E-E-A-T

What it is

Google employs human quality raters. They are given sample search results and a long manual, the Search Quality Rater Guidelines, and asked to judge whether a page deserves the position it holds. E-E-A-T is the section of that manual dealing with trust. Four properties are named, and a rater is asked to consider all of them together: Experience, meaning first-hand contact with the subject; Expertise, meaning knowledge of it; Authoritativeness, meaning standing as a source others defer to; and Trustworthiness, which the framework treats as the property the other three exist to support.

The dates matter, because the acronym has changed shape and a lot of writing about it is stale. E-A-T, with three letters, has been in Google's rater documentation since 2013 and was published openly in 2014. The first E, Experience, was added on 15 December 2022. Its arrival was explicit about what it was for: rewarding first-hand knowledge, having actually used the product or been to the place, over competently rewritten desk research. That was a targeted move against a specific kind of page, the review that reads well and was written by someone who never held the object. Google clarified its YMYL definitions again in September 2025.

The single most important thing to understand about E-E-A-T is what it is not. It is not a ranking factor. There is no E-E-A-T score sitting inside the ranking system, and the quality raters do not change the ranking of the pages they rate. Their judgments are used to evaluate whether algorithm changes are working, which is a different job entirely: they are a measuring instrument, not a control. Treating E-E-A-T as a dial to be turned is a category error, and it is one of the most common errors in SEO writing. The framework is best read as a description of what Google is trying to approximate with machinery that cannot directly observe any of the four properties.

That distinction has a practical consequence rather than a merely pedantic one. If E-E-A-T were a score, the correct response would be to optimise the signals it is measured by. Because it is a description of an intention, the correct response is to actually be the thing described, on the assumption that the ranking system is being tuned, imperfectly and over years, to find it. The two strategies diverge fast. One produces author boxes containing invented credentials. The other produces a named author who has done the work.

The framework carries most weight in YMYL territory, Google's classification for pages that can affect a reader's health, financial stability, safety or major life decisions. Outside that territory the scrutiny is lighter, which is why a well-made page about typography faces a different bar from a page about a supplement.

For a publication built on citing research, E-E-A-T and the vault's own epistemic standards point in the same direction, which is convenient but not accidental. Citing a study that later collapsed is the fastest available way to lose the authority a site of this kind is built on. That is a trust failure before it is a ranking failure, and it is the practical reason this publication tracks the replication status of everything it repeats. See The Replication Crisis for the underlying problem, and Publication Bias for the mechanism that keeps producing it.

In effect

15 December 2022, when the first E arrived

The clearest documentary case is the change itself. Google added Experience to a nine-year-old three-letter framework and said what it was for: first-hand knowledge, having used the product or been to the place. That is worth reading as a diagnosis rather than an announcement. A framework only acquires a new letter when the existing letters have stopped separating good pages from convincing ones, and by 2022 Expertise and Authoritativeness were both cheap to simulate. Experience is harder to fake in prose, because a person who has used the thing writes about it differently. The addition is also the strongest available evidence that the guidelines are revised in response to what raters could no longer distinguish.

Judging a supplement review at a glance

An ordinary reader lands on a page comparing three magnesium supplements. The E-E-A-T questions are the ones that reader is already asking without the vocabulary: does anyone's name appear on this, and does that person have any reason to know; has anyone here actually taken the product, or is this three product pages rewritten into paragraphs; does the page cite the trial it refers to, or does it say "studies show"; and would this site be recoverable if the advice were wrong. This is YMYL territory, health, so the bar is the highest the framework sets. The everyday version of E-E-A-T is simply the set of things a careful reader checks before believing a stranger about their body.

The Why We Project's own editorial setup

This publication earns its E-E-A-T signals through specific and auditable choices rather than through markup. Real named author bios carrying credentials, a photograph and professional history. Citations that go to primary sources rather than to aggregator sites. A visible editorial-standards page. Credentialed reviewers on YMYL clinical content. The Experience signal, the one added in 2022, is the one a practitioner-led voice earns most naturally: years of applying psychology in marketing and product work is first-hand contact with the subject in exactly the sense the guidelines describe. Note what is absent from that list. There is no schema field, no plugin and no score, because there is nothing to fill in.

What it does not say

E-E-A-T is not a ranking factor, and this is the claim most often got wrong. There is no E-E-A-T score, and the quality raters do not change the ranking of the pages they rate. Their judgments feed back into evaluating algorithm changes. Any article, tool or agency that offers to raise your E-E-A-T score is selling a thing that does not exist.

It is also not a research finding, which is why this page is tagged not-a-research-claim. It is a proprietary framework and a term of art, published by a company, verifiable against that company's own documentation rather than against a literature. Replication does not apply to it. Neither does effect size. It can be described accurately or inaccurately, and it can change without notice, as it did in 2022 and again in September 2025.

One further correction belongs on the page rather than buried in a changelog, because the site's whole argument is that publications should show their errors. This vault previously credited E-E-A-T to Adam Alter. That is simply wrong: Alter has no connection to the framework, and the attribution was an instance of a failure mode this vault has now caught seven times, a name lifted from a neighbouring note and attached to a concept it has nothing to do with. Corrected on 2026-09-01. The originator is Google, in its Search Quality Rater Guidelines.

Who developed it

Google, in its Search Quality Rater Guidelines. E-A-T appears in that documentation from 2013 and was published openly in 2014; Experience was added on 15 December 2022. There is no individual originator to link to, and no academic paper behind it. Attribution corrected on 2026-09-01.

---