E-E-A-T: a rater framework rather than a score, and what actually changes on a page

Where the four letters come from, the four things E-E-A-T is not, what each one looks like on a real page, the half that happens off your site, and a test for whether this is even your problem.

Published by AI Knows Us · Updated 19 October 2026

E-E-A-T stands for Experience, Expertise, Authoritativeness and Trustworthiness, and it comes from Google's Search Quality Rater Guidelines, a public document written for the human raters who assess search results. That origin is the single most important thing to know about it, because it tells you what E-E-A-T is not. There is no E-E-A-T score attached to your site, no rater looks at your page and moves it up or down, and no setting anywhere accepts a value for it. What it is, is a description of what Google wants its results to look like, written plainly enough to be used as a checklist. The guidelines name Trust as the most important of the four. The second E, Experience, was added at the end of 2022 and is the easiest of them to act on, because it asks a question with a factual answer: did whoever wrote this actually do the thing?

This page covers what the letters mean, the four things people wrongly believe about them, what each one looks like on a real page, and how to tell whether this is actually your problem. For the primary sources behind all of it, see the documentation Google publishes itself.

The four letters, and where they came from

LetterThe question it asksWhat answers it on a page
ExperienceHas the creator actually done, used, visited or lived this?Original photographs, your own measurements, what you tried that failed, the detail nobody could invent
ExpertiseDoes the creator have the knowledge or skill for this subject?A named author with relevant credentials, qualifications or a track record, stated where a reader can see it
AuthoritativenessIs this person or site a recognised source on the subject?Being cited, quoted, listed or linked by others in the field, and a focused body of work rather than a page on everything
TrustworthinessIs this accurate, honest, safe and transparent?Who you are, how to reach you, how you make money, when the page was updated, and corrections when you get something wrong

The framework was E-A-T for years and became E-E-A-T when Experience was added in December 2022. Trust sits at the centre of the group in the guidelines' own description, which is the practical hierarchy to work in: an expert page on an untrustworthy site is worth less than a trustworthy page with modest expertise.

Four things E-E-A-T is not

  • It is not a score. Nothing in Search Console reports it, the value exists nowhere on your site for a tool to find, and any product selling you an E-E-A-T number has invented the number.
  • It is not a ranking factor you can set. It is a description of the qualities Google's systems are built to approximate, which means you influence it only by changing the things a human or a machine could observe about you.
  • Raters do not rank your page. They assess results so Google can tell whether a change to its systems made things better or worse. Nobody is scoring your site individually and nobody will visit it on request.
  • The bar is not the same for every topic. The guidelines set a far higher standard for what they call Your Money or Your Life subjects, where bad information can damage somebody's health, finances, safety or rights. A recipe and a page about drug interactions are not held to the same standard, and they should not be.

Trustworthiness, which is the one to fix first

Trust is the cheapest of the four to improve and the most commonly missing. Every item here is a fact about your organisation that either appears on the site or does not.

  • Who publishes this. A real about page naming the business, where it is and who runs it. Not a paragraph of mission language.
  • How to reach a person. An address, a phone number, a monitored email. A contact form alone is weaker than any of them.
  • How you make money. If pages carry affiliate links, sponsorship or advertising, say so where the reader meets it.
  • When this was last checked. A visible date, and an actual review rather than a touched timestamp.
  • What happens when you are wrong. A correction note on the page beats a silent edit, and it is the single most credible thing a publisher can do.
  • Evidence for claims. A figure with a source and a date, rather than a figure.

Experience, the newest letter and the easiest to act on

Expertise is slow to build and experience is often already in the building. The question is whether the page shows it. Three substitutions turn a generic page into one that demonstrably comes from doing the work:

  • Your own photographs instead of stock. The job in progress, the premises, the actual product, the screen you are describing.
  • Your own numbers instead of industry figures. What you measured, over what period, with the method stated. One honest number of your own outweighs five borrowed ones.
  • What went wrong. The failure modes, the thing that did not work, the case where you would advise against your own service. Nobody who has not done the work can write that paragraph.

The half most people skip, which happens off your site

The guidelines send raters to look for what other people say about a site and its creators, using independent sources rather than the site's own claims. That means a meaningful part of this is not on your website at all.

  • Independent coverage and citations. Being quoted, listed or referenced by organisations in your field.
  • Reviews and their responses. Not the average alone. Whether complaints are answered, and how.
  • Professional registers and licences. In regulated work this is the strongest single trust signal available, and it is verifiable by a stranger.
  • A consistent identity across the places that describe you, so the business a researcher finds elsewhere is recognisably the same one. The mechanics of linking a person or organisation to its other profiles are in author credentials and citation mechanics.

