About the Client

A telecom provider in Europe, with approximately 10 million customer base

Business Need

  • Identify the factors responsible for the high customer churn rate; design and implement appropriate steps to reduce the churn
  • Reduce customer campaign cost.
  • Launch focused campaigns based on propensity of a subscriber to churn

Approach and Methodology

  • Various data points such as demographics, contact information, product and services availed, offer/plan detail, months in service, bill amount, payment history, inbound and outbound call details, handsets used and service features were analysed
  • Data preparation through exploratory data analysis, missing value and outlier treatment. Checked correlation and muticollinearity relationships for different parameters
  • Developed a model using various techniques such as logistic regression, decision tree, concordance, etc.
  • Tested and validated the model using a test sample; checked key statistics and performed out-of-time validation
  • Implemented the model to identify customers with high propensity to churn
  • Reached out proactively with better offers to a targeted subscriber segment – only the high-risk customers

Business Outcomes

  • 20% reduction in customer churn rate
  • 35% decrease in customer campaign cost

—Related content

Ready to talk?