The Future of High-Precision Location Data for Insurance

Challenges with inaccurate geocoding data

All industries can appreciate the power of accurate data when making critical decisions. As a property and casualty (P&C) insurer, long-term success is closely correlated with your ability to correctly assess and manage risk. And, to assess and manage risk for property insurance, you must first understand the precise location of each building you are assessing, along with the risks associated with that location.

Over the years, P&C insurers have relied on zip code or street-level geocodes as a proxy for understanding building location at a given address. However, these coordinates often result in substantial inaccuracies. Thus, misrepresenting the true location of a building and negatively affecting all downstream risk analysis.

In response to these challenges, the industry moved to adopt parcel-centroid-based geocoding, using a parcel center to represent a building’s location. Although this method improved accuracy compared to antiquated street and zip code based geocodes, the use of parcel-based geocoding still results in error-prone underwriting scenarios. This blog focuses on the challenges of using traditional location data solutions, how the industry is rapidly moving towards geocoded building footprints as the new standard, and how Ecopia AI's (Ecopia's) Building-Based Geocoding is helping solve the age-old problem of building location.

The importance of accuracy

Why ‘close enough’ is a gamble

As an insurer, input data should allow you to identify high-risk opportunities to adequately price risk you decide to underwrite. However, evaluating the risk of a peril for a given property begins with an accurate understanding of the location of the buildings on that property. Consider flood risk: how can you accurately assess a building's flood risk if you're unsure whether the building even falls within a flood zone?

The graphic below highlights the dangers of inaccurate geocoding data in a flood risk scenario.

The cost of being ‘close enough’

The threat of catastrophic risk

Ecopia conducted an analysis revealing the potential risks of using parcel-based geocoding as a source to evaluate property flood risk across the United States. To do this, we assessed the location of each parcel-centroid across the US compared to the respective building footprints on each property, identifying situations where parcel-centroids would incorrectly categorize a property as low risk (resulting in underpricing) or high risk (resulting in overpricing).

Ecopia found that, by using parcel-centroid based geocoding, over 1 million buildings across the US would be underpriced. Across each of these properties, the parcel-centroid was outside the flood zone but, in reality, these properties contained buildings within the flood zone.

In aggregate, these buildings represent an incredible value at risk across the US. Considering the average amount of a flood insurance claim from 2009 to 2018 ($43,846), the total claims from this risk would amount to approximately $43 billion. For the largest insurance carriers, this could translate into billions in potential claims across their portfolios. This is only considering the effect of geocoding on flood risk - when considering many other perils, the risk is likely much higher.

In addition, over 600,000 properties were identified as overpriced, as the parcel centroid was within the flood zone, but the building footprints were all outside of the flood zone. This effect could lead to customer churn and lost revenue.

The below heat map plots each of the potentially underpriced properties across the US.

Challenges with Building-Based Geocoding

Why all building-based geocodes are not made equal

Due to the challenges presented by using parcel-based geocoding, the industry is rapidly shifting towards building-based geocoding. However, achieving accurate building-based geocoding remains highly challenging for several reasons. 

The first reason is the difficulty in creating or accessing a complete database of every building footprint across the US. Many data providers leverage open data. However, a previous analysis done by Ecopia revealed that this dataset was missing 48 million of the country’s over 173 million buildings that were present in Ecopia’s enhanced geocoded footprints product at the time. Of the building footprints that were captured in this open data release, 34% of them had severe accuracy issues (defined as positional errors of 30% or greater).  

The second challenge, as shown in the image below, is linking address data with the correct building footprints. For example, how do you assign an address to a building when the buildings span multiple parcels, or when there are multiple buildings in one parcel? Many building-based geocode providers fail to address these challenges. The third challenge is updating building-based geocodes quickly and cost-effectively. Open data sets often rely on outdated imagery and fail to account for recent construction or demolition, making them unsuitable for accurate geocoding.

Ecopia’s Enhanced Geocoded Footprints

How Ecopia has achieved the gold standard for building-based geocoding

Ecopia has over a decade of experience leveraging artificial intelligence (AI) to digitize geospatial imagery, rapidly generating high-definition vector maps at scale with the accuracy of a trained GIS professional. Today, Ecopia's Building-Based Geocoding offers the first and only complete building footprint collection in the US, paired with best-in-class address data. We leverage the most up-to-date geospatial imagery available to create a unique source of ground truth, ensuring that our data provides the most accurate representation of reality so you can streamline master data management assess risk with more confidence than ever before.

Proprietary geocoding engine to match the correct address to the correct footprint

Ecopia's proprietary geocoding engine, designed to handle vast amounts of geospatial data, applies a unique machine-learning based parsing system to match each address to the correct building - resulting in the most comprehensive rooftop-level geocoding across the US, with over 270M+ unique primary and secondary addresses.

Annual updates

Ecopia leverages our partnership with leading geospatial imagery providers to source fresh high-resolution imagery of the US each year. This imagery is mined using Ecopia's AI, providing an update of building footprints every year - the updated building footprints are then further enhanced with the most up-to-date address data available. Our building footprints are pre-appended with year-over-year change detection status for deeper insights.

Street-level and parcel centroid geocodes can yield vastly different results for the same address, sometimes placing geocodes miles away from the buildings associated with a policy. This can lead to inaccurate risk assessments, incorrect premiums, and challenges in claims management. Ecopia’s building-based geocoding assigns geographic coordinates directly to building structures, ensuring accurate detection for underwriting, claims, and more. That's why insurers like Hartford Mutual, Standard Casualty, and SageSure are leveraging Ecopia's high-precision building footprint data for policy underwriting, pricing, risk assessment, and other critical decisions.

Get started today

Ecopia’s Building-Based Geocoding takes the guesswork out of location, resulting in best-in-class geocoding data to power accurate underwriting, risk assessment, and improve master data management. To learn more about Ecopia and our geocoded building footprints, get in touch with our team.