How Our Data Is Built: Methodology

In short. We collect registration and trade records from official publications and licensed or partner sources, map every make, model and variant to one naming list, band engine capacity the same way in every market, check totals against published industry figures where they exist, and tell you the basis and the lag of every data set.

1. Source types

Sources differ by market. We state the source type for each data set in the sample you receive.

Source types by market
Market	Source type	How we obtain it	What we state with the data
India	Official vehicle registration records	Authorised or licensed access to registration records, aggregated before delivery	Basis, period covered, latest month and lag
China (exports)	Customs and industry export statistics	Public releases and licensed trade data	Basis (customs or industry body), period and coverage of destinations
USA and Europe	National and industry registration statistics	Public releases from official bodies and industry associations, and licensed partner data	Whether the figure is new registrations or registered stock
Africa and ASEAN	National registration or import statistics, customs records, industry association sales	Public releases, customs data, association reports and licensed partner data	Whether the figure is registrations, imports or association sales, and which countries are totals only

Some markets publish only totals, or only import statistics. Where that is the case we say so on the coverage page and in the sample, and we do not present totals as model-level data.

2. Standardisation

One naming list. Make, model and variant spellings from every source are mapped to a single name, so the same model is not counted under two spellings.
Engine capacity. We keep exact CC where the source gives it and also add a band, using the same breaks as the customs classification for motorcycles: up to 50 cc, over 50 to 250, over 250 to 500, over 500 to 800 and over 800.
Powertrain and segment. Each record is assigned a powertrain (petrol, diesel, CNG or LPG, hybrid, electric) and a segment (two-wheeler, three-wheeler, passenger car, bus, truck and others) on a fixed definition list. See the data dictionary.
PIN codes (India). A PIN code is checked to be six digits and to match a known postal area. Records that do not match are flagged, not silently dropped.
De-duplication and outliers. Duplicate records are removed and unusual values are reviewed against the source before they are kept.

3. Quality checks

Monthly totals are compared with published industry or official totals where they exist (for example retail registrations from dealer associations or factory dispatches from manufacturer associations), and differences are explained by definition, not hidden.
The top models in each segment are spot-checked against the source each release.
Month-to-month movements outside a normal range are reviewed before delivery.
Issues found and corrected are listed in the release notes that accompany each delivery.

4. Refresh, lag and revisions

Refresh frequency and the lag between a month ending and the data being available depend on the market and are stated for each data set. If a source revises its figures, we reissue the affected months and note the change.

5. What the figures measure

Registrations, retail sales, factory dispatches and registered stock are different measures. We always state which one a number is. Our guide Registrations vs sales vs exports explains the differences.

6. Known limitations

A registration can lag the sale or delivery by days or weeks.
Fields available differ by country. PIN code and colour are for India.
Some segments are reported as combined categories in some countries.
We confirm coverage for your countries before you buy, rather than assume it.

Questions

Write to info@autodatainsights.com or use the contact form.

How our data is built