Google E-E-A-T

Google employs human quality raters who evaluate search results against a long published document, the Search Quality Rater Guidelines. E-E-A-T is one of its central concepts. Experience asks whether the author has first-hand experience of the subject. Expertise asks whether they have relevant knowledge or skill. Authoritativeness asks whether the creator or site is recognised as a go-to source. Trust, which the guidelines treat as the most important of the four, asks whether the page and site can be relied on.

Raters do not adjust rankings. Their scores tell Google whether a change to its systems improved results or made them worse. This is the distinction that most E-E-A-T content elides, and it has practical consequences: there is nothing to submit, nothing to score, and no mechanism by which adding an author box moves a page up.

Because Google's ranking systems are developed and evaluated against what raters reward. If raters consistently mark content from identifiable, qualified, demonstrably experienced sources as higher quality, systems that promote such content score better and survive. The relationship is indirect, slow and real.

It is also useful in a second way that has nothing to do with Google. The signals that make a page look trustworthy to a rater — a named author with a reason to be writing, a visible review process, clear sourcing, an organisation that can be identified and contacted — are the same ones that make a reader take the page seriously. That return does not depend on anyone's algorithm.

The weighting is not uniform. The guidelines ask raters to apply far more scrutiny to pages that could affect health, finances, safety or major life decisions. For a recipe, light credentials are fine; for a page about medication, a property transaction or an investment, they are not. Sites in those categories carry a materially higher bar, and in our markets that covers healthcare, property and financial services directly.

Name the author, and make the name mean something. A byline attached to a profile that explains why this person is writing about this subject is a signal; a byline attached to nothing is decoration. Where the writer is not the expert, say who reviewed it and when — which is more honest than attributing the article to an expert who did not write it.

Publish the review process rather than implying one. A short editorial policy stating how content is drafted, who edits it and who signs it off is more credible than any badge, because it is checkable. If drafts are produced with AI and edited by people, say that; the alternative is a claim that does not survive scrutiny.

Show the experience. First-hand detail — what was actually done, what the numbers were, what went wrong — is the part that cannot be synthesised from other articles on the same topic, and it is exactly what the Experience criterion asks about. Where figures are quoted, say how they were measured. Our own approach to that is on how we measure.

Make the organisation identifiable. A real address, a working phone number, named people, and consistency between what the site says and what independent sources say about the business. For Trust, the bar is low and a surprising number of sites fail it.

Not inherently. Google's published position is that it rewards helpful, reliable content produced for people, whatever the production method, and that using automation to generate content primarily to manipulate rankings is against its spam policies. Those two statements sit together: the method is not the issue, the purpose and the quality are.

Where AI-produced content does fail is predictable. It restates the consensus of the pages that already rank, which means it contains no first-hand experience and nothing a reader could not get elsewhere. It states specifics confidently without sources. And it is often published at a volume that makes genuine review impossible, which is itself the signal that nobody was accountable for it.

The workable position is to be explicit about the process and to make sure a person is accountable for the output. That is what our own editorial policy sets out, and it is a more defensible answer than claiming no AI is involved.

That there is an E-E-A-T score. There is not. Tools selling an E-E-A-T score have invented a metric and are grading against their own invention.

That author schema improves rankings. Person and author markup helps machines associate content with an identity, which is sensible. It is not a documented ranking input, and adding it to a page with nothing behind it changes nothing.

That you need formal credentials to rank. The guidelines explicitly recognise everyday expertise — someone who has genuinely done the thing can be the right source for it. For health, finance and safety topics the bar is higher, but outside those, demonstrated experience counts.

That E-E-A-T explains a traffic drop. It is a framework, not a mechanism, so it cannot be the cause of anything on its own. A drop has a traceable cause — an indexation change, a core update, a competitor's work, a technical fault — and attributing it to E-E-A-T is usually a way of avoiding the diagnosis.