Telematics and UBI: The data challenges
Usage-based insurance needs to reach higher levels of market penetration for sustainable success. Jessica Royer Ocken explores how data collection and management strategies can help
As usage-based insurance (UBI) enters the mainstream, the methods used to collect and analyze driver data present potential challenges. So insurance companies must clearly articulate the goals of their UBI programs.
Those seeking to better customize policies will take a different path than those seeking to improve driver behavior. The nature of the UBI program will also determine the answers to other questions, such as: What types of data are needed? How should data be collected? And crucially, says California-based navigation, telematics, and probe vehicle consultant Christopher Wilson, insurance companies must determine whether they want to be in the “data processing and retaining business.”
In other words, will they manage the data on their own or choose a third party to assist?
Although the transition can be daunting, UBI can be cost-effective and appealing to drivers as well as potentially making the roads safer. “There’s a huge value from this,” says Wilson. “There are ways to really differentiate between good and bad driving and ways to feed that back.”
UBI is not a one-size-fits-all solution, so insurers should begin by defining their goals. Do you want to leverage policies based on the number of miles someone drives? How much weekend driving they do? Whether they drive at night? Whether they drive on highways versus residential streets?
Or will you examine the way your insured drivers behave behind the wheel? This approach could be used to weed out “nasty drivers,” notes Cyril Zeller, senior sales director for the global telematics segment of Telit, or to help drivers improve. Depending on the choice, insurers will need different equipment, different applications, and different data plans.
There are a number of data collection schemes. GMAC uses a minimalist approach that examines little more than where drivers are located and what time they are on the road. That’s not a lot of data, and it can be compressed for transmission.
Other companies add acceleration data. Do drivers accelerate a lot? Do they decelerate rapidly? These factors can be used as a “surrogate for accidents,” according to Wilson. That data collection is also fairly simple. The insurer receives a report saying a driver accelerated hard X times a day or X times in 100 driving hours.
However, to analyze detailed driving patterns, more complex data is needed. This is a larger undertaking, but it will “enable insurers to discriminate really good versus pretty good and fairly good drivers,” says Wilson. This, in turn, will allow them to tweak policy rates and provide feedback to drivers—perhaps through a green/amber/red light on the dash and via follow up reports highlighting problems.
“It’s a lot more data coming in, which means more ability to compare data to actuarial records and claims history and across drivers,” Wilson notes.
Data collection methods
Having such data can be an asset to insurers, but there must be a convenient, cost-effective means by which to collect it. Devices that require professional installation may work for commercial UBI applications, but individual drivers are less likely to comply. Instead, the OBD dongle looks promising for consumer-level UBI, according to Zeller: “For the first time ever, you can ship a telematics box by mail. That opens a huge new opportunity.”
In addition to providing diagnostic data, the OBD system reports the car’s VIN, which can help insurers verify that the car they have insured is the car from which they are receiving data. Although smartphone apps offer some possibilities for insurance data collection, their portability can make vehicle verification and accurate sensor data difficult. In some cases, the connection may not be reliable enough. And what happens if the driver forgets his or her phone—on the day of an accident? (For more on smartphones, see Smartphones as an incentive for insurance telematics.)
Another consideration, Zeller adds, is the ability to filter and preprocess data. It takes a lot to do driver profiling—GPS, gyroscope, acceleration, and perhaps even more. “And you don’t want to send [all that data] over the air,” says Zeller. Instead, the in-vehicle device should be sophisticated enough to differentiate between a reading of 0.01 G (a non-event) and 3 Gs—and only transmit the latter.
Some companies have a ‘black box’ they ask drivers to keep in their vehicles; GMAC collects data through OnStar. It’s not a deeply integrated system at this point, but it could be. OEMs might help with this process, by integrating more sensors into the embedded system or OBD port, for example, in exchange for a cut of the insurance. Most are watching GMAC, though, to see if a business model emerges. (For more on data and revenue opportunities, see Telematics and the value of data and Telematics and probe data: The revenue opportunities.)
The final piece of the data puzzle, and perhaps the most daunting for insurance companies, is how to deal with all the collected information. Insurers keep everything for backup and liability reasons, but the influx of UBI data could be overwhelming. “Do you want a turnkey solution you will brand?” Zeller asks. “Or will you be involved in creating the solution?”
Some insurers opt to have a third party collect the data, process it, and develop a ‘score’ to rate their drivers. Several companies are now filling this niche. In some cases, the insurance company provides the algorithm to be applied to the data; in other cases, the partner handles that as well. One still unresolved question: How does this score translate into rates? “There’s a lot of work to be done on the actuary side,” Zeller explains.
Whatever their goals for UBI, insurance companies are “realizing with more data, and more sophisticated data, you can do a much better job, but it will cost more,” Wilson says. “The benefits will play out over tens of years. But they have to make the choice now.”
Jessica Royer Ockenis a regular contributor to TU.
For more on insurance telematics, see Special report: Insurance telematics.
For exclusive insurance telematics business analysis and insight, read TU’s Smart Vehicle Technology: The Future of Insurance Telematics report.