Helpful content & E-E-A-T
Google's Spam Policies: What Content Teams Need to Know
The Google spam policies that content teams can actually break, explained with the line between normal practice and abuse: scaled content, site reputation abuse, expired domains, doorways, thin affