Marketers Turn to Data Mining to Fine-Tune Product Pitches
ACTION ITEMS

ComputerWorld has reported that data mining and database marketing software vendors are forming a union that corporate marketers said should help them tailor promotions to groups of like-minded customers.

Two vendors that make software for managing marketing campaigns last week announced plans to link up with data mining and statistical analysis tools used to build predictive models of customer buying habits.

The promised integration has whet the appetites of users, who hope to get big paybacks by using data mining to match their product pitches to the right customers.

"This could really tie data mining into the business process," said Mike Eichorst, vice president of predictive modeling and data mining at The Chase Manhattan Bank Corp.'s consumer credit unit in New York. "If the marketplace is shifting, you could discover that and react to it quickly."

Chase Manhattan's credit operation runs several data mining applications and is evaluating packaged database marketing software as an alternative to its own custom-built process for managing marketing campaigns, Eichorst said.

Exchange Applications, Inc. in Boston announced a deal to link its ValEx campaign management software with SAS Institute, Inc.'s upcoming Enterprise Miner tool.

The integration, due by March, will let users hone in on selected groups of customers for predictive modeling rather than having to crunch through an entire database, the companies said.

Paragren Technologies, Inc. in Reston, Va., also introduced database marketing software with built-in ties to statistical analysis and data mining tools. Paragren officials said they initially will support Cary, N.C.-based SAS' namesake analysis software and products from SPSS, Inc. in Chicago.

Building and refining customer models is "a very slow and clunky process" for marketing departments, said Jeff Johnson, director of consumer marketing at American Security Group, Inc. in Atlanta. But the marriage of data mining and database marketing should make it possible to update models overnight, he said. That would allow users to "make day-to-day decisions on how you want to set up your marketing prgrams," Johnson said.

British Columbia Telecommunications, Inc.'s telephone unit already relies quite heavily on predictive customer modeling. "But one of the problems you have with a large customer base is getting the models into play," said Bob Boroski, manager of database marketing at the Vancouver-based BCTel phone operation.

BCTel now can only update its predictive models monthly a process Boroski said he hopes to shorten by using a combina-tion of ValEx and Enterprise Miner.