Transcript
Intro
On the fifth of October 2026, Google quietly added a sentence to its helpful content guide that a lot of sites should read twice. Fabricating creator profiles, with AI generated headshots, made up names or false credentials, is a form of deception. And any deception, Google says, is a signal of a low quality page. Somewhere, a stock photo expert just lost his only job.
For years the advice ran the other way. Add an author box, add a face, add credentials, because E-E-A-T. Plenty of sites filled those boxes with people who don't exist. Google has now named the practice in writing, which tells you where it's looking.
So today: Google's current wording on helpfulness, the three spam policies that actually get sites removed, and what E-E-A-T really is. By the end you'll be able to tell a quality problem from a policy problem, and they need very different fixes.
Helpfulness is core ranking now
First, a correction to a lot of current advice. There's no separate helpful content update anymore. Google's ranking systems guide files the helpful content system under retired systems. In March 2024 it became part of the core ranking systems. The last update with that name on Google's status dashboard was September 2023.
Why does that matter? Because people are still waiting for the next helpful content update to recover. It isn't coming. Helpfulness now gets assessed whenever the core systems run, so gains and losses both show up with core updates, mixed in with everything else Google changed that week.
So what does Google mean by helpful today? The guide is unusually concrete now. Quality raters are trained to judge four things about a page's main content: effort, originality, talent or skill, and accuracy. These aren't secret ranking factors. They're the yardstick Google uses to check whether its ranking systems are doing their job.
The same guide gives you a quick self test: who, how and why. Who made the content, and is that clear to a reader? How was it made? Did someone actually test the product? And why does it exist? If the honest answer to why is to attract search visits, Google says that's not what its systems want to reward.
Then there's a list of warning signs, and some of them read like a description of half the affiliate sites you've ever seen. Did you enter a niche topic without real expertise, mainly because you thought you'd get search traffic? Are you publishing lots of content on many different topics, hoping some of it performs? Both describe a strategy, not a page.
Effort first. Google gives examples. Original analysis, translating a poem by hand, building a custom interactive map for a game: high effort. Generating pages from feeds, or pumping out large amounts of AI text with no oversight: little to no effort. And one line people skip: crediting your sources doesn't replace original effort.
Accuracy has a stricter version. For topics that can seriously affect people's lives, what Google calls your money or your life, content must be highly accurate and match established expert consensus. The rater guidelines added a category in September 2025 for government, civics and society. A film site rarely lives here. A clinic, a law firm or a finance blog always does.
Let's score a page. This is a ReelRadar where to watch page, from our fictional film site, the kind every film site has. Streaming services pulled from a feed. A synopsis rewritten from the studio's press text. The cast list. And a release date for the sequel, which the studio hasn't confirmed.
Against Google's four attributes, it struggles. Effort: the feed did the work. Originality: every competitor has the same synopsis and the same service list. Skill: nothing on the page needed any. And accuracy is the worst, because the sequel date is a guess dressed up as a fact.
That last one isn't our invention. Google's own warning signs include this exact case: suggesting there's a release date for a movie or TV show when one isn't confirmed. Of all the examples Google could have picked, it picked movies and TV shows. We're trying not to take it personally. Verdict on guessing sequel dates for traffic: placebo.
Now the rebuild. ReelRadar checks availability by hand on a stated date and notes where the UK and the US differ. It adds which version each service carries, like the theatrical cut or an extended one. It writes its own short take on the film. And it swaps the sequel date for what's actually confirmed, with a link to the studio's announcement.
Same page, same topic. The difference is that it now holds things you can only know by doing the work, which is exactly what those four attributes reward. Notice we didn't add a single word for length. Google's guide settles that one in brackets: does Google have a preferred word count? No, we don't.
One more warning sign from that list, because we still see it every month. Are you changing the date of pages to make them seem fresh, when the content hasn't substantially changed? Bumping the date isn't an update. If you change the date, change the content. Verdict on date bumps: placebo.
AI content and fake authors
Now the claim that refuses to die. Google penalises AI content. Write it with AI and it gets buried.
