content/eeat-signals checks four trust categories on every page — an about-page link, an author byline, a published date, and a sources or references marker — then fires at info severity for any URL carrying fewer than 2 of the 4, the anonymity pattern Google's E-E-A-T framework has weighed against pages since its December 2022 Quality Rater Guidelines update.
Optimize your content to trigger Google AI Overviews and answer engine summaries:
content/eeat-signals scores each page against four independent trust categories and counts how many it carries. The first is an about-page link: the rule scans the page's resolved hrefs for any URL matching '/about'. The second is an author signal, satisfied if the page exposes a non-empty author meta tag, a schema.org author, a byline element, or a rel=author link. The third is a published date the parser could extract. The fourth is a 'sources' category, matched when the raw HTML contains any of five patterns: 'last updated', 'last modified', 'reviewed by', 'sources:', or 'references:'.
A page passes if it carries 2 or more of those 4 categories. Any page below that floor is flagged. The rule never inspects the quality of the byline or the accuracy of the date — it only asks whether the markers of accountability are present at all. The point is structural: a page that names nobody, dates nothing, links to no about page, and cites no source is anonymous by construction, and anonymity is the baseline condition Google's trust evaluation reads first.
E-E-A-T — Experience, Expertise, Authoritativeness, Trustworthiness — is how Google's raters decide whether a page deserves trust, and trust starts with knowing who is speaking. A page with no author, no date, and no sources gives a rater nothing to evaluate, so it defaults to the floor. This rule catches the corpora most prone to that failure: programmatically generated pages, where the template binds entity data into the body but forgets that a real publisher signs its work.
The cost is highest on Your-Money-or-Your-Life topics — health, finance, legal, safety — where Google's guidelines demand visible expertise before a page can rank. But the markers are cheap to add and the absence is conspicuous at scale: ten thousand undated, unsigned pages on one template is a clean tell that no human stood behind any of them. The rule fires at info severity because a single missing signal is guidance, not a verdict — but a whole corpus stuck below the 2-of-4 floor is a structural credibility gap that pairs badly with thin-content or near-duplicate findings on the same pages.
/guides/how-to-refinance-a-mortgage on a programmatically generated finance site. The body is 1,200 words of real advice, but there is no byline, no published or updated date, no link to an about page, and no sources or references block anywhere in the HTML. The page carries 0 of the 4 trust categories. The rule fires at info: '/guides/how-to-refinance-a-mortgage has fewer than 2 out of 4 E-E-A-T signal categories.'
The same refinance guide, reissued with accountability attached: a byline reading 'Reviewed by Dana Okafor, CFP' resolves the author category, a visible 'Last updated March 4, 2026' line satisfies both the date and the 'last updated' sources pattern, and a footer link to /about-our-editorial-team adds the about category. The page now carries 4 of 4 categories, clears the 2-of-4 floor with room to spare, and a rater can see exactly who stands behind the advice.
E-E-A-T is not a spam rule and not a ranking factor you can game — it is the conceptual frame Google's Search Quality Rater Guidelines use to describe trustworthy pages, expanded from E-A-T to add 'Experience' in the December 2022 guidelines update. The raters who apply it are not the algorithm, but their judgements train the systems that are, which is why the visible markers of accountability matter even though no single tag is a direct signal.
content/eeat-signals (in @pseolint/core, MIT-licensed at github.com/ouranos-labs/pseolint) is the gentlest expression of that frame in the suite — it fires at info severity and never blocks a verdict, because the presence of an about link or a byline proves nothing about quality on its own. Its job is to catch the structural anonymity that programmatic templates fall into: pages that carry real content in the body but none of the who-wrote-this, when, and on-what-authority markers a credible publisher attaches by reflex. When that anonymity coincides with thin or templated content, the corpus reads as mass-produced and unaccountable at the same time.
Tamsin Roark launched Ferrocite Legal Guides, a 740-page regulatory-explainer site covering building-code variances across 38 US states. pseolint flagged 612 pages at info severity: each scored only 1 of the 4 E-E-A-T trust categories — a parseable published date — but carried no about-page link in resolved hrefs, no schema.org author or byline element, and no sources or references section. Tamsin added a /about page linked from every template header, author JSON-LD naming staff attorney Declan Osei-Mensah, and a 'Sources' anchor block citing the relevant state administrative code. A second audit on 2026-02-14 confirmed 98% of pages now clear the 2-of-4 threshold.
Want to know whether this rule actually fires on your site?
Run pseolint against your sitemap. The audit is free, takes about a minute, and returns a per-URL list of every rule that fired — including this one — with the exact metric values so you can prioritise the fix queue.