the company can send marketing messages to people who will be receptive, he adds.
Modell’s data analytics team takes large extracts of clients’ data warehouses, matches them with external databases and then uses mining tools from SAS Institute Inc., SPSS Inc. and Brio Software Inc., as well as direct SQL queries.
Some results are almost common sense. For example, the best markets for toy sales are places where there are lots of children and lots of toy purchases, such as Little Rock, Ark., and Macon, Ga., says Warren Foster, director of marketing intelligence at The Martin Agency, a Richmond, Va.-based advertising agency. He uses mapping software and data sets from MapInfo Corp. in Troy, N.Y., to pinpoint areas where clients should advertise.
But some results are counterintuitive. For example, Foster says Phoenix isn’t a good place for selling golf clubs, despite the many golf courses nearby. Why? Tourists and conventioneers play in Phoenix, but they don’t buy their clubs there. It turns out the best places for golf club sales are Rochester, N.Y., and Detroit, where there are many avid golfers who buy their clubs locally.
Perhaps the most important value of data analytics is that they help business managers make decisions based on facts rather than on old assumptions or gut feelings, many users say.
New York-based cosmetics and fragrance firm Coty Inc. uses financial analytics from Hyperion Solutions Corp. in Sunnyvale, Calif., to settle internal debates about resource allocation for various product lines, says Jim Shiah, senior vice president and corporate controller. The facts about sales and profitability of each product line have put an end to the “democratic anarchy over resource allocation and elevated the level of discourse,” Shiah says.
Ace Hardware Corp. in Oak Brook, Ill., finds that the facts gleaned from its data warehouse can persuade reluctant store owners to raise or lower their prices. “We had one store that only sold one wheelbarrow a year, but when he lowered the price, he sold four in one month and made more money than he did the previous year,” says data warehouse architect Mark Cothron. Ace uses analytical software from Informatica Corp. in Redwood City, Calif.
And conventional wisdom held that batteries and light bulbs have to be priced very competitively because they can be bought at many different places. But data mining found that shoppers don’t really do a lot of comparison shopping for those items, says Diane Flynn, technology business manager at Ace. “People have their long-held theories, but with the facts, you can prove them right or wrong.”