Read what Google actually wrote. Its generative AI guidance, updated in October 2026, says AI can be particularly useful for researching a topic and adding structure to original content. The problem it names is generating many pages without adding value for users. And the rater guidelines say the use of AI tools alone does not determine the level of effort. The axis is value and scale, not the tool. Verdict: placebo.
That doesn't make AI safe by default. Google is blunt: using automation, AI included, to produce content mainly to manipulate rankings is a spam violation. And its AI guidance says fact check everything, including titles, meta descriptions, structured data and image alt text, because those show up in search too.
Which brings us back to fake authors. The logic is simple. A real editor with a real name and history is something Google actively encourages: add accurate authorship information, such as bylines. A made up expert is deception, and deception makes the whole page untrustworthy, to readers and to Google's automated systems.
The three site-killers
Now the part that matters most for risk. Most quality problems cost you rankings slowly and quietly. A small set of spam policies can get a site or a section removed. For content sites today, three of them do most of the damage: scaled content abuse, expired domain abuse, and the site reputation policy.
Scaled content abuse first. Google defines it as many pages generated mainly to manipulate rankings and not to help users, and the key phrase is no matter how it's created. Its examples: AI tools churning out pages with no added value, scraped content that's synonymised or translated, and content stitched together from other pages.
For a site like ReelRadar, this is the line to watch. A page for every film in a database is fine if each page carries something real, like checked availability. The same template filled with rewritten synopses for fifty thousand films is the textbook case. And human writers doing it at scale fail the exact same test.
Translation is the other common trap. Machine translation isn't banned, but Google's scaled content examples include automated transformations like translating where little value is added. We've rebuilt this from the other side: a dental clinic that came to us running eleven languages of machine translation. Proper translations were the foundation of its growth, which you saw in the last episode.
Expired domain abuse second. Buying an old domain is allowed. Google said so in March 2024: it's fine to use an old domain for a new, original site built to serve people first. The abuse is buying one and filling it with low value content to cash in on its past. Google's favourite example: casino content on a former elementary school site.
Why does this one matter beyond content sites? Because expired domains are also a link building tool, and that's a different conversation. Consolidating a relevant old domain's strength into a site is precision work we cover in Part 2, risks included. What this policy targets is narrower: an old domain kept alive with junk to exploit its history.
Third, and the one that changed most this year: the site reputation policy. It covers third party content published on a host site mainly to borrow that host's ranking signals. Coupon sections on news sites, casino pages on medical sites, you know the genre. Google is clear that third party content on its own isn't a problem.
Here's what changed. In November 2024, Google's line was hard: no amount of first party involvement alters the third party nature of the content. That sentence is gone. Since August 2026, a reviewer asks whether the host had enough input, editorial oversight or contribution for the content to be fully integrated with the site.
The factors Google lists: how the content is presented, whether its quality matches the main site, whether authorship and responsibility are clear, and whether the same content appears on other sites. Google adds that none of them is necessary or sufficient on its own. The reviewer weighs them together.
And there's now a regional split. Following discussion with the European Commission, from the thirtieth of August 2026 a site reputation manual action no longer applies for people searching in the European Economic Area. There, the section may be separated so it ranks on its own over time. Outside the EEA, a manual action still hits that section directly.
These policies are enforced two ways: automated systems and, when needed, manual review. Spam updates are when the automated side shifts. Google's status dashboard lists four spam updates in 2026: March, June, August and September, and the September one took almost two weeks to roll out. Note the dates. They come back when we diagnose traffic drops next episode.
Quiz: would this get actioned?
Game time. Four cases, straight from Google's own policy page. Pause, make your call, then watch the stamp. Case one: a publisher runs a coupons section with a specialist partner, on a different content system, in a subfolder, linked from the homepage, curated, commercial nature disclosed, publisher named as responsible. Safe, or actioned?
Google's answer: unlikely to take action. Look at what didn't matter. A different content system and a third party partner were both fine, because the section is curated, disclosed and integrated into the publication.
