Site reputation abuse and scaled content rules

A section the newsroom does not write and a program of generated pages do not look alike. In Google Search they can land in the same spam policy document. The first is what teams usually look up as site reputation abuse: coupons, loans, casinos, affiliate reviews, or a whole vertical signed by a partner and hosted on a site that already ranks. The second is a pile of URLs built with templates, with generative AI, or with both. The heading for that one is scaled content abuse.

Before you publish, or before you answer Search Console, the question is whether this project fits one of those two policies. Third parties are not banned. Neither is automation. Both are global rules for Google web search. They are not Spanish law, and this article is not legal advice. If a 2024 announcement and the current policy page disagree, the policy page wins. The English page I opened was last updated on 28 August 2026 UTC. The Spanish page of the same document was last updated on 1 September 2026 UTC (English spam policies, Spanish spam policies). The operative rules match. Sentences that do not are marked below.

Two labels, one current policy

On the current page the heading is “Site reputation policy.” The manual actions report uses that name too. Site reputation abuse is how a team searches for the entry. Quoting a 2024 definition as the working text is a mistake when the current page says something else.

Google places the introduction of the policy in 2024, to stop third-party content being published on a trusted site just to exploit that site’s reputation (post of 28 August 2026). That post explains the change for the European Economic Area. It does not replace the criteria on the policy page.

Scaled content abuse sits in the same document, under another heading. It is not the update calendar. If you want that list, it is in the 2026 guide to Google algorithm updates.

Sites that violate the spam policies may rank lower in results or not appear in results at all. Google detects violating practices with automated systems and, when it needs to, with human review that can end in a manual action (spam policies). The policies cover common practices, but Google says it may act against any spam it detects, including practices it does not name on the page. No manual action in the report is not proof that the pages are fine.

What site reputation abuse means now

The site reputation policy applies where third-party content is published on a host site mainly because of that host’s already-established ranking signals, which it has earned primarily from its first-party content. The goal of this tactic is for the content to rank better than it could otherwise on its own (site reputation policy).

Third-party content is content that’s created by an entity that’s separate from the established host site. Examples of separate entities include users of that site, freelancers, white-label services, and content created by people not employed directly by the host site (definition of third-party content). Paying an agency neither clears the case nor condemns it. “Not employed directly” is a policy example, not Spanish labor law. The Spanish page says “entidad independiente” and “personas que no están contratadas directamente.” This guide follows the English wording above.

Having third-party content alone isn’t inconsistent with the site reputation policy. It is inconsistent only if the third-party content is published on a host site mainly because of that host site’s already-established ranking signals (when there is a violation).

That “mainly because of” is the decision. It does not come out of a Spanish case file. Google’s examples of a violation are an educational site hosting sponsored payday-loan reviews written by a third party and distributed to other sites, and a medical site hosting an unintegrated, low-quality third-party page about “best casinos” so it can use the established site’s signals (examples that do not comply).

What the policy says is not this violation

Google also lists cases it does not treat as a violation of this policy. The list is not closed. The page says the examples are not exhaustive and that there will be other instances where it does not take action (examples that are not a violation).

Left outside this violation are wire services and press-release services, syndication between news publications, forums and comment sections built for users, columns, opinion pieces, and other editorial work, native advertising or advertorials whose purpose is the reader of that publication rather than ranking, and affiliate links treated properly or third-party ad units embedded on a page (exclusions from the site reputation policy).

Do not read that list as a pass. A forum can sit outside this policy and still collect user spam. The report treats that as its own problem, user-generated spam, not as site reputation (actions for user-generated spam).

“Links treated appropriately” points, in the same document, at the link spam policy. Buying or selling links for ranking violates that policy. It does not violate it to have paid links if they are qualified with rel="nofollow" or rel="sponsored" (link spam). That qualification does not answer why the section is on this domain.

Google generally applies a presumption that individual pages, including new pages, match the overall quality of other pages on the domain. If it detects that a portion of the site may be out of line with the site reputation policy, the site goes to human review. What happens in results depends on where the user is (quality presumption and human review).

What that review looks at, when it happens

According to Google, the human review is rare. The goal it describes is specific: to see whether the content on that portion was created with enough input, editorial oversight, or contribution from the host to count as integrated with the main site, so that ranking reflects how the page is presented and how a reader understands it. The review applies worldwide and uses objective factors about the control the host domain has over the page (human review instructions).

