

InfiniteInsight® puts us in the driver's seat. ![]()
John Gulliford
Manager, Analytics and Circulation, Online BU
Sears, Roebuck and Company Boosts Productivity with InfiniteInsight®
Challenges
- Like many retailers, Sears realized it needed to find a way to compete in the volatile world of modern retail fueled by technology and flooded with competition.
- Sears modeling productivity was crippled by dwindling analytical resources.
- Deploying predictive analytic scores from the Teradata data warehouse into production required manual, redundant and time-consuming processes that were prone to human error.
Solutions
- Standardized on InfiniteInsight® to optimize marketing processes including customer segmentation and targeted mailings.
- Automated the predictive analytics process including data preparation, modeling and score deployment.
- Built a 900-attribute data mart within the Teradata warehouse that combined information which spanned multiple channels, hundreds of brands combined with a wealth of customer data including credit history, market demographics, and external sources such as mail-order overlays.
- InfiniteInsight® allows Sears to apply scores directly in-database to their Teradata data warehouse.
Results
- Eliminated labor-intensive, high-maintenance and repetitive tasks.
- Reduced the time to build accurate, robust models to a few hours rather than a few weeks.
- Using a central Teradata data warehouse gives the catalog team a highly scalable computing and storage environment while reducing operational costs.
- Modeling and scoring directly within their Teradata warehouse results in hundreds of thousands of dollars in annual savings.
- Able to score 75 million customers in 30 minutes, down from several hours which results in significant savings.
- Realized a welcome and unexpected benefit: business partners, analysts, and merchandising colleagues are asking more “what if” questions because answers are immediate – often within 10 minutes.
- Automation allows analysts to build more sophisticated models than they could with conventional predictive analytics solutions.

