Demographic Segmentation

Computerworld - To gain an edge in nabbing profitable clients, savvy companies often use segmentation techniques to slice and dice customer data in an effort to match the best sales prospects to specific products and services.

All customers leave data "footprints" about themselves in each transaction they make, which companies can in turn analyze to help determine future buying preferences. These bits of customer relationship management (CRM) information - such as the date of purchase, the buyer's location and the types of products purchased - can be used by a retailer to create a sales and marketing campaign for, say, women between the ages of 35 and 55 in the Chicago area whose household income is in the top 10 percentile and who show a predisposition for buying upscale shoes.

To mine such data electronically, companies use online analytical processing (OLAP) or data-mining applications from vendors such as SAS Institute Inc. in Cary, N.C., and Information Builders Inc. in New York.

But there are big challenges in gathering and making use of this data. The data itself must be solid and reliable, the categorization and definitions in the data must be consistent, and the folks who launch these queries must be flexible in the types of methodologies they use, says John Thompson, vice president of marketing at WhiteCross Systems Inc., a maker of analytical software in San Francisco.

Roots of Segmentation

The concept of using automation to conduct segment analysis goes back to the turn of the past century, when the government employed punch cards to tabulate the U.S. census, says Aaron Zornes, an analyst at Meta Group Inc. in Stamford Conn.

When the first commercial computers were introduced in the 1950s, customer segmentation was one of the first applications for which they were used. Pioneers in this area include catalog companies such as Minnetonka, Minn.-based Fingerhut Cos. and others.

As recently as the mid-1990s, companies would typically ship customer data to a third party, such as San Antonio-based Harte-Hanks Inc., for analysis and segmentation. The third party would send the results back to the originating company for use in direct-mail campaigns. But the advent of relatively affordable OLAP and data warehousing tools has made it possible for retailers and other companies to run these analyses themselves.

Now, enterprise managers prefer to keep that data in-house, merge it with accounting and sales figures and run queries from their own desktops, says Zornes. Among the companies providing the applications are Unica Corp. in Lincoln, Mass.; Hyperion Solutions Corp. in Sunnyvale, Calif.; and Chordiant Software Inc. in Cupertino, Calif. This type of analytic software generally starts at $250,000 for an enterprise license.