Sources and corrections
If something here is wrong, I want to know, and the fix gets published where you can see it.
Reporting a mistake
Email elin@rankpedia.org. Tell me which entry, which sentence, and — if you have one — a source. You do not need to be polite about it.
What happens next: I check it against primary documentation. If you are right, the entry is corrected, its last-checked date is updated, and the correction is added to the changelog with the date and what changed. If I think you are wrong, I will tell you why. If it turns out to be genuinely disputed, the entry gets a line saying it is disputed, which is usually the most accurate outcome anyway.
Corrections are never quietly overwritten. A reference site that edits its history is not a reference site.
What counts as authoritative here
In rough order of weight:
- Search engine documentation. Google Search Central, Google's Search Quality Rater Guidelines, Bing Webmaster documentation. These describe intended behaviour, which is not always actual behaviour, but they are the only first-party statements that exist.
- Standards bodies. W3C, WHATWG, IETF RFCs, schema.org. For anything about HTML, HTTP, structured data or accessibility, the standard is the source and everyone else is a commentary on it.
- Vendor documentation, for entries about that vendor's product. How a tool calculates its own metric is a fact only the vendor holds.
- Published research and large-scale studies, cited directly rather than through someone's summary of them.
- Consistent, documented practitioner observation, clearly labelled as observation. Plenty of real search behaviour is undocumented; where the field has converged on an explanation through testing, that is worth reporting, as long as the entry says that is what it is.
What does not count
Recycled listicles, agency blog posts that cite other agency blog posts, conference-slide screenshots without a primary reference, and anything whose only support is that it has been repeated for a decade.
A specific case worth naming: quoted "ranking factor" percentages. Numbers of the shape "backlinks are 31% of the algorithm" have no verifiable origin, and where an entry mentions one it is to explain where the number came from and why it does not mean what it appears to mean.
Outbound links
All external links carry nofollow and open in a new tab. This is a deliberate policy rather than an oversight: it means nothing can be gained by persuading me to cite a page, so citations stay about whether a source is the best available.
For the same reason, no site can pay to be cited, no guest posts are accepted, and link-exchange emails get no reply.
Corrections policy in one paragraph
Factual errors are fixed as soon as they are confirmed and logged publicly. Wording improvements and clarifications are made continuously and are not logged individually. Entries that turn out to be substantially wrong are rewritten rather than patched, and the changelog says so. Nothing is deleted to hide a mistake; if an entry is retired, the reason is stated and the URL keeps working.
Independence
This site takes no advertising, no sponsorship, no affiliate revenue and no payment of any kind for coverage. It is funded by unrelated consulting work, described on the about page. No client of mine is named in any entry, and no entry has been written or altered to benefit one.
Entries about commercial tools exist because people search for them and deserve a neutral explanation. They are descriptions, not recommendations, and no vendor has been given a preview or a right of reply beyond the same correction process as everyone else.