AUCN Data: Methodology, Definitions and Limitations

Every dataset has rules, and the numbers in AUCN's data articles are only useful if you can see how they were produced. This page sets out the definitions, the calculation method, the limitations and the citation format in one place.

What the data is

AUCN publishes two families of numbers, and they answer different questions:

External figures — for example national theft and fleet statistics — are always attributed to their source in the article's Sources list, which links to the original government publication.

Definitions

Risk flags are mutually non-exclusive: one vehicle can carry finance owing and a written-off record, so segment percentages do not sum to the headline figure.

How the risk percentages are calculated

1. Each check produces a set of flags. A vehicle is counted once per check.
2. A segment percentage is the share of checks within that segment that carried the flag — for example, the share of motorcycle checks with a stolen record.
3. Segments are only published when the underlying sample is large enough to be meaningful. Smaller categories are labelled as approximate rather than dropped, so readers can still see the pattern.
4. Percentages are rounded for readability, so small differences between adjacent figures are not significant.

How the valuation percentages are calculated

• Residual curve: average advertised price by model year, expressed against the newest model year in the sample as 100%. This compares model years, not calendar ages.
• Brand retention: for each brand, the average advertised price of 2019-model vehicles as a share of the brand's 2025-model average. Only brands with sufficient volume in both years are ranked.
• Fuel-type index: average advertised price by fuel type, expressed against petrol as 100.
• Kilometre benchmark: average advertised odometer by model year, divided by the age of that model year.

All market figures are drawn from listings with a sane price and odometer range. Extreme outliers, parts listings and mispriced advertisements are excluded before averages are taken.

What the data does not measure

• It is not a random sample of the Australian fleet. Vehicles are checked because someone is buying, selling or financing them. That self-selection is a feature for buyers — it describes the cars actually being traded — but it is not a census.
• Registration of a write-off is not instant. A recent accident may not yet appear on a register at the time of a check.
• NEVDIS coverage varies. Some imported and older vehicles have thinner identity records, which is why category and brand segments note sample limitations.
• Brand figures reflect model mix as well as brand. A brand whose sales are dominated by large SUVs will look different from one selling small hatchbacks, even with identical engineering.
• Market averages hide condition. Two cars of the same year and model can differ by a wide margin once service history, damage and kilometres are accounted for.

How often it is updated

The percentages are reviewed every month against the latest check and listing data. Each article carries an "About this data" note with the month it was last refreshed, and figures are restated when a new month materially changes a segment.

Last reviewed: 28 September 2026.

How to cite AUCN data

AUCN Data (2026), article title, aucn.net.au, accessed month/year.

When you cite a figure, please include the month it was published so readers know the vintage of the number. For media enquiries, data requests or corrections, contact AUCN through the AUCN website.

Corrections

If a published figure is wrong, AUCN will correct it in place and note the change in the article. Corrections take priority over consistency with earlier reporting.