The challenge of insurance policies in force in insurance analytics

There are a couple of challenges that are faced when trying to bring Insurance In Force data into any analytics platform.  

First is the challenge with Policies In Force (PIF) being a point in time metric.  PIF is a snapshot at a specific point in time of policies that are In force as of that moment.  This specific point in time makes it difficult to perform calculations on as it changes over time and is usually accomplished by running multiple reports and building different spreadsheets.  The creation of this information typically isn’t very dynamic and takes time to create and manage.  

Users also expect that we can compare multiple periods or do time series analysis which can be difficult when you have to regenerate the data for each period in the series.  The challenge is further complicated because we don’t know what that series will be in advance, so we can’t prepare the data in advance.  Our users want to select based on their business questions of the day and not what was determined by the IT Department months or years ago when the solution was built.  With analytics solutions, we want to be able to pick any period and get instant results. 

The second challenge of Policies in Force (PIF) in analytics is merging or integrating in force data with transaction data to allow for a full analysis.  With In force data being a snapshot of the business at a point in time it behaves differently than transactional data.  Where as transactional data builds a story of the business over time.   In force data tells a story at a specific moment but may not be relevant before or after that moment.

The integration of the two is important to allow for the discovery of the whole story in one spot.  When looking at a dashboard the user wants to see metrics that tell how many policies are currently in force but also what premium has been written and earned for the period or year to date.  All three of these are different types of calculations that would traditionally be done in separate reports and would not be dynamic and certainly not be together.  The expectation of analytical solutions is that they are dynamic and linked.  That we can select different periods or dimensions and have results brought back to our screen without having to request a new report to be run.  

So how is this problem solved?  I’ll be honest, this had us scratching our heads for a while.  Over the period of a couple years we experimented with several different solutions but found that they didn’t meet the criteria that we had.  The solution needed to allow the selection of any period (year, quarter or month range) and it needed to handle new periods being added to the data without being rebuilt.  Therefore, pre-building the metrics was out of the question if it was to be dynamic for the user.  The solution also had to handle large data sets and still provide a responsive result to the user.

Then after a period we did solve this problem using Qlik with its associative engine (either QlikView or Qlik Sense). Using the combination of advanced data modelling and visualizations with formulas we are able to build a Policy In Force solution that integrates with transactional data.  Our users are now able to explore their Policies in Force and immediately look at other aspects of their business such as written and earned premiums or incurred claims all in one spot.  This allows them to extract more meaning and do a deeper dive into the business to see the whole story in their data.  With a deeper understanding of their business, they are able to support their decisions with facts instead of intuition.

One of the additional benefits of our method for Policies in Force (PIF) calculations is Business Retention analysis.  With the ability to do PIF for any period, we can do accurate PIF Retention calculations.  This method is better than the traditional way of calculating retention using just transactional data.  With PIF information for a point in time, mixed with rolling period analysis we get an accurate business retention.  Policy retention analysis over periods of time allows users to see trends that may have not been visible in the past using traditional tools and methods.