Case two: a globally known business publication hosts an article linking to a marketplace that sells CBD oils. No author, no responsible editor, presentation that doesn't match the publication, and the text is copied from another site. Safe, or actioned?
Google's answer: likely to take action, for searches outside the EEA. All four factors fail at once. This is what parasite content usually looks like in the wild.
Case three: a news site starts a cooking section with affiliate links to grocery stores. A freelancer writes it, interviewing guest chefs. The freelancer is named, the site carries a statement of editorial responsibility, and the same freelancer writes for other publishers too.
Google's answer: unlikely to take action. The content is original to that site, authored and overseen. Writing for other publishers doesn't change that. Case four is the easy one: an educational site hosting sponsored payday loan reviews that the same third party spreads across the web. Google lists that as inconsistent with the policy.
So the old shortcut, putting third party content in a subdomain or subfolder and calling it separate, is a placebo. Location was never the test. Integration is.
E-E-A-T: a rubric, not a score
Now E-E-A-T, which has its own folklore. The claim: Google gives every site an E-E-A-T score, and you raise it with author boxes, an About page and a wall of credentials.
E-E-A-T comes from Google's search quality rater guidelines, a hundred and eighty two pages, current edition from September 2025. Raters use it to judge sample results, and Google says their ratings don't directly impact ranking. They benchmark quality and help test changes. In 2023 alone, Google ran over seven hundred thousand search quality tests.
And Google's helpful content guide says it directly: E-E-A-T itself isn't a specific ranking factor. Google uses a mix of factors that can identify content with good E-E-A-T, and weights them more on topics touching health, money or safety. The 2024 leak, which shows what Google stores but not how it's weighted, documents no author reputation score either, only a flag marking which entity wrote a document. Verdict: placebo.
So which part of E-E-A-T should you care about? The guidelines answer that. Trust is the most important member of the family, because untrustworthy pages have low E-E-A-T no matter how experienced, expert or authoritative they seem. Their example is a financial scam run by a highly skilled scammer. Expert, yes. Trustworthy, no.
Which is why the About page fix is a placebo. An About page is useful when it tells the truth about who runs the site. It can't create trust the content hasn't earned. Raters look at what a site says about itself, at what others say about it, and at whether the content holds up.
The guidelines get specific. Raters must do reputation research on every page quality task, looking at what independent sources say about the site and its creators. And when a site's description of itself disagrees with reputable independent sources, raters are told to trust the independent sources. Your About page doesn't get the last word.
What actually builds trust is slow and unglamorous. Content that shows the work. Facts that hold up. A real person or a real team accountable for it. And other sites talking about you without being paid to.
Your brand as an entity
One last piece ties this together. Your brand is an entity too, just like a film or an actor. Search engines and AI assistants build their picture of it from everything written about it: your site, your profiles, directory listings, press, reviews. If those descriptions disagree, the picture comes out blurry.
So write one plain sentence that says what you are, and use it everywhere you control. Same name, same category, same audience. That's our field work, not a Google rule. In our experience it's the cheapest way to stop assistants describing a brand wrongly, and it takes an afternoon.
Then ask the question that decides your fix. Quality problem or policy problem? A quality problem is a slow, quiet decline: improve the content and wait for the core systems. A policy problem usually arrives in Search Console as a manual action: fix the cause, then file a reconsideration request.
One catch, so that split doesn't fool you. Google detects policy violations with automated systems and, when needed, human review. Only human review produces a manual action message. So a site can be demoted by spam systems with no message at all, and it'll look like a quality decline. If the drop lines up with a spam update on the dashboard, take the policy question seriously, message or not.
Wrap up
Here's the episode on one screen. Google's current wording is more concrete than it's ever been, and most of the folklore around it doesn't survive a first reading of the actual documents.
The helpfulness self audit sheet, built from Google's own questions, is free on the episode page at irongoo.com. Run your top ten pages through it before your next content sprint, and fix the ones that fail on accuracy first.
