How we source spec data
Every figure we publish comes from a document we can name. This page explains which documents, in what order of authority, and what we do when they disagree.
- Effective date
- 1 October 2026
- Last updated
- 1 October 2026
- Editorial team
- 7
- Market covered
- United States
- Report an error
- contact@featureandfit.com
Most vehicle specifications on the internet have been copied from a site that copied them from another site. Somewhere in that chain the configuration fell off, a figure got rounded, or a mid-year revision was never picked up. The number survives; the context that made it meaningful does not.
We work the other way round. Every table on this site is rebuilt from the original document, tied back to the configuration it describes, and checked by a second person before it goes live. This page documents that process so you can judge it rather than take our word for it.
1. Our source hierarchy
Sources are ranked. A figure from a higher tier always takes precedence over one from a lower tier, and we record which tier every published figure came from.
Manufacturer primary documentation
The first place we look and, for most specification figures, the last. These are the documents the manufacturer produces for its own dealers and press:
- Dealer order guides and ordering documentation for the specific model year
- Published towing and payload guides
- Press kit specification sheets from sources such as the Ford Media Center
- Owner’s manuals and supplements
- Technical service bulletins and published revisions
United States federal and regulatory records
Independent of the manufacturer, published on a different schedule, and often the first place a mid-year change becomes visible:
- FuelEconomy.gov for EPA-rated fuel economy and emissions by engine and drivetrain
- NHTSA for crash test ratings, recall records, safety complaints and VIN decoding
- Federal Motor Vehicle Safety Standards where a requirement affects a published figure
Independent testing and standards bodies
Used to explain what a figure means and how it was measured, rather than to source the figure itself:
- IIHS for crashworthiness and crash avoidance evaluations
- SAE International for test procedures, including SAE J2807 for tow ratings
Vehicle-specific documents supplied by readers
Window stickers, build sheets and door jamb labels sent to us by owners. These are authoritative for that one vehicle and nothing else, so we use them to check our tables and to identify gaps, never as the source for a general figure.
Everything else
Nothing outside tiers 1 to 4 becomes a published figure. See the next section.
2. Sources we never use
Being clear about what we exclude is as useful as listing what we include. No figure on this site comes from:
- Specification aggregator sites, comparison databases or car listing portals
- Other automotive publications and blogs, including large and well-known ones
- Enthusiast forums, social media posts or comment threads
- Dealer listings, classified adverts or auction descriptions
- Wikipedia or other user-edited encyclopaedias
- Generative AI tools, which cannot cite a document and routinely invent plausible figures
- Our own memory of a previous model year, however similar it looks
Some of these are useful for finding out that a figure exists, or that readers are confused about something. None of them is ever the source we publish from. If a forum thread points us at a discrepancy, we go and find the primary document before we write anything.
3. The four-step process
The same four steps run on every article, without exception.
Pull the primary document
We locate the manufacturer’s own published material for that exact model year and identify which document governs the figures we need. If we cannot find a tier 1 document for a claim, we do not proceed to write the claim.
Cross-check against federal records
Fuel economy, emissions class, safety ratings and recall history are verified independently against United States government databases. Where a mid-year revision exists, this is usually where it surfaces first, because federal data is republished on its own schedule rather than the manufacturer’s.
Rebuild the table by configuration
Every figure is re-entered against the specific build it belongs to. Nothing is averaged across trims, carried over from a neighbouring configuration, or inferred from a pattern. If the source gives twelve rows, we publish twelve rows.
Review before publishing
A second member of the team checks every number in the finished article against the source document, working from the document rather than from the draft. Any figure that cannot be confirmed is removed and marked as unavailable. The reviewer is never the person who wrote the article.
This is slower than aggregating. It is the reason we publish fewer articles than sites of comparable age, and the reason we are willing to leave cells in a table empty.
4. When sources disagree
Sources conflict more often than you would expect. These are the rules we apply, in order.
