SafeGraph's Guide to Precision Insurance Risk Assessment

Location data can help insurance companies both minimize risk while also creating new kinds of meaningful and relevant customer value.

Insurance Companies Take on an Enormous Amount of Risk by Default

But that doesn’t mean they should willingly take on more risk than necessary. Rather, they should be focused on finding new ways to get a leg up on the competition, all while minimizing risk. Recently, this has involved bringing new technologies and solutions into the mix to provide customers with more long-term value and also a better overall customer experience. 

For insurance companies to be truly successful at what they do, however, they need access to highly accurate data. This is essential for pricing insurance policies with confidence. Fortunately, as an already incredibly data-driven and data mature industry, insurance companies not only understand the importance of data but also know how to use it to their advantage. 

In many ways, actuaries are the world’s original data scientists. They know how to analyze multiple factors, including historical losses, to predict the potential for future loss. And while their ability to do predictive modeling is already quite sophisticated as it is, there’s always room for improvement. This is where geospatial data comes into the picture. 

Points of interest (POI), building footprints, and foot traffic patterns datasets are making their way into insurance risk models. And rightly so because they can add a new and valuable dimension for measuring the intersection of human and environmental risk. This added geospatial element can make a world of difference in the overall accuracy of insurance policies. 

In this guide, we’ll look at how insurance companies can intelligently leverage location data to price policies that minimize risk and accurately reflect real-time circumstances. 

Key takeaways at a glance

  • Location data can add a new dimension of accuracy, precision, and real-time relevancy when it comes to writing and pricing all kinds of insurance policies. 
  • There are multiple policy types for which geospatial data can give insurance companies a cutting edge, from general liability to co-tenancy risk and usage-based insurance to understanding new social determinants of health. 
  • SafeGraph Places data provides insurance companies with access to highly accurate location data for writing better policies and providing better customer experiences.
New data sources are transformative for the insurance industry because they can make customer interactions seamless to increase brand loyalty, make critical business processes such as claims management efficient and even help implement preventive practices that can improve the overall profitability of the industry.

5 Ways Data is Transforming the Insurance Industry (Towards Data Science)

Insurance Companies Need to Start Using Data Differently

The insurance industry is quite complex and oftentimes hard to navigate, especially for insurance purchasers. This is one of the many reasons why data has traditionally played such a big role in this space. As nebulous and overwhelming as many insurance policies may seem, for those of you who actually read the fine print, actually getting to the point of writing and pricing an insurance policy requires analyzing a number of factors. Many of those are data-driven. 

So, we don’t need to explain or even reinforce the important role that data plays in the world of insurance. That’s likely a given by now. However, that doesn’t mean insurance companies can’t cast a wider data net to make policy underwriting more accurate and less prone to risk. 

Traditionally, insurance companies have relied heavily on geocoding and environmental data to assess risk. But unfortunately, as accurate as that may be as a predictor of environmental risk, it fails to capture either the real world or human side of risk, both of which can radically impact the accuracy of an insurance policy. 

That’s why now’s the time for insurance companies to begin using data differently. By joining location datasets—like POI, building polygons, and foot traffic patterns—to existing environmental datasets, insurance companies can begin to find the “sweet spot” that brings the human aspect of policy writing into focus. While the goal here is ultimately to minimize risk, as no insurance company wants to pay out more in claims than absolutely necessary, location data can also be used to create better, more personalized customer experiences. And although those seeking to be insured often aim for the most well-priced policies, added value programs and incentives in addition to standard policy benefits can sometimes tip the balances in favor of one policy or another, where the winning policy doesn’t always necessarily win on price alone. 

The big takeaway here is simple: Insurance companies today are facing more competition than ever before—and that competition is growing every single day—which means they must start using all data sources in more sophisticated ways to stand out and, of course, minimize risk. From our perspective, location data is the next frontier for improved insurance policy writing that takes important contextual or human (behavioral or demographic) elements into account. 

This is not merely the future of policy writing; this is already happening now. But it’s about time all insurance companies got on the location data bandwagon because failing to do so is akin to leaving money on the table—which, in the insurance world, is never really an option. 

