Skip to main content

Last updated: 2026-05-24 · Written by Bertram Sargla, founder

What Carlytics knows — and what it doesn't

Most VIN-check companies will tell you they are the most accurate, the most complete, the largest in Europe. I am not going to do that. The honest version is more useful: here is exactly what each Carlytics report does, where we have strong data, where we have weak data, and when you should not pay us EUR 8.90 because we will not help you. If after reading this you still want a report, you will know what you are buying. If you decide we are not the right tool for your situation, you have saved yourself a transaction.

Run a free decode first, then read on:

1. What we know

A Carlytics report stitches together evidence from several independent layers. Each layer answers a different question, and each layer can succeed or fail independently of the others.

National vehicle registries. Across most of the European Union, government bodies keep records of registered vehicles, periodic technical inspections, and ownership changes. Some publish this data openly; some make it accessible to authorised processors; some keep it private. Where we have access to a country's registry, the report can show mileage timelines from inspection records, body and engine details that match the registration entry, and damage or write-off events when the registry publishes them. The depth varies by country. We do not name the specific registries on customer-facing pages because the data is licensed and the agreements evolve, but the principle is straightforward: we use what governments choose to publish.

Periodic technical inspections. In countries with strong mandatory inspection regimes — Germany, Czech Republic, the Netherlands, the United Kingdom, the Nordic states — every car is inspected on a one- or two-year cycle, and the mileage is recorded at each inspection. This turns into a multi-year timeline showing odometer movement, which is the single best evidence against mileage rollback. A car with eight years of monotonically-increasing inspection readings is hard to fake.

Stolen-vehicle databases. Police and Interpol-aligned stolen-vehicle records are queried for each report. This check answers a yes/no question against the 17-character VIN. It does not depend on whether we correctly identified the engine displacement — the lookup is on the raw VIN.

Safety recalls. EU and US regulators publish manufacturer safety-recall campaigns that identify VIN ranges. We match your VIN, model, and production year against the open campaigns on file. If the report shows no recalls, no campaign matched. If a campaign matched, you see it.

Market signals. Where we observe public listing prices for vehicles of the same year, model, mileage band, and country, the report shows a market value estimate. This is a statistical signal, not a transaction record, and we label it as such.

VIN decoding. The part most customers see first — turning a 17-character string into "2018 Volkswagen Golf 1.6 TDI" — uses manufacturer-specific decoding logic layered on top of regulatory VIN structure. It is also the part most likely to be partly wrong on any given VIN, for reasons explained below.

2. What we don't know

Some gaps are structural. We could not fix them by trying harder; the data does not exist in a form we can access. Being clear about this is more useful than pretending otherwise.

Country-by-country coverage is uneven. Mileage timelines are very strong for cars inspected in Germany, the Netherlands, the United Kingdom, Czech Republic and most Nordic countries. They are thinner in southern and south-eastern Europe where periodic-inspection records are less openly published. Accident and damage events are well-covered in some markets and almost invisible in others. If your VIN is from a country whose national registry does not publish inspection or damage data, the corresponding sections of the report will be sparse. The section still renders — we put a dash where the data does not exist — but a dash is honest and a fabricated reading would not be.

Some VIN families are deliberately ambiguous. Two different trim levels can share the same character at position 7. Model-year codes recycle on a 30-year cycle. Manufacturers reuse plant codes across decades. When our decoder cannot be certain which interpretation is correct, our policy is to abstain — show a production-year range rather than a confidently wrong single year, or leave a field blank rather than guess. A null year is better than a wrong year. A blank trim line is better than the wrong trim.

Live battery health and live FSD status are not observable. For electric vehicles we can report on chemistry (which varies by plant of origin), warranty terms, and typical degradation curves for the model and year. We cannot tell you the actual current state-of-health of a specific battery without a physical diagnostic read from the car. We say so in the report rather than pretending we can.

Brand-new VINs have almost no history. A car registered in the last 60 days has not accumulated inspection records, service history, or market resale signals yet. The report will be thin not because we missed anything but because there is nothing yet to find.

Cohort fallbacks happen. For a small fraction of VINs we cannot resolve to an exact build sheet and return typical values for the model+year cohort instead. Where we do this, the field is labelled as a cohort estimate rather than a specific observation.

3. How we score confidence

Every field in a Carlytics report sits in one of three states. We chose this taxonomy deliberately because it forces a binary decision on every field: be honest, or abstain.

High confidence — the value is supported by direct evidence in a primary source: an inspection record, a registry entry, an open recall campaign matched against your VIN range. We display the value plainly.

Low confidence — the value is supported by secondary evidence (cohort statistics, VIN-pattern inference, market-listing inference) but not directly observed for your specific vehicle. We display it with a label such as "typical for this model" or as a range rather than a point estimate.

Skip — we do not have evidence either way and will not guess. The field renders as a dash, with a short note saying the data was not available for this VIN in this market. The section structure still renders. This matters: a missing field tells you something true (we do not know), whereas a fabricated field tells you something false (a number that looks observed but is not).

The reason we built it this way is plain. A wrong production year on a cover page is a worse customer experience than a null year shown as a range, because the wrong year is invisible to you until it bites you. The null is annoying but it cannot mislead you.

4. When you should not buy our report

There are situations where we will not give you what you need. We would rather you read this and not buy than buy and be disappointed.

You need a current physical condition assessment. A VIN report describes the history attached to a vehicle identifier. It cannot tell you whether the timing belt is about to snap, whether the clutch is at end of life, or whether the bodywork has been recently re-sprayed. For those questions you need a pre-purchase inspection by a mechanic. The two products are complementary; a Carlytics report is not a substitute.

The vehicle is a motorcycle, agricultural machine, or truck. Our coverage is built for passenger cars and light commercial vehicles. We have data on some other categories but it is thinner and less reliable. If the vehicle is not a passenger car or light van, expect a sparser report.

You need a legally certified document for a court proceeding or an insurance claim. A Carlytics report is a commercial summary of public and licensed data. It is not a certified document. If a court or insurer requires certified registry extracts, request them from the relevant national authority directly.

The vehicle was registered in a country where we have very thin data. Some markets simply do not publish enough for us to build a useful report. If the free decode shows we cannot identify the vehicle confidently, the paid report is unlikely to help you much either. Run the free decode first; if it looks weak, do not pay for the upgrade.

You bought the car already and are looking for validation. Most customers say a VIN report helps them before purchase, when the information can still change a decision. Some report it as useful after purchase too — for insurance verification or for reselling later. But if you are looking for reassurance about a car already in your driveway, weigh whether the EUR 8.90 actually changes anything for you. If it does, buy. If it does not, do not.

Why I wrote this page

Vehicle history products have a long history of overpromising. The category convention is to claim the largest database, the most accuracy, the best coverage. We do not get to claim those things honestly, and I would rather sell fewer reports to better-informed buyers than more reports to people who later feel misled. A report is EUR 8.90. We refund when one disappoints. The fact that we refund is not a marketing claim — it is the natural consequence of a product that tries to be honest about what it can and cannot do.

If you want to see how this looks in practice, start with the free decode. If the report looks like the right tool for your decision, unlock the paid sections. If it does not, you have not paid anything.

Related reading

Signed,
Bertram Sargla — founder, Carlytics