Susan Kuchinskas looks at how fleets, auto manufacturers and telcos can make sense of the data flowing from connected cars
Insurers and telematics companies are racing to get usage-based insurance (UBI) into the market; simply plugging a device into the diagnostic port starts the data flow. What they will learn about how to manage all that data—and how to glean insights from it—will drive new business models for fleets, auto manufacturers, telcos, and probably other new entrants to the game.
Clearly, getting data from a vehicle is not the problem. But companies need to turn that data into usable business information—and that's a challenge. (For more on business models, see Telematics and the business case for UBI and Telematics and customized UBI business models.)
Two recent announcements from Sprint highlight divergent approaches to analytics: onboard or in the cloud.
Sprint’s Integrated Insurance Solutions
In July, Sprint announced Integrated Insurance Solutions, a low-cost, turnkey trial program for insurance carriers to start their own UBI trials and pilot programs. A device that plugs into the diagnostic port captures vehicle information and driver behavior data, which is transmitted over the Sprint wireless network. A cloud-based system analyzes the data with driver scoring software that enables insurance carriers to improve driver risk assessments, reduce costs and improve profitability.
That followed an announcement that Sprint was teaming with Agnik for advanced data mining of vehicle and performance data. Agnik, which is also a partner in Integrated Insurance Solutions, offers several products for the UBI market. Its technology captures vehicle and driver behavior, performing statistical data mining, clustering, segmentation, trend analysis and predictive modeling, and then sends the resulting data analytics over the Sprint wireless network to a remote server that can be accessed by actuarial managers and policyholders. In this case, Agnik analyzes the data onboard the vehicle before sending it to the data center. (For more on UBI and driver behavior, see Telematics, UBI, and driver education.)
Moving data analysis to the cloud applies massive computing power to data generated by the car. On the other hand, that can be a lot of data to move…
Large data streams
According to Ronnie Beggs, vice president of marketing for SQLStream, a company that specializes in real-time analysis of very large data streams, automotive use cases may require the analysis of millions of events per second, compared to hundreds of thousands of events per second for other applications.
"You need to be able to do two things: capture all the streams and then clean and filter," Beggs says. For example, while GPS systems tend to be highly accurate, a percentage of the information will have errors. SQLStream's solution also applies trend data to the current data stream to improve its accuracy. It stores transportation-derived data for as long as 28 days to compare it to the real-time data flow.
We still don't know which data will be valuable and useful, says Blair Currie, vice president of marketing for IMS. That's why analytics partners will be crucial for UBI and other services. "When we talked about an end-to-end solution, we used to mean everything from R&D to design to installation and customer care. Now, realistically, that's with an analytics partner included," Currie says. (For more from Blair Currie, see Q&A: How to make money with insurance telematics. LINK)
In fact, different markets may require different data sets and custom analysis. For example, insurance companies are hoping for more than simple risk scoring, according to George Kandt, M2M marketing manager, Sprint: "They also want help with customer risk segmentation, vehicle location and stolen vehicle recovery, and fraud detection. We've heard a lot about enhancing their bottom line, in addition to reducing the costs."
Policyholders, on the other hand, while they also desire vehicle location, want to be able to monitor in real time how teenagers are driving. Fuel usage analysis and vehicle health monitoring are also on their wish list, Kandt says. (For more on UBI data, see Telematics and UBI: The data challenges, Telematics and the value of data and Telematics and probe data: The revenue opportunities.)
Agnik's MineCar combines the real-time telematics vehicle performance analytics with non-telematics data, such as historical data about the car, self-reported lifestyle information about the driver, repair details and current market data about the car.
While the basics of insurance ratings in pay-how-you-drive (PHYD) products may be relatively simple—speed, braking, time of day—those may not ultimately provide a robust enough picture. Currie notes that all of us may drive better on some days than on others, while installing a temporary telematics device to measure driving may create a placebo effect: You drive better while you're being monitored, but then revert when the device comes out. (For more on consumers and driving behavior, see Consumers and UBI: The power of value-added services, Telematics and UBI: How to increase consumer acceptance and Telematics, UBI, and driver education.)
Beggs adds that the ability to identify trends and patterns in real-time can address changes in a driver's style. "You have to look for lots of different patterns in real-time, not just apply one rule," Beggs says. "You need to be able to apply 700 rules and patterns in real time, and be able to keep the system running while adding in new rules."
According to Currie, companies that have better consumer insight will be best positioned to meet the demands of today's more empowered consumers. Consumers with unprecedented access to information and choices are upsetting the traditional business model of building things cheaper and faster. Now, Currie says, "The companies that have more influence are the ones that have more consumer insight. We're trying to move into insights and better intelligence by using our data in intelligent ways."
Susan Kuchinskas is a regular contributor to TU.
For more on insurance telematics, see Special report: Insurance telematics.
For all the latest telematics trends, check out Content and Apps for Automotive USA 2012 on December 4-5 in San Diego.
Coming up in 2013: Consumer Telematics Show 2013 on January 7 in Las Vegas, V2X for Auto Safety and Mobility Europe 2013 on February 19-20 in Frankfurt, Telematics for Fleet Management Europe 2013 on March 19-20 in Amsterdam, Insurance Telematics Europe 2013 on May 8-9 in London and Telematics India and South Asia 2013 on June 5-7 in India.
For exclusive insurance telematics business analysis and insight, read TU’s Smart Vehicle Technology: The Future of Insurance Telematics report.