How is geospatial data used in insurance?

Geospatial data is increasingly used across industries, including insurance. Having data that accurately represents the physical world is an integral part of assessing risk, understanding property relationships, and more as property and casualty (P&C) carriers manage their books of business. In this guide, we’ll explore the different ways insurers leverage geospatial data, the types of data most often used, and where to get it.

Risk assessment

The most obvious application of geospatial data in P&C insurance is in visualizing and analyzing physical features to calculate risk, and determine how that risk impacts policy prices, replacement cost estimates, and other elements of the carrier’s portfolio. For example, with an up-to-date map of a property and its surroundings, insurers can measure how far the structure is from a water body, and thus how susceptible it is to possible flood damage. Additionally, detailed property attributes can provide further context for risk assessment, providing information like what type of material a structure is made of, which can also determine how resilient a property is to certain risks. There are many different ways insurers use geospatial analysis to understand risk, especially given the complexities brought on by climate change. The insights derived from this risk assessment inform a variety of decision-making use cases, which we’ll explore later in this guide.

Master data management

Insurers also use geospatial data to underpin their master data management (MDM) strategy. Carriers work with an exorbitant amount of data, from customer contact information to addresses and parcels to historical claims and beyond. Managing all of this data, and effectively tying it together to reflect real-world relationships, is equal parts essential and challenging for P&C carriers. However, geospatial data provides the location context that is often the missing component of MDM strategies, helping to link different elements of a property and its risk profile together and streamline operations. A good way to think of this is to consider a single family home with a detached garage; when a policyholder submits a claim for damage to the garage, the claims department can easily look up the garage’s relationship to the main home, and determine the correct policy it is covered by. Believe it or not, different departments within a single carrier often use disparate datasets, so these types of data-driven connections are not always possible. With a single source of truth for geospatial information, however, insurers can standardize their MDM and reduce tech bloat.

Top 5 geospatial datasets for insurance

If you ask a carrier what they need to know about a property to insure it, they’ll likely respond with “everything!” While that’s certainly the ideal scenario, there are a few key geospatial datasets to focus on first when implementing geographic information for better risk assessment and master data management. Here are the top 5 types of geospatial data we see insurers leveraging in their P&C workflows.

Geocoding

Geocoding is the process of plotting a street address onto a map with latitude and longitude coordinates. This is needed to understand the true geographic location of a property, and is the foundation of P&C insurance. However, there are a variety of geocoding methods out there that vary in precision, so carriers should be mindful of which they choose. If addresses of properties are not correctly pinpointed on Earth’s surface, the analytics used to determine policy price, replacement cost, and claims responses will not be accurate. 

Building footprints & 3D buildings

While geocoding pinpoints the location of a property, it provides just that - a point. Properties are not single points of latitude and longitude, but rather three dimensional structures with nuances and unique characteristics. Building footprint and 3D building data are often used by insurers to understand the true geographic extent of a property, as they add much needed context to geocodes for risk assessment. The best building data is appended with unique identifiers to link these structures back to the geocode, so carriers can not only locate the property or its policy but also analyze its risk profile.

Parcels

Another key component of property data is the parcel. Parcel data refers to the entire lot of land a property encompasses, which is also important for P&C insurers to understand as they assess risk and effectively manage data about their books of business. Analyzing all of the structures on a parcel, not just the main building an address is associated with, is critical for insurers to see the full picture and make policy decisions accordingly. For example, some but not all of a parcel may be in a flood zone, which the carrier needs to factor into its risk scoring. Once again, unique identifiers connecting parcels to their related buildings and address points helps insurers make connections between different property elements, and better manage data across their organization.

Property attributes

Geocodes, building footprints, and parcel boundaries are all foundational to geospatial analytics in insurance, but can be further augmented by detailed attribution. Understanding where a property is is important, but the most innovative insurers take their analysis a step further by layering in details about the property’s roofing material, room count, and presence of amenities (like a swimming pool). These property attributes help carriers determine the value of a property by factoring in context not provided by property size alone, which informs replacement cost estimation, policy pricing, and claims response.

Hazard data

Hazard information also adds a layer of necessary context to property data. As property attributes include details about characteristics of a specific parcel or structure, hazard data provides information about local environmental conditions that may impact a property’s risk profile. Flood zones have historically been leveraged in underwriting and replacement cost analytics by P&C carriers, and today many are beginning to use detailed flood models that factor impervious surfaces and elevation into the equation. Insurers also use data about wildfire, earthquake, wind, and other hazards in their risk assessment, which is increasingly critical in the face of climate change.

4 most common use cases for geospatial data in insurance

The information provided by all of these data types can be applied in a variety of ways useful for insurers, but there are four use cases that stand out for P&C carriers. In this next section of our guide, we’ll explore the most common ways geospatial data is leveraged in insurance. 

Underwriting

Policy underwriting is one of the most popular use cases for geospatial data in insurance. When a potential customer reaches out to a carrier about opening up a policy, the carrier must input the customer’s address and assess its risk profile to determine the appropriate policy price. For instance, a carrier may input an address into their geocoder, see the parcel and building structures associated to that address, and use property attribute and hazard data to calculate its risk score and determine a rate. 

However, that is easier said than done. Inaccurate data or ineffective MDM can greatly impact the policy price ultimately quoted to the customer, leading to underpriced policies that open the carrier up to greater financial risk, or overpriced policies that provide a poor customer experience. Despite the increasing availability of geospatial datasets for insurers to choose from, many policies are not accurately priced. In fact, we recently found that there is over $43B in uninsured flood risk in the US due to inaccurate data analytics in underwriting.

