What is E-E-A-T and what status does it actually have?

E-E-A-T is the acronym Google uses to summarize four qualities a human rater is expected to recognize in a page: Experience, meaning the writer's first-hand exposure to the subject, Expertise, actual command of the topic, Authoritativeness, the standing of the source, and Trustworthiness, how far what you read can be relied upon. The part almost no popular article states plainly is where the acronym lives: in the Search Quality Rater Guidelines, the manual Google uses to instruct the people it pays to judge the quality of results. Those judgments exist to tell whether a ranking system works, not to position an individual page, and the official documentation says so without hedging: E-E-A-T is not a ranking factor and no URL receives an E-E-A-T score. It is vocabulary describing what the algorithms try to approximate, not a lever anyone can pull from outside.

An evaluation criterion, not a lever

The second E is the recent one: Google added it in December 2022, turning E-A-T into E-E-A-T to recognize people who have actually done a thing rather than only studied it. The four letters do not carry equal weight: the Search Central documentation calls trust the most important of them, with the other three contributing to it. The overall weight rises on the topics Google labels YMYL, those that can affect people's health, financial stability or safety.

The interesting question today is a different one: when a generative system decides whom to cite, on what basis does it pick the source? There is no official Google position on how its AI experiences analyze the reliability of a site before citing it. What the organization has published repeats the long-standing fundamentals, original content, a page people can navigate, technical accessibility, and introduces no separate criterion. Anyone selling a method to "optimize your E-E-A-T inside AI answers" is selling an inference, not a specification. The practice of getting cited is a different territory, the one owned by GEO and AEO, and it is worth keeping apart from the criterion by which a source is judged.

An enterprise example

A machinery manufacturer has published application notes for years, signed with the company name alone, never saying who wrote them or on which installations they were verified. The content is accurate and no rater would call it wrong: it simply contains nothing that ties it to someone who has commissioned that machine. The first E is missing, and with it the only element separating that page from any other produced at zero cost on the same subject. This is the boundary with AI slop: not the grammar, but whether the speaker can be verified.

Why it matters for decision makers

It matters because it dismantles a request suppliers hear often, "raise our E-E-A-T". It is not a metric, it has no threshold, and no single intervention moves it directly. What does exist is a question a company can either answer or not: who wrote this page, what have they actually done, and how do they show it. In organizations where the answer is "the marketing department", the problem is not editorial but procedural, because the competence sits with the people who never write. Decision makers should insist that every technical piece ships with a name, a role and a verifiable reference, and treat topical authority as a consequence of that process, never as its starting point.