Data holds the key for future insurance

The evolution of auto insurance is sure to include two primary components – those born from within (such as UBI) and those created by manufacturers and suppliers (smart mobility, autonomy, etc.). This will lead to a host of changes over the next several years as automakers carefully shape the future of transportation.

Insurance will be an important part of that future, whether acquired through an industry stalwart like Progressive or a newcomer like Root Insurance. Many questions remain, however, including the mystery surrounding who will be responsible for insuring tomorrow’s automobiles.

“I think there’s still a fair amount of uncertainty regarding how this will evolve,” said Dave Pratt, general manager for UBI at Progressive Insurance. “For example, who’s actually going to own the car? Is it an individual person who owns the autonomous car, or is it a big company that owns a fleet of cars and you’ll just buy a subscription or order up a ride when you need one? That will have a big implication for how the cars are insured.”

Alex Timm, CEO of Root Insurance, a start-up that uses an app to weed out bad drivers and offers discounts to everyone else, said that his company aims to “win” no matter the scenario. He has a flexible plan for autonomous vehicles, paving the way for a fleet owner like Uber or a manufacturer that decides to take responsibility for its product. Said Timm: “Do you understand the data and the risk that these individuals are undertaking? That’s really going to be key. I don’t believe full autonomy is right around the corner. I do believe it’s going to get there and I think, when it does, no one will have wished that they had taken a ‘wait and see’ approach.”

Dan Preston, CEO of Metromile, a pay-per-mile insurer, is excited for what autonomous technology will do for transportation. He’s particularly intrigued by the potential improvements in safety. However, he doesn’t expect driverless vehicles to arrive all at once and eliminate steering wheels altogether. Said Preston: “It’s unlikely that we’re all of a sudden going to have cars that have no steering wheel and will just be driving themselves. What we are going to see is more of the vision from Tesla, for instance, where the routine miles – the ones that are typically on a highway, the ones that are well-known – will be driven by the car. And then the harder ones, typically within cities where there’s lots of traffic and unknown events, will be driven by the person in the car.”

Technology is also expected to have a positive impact on the way car insurers conduct business. Start-ups are rising up, hoping to take the lead as either a new insurer or a partner. Arity, a start-up from Allstate created to analyse driving data and generate insights, is chief among them. “We think the transportation system is hugely inefficient and that it’s going to change over time,” said Gary Hallgren, president of Arity. “When you think of how it operates today, it just doesn’t make any sense for it to be as inefficient as what it is. There’s going to be an evolution as we move toward autonomous vehicles.”

Hallgren anticipates a series of steps as ride sharing and hailing become more prevalent, all the way through to semi-autonomous and, eventually, fully autonomous vehicles. He hypothesised that an algorithm may one day determine which vehicles are the safest to insure. “It’s those kinds of models and algorithms that we’re still going to need to have because there are still going to be accidents, there are still going to be risks,” said Hallgren. “Insurance won’t necessarily go to people. Maybe it turns into more product liability. But I think that the whole notion of these models and algorithms of how to keep people and vehicles safe is still going to be as relevant.”

The beginning of UBI

Pratt is optimistic that UBI is still in its infancy and will continue to evolve. He believes that mobile phone data will usher in a major change. “In addition to information about how you drive, we’ll start to get information about distracted driving,” said Pratt. “We’ll be able to learn just how important that is and hopefully give people engaging feedback through the phone that encourages them to be safer. To me, that’s an area where I think UBI will go from just being a tool (to more accurately price risk) to starting to actively help people be safer and actually avoid accidents. I’m pretty excited about that.”

Data will also be essential in identifying risky drivers. Pratt added: “As more cars are connected and there’s more data available, I think our understanding of what kinds of behaviours are risky and how the car itself can help people avoid risky situations will grow.” Hallgren estimates that today’s UBI, which determines the fee after evaluating a driver’s behaviour, is just the first step in the transition. Said Hallgren: “I think it’s the early, early stages of it. Inside of Arity we have next-generation models that are far more predictive than even the current models that are taking into account more variables. Instead of a couple it might be 20 or more that we would be able to pull information from and create insights that are that much better.”

In time this will lead to more contextual results. “You start thinking of a trip during the day on the interstate on dry roads relative to a trip that is in the winter time at night on side streets,” Hallgren added. “You can foresee the future of a trip-based risk, as opposed to just an overall assessment of your risk. I think it can get to that point but it’s going to take some time. It’s those models and insights that I think are really going drive the innovation in the industry going forward.”

