The site reputation abuse playbook for multi-region sites
On August 30 Google split its site reputation abuse policy along a geographic line: the same manual action now suppresses a section for a searcher in Chicago and does nothing to it for a searcher in Cologne. Here is how to govern third-party content when enforcement has two answers.
PLAYBOOKCONTENT GOVERNANCE5 PLAYS
Google's site reputation abuse policy stopped being one policy on August 30, 2026. Manual actions still get issued, notifications still arrive in Search Console, reconsideration requests still exist. But the consequence of the action now depends on where the person searching happens to be. Outside the European Economic Area a manual action directly suppresses the affected section. Inside it, the suppression does not apply, and Google instead separates that section in its systems so it can rank independently over time.
TL;DR · 60 SECONDSOne policy, two enforcement outcomes, split by searcher location. Non EEA searchers see the section suppressed. EEA searchers see it competing on its own merits, detached from the host domain's authority. Prior EEA manual actions are being lifted, and Google says the separation itself is not a ranking signal. The practical consequence for any company hosting third-party, partner or affiliate content is that a single global governance rule no longer describes your risk, and a section that looks fine in your European reporting can be suppressed everywhere else.
If your site hosts partner deals, sponsored subfolders, affiliate reviews, syndicated marketplace listings, a licensed coupon section or a white labeled comparison engine, this is now a governance question with a map attached. And the pressure runs in an uncomfortable direction: the region where enforcement got softer is also the region where the softening came from regulatory scrutiny rather than from Google deciding the practice was fine.
Aug 30, 2026
the date the geographic split in manual action enforcement took effect
4
factors Google publishes for evaluating whether hosted content is site reputation abuse
2
enforcement outcomes now possible from one manual action, decided by searcher location
0
change to the underlying policy definition of what counts as abuse
Read that last number carefully, because it is the one most teams will get wrong. Nothing about what Google considers abusive changed. What changed is what happens to you in one region when Google decides you did it. Treating this as permission is the expensive misreading available this quarter.
What Google changed on August 30, and what it did not
The documentation now carries two branches. For searchers outside the EEA, Google's wording is that a manual action regarding the site reputation policy will directly affect search results for the portion of the site affected. For searchers inside the EEA, the impact of the manual action will not apply, and the affected section of the site may be separated in Google's systems so that, over time, it ranks independently.
That separation is the part worth slowing down on. It does not delete the section from the index and it is not a penalty in the ordinary sense. It severs the section from the host domain's accumulated authority and makes it compete against comparable content on its own signals. A casino subfolder on a national newspaper stops borrowing the newspaper's standing and starts competing against other casino content. For most parasite arrangements that is a larger loss than a ranking suppression would be, because the borrowed authority was the entire commercial premise.
ELEMENT
OUTSIDE THE EEA
INSIDE THE EEA
Manual action issued
Yes
Yes
Search Console notification
Yes
Yes
Ranking impact of the action
Direct suppression of the affected section
No suppression applied
Section separated from host domain in Google's systems
Not the stated mechanism
Yes, ranks independently over time
Prior manual actions
Stand
Being removed, with no lingering disadvantage stated
Reconsideration request available
Yes
Yes, plus access to mediation
Requirement to noindex the affected section
No, but suppression is the practical outcome
Explicitly not required
Google's public position on why is short. A spokesperson framed it as unchanged conviction rather than retreat, saying European users are no less frustrated by parasite SEO and other deceptive, pay to play tactics that degrade search results, and that the company stands by its site reputation policy. The trigger was the European Commission's scrutiny of how the policy was being applied. Google has also been explicit that it is not pausing or weakening core spam enforcement in Europe, only adjusting how manual penalties land there. Search Engine Land documented the change as it went live.
THE TRAP IN ONE SENTENCEA regulatory carve out is not an editorial endorsement, and the enforcement mechanism that replaced suppression in Europe removes the exact thing parasite arrangements are bought for. If your partner content only works because it sits on your domain, separation ends it just as surely as a penalty would, only more quietly and with a worse diagnostic trail.
Why the site reputation abuse policy now has two answers
Most enterprise sites are not running deliberate parasite schemes. They are running things that grew: a deals hub a commerce team licensed in 2023, a partner directory marketing inherited, a syndicated jobs feed nobody owns, a reviews section built by an agency that no longer has the login. None of it was designed to abuse anything. All of it is now sitting inside a policy whose consequences differ by continent.
