Better data for better decisions

While the ability to predict the future is valuable in business in general, it is particularly useful to insurers. Munich Re is well aware of this and aims to offer primary insurers actionable information and insights by providing prognosis expertise as well as state-of-the-art tools to guide and grow their business. What remains crucial is determining how much can be drawn from the data at hand – and how it can best benefit a client’s processes.


“In insurance, you don’t know the costs in advance, as in many other industries. You have to predict them. To do this, you need quality data and predictive models,” explains Massimo Cavadini, Head of Actuarial Consulting and Data Analytics at Munich Re. "The role of actuaries, statisticians and data scientists is to do exactly that: to make use of the data to enhance and improve the prediction of future costs.” Such predictions allow Munich Re’s insurance pricing and underwriting team of experts to help insurers establish pricing that maximises profits by covering the risks without charging more than the customer will bear.

When the proper data-driven tools are put into place, clients can enhance their business’s performance and cut costs along the entire value chain. However, in addition to pricing and underwriting, claims analytics and customer analytics remain crucial to reducing risk and delivering a higher level of transparency to the mix.  

The perils of overlooking key steps in data comprehension

The ability to collect, analyse and draw actionable conclusions from data allows primary insurers to put claims analytics tools in place to control costs and recognise problems. But these tools, like advance fraud detection for example, require a foundation of quality data in order to be optimised. “Many clients want to further enhance, digitise and automate their claims process, and we often have clients who say to us, ‘Now we want to implement fraud detection software,’” says Markus Gützlaff, Senior Motor and Mobility Consultant at Munich Re. “We say, ‘Okay, but let’s first check if your data is sufficient and your claims processes are good enough that such software makes sense.’” Something like fraud detection software is just software, and this software requires data. If there is not sufficient data available, the software simply won’t work properly, and we don’t want to offer primary insurers support in the implementation of something that won’t work for them.”

Many primary insurers don’t anticipate this step, and some initially struggle to see its value. “A lot of people underestimate the importance of quality data,” continues Gützlaff. “Data analysis is not just a matter of dropping in and then analysing the data; we often have to spend time on data cleansing.” This fact frequently requires a conversation to align primary insurers’ goals with their best interests. Since some clients are very solution oriented, as opposed to more service oriented, they will often want to implement a fraud solution right away. Gützlaff adds, “They really want to start with the solution without having discussed all the additional work and requirements. Of course, we are ready to help with this but, as a first step, we have to work on the data so that they’re prepared for fraud software that fits their needs and that really generates meaningful savings.”

The domain knowledge as key differentiator

Munich Re’s data analytics teams can advise primary insurers on changes they can make to boost the quality of available data. Plausibility checks and data cleansing can help improve the quality of existing data, while adding and refining data fields – eliminating free fields and making more fields mandatory – can quickly result in the collection of greater usable data. If there are limitations that prevent primary insurers from employing ideal data collection methods, Munich Re can still help. For most of Munich Re’s primary insurance clients, it’s useful to work with the reinsurer to improve their data quality instead of leaving them to do it by themselves. Munich Re has the tools and expertise already in place as well as the ability to efficiently access external data for different markets. Tapping into these pools, especially for small- and medium-sized primary insurers, would normally require huge overhead.

Munich Re not only has the breadth and depth of experience to help primary insurers optimise their available data; they also have the expertise and domain knowledge to truly understand that data and glean meaningful insights from it. “Without data, you can’t do any prediction by definition, and the more data you have, the better,” says Cavadini. “But the main difference is the domain knowledge: the way you treat the data and the way features are engineered on top of the data. The predictions you run are only part of it; the soul of our data analytics is our domain expertise. With this, we are able to use prediction to transform your business.” Domain knowledge sets Munich Re apart from its competitors. "The prediction coming from algorithms is becoming a commodity. The way you transform this prediction in a business decision is the real key differentiator.”