The factors it names are presentation, quality against the main site, stated or implied authorship, and whether the same text appears, identical or nearly so, on other sites. An author box can make a byline look credible. It does not clear the section. E-E-A-T and content authority is a different question. Not one of these factors is either necessary or sufficient on its own. Other evidence can matter, for example pages that are a third party’s marketing material aimed at ranking (no single factor is enough).

The contrast is three illustrations, not three case files. A coupons section built with a provider, even on another CMS, but tied into the homepage, with its own categories, a commercial notice, an editorial owner, and a contact for broken offers: unlikely that Google acts. An affiliate article on CBD oils, no author, no disclaimer, unreachable from the sections, and a copy of a marketplace page: likely that it acts on results outside the EEA. A cooking section by a freelancer who also writes elsewhere, with host oversight, a named author, and recipes chosen for that site: unlikely that it acts, even when the affiliate links look familiar (three illustrations in the policy).

“Unlikely” and “likely” do not predict your directory. If the pages look like the CBD article, authorship and the duplicate matter more than the story in the contract.

What changes for results shown in the EEA

On 28 August 2026, the Search Quality team announced an adjustment after talking with the European Commission. Beginning 30 August, manual actions under the site reputation policy have a different effect for people searching in the EEA than for people searching outside it (announcement of 28 August 2026).

According to that English post, for users inside the EEA the impact of the manual action will not apply. The affected section may be separated in Google’s systems so that, over time, it ranks independently from the rest of the site. Outside, the action affects the flagged portion, not the rest of the site. The same URL can carry the action, and that change only alters results shown to users outside the EEA (different effect inside and outside the EEA).

The policy page talks about results shown, not about where someone lives. Within the EEA, the pages may be categorized as separate from the main domain and are not subject to the manual action. The example is casino content ranking against other casino content (effect on EEA results). The Spanish version of the same post does use residence. It says that, for users who reside in the EEA, the manual action will not have an effect (Spanish announcement). The English post does not say “reside.” It says “searching,” “inside the EEA,” and “results shown.” Keep both sentences if you also read the Spanish documentation. They are not interchangeable, and neither page explains how Google locates the user. A VPN is not defined as proof.

Spain has been an EU member state since 1 January 1986 (Spain’s page on the EU site). The European Economic and Social Committee describes the EEA as the space that brings together the EU member states and Iceland, Liechtenstein, and Norway, and says the Agreement entered into force on 1 January 1994 (description of the EEA). A search from Spain is inside that group. A search from outside the EU and those three states is not. This is a search policy, not a legal exemption. The English post mentions the Digital Markets Act (post of 28 August 2026). I am not interpreting that law.

The impact of the site reputation manual action does not apply to search results shown to users in the EEA (manual actions report). Separating a portion drops the presumption that each page shares the domain’s quality. That separation does not mean the separate portion immediately loses the ranking signals of the main site. According to the same page, the fact that this categorization has been applied is not used as a ranking signal (what independent categorization means).

According to Google, it will lift previous manual actions taken under this policy on pages that appear in search results shown to users in the EEA. Classifying those pages as a separate part of the site, from that point on, is not automatic (lifting earlier actions in the EEA).

There is no obligation to apply noindex to content that is subject to a manual action outside the EEA, and not doing so is not taken into account when ranking that content in the EEA, nor as evasion or a repeated violation (FAQ on noindex and the EEA).

For sites in the EEA, Google describes a reconsideration that answers in a short timeframe and with more detail. It does not say how many days. Sites that qualify can go to mediation after the reconsideration request (reconsideration and mediation in the EEA). It does not list who qualifies. Mediation is not a trial in Spain. This geographic split is not repeated under scaled content abuse. Do not extend it.

How it shows up in Search Console

Google issues a manual action when a human reviewer concludes that pages on the site do not comply with the spam policies. The report opens like this: if a site has a manual action, some or all of that site will not be shown in Google search results. The notice arrives in the report and in the message center (what a manual action is).

For site reputation, that sentence sits next to the EEA exception. If the violation is repeated, there can be further actions, or the site’s overall ranking can be affected (repeat of the site reputation policy). A green check only says there is no manual action.

