Using data modeling and predictive analytics, we worked with a large life insurance company to help them understand the lapse behaviour of their protection business. This work enabled the company to improve their cashflow projections for its existing book, identify customers at risk of lapse, and target high-quality customers in its marketing campaigns.
Situation
A life insurer had a higher-than-expected lapse rate and wanted to understand the key drivers of lapse are in their protection book.
Solution
By applying survival analysis techniques, and socio-economic segmentation, we predicted expected lapse probabilities for each policy, allowing us to discover the effect of policy characteristics on the lapse rate, and how those effects change over time.
Lapse Curves by Policy Type
Business Outcomes
Visual exploration of policy data
Thorough data validity and integrity checks
Identification of key drivers of the policy lapse rate
Per-policy prediction of expected lapse rate
Improved quality of business through targetted marketing for new business
Model Outputs
Key Lapse Drivers
Gender of the primary policyholder
Smoking status of the primary policyholder
Coverage type of the policy
Socio-economic category based on address of the primary policyholder
Age of policyholder at inception
Source of business (broker, IFA, etc)
Implications
Key drivers of lapse changed over lifetime of the policy - even reversing for some
The effects of the credit crisis affected policy lapse from 2008 to 2012
Policies with a large sum assured (over 1 million euro) needed separate treatment
Key Benefits for Client
Data quality and integrity rigorously tested
Improved cash-flow projections for the in-force business
Improved quality of independent broker business
Identification of profitable policy cohorts for future marketing