The reason a split policy creates real operational risk rather than mild inconvenience is that reporting and enforcement no longer line up. A European marketing lead pulls the numbers, sees the section performing, and reports it as healthy. The same section is suppressed for every searcher in the United States, Canada, Australia and Japan. Unless someone segments by country, the average hides the split, and averages are what most dashboards display by default. This is the same measurement failure we wrote about when rank tracking data accuracy quietly degraded: the number is not wrong, it is answering a different question than the one being asked of it.
EEA traffic, separated section competing on own signals55%
Non EEA traffic, suppression applied12%
Blended global average as most dashboards report it34%
Pre action baseline100%
Illustrative scenario for one hosted deals subfolder receiving a manual action, showing indexed sessions by region against a pre action baseline of 100. Regional shares are illustrative and not measured data.
The blended figure is the dangerous one. It looks like a bad quarter rather than a policy event, which means it gets escalated to the content team as a quality problem and answered with more content. Six weeks later the section is bigger, the manual action is still there, and nobody has read the Search Console message because the person who owns that property left in March.
01Ownership is the usual root causeHosted third-party sections almost always sit outside the org chart that governs the rest of the site. They were bought by commerce, built by an agency, and are monitored by nobody. Before any technical work, name a single accountable owner with access to Search Console for the property. Half the remediation problems on these sections are actually access problems.
02The four factors are testable, not vibesGoogle publishes what it evaluates: presentation consistency, quality standards, authorship clarity and content duplication. Every one of those can be checked by a person with a browser in under ten minutes per section. Treat them as an inspection checklist, because that is what they are.
03Separation is harder to detect than suppressionA suppressed section falls off a cliff in one week. A section separated from domain authority decays over weeks as it loses the borrowed signal, which reads as a slow content decline. If you are only watching for sharp drops, the European failure mode is invisible to you.
04The regulatory position could move againThis change came from Commission scrutiny, not from a shift in Google's stated view of the practice. Google says it stands by the policy. Building a 2027 revenue plan on the assumption that the EEA carve out is permanent is a bet on a regulatory process, not on a search strategy.
The four factors behind every site reputation abuse policy call
Google's own criteria are the most useful thing published on this subject, and they are underused because they read like boilerplate. They are not. Each one maps to a concrete artifact you can inspect on a page.
Presentation consistencyDoes the section look and behave like the rest of the site? Shared header, footer, typography, navigation and page furniture. A subfolder that renders in a different template with a different grid is announcing that it was bolted on, and that announcement is legible to a human reviewer in about two seconds.
Quality standardsDoes the content in this section meet the bar the main domain sets? Thin pages, spun descriptions, machine translated copy and unmaintained listings all read as a different quality regime operating under the same roof. The comparison is against your own site, not against the category.
Authorship clarityIs it obvious who made this and who is responsible for it? Named authors, an editorial contact, a disclosed commercial relationship. Content that hides its origin or, worse, contradicts itself across a byline and a disclosure, is the single clearest signal in the set.
Content duplicationDoes this appear identically or near identically on other domains? Syndicated marketplace inventory and reseller feeds fail here routinely, and they fail without anyone acting in bad faith. If the same 4,000 listings appear on eleven partner sites, none of those eleven is the answer to anything.
Google's published examples make the boundary concrete in both directions. A publisher that curates partner deals, discloses the relationship plainly, keeps editorial control, integrates the section into site navigation and collaborates meaningfully on the content is described as the safe pattern. So is affiliate content written by a freelancer specifically for that publication, with an identified author, editorial responsibility, consistent branding and clear affiliate disclosure. The failing pattern is content with no identified author, hidden commercial intent, placement outside normal editorial sections, weaker quality and design than the host, duplication from a third-party marketplace, and no route to it from the main navigation.
“Our European users are no less frustrated by parasite SEO and other deceptive, pay-to-play tactics that degrade search results, and we stand by our Site Reputation Policy.”
Note what separates the two lists. It is not whether money changed hands, and it is not whether the content is commercial. It is whether the host is actually publishing the content or merely renting the address. Every play below is a way of moving a section from renting toward publishing, and the ones that cannot make that move should be shut down rather than optimized.
The five plays
1WEEK 1Inventory every square metre of your domain you do not editorially control
THE MOVES
Crawl the full domain and diff the URL set against the pages your CMS actually owns. Anything in the crawl that your CMS cannot account for is hosted third-party content, whether or not anyone calls it that. Reverse proxies, subfolder integrations and headless partner apps all surface this way and only this way.
Add the sections your crawl cannot see: subdomains, country folders, licensed tools, embedded marketplaces, jobs and events feeds, and any subfolder served by a partner's infrastructure. Ask finance for a list of vendors billing against web properties, because contracts find sections that crawls miss.
