Editorial policy

I do. I pull the data, write up the findings and publish them.

I sell SEO services, so I gain when readers trust SEO data. Where a guide lists my own business, it says so next to the listing. The commercial disclosure in my UK SEO companies guide is an example.

An original study uses data I collected from my own Search Console and GA4 properties, my own Ahrefs pulls or my own checks of public pages. The research hub lists every one.

A roundup collects figures that other organisations measured. Each figure names who measured it, when and on what sample. When an original study quotes someone else's benchmark, it names the source and date beside it. The data catalogue marks files I compiled and does not present them as my own measurements.

The roundups sit in the statistics library.

No provider has paid to be listed in a guide on this site. Where a listing came from a swap or a relationship, the guide says so.

If one ever did, the listing would read Paid inclusion and its links would be marked sponsored. My AI search agencies guide already supports that label. I would not publish a paid place without it.

A guide that lists my own business is an editorial choice and not a recommendation I have been paid for. It is labelled as mine.

I fix the page, add a dated note and leave the old figure visible beside the new one.

If I cannot trace a figure to its source, I remove it. I do not round it or soften it. When I withdraw a study, I keep it online, mark it withdrawn and give the reasons.

If you spot an error, tell me and I will check it. Past corrections are listed in the methodology.

Every page carries a published date, and an updated date when I have re-checked its figures.

An updated date moves only after the figures were re-checked at the source. A weekly script lists any content page whose review date is older than 25 days. Each study states its data window and the day I pulled the data. Data files carry the date or edition in the file name.

The update tracker is a recurring study. I add a row about four weeks after Google marks an update complete, so the data has settled.

I do not publish client data that identifies a client without permission. One study, on multimodal search, includes anonymised figures from client properties, described by type only.

Client sites are not in the 100-site Search Console sample or in the GA4 frame.

Yes, with credit. The files use CC BY 4.0. Keep the window and sample with every figure you quote, and say whether the figure comes from my sites or from a source I compiled.

Each file has suggested citation text in the data catalogue.