This year, after more than 15 years of writing about it, I decided to take the plunge on telematics-based auto insurance. I switched from one tier-one carrier to another seeking some financial relief from escalating costs, and at onboarding, my new insurer offered an on ramp to their telematics offering.

I was already saving a considerable amount of money by making the switch, so it wasn’t really the additional percentage in savings that attracted me to it. I was interested in the experience. Would I feel different driving knowing that my insurance company “had eyes” on me? Was I a better or worse driver than I think myself to be?

What I’ve learned over a month has been interesting. First of all, my car is the secondary car in the family: For trips with all of us, we use the other vehicle. My insurer, therefore, sees a guy who basically goes to his kids’ school, the gym, the grocery store, and maybe a couple of other places. In the week leading up to writing this blog – and I’m writing it on a plane out of town, so I won’t be adding to this total – I’ve driven less than a hundred miles in my car. I imagine most families have a secondary car that is about the same.

Second, my “adverse events” seem to come in bursts. The first time the device beeped at me for a hard brake, I recorded a second one on the same trip. A few days later, I had two, in the same day, on separate trips. About a week after that? Three in one trip. In between, no drama. My score is exemplary, but I wonder how many people have similar patterns. Sometimes, you just have a less than optimal day behind the wheel.

What can insurers learn from this? On the first point, perhaps targeting multi-car households and offering a break on the car that is driven less (as long as you can confirm the lower frequency via telematics) might be an enticing entry point to expansion of the market.

On the second one, though, I wonder if there’s more to say about the story the data tells. On the three-in-one day, I was behind a driver who was clearly lost, and themselves braking hard as they snaked around my (admittedly confusing) neighborhood. Was this a bad day for me, or a bad day for them? And what about a geography – if there are areas where the drivers are prone to bursts of “bad” habits in between long stretches of good driving, could there be an environmental factor causing it?

Maybe that doesn’t matter too much for insurance underwriting. But at a time when every little bit of data can help insurers price policies optimally – in ways that both maintain a tenuous bottom line and retain customers at a time of premium shocks -- it’s something. And I’m simply looking at a public-facing report and interpreting it offhand. My guess is that there’s lots of hidden information about populations and locations in what has been typically thought of as a personalization tool. What are some things your company has found in new data sets that you weren’t expecting? I’d love to chat about them.