kentoh - Fotolia

Munich Re puts data analytics into services for a ‘riskier world’

Re-insurer combines a Hadoop data lake, SAS analytics and SAP Hana to create new cyber and other insurance services

We live in a risky world, getting riskier all the time, and Marcus Winter, head of re-insurance development at Munich Re, sees a world to be insured and re-insured as it becomes ever more complex and digital.

The Munich-based re-insurance firm has a global reach, and is seeking to develop newer services, including in cyber insurance, that draw on data analytics – services that would not have been possible even five years ago, he says. These services are built on a data management infrastructure of a Hadoop data lake, based on a Hortonworks distribution of the open source data store, SAS analytics, and SAP Hana for structured data.

Winter has a global strategic role, reporting to Munich Re’s CEO. The data analytics side of what he does is delivered by a 100-strong data science and business development team, which he leads. They work up services of a new kind for insurance companies and large corporations, says Winter. “We insure insurers, but also provide risk management to large corporates, for example on the cyber side.”

Winter describes cyber insurance as a $2.5bn market, encompassing indemnification of losses from malware and hacking. An example would be if a bank has to replace customer credit cards because of data loss.

Munich Re, which had revenue of €48.9bn in 2016, has 7-8% of the cyber market and anticipates growth in that market as the EU’s General Data Protection Regulation (GDPR) approaches next May.

“Cyber insurance is different from other forms of risk transfer insurance, for example flood risk to housing,” says Winter. “Those change very slowly over time, and so past experience is a good guide for prediction. That is not so in cyber – there is a greater reliance on technology due to the real-time dimension.”

This has entailed building a data management architecture with a data lake, the Hana database technology, and SAS analytics.

“That is also the reason for being cloud-based,” he says. “To deliver modern risk services, you need a big toolbox, so you need pre-defined interfaces in cloud [environments] where you have standardised software, so you can deliver services in a modular and agile way – and also update those modules centrally. That is a big change. It’s not that the cloud technology triggers these changes, it is more that there is a need that can be addressed by these technologies.”

Read more about data analytics in insurance

As well as cyber, Munich Re is adding more “value-added services”, says Winter. These include risk assessments for natural catastrophes or explosions that can affect global supply chains and need to be pushed out to clients quickly.

He gives the example of a manufacturer that is shipping high-value components to customers, or to their factories, that are using just-in-time production. Managing the risks in global supply chains requires more alacrity as the world becomes ever more risky, says Winter.

“Our co-operation partners there, including SAP, are important in terms of delivering those services,” he says. “This is a new area, at least in terms of doing it in a structured way. Computational power has improved so much that we can deliver services we could not have done even five years ago. And the cloud helps with that since we can access more power than we could before on-premise.”

Munich Re has “a complex IT landscape including in-house developed IT”, says Winter, adding: “SAP is strategic for us. There is close cooperation at all levels with them.”

Most of Winter’s staff are in Munich, but the firm also has data analytics workers in Beijing and New York, as well as “scouts” in Tel Aviv and Silicon Valley. “When you do cyber security, you have to include the ecosystem in Israel, just as when you do fintech, you need an ecosystem in London,” he says.

Winter spends half of his time on the data analytics programme and the rest on strategy and innovation management. “If you add it up, it comes to more than 100%,” he says.

Read more on Artificial intelligence, automation and robotics