Expand the description and you get patterns. Not every page that matches those patterns is necessarily affected. The English report’s example is https://example.com/real-estate/* (patterns of affected pages). The Spanish report localizes that sample path to pisosenventa. Use the message you received.

Scaled content abuse does not have its own row in the list I read. It sits inside “Major spam problems,” with cloaking and other repeated or serious violations. The advice is to stop violating, request a review, and show what you removed and what you improved (major spam problems). A different row, “Thin content with little or no added value,” covers thin affiliate pages, content taken from elsewhere, and doorway pages (thin content with little or no value). A weak vertical can touch more than one. They are not the same action.

If you just bought a site that was already violating the policies, the report asks you to fix the problems and to say in the reconsideration that the acquisition is recent and that the site no longer violates them (a site bought with a prior action). A drop with no message in the report is a different job: measurement, demand, an update, or a technical fault. The order of that work is in how to diagnose organic traffic drops. Do not request a reconsideration for a decline that Search Console has not labeled.

What to settle before you publish the third-party section

Judge the section with the policy’s test, not with the partner’s prestige. There is a violation only if the third-party content is published on the host site mainly because of ranking signals that site already has (main test in the policy). If the honest answer is “this folder would not rank on a new domain,” you are describing the goal the policy names. If the answer is “our readers use these coupons, and the newsroom selects them, answers for them, and links them from articles,” you are closer to the example where Google sees action as unlikely. The difference is not the topic. It is the use of the host’s signals.

Before you create the directory, write down who writes it, why it lives on this domain, whether the same text is elsewhere, who signs it, and who handles a complaint. If the partner is selling the publisher’s domain, the “mainly because of” is already in the pitch. Three of that partner’s URLs on other domains beat a summary. The same header does not integrate a section. In the coupons example a broken code reaches the editor. In the CBD example there is no author and no disclaimer.

No single check is enough. A copied article in the publication’s design is still the second example. Native advertising leaves this violation when the aim is that publication’s readers, not ranking. A campaign should not become a permanent SEO section, and an invented byline is not an exit the report lists.

Where scaled content abuse comes in

Scaled content abuse is when many pages are generated for the primary purpose of manipulating search rankings and not helping users. Google describes that practice as large amounts of unoriginal content that provides little to no value for users, no matter how it is created (scaled content abuse).

The tool is not the type. The examples are not a closed list: many pages from generative AI with nothing added, scraped feeds rewritten with synonyms, translations, or other obfuscation, text stitched from several pages, extra sites opened to hide the volume, and pages that say little while repeating search keywords (examples of scaled content abuse).

A template program can fall in or stay out. The dataset and the template are discussed in how to scale pages with programmatic SEO. If the main purpose is to cover combinations so the pages can manipulate results, the program fits even when a person reviews it. If the pages help, the template or the AI is not, by itself, this abuse. That is not a promise either way. Google can act against spam that is not on the list.

Three neighboring policies block a quick self-pardon. Scraping republishes someone else’s content to manipulate ranking, without value of its own (scraping). Rewriting the manufacturer’s sheet lands there. Doorway pages target narrow queries and send people to a less useful intermediate page. Cities or regions that funnel users to one page are the example Google gives (doorway abuse). A grid of “service plus city” URLs aimed at one form can be that. Thin affiliation copies the merchant’s description. It does not cover a site that adds its own reviews or comparisons (thin affiliation).

If the site hosts the kind of content the policy calls scaled content abuse, Google asks you to exclude it from Search (exclude that content from Search). The page on controlling what you share with Google says that if users can create content, some of it may be low quality or even spam, and that indexing it can hurt ranking. Separately, it talks about a very large site, over hundreds of thousands of URLs, where you might keep the crawler off less important pages so it spends time on the rest (control what you share with Google). That is not a number of pages above which a program becomes acceptable.

What to ask a page program before you generate it

Take twenty future URLs, not the schema. If a page exists only because the combination fit the template, that is the purpose. If the searcher gets a fact that was not in the feed, the sign points the other way. Twenty URLs do not clear the other two thousand.

