E-E-A-T: what it is and how Google decides who to believe

Greg K.SEO and analytics at Adcore · LinkedIn

E-E-A-T stands for experience, expertise, authoritativeness and trust: four qualities Google uses to judge whether a page and its author can be believed. There is no E-E-A-T score and no algorithm by that name, although plenty of articles describe one. Still, these letters explain why one site gets read and cited while another just sits on the internet. Here is how it works, using an example you have lived through yourself.

What E-E-A-T is: four phone calls you have already made

Picture a Thursday evening. Your mother has been told she needs knee surgery, and now you are the one looking for a doctor. Website rankings mean nothing to you right now. You pick up the phone and, without noticing, ask four different people four different questions.

The first call goes to Aunt Larysa, who had the same operation two years ago. She is not a doctor and will not advise on painkiller doses, but she knows the third night is the worst and when you can use the stairs again. That is experience.

The second call goes to the clinic: who operates, what is their specialty, how many of these operations a year. That is expertise.

The third call goes to a doctor you know: “Who would you take your own mother to?” They name someone, and you have already heard that name from a person who has never met them. That is authoritativeness.

Then comes the consultation. The surgeon says: “You can manage without surgery for now, let’s try something else first.” The price is stated up front, and you may take the contract home to read. That is trust. And that is why you will agree when surgery really is needed.

Google does the same, except that instead of one mother it has billions of searches, and instead of a phone it has signals from pages and from the whole web. E-E-A-T is simply the name for those four questions.

ExperienceHas the author been where they write about?

ExpertiseDo they know the subject?

AuthoritativenessDo people come to them for this topic?

TrustCan they be relied on?

Experience: has the author been where they write about

Experience means the author has been through it: used the thing, visited the place, lived with it. Google asks its raters plainly: who would you trust, a review from someone who used the product or a “review” from someone who never saw it? This letter was added last, in December 2022. The handbook now gives a restaurant review by someone who never ate there as an example of low quality. On a page, experience shows in original photos and details no brochure contains.

Expertise: does the author know the subject

Expertise means knowledge and skill. You would take advice on rewiring an old house from an electrician, not from an old-house enthusiast: that is Google’s own example. A diploma is not always needed, since a hairdresser’s skill shows in a video of them cutting hair. But the more dangerous a mistake is, the less room there is for the self-taught, and that is a separate story: what YMYL is.

Authoritativeness: do people come to you for this topic

Authoritativeness means you are known as the source people go to for a topic. For getting a passport there is one such source: the government website. For what is on offer today at a café, it is the café’s own social media page. Authority does not carry over between topics: Google gives tax forms on a cooking site as an example of a page without the right E-E-A-T, and a good reputation does not fix that.

Why trust sits at the centre of E-E-A-T

Trust matters most because a page without it has low E-E-A-T, however experienced and expert the author may seem. The rater handbook calls Trust the most important member of the family and the other letters what helps to assess it. Trust here means something specific: how accurate, honest, safe and reliable the page is.

The example in the document deserves a verbatim quote: a financial scam is untrustworthy, “even if the content creator is a highly experienced and expert scammer who is considered the go-to on running scams!”

You do not often read that in corporate documentation. The experience is there, the expertise is there, and so is authority in certain circles. The rating is still the lowest.

Google says directly that experience, expertise and authoritativeness can overlap, and that trust sits at the centre.

Three things follow from this, and they save a lot of money on “E-E-A-T optimisation”:

Is E-E-A-T a ranking factor

No: E-E-A-T is not a separate ranking factor, and there is no algorithm by that name. Google puts it in one sentence:

If you have searched this topic before, you have seen headlines about “the EEAT algorithm”. Gary Illyes of Google explained it differently at the Pubcon conference back in 2019: there is no internal E-A-T or YMYL score, only millions of small algorithms, which he called baby algorithms, each with its own narrow job. Together they can at most be loosely described by these letters.

To keep it straight, picture two separate tracks.