- Two manufacturer documents disagree The later publication wins. If both carry the same date, the more specific document wins — an order guide over a press release.
- Manufacturer and federal data disagree For fuel economy and emissions, EPA data wins, because EPA figures are the rated values. For capacities and dimensions, the manufacturer wins.
- A safety rating differs between NHTSA and IIHS Both are published side by side. They test different things and neither is a correction of the other.
- A reader’s build sheet contradicts our table We treat it as a signal that a configuration is missing or mis-scoped, return to the primary document, and correct our table if it is wrong.
- The conflict cannot be resolved We publish both figures, name the document each came from, and say plainly that the sources disagree rather than picking one silently.
We never split the difference between two conflicting figures, and we never publish an average. A number that appears in no source document is not data, whatever arithmetic produced it.
5. Tying figures to configurations
A specification without its configuration is close to useless. Every figure in our tables is tied to as many of these as the source document specifies:
Where the source document does not break a figure down this finely, we say so rather than implying a precision the document does not have.
6. Units, rounding and precision
Our readers are in the United States, so US customary units lead throughout.
We reproduce the precision of the source and do not add to it. If a document says 13,500 pounds, we publish 13,500 pounds — we do not convert it, round it to 13,000, or present it as “about 13,500”. Converted values are marked as conversions so you always know which figure was the rated one.
7. Model years and mid-year revisions
A model year is not a fixed object. Manufacturers revise ratings, change standard equipment and re-scope packages partway through production, sometimes without any announcement.
Where we can establish that a figure changed mid-year, we publish both values with the revision documented. Where we can see that a figure changed but cannot establish when, we say that plainly rather than guessing a date. Where a model year carried over unchanged from the previous one, we confirm that against the new year’s document rather than assuming it.
8. What we will not publish
An empty cell in one of our tables is a deliberate choice, not an oversight.
- Estimated figures. If a document does not state it, we do not calculate it and present it as a specification.
- Averaged figures. A number spanning several configurations describes none of them.
- Carried-over figures. Last year’s number is not this year’s number until a document says it is.
- Interpolated figures. If a document gives ratings for two axle ratios and not a third, the third stays empty.
- Unattributable figures. If we cannot name the document, it does not go in.
When a figure is missing we distinguish between not documented, meaning the configuration exists but no published source gives the number, and not applicable, meaning the configuration cannot be ordered. These are different facts and we label them differently.
9. How we show our sources
Articles state which document type a table was built from and which model year it covers. Where a figure comes from a federal database rather than the manufacturer, that is noted against the figure. Where sources disagree, both are shown with their origins named.
We link to primary sources wherever they are publicly reachable. Some manufacturer order guides are distributed to dealers and are not published at a stable public address; in those cases we name the document and its date rather than linking to a copy we cannot vouch for.
10. Review and update cycle
11. Who does this work
Feature and Fit is produced by an editorial team of seven, working from Risali in Chhattisgarh, India, for the United States market. Research, verification and editing are handled by different people, which is what makes step 4 meaningful — the reviewer is never checking their own work.
Our editorial team works from published documentation rather than road experience. We do not write driving impressions, because we would be inventing them. What we do is documentary: find the record, verify it, and present it accurately. That is a narrower claim than a road-test publication makes, and it is one we can support on every figure.
Editorial standards are set by Sameer Patel, who reviews specification tables before release. More about the publication is on our About page, and our bylines, independence and funding are covered in our Editorial policy.
12. What we do not do
The limits of how this data should be used, particularly for towing and loading, are set out in our Disclaimer. Please read it before relying on any figure here.
13. Corrections
Documents get revised, packages change, and we make mistakes. A reader who has found a wrong number is doing us a favour.
Write to contact@featureandfit.com with the article URL, the figure as published, the configuration it relates to, what you believe the correct value is, and the document you are comparing against. We check every report against the primary source, correct the article, and record what changed and when.