Main content, effort, invented authors: Google's quality documentation from October 2026

In a revision dated 1 October 2026, Google expanded its page "Creating helpful, reliable, people-first content" (opens in a new tab) — without an entry in the changelog. The German SEO news site SEO Südwest reported it on 4 October (opens in a new tab). I laid the new version sentence by sentence against the old one and against Google's quality rater guidelines. In short: three additions, and almost all of it was already in the rater guidelines. What is mainly new is that Google now says it in its documentation for site owners. Genuinely new in substance is half a sentence: Google explicitly ties invented authors to its automated quality systems.

If you are short on time: read the section on invented authors and what it means for ranking. The rest is the comparison with the sources.

What changed

I compared the current English page ("Last updated 2026-10-01") with the English version before it (dated 2025-12-10, archived in the Wayback Machine (opens in a new tab)). As a cross-check I used the German translation, which still shows the version dated 2025-12-18. The comparison yields exactly two inserted blocks; otherwise the text is unchanged. On top of that comes the sister page on AI-generated content, which was changed with the same date.

Where	What is new	Where it comes from
Helpful content, at the end of the E-E-A-T section	Transition: alongside E-E-A-T, the quality of the main content is one of the most important factors	Rater guidelines, section 3.2
New section "Main content"	Definition of main content with five categories	Rater guidelines, section 2.4.1, list restructured and expanded
New section "Main content"	Four attributes: effort, originality, talent or skill, accuracy	Rater guidelines, section 3.2, supplemented with statements from 4.6.6
"Who", section on authorship	Warning against invented author profiles as a signal for "automated quality systems"	Rater guidelines, section 4.5.3 — but with new reach
Page on AI-generated content	Note on hallucinations; Google calls it "critical" to manually review AI content before publishing	newly worded, not found in the rater guidelines

The official changelog (opens in a new tab) lists only the page on AI-generated content for 1 October: updated "with information from the Search Quality Raters guidelines" to align the documentation with the presentations Google gives at its developer events. For the hallucination sentences, however, I find no source in the rater guidelines. The change to the helpful content page is missing from the changelog entirely. Search Engine Journal (opens in a new tab) noticed that too. In terms of timing, the reasoning fits Search Central Live Deep Dive Europe (opens in a new tab), which Google had announced for 30 September to 2 October in Barcelona. SEO Südwest, citing Marie Haynes, reports that the criteria were presented there. I have not seen the slides; the connection is an obvious reading, not evidence.

The section numbers refer to the version of the rater guidelines (opens in a new tab) dated 11 September 2025, which Google was serving at the time of my comparison.

Main content: a definition from the rater guidelines

Google now defines in its documentation for site owners: "Main content" is "any part of the webpage that directly helps the page achieve its purpose". Depending on the purpose of the page, the main content "could include" five categories:

Main text and media: text, articles, images, audio, video.
Interactive features: calculators, online tools, games, search functions.
User contributions: forum discussions, customer reviews, comments — if they directly serve the purpose of the page.
Content behind tabs or expandable sections: for example product data, safety information or reviews.
Page titles and headings: visible headings that summarize the topic and help decide whether the visit is worthwhile.

The definition sentence is almost word for word section 2.4.1 of the rater guidelines; the list is restructured and expanded. Two differences lie in the detail. First, the guidelines say content behind tabs could be main, supplementary or advertising content, depending on what is there. The page for site owners leaves out that qualification and only says in general terms what the main content can include depending on its purpose. I read this as a simplification, not as a rule change. Second, the guidelines count only the title at the top of the page as main content; the new page also counts the headings throughout the text.

In practice this means two things for you. First: whatever sits behind an expandable section can fully count as evaluated content — an FAQ in an accordion is not an accessory if it serves the purpose of the page. Second: headings explicitly belong to the main content, and all of them, not just the title. A heading that promises more than the section delivers is therefore no longer a style problem but a quality problem — which matches the long-standing question on the same page of whether the title and heading "avoid exaggerating or being shocking in nature".

What the definition does not say: how Google recognizes the main content in the markup. There is not a word about <main> or other HTML elements. Why I still set <main> properly is explained in HTML landmarks — the reasons are reader mode, screen readers and extraction libraries, not a ranking promise.