How to tell whether this is actually your problem

E-E-A-T is the most over-diagnosed thing in search. Before spending a quarter on author bios, run this test. Search the query you are losing, open the three pages above you, and ask which of these describes them.

  • They show first hand evidence you do not. Their own photographs, their own data, specifics from having done it. That is an experience problem and the fix is the three substitutions above.
  • They are a recognised name in the field and you are not. An authority problem, which is a two year programme rather than a page edit, and the route is doing work that gets cited.
  • They are transparent about who they are and you are anonymous. A trust problem, and the cheapest one on this list to fix this week.
  • None of the above, and their page is simply a better answer. That is a content problem, not an E-E-A-T problem.
  • Your page is not indexed, or is slower, or targets the wrong query. Then no amount of credentials will move it, and the diagnosis belongs in technical and keyword work instead.

E-E-A-T: the questions that keep coming up

Is E-E-A-T a ranking factor?

Not in the sense of a measurable input you can adjust. It is the standard Google's quality raters are asked to apply when judging results, and Google builds and tunes its systems to produce results that match that standard. So the qualities matter and the acronym itself is not a dial. Treat it as a description of the destination rather than as a lever.

What changed when the second E was added?

Experience was introduced alongside the existing three in December 2022, recognising that for many subjects first hand use matters more than formal qualification. A review of a vacuum cleaner written by somebody who owns one beats a review by a qualified engineer who has never touched it, and the framework now says so.

Do I need an author bio on every page?

You need the reader to be able to tell who stands behind the content, which an author bio is one way to achieve. For an organisation speaking in its own voice, a clear about page, named leadership and real contact details do the same job. Where the subject is one where the person matters, such as medical, legal or financial advice, the named author with credentials is the stronger answer.

Does adding schema markup create E-E-A-T?

No. Structured data makes facts that are already true machine readable, which is useful and is not a substitute for the facts. Marking up an author who has no credentials and no track record states that absence more clearly.

Is YMYL only about medical and financial pages?

It is wider than that. The guidelines cover anything that could significantly affect somebody's health, financial stability, safety, or the welfare and wellbeing of people and society, which pulls in legal information, civic information, safety instructions and major purchases. If being wrong on your page could cost somebody real harm, you are in that territory.

How long does improving this take?

The trust items are a week of work and they often matter most. Experience shows up as fast as you can publish your own evidence. Authority is the slow one, measured in years of being the source other people cite, and there is no way to buy it that does not eventually look bought.

The same research step, run by a machine

The rater guidelines tell a human to go and find out what independent sources say about a site and the people behind it, rather than trusting the site's own claims. An assistant answering a buyer's question does a recognisable version of the same thing: it assembles a view of you from whatever else the web holds, and it reaches for sources it treats as independent. That is why the off-site half of this page matters more than it used to, and why a business with excellent pages and no external trace gets described thinly or not at all. Our record of what assistants actually cited across 597 readings shows which kinds of source those answers were built from.

Read next

AI Share of Voice: Three Legitimate Denominators, and Why Ours Reads 0.5% or Six in 20,143A share needs something to be a share of. One run of 1,260 readings produces three defensible denominators, and the same six citations look sixteen times smaller under one of them than under another.Simon-Kucher pricing strategy, what the firm publishes about its own method and what a Singapore buyer should ask before briefing itThe firm publishes its approach, its scale and four named service lines, and publishes no fee, so the useful work before any call is reading what it claims and deciding which decision you are buying.International pricing strategies for a Singapore company, and the corridor that stops one market undercutting anotherA converted home price is the one answer that is almost always wrong, so set a price per market and keep the gaps narrower than the cost of moving a unit between them.Premium pricing strategy for a Singapore business, and the fence that makes the higher number holdCharging permanently more than the credible alternatives works when the buyer can name what the extra money buys, and the published Singapore writing ladder shows exactly how that is built.Red Antler publishes no price, so here is what its own site does publish and how to cost the same scope from SingaporeRead on 29 October 2026, three pages of redantler.com carry a complete deliverable list, three business stages and not one figure about what any of it costs.Hiring a full service content agency, the five stages a retainer has to cover and the two handoffs where the money leaksMost content retainers buy production and quietly leave strategy and distribution with you, so price the stages rather than the word count.

Or browse the whole archive.

See what AI says about you.

The first scan is free, it needs no account, and it asks the questions your buyers ask.

Free. No card. We ask 5 real buyer questions on 2 AI apps.

Already sure you want this? Start free. No card needed, and the free plan stays free. Pay only when you want the work done. Or talk to a person if you have a question first. Or read how we fix what it finds.