For each section record six fields: URL pattern, page count, who writes the content, who profits from it, which contract governs it, and whether it appears in the main navigation. The last field predicts the outcome more often than any other.
Pull Search Console manual actions for every property and every domain property variant, not just the primary one. A manual action delivered to a property nobody has opened in a year is the most common way this goes unnoticed for months.
Rank the inventory by indexed pages times organic sessions. You are looking for the sections that are large and load bearing, because those are the ones where a wrong call is expensive in both directions.
DONE WHENA single sheet lists every hosted or third-party section with all six fields populated, an owner named for each, and any live manual actions attached. Evidence: the sheet, plus a screenshot of the manual actions panel for each property.
2WEEK 2Score every section against the four published factors
THE MOVES
For each section, open three representative pages side by side with three pages from the core site. Score presentation consistency, quality standards, authorship clarity and duplication as pass, borderline or fail. Do it with a person, not a tool, because a reviewer at Google will do it with a person.
Test duplication concretely: take three exact sentences from a listing page and search them in quotes. If the same text appears on other domains, the section fails this factor regardless of licensing agreements, because Google is evaluating the result page rather than your contract.
Test authorship by asking someone outside marketing to say who wrote a page and who benefits from it, using only what is on the page. If they cannot answer in fifteen seconds, that factor fails.
Test navigation reachability by starting at the homepage and clicking. If a section is only reachable from a sitemap or an external link, it is outside the editorial structure of the site in the way the policy describes.
Record a one line verdict per section: integrate, remediate, or exit. Three buckets only. Sections with two or more fails go straight to exit unless someone will fund a genuine editorial rebuild this quarter.
DONE WHENEvery section carries four factor scores and one of three verdicts, signed off by the accountable owner. Evidence: the scoring sheet with page-level screenshots for any borderline call.
3WEEKS 3-5Rebuild the keepers into content you can honestly claim as your own
THE MOVES
Move the section onto the site's own templates. Same header, footer, navigation, typography and page furniture as the core site. This is usually a two day front-end job that has been deferred for two years, and it resolves the presentation factor by itself.
Put a real named author or an editorial team byline on every page, with a link to a profile that exists, and add a plain disclosure of the commercial relationship above the fold rather than in a footer. Write the disclosure in the same voice as the rest of the site.
Replace duplicated third-party copy with original material. Where inventory has to stay syndicated, add original editorial around it and make the page's value the editorial, not the feed. If nobody will fund original writing for a section, that is the section telling you it belongs in the exit bucket.
Wire the section into the site's internal linking so it is reachable from relevant category and service pages, and links back out to them. Apply the same discipline described in [our guide to internal linking for AI search](/blog/internal-linking-guide-ai-search), because a section with no inbound internal links is orphaned from the reader's perspective as well as the crawler's.
Set a maintenance owner and a review cadence in the CMS, not in someone's calendar. Sections fail this policy by drifting, and drift is what happens to content with no owner.
DONE WHENEvery keeper section passes all four factors on re-inspection by a reviewer who did not do the remediation, and is reachable from the homepage in three clicks. Evidence: before and after screenshots plus the re-inspection scoring sheet.
4WEEK 5Exit the sections that cannot be honestly claimed, cleanly
THE MOVES
Give notice under the contract before touching anything technical. Pulling a partner section without notice is a commercial problem that will land on someone senior, and it converts an SEO decision into a legal one.
Decide per URL between redirect, consolidate and remove. Redirect to the closest genuinely relevant page where one exists. Where nothing relevant exists, return 410 rather than 404, so the removal is a statement rather than an ambiguity.
Never redirect a large removed section wholesale to the homepage. It is the reflex, it teaches Google nothing, and it moves the quality question onto your most important URL.
If the section carries a live manual action, file the reconsideration request only after removal or remediation is complete and verifiable, and describe what was done in specifics rather than intentions. Inside the EEA, note that mediation is also available as a route.
Track recovery with a fixed URL set rather than sitewide totals, and split the reporting by region from day one so the two enforcement outcomes stay visible as separate lines.
DONE WHENExited sections return 410 or redirect to a genuinely relevant page, contract notice is on file, and any reconsideration request is submitted with a specific remediation description. Evidence: a URL-level status export and the submitted request text.
5ONGOING, MONTHLYRun region-split monitoring so the two policies stay visible
THE MOVES
Split every report on hosted sections by EEA and non EEA. One blended line is worse than no line, because it converts a policy event into a vague performance story. Country dimension in Search Console does this in about ten minutes of setup.