Four attributes: effort, originality, skill, accuracy

To evaluate the main content, Google names four attributes that quality raters are trained on. The rater guidelines have had them in section 3.2 for a while. The version for site owners differs in a few places, and that is exactly where reading closely pays off: it pulls statements from other sections of the guidelines into the definition, it eases off on skill, and for accuracy "consider the extent" becomes "must be highly accurate".

Effort: how much human work went into the content or the systems behind it. Two sentences are new in the definition, but not in substance. The first names generative AI directly: Google describes producing large amounts of text with AI "without manual oversight or curation" as "little to no effort". The second is uncomfortable for summaries and curator sites: "Attribution or giving credit to other sources doesn't replace the need for original effort." Both are already in section 4.6.6 of the rater guidelines in essence. There, the lowest rating applies even "if the page assigns credit for the content to another source". What is new is that Google now tells site owners directly. A clean list of sources does not turn a summary into original work.

Originality: whether the content offers information or perspectives that do not already exist elsewhere.

Talent or skill: whether the content shows the skill needed for a satisfying experience — clear language, well-produced video, working tools. New in this definition is a reassurance, together with an example: not every piece of content needs specialist knowledge. The guidelines already know the idea ("Informal expertise is equally important"), but not the example: someone sharing their experience with a new method for clearing snow from their driveway.

Accuracy: informational pages should be factually correct. For YMYL topics — health, finance, safety, that is, topics that can significantly affect people's lives — the content has to be "highly accurate and consistent with established expert consensus". The guidelines phrase this as a question to consider ("consider the extent"), the page for site owners as a must.

I check this article against exactly this list, because otherwise it would not be credible. The effort lies in the comparison: I put the old and the new version side by side and looked up every new passage in the rater guidelines. I have not found this comparison of all three additions in the reports I know of. Search Engine Journal did at least check the author paragraph against the guidelines. Accuracy is secured by an adversarial fact-check before the text goes online. A list of sources alone would be too little by Google's sentence — and I see it the same way.

Invented authors: the genuinely new sentence

In the "Who" section, Google already asked whether your visitors can tell who created a piece of content, whether there is a byline, and whether it leads to further information about the author. This paragraph is new:

However, avoid using deceptive authorship information. Fabricating creator profiles (such as by using AI-generated headshots, made-up names, or false credentials to make content appear as if it was written by human experts) is a form of deception. Any form of deception makes a page untrustworthy to both users and our automated quality systems, and is a signal of a low-quality page.

In substance, this too is for the most part not new. Section 4.5.3 of the rater guidelines already names invented author profiles with AI-generated images and false claims about qualifications as deception, and deception gets the lowest rating there. What is new is the addressee. The rater guidelines describe how people evaluate search results; Google itself stresses that these ratings do not directly influence ranking. The new sentence, by contrast, explicitly speaks of Google's "automated quality systems" and calls deception "a signal of a low-quality page".

As far as I can see, this is the first place in Google's documentation for site owners where invented authors are explicitly tied to the automated systems. Google does not say how this signal is technically detected or weighted. Nor does the page announce a penalty, an update or a manual action.

For my article Author pages without external signals this means: the core thesis stands. An author page does not create reputation, and you need an author identity right away. But Google has now drawn the line between thin and honest and invented itself. I have updated the article in the affected places.

Accuracy and AI: fact-checking is now Google's wording

The second changed page is Google's guidance on AI-generated content (opens in a new tab). It now says:

Keep in mind that generative models don't retrieve facts, but predict a likely sequence of words based on their training data. Because of this, generative AI outputs may contain inaccuracies (also known as hallucinations). It is critical to manually factcheck and review all AI-generated content for accuracy and trustworthiness before publishing.

The sentence after that explicitly extends the review to metadata: <title>, meta description, structured data and image alt text, because they can appear in search results. That list existed before. What is new is that it now hangs on the manual review: "This includes metadata like …" became "This review also applies to metadata like …".

For me this confirms my approach, and I take the metadata list literally. My fact-check examines every verifiable statement in the text. For this article, the description in the frontmatter is also explicitly part of the review brief, because that is exactly what ends up as the meta description in search results.

What follows for ranking — and what doesn't

This is where it pays to take Google's own boundaries seriously, because the first reports like to leave them out.

