MarketSoft1

Problem

A leading decade-old life insurance firm had a persistent problem: several policy holders discontinued paying their life insurance premiums after the first year. There was no good way to identify them ahead of the event. Initial simple analysis could not identify any correlation between im-persistency and a segment of policy holders (demographics), agents, or geography.

Solution

Marketsof1 engaged for Persistency Modelling and maintenance, used advance analytics - a supervised machine learning program - to identify every month who will not persist. It helped the firm, achieve several of the objectives and importantly, reduce the im-persistency significantly. The insurance company has reached the top industry decile in persistency from the being in the 5th decile when the modelling commenced.

Benefit

Ability to identify so had twin benefits. One could boost current profitability. One could treat the ‘potential to become im-persistent’ customers with different marketing programs: for example, to motivate them to create a standing instruction with a bank or switch to a different product. Or communicate at different intensities, tone and content.

The other benefit is to boost future profits by acquiring the right type of customers and pre-empting customers who are likely to be im-persistent. It could be achieved by deploying a lead scoring engine basis learning from the Persistency Model. Lastly, one could devise a different product for low persistency customers.