Business Challenge

This telecom provider wanted to know the potential value of their customers to design strategy around available headroom. While they knew current revenue generated from a customer, they wanted to calculate additional revenues that can be generated from the same customer using wallet size, in order to cross-sell and up-sell value added services optimally.


We deployed a strategic segmentation & predictive modeling solution:

  • Segmentation analysis was developed using Call Data Records and customer demographics data
  • Customers were grouped into Low, Medium & High revenues categories basis their current value
  • Multinomial logistic regression model was developed for customers in each bucket and probabilities to migrate from a lower bucket to higher bucket was calculated
  • Scoring algorithm was deployed for periodic scoring of customers


  • More Targeted Campaigns: Potential value of a customer based on revenues helped the client make design more targeted campaigns.
  • 8% Reduction in Targeting Cost: This was achieved by targeting the right VAS offers at the right customers, based on wallet size.
  • Focus on Profitable Customers: Our scoring algorithm helped score customers frequently and understand potential value of new customers at inception, allowing shifting of marketing focus to more profitable customers.