AI Content and E-E-A-T: What 'Experience' Means for a Review Site That Hasn't Touched the Products
Google added the first E to E-A-T in December 2022, and "experience" has been the word that worries AI-assisted review sites ever since. If the model wrote the guide and nobody on the team has held the product, does the page have experience? What does Google even mean by it? And what happens to sites that fake it?
This post sets out what Google's own documents say, what counts as experience on a review site that works mainly from listing data, the signals that fake it and get caught, and the author and about pages that carry most of the weight. It follows on from the Google AI content post, which covers the policy; this one is about the quality bar.
The short answer
Experience means the content was made by someone who has first-hand knowledge of the subject, and Google wants that shown, not claimed.
A review site that hasn't used every product can still have it: say what you compared and how, show what you do own, add judgement the listing can't supply, and put a real named author behind the page. Don't write "we tested" when you didn't; that's the one thing that gets a site classed as fake.
What Google actually says
When Google announced the extra E, it described experience as whether the content "was produced with some degree of experience, such as with actual use of a product, having actually visited a place or communicating what a person experienced".
The quality rater guidelines add that for some topics the most valuable content comes from first-hand experience rather than expertise, and give product reviews as the example.
Google's review best practices are more concrete. They ask reviewers to "provide evidence such as visuals, audio, or other links of your own experience with what you are reviewing", to "discuss the benefits and drawbacks of something, based on your own original research".
They also ask reviewers to "explain what sets something apart from its competitors", and, when calling something the best, to "include why you consider it the best, with first-hand supporting evidence".
Note what that list doesn't say. It doesn't say every product must be bought and used. It says that claims of experience need evidence, and that the review needs original research and comparison. Those are different requirements, and a site can meet the second without lying about the first.
What counts as experience on a review site
| Signal | Counts? | Why |
|---|---|---|
| Your own photos or video of a product you own | Strongly | The evidence Google's guidance names first |
| Specific facts the listing hides (what's in the box, a measured weight, the firmware quirk) | Yes | Original research; can come from owner reviews and manuals, honestly attributed |
| A clear, defended "why this one" with a trade-off | Yes | Judgement a listing can't supply |
| "Based on specs, 400 owner reviews and the manual" stated plainly | Yes | Honest method; raters are asked to judge whether experience is credible, not whether it's exhaustive |
| A named author with a fitting background | Yes | Who is responsible, and are they credible for this |
| "We tested" with no evidence | No, and it's harmful | A claim with nothing behind it is what raters are told to distrust |
| Stock photos of the product | No | Not your experience; also outside Amazon's image licence |
| Generic pros and cons that fit any product | No | Signals no one looked at this product specifically |
Honest wording when you haven't used it
The simplest fix is to say what you did. "We compared the eight most-bought models on their specs, the manuals and several hundred owner reviews, and we own the top pick" is a method statement, and it's one a rater can respect. It also reads as more credible than "we spent 40 hours testing" on a site with no photos of anything.
Then add what only a person could: which product you'd buy, what most buyers overrate, which spec matters and which is marketing. One or two sentences of that per product. It's the editing pass the fact-check post comes before and the editing post describes.
What fakes experience and gets caught
- Invented testing. "In our tests the battery lasted 9 hours" when there were no tests. One contradiction with an owner review and the page is a fabrication.
- Fictional authors. A stock headshot and a made-up name with "10 years in outdoor gear". Raters check; so do readers, and so do competitors.
- Experience boilerplate. "We tested every product on this list" repeated across two hundred guides on a two-person site.
- Hands-on language from listing data. "The grip felt secure" about a product the writer has never seen. The model does this if you let it; the editing pass removes it.
None of these need a penalty to hurt. They're what Google's scaled content abuse and reviews guidance describe, and they're what makes a reader leave.
The pages that carry the signal
Author page
A real name, a photo, two or three sentences of specific background ("has run a bike shop's service desk since 2017"), and a list of their posts. Link every post to it. If the owner writes under a pen name, keep the background true even if the name isn't.
About page
Who runs the site, why this topic, how products are chosen and compared (the method statement again, in full), and how the site makes money. A reader who checks and finds an honest page stays; a rater who checks and finds nothing marks the site down.
The guide itself
The method statement near the top, the author's judgement in the verdict, any real photos where the product is discussed, and specific facts in each section rather than adjectives. That's the whole of E-E-A-T on the page: shown, not claimed.
Doing it with AffSERP
AffSERP's X Best Review tool is built to produce the honest version by default: it works from live product data and owner reviews, keeps to facts the listing supports, and doesn't write hands-on claims the site can't back.
The author set per site goes on every guide, and the guide opens in the editor after generation so the method statement, your judgement and your own photos go in before publishing. The experience itself has to come from you; the tool's job is to leave room for it rather than fake it.
FAQ
What is E-E-A-T?
Experience, Expertise, Authoritativeness and Trust: the four qualities Google's quality raters are told to look for when judging whether a page deserves to rank. It's a framework for human raters, not a direct ranking factor, but Google's systems are built to reward the same things. Experience was added in December 2022.
Can an AI-written review site have E-E-A-T?
Yes, if the experience and expertise are real and shown. The model writes the text; the site owner supplies what they actually know, photos of what they actually own, and a named author with a background that fits. What fails is an AI site pretending to have tested products it never saw.
Do I have to buy every product to review it?
No. Google's review guidance asks for evidence of your own experience where you claim it, and for original research and comparison. A guide that compares products from listing data, specs and owner reviews, says so, and adds the writer's judgement is honest and can rank. A guide that says 'we tested' and didn't is the problem.
Does an author page actually matter?
Yes. Quality raters are told to look for who is responsible for the content and whether that person is credible for the topic. A named author with a short, specific background, a photo and a page listing their posts is the simplest signal a small site can send. 'Admin' as the author sends the opposite.
Keep reading
