Predictinginsuranceindustrychurn-useR!2014.pdfVIP

Predicting insurance industry churn Pramod Kunju1 1. Founder and CEO, Dataversal Inc *Contact author: pramod.kunju@ Keywords: Predictive analytics, Customer retention, Customer churn, Insurance churn, Big data Customer retention is a critical success factor for insurance industry, much more so than for other industries. Losing one customer can result in lost revenue for several years. For example, an auto policy might last for 10 years, and a home policy for 30 years. An existing customer canceling a policy after 1 year of inception can result in lost revenues of 9 years and 29 years for auto and home policies, for example. Customer retention in insurance industry is very complicated. For example, let’s assume that a customer is identified as a churn prospect, and a call is made to her to assess how happy she is with her insurance. If she is really a churn prospect, the insurance company can try to address any concern she has, and proactively retain her. However, if she is a perfectly happy customer, that call can potentially trigger a churn as well, since calls from customer service agents are not something everyone looks forward to. So, it is crucial that any customer churn prediction is accurate. The presentation will cover real-life examples of how R is instrumental in accurately modeling customer churn for insurance companies. The typical output of such a model-based analytics solution is a list generation. The list will contain existing customers who are likely to churn, and the reasons why they are churn candidates. The input data to R will come from internal transactional systems, and external sources such as social media. In fact, an increasing percentage of the input data is from a big data platform. The presentation will lay out a business approach, and an architectural pa