Combining other people’s pages, translating a feed, or swapping synonyms are examples in the policy. A model in the middle does not invent a new category, and another domain opened to hide the volume is an example too. Keyword-heavy text that does not help still misses the cut. Exclude those URLs from Search. Leaving them indexed in case they rank is the purpose the definition names. Where a part should exist as your own content, the patterns for citable content ask for a source, a scope, and a limit. They do not launder a feed, and the policy does not say they block a manual action.

If the manual action is already in the report

A reconsideration request is how you ask Google to review the site after you have fixed what a manual action or a security issue notification identified (what a reconsideration request is). It is not a letter you send before anything is wrong, and it is not an appeal against a drop with no action.

Expand the description, read the pattern, and fix every affected page. You have to fix the problem on all affected pages: doing it on only some of them does not earn a partial return to search results. If there are several actions, fix all of them, then select “Request Review” (how a manual action is fixed).

Moving the content to a new domain requires nofollow from the old site and forbids redirecting the old URLs, because the redirect can bring the problem back (move to a new domain). Use the guide to 301 and 302 redirects when a redirect is actually the right tool. Here it is not. Google generally does not follow those rel attributes, but the destination page can still be found by other means, such as a sitemap or links from other sites, and can end up crawled (rel links and discovery). That English page was last updated on 10 December 2025 UTC.

The other route is noindex. For the noindex rule to be effective, the page must not be blocked by a robots.txt file or by anything else that stops the crawler from seeing it. If the crawler never sees the rule, the URL can still appear because other pages link to it (noindex and robots.txt). Putting noindex inside robots.txt is not supported. The reputation section says the same practical thing: do not block that content with robots.txt if noindex is how you exclude it (noindex on the reputation action). A bad robots.txt deploy is a separate fault, covered in robots.txt errors that block Googlebot.

In the general process, affected pages should not require a login, sit behind a paywall, or be blocked by robots.txt or noindex, so the reviewer can see them (access during a general review). On this action, noindex is also a listed correction. Leave the page visible if you believe it now complies. Exclude it with noindex, and leave robots.txt out of the way, if exclusion is the fix. Do not do both on the same URL. I am reading two instructions on one report as applying to different fixes. There is no line that says “never combine them.”

Rebuilding the page as yours fails if the partner is still the author and only the template changed. Remove the page when it should not have existed. A large archive is a later audit: pruning outdated content.

Moving the content to a subdomain or a subdirectory of the same site may be treated as an attempt to circumvent the spam policy, and it can widen the actions. Taking it to another site that already has a reputation can move the problem if the content is still third-party. In the EEA, failure to take these steps is not, by itself, considered an attempt to circumvent or avoid this policy (warning about moving the content). That note does not license other spam. Circumvention can restrict Top Stories or Discover, or remove more sections from results (policy circumvention).

A good reconsideration request explains the exact quality problem, describes the steps taken to fix it, and documents the outcome of those corrections. In most cases the review takes several days or several weeks. Some take longer, and the report cites links as the example. Do not send another request while the pending one has not been resolved (reconsideration timing).

Google revokes the manual action if, after the review, it determines that the site is no longer violating its spam policies (condition for revoking the action). Revocation is conditional. It is not a date you can put on a calendar, and the short timeframe for EEA sites does not supply one. Outside the EEA the URLs can stay affected while Spanish traffic sits still. Measure the two audiences apart. A flat line in Spain does not prove the action is gone.

An author box does not declare compliance. The notice still lives in Google Search Console. The request has to name pages, the pattern in the message, and a change already in production.

The next step

If you have not published yet, do not request a review. Note who creates the content, whether the site employs them directly, whether the text sits on other domains, who signs it, and who answers a complaint. For a page program, paste twenty URLs and write what the searcher gets that is not in the feed. If that line will not come, those URLs stay out of the index.

If the action is already there, open the message before the CMS. Note the pattern. Do not mark every matching URL as affected, and do not ignore the ones the message names. Pick a correction the report lists, not a new subdomain on the same site. If you use noindex, confirm the crawler can see it and that robots.txt is not in the way. With the whole pattern fixed, request the review once: the problem, the steps, and the result. Wait for the answer.

This is an editorial decision. It does not predict that Google will return the URLs, and it does not rule on the contract. A valid contract can still violate the policy, and a bad contract can still describe a section the examples treat as unlikely to draw action.