How geocoding improves risk management – and what it looks like to improve it

Geocoding is quickly becoming an important tool in the insurance industry, allowing for more accurate risk assessment by pinpointing the exact location of insured properties. As insurers move beyond postal code-based assessments, geocoding allows asset-level analysis, providing in-depth information on geographic risks such as floods and wildfires.

With increased adoption driven by climate change and the need for more accurate data, geocoding is reshaping the way insurers manage risk.

In a recent interview with Insurance Business, Mark Varley (pictured), founder and CEO of Addresscloud, revealed how he founded the company in 2015 after working for nearly 20 years in the insurance industry.

His focus on geographic risk began during an important project with RSA Insurance Group, where he was part of a team that integrated geographic data into the underwriting process.

“RSA wanted to bring geographic and geographic risks to the table,” said Varley. “They were one of the first companies in the UK to do that, particularly in the residential, retail and commercial space.”

According to Varley, the use of geographic map data at the time of writing was difficult at the time.

“Before that, most companies in the UK relied on postcodes to assess risk. But RSA went on to assess the level of buildings, which means that each area has its own risk base,” he said. “You could no longer aggregate properties based on postcodes – it was about getting granular, down to the local level.”

This experience sparked Varley's interest in technology that creates environmental risks.

“I found the technology interesting – the idea that you can take a map, put it in a database, and ask questions about it,” he said. “Even now, 20 years later, he still impresses me.”

Addresscloud and geocoding

Varley defined geocoding as the process of converting an address into its exact location on a map. He likened it to what happens when someone uses Google Maps. “Geocoding is important in risk analysis, especially in insurance. Whether you're assessing risk, property characteristics, or managing exposure, finding the exact location of a property is important,” he said.

Varley also stressed the importance of accuracy in the insurance industry: “We need to know the exact roof of the building we are insuring, not just the general area.”

When Addresscloud launched, geocoding was its first service. “We created our geocoding system from scratch in the UK and Ireland,” Varley said. The company aggregates data from sources such as Ordnance Survey in the UK and Royal Mail, as well as Ordnance Survey Ireland and An Post in Ireland.

Addresscloud has also developed a custom address resolver that uses fuzzy matching to clean up incomplete or erroneous data. “We built a custom address book to handle these situations,” Varley said.

Addresscloud works globally, although the accuracy of its geocoding varies depending on the data available in each country. In some places, such as France, government geocoding data is freely available, making it easier to deliver accurate services.

“We offer a global network of API partners, and we recently launched a service where we call third-party APIs like geocoding in countries like France or Australia,” Varley said.

Varley also pointed out that Addresscloud has a map of 2.4 billion building steps around the world, which helps improve the accuracy of its services, even in regions with incomplete data, such as India.

The company has a comprehensive map of flood and wildfire hazards across continental Europe, the US, and Canada. By the end of the year, Addresscloud plans to expand to Australia, New Zealand and South Africa, with further expansion into Asia and South America next year.

“Customer demand is driving our growth,” Varley said. “We are a small team, but if a client needs to be served in a certain country, we have the tools and data to move quickly.”

Geocoding and future insurance challenges

Varley also addressed data privacy concerns, saying Addresscloud does not handle personal data. “We don't deal with personal data,” he said. “For example, we recently licensed US tax inspector data, which includes structural features but also personal financial information. We're removing personal data and focusing only on geographic information.”

He noted how Addresscloud works primarily with private companies, including insurance companies, banks, and utilities, as well as partners with commercial data providers such as JBA for flood modeling and mapping. “There is a growing network of private companies like ours operating in this space, using technology to provide critical risk analysis and mapping services,” said Varley.

Looking ahead, Varley sees improvements in geocoding as governments open up more accurate data and as AI is more integrated into the process. “We are working to train AI models to better understand and classify addresses, which will help improve the accuracy of geocoding,” he said.

Landscape risk analysis is also expected to become more sophisticated, with the use of satellite imagery, sensor data, and detailed structural traces to help assess hazards such as floods and wildfires.

On the broader issue of climate change, Varley believes that insurance companies are well equipped to deal with the growing threats from natural disasters but the industry is evolving.

“Most insurers already use catastrophe models and simulations to assess risk,” he said. However, he predicts more real-time use of data in the future. “I think we're going to start seeing more real-time data being used, with insurers using models earlier in the process, and before issuing quotes.”

Varley also sees regulatory changes pushing insurers to take a longer view of weather risk. “Although insurance providers are prepared for short-term risks, I think the regulatory changes will push them to look at long-term climate risks, similar to how lenders do,” he said.

What are your thoughts on this matter? Please feel free to share your comments below.