Taking to
the Sky

Insurers have recently been driven to unusually fast innovation by factors both on the financial and technical front. On the financial side, low interest rates have eroded investment income and raised the present value of future liabilities. These changes have posed a major threat to profitability and have made innovation necessary.

On the technical side, insurance companies are driven by data. Lots of new information, such as weather and geodata, has been made publicly available, digitally accessible and therefore suitable for enriching insurers’ own data assets. That availability combined with the rise of advanced data technology, have made digital innovation possible.

In that context, we joined forces with a global reinsurer and other external experts to develop a new line of business: Setting up a new analytics offering, delivered as Software as a Service, which is globally accessible to any primary insurer thereby opening up new markets, customers and revenue streams to our client.

Overtaking Fraudsters

To grasp newly arising potential for innovation, consider a reinsurer who can automatically combine satellite images of a disaster area with locations of its insured buildings to predict possible damages. In the past, the company faced the problem of having to assess damages long after they had occurred and often already been cleared in reconstruction.

Now, receiving timely alerts, the reinsurer can proactively assess damages. That gives it a leg up in its efforts against insurance fraud.

However, to exploit the potential of such innovation, insurance companies face one technological obstacle: They require state-of-the art systems that coherently combine, structure and analyse huge amounts of complex, partly unstructured data. Here, we come into play.

Big is Bountiful

As technical basis for the reinsurer’s new analytics service, we first set up a big data platform that combines different cutting-edge technologies. It is built on a Hadoop platform. That software framework combines several servers that can be scaled to allow for distributed storage and processing across clusters of computers.

Then we strove to improve data processing on the platform and to enable external parties to better access certain sets of this information. To achieve these goals, we are migrating the data from the reinsurer’s internal system into the cloud.

“For me, the project is exciting because big data and the cloud are developing so fast that we’re permanently exploring new approaches together with our client. So we’re all on a steep learning curve at technology’s cutting edge.”

Jan Thorben Barenthien, Senior Data Engineer

For our client, the cloud structure provides three advantages that make its new analytics service possible: First, it can attach different levels of confidentiality to specific data sets. Based on that segregation each external party gets more flexibly tailored and faster access to exactly the information it needs.

Secondly, the cloud allows to efficiently distribute the workload: Information is stored on data servers and processed on application servers. Such specialisation improves the performance of the overall system. Plus, systems can swiftly be scaled without hardware limits by adding extra resources for storage or processing if the need arises.

The most important advantage is that, the cloud simplifies offering Data Analytics as a Service. Specifically, the reinsurer first integrates data of primary insurers; then it validates, cleanses and enhances the data by supplementing it with its own information; finally, it develops the analytics tools to analyse it. These tools allow primary insurers to assess their data and to visualise the results on a bespoke dashboard.

The range of analytics thus available to the primary insurer include evaluations of various parameters: customer lifecycle value, which describes the future value of a specific customer; so-called churn, which describes the probability that a certain customer will change the insurance provider; and cross-selling models, which show the probability that a customer will take out other policies with the same reinsurance provider.

One Platform for the World

Besides the tricky infrastructural groundwork, there are two further elements that have made this project a fascinating challenge. First, it relies on sometimes unprecedented combinations of cutting-edge technologies that all evolve dynamically. Experimenting in such fast-moving environment requires flexibility, mastery of the technologies and extra efforts. But it pays off as it enables our client to stay ahead in a competitive market.

Plus, this system not only has to continuously advance while running smoothly throughout all these updates and migrations. It also needs to be made accessible from all the reinsurer’s offices across the globe and for all the desired data science and big data applications by diverse users. Only such flexibility enables our client to globally provide its analytics service in different markets and to all kinds of insurers.

In all these challenges we have successfully relied on our experience in developing and running complex applications. By raising the capacity, flexibility and efficiency of our client’s data storage and processing we have enabled it to develop a new offering, provided as Analytics as a Service. It is at once sophisticated, innovative and user-centric, giving our client an edge over its competitors.