People-First Content: The 38 Questions in Google's Own Guide

Google’s guidance on people-first content contains 38 self-assessment questions across six groups. They are a self-assessment instrument rather than a list of ranking factors, and the most quoted line in the document is routinely cut in half.

Google writes that among experience, expertise, authoritativeness, and trustworthiness, “trust is most important.” The sentence continues: “The others contribute to trust, but content doesn’t necessarily have to demonstrate all of them.” So E-E-A-T is not four requirements. It is one outcome with three contributors, and you do not need all three.

What Is People-First Content?

People-first content is content created primarily to help people solve a problem, answer a question, or make an informed decision - not simply to rank higher in search engines.

Does that make SEO search-engine-first?

No, and Google says so directly in a section headed “What about SEO? Isn’t that search engine-first?” Its conclusion is verbatim: “SEO can be a helpful activity when it is applied to people-first content, rather than search engine-first content.”

The difference is sequence. Deciding what to write because a reader needs it, then making it findable, is people-first. Starting from a keyword and finding a reader afterwards is the other thing.

What Does Google Actually Say About E-E-A-T?

More than the quote that usually circulates. The full passage:

“To do this, they identify a mix of factors that can help determine which content demonstrates aspects of experience, expertise, authoritativeness, and trustworthiness, or what we call E-E-A-T. Of these aspects, trust is most important. The others contribute to trust, but content doesn’t necessarily have to demonstrate all of them.”

Three things follow, and the third is the one almost nobody acts on.

Trust is the outcome. Experience, expertise, and authoritativeness are not parallel goals; Google describes them as contributors to trust.

You do not need all of them. “Content doesn’t necessarily have to demonstrate all of them” is explicit permission to be strong in one contributor and absent in another. A first-hand account from someone with no formal credentials can satisfy this. So can a credentialed review of something the author has never personally used.

So the common audit is wrong. Scoring a page against four separate E-E-A-T boxes and treating gaps as failures invents a requirement Google disclaims in the same sentence it names the priority.

What Does Google Say About Effort and AI Content?

It defines Effort and sets one explicit standard for generative AI. In the “Main content” section, Google defines effort as “The extent to which human work went into creating the content or the systems powering it.”

Then the line that matters. Google states that “using generative AI to produce large amounts of text without manual oversight or curation represents little to no effort.”

What does “or the systems powering it” mean in practice?

That the human work does not have to be the writing itself. Google’s phrasing explicitly credits work that went into the systems producing the content, not only work on the page.

Read the two sentences together, and Google is not saying AI output is effortless. It names three conditions that make it so: large amounts of text, no manual oversight, no curation. Remove any of them and the sentence stops applying.

This is the most useful standard in the entire document for anyone running an AI-assisted process, because it tells you what to build. A verified pipeline with human review at the point of publication is work on “the systems powering it.” Bulk unsupervised generation is the thing named.

Calibration note. Google states this as a judgment about effort, not as a ranking penalty, and the document contains no thresholds for what counts as “large amounts” or sufficient “oversight.” Treating a specific review ratio as the compliant answer invents precision Google did not give.

What Are the 12 Content and Quality Questions?

Verbatim, complete. These are the questions about the work itself.

1. Does the content provide original information, reporting, research, or analysis?
2. Does the content provide a substantial, complete, or comprehensive description of the topic?
3. Does the content provide insightful analysis or interesting information that is beyond the obvious?
4. If the content draws on other sources, does it avoid simply copying or rewriting those sources, and instead provide substantial additional value and originality?
5. Does the main heading or page title provide a descriptive, helpful summary of the content?
6. Does the main heading or page title avoid exaggerating or being shocking in nature?
7. Is this the sort of page you’d want to bookmark, share with a friend, or recommend?
8. Would you expect to see this content in or referenced by a printed magazine, encyclopedia, or book?
9. Does the content provide substantial value when compared to other pages in search results?
10. Does the content have any spelling or stylistic issues?
11. Is the content produced well, or does it appear sloppy or hastily produced?
12. Is the content mass-produced by or outsourced to a large number of creators, or spread across a large network of sites, so that individual pages or sites don’t get as much attention or care?

Question 9 sets the hardest standard in the document, because value is defined comparatively against what already ranks. A competent page covering the same ground as ten competent pages provides no substantial value by this framing.

Question 12 describes the operating model of a large part of the content industry.

What Are the 4 Expertise Questions?

Verbatim, complete. Note that three of the four are about trust rather than knowledge.

1. Does the content present information in a way that makes you want to trust it, such as clear sourcing, evidence of the expertise involved, background about the author or the site that publishes it, such as through links to an author page or a site’s About page?
2. If someone researched the site producing the content, would they come away with an impression that it is well-trusted or widely recognized as an authority on its topic?
3. Is this content written or reviewed by an expert or enthusiast who demonstrably knows the topic well?
4. Does the content have any easily-verified factual errors?

Two details worth extracting.

Question 1 names the mechanism. Clear sourcing, an author page, an About page. These are not inferred best practices; they are the examples Google gives.

Question 3 says “expert or enthusiast.” Not a credentialed expert. An enthusiast who demonstrably knows the topic qualifies, and “or reviewed by” means the knowledgeable person does not have to be the writer.

Question 4 is the only factual standard in the document, and it is narrower than it looks. Not whether the content is true in some unverifiable sense, but whether it contains errors that are easily verified. One checkable wrong figure is a trust failure and the cheapest possible thing to avoid.

What Are the 11 Search-Engine-First Warning Signs?

Verbatim, complete. Several describe practices sold as standard.

