Jeff Briglia is our Chief Insurance Officer.
– – –
It’s no secret that mobility is on the cusp of a revolution. Increasingly, autonomous vehicles are transforming the role of the driver, just as ride- and car-sharing platforms are changing car ownership.
What’s less obvious is that these same technologies are also upending automobile insurance. If a car is driving itself half the time, or all of the time, should the driver be held liable for an accident that is caused when the vehicle is in autonomous mode? If a person travels mostly by ride-sharing, and uses her own car very little, shouldn’t she pay a lower premium?
That’s just the tip of the iceberg. An even bigger revolution is being driven by information technology — real-time driving data, or telematics, that makes it possible to know both how much and how safely each individual person (or car!) is driving.
That promises to profoundly change the century-old approach for how insurers rate drivers for safety and how much they charge.
Put simply, instead of lumping drivers into different risk-groups based on proxies for risk like age, education, and even gender, telematics makes it possible to rate drivers based primarily on their actual, individual driving practices.
Done correctly, this will greatly increase fairness for consumers. Imagine this: one of the most expensive times to have auto insurance is when you have a teenage driver on your policy. Taken as a group, teens get into accidents at a much greater rate than adults. Today, your teen is assumed to drive like the average teen and, as a result, you pay for it. In the future, your teen driver’s insurance rate will be based on how she actually drives; not only is this a fairer way to price insurance but it also creates stronger accountability for people of all ages to drive more safely.
Equally important, the new technology will enable insurers to recognize the kaleidoscope of new approaches to mobility. People who rely heavily on ride-sharing platforms, and only seldom drive their own cars, can now pay less because they drive fewer miles.
We already have the fundamental technology. Metromile, which insures drivers based on the number of miles they drive and the quality of those miles, gives customers a device that plugs into the car’s diagnostic port and transmits a real-time stream of data about what’s happening. How many miles does each customer drive? Is any rapid acceleration or hard braking taking place? Is it a riskier time of day? Other technologies make it possible to spot signs of distracted driving or falling asleep at the wheel.
In fact, I can imagine a day when we don’t price at all on the basis of which risk-groups a person is in. If we have access to individual data and understand what it means, an extremely careful twenty-something millennial could end up paying less for car insurance than a 45-year-old Gen X’er who appears shaky behind the wheel.
Telematics can also transform the way we, as insurers, interact with our customers. If we see signs of less-than-ideal driving practices, such as a lot of sudden accelerations and hard stops, we can gently advise them on ways to improve safety and get better mileage as well. If drivers know they may be rewarded for following better practices, and they are receiving concrete tips, they are likely to become safer drivers.
Until recently, the biggest hurdle to this kind of personalized insurance has been knowing how to process all the incoming data. We’re talking about staggering amounts of data sent from millions of cars every minute of the day. Simply storing all that was a gargantuan task, and analyzing it was many times harder.
That has now changed. The plunging cost of computer processing power, and the explosive advances in artificial intelligence and “deep learning” computers, enable us to make sense out of what would have seemed like chaos just a few years ago.
To be sure, we have a ways to go.
For all the computing and algorithmic power at our fingertips, this is an entirely new approach to modeling driver safety. New approaches require new principles for risk analysis. The leaders in this new field will be companies that can both collect all that data and then actually learn the digital traits of safe drivers. Even after accomplishing all that, insurers will need to translate those risk-profiles into practical underwriting principles.
It’s also important to acknowledge that insurance is a regulated industry, and most state regulators are still grounded in traditional group-based insurance models. We are already working with several states to test-drive new kinds of policies, but this is a learning process for them as well as for consumers.
Metromile is in a strong position here. Because we started from the ground up with a model based on individual driving rather than group patterns, we enjoy a big headstart in understanding almost every conceivable kind of driver in almost every conceivable situation. We have data on more than 2 billion miles of driving so far…and we’re still in the early stages.
As a 20-year insurance industry vet, this is by far the most exciting time for innovation and improvement that I have ever experienced.
We’re putting drivers back in the driver’s seat.