Does Telematics Work? Twenty Years of Evidence From 35 Million Policies
The insurance industry loves a good tech story. We've watched wave after wave of Insurtechs promise to revolutionise everything from distribution to underwriting. Most deliver incremental improvements dressed up as transformation.
Telematics is different. The unglamorous business of putting black boxes in cars, or giving an app to customers to monitor how they actually drive, has quietly done something most innovations only talk about: delivered tangible, measurable results.
After twenty years and nearly 35 million policies globally, we can answer the question properly.
Does telematics work? Yes, it does. But not everywhere, and definitely not the way most sales pitches suggest.
Three Markets. Three Very Different Stories.
The global telematics picture is best understood by looking at three markets that took fundamentally different approaches: Italy, which placed the portfolio bet on it; the United Kingdom, which introduced it as a niche solution; and the United States, which has scaled volume at the expense of depth. Each tells us something different about what telematics can and can’t do.
Italy: Where Telematics Actually Worked
Italy hasn't just adopted telematics. They've made it boring, which is the biggest compliment you can pay any insurance innovation.
Around 9 million policies have been sold—roughly a quarter of their entire motor market. UnipolSai alone runs 4.5 million black boxes — half the country's telematics market. These aren't pilot programmes or niche segments. This is how Italian motor insurance works now.
The Numbers
UnipolSai's five-year average combined ratio sits at 92%. In 2020, they hit 85%. To understand what that means: the average European motor insurer struggles to get below 95%. Claims frequency runs 20% lower than traditional policies on a risk-adjusted basis. Young drivers — the segment that usually destroys a motor book’s profitability — show a material improvement with telematics. Frequency drops 10%, claim costs fell 14%, and combined ratio improvements have reached 12-13 percentage points.
These aren't rounding errors. These are the numbers that change P&L statements, board conversations, and strategic plans.
Fraud and theft detection have been equally significant. Italy cut vehicle theft by 50% through telematics-enabled recovery systems. Claims handling costs dropped 55% as remote diagnostics replaced much of the manual investigation process.
Why Italy Got There
Several things aligned that didn't elsewhere.
High fraud and theft rates created genuine regulatory pressure to do something different. Concentrated agency networks could distribute hardware at scale without the distribution headaches that plagued UK and US implementations. Premium variation was extreme — €565 in Naples versus €196 in Bolzano — so the value exchange was obvious to customers who understood they were being priced on postcode rather than behaviour.
But the most important factor was a strategic commitment. UnipolSai didn't test the waters. They built an entire ecosystem. Roadside assistance integration. Claims processing redesign. Fraud detection infrastructure. Toll integration. Stolen vehicle recovery. Telematics became the infrastructure, not a product feature.
That distinction matters more than any other single factor in explaining why Italy succeeded while others stalled.
United Kingdom: Technically Successful, Commercially Contained
Britain took a different path, and the results reflect it.
About 1.5 million policies. That’s around 3-5% market penetration after fifteen years of effort. The industry made a strategic choice to position telematics as the young driver solution rather than a mainstream pricing and underwriting tool.
That choice did, however, have some logic. Young drivers are the highest-risk segment. By introducing telematics, insurers could, in theory, prove individual driver behaviour and price accordingly. With telematics, young drivers can save over £1,000 a year on average. The value exchange was clear: drivers are willing to accept monitoring and give their data in exchange for affordability.
But the framing created a ceiling. Telematics has become associated with young drivers, black boxes, and restricted driving hours. It never escaped the perception of being a product for people who couldn't get standard insurance. Mainstream adoption has never really followed.
The lesson from the UK isn't that telematics failed. It's that positioning determines scale. A niche solution stays niche.
What the UK Got Right
Where the UK excelled was in demonstrating that telematics could change driving behaviour, not just price it. Young drivers on telematics programmes show measurably different driving patterns — less night driving, lower speeds, smoother braking. The feedback loop works.
The UK also developed sophisticated scoring models that the rest of the market is still learning from. The actuarial and data science capability built around UK telematics programmes has influenced how carriers everywhere think about behavioural pricing.
United States: Scale Without Depth
America has the numbers, the scale. 20 million policies, heading towards 26 million by 2029. But that's still only 6-8% of the market, and the US approach diverged from Europe's in one critical way: they largely abandoned hardware for smartphone apps.
Progressive's Snapshot programme. State Farm's Drive Safe & Save. Allstate's Drivewise. All smartphone-based. All optimising for adoption over data quality.
Whether phone-based telematics delivers the same loss ratio improvements as Italy's hardware approach remains genuinely unclear. Accelerometer data from smartphones introduces noise that dedicated OBD hardware doesn't. Location accuracy varies. Battery optimisation on modern smartphones means data isn't always captured consistently.
The US market has prioritised reach over rigour. That may prove to have been the right trade-off at scale — the data is still coming in — but it means the US cannot yet point to Italy-equivalent combined ratio improvements.
