The search quality rater guidelines get quoted in almost every conversation about E-E-A-T, usually as if a rater somewhere had marked a site down and its traffic followed. That is not how Google describes the process. Raters are part of how Google tests its own systems, and the guidelines are the instructions they follow while doing it.
This guide sets out what Google says raters do, how their work feeds into testing, what the published guidelines are, and how to use them on your own site without mistaking them for a scoring sheet. The facts come from Google’s How Search Works page on rigorous testing.
What search quality raters actually do
Google’s wording is direct: “We work with external Search Quality Raters,” and “Raters assess how well website content fulfills a search request.” They judge quality by looking at the expertise, authoritativeness and trustworthiness of the content in front of them, against the query that produced it.
The key word is external. These are people outside Google’s engineering teams, looking at results the way a searcher might, with a written standard to keep their judgements consistent. Google does not publish how many raters there are or where they work, and any figure you see for either did not come from that page.
A large part of the work is comparison. In side-by-side experiments, raters see “two different sets of Search results”, one of them produced with a proposed change, and say which they prefer and why. The question is never “is this site good?” in isolation; it is whether a change to Google’s systems produces a better set of results.
Do rater scores affect rankings?
Not directly, and Google states this in so many words: “These ratings do not directly impact ranking.” Ratings help Google benchmark the quality of its results, which tells engineers whether a proposed change is an improvement worth launching.
The scale of that testing is set out on Google’s rigorous testing page, which gives figures for 2023:
| Type of test (2023) | Number reported by Google |
|---|---|
| Search quality tests | 719,326 |
| Side-by-side experiments | 124,942 |
| Live traffic experiments | 16,871 |
| Launches | 4,781 |
Google summarises it as “we ran over 700,000 experiments that resulted in more than 4,000 improvements to Search.” Read the funnel from top to bottom. Most proposals are tested and dropped; a small share ship. Raters sit inside that funnel, helping decide which changes survive.
The indirect effect is real, though. If raters consistently prefer results that show clear expertise and trustworthiness, the changes that ship tend to reward those qualities. That is why the guidelines matter to you, even though no rater ever moves your page. Our explainer on how Google ranks websites covers the systems that do.
What is in the search quality rater guidelines
Google publishes the Search Quality Rater Guidelines in two forms: a short overview and the full version, both as PDFs. If you have searched for the search quality evaluator guidelines 2026 PDF, the honest answer is that Google hosts the current document and updates it periodically. Download it from Google rather than from a mirror, and check the date on the cover, because copies of older editions circulate widely.
We will not summarise what any particular edition changed. Edition-by-edition commentary dates quickly, and the substance worth knowing has been stable: page quality, how well a result meets the need behind a query, and the qualities raters look for.
Those qualities are usually written as E-E-A-T: Experience, Expertise, Authoritativeness and Trustworthiness. The rigorous testing page itself names the expertise, authoritativeness and trustworthiness of content; the experience element is covered in our E-E-A-T examples, which show what each signal looks like on a real page. The guidelines also give extra weight to topics that could affect someone’s health, money or safety. We define that category in what YMYL means rather than restating it here.
How to use the guidelines as a self-assessment lens
The useful move is to read the guidelines as a description of what Google’s systems are tuned towards, then ask whether your pages would win a side-by-side test. That is a better question than “which boxes have I ticked?”
- Pick a real query. Choose one your page targets and look at what currently ranks. Would a rater, comparing your result with those, prefer yours, and could you say why?
- Check who stands behind the content. Is it clear who wrote it and why they are qualified? Expertise that cannot be seen cannot be assessed.
- Test trust before anything else. Accurate claims, clear ownership and honest commercial disclosure matter most on pages where a wrong answer costs the reader something.
- Read reviews and comparisons harder. If you publish them, Google’s product review guidelines set a parallel, more specific standard.
What the guidelines cannot do is diagnose a traffic drop. If you lost visibility in an update, the guidelines describe the direction Google was heading, not the specific signal that moved your pages. Start with the timing and the pages affected, as set out in reading a Google core update, and look at what the sites that held up have in common. The helpful content update aftermath is a worked example of why that comparison matters more than any single guideline.
Where the guidelines get misread
Three misreadings come up again and again. The first is that a rater visited your site. Raters rate results for queries in tests; Google does not describe them reviewing individual sites to change their rankings. The second is that adding an author bio is enough. A bio supports expertise that the content already shows; it does not create it. The third is treating the guidelines as a complete list of ranking factors. They are rater instructions, and Google says plainly that the ratings do not directly impact ranking.
A quieter misreading is treating the guidelines as a reason to rewrite everything. If a page already answers its query well, is accurate and makes clear who stands behind it, a rewrite aimed at the guidelines is more likely to dilute it than improve it. The better use of the document is triage: find the pages where a side-by-side comparison would go against you, usually thin pages on topics where readers need to trust the answer, and fix those first. A short list of honest weaknesses beats a sitewide sweep of author boxes and trust badges, and it is easier to check whether the work made any difference once the next update has settled.
A stated limitation: this guide relies on Google’s own description of its testing process. Google does not publish how rater preferences translate into specific launches, so nobody outside the company can trace a particular ranking change back to rater feedback. Treat anyone who claims to as guessing.
Frequently asked questions
Do quality raters change my site’s rankings?
No. Google says the ratings “do not directly impact ranking”. They help benchmark the quality of results so Google can judge whether proposed changes improve Search.
Where can I get the search quality rater guidelines 2026 PDF?
Google hosts the current guidelines, as an overview and a full version, and updates them periodically. Download from Google directly and check the edition date on the document.
What do raters judge?
How well content fulfils a search request, judged on the expertise, authoritativeness and trustworthiness of the content. In side-by-side tests they compare two sets of results and say which they prefer and why.
How much testing does Google run?
For 2023 Google reported 719,326 search quality tests, 124,942 side-by-side experiments, 16,871 live traffic experiments and 4,781 launches.
The takeaway Raters test Google’s changes, not your site. Read the guidelines to understand what Google is aiming for, then ask whether your page would win a side-by-side comparison for the query it targets.