Raters do not directly influence ranking. The helpful content page says: "Search raters have no control over how pages rank. Rater data is not used directly in our ranking algorithms." The four attributes describe what raters are trained on — not a formula that an algorithm works through.
E-E-A-T is not a single ranking factor. "While E-E-A-T itself isn't a specific ranking factor" — that sentence was already in the old version and is still there unchanged.
No update announced. The change is a documentation change. Google names no core update, no new spam policy and no date from which anything will be evaluated differently.
The exception: in the paragraph on invented authors, Google explicitly speaks of the automated systems. If you equip your content with invented people, you can no longer claim that this is "only" a matter for the rater guidelines.

Self-test: this page against "who, how and why"

I put the questions from the new version to this page. Not as advertising, but because it is the fastest way to show where the questions catch in practice.

Who: every blog article has a byline with my name, linked with rel="author" to my main site jpkc.com. The JSON-LD of every article page contains a complete Person node with sameAs pointing to my external profiles. There is deliberately no dedicated author page. Google's questions do not require one either: the byline should lead to "further information about the author", and elsewhere the page names "links to an author page or a site's About page" as equivalent examples.

How: where I use AI is stated on the AI transparency page, which is linked in the footer of every article page. What is missing there is a statement per article. Google phrases disclosure as a question ("self-evident to visitors through disclosures or in other ways?") and suggests it "when it would be reasonably expected". The questions are asked for content where automation contributed "substantially". Whether a site-wide note in the footer is enough is a judgment call, not a clear rule. I consider it an open item on my list.

Why: there is no advertising on this site, and the articles grow out of what I come across in my daily work with the web. I can claim that. Whether it is true, you have to check against the texts themselves.

The German translation lags behind

If you open Google's page in German, you do not see the change. When I compared them, the German version of the helpful content page was still dated 2025-12-18, the page on AI-generated content 2025-12-31. The new sections on main content, the four attributes and invented authors are missing there, as is the sentence on manual review.

My general tip: on Google's Search Central documentation, always check the "Last updated" date below the text. Translations can lag behind the English original.

What I am not claiming

Not: "Google has introduced new ranking factors." The four attributes come from the rater guidelines, and according to Google, rater data does not flow directly into ranking.
Not: "Content behind tabs is now fully weighted." The documentation defines what main content is. It says nothing about its weighting in ranking.
Not: "Invented authors lead to a penalty." Google calls them a signal of low-quality pages. Whether and how strongly that has an effect in an individual case is not documented.
Not: "The four attributes were presented in Barcelona." SEO Südwest reports that, citing Marie Haynes, and the changelog justifies the change to the AI page with presentations at developer events. I do not know first-hand what was shown there.
Not: "AI-generated content is now prohibited." Google calls manual review "critical" and warns against AI mass content without oversight. The page does not prohibit the tool itself.

FAQ

What is new in Google's documentation on helpful content?

Since 1 October 2026, Google defines in its documentation for site owners what main content is and names four attributes that quality raters are trained on: effort, originality, talent or skill, and accuracy. On top of that comes a warning against invented author profiles. Almost all of it was already in the rater guidelines. What is mainly new is that Google explicitly names invented authors as a signal for its automated quality systems.

What counts as main content according to Google?

Any part of the page that directly helps the page achieve its purpose: main text and media, interactive features such as calculators or tools, user contributions if they serve the purpose, content behind tabs or expandable sections, and page titles and headings.

Are the four attributes now ranking factors?

No, not in the sense of a formula. Google describes them as criteria that quality raters are trained on, and writes on the same page that rater data does not flow directly into the ranking algorithms. They do show, however, what Google is trying to reward with its systems.

Can I still use AI for writing?

Yes. But Google explicitly calls it "critical" to manually review AI-generated content for accuracy and trustworthiness before publishing, including titles, meta descriptions, structured data and alt text. Google describes large volumes of AI text without human oversight as "little to no effort".

Do I need an author page now?

Not necessarily. Google asks whether the byline leads to further information about the author, and names an author page or an About page as examples. What matters is that the information is true: Google treats invented names, AI-generated headshots or false credentials as deception. When a dedicated author page starts to pay off is covered in Author pages without external signals.