Why This Should Work At All: The Human Behaviour Argument
Before examining implementation, it's worth being clear about why telematics works when it works.
You've heard that 94% of crashes are caused by human error. This figure comes from a 2008 US NHTSA study, and it's routinely misquoted. What NHTSA actually found: in 94% of cases, a human action was "the last failure in the causal chain of events leading up to the crash." Their report explicitly says this shouldn't be "interpreted as the cause of the crash nor as the assignment of fault."
There is, however, more precise evidence, albeit from quite a few years ago. The 1979 Tri-Level Study found human factors were a definite or probable cause in 90-93% of incidents. Distracted driving accounts for 8% of fatal crashes and 13% of injury crashes according to NHTSA's current data. Driver inattention more broadly appears as a factor in approximately 80% of all accidents.
The case for telematics doesn't rest on a single misquoted statistic. It rests on consistent evidence across multiple studies spanning decades: human behaviour is the dominant variable in motor accident causation. If you can measure behaviour objectively and price it accurately, you can underwrite motor insurance better than anyone using proxies: postcode, age, use and so on.
LexisNexis data demonstrates this clearly. Drivers with the worst UBI scores produce loss ratios around 135%. The best scores? 38%. That's not a different rating class. That's a different business.
What "Working" Actually Means in Practice
For Carriers
The evidence from mature telematics programmes points consistently to:
Claims frequency reduction of 10-20% versus traditional policies on equivalent risk
Average claim cost reduction of approximately 14%
Combined ratio improvements of 15-20 points for carriers with genuine portfolio commitment
Fraud detection that creates material bottom-line improvement (Italy's 50% theft reduction)
Better risk selection that compounds over time as high-risk drivers self-select out
For Policyholders
The fairness argument is underappreciated in carrier discussions of telematics economics.
A 19-year-old who drives carefully — low speeds, smooth braking, daylight hours, modest mileage — shouldn't subsidise a 45-year-old who regularly exceeds the speed limit whilst checking their phone. Traditional rating cannot distinguish between them with any precision. Telematics can.
The feedback loop matters too. Drivers who can see their driving scores in real time drive differently. Italy's data showing 30% reduction in accident risk isn't just a pricing improvement. It's fewer crashes actually happening. That has value beyond the insurer's loss ratio.
For the Market
Reduced accidents mean reduced claims volumes across the industry. Better priced risks mean more sustainable premiums. Carriers who build genuine telematics capability create competitive advantages that compound as data accumulates.
The Implementation Requirements No One Talks About Enough
Telematics works. The evidence is unambiguous on that point. But it doesn't work as a bolt-on.
Carriers that have achieved Italy-equivalent results invested seriously in:
Data infrastructure. Raw telematics data is meaningless without the processing capability to turn it into actionable pricing variables. This requires data science investment that most motor insurers haven't historically needed.
Pricing model reconstruction. You cannot simply add a telematics score to an existing rating structure and expect Italy's results. The pricing model needs to be rebuilt around behavioural data as the primary variable.
Claims process redesign. The 55% claims handling cost reduction Italy achieved came from fundamentally changing how claims are investigated and processed. Telematics data changes the first notification of loss, liability assessment, and fraud investigation simultaneously.
Customer communication. The value exchange needs to be transparent and credible. Customers accept monitoring when the benefit is clear and the discount is real. Italy's high penetration reflects trust in the value exchange. That trust is built deliberately, not assumed.
Three-year horizon. The carriers that have abandoned telematics programmes typically did so because they expected results within eighteen months. Italy's ecosystem took years to build. The data quality that drives UnipolSai's combined ratio took years to accumulate.
The Verdict
Does telematics work? Yes — if "work" means materially improving loss ratios through better risk selection and genuine behavioural change.
Italy provides fifteen years of evidence across millions of policies. Combined ratios improve 12-15 points. Claims frequency drops 10-20%. Young driver expected losses improve 23%. These aren't projections or pilots. These are audited financials from carriers maintaining 85-92% combined ratios in a mature, competitive market.
Britain shows telematics can be technically successful and commercially contained simultaneously. The data science capability is world-class. The market penetration reflects strategic choices, not technical limitations.
America shows that optimising for adoption at the expense of data quality has consequences that take years to fully understand.
The fundamental insight hasn't changed since the first black box went into a car: 90%+ of motor accidents involve human behaviour as a contributing factor. Telematics gives us objective, granular data on that behaviour. Properly implemented, it lets carriers price risk accurately, reward safe drivers, charge risky drivers appropriately, and create feedback loops that reduce accidents.
Twenty years ago, putting tracking devices in customer vehicles felt like overreach. Today, Italian drivers queue up for black boxes because the value exchange is transparent and the savings are real.
That's not marketing. That's mathematics applied to human behaviour at scale.
The question was never whether telematics works. The question is whether you're willing to do the work required to make it work for you.
