InformationWeek

Quadstone's software collects customer behavior data from multiple sources, such as store retail systems, call centers, and mail-order systems--a capability accomplished through the software's support for a broad range of relational databases. The software applies data-mining capabilities, including regression, decision tree, and optimization algorithms, to detect patterns in customer behavior and make predictions about future behavior, such as what products a customer might be inclined to buy based on past purchases.

CustomerLinx Corp. uses an earlier version of Quadstone's software to analyze data about customer buying behavior to develop lists of sales prospects for its telemarketing and direct-mail marketing clients. "The advantage the Quadstone tool offers is its robustness," says Dick Selmeier, VP and general manager of the San Francisco company. "It's able to analyze hundreds and hundreds of characteristics of the individual leads."

Quadstone can also be used for a range of other analytical applications, including customer retention, risk assessment, cross-selling, and up-selling.

Until now, companies needed a separate Quadstone product, CustomerConversion, to gather and analyze clickstream data detailing the behavior of Web customers. That ability is now built in to the Quadstone System, making it possible to mix customer behavior data from multiple channels for analysis.