Churn Prediction
InsuranceBusiness Impacts
40%
Improvement in model accuracy
2X
Uplift achieved from existing baseline Risk Score Model
Customer Key Facts
- Location : North America
- Industry : Insurance
Problem Context
The customer is an American insurance company that provides financial protection to more than 50 million people worldwide. Their financial planning and analysis teams create projections on customer churn at a quarterly level albeit without any statistical understanding. They wanted to maximize their ability to predict groups with a high probability of churn and identify the key reasons that influence the probability of churn.
Challenges
- Identifying the right set of control groups
- Sparse and missing attributes within the data
- Different churn models had to be created for different demographic levels and quarters
Technologies Used
Google Cloud Compute Engine
Google Cloud Storage
Google's BigQuery
Tableau
Developing a One-Stop Platform to Identify Customer Groups Likely to Churn
Solution
Quantiphi developed a churn prediction model and Tableau dashboard that enabled the business to better mitigate churn by developing a sound understanding of the factors responsible.
Result
- Accurate prediction led to better targeting of customer groups
- Dashboard was leveraged to democratize model results among both analyst and business user groups to highlight key factors