4 Ways to Use Location Data for Writing Insurance Policies

Using location-based data for both risk modeling and improving the customer experience can help insurance companies write more accurate and risk-resistant policies that actually address real-time needs. Here are a few use cases where this data can make a big difference:

1. General liability insurance

The more people who visit a POI, the greater the risk that something human-related might happen. This is often the case when you factor in seasonality, especially in parts of the world where there is a true rain or snow season. These adverse weather patterns increase the possibility of customers slipping, falling, and potentially injuring themselves on a business’s property. The only way to know what that influx of traffic looks like is by leveraging mobility data to understand real-time foot traffic patterns.

Foot traffic data informs general liability analysis.

An insurer may find, after analyzing this foot traffic data, that a business seeing greater foot traffic during adverse weather months will need to pay a higher premium for general liability insurance—versus another nearby store that sees less foot traffic during those months—simply because of the added human risk that quickly becomes amplified by various environmental factors. At the end of the day, assessing risk is a question of knowing how many people might be exposed to those potential risk factors. Real-time foot traffic data, as opposed to predicting foot traffic based on the lagging factor of revenue, is a better and more accurate way for insurance companies to know when businesses are at greatest risk.

2. Co-tenancy insurance

When insurance companies price policies for a specific business, they don’t always know, in granular detail, what other businesses are in the vicinity that could drive up its risk. This is especially the case when a relatively low-risk business shares a wall with an indisputably high-risk business. In this instance, the low-risk business, almost by default, will assume some of the risk of its neighbor, even though their businesses have nothing to do with each other. 

POI data is critical for co-tenancy insurance analysis.

For example, if you run a gym or a restaurant that sits next to a fireworks store, your business runs the risk of being adversely impacted should an explosive go off accidentally one day, causing all sorts of property damage or, even worse, human casualties. 

Unfortunately, insurance companies don’t always have access to the right data and information to help them accurately assess this kind of co-tenancy risk. However, equipped with building footprint data, insurers can leverage precise polygons to price a policy more accurately based on the potential for adjacent or shared risk that exists.

3. Usage-based insurance

Usage-based insurance (UBI), where policy pricing is directly related to consumer activities, has become a growing phenomenon over the years. It’s essentially a way to incentivize consumers to trade-off less risky habits in exchange for lower policy premiums. This is quite common with car insurance companies that allow drivers to insert a connected device into their cars that measures all sorts of things, like time spent driving, distance traveled, and average speed. 

POI data informs usage-based insurance models.

While that is a great way to measure how respectfully a consumer is driving, it doesn’t provide any details around the potential risk associated with where consumers are driving to. For example, if an insurance company knows that a specific driver frequents streets and other public spaces with a lot of different POIs—and, therefore, greater foot traffic overall—then it can be assumed that this driver is open to greater risk from other drivers pulling in and out of parking lots or simply exhibiting poor driving habits. 

Without location data to provide these real-world insights, there is no way for insurance companies to identify additional factors that put drivers at greater risk. In the example above, the absence of location data basically means, from the perspective of an insurance company, that all streets look essentially the same and that potential risk can only be measured as a factor of that driver’s driving habits and not in relation whatsoever to the surrounding environment. Being able to distinguish these factors can help insurance companies price policies accurately. 

4. Social determinants of health

Because health insurance providers don’t want to pay on claims more than they absolutely need to, they are perennially focused on providing customers with added value health and wellness programs—like those targeted to weight loss or smoking addiction. The general thinking is that the healthier people are, the less financial risk they present to health insurance providers. 

Social determinants of health are integral to insurance modeling.

While these personal behaviors, otherwise known as “social determinants of health,” obviously play a big role in overall health outcomes, in recent years, insurance companies have become interested in understanding other social and demographic factors that could contribute to a person’s overall health risk. For example, POI and foot traffic data can help health insurance companies get a better understanding of where their members live as well as how their surroundings can contribute to lifestyle choices. 

