• A specialty multi-channel apparel retailer was using both brick-and-mortar and eCommerce to sell to its customer base. The client wanted to understand the “characteristics” of best customers both from behavior as well as from demographics perspectives.


  • Extracted and cleansed the customer behavior data across all channels and appended demographic variables from Experian
  • Developed multi-variate regression model to identify drivers of sales and validated model using CHAID
  • Used “Significant Drivers” from regression and CHAID to segment customer base using cluster analysis
  • Recommend segment specific strategies to increase sales


  • Best Customers are multi-channel buyers (Catalog and in-store) from New England or the Northwest that lived more than 20 miles from the nearest store
  • Highest value segment represents a 523% gain over the average customer
  • Migrated 6% of the audience from average customer segment to higher value segments leading to 11% increase in overall sales