Results Of The Use Of Predictive Analytics In Life And Health Insurance

Predictive analytics involves the analysis of large data sets ie big data ,to make inferences by identifying meaningful relationships between different variables and using these relationships to forecast what might happen in the future with an acceptable level of reliability.Predictive Analytics includes what-if scenarios and risk assessments.

Now, insurance is an industry where intelligent use of data can provide huge competitive advantages.So over the last dozen years, insurance companies world-wide have tried to be early adopters in using the emerging science of predictive analytics in life and health insurance to get ahead of their competitors.They have attempted to use predictive analytics to segment and underwrite their risks in a more accurate, reliable and cost-effective way.

However,as the Willis Towers Watson’s North American Life Insurance CFO Survey On Big Data and Predictive Analytics shows, even in the more advanced economies like those of North America,only 58% of the respondents know a little bit or understand the basics of predictive analytics and none of the respondents considered themselves experts.So there is plenty of scope for the industry as a whole to make greater use of predictive analytics.

Now, one of the great advantages of being a second mover is learning from the experience of all those who have worked in the area before and not having to make all the heavy investments of resources in areas that might not give the best results.So, following are some valuable insights into the results of the experiments with predictive analysis by health and life insurance companies world-wide.They originate from the decadal study of Clair Nolan and William Trump who spoke about their insights at the at the Swiss Re Centre for Global Dialogue,Zurich, in June 2016. Ms. Claire Nolan is currently Head of L&H Underwriting UK & Ireland ,Middle East & Africa at Swiss Re and Mr. William Trump is Customer Behaviour Consultant at Swiss Swiss Re and Behavioural Science Advisor at iptiQ by Swiss Re.

Insight 1:
Some of the best predictive models in life and health insurance have come through bank assurance.This is because banks have good data sources on their customers.
However,if bank data is not readily available at hand,there are many alternative sources of data that are predictors of health such as such as credit scores, social media and data from wearables. The challenge for companies lies in determining how much of this data can actually be used and how predictive this data actually is.

Insight 2:
An effective predictive underwriting model is built on rich data sources in the form of credit scores, demographics, credit card use , visits to particular department stores etc.
And a good predictive model can provide pre-approved underwriting for the top 40% segment of the population and so they are only required to sign one good health declaration. Thus the underwriting time for this segment falls from 1 hour to just 15 minutes.

Insight 3:
Sometimes less is genuinely more.And in case of the variables considered in predictive analysis in the life and health insurance sector,around 30 data points can form good correlations.
Data points can be used more intelligently to inform the underwriting process. For example: A customer taking out a business loan from a bank, is likely to be in good health. That provides a potential point of the sale for a life insurance policy without requiring extensive underwriting. In the same way, mortgages provide an opportunity for insurance companies to cross sell life and health insurance policies with minimal underwriting.

Insight 4:
Preapproved underwriting means that insurance agents don’t have to spend time asking the sometimes difficult questions of prospective insurance buyers. So their conversations with clients become much easier.So sales teams have a better chance to convert leads.

Insight 5:
Unfortunately, analysis of the data collected during the decadal study suggests that predictive underwriting did not make any material change to demand.