Geocoding Examples for Next-Generation Insurance Analytics

What is geocoding?

Geocoding is the technical process of associating an address string to latitude and longitude coordinates that represent a point on Earth’s surface. This enables a property to be located and analyzed for a variety of purposes including navigation, logistics planning, broadband expansion, and property insurance analytics. For these applications to be successful, geocoding must be as accurate as possible. As an example, inaccurate geocoding could result in a delivery vehicle being routed to the incorrect property, or a home being deemed “uninsurable” due to misunderstood risk factors.

Types of geocoding

There are a few different methods for geocoding, each with their own accuracy implications to consider:

• Street-level geocoding: street-level geocoding plots an address point on an approximate location along the street segment where the property is located
  
  
• Parcel centroid geocoding: parcel centroid geocoding uses the latitude and longitude coordinates of the geographic center of a parcel of land as the location of the address
  
  
• Rooftop or building-based geocoding: rooftop-level or building-based geocoding derives latitude and longitude points from individual property structures

Each method has its own advantages and disadvantages depending on how it is being used. For instance, many open-source or free geocoders are street-level or parcel centroid-based, making them easy to integrate into an application. However, these methods are best suited for situations where approximate location is acceptable, as they do not accurately represent where an addressable property structure is on Earth’s surface. For instance, street-level geocoding is often sufficient for consumer navigation applications, while parcel centroid geocoding can support use cases that only require an understanding of the land associated with an address.

Building-based methods of geocoding are the most accurate for locating and understanding properties with structures, but are only offered by a few data providers. This is because rooftop-level geocoding is built using high-precision building footprints in addition to comprehensive address data, both of which can be difficult to source or create in a rapidly changing world. Despite the complexities in developing building-based geocoding solutions, they remain the preferred method for linking address strings with geographic coordinates for applications requiring high-precision. For example, insurance companies managing their need to see exactly where a structure is located on Earth in order to price its policy accurately, respond to claims, or understand its risk profile.

Reverse geocoding

Latitude and longitude coordinates can also be associated with addresses, a process known as reverse geocoding. Reverse geocoding is commonly used to understand which addresses are associated with a specific location on Earth’s surface. For some use cases, such as locating emergency callers, reverse geocoding takes the geographic coordinates of a point location and translates them to an address string. In other applications, address strings can be reverse geocoded from inputs of parcel boundaries, building footprints, street segments, or other larger geographic entities. This is often the case in insurance, when property and casualty (P&C) carriers are investigating property relationships and their impact on risk profiles. 

Examples of geocoding in P&C insurance

Geocoding is foundational to P&C insurance, providing an essential link between address and geospatial property data to power a wide range of applications. Ecopia AI (Ecopia) works with 7 of the top 10 US P&C carriers, the top two global reinsurers, and many smaller insurance and insurtech firms, exposing us to the many different ways geocoding is fueling innovation in property analytics today. Among the most common applications of geocoding in insurance are:

Property risk assessment

Whether balancing risk across an entire portfolio or assigning a risk score to an individual property, P&C carriers frequently use geocoding data to model and understand hazards facing properties in their book of business. Risk assessment involves first locating the property and then layering in additional data to see how the property’s location and condition impact its overall risk profile. For example, P&C underwriters pricing a policy will often compare a property’s location with flood zone, historical wildfire, and other climatological data to derive a quote that reflects these hazard risks. Similarly, property attributes related to how many structures are present or roof type and condition can factor into a risk score, as they indicate how well a property can withstand a hazard or what the overall impact will be. Because risk scores can drastically change when a property is misrepresented by only a few feet, the most accurate property risk assessment is performed using building-based geocoding.

Openly, an insurtech founded in 2017 to provide data-driven solutions to both homeowners and independent agents, leverages Ecopia’s Building-Based Geocoding within their property risk assessment models. Along with geocoded building footprints, parcel boundaries, and unique identifiers linking these property elements together, the Building-Based Geocoding API also delivers US Census data to Openly, enabling their models to factor demographic trends such as population density into their risk assessments. 

Learn more about Openly’s property risk assessment using Building-Based Geocoding

Replacement cost estimation

When P&C carriers are analyzing properties, a key metric they define is how much they would cost to replace in the event of a claim. Non-building-based geocoding methods used for replacement cost calculation often underestimate costs, as they do not leverage structures themselves to locate addresses. Across the US, 45% of occupied land parcels contain more than one structure. Street or parcel centroid geocoding methods typically aggregate these multi-structure properties into one address point each, leading insurers to price premiums based on inaccurately low replacement cost estimates and opening them up to paying out larger claims than expected.

