Early Claims Forecasting - I in Anatomy of Decisions

As you learned in previous segments that any new policy written by an insurer is a drain on the books for the initial period of time. The insurance companies only start making money when a customer has been with them for a long period of time. In the previous segment, we learnt how analytics helps insurance companies in reaching out to customers and make sure they renew the policy.

However, there is one case where none of the models will apply. You might be wondering what is it. The answer is simple - death. The insurance policy will have no one to renew if unfortunately, the customer passes away. Apart from the loss of a customer, this will lead to a huge payout for the organisation and hence, every insurance company would like to onboard people less risky to make a claim early on. Analytics is strongly used in insurance companies to help predict risky customers so as to charge them a higher amount for the insurance product.

In the following video, Ashish will walk you through the process of identify risky customers and how analytics can be used to onboard safe customers.

As you learnt, every insurance company would prefer to onboard a non-risky profile as far as possible. You also learnt about the series of fundamental hypotheses which are verified using analytics.

• Certain pin codes are risky where the mortality rates are very high.

• Certain professions are risky and should be avoided or at least every policy with the customer having this profession should be investigated.

• Date of birth as 01/01 is a red flag.

You can see that all of these hypotheses are logical and intuitive. Furthermore, you learnt that the data used to solve this problem is the claims data which is readily available with any insurance provider.

In the next segment, you will learn the details of the analytics process as to how the dataset is used to decide whether a potential customer is non-risky and should be onboarded.