Telematics was going to reprice motor insurance around behaviour. A decade on, the results are real but selective — and the lessons matter for every data-rich distribution channel.
Ten-plus years ago, telematics was pitched as the end of motor insurance as we knew it: price the driver, not the demographic. The dongle gave way to the smartphone, the smartphone to the connected car, and market analysts now track usage-based insurance as a multi-tens-of-billions market growing at double digits. So did behaviour-based pricing win? The honest answer: it won specific battles, lost others, and the pattern of which is which is the most useful market intelligence in motor insurance.
What actually worked
Self-selection did more work than the sensor. The customers who opt into monitoring are disproportionately the careful ones — programmes profit from who joins as much as from how they drive. Insurers learned to love this quietly: the discount attracts a better book even before the first mile is scored.
Claims context, not just pricing. Crash detection, first-notice-of-loss triggered by the event itself, reconstruction data that resolves liability disputes — the claims-side value of driving data proved at least as durable as the pricing-side story.
Commercial fleets. Where the vehicle owner is a business, monitoring is a management tool rather than a privacy intrusion, and telematics-based fleet cover became standard practice with far less friction than personal lines ever managed.
What underdelivered
Mass-market behaviour change. The vision of drivers coaching themselves to lower premiums met consumer reality: surveys consistently show a stubborn share of drivers refuse monitoring at any discount, and privacy sentiment moves with every data-sharing news cycle. Participation grew — especially where money is tight and discounts are meaningful — but the fully-telematic personal motor market did not arrive.
Data without a use. More than one insurer collected billions of miles of driving data before working out which decisions it would change. Storage is cheap; a signal that never alters a price, a claim or a renewal is just liability with a dashboard.
The deeper lesson: context beats surveillance
Read the decade carefully and the winning pattern is not "monitor everything." It is: use data the customer already generates for their own reasons, at the moment it is decision-relevant. Fleet owners already track vehicles; the insurance rides along. The connected car already reports the crash; the claim opens itself. That framing — data in context, at the moment of need — is exactly the thesis of embedded distribution, applied to pricing and claims instead of sales.
What this means for embedded motor
Embedded motor journeys start with an advantage telematics spent a decade chasing: verified, structured truth about the vehicle and the buyer, present at the exact moment of sale, supplied willingly because it serves the purchase itself. No dongle, no monitoring consent, no selection fight. The strategic question for the next five years is whether connected-car data joins that moment — a panel pricing a policy against the manufacturer's own vehicle data at checkout would combine the best of both stories. The insurers that navigate it well will be the ones that learned telematics' real lesson: customers trade data for value at a moment of need, and resent trading it for surveillance.