Using Building-Based Geocoding, Tokio Marine North America Services (TMNAS) avoids these pitfalls by truly understanding the buildings associated with each address submitted for a quote. Ecopia’s comprehensive, accurate, and up-to-date database of geocoded building footprints enables TMNAS to calculate the replacement cost of each structure on that property, informing more accurate policy prices and reducing the likelihood of unexpectedly high claims payouts. 

Read the full case study about TMNAS’s shift to a true building-based database

High-precision underwriting

Geocoding is integral to P&C insurance from the moment a prospective policyholder requests a quote. When a customer submits their address for a quote, that address string is used to locate where the property is, assess its risk, estimate its replacement cost, and perform similar analytics to ultimately generate a policy price. If the price quoted is too high, insurers run the risk of losing business to another carrier that can insure the property for less. If it is too low, the carrier opens its portfolio up to greater financial risk. Both types of inaccurate pricing during the underwriting process are often due to incorrect geocoding that either over- or underestimates the overall risk associated with the property. For instance, a street or parcel centroid geocoder can misrepresent where the actual structure being insured is, which can impact how proximity analyses for nearby properties, flood zones, and similar hazards are factored into the policy price. In fact, non-building-based geocoding methods are only accurate 58% of the time, opening carriers who use those methods up to greater financial risk throughout the P&C customer lifecycle.

To avoid these effects of inaccurate geocoding during the underwriting process, flood insurance provider Neptune leverages Ecopia’s Building-Based Geocoding. Flooding is typically considered an “uninsurable peril” by many P&C carriers due to its increasing intensity and frequency, as well as the complexities involved in flood risk assessment. Neptune fills this market gap, providing flood insurance coverage powered by a building-based approach to geocoding integrated into their Triton AI underwriting engine. Ecopia’s geocoded building footprints enable Neptune to accurately understand individual property flood risk and generate quotes based on real-world conditions, providing customers with fairly priced policies and flood insurance coverage without opening their firm up to undue risk.

Learn more about Neptune’s building-based approach to flood insurance underwriting

Reinsurance strategy & analytics

While P&C carriers provide insurance to policyholders, they also take out insurance on the properties in their portfolio to better manage their own financial risk. This strategy, often outlined in a reinsurance treaty, identifies a threshold of property values that the carrier is responsible for paying out claims for themselves. Any property with a value falling outside of that range is then insured by a reinsurance carrier, passing the financial responsibility in the event of a claim onto that organization. Each P&C carrier has its own strategy for defining that threshold, but it is almost always driven by a geocoding foundation. This is because property-based analytics such as risk assessment and replacement cost estimation are important considerations for reinsurance treaties, as they indicate how much a carrier is liable for in a claims payout. The same drawbacks of street-level and parcel centroid geocoding apply to reinsurance analytics; unless an addressable property is analyzed from a building perspective, any risk assessment or replacement cost estimation is approximate at best.

Harford Mutual Insurance Group relies on Ecopia’s Building-Based Geocoding to not only locate all of the buildings associated with a single address, but also strategically group nearby properties to optimize their reinsurance strategy.  For example, if a hazard at one property results in damages to both that property and another within 100 feet of it, their reinsurance treaty will cover the costs for both buildings, as they are underwritten as the same “value subject.” Thanks to a custom reinsurance analysis platform powered by Ecopia’s building-based approach to geocoding, Harford Mutual was able to reduce property location mapping time by 75% and automatically group properties based on proximity, greatly increasing the efficiency and accuracy of their reinsurance analytics.

Read the full case study to learn more about Harford Mutual’s strategic property grouping

Ecopia’s Building-Based Geocoding: a digital source of truth for the physical world

Geocoding is foundational to the mission-critical property analytics conducted by P&C insurance carriers and insurtechs, so any inaccuracies in locating addresses on Earth’s surface can have serious implications for both policyholders and insurers alike. A building-based or rooftop-level approach to geocoding is the only way to truly understand where properties are located, how many structures are associated with an address, and a property’s overall risk score.

Ecopia’s Building-Based Geocoding, created using the first and only complete map of buildings in the US, is derived freshly each year from high-resolution geospatial imagery to capture all structures. These buildings are then appended with best-in-class address data to produce rooftop-level geocodes so carriers can instantly see which structures are associated with a single address, and how those structures relate to potential risk factors. 

Ecopia’s address points, building footprints, and parcel boundaries are also appended with a system of unique identifiers to indicate property relationships, streamlining master data management practices so P&C insurers can be confident that their analysis is comprehensive. What’s more, Building-Based Geocoding now includes FEMA flood zone classifications and change detection status fields to provide even deeper insight into property risk, eliminating the need for carriers to perform their own post-processing in order to derive this critical information.

To learn more about Ecopia’s Building-Based Geocoding and see how it can improve the accuracy of your own property analytics, get in touch with our team.