By analyzing this location-based data, insurers can see whether a member lives in a so-called “food desert,” where access to fresh, healthy food might be limited or at a high cost, or far away from access to preventative healthcare, meaning that certain health issues may go overlooked based on convenience alone. Knowing this, insurance companies can not only write more accurate policies based on the real potential for health risks but also target health and wellness programs or other incentives to reduce the number of claims from at-risk populations. This has the potential to be a real game-changer for both the health insurance industry and the general wellbeing of the people they cover. 

Big data implementation results in 30% better access to insurance services, 40-70% cost savings, and 60% higher fraud detection rates, which is beneficial for both insurers and stakeholders.

How Big Data Impacts The Insurance Industry And Beyond (Yes Magazine)

Precision Policy Writing with the Help of SafeGraph Places Data

The SafeGraph Places dataset, updated monthly for utmost accuracy and precision, provides the granular POI, building footprint, and foot traffic data needed to write more comprehensive insurance policies, those that take into account both human and environmental risk factors.

Here’s a quick overview of the three types of location-based data within this powerful dataset: 

Points of Interest (POIs) 

Core Places data provides millions of global POIs for insurance data analytics.

Places includes base information—such as location name, address, category, and brand association, and accurate geocodes—for 8.3M places in the United States, Canada, and the United Kingdom. It also provides details around whether locations are currently open or permanently closed. And finally, it can also help shed in-depth light on the relationship that may exist between adjacent POIs.

When applied to the insurance industry, specifically, POI data can set a new baseline for understanding the business activities happening in and around a given location as well as how the broader business ecosystem can potentially have an impact on an insurance company’s approach to policy pricing. It, therefore, provides a ‘birds-eye view’ of business’s surrounding environment to help underwriters look at risk from a more people-oriented perspective as an added layer to other environmental factors, like the potential for damage caused by flooding or wildfires.

Building Footprints

Polygon data is the basis of insurance risk analytics.

Geometry provides precise building footprints for 7.8M POIs in the United States, Canada, and the United Kingdom, coupled with spatial hierarchy metadata and accurate geocodes. Because the Geometry dataset is not built with a centroid radius—and instead represents the true structural boundaries of buildings—these polygons are 96.3% accurate. From an insurance company’s perspective, this provides greater detail around potential co-tenancy risk, especially in cases where businesses share a wall and, as a result, can be affected by each other’s activities. Knowing this information can dramatically change a policy’s price as well as an insurer’s willingness to take on certain levels of risk. 

Foot Traffic Patterns

Heat mapping mobility patterns helps visualize consumer behavior.

Patterns data precisely measures hourly foot traffic to and from 4.5M POIs in the United States and Canada. This is powered by aggregated anonymized mobile device activity and other demographic data to paint a clearer picture of how often people visit certain POIs, how long they stay, where they came from, where else they go, and more. Although this data has not been traditionally used by insurance companies in the past, mobility data can help insurers understand the potential human risk associated with certain POIs and, thus, allow them to price policies accordingly. 

The explosion in available customer data (both personal and commercial), the growth in analytics technologies and the rapidly declining cost of computing power and data storage are prompting companies to invest in data analytics as a means to innovate.

Harnessing the Potential of Data in Insurance (McKinsey and Company)

Better Policy Writing, Fueled by Location Data

Although the insurance industry already understands the importance of accurate data, human-related geospatial data is a relative newcomer into this mix. However, by leveraging POI, building footprint, and foot traffic patterns datasets in relevant and thoughtful ways, insurance companies can not only write better, more accurate policies that minimize short- and long-term risk but also create new value for customers along the way. 

Location data is opening up new opportunities for the insurance industry. However, we realize that using it may seem a bit intimidating at first, especially if you’ve never joined geospatial data together with your other datasets. Fortunately, that doesn’t have to be the case whatsoever. With the right tools and techniques in place, location data can give the most data-savvy insurance companies a leg up amidst constantly growing competition. And of course, if you’re still not quite sure where to start, our team is always here to help.


Want to try it on your own? Check out our tutorial below!