Methodology

Last updated 2026-07-01. Dataset: DVSA MOT testing data, release(s) 2021,2022,2023,2024,2025.

Where the data comes from

All figures are derived from the DVSA anonymised MOT testing dataset, published on data.gov.uk under the Open Government Licence v3.0. It records millions of MOT tests, including make, model, year of first use, fuel type, recorded mileage, the result, and the specific defects ("reasons for rejection") and advisories noted by testers. We use class 4 tests (cars and small vehicles). This site is not affiliated with or endorsed by DVSA or DVLA.

What the MOT does and doesn't measure — read this first

The MOT is a roadworthiness test of testable defects: tyres, brakes, steering, suspension, lighting and signalling, visibility, bodywork/structure and corrosion, seat belts, and exhaust emissions. It does not assess engines, gearboxes, turbos, clutches, infotainment or general electrical reliability. A high MOT pass rate is not proof that a car is reliable overall — only that it tends to pass the testable checks. Treat these figures as one input, not a verdict.

How we clean and group makes and models

Raw records contain inconsistent spellings and trim names mixed into model fields. We normalise spelling and case, map manufacturer aliases to a single name (e.g. VW → Volkswagen), reduce models to a base model (e.g. Fiesta Zetec → Fiesta) so years are comparable, and assign each model a segment so "similar cars" and comparisons are like-for-like. These rules are deterministic and applied identically on every refresh, so groupings and page URLs stay stable.

How we calculate the headline figures

Pass rate = (passes + passes after rectification) ÷ total tests.
Fail rate = outright failures ÷ total tests.
vs average compares each model-year against cars of a similar age in the same segment — not against all cars of all ages, which would mislead.
Defect category share is the percentage of tests with at least one failure in that category, grouped from the detailed reasons for rejection.
Mileage bands and fuel splits are shown only where each group has enough tests.

How we handle low-sample and thin pages

Small samples give unreliable percentages, so model-year pages with fewer than 250 tests are not indexed; any breakdown cell below a minimum size is hidden; a model-year that is statistically indistinguishable from the model's overall figures points to the model page rather than duplicating it; and comparisons between cars that are not meaningfully different are not indexed. We would rather show fewer, more reliable pages than many weak ones.

How often the data is refreshed

DVSA publishes the dataset annually. We re-process it on each new release and update the "data last updated" date shown on every page.

Limitations

It is aggregate data: individual cars and maintenance histories vary widely.
Mileage is self-recorded at test and occasionally implausible; we filter extremes.
Model grouping merges trims and engine variants.
Regional, usage and demographic factors are not controlled for.
The dataset only reflects vehicles that reach test age.

Corrections & contact

Spotted something wrong, or want a model grouped differently? Email [email protected]. We log corrections and apply them on the next refresh.