Sponsored by: ?

This article was paid for by a contributing third party.

Spotlight: Data - Ordering data in

Data

Insurers continue to hang their success on the quality and use of data, with businesses continually reviewing how to gain a commercial advantage. Edward Murray looks at the benefits of using external data providers and why some peril risks have proven more popular in terms of being profiled at a more granular level than others

Property insurers are desperate to spin straw into gold and use the mountains of data available to generate valuable commercial insights that sharpen their pricing, deliver better loss ratios and improve customer experience.

But the abundance and availability of data makes using it a bit like trying to drink from a firehose. To make it manageable, insurers must identify the individual data sets that will improve their risk understanding and which they can absorb effectively into their processes.

“We formally review the external data market on a regular basis to understand the latest capabilities and products available,” says Chris Wyard, head of underwriting data at Allianz.

He adds: “We engage with vendors to explain our motivations and look to thoroughly test the performance of their products to ensure that we licence the best data in the market. Ultimately, we will make purchase decisions based on data coverage, granularity, quality, scalability and benefits realisation.”

This is a tried and tested approach and one endorsed by Plum Underwriting. Toby Daley, underwriting director at the managing general agent cites, “accuracy and granularity,” as key factors in buying external data and says it must be possible to host the data and establish a positive balance between its cost, usability, and creation of added value.

Future outlook

Data aggregation, data analysis and greater data sharing will be the key to future success in the commercial vehicle and motor fleet markets for both insurers and manufacturers, says Roger Ball, head of motor at QBE.

Insurers are all looking to use data differently, depending on their core underwriting appetite and commercial objectives. Here’s how some of them see the future panning out.

Tom O’Connor, CEO of CLS Risk Solutions: “Our underwriters of the future will need to be very tech-savvy, creative thinkers who work well in teams, collaborate with clients and who can adopt continual learning behaviours. So, as we move along that journey of fusing technology and underwriting nous, we know that if we can exploit the technology that we already have available to us, our underwriters will begin to perform better and better. We know that for complex risks we will always need human judgement: but we can also provide powerful technology to refine and contextualise the judgments that they need to make.”

Annarita Roscino, head of predictive analytics at Zurich: “We believe that the external data solutions will continue to expand with more external data offerings and perhaps more offerings which are tailored to the insurance industry. It is going to be really imperative to have robust internal processes to identify high-quality data at a reasonable cost, to combine this data with internal sources and to seamlessly implement those new sources to gain a competitive advantage in the market.”

Chris Wyard, head of underwriting data at Allianz: “There will be further innovation and challenge to the traditional market from insurtech players and I expect the market to move beyond arbitrary use of postcode and address level products (which have some inherent inaccuracies) to deliver richer data that accounts for individual building footprints.”

Toby Daley, underwriting director, Plum Underwriting: “With greater reliance on data by mainstream insurers increasing, there is an opportunity for niche providers that are able to assist customers who fall within areas deemed undesirable by those investing heavily in external data use. Increased ‘better use’ of data will provide greater opportunity for those specialists willing to rely on underwriting skill to help a growing number of customers who do not fit the boxes of the data driven mass market.”

Striking this balance is difficult enough for internal data, where insurers must devote significant resource to uncovering valuable commercial insights and then implementing them into their day-to-day operations. But finding the right tipping point for investment in external data carries additional considerations.

James Todd, head of home pricing at Direct Line Group, explains: “When this insight sits outside of the business, we have the additional barriers of contract negotiations, integrations and costs that clearly need to be considered against the benefits they are expected to bring.”

External data can help deliver better underwriting results and this is clearly a major priority for the market. Indeed, recent research conducted on behalf of Verisk by Post identified the soft market and competitive nature of the property sector as the biggest challenge facing underwriters.

Put to good use, external data can help alleviate that pressure. Daley comments: “Peril data has been successfully integrated into postcode/rating files for many years and has protected our core non-standard business and consistently delivered target loss ratio for a decade.”

Spotlight data 1

Difficulties with data

In addition to finding the right balance between the cost of data, the analytical expertise required to extract value, and the financial benefits generated, insurers must define very precisely the data they need.

But finding the penny that fits the slot is not easy and even when insurers do, it is not always available at the granularity or in the locations required.

The research identified escape of water as the peril that will have the greatest impact on pricing and underwriting in the next two years.

It was also cited as the peril that has generated the biggest increase in frequency and severity over the last two years.  And so, if data can illuminate the true nature of an escape of water risk at an individual property level, there are significant upsides for insurers.

But it is a tough nut to crack, as Annarita Roscino, head of predictive analytics at Zurich, explains: “The definition of escape of water is very broad and includes very different claims scenarios: from frozen pipes, through water leakage to appliances issues. It is, therefore, much more difficult to identify a unique data source that can explain such a variety of events.”

