Prospective Customer & Response Using Classification Modelling

October 27, 2020

A Leading Retail organization

Problem Statement

• Find look-a-likes of the current successful customer and sell aggressively to qualified leads
• Identifying responders to marketing campaigns and extend offers to responders

Solution Approach

• Multisource input data

• Lasso Regression &  Multivariate Adaptive Regression

• R Studio & R Shiny

Business Impact

• 20% improvement in response with respect to existing campaigns

• Out sample prediction accuracy at 81% along with customer classification model.

Critical Success Factors

• Value & Risk Based classification
• Geographical distribution of projects