Telecommunication Company
Analyze mobile data usage patterns to advise customers on how to lower their subscription costs. Segment customers, find those who are about to churn and the potential reasons why, and also estimate the time to churn to see if intervention is possible.
Challenge
A mobile virtual network operator in North America wanted to understand their customers better: how different apps use foreground and background data, what typical usage patterns look like, and how data usage can be optimized. A large group of volunteers had app-level usage data collected over months, providing a unique and comprehensive dataset.
In a subsequent project, the focus shifted to understanding why and when customers churn. This requires explicitly handling censoring in datasets — in the long term, all customers eventually churn; you just haven’t observed it yet. The business also needed to know how much time remains before predicted cancellation, to try to influence the decision.
Solution
We analyzed per-device and per-app usage data totaling hundreds of millions of data points, using geospatial visualization, customer segmentation, churn modeling with interpretability features, and survival analysis to address retention challenges.
Results
The analysis provided actionable insights into customer usage patterns, enabled targeted retention strategies through churn prediction, and delivered survival models accounting for right-censored data to estimate intervention windows.