Timm said that insurance will become much broader than UBI and tap into the IoT. “We’re seeing the IoT take effect where everything is connected and everything is generating data,” said Timm. “In that future it’s going to be imperative to understand how to collect the data. It’s going to be imperative to be very close to that source of data, which means being very close to your consumer. For us we think it’s the smartphone. You can’t get much closer to a human being in an electronic device than a smartphone. It’s always on them, some people even sleep with it.”

As IoT comes online, Timm envisions a future where all smart components synchronise together because many of them will be helpful to drivers. They could gain new insights into their vehicles’ health while insurers learn more about the people behind the wheel. Wearable devices could also be used to communicate consumer health information (ex: a spike in blood pressure) to the vehicle. No, cars are not doctors but they might benefit from knowing when a driver isn’t in the best condition.

“We think we’ve just begun with the data where sensors and connectivity and IoT are going to continue to push the number of data sources,” said Timm. “The important thing for insurance companies is to remember that you don’t have the option to use data anymore – you must use it. If a core competency of yours is not data science, I think you will be rendered irrelevant in the next 30 to 40 years.”

Data use and ownership

Big data may be essential to the future of car insurance but consumers are only just beginning to see the ways in which it might be used. All four men spoke about the importance of being open and honest with consumers before any data is collected. “If you were a customer of ours, one thing that we make explicitly clear is that data is always used in a customer’s benefit,” said Preston.

Pratt advised insurers to be “completely transparent” with customers about the information that’s collected and how it will be used. “And then we have to have good information security so that we actually live up to our promises,” said Pratt. “In our case we ask the customer to share a three-month sample of their driving data from the car. It’s really one snapshot of your driving. We tell you what your discount is and then we’re done. We don’t collect any more information about you.”

Once consumers are comfortable with sharing their info (and some already are), insurers can start to collect all of that data and learn new things about their drivers. Said Pratt: “Let’s say we have a new hypothesis that we want to test: do people who have more hard accelerations have more accidents? A couple years ago it would take us nearly a month to do that analysis just to go through all the data, count up each participant and how many hard accelerations they’ve had. And then do the statistical analysis and look at that summary and see if it actually predicts future accidents or not. With the new big data tools we can do that same analysis in nine hours. It lets us evolve our thinking and our models much more quickly.”

Pratt added that this is especially important with the advent of mobile phones, which provide significantly more data than what was previously available. “When we have the OBD device, we just collect speed from the vehicle once a second,” said Pratt. “You have one data point per second from the car. In mobile, we’re collecting data from all these sensors that the phone has – GPS, accelerometer, gyroscope – and, in order of magnitude, it’s more data that we’re collecting from each customer. Just having the tools to be able to take advantage of all that data is really important. Until a couple of years ago we just didn’t have the tools.”

Technology’s impact and preventing accidents

When considering AI, autonomous driving and connected mobility, Pratt expects the latter to affect car insurance before the rest. “I think it will give drivers a way to save money on their insurance based on their good driving and it will just be easy and seamless for them,” said Pratt.

From an insurance perspective, he would also like to see a database that shows which vehicles have each safety feature. He added: “I’ll use forward collision avoidance technology as an example. That’s where the car will brake for you if you’re about to have a crash. In today’s world it’s really hard for us to know which cars have that and which cars don’t. We’d love to have some reliable data source where, if you knew the VIN of the car, you could look it up and find the answer. Then that would probably allow us to give bigger discounts on insurance for the cars that have this stuff.”

Preston had similar ideas. He looks forward to the day when “dynamic insurance products” (those which can be priced differently based on the driving mode being used) become available. “I believe there’s an interesting opportunity to help transition this industry for what I guess will be many years as we get to fully autonomous,” said Preston. “It’s hard to say exactly where that end state will end up. Insurance will still have a role in that but it’s obvious that in the long-term auto insurance rates will go down if everything becomes autonomous in that way. You will start to insure the long-tail events, ones that are sort of more unexpected but higher in severity, rather than what’s more common now, which is high frequency, low severity.”

That’s one way to help consumers save money but Pratt would like automakers to re-evaluate the distracting technology found in most modern cars. By reducing those distractions, he believes there could be fewer accidents. “I think an important area that doesn’t get as much press is designing the various systems in the car to reduce driving distraction,” said Pratt. “If cars have big screens with lots of fun stuff to play with on the screen, my impression is that there would be a lot of value in more research and effort into how to make those user interfaces effective so that they don’t distract the driver.”


Leave a comment

Your email address will not be published. Required fields are marked *