Track one: people. Roughly 16,000 raters in different countries follow a 182-page handbook and rate pages. Those ratings do not move any site: the handbook itself says a single rating cannot raise or lower a page, and humans could never check the whole web by hand. Google compares them to feedback cards in a restaurant. The chef reads the cards to find out whether the new menu works, and no guest’s bill changes because of them.

Track two: systems. Automated systems look for signals that together point to experience, expertise, authoritativeness and trust. Google does not publish the full list. It has named one example itself: whether other prominent sites link to or mention the page. On topics where mistakes are costly, such signals get even more weight.

A rater’s score does not move your page. It shows Google whether its changes work.

The practical conclusion follows. You cannot “optimise for E-E-A-T” with a plugin or a checklist, because there is nothing to optimise for. You can only become someone these signals find, and that is the ordinary long work that SEO consists of. The same goes for AI Overviews.

How E-A-T became E-E-A-T

Where Google looks for proof: three places

Raters look for E-E-A-T proof in three places: what a site says about itself, what others say about it and what is visible on the page. This may be the most useful cheat sheet in the whole document, because it works for a real customer too.

The easiest way to show how this looks together is with two pages. Their content is nearly identical: same topic, same length, even the same advice. The only difference is whether you can verify who wrote it and how they know.

An illustrative example. The content is similar, but only the second page can be checked.

The E-E-A-T cargo cult: what does not work

On Pacific islands after the Second World War, people built bamboo runways and wooden planes, then waited for the cargo to arrive. E-E-A-T can go the same way: the form gets copied, and somehow the cargo never lands.

What our measurements showed

In the blog of one dental clinic in Kyiv, 65 articles follow one template: a doctor as author, a doctor as reviewer, dates, tables, questions. We compared the articles that took off with those that stayed at the bottom. The author box was the same, since the template is. Length and number of blocks barely differed either.

What differed was this: does the article answer what the person actually types, and does it open with the answer. Of the 40 articles whose first sentence answers the query directly, 20 reached the top 10 for their main query. Of the 21 that opened with a story, only 4 did. The sample is small, so this is a pattern rather than proof.

The conclusion is unpleasant for checklist lovers, but honest: an author on the page is there so that you are believed, but the query is won by answering it. When both are done together, the result is different. At another dental clinic we rewrote the whole blog to this standard: doctors as authors, reviewers, dates, real images, answers to what people search for. Daily clicks from search grew roughly thirtyfold.

It is impossible to measure the share of E-E-A-T alone here, because everything changed at once. So we do not claim the author box did it.

Test: how much trust your site shows

Twelve questions, three per letter. Tick only what you can point to on the site, not what you know about yourself. This is our cheat sheet, not a Google score: as we said, no such score exists.

This article sits the same exam

Google suggests asking three things about any content: who, how and why. It would be odd not to do that with a text about trust.

E-E-A-T FAQ

How does E-A-T differ from E-E-A-T?

By one letter: in December 2022, experience was added to expertise, authoritativeness and trust. The principle has not changed, so older articles on E-A-T mostly still hold.

Do I need E-E-A-T if my site is not about health or money?

Yes, but the bar is lower. A vacuum cleaner review only needs experience and honesty. To see how high the bar is for your topic, use the YMYL risk scale.

Can articles be written with AI?

Yes, AI in itself is not a violation. Google looks at the purpose: if a text is made primarily for search positions, that is spam. AI has no first-hand experience either, so images, cases and an expert’s review still have to come from a person.

Does E-E-A-T affect AI Overviews?

There are no separate requirements: Google says the same SEO best practices apply to appearing in AI Overviews. So whatever helps a page in regular results helps there too.

Do I need author markup?

It helps systems understand who wrote a text, but it does not make anyone an expert. First a real author and a profile, then the markup. The other way round works about as well as a wooden plane.

Where can I see my E-E-A-T score?

Nowhere: Google does not calculate one, and numbers from any tool are that tool’s own estimate. It is more useful to take the test above and fix the weakest letter.

Sources