Improving Customer Conversion


A leading apparel retailer wanted to improve personalized marketing to drive additional transactions. They had a high (80%) one-time purchase rate. It was costing them 5X more to acquire new customers compared to retain existing customers.


Our approach to personalization started with building two Machine Learning models leveraging all of the retailers digital customer data:


1.Purchase Propensity:
Likelihood of a customer purchasing in the next two weeks

2.Category Affinity:
Product category the customer will most likely purchase next

Models generated predictive information used for personalization across channels


•Improvement of test over control groups in nearly every KPI measured

•Over the course of a year, the conversion lift represents an incremental 10% lift in sales in this channel alone, applied to all customers