To be clear, I was saying I *wouldn’t* do it onboard. The big upshot of the smartphone era is that everyone now has “the hardware” in their pocket, so it’s almost always cheaper to design an app.
Your use of the demonstrative pronoun was not clear. The errors in measurement that you complain about appear to be related to mobile devices. Hence my misunderstanding.
The errors aren’t inherent to the type of device, but to the simplicity of the KPI. Regardless of whether it’s on a smartphone or an onboard device, an accelerometer cannot determine whether I’m a bad driver who constantly tailgates and rides my brakes or a safe driver who occasionally has to slam the brakes because of a road condition I couldn’t control.
Which makes it a dumb KPI. You want to charge the tailgater more and the safe driver less, but that’s not what the KPI accomplishes.
You are not addressing two other complaints. If I understand, you complained that the app sometimes confused your driving behavior with your behavior when you were not driving. If so, that’s a mobile device problem.
Insurance companies use your phone and not some device mounted in the car? That's absurd.
As an engineer, I’d totally design it that way, actually, because it has all the sensors you need AND it monitors touchscreen interactions.
You would design an onboard system? Is using phones a cheaper and less effective method?
To be clear, I was saying I *wouldn’t* do it onboard. The big upshot of the smartphone era is that everyone now has “the hardware” in their pocket, so it’s almost always cheaper to design an app.
Your use of the demonstrative pronoun was not clear. The errors in measurement that you complain about appear to be related to mobile devices. Hence my misunderstanding.
The errors aren’t inherent to the type of device, but to the simplicity of the KPI. Regardless of whether it’s on a smartphone or an onboard device, an accelerometer cannot determine whether I’m a bad driver who constantly tailgates and rides my brakes or a safe driver who occasionally has to slam the brakes because of a road condition I couldn’t control.
Which makes it a dumb KPI. You want to charge the tailgater more and the safe driver less, but that’s not what the KPI accomplishes.
You are not addressing two other complaints. If I understand, you complained that the app sometimes confused your driving behavior with your behavior when you were not driving. If so, that’s a mobile device problem.