See if your page
looks trustworthy.

What E-E-A-T actually is

E-E-A-T stands for Experience, Expertise, Authoritativeness and Trustworthiness. It comes from Google's Search Quality Rater Guidelines, the handbook human raters use to judge whether results are any good. Trustworthiness sits at the centre; the other three exist to support it.

Crucially, it is not a score Google calculates and applies to your page. There's no E-E-A-T number in the algorithm. What exists are many signals that correlate with these qualities, which is exactly what this tool checks: the observable, on-page evidence that a real, qualified, accountable source is behind the content.

The raters themselves don't move your rankings either. Thousands of them rate sample results, and Google uses those ratings to test whether changes to its ranking systems surface better pages. Your page isn't scored by a person and then demoted.

Is there a Google E-E-A-T checker?

No. Google doesn't offer an E-E-A-T checker or report one in Search Console, and it has said E-E-A-T isn't a single ranking factor. Every "E-E-A-T score" you'll find, including ours, is a proxy: a third party picking signals it can detect and turning them into a number.

That makes the score useful for one thing: spotting missing evidence. A page can score 100 here and still be thin, wrong or unhelpful, and a page with no schema can still rank on reputation alone. Use the score to find gaps, not to predict rankings.

How the score is calculated

The tool fetches the page's raw HTML and runs 13 yes-or-no checks: three for Experience, three for Expertise, three for Authoritativeness and four for Trustworthiness. Each pillar scores the share of its checks that pass, from 0 to 100, and the overall score is the average of the four pillars, so each pillar carries equal weight.

Every check is a fixed rule, not an AI judgement, so the same page gives the same result every time. The overall score falls into one of four bands:

80 to 100: Strong signals

Most of the evidence is in place. Look at the remaining failed checks and decide whether each one applies to this page.

60 to 79: Good signals

The basics are covered, usually with one pillar lagging. Fix that pillar first; it moves the total the most.

40 to 59: Mixed signals

Typical for a page with sitewide schema and footer links but no byline, date or sources on the content itself.

0 to 39: Weak signals

The page reads as anonymous. Start with the quick wins: a named author, a visible date and links to your sources.

Every signal in your report, explained

Here is exactly what each check looks for, so you know why a signal passed or failed.

Visible author byline

The word "by" (as in "written by" or "authored by") followed by a capitalised name, such as "Written by Jane Smith", somewhere in the HTML.

First-hand language

Phrases such as "we tested", "we found", "in our experience", "hands-on" or "I tried" in the page text.

Supporting imagery

At least two images on the page. The tool counts them; it can't tell an original screenshot from a stock photo, so that part is on you.

Author (Person) schema

Person JSON-LD anywhere on the page, including a Person nested as the author of an Article or BlogPosting.

Author or about page linked

A link to a URL path such as /author, /about, /team or /our-story.

Stated credentials

Words like PhD, MD, CPA, CFA, certified, licensed, accredited or board-certified, or a phrase like "10 years of experience".

Organization schema

Organization (or Corporation) JSON-LD identifying the business behind the site.

Linked brand profiles (sameAs)

A non-empty sameAs list inside that Organization schema, pointing to your profiles elsewhere.

Cites external sources

Links to at least two other websites. Social networks don't count, because they're brand profiles, not sources.

Served over HTTPS

The page loads over a secure connection.

Published or updated date

A visible "Updated", "Published" or "Reviewed" label followed by a date such as "July 12, 2026", "07/12/2026" or "2026-07-12", a <time datetime> element, or a published or modified time in the page's meta tags.

Contact page linked

A link to a URL path such as /contact, /support or /help.

Privacy or terms linked

A link to a privacy, terms, legal or cookie page.

Why it matters more in the AI era

AI assistants have the same problem search engines do, at higher stakes: they have to decide which sources to trust enough to repeat. Pages with clear authorship, dates, citations and entity markup are easier to verify and safer to cite. Nobody outside the AI companies knows exactly how much weight each signal gets, but anonymous, undated pages give a model the least to go on.

That makes trust signals do double duty. They help you rank, and they help you get named when someone asks ChatGPT or Perplexity about your category.

How to strengthen your signals

Most sites lose points on the easy stuff, not the hard stuff: no byline, no date, no schema, no outbound sources. Those are all fixable in an afternoon, and they move you further than another 500 words.

1. 1Put a real, named author byline on every substantive page
2. 2Add Organization schema with sameAs, plus Person schema for authors
3. 3Show a published or last-updated date, and actually keep it current
4. 4Link out to primary sources for the claims you make
5. 5Publish about, contact and policy pages, and link them sitewide

YMYL: where the bar is highest

Your Money or Your Life (YMYL) topics are ones where bad information could hurt someone's health, finances, safety or legal standing: medical advice, investing, tax, insurance, legal guidance. The rater guidelines hold these pages to the strictest standard of trust.

On a YMYL page, a named author with real, stated credentials matters far more than on a recipe or a product roundup. So do citations to primary sources such as studies, regulators and official guidance, and a clear way to reach the business. If your site covers YMYL topics, treat a failed byline, credentials or citations check as urgent.

E-E-A-T and helpful content

Google's helpful content system, launched in 2022 and folded into its core ranking systems in March 2024, targets content made mainly to rank rather than to help people. E-E-A-T is part of how Google describes helpful content, but they aren't the same thing. A page can have every trust signal and still be low-value: thin, repetitive or written around a keyword instead of a question.

This tool checks the trust signals, not whether the writing is helpful. To find low-value content, look in Google Search Console for pages that get impressions but almost no clicks, and pages that are crawled but not indexed. Then improve them with first-hand detail and a direct answer, merge them into a stronger page, or remove them.

An E-E-A-T check vs an E-E-A-T audit

A check, like this one, detects what's visible in the HTML: a byline, a date, schema, links to sources. An audit is a human reading the page and asking harder questions. Is this author actually qualified on this topic? Is the experience real? Do the sources support the claims? What do other sites and reviewers say about the brand?

A lot of E-E-A-T lives off the page, in your backlinks, brand mentions, reviews and reputation, and no single-page check can see it. Run the check first to clear the easy gaps, then audit your most important pages by hand.

Why a signal you have might not show up

The tool reads the HTML your server sends back, the same first look a crawler gets. A few things can make it miss a signal that's really there:

JavaScript-rendered content

If your byline, date or schema is added by JavaScript after the page loads, it isn't in the HTML the tool fetches, so the check fails. Server-rendering these signals is safer for crawlers too.

Unusual byline wording

A name in lowercase, a name shown only as an image, or a name on its own with no "by" in front won't match. "Written by Jane Smith" will.

Dates the tool can't read

A bare date with no "Updated" or "Published" label, or a date written only in words ("last spring"), isn't picked up from the visible text. It still counts if the page has a <time datetime> element or a published or modified time in its meta tags, which most CMSs add.

Pages that block the fetch

Pages behind a login, a bot wall or an error can't be analyzed at all, and you'll see an error instead of a score.

How often to check

Run it on a new template once, then whenever you change it: a theme update or new CMS can quietly drop schema or footer links across the whole site. Check individual articles when you publish or revise them, since bylines, dates and citations vary page to page. For a content library, a quarterly pass over your top pages is a sensible rhythm, more often for YMYL topics.