Editorial policy

• Invent compatibility data. If we have not read it on a manufacturer source, it is marked unverified or it is not here.
• Claim hands-on testing we have not done. Guides state plainly whether anyone performed the procedure. Most of ours are compiled from manufacturer documentation, and they say so.
• Publish fake credentials or bylines. A guide cannot be published under a placeholder author — the build refuses it.
• Show star ratings or reviews we did not collect. We collect none, so there are none, and we emit no review or rating structured data.
• Quote a price we cannot verify. See how we verify fitment.
• Mass-produce pages. A page exists because someone has a question it answers with evidence, not because a URL pattern allows it.

1. We identify a real question with a distinct answer.
2. We collect primary sources and record each one with the date we read it.
3. We enter the structured data — parts, fitments, serial ranges.
4. A separate review pass re-reads every record against the source it cites — a different reviewer from the one that entered it, working from the manufacturer page rather than from our own notes. Until that pass completes, the record stays in review and cannot be indexed.
5. We write the original analysis: what actually decides this purchase, and what people get wrong.
6. The page publishes only if it passes the evidence gate. If it does not, it stays visible but unindexed, and the reason is recorded.
7. We re-verify on a schedule and retire what is no longer accurate.

This site is compiled with AI tools, and we would rather say so plainly than bury it. They read manufacturers’ published parts listings, record what those listings show, and draft the analysis around them.

The failure mode is specific and it matters more here than on most sites: a language model will produce a confident, plausible, wrong part number, and someone will buy it. Everything below exists because of that one sentence.

So no record is published on the say-so of whatever compiled it. A second, independent review pass re-reads each claim against the manufacturer page it cites, working from the source rather than from the first pass’s notes, and it is run by a different system from the one that wrote the record. Disagreement between the two is treated as a defect, not resolved by preference. Until that pass completes, a record stays in review status and is served noindex.

Two structural safeguards back that up. Every fitment claim carries the manufacturer URL it came from and the date that page was read, so you can check us rather than trust us. And where a manufacturer simply does not list a part for a serial range, we record that as unknown rather than as a fit or a non-fit — the gap in the evidence stays visible instead of being filled in with a guess.

Which retailers we list, and the order we list them in, is decided on usefulness to the reader — stock depth, shipping, and return terms. It is not decided by commission rate, and no retailer can pay for placement or for a better fitment verdict.

Fitment verdicts are derived from source data. There is no mechanism by which a commercial relationship could change one, and we would not build one.

Full affiliate disclosure

If you find an error, tell us and we will fix the record and update its verification date. Corrections to fitment data are a priority — a wrong fitment claim costs a reader money.

Report a correction