Who writes GEOhandbook

Peter Adams, Editor

Peter Adams is a senior web developer with more than 20 years of experience building websites. He has long treated quality content and technical SEO as part of the job, and now applies the same discipline to generative engine optimisation. He writes and maintains GEOhandbook. Every claim on the site links to primary documentation or published research rather than to commentary repeating it, and the site is audited against its own checklist, including the items it currently fails, which are listed openly rather than quietly dropped.

Verify this byline elsewhere:

These are the same URLs given as Person.sameAs in this page's structured data. rel="me" states the same claim in the markup: that the profile and this byline are one person. The claim is only worth something if it holds in both directions, so each profile should link back here.

How claims get sourced

Every non-obvious factual claim on this site links inline to a primary source — the organisation that published the data, not an aggregator repeating it. In practice that means:

Vendor documentation over commentary. When the question is how OAI-SearchBot treats robots.txt, the source is OpenAI’s own crawler documentation, not a blog post summarising it.
Named studies with their numbers intact. Where a figure appears, the figure and its source are both given, so a claim can be checked rather than taken on trust.
Primary research for mechanism claims. Assertions about how retrieval actually behaves cite the papers that measured it. The paragraph-structure rules in the checklist, for instance, cite the retrieval benchmarks they come from rather than repeating a number in circulation.
Dated sources. GEO moves quickly enough that an undated claim is close to worthless. Where a source has a date, it is stated.

If a page makes a claim with no link, it is either self-evident, an opinion clearly framed as one, or a mistake worth reporting.

How articles are made

Articles here are researched and drafted with the help of AI tools. That is stated on this page, and linked from the end of every article, because a reader weighing a claim is entitled to know how it was produced, and because Google’s own guidance on AI-assisted content recommends exactly this: “Sharing information about how a piece of content was created can help give your readers more context” (Google Search Central).

What that means in practice:

The topic and the argument come from the editor. Each article starts from a specific question a marketer or developer actually has to answer, not from a keyword list.
AI does the legwork, not the judgement. The tools are used to find candidate sources, to draft, and to restructure. They are not treated as a source. A claim is never published because a model said it.
Every cited source is opened and read. Before a page ships, each linked source is fetched and the claim next to it is checked against what the source actually says. Where a figure could not be traced to its origin, it is removed rather than softened.
The editor is accountable. Peter Adams decides what is published and is responsible for every page, however it was drafted. Mistakes are the editor’s, and are corrected as described below.

What AI assistance does not replace is first-hand evidence: tests run on real sites, numbers from real Search Console accounts, results from real client work. That is the kind of material that makes a page worth reading when the documentation it cites is available to everyone. Articles are being extended with it, and an article that includes original testing says so and shows the method.

What this site will not do

No affiliate links, paid placements, or sponsored posts. Tool mentions are editorial judgements. The full position is in the disclosure.
No invented statistics, and no round numbers presented as research findings.
No claims about rankings or citation rates that cannot be traced to something measurable.

Corrections

Errors get fixed on the page, not quietly deleted. If a correction changes the substance of a claim rather than its phrasing, the page says so.

To report one, email hello@geohandbook.io with the URL and what is wrong. Specific corrections with a source attached get handled fastest.

How pages stay current

Every page shows an Updated date whenever its source has actually changed since publication, and that date is derived from the site’s version history rather than typed in by hand — so it cannot be inflated to fake freshness. Pages that have never been edited show only their publication date, deliberately.

The same dates feed the sitemap, which means the date a reader sees and the date a crawler is told are always the same value from the same source.

Practising what the checklist preaches

This site is audited against its own checklist, and the failures are treated as a work queue rather than quietly dropped. Some items are still open — off-site authority signals in particular, and this named-author gap. Where the site does not yet meet its own standard, the honest position is to say so on the page rather than to soften the standard.