AQUA Intelligence - The science behind the business  

The insurance sector ischaracterized by sparse customer interaction. Churn is therefore difficult topredict and cancellation forecasting has been traditionally based on the presenceof highly obvious hints, such as customer statements of dissatisfactionresulting from negative claim resolutions or any other types of complaints.

In this uphillcontext, the challenge was to develop a system allowing for significantimprovement in predictive power of customer churn. The use of historicalcustomer data became necessary for this purpose.

The innovative solution contributed by Neo Metrics was based on a particular fact from the business domain: A policy may reveal relationships among several individuals. Thus, the subscriber need not coincide with the insurance user, who can in turn undertake another policy contracts with other beneficiaries.

This simple scheme allowed defining relations among several customers (sharing a contract or a home address, or prescribing for other individuals, etc.), enabling identification of communities also known as influence spheres of each customer.

The end result was to double the predictive power of traditional solutions.