CRO
CHIEF
REPUTATION
OFFICERS
SEO Reputation Management8 min read

What Google's Raters Actually Check: The Reputation Test Inside the Quality Guidelines

Google employs thousands of people whose job is to review web pages and assess them. They do not directly change any single ranking. They grade the quality of Google's results so the engineers can tune the algorithms that do. The manual these raters follow, the Search Quality Rater Guidelines, runs to roughly 180 pages and is public. Most of the SEO industry has read the part about E-E-A-T and turned it into the same advice on repeat: build authority, write good content, add an author bio. That reading misses what the document actually is. Underneath the content-quality language, the guidelines are a reputation manual that tells raters to judge a business the same way its customers do: by looking it up.

The instruction almost nobody quotes

Buried in the reputation section is a specific method. Raters are told not to take a website's word for anything important. They are instructed to research the site and its creators using independent sources, and the guidelines even suggest the following searches: look up the business while excluding its own domain, search for its name alongside the word reviews while excluding its own domain, and look for news coverage, expert recognition, and independent references. Then comes the line that should be printed on the wall of every reputation team. When there is disagreement between what a website says about itself and what independent sources say, raters are told to trust the independent sources.

Read that again with a marketer's eye. Google's own quality reviewers are told to run the exact searches a reputation professional runs, to weigh what strangers say about you over what you say about yourself, and to let that off-site picture shape their judgment of your page. The polished About page is a starting point, nothing more. The verdict is assembled from everywhere else.

Trust is the only part you cannot manufacture

The guidelines are unusually direct about how the four letters of E-E-A-T relate to each other. Experience, Expertise, and Authoritativeness are not four equal boxes to tick. They are three supports for a single load-bearing beam: Trust. The document calls Trust the most important member of the family and states plainly that untrustworthy pages have low E-E-A-T no matter how experienced, expert, or authoritative they may seem. Trust overrides the rest.

This is where the usual E-E-A-T checklist quietly falls apart. Three of the four qualities can be staged on your own page. You can write an experienced-sounding first-person account. You can list expertise in a bio. You can imply authority with schema markup and credential badges. Those are on-page signals, and the guidelines treat them as decoration when the fourth quality is missing. Trust is the one dimension assessed from outside your walls, based on the reputation research above. You do not get to declare yourself trustworthy. Everyone else decides that, and Google goes looking for their verdict.

A bad reputation is not a soft factor. It is a floor.

Most writing on the guidelines is aspirational, aiming to reach the top of the scale. The more useful half of the document is the basement, the specific conditions that drop a page to the lowest possible rating. Among the named triggers, alongside deceptive content and pages built only to make money, sits this one: an extremely negative reputation. If independent sources establish that a business engages in scams, fraud, or other conduct that harms its customers, the guidelines direct raters to assign the lowest rating regardless of how the page itself looks.

That is the entire thesis of reputation management stated in Google's own manual. A reputation problem is not merely bad for morale or awkward at the sales meeting. In the framework Google uses to train its quality models, a sufficiently negative reputation is disqualifying on its own. The clean website and the well-written page do not rescue it, because the rating was decided somewhere outside the business's control. The same logic runs in the other direction: a genuinely strong independent reputation is the asset that on-page tactics cannot fake or replace.

What the guidelines mean by quality, precisely

There is a running joke that the entire content half of the SEO field can be summarized as "write good content," advice so vague it commits to nothing. No one is trying to write bad content. The guidelines are more useful than the people who quote them, because Google does not use the word "good". It identifies four measurable dimensions of the main content and expects raters to weigh each.

The first is effort, defined as the extent to which a human being actively worked to create something satisfying. The second is originality, whether the content offers something not already available elsewhere, and whether this page is the original source. The third is talent or skill, whether it was made with enough competence to satisfy the intended audience. The fourth, for informational purposes, is accuracy: consistency with well-established expert consensus. The bar scales to the purpose. A short social post is not held to the standard of a documentary, but both are expected to show enough effort to serve their purpose.

None of that is writing good content. It is a rubric. You can hold a page against those four dimensions and find the specific dimension it fails on. That is the difference between advice and a standard, and it is why hiring a capable writer is a means to the end, not the whole answer. The writer has to demonstrate effort, originality, skill, and accuracy that an outside reviewer can recognize, because that reviewer is exactly who will grade it.

The AI question was never about AI

The guidelines address generative AI without panic, and the distinction they draw is worth taking to heart. AI is treated as a tool that can help create content or can be misused, like any tool. What earns the lowest rating is not the use of a machine. It is the absence of effort, originality, and value. The document gives a concrete example of what no effort looks like: taking freely available content and running it through translation software with no oversight, curation, or added value. The failure there is not that a machine was involved. It is that no human meaningfully worked, nothing original was added, and the result serves no one.

The practical line for any business is the same whether a person or a model produced the draft. Did someone with judgment actually work on this, add something that was not already on the web, and get it right? Volume produced to fill a template fails that test in either hand. The guidelines call the industrial version of it scaled content abuse, and place it among the lowest-quality markers.

Manufactured signals are the ones Google distrusts by design

There is a final irony worth sitting with, because it undercuts an entire category of reputation-for-sale offers. The guidelines specifically name manufacturer self-reviews and paid influencer endorsements as less trustworthy, because the conflict of interest is obvious. The very signals that manipulation vendors sell (planted praise, coordinated reviews, endorsements that look independent but are bought) are the signals Google's raters are trained to discount. A reputation assembled from manufactured enthusiasm is not just fragile in the face of exposure. It is aimed at reviewers who have been told to look straight through it.

What to take from this

The Search Quality Rater Guidelines reward what cannot be shortcut. Build a real off-site reputation because Google's reviewers look for it and trust it over your own page. Concentrate on Trust, because the other three qualities are decoration without it, and Trust is decided by other people. Fix the underlying business problems that generate negative independent coverage, because a bad enough reputation is a rating floor no on-page work can lift. Hold your content to effort, originality, skill, and accuracy rather than to the word good. And do not buy manufactured signals, because they are pointed at an audience trained to ignore them. The guidelines are not really a content manual. They are a long, detailed argument that reputation is the substance and the page is only the surface.


Part of the SEO Reputation Management pillar. For the framework itself rather than the rater's procedure, see E-E-A-T and Branded Search. Related: Brand Reputation Management, Review Management, the DON'Ts of Reputation Management, and the Reputation Is Not Visibility and Fake Grassroots, Real Blowback breakdowns.

Related Topics