A retailer, with over 4,000 stores and $40 billion in annual revenues, was looking to engage their customers in mature upsell and cross-sell activities through its loyalty program. The problem — unidentified customers accounted for 40 percent of in-store transactions which amounted to approximately $35 million on a daily basis.
Our team members implemented a solution that delivered loyalty program sign-up messages and bounce back offers at check out, enabling the retailer to test various messages and optimize their messages through adaptive models based on customer responses.
By using real-time market basket analytics they were able to provide customers with recommendations based on product affinity models. The retailer was able to deliver three category level offers from a variable catalog of thousands of offers set to expire on the same day.
The solution used centralized IBM decision management and handled thousands of model executions across all stores. Products used in the solution included IBM Interact, SPSS, IBM DB2, Oracle Exadata and point of sale systems.
After implementing the solution, the retailer experienced $6 million in incremental margin within the first three weeks, inclusive of the day after Thanksgiving where 800,000 offers were delivered. The average response rate, which was approximately five percent at launch, increased to approximately 20 percent during that same time frame. The average basket size increased from $80 before the offer to $125 after the offer—approximately a 56 percent increase.