Blogs
Running a Tight Ship Through Data and Analytics
Last week, I attended the virtual roundtable All Aboard for Data Analytics, as part of the Marine Insurance Nordics online conference. Moderated by our own Thomas Anderson, participants, Lars Dueled, Chief Underwriting Officer, Skuld; Sigvald Fossum, Vice President, Head of Analytics, Gard AS; Ranjith Kanipayur, Head of Business Transformation, The Shipowners’ Club and Callum Shaw, Pricing Lead, RSA, explored how data and algorithmic underwriting are impacting the marine insurance market. They also discussed if current platforms and systems are fit for purpose and agile enough to enable marine insurers to improve risk profiling and add new lines of business and products.
With the challenges around securing an increase in P&I premiums, reducing exposure in certain areas, and static market conditions, marine insurers seek to drive premium growth and increase market share through portfolio diversification. Many marine insurers operating on legacy platforms may miss out on leveraging data assets and using analytics to improve underwriting decisions and risk selection. Data-driven decision-making is the new frontier for insurers.
In this roundtable, participants debated algorithmic-based underwriting’s place in the marine industry and if it can help marine insurers drive underwriting profitability and efficiencies through using multiple sources of data to maximise data insights.
Deep-sea data diving
The session kicked off with the obvious question – how coronavirus has affected marine insurers and their ability to shift from a technology perspective. A panellist started the discussion by explaining the effect on pricing since the outbreak and that they now rely on analytics daily, which they had not done before covid.
One panellist said that as covid has upended the world economy and the marine industry, his company has to deal with an influx of new data and business challenges without the benefit of having historical data or viewpoints. He explained that his organisation has not only tightened the clauses around covid but is now making sure that they are robust in the business they write.
On the upside, several panellists agreed that trading patterns are much more available compared to the past because of data and analytics, with one participant saying his company is “having a much better grip on the risk profile through data and analytics, and during the last 18 months, data and analytics have really helped us manoeuvre through these times.” One panellist said that “our customers are screaming for data”, which has resulted in his organisation increasingly sharing data with clients to help prevent incidents.
The sharing conundrum
As the conversation turned to data sharing within the industry, the overwhelming sentiment was that marine insurers must do more to share data amongst themselves – but how much data should be shared and what type of data to be shared elicited different responses. One panellist said that the industry must “sit down and figure out how do we use this massive amount of data from an environmental and crew point of view… and that we need to join forces to meet social obligations that are put in our hands.” The same spokesperson said that “we have the possibility of doing it, but in order to have the maximum gain of the data in our industry, we need to join forces.” Others agreed, saying that competitive data should not prevent insurers from delivering for the common of the marine industry, but that competitive data should not be shared.
One spokesperson was a little wary of data sharing, saying that “sharing data for the sake of sharing data is inherently difficult” and that the first step was to figure out the given problem and understand if sharing data is the solution. The group discussed the potential unfairness of one company sharing 90% of data, one sharing 10%, and another sharing nothing at all. All participants agreed that sharing data should not risk a company losing its IP and that sharing parameters need to be set up. One spokesperson pointed out that the motor insurance industry has nailed sharing pricing data and that lessons can be learned from evaluating how they have approached the issue so everyone can compete fairly.
One final point around data sharing was that pricing standards are non-existent in the marine insurance industry and that each marine insurers categorise claims differently. All agreed that it is essential for the industry to establish standards to facilitate data sharing.
Rocking the boat
The emergence of new players with no legacy systems entering the market was discussed, and the risks and benefits they bring to the market. One participant said that “we shouldn’t underestimate these new players… they will challenge us, and that’s a good thing.” Another participant stated that these new companies would find it much easier to hire talent, especially data scientists who do not want to work with legacy systems or companies that are “stuck in the past.”
However, one participant pointed out that established marine insurers have the advantage of history behind them, and new entrants lacking domain experience will struggle. Another speaker added that building up expertise takes decades and the complexities of the industry make it complicated to penetrate.
Shipshape for the future
In closing, participants were asked for their views on the future and how marine insurers can embrace innovative technologies such as artificial intelligence, machine learning, and IoT to drive forward profitable underwriting. One participant said that to make a change, “we need to look at the loss side… if we can limit the number of medium to small-sized losses by getting all that data information to our clients, that is our main obligation going forward.”
One participant stressed that the industry must work collaboratively, ensuring that it makes use of real-time analytics, “we need to monitor the business and feedback to those companies (clients), and tell them what they are doing differently and any risks involved so they can change their behaviours.”
Overall, the group expressed optimism for the future and said they expect to see huge changes over the next three years. However, success will largely depend on how marine insurers use data and analytics, especially as many of them run on legacy systems. As one participant said, “we have the upper hand of the data but also the lower hand as we have legacy systems; I guess the problem I find is that the systems are not adaptable, we have so much data which is brilliant but how do we integrate that data with our systems… some systems are clunky and there is a real debate going on whether we need to bin our systems or whether we need a new system to adapt.”