Replacement cost estimation

Similarly, P&C carriers leverage the insights from geospatial data sources to calculate replacement costs for properties in their books of business, which informs the underwriting process as well as reinsurance strategies. Here’s an example of how Tokio Marine North America Services uses Building-Based Geocoding to upgrade their replacement cost and risk estimation practices from their historical approach of using address aggregations.

Having comprehensive, accurate, and up-to-date data about each property in a portfolio is essential for precise replacement cost estimation, as many factors impact a property’s value. If a property is not geocoded correctly, appended with relevant attributes, or analyzed in relation to nearby hazards, carriers run the risk of over or underpricing a policy, or not optimizing their own reinsurance strategy for that property. Many insurers have been turning to geospatial data -powered MDM to inform their reinsurance decisions, identifying groups of properties to reinsure together and better balance the risk across their portfolio.

Claims management

Geospatial data is useful for carriers even after a policy is in place. When a policyholder submits a claim for property damage, insurers must investigate the claim to determine the payout to the customer. While part of the information used in this investigation is submitted by the policyholders themselves, often in the form of pictures of the specific damage, carriers augment their analysis with the geospatial data on file for that property. This helps insurers assess the possible damages and respond to claims more quickly and accurately, as well as flag potentially fraudulent claims.

As we mentioned before, many carriers struggle with tech bloat and geospatial lineage, with disparate departments working from different sources of information. This can cause issues in the claims process, especially when data used to investigate a claim is contradictory to the data used to underwrite the policy in the first place. With effective MDM across departments, insurers can develop a single source of truth for all of their operations and avoid costly mistakes in both pricing and claims management. 

Property analytics

P&C insurance carriers perform a wide array of analytics that go beyond underwriting, replacement cost estimation, and claims management. As more geospatial data becomes available and the P&C landscape becomes more complex, carriers are continuously innovating new solutions to beat out the competition, strategically plan market expansion, and adapt to a changing climate. 

For example, the increased frequency and intensity of national catastrophe (nat cat) events in some regions of the US has caused some carriers to cease operations in those areas altogether. Those decisions were likely backed by careful geospatial analysis that indicated the financial risk taken on to insure those properties was too great. However, other carriers are using geospatial data and property analytics to adapt their strategies to the new risks posed by climate change, leveraging the deep insights data can provide to enhance their own resilience and endurance.  

Where to get geospatial data for insurance

It seems as if an abundance of geospatial data exists for insurance, but the reality is that not all of this data is valuable for the high-stakes analytics P&C carriers perform. Here’s a quick breakdown of different geospatial data sources for insurance and what to expect from them.

Open source

Some of the data insurers use is freely available to anyone on the internet, most notably parcel data. US parcel data is generally made available in some way by municipalities, but the freshness and format is never consistent and can sometimes require a lot of data manipulation on the carrier’s end. Other open government data sources can also be helpful for insurers, such as FEMA flood zone boundaries or NOAA coastal change analysis data. Some open source building footprint datasets exist, but they are usually intended for consumer mapping applications or high-level visualizations, and lack the accuracy, completeness, and freshness insurers require for property analysis. As a result, most insurers rely on a mix of their own first-party data and carefully selected third-party sources.

First-party data collection

Insurers collect and maintain an enormous amount of data on their own simply by conducting business. Books of business contain addresses, policyholder names, and other key data points that are all used at some point in underwriting, replacement cost estimation, and claims workflows. But while this data is essential, it lacks the full geographic context P&C carriers need to successfully perform their various analytics processes.

Creating geospatial data in-house is an option for carriers looking to extend their analysis capabilities, but requires a lot of work and can quickly become expensive. For example, to develop an accurate building footprint dataset to associate with all of their policies, insurers need to manually digitize features from satellite or aerial imagery. This is not only an extremely resource-intensive process, but also is difficult to maintain with how often buildings change in the real world. Similarly, building a geocoding engine is one of the most complex geospatial endeavors, requiring deep knowledge of address validation theory, geographic information systems (GIS), and other highly technical solutions. For this reason, most insurers license geocoding and geospatial datasets from reputable third-party providers.

Third-party data providers

P&C carriers seem to have their pick of third-party data providers in today’s market. Data, SaaS, and insurtech companies are all competing with each other to power the property analytics at US carriers, claiming to offer the best data and insights for decision-making. There are a lot of great solutions out there, but a couple things to keep in mind when ultimately selecting a data partner.

For instance, make sure to evaluate the level of precision of the geocoder. Is it street-based, parcel centroid-based, or building-based? Only Building-Based Geocoding can provide the granularity needed to avoid mispricing policies or miscalculating risk. Similarly, how is the data created and when is it refreshed? If data is created completely by machines, it’s probably too good to be true; but if it is generated from manual digitization, it’s most likely out-of-date. The best third-party data providers combine artificial intelligence (AI) and human annotation to deliver insurers with an accurate and up-to-date reflection of the real world. And last but not least, assess how interoperable the datasets you are evaluating are, and determine if their system of identifiers is conducive to effective master data management.

Get started with Ecopia

Ecopia is building the foundation of a digital twin to provide P&C carriers with a single source of truth for property intelligence. Spun out of PhD research ten years ago, we are solely focused on digitizing the world with a mix of our AI-based mapping systems and world-class team of geomatics engineers. Through our network of unique and persistent identifiers for parcels, buildings, and addresses, all derived from the first and only complete map of buildings in the US (which we made!), Ecopia is empowering the insurance industry with the data they need to adapt to increased climate risk, rising consumer expectations, and a highly competitive landscape.

To learn more, get in touch with our insurance team.