Know Thy CustomerKnow Thy Customer

SPSS Inc. this week will unveil predictive-analysis software that a broad range of marketers can use to anticipate customer behavior. Most predictive-analysis applications are used by business analysts with expertise in data mining and statistics. SPSS PredictiveMarketing hides that complexity using pre-built models that help marketers answer questions such as which customers are most likely to respond to a sales offer.

The software can be used to identify customers who are likely to abandon a business, are more likely to respond to a new offer, or have the potential to generate greater profits for a company. SPSS's Clementine software can be used to build additional predictive models for the tool.

BankFinancial uses SPSS PredictiveMarketing to identify customers who may be about to leave. The plan is to provide branch managers with a "red list" of such customers so they can try to keep them, says William Connerty, assistant VP of research. Every month about 1,200 customers respond to promotional offers, become BankFinancial customers, then leave. Connerty would like to cut that number by at least a third.

In the future, the bank will use the software to identify customers who are likely to use more bank services and generate more profits. "It automates a lot of the data manipulation and data flow," Connerty says, adding that the pre-built models and easy-to-use interface eliminate the three to four days needed to develop predictive models. "It really increases our staff efficiency. It's like data mining for dummies."

SPSS PredictiveMarketing will ship later this month. Pricing varies according to use but starts around $80,000 and averages $200,000. SPSS is also developing a version of the software that works with Siebel Systems Inc.'s marketing apps.