Watch for the separation signature specifically: a section holding position in EEA countries while decaying in the United States, Canada, Australia and Japan over four to eight weeks. That divergence pattern is diagnostic and it does not look like an algorithm update.
Re-score any section against the four factors whenever its template, its author model or its content source changes. A partner switching feed providers is a policy event even though it arrives as a ticket.
Add a standing agenda item wherever new partner and affiliate arrangements get approved, and make the four factors a gate on signing rather than a review after launch. The cheapest version of this work happens before a contract exists.
Keep a dated log of what Google publishes on this policy. Enforcement moved once this year under regulatory pressure and can move again, and a written history is what stops the next team from treating the current state as permanent.
DONE WHENRegion-split reporting is live for every hosted section, the four factor gate is written into the partnership approval process, and a dated policy log exists with a named keeper. Evidence: the dashboard, the approval checklist, and the log.
How to measure a section that ranks independently
Separation creates a measurement problem that suppression does not. When a section is suppressed you get a cliff, and cliffs are easy. When a section is detached from domain authority you get a slope, and slopes get attributed to whatever else was happening that month.
Build the baseline before you need it. For each hosted section, record current position distribution, indexed page count and organic sessions, split by region, and keep the raw export rather than a dashboard view. Then watch three things: whether the region split widens, whether the section's rankings converge toward comparable standalone sites in its category, and whether its pages stop appearing for head terms while holding long tail. That third pattern is the clearest sign that borrowed authority has been withdrawn, because borrowed authority is what wins head terms in the first place.
WHAT YOU OBSERVE
MOST LIKELY CAUSE
FIRST CHECK
Sharp drop across all regions in one week
Manual action with suppression, or a technical break
Search Console manual actions panel and server logs
EEA holds, non EEA falls over four to eight weeks
Manual action under the split enforcement
Manual actions panel, then region-split performance export
Gradual decline in all regions, head terms first
Loss of borrowed domain authority, or a core update
Compare against category competitors on the same terms
Impressions steady, clicks falling
Result presentation change, not a ranking event
Query-level CTR by position, not sitewide averages
Section absent from reporting entirely
Property access or filter problem
Confirm the section is inside the property you are reporting on
Two of those five rows are instrumentation rather than ranking, which is the usual ratio and the reason the first move is always to check the panel rather than to commission a content audit. We make the broader case for stating uncertainty in visibility reporting in the argument for error bars in AI search measurement, and a region-split policy is the strongest practical argument for it we have seen this year.
What this changes about your 2027 content plan
The strategic read is that hosted third-party content just became a governed asset class rather than a growth tactic. It has a compliance surface, a regional variance and a documented set of criteria, which is roughly where paid link buying was fifteen years ago before it stopped being a strategy anyone would write down.
For most enterprises the correct response is not to exit partner content. It is to publish it properly. Every one of Google's four factors is satisfied by doing the thing an honest publisher would do anyway: use your own templates, put your name on it, say who paid, write something original, and link it into the site like you mean it. That work also happens to be what earns citations in generative engines, which read hosted junk with exactly as much enthusiasm as a manual reviewer does. If you want the same content to work in AI answers, it needs the structure described in the enterprise standard for structured data in AI search, and structure is not something a rented subfolder tends to carry.
There is a commercial argument here too, and it is the one that usually moves the decision. A partner section that only performs because it borrows your domain's authority is a section whose value you are subsidizing and whose risk you are carrying. The revenue share rarely reflects that. Repricing those deals around what the content is actually worth on its own signals is a conversation worth having this quarter, and the separation mechanism now gives you a number to have it with.
The commerce and retail case is the sharpest version of it, since deals, coupons and marketplace inventory are where most of this content lives. We work through that tension regularly in ecommerce and retail engagements, and the pattern is consistent: the sections that survive scrutiny are the ones a real editor was already responsible for.
DO THIS NEXTOpen the manual actions panel for every Search Console property you own, including the ones nobody has logged into this year, and check the date on each message. Then run Play 1 and get the inventory onto one sheet with an owner per section. If you find a live action, split your reporting by region before you do anything else, because the shape of the damage tells you which enforcement outcome you are living with. This is standard scope in a technical SEO and GEO audit, and it is usually the first place we find revenue attached to content nobody in the building will claim.
See where you are cited today
A free snapshot audit of your rankings and AI citations before we ever talk.
Tyler leads work at the intersection of SEO and generative engines at Something Inc., helping B2B brands get ranked and cited across every major AI engine.
