Telematics data offers big challenges — and big opportunities
The auto insurance industry as we know it is going through its third revolution. Following the direct insurance and credit rating revolutions, telematics data is now driving the usage-based insurance (UBI) revolution. Rich data from telematics devices has become so available and affordable that insurance carriers can use it to understand driving behaviors. Instead of relying exclusively on traditional rating variables — which are only indirect indicators of risk — insurers can now directly measure how people actually drive.
Progressive Insurance, for example, is already capitalizing on this revolution. The primary auto insurance product offered on the Progressive website (the number one car insurance website, according to Keynote Competitive Research), is a UBI product, Snapshot. Customers can either drive with it for 30 days and get a better rate when joining or opt in to continuous monitoring and get additional discounts according to how safely they drive. Progressive is using a few simple parameters, including braking frequency, mileage, and night driving. Over the last two to three years, Progressive reportedly sold more than a million Snapshot policies. That huge success effectively paved the way for the UBI revolution and pushed nearly every large insurance carrier to offer its own version of a UBI program.
Data is everywhere
Telematics data is now available through a variety of devices. While professionally installed equipment is still relevant (mostly for commercial vehicles), the main type of UBI device used today is a small gadget you can plug in yourself to your car’s on-board diagnostics (OBD) port. The devices are relatively affordable, but together with the cost of cellular communications and the logistics associated with their provisioning and management, they still require insurers to make a significant investment when introducing UBI. Until now, those factors have limited UBI to insurers able to build and fund such programs themselves.
Car manufacturers embed other telematics equipment in vehicles. GM’s OnStar and Ford Sync are major examples. Traditionally, data from those devices was either unavailable for insurers or of insufficient quality for insurance purposes. The recent proliferation of smartphones is now driving some car manufacturers to allow drivers to link their phones to car systems for infotainment purposes. That new approach holds promise for cars to collect more data and have that data available for drivers and their insurers. In January, Google launched the Open Automotive Alliance, aimed at connecting its Android smartphones with major car brands, including Audi, GM, Honda, and Hyundai. Similarly, Apple plans to introduce iOS in the Car, a platform that will allow car manufacturers to enable iPhones and iPods to stream apps, music, and other information to and from automobiles.
Smartphones alone are best-of-breed telematics devices. Most smartphones are equipped with a variety of relevant sensors — GPS, accelerometers, and gyroscopes — with nearly infinite data storage and communication capabilities. They also offer a platform for distributing new applications and communicating with drivers. Some nimble developers are already offering apps that collect driving and vehicle data and provide a more personalized connected-car experience. While it’s still difficult to rely on such data and the availability of smartphones in cars, those trends are likely to open an opportunity for insurers to collect more data faster and at lower costs while maintaining a more personal and real-time relationship with drivers.
Rich data, poor data
Abundant and low-cost driving data is around the corner. Nevertheless, the insurance industry is still trying to make sense of first- and second-generation UBI programs. Those programs use exposure-related driving variables, such as mileage, duration of driving, and number of braking or speeding events. Traditionally, telematics vendors made those data variables available to insurers simply because that was the only data available. Unfortunately, those variables are just secondary contributors to risk.
Understanding risky driving behavior requires richer data. Rich data relates to the quantity and quality of data. In today’s UBI landscape, insurers need to find a way to differentiate and make sure their UBI program collects data that allows them to remain competitive over time. Collecting the wrong amount or type of data today means the current UBI program would have limited benefits and would require redevelopment in the future.
The only sustainable solution is rich data. Rich data ensures that even in years to come, and as analytics continually improve, the data could support such enhancements.
Rich data challenges
Taking UBI into the third generation will require insurers to handle even greater volumes of data. With traditional driver rating plans, insurers can collect just a few dozen data points per driver over the lifetime of a policy. But telematics devices are capable of collecting dozens of data points per driver per day. To mature into the third generation of UBI, insurers need to find a solution for collecting, communicating, managing, and analyzing unprecedented quantities of data.
An additional challenge is the harmonization of data across different devices, platforms, and data sets. Building a sustainable UBI program requires an insurer to become independent of a specific device vendor and offer solutions for smartphones and connected cars. A program needs to structure, validate, clean, and bring together each such data set into a unified platform that can analyze the data regardless of how it’s collected or how rich it is.
Another challenge is sensor calibration. Rich data sensors, such as accelerometers, gyroscopes, and car computers, need calibration so programs can interpret their data correctly. Three-dimensional accelerometers are considered the critical element in monitoring how people brake, accelerate, turn, and negotiate a curve, ramp, or traffic circle — the basic elements of driving behavior analytics. Accelerometers are by nature uncalibrated and require extensive processing before insurers can interpret their data.
Once a UBI program collects, validates, harmonizes, and calibrates data, the biggest challenge lies in the ability to make sense of the data. First- and second-generation UBI used simple counting of a few basic driving events, but introducing the third generation mandates a much deeper understanding of driving events and their context. No one braking event is similar to another. For example, it’s relatively easy to explain why braking in low speed on a rural road is much less risky than high-speed braking on a freeway. But it can be difficult to make sense of other types of events and the permutations of their environmental characteristics. A good analytics platform should differentiate between the different types of braking events while also taking their context into consideration: where they occurred and the implications of that location, the conditions on the road (traffic, weather, infrastructure), and their overall contribution to risk.
Rich driving variables are the way to understand how people drive and to sustain UBI risk models over many years.
Big opportunity for commercial lines
Commercial lines auto insurers can achieve their telematics goals relatively easily. Commercial vehicle owners and fleet managers already use telematics devices for a variety of operational reasons. Most of the commercial vehicles in the United States are still not equipped with telematics, but installation rates are continually growing. That creates a huge opportunity for insurers. While UBI has been more available for personal lines, opportunities are increasing for commercial lines insurers to introduce new insurance schemes for their fleet customers to benefit from telematics. UBI may be one of those models, but there’s also much to gain from other models, such as advanced fleet safety and risk management solutions.
Asaf Tamir is head of Telematics Innovation for Verisk Telematics, a unit of Verisk Analytics. In 2004, he cofounded and led Sensomatix, a technology company that developed the Driving DNA analytics platform for usage-based insurance and the Safety Scoring model. Verisk Analytics acquired Sensomatix in 2013.
What is Verisk doing?
Verisk recently formed Verisk Telematics, a business unit devoted to the development and application of telematics and analytics for insurance purposes. We’ve been pursuing telematics for many years, but this new unit further highlights our belief in the potential of telematics for the insurance industry.
In 2013, Verisk acquired the assets of Sensomatix. Over the last decade, Sensomatix developed Driving DNA®, an advanced analytical platform for UBI, and the Safety Scoring® rating model for driving behavior. Key Sensomatix personnel joined Verisk Telematics and formed the Verisk Telematics Innovation Center. This research and development facility is part of Verisk Telematics and leads the technological development of UBI programs and models for the company.
Verisk continues to file its GeoMetric® location-based rating rule and Safety Scoring behavior-based rating rule in many states, and we’re working with customers to use those models for their UBI programs.
For more information on Verisk Telematics and our programs, e-mail firstname.lastname@example.org.