What Telematics Actually Scores

Usage-based insurance has moved from novelty to default offer. An app or a plugged-in device observes how the vehicle is driven and feeds a score into the rating. The pitch is straightforward: safe drivers should pay less, and now the carrier can tell who they are.

The pitch is not dishonest. But the gap between "how you drive" and "what the device can measure" is where the difficulties live, and it is worth understanding before you agree to be measured.

The measurable set

Programmes differ, and the specific inputs and weights are set out in each programme's own terms. Broadly, the categories in use are these.

Mileage. The most defensible input. Less exposure genuinely means less expected loss, and it is measured accurately.

Hard braking and rapid acceleration. Detected by accelerometer against a threshold. This is the input that generates the most complaints, for good reason — see below.

Time of day. Late-night driving is treated as higher risk in most programmes.

Phone handling. App-based programmes commonly detect screen interaction while the vehicle is moving.

Speed, or speed relative to limit. Implementation varies considerably; some programmes score absolute speed bands, others attempt to compare against posted limits.

Cornering forces. Lateral acceleration above a threshold.

Notice what is absent. The device cannot see following distance, mirror discipline, lane positioning, hazard anticipation, or whether the braking event it just logged was you being inattentive or you being the only person paying attention. It measures the physics of the vehicle, not the judgment of the driver.

The hard-braking problem

This deserves its own treatment because it is the single most common source of unexpected scores.

A hard-braking event is a deceleration exceeding a threshold. It does not know why. Someone pulling out of a side road, a light changing on a short amber, a cyclist appearing from between parked cars, heavy stop-start traffic — all generate events, and all of them represent a driver braking appropriately in response to something.

The structural consequence is that dense urban and suburban arterial driving generates more events than open highway driving does, at equal or better levels of attention. A driver who moves house from a quiet lane to a busy town can see their score deteriorate without changing a single habit.

This does not make the input useless. Aggregate hard-braking frequency probably does correlate with claim frequency across a large population, which is what the rating plan cares about. But it means an individual score is not a verdict on your competence, and it means the programme suits some driving environments considerably better than others.

Three questions to ask before enrolling

Can the rate go up? Some programmes are discount-only: you can improve your price or stay where you are, but you cannot be surcharged. Others allow the score to increase the premium at renewal. These are materially different products and the difference is not always prominent in the marketing. Get it in writing from the programme terms.

What happens to the data? Ask how long driving data is retained, whether it is shared with third parties, and whether it can be requested in a claim or a legal proceeding. Data practices are governed partly by state law and partly by the programme's own terms, and they vary. This is a reasonable question to ask before you generate the data, and an awkward one to ask afterwards.

Can I leave? Withdrawal terms differ. Some programmes let you exit at any point with the price reverting; others hold the score for the remainder of a term. If a monitoring period has already run, ask specifically whether exiting mid-score removes the effect or locks it in.

Add a fourth if the household shares vehicles: find out how the programme attributes trips. An app tied to a phone may score you for a journey you were a passenger on, and a plugged-in device cannot distinguish drivers at all. Programmes handle this differently and some offer a correction mechanism.

Who it actually suits

The honest summary is that telematics rewards a driving profile, not a driving standard.

It suits low-mileage drivers, drivers who avoid late nights, drivers with predictable open-road commutes, and drivers who genuinely leave the phone alone. For that group the saving can be real and worth having.

It suits urban drivers, shift workers, and anyone in stop-start traffic considerably less well — not because they drive worse, but because the proxies penalise their conditions.

If your profile is uncertain, the sequence that costs least is to check the simpler things first. Confirm the reported mileage and use classification are accurate, since that is a rating input you can correct with no monitoring at all — the method is in mileage bands and the commute you no longer make. Then compare identical coverage across several carriers, because the variation between rating plans is frequently larger than anything a telematics score will produce; the underlying mechanics are in how carriers price you.

Enrol if the terms are discount-only and your profile fits. Read carefully if they are not.