Predicting Customer Churn for large SaaS company

About our customer

The client is a reputed unicorn SaaS company based out of North America.


Insurance, SaaS, HR

Use Case

Marketing, Customer Churn

Implemented a data analytics and lead enrichment platform for activating dormant customers for the biggest private bank in India.

The Challenge

Clients are from various geographical zones, of varied sizes and includes various industries. This requires grouping of customers sharing common traits.

Also due to the way the data warehouse had been set up by the client, all historical data was not available for all clients.

The Approach

We approached the problem by doing robust data analysis on the assumed churn hypothesis. Continuous review feedback from business made the dedication valid churn prediction features. Reverse cohorts features were made on cleaned data for model building.

The final model comprised of four minor independent model for each data availability case. This resolved the problem of mission data in AWS warehouse giving very satisfactory results.

Further the scores from each model were standardised based on a common scale to form uniformity in risk score. Further churn profile was developed to gain feature insights highlighting reasons for Client Churn.

The Outcome

The client was able to take prevent considerable churn losses. We are now deploying this solution in their HR product.

Capabilities Enabled

Implemented a Client churn analytics solution which helps identify the riskiest clients to enable business make necessary early action to prevent churn.

Impact Created

Identified 61% future churned client with very high precision.