Insurance, Tech and People

Aviva is trialling a smartphone App that will assess driving style. Using the phone’s built-in accelerometer, the app will measure acceleration, braking and cornering.

In many ways this is no different from other permanently installed tracking systems increasingly used for underwriting and assessing risk. Differences include:

  • Since it utilises the policyholder’s existing phone, no installation is necessary and no additional hardware costs are incurred
  • the App can be updated remotely quite easily and uninstalled by the policyholder at any point. A range of issues arise from this, including potentially greater acceptability to policyholders since they have the control. On the flip side, insurers will have to decide what it means if the driver is driving without the app installed and running. The rules can’t be too draconian here since battery life issues and others mean it won’t always be practical to have the app running.
  • Interestingly, the way Aviva side-steps this issue is by requiring the app to run for a 200 mile test period only. Of course this adds any number of additional issues. How representative is that 200 mile sample? How easily can thus be gamed by policyholders who are prepared to drive like model citizens for 200 miles but no further?
  • The simplest way to test this would be do both. Use the app for 200 miles on a sample who also have a gps tracker. Double blind requirements aside (which may be dangerous while driving) this would provide the data necessary to evaluate the predictive power of the app for the tracks. A way around the behaviour monitoring nature of the tracker is to run the get on some people who aren’t aware of the tracker. With policyholders this is almost definitely going to be unethical under every imaginable scenario, but with employees or test participants there may be a way to do it.

So a range if interesting questions in need of answers. Whatever the answer, this is clearly part of the new phase of tech and data supporting motor underwriting around the world.

The next interesting observation for me was the comments of people on hearing abut this specific system. Many were directed at telematics in general, which suggests this still isn’t widely known or accepted. Insurers will always have an uphill communication battle, and while sophisticated new rating systems may improve underwriting, it only makes this “people” challenge more difficult.