Churn Prediction Model for an American Insurance Company
InsuranceBusiness Impacts
Preventing loss of customer groups
Acceleration of business growth with intuitive recommendations
2.7x improvement over the existing model
Customer Key Facts
- Country : United States
- Industry : Insurance
Problem Context
The customer, an American insurance company and the largest provider of supplemental insurance in the US, wanted to identify the customer groups that have a higher chance of churning and better understand the reasons behind it.
Challenges
- Low volume of data
- Highly dependent on data for analysis
- Imbalanced dataset
Technologies Used
Rstudio
Apache Zeppelin
Cloud Spanner
Cloud Compute Engine
Dmlc XGBoost
Identifying customer groups that are likely to churn and the potential reasons behind it
Solution
Quantiphi created a model that outperformed the client’s existing model to identify customer groups that are at a high risk of churning out.
The model suggested reasons for the churn.
Results
- High retention of customer groups with creation of action list. This helped ascertain the reasons behind increasing rates of churn.
- Boost in business growth through specially curated instinctive recommendations.
- 2.7x improvement over the client's existing model.