1. Is the content primarily made to attract visits from search engines?
2. Are you producing lots of content on many different topics in hopes that some of it might perform well in search results?
3. Are you using extensive automation to produce content on many topics?
4. Are you mainly summarizing what others have to say without adding much value?
5. Are you writing about things simply because they seem trending and not because you’d write about them otherwise for your existing audience?
6. Does your content leave readers feeling like they need to search again to get better information from other sources?
7. Are you writing to a particular word count because you’ve heard or read that Google has a preferred word count?
8. Did you decide to enter some niche topic area without any real expertise, but instead mainly because you thought you’d get search traffic?
9. Does your content promise to answer a question that actually has no answer, such as suggesting there’s a release date for a product, movie, or TV show when one isn’t confirmed?
10. Are you changing the date of pages to make them seem fresh when the content has not substantially changed?
11. Are you adding a lot of new content or removing a lot of older content primarily because you believe it will help your search rankings overall by somehow making your site seem ‘fresh’?

Question 6 is the most useful test in the whole document, because it is answerable by one person reading one page.

Questions 2, 7 and 10 describe a publishing cadence, a word-count brief and a date refresh. All three appear in standard content operations, and two of them are in most SEO tools’ default output.

Where Do Google’s Guidance and the Measurements Agree?

On length and volume, Google’s warnings and independent correlation data point the same way.

Ahrefs studied 75,000 brands for the correlation figures, and separately found that adding excessive content diluted focus and harmed rankings.

Calibration note. Ahrefs states that correlation is not causation and that all its measured correlations are moderate to weak. Two independent sources pointing the same way is stronger than either alone, and it is still not proof of mechanism. Treat it as a well-supported bet.

For how this plays out in AI answers, see how the CITE framework gets content cited in AI.

What Are the Who, How, and Why Questions?

Nine in total across the three groups, and the How group is the most current material in the document.

H3: Who created the content - 3 questions:

1. Is it self-evident to your visitors who authored your content?
2. Do pages carry a byline where one might be expected?
3. Do bylines lead to further information about the author or authors involved, giving background about them and the areas they write about?

Google names the specific tactic it considers deceptive: fabricating creator profiles “by using AI-generated headshots, made-up names, or false credentials.” And it states the consequence in full: “Any form of deception makes a page untrustworthy to both users and our automated quality systems.”

That closing clause is worth reading twice. Google is saying deception costs you with readers and with its ranking systems, which puts invented authors in the deception category rather than the optimisation category.

H3: How the content was created - 3 questions:

1. Is the use of automation, including AI-generation, self-evident to visitors through disclosures or in other ways?
2. Are you providing background about how automation or AI-generation was used to create content?
3. Are you explaining why automation or AI was seen as useful to produce content?

Question 3 is the one that gets missed, and it asks for more than disclosure. Google is asking you to justify the choice, not merely admit it. A note saying a piece was AI-assisted answers question 1. A note explaining that AI was used to process a large dataset that a human then verified answers all three.

None of these prohibit AI. They ask that the reader knows, understands how, and understands why.

Why it was created. The purpose should be helping visitors rather than manipulating rankings, which is the question the other eight serve.

What Does Google Say About Page Experience?

That breadth matters more than any single metric:

“Google’s core ranking systems look to reward content that provides a good page experience. Site owners seeking to be successful with our systems should not focus on only one or two aspects of page experience.”

Read that against how page experience is usually audited, which is a Core Web Vitals score and little else. Google’s own framing is explicitly against optimising one or two aspects.

How Should You Actually Self-Assess?

Seven steps. Step 7 is Google’s own instruction and the one almost nobody follows.

1. Run the questions against one real URL, not in the abstract. They are vague as principles and sharp against a specific page.
2. Start with expertise question 4. Check every number, name, and date for easily-verified errors. Cheapest high-value check available.
3. Answer content question 9 honestly. Open the top three results and ask what your page adds. If nothing, formatting will not save it.
4. Apply warning sign 6. Would a reader need to search again? If yes, find what is missing.
5. Check the Who. A real named author, a byline, and a profile that gives background. If invented, fix that first.
6. Answer all three How questions, not just the first. Disclose automation, explain how it was used, and explain why.
7. Get an outside opinion. Google writes: “Beyond asking yourself these questions, consider having others you trust but who are unaffiliated with your site provide an honest assessment.”

What Is This Document Not?

• Not a list of ranking factors. None of the 38 is a confirmed signal, and Google does not claim otherwise.
• Not a compliance checklist. Treating 38 questions as boxes to tick to rank is the search-engine-first behaviour warning signs 1 and 7 describe. Most “helpful content checklist” articles sit inside that contradiction.
• Not a requirement to satisfy all of E-E-A-T. Google states the opposite in the same sentence.
• Not quantified. No thresholds, weights, or percentages anywhere, including for effort and oversight.
• Not about the Helpful Content Update. That was a specific ranking system since it folded into core ranking. This is standing guidance.
• Not measurable. You cannot report movement on any of these, which is why they lose to metrics that move.

Conclusion

The article is 38 uncomfortable questions and two sentences most people only half-quote. Trust is the outcome; experience and expertise and authoritativeness feed it, and you do not need all three.

The newest useful part is the effort standard. Google does not call AI-assisted content effortless; it names bulk generation without oversight or curation as the thing that is, and it credits human work on “the systems powering it” as effort. That is a buildable instruction rather than a warning.

Take one page and answer two questions. Does it contain an easily-verified factual error, and would a reader need to search again afterwards? Both come straight from Google’s list, both take ten minutes, and between them they catch most of what is actually wrong.