E-E-A-T gets treated as a box to tick in your page template. Add an author bio, add some schema, done. It isn’t that, and the chapter on it in our book, The Digital PR Playbook, explains why better than most of the guides you’ll find online.
Andrew Holland, who wrote the chapter, sums up the whole idea in one word: “brand”. E-E-A-T (experience, expertise, authoritativeness and trustworthiness) is how Google works out who you are, why you know what you’re talking about and why anyone should believe you. Your web team can’t fix that on its own. It’s the sum of everything your marketing has done. Below is the formula from the book, its three quick tests and a fourth test we’d now add for AI search.
The formula: evidence from the past and for the present
The book puts it as a simple sum: “Evidence from the Past and Evidence for the Present = E-E-A-T.”
Evidence from the past is what you’ve already done that proves you’re credible. Reviews on independent sites. Coverage in publications your customers read. Awards, qualifications, years spent doing the thing well. The chapter suggests picturing yourself as an expert witness in court, with Google as the judge. Nobody takes an expert witness at their word. They get vetted, and so does your website.
Evidence for the present is the marketing you’re doing now to show you’re an expert: comment in the press, data and research about your category, podcast appearances, useful content and social profiles that actually say something. If that sounds like everyone’s job, the book’s answer is that it is, whether the rest of the marketing team knows it or not.
One detail worth knowing. E-E-A-T comes from Google’s search quality rater guidelines, the document its human raters use to judge whether results are any good. Raters don’t rank your site. Their feedback shapes the systems that do. The second E, for experience, was added in December 2022.
Test 1: the brand search test
This one takes seconds. Search Google for your brand name and look at what comes up. Is there a knowledge panel? Is it your site, then your social profiles, then reviews and coverage? Or is it a competitor, an old complaint thread, or a different company with a similar name?
You can’t control that page directly. What you can do is give Google more good things to show: coverage, interviews, podcast appearances and social profiles worth clicking. The aim is simple. Anyone checking you out should find positive signals and easy ways to learn more.
Test 2: the news test
Search your brand again, then click Google’s News tab. Are you there? Keep scrolling. If there’s still nothing, the book is blunt about what that means: it “should serve as a warning that you need to do something newsworthy.”
For a consumer brand this is often the most revealing of the three. Plenty of brands with strong sales and good reviews have almost no independent coverage, which leaves very little third-party evidence that they’re experts in anything. Newsworthy doesn’t have to mean a stunt. A data campaign built on what you already know about your customers counts. So does a sharp expert comment on the news journalists are covering today.
Test 3: the instant reaction
This is the one Andrew likes best. Pick five websites from different sectors, one of them yours. Ask someone, ideally a stranger, to look at each for 30 seconds. Then ask what they think the business does, whether they’d trust it and what they make of the site.
It isn’t scientific, and it isn’t meant to be. You want gut feelings rather than a considered review, because gut feelings are what a first-time visitor runs on. And if you can’t find a stranger, the book has a fallback: “this is one of those things that parents are good for.” You’ll be surprised by what people pick up on. A tagline nobody understands. A homepage that never says what you sell. No sign of a real person anywhere.
The chapter makes one more point that applies to all three tests: get someone else to judge. You’re too close to your own brand to see it the way a customer, a journalist or a search engine does.
Test 4: ask the AI tools
This test is new. The book was written in 2024, when Google’s AI search was still being called SGE, and if we were writing the chapter today we’d add it. Ask ChatGPT, Gemini, Perplexity and Google’s AI Mode the questions a customer asks before they’ve heard of you: “best [your category] brands UK”, “is [your brand] any good”, “[your brand] vs [competitor]”. Note whether you’re named, how you’re described and which sources each tool leans on.
The logic is the same as the brand search and news tests, just with a different reader. AI tools build their answers from what trusted sources already say about you. Muck Rack’s May 2026 study of more than 25 million links cited in AI answers shows how much of that comes from other people. So if the news test came back empty, don’t expect this one to go much better.
of the links AI answers cite come from earned media, sources a brand neither owns nor pays for.
Muck Rack, May 2026of the links AI answers cite are journalism, from more than 20,000 newspapers, magazines and news sites.
Muck Rack, May 2026What to do with the results
Most brands that run these tests find the same gap. The website is in reasonable shape, but the evidence elsewhere is thin: few independent mentions, little recent coverage, no visible experts. That’s the part your own site can’t fix, because the whole point is that other people are saying it.
The chapter ends on the line we come back to most: “Experts display expertise, experts get cited, experts provide expert commentary, and experts also provide data and analysis on their industry. And this is why digital PR is the fuel for better E-E-A-T.”
In practice that means three things. Put your real experts forward, with names, roles and credentials, and make sure your site says who they are. Give journalists something worth covering, whether that’s original data or a quick, useful reaction to news that’s breaking. And keep going, because evidence for the present only counts while it’s current. Rerun all four tests every quarter and see what’s moved.
AI tools lean the same way. Most of the journalism they cite is recent, so a burst of coverage two years ago won’t carry you for long.
of the journalism AI answers cite was published in the last 12 months, where the publish date is known.
Muck Rack, May 2026The E-E-A-T chapter is one part of The Digital PR Playbook, which covers the rest of the job too, from ideation and outreach to measuring what digital PR is worth. If this was useful, the book goes into a lot more detail.