She adds: “There is still some work that needs to be done to better classify those claims before we can understand what data sources can support insurers in better predicting and preventing escape of water claims.”

Until insurers codify escape of water claims in tighter categories, using data to generate meaningful underwriting insight will be incredibly difficult. It is also true that the frequency and severity of escape of water claims are driven by property attributes that are not always easy to record.

Spotlight data 2

“In order to pin down our understanding of the risks individual properties represent, we need to have a detailed understanding of the build quality and materials used in the construction and renovation of buildings,” says Todd.

He continues: “This type of information is not as readily available as information relating to other types of risks, such as theft. In addition to this, our customers often do not know the details themselves, particularly when some of the works carried out on buildings occurred before they purchased the property.”

The desire to get hold of this data was reflected in the research, which found that property attributes were listed as the most important data for underwriting property risks in the future.  

While data may unlock some of the challenges around escape of water risks, there are a number of physical and legislative steps that would make a major impact overnight.

Daley comments: “The escape of water issue is wider than peril data; it is about buildings being resilient to freeze. Better insulation helps, but better building regulations around incorporating water shut off devices in properties would help, along with supporting customers to improve escape of water protection in existing housing stock.”

Data has its role to play, but it cannot come at the expensive of obvious and immediate solutions that are often more readily available and can generate market-wide benefits.   

Spotlight data 3

Partnership approach

While insurers are keen to buy external data and beef up internal analytics capabilities, they have also shown an immense appetite to invest in and partner data/technology specialists through standalone operations such as Axa Venture Partners, Aviva’s Digital Garage and Allianz X.

Zurich is no different and Roscino says: “Partnering with digital disrupters and gaining early access to the data they generate or new products they create is going to be a big game changer when talking about harnessing the power of external data sources. This would allow insurers to not only make better decisions about risks, create new products and generate opportunities to grow revenue, but most importantly to differentiate themselves from their competitors.”

The benefit runs both ways and by understanding what insurers want, data vendors can better fit their products to the practical needs of the market.

For example, Wyard says: “It is a constant disappointment when vendors say it will take years to build an equivalent dataset in another territory.” If vendors can prove the value of their data quickly and partner with deep-pocketed insurers early, it should be possible to reduce these development times and drive quicker returns for both parties.

Horses for courses when it comes to collecting and using data

Property insurer FM Global targets the world’s biggest companies and the size of the risks it underwrites has steered it away from a statistical approach to underwriting.

Owen Lewis, group manager for account engineering at FM Global’s London operations, says risk surveyors visit 60,000 sites annually and conduct around 100,000 surveys as some sites are visited more than once. Surveyors collect between 700 and 1,000 data points each time, totalling around 70 million data points.

He comments: “In terms of underwriting decisions, we typically do that on the basis of having sent our engineers to site. We train them all to the same way and it is the best way to assure consistency in what is a pretty inconsistent world of risk evaluation.”

“We have a very rigorous system of standards that we apply, and which are published,” says Lewis. “It is not statistical modelling where you can change one variable and end up with very different outcomes. We are always a little bit cautious of that and it is one of the reasons why we put so much store in the risk engineering and we release insurance capacity in very different ways depending on that. If we have sent engineers to site and been able to do what we would like through our normal engineering process, we can be very confident.”

Lewis accepts that this model is not appropriate for insurers that have thousands and thousands of customers who fit into different classes and profiles. In this market a statistical-led actuarial approach can deliver quick and accurate decisions at the point of sale. But if you have a complex commercial site with a value running to billions of pounds, a purely statistical approach is ineffective because there are so few comparables.

FM Global does use external data in many of its risk mapping and modelling tools, but when it comes to underwriting, it puts a priority on the data collected during physical surveys.

This sort of partnership approach will also help vendors set their priorities and develop efficient ways to access the most granular levels of data, which the research highlighted as the second biggest problem facing property underwriters.

For perils such as flood, getting even more granular data could prove invaluable with it being suggested that using building footprints – as opposed to address location – could identify an additional 400,000 properties which are at potential risk of flooding.

This level of detail will either let mainstream providers avoid these risks or price them more accurately, while it also throws up opportunities for start-ups or niche providers. Whatever the future of property underwriting looks like, data will sit front and centre.

Spotlight data 4
  • LinkedIn  
  • Save this article
  • Print this page  

You need to sign in to use this feature. If you don’t have an Insurance Post account, please register for a trial.

Sign in
You are currently on corporate access.

To use this feature you will need an individual account. If you have one already please sign in.

Sign in.

Alternatively you can request an individual account here: