Production

Customer Churn Prediction & Segmentation

Biznesni Rivojlantirish Banki

⚡Customer attrition was detected retrospectively - by the time it showed in the numbers, the relationship was already lost.

⚠️Not all customers are worth the same retention spend; there was no way to prioritise who to reach and how.

⚙️Built classification models predicting churn probability from transaction frequency, balance trends, product usage, and engagement decay.

🛡️Added segmentation of the customer base to tailor retention strategy per group and focus effort where predicted churn and customer value are both high.

🚀Gave retention teams an early-warning, prioritised list - turning churn handling from reactive to proactive.

↳ Predicts which customers are likely to leave and segments the base for targeted retention - proactive instead of reactive

Churn ModelingCustomer SegmentationClassificationScikit-learnFeature EngineeringPython