End-to-end data analytics

Data science can significantly improve the management of in-force customers.  In the process it can also add significant value for the shareholder.

But many insurers (and banks) have struggled with data analytics. Progress has been slower than data experts hired from other industries would have predicted. Here are two of the common pitfalls:

  • The ‘we need to organise our data first’ pitfall. An insurer starts a large multi-year data warehousing project to bring together datasources, categorise and clean the data before we can start the analytics. This approach misses the fact that we rarely need perfect or deep data to get started with data analytics. There is a real risk that the data warehousing project is cut back or shut down before the data is delivered and so analytics makes no progress at all.
  • The ‘analytics delivers on paper only’ pitfall. An insurer runs a project with expert help to identify, describe and quantify the analytics opportunities in its business. Whilst the work is often very good and the final report is compelling, too often these analytics reports remain on the shelf without actions being delivered in the business.

We find this maxim helpful:

Data analytics will have zero impact unless our customer-facing staff act differently as a result of the insights

In in-force management, this means that we need to get the insights into the hands of the staff managing the in-force customers. We need to do this in a way that makes it easy for them to have better conversations with customers every day.

We need to start with simple end-to-end data analytics…

End-to-end data analytics

Following the above diagram, we should build all of the pieces in parallel so that we have an end-to-end solution to a business problem. We will use just enough data to get started. We create simple but useful analytics. And we create front-end user applications to get the right insights and suggested actions to the customer-facing staff at the right time. Until all three are in place, no value will emerge.

We recommend you start with an end-to-end analytics pilot with a small group of customer-facing staff. Start with the simpler business problems that make incremental improvements to the way staff work. Once the analytics is proven end-to-end for the simple, but nevertheless valuable business problems, we can build more sophisticated analytics and use deeper data sets over time.

The largest gap today at many insurers is the front-end. The front-end tools for customer-facing staff need to be intuitive, attractive and fun to use. Staff want to use tools that improve their work day and help them have better conversations with customers. We work on the philosophy that the user front-end design should be shaped by the front-line, for the front-line.


You can get in contact with Matt here.

Click here to download diagram.