Churn Prediction and Customer Cohort -
Date
October 2023
Service
Employer
Paytm Money
Project Overview
This project pinpointed those users on the Paytm Money platform that were active the previous month but were predicted to churn in the upcoming month. Insights helped marketing and growth teams proactively target these users and implement retention strategies.
Key Highlights
To identify the users who traded in the platform the previous month but are likely to churn for the current month to assist the marketing and growth team targets the users and plans.
Used Python to implement clustering techniques like KNN and DB-SCAN to create cohorts of users that helped to segment the user based on trading habits.
ML methods like Random Forests, Gradient Boosting, and Deep Neural Networks were used to predict and identify the churned users with a Recall of 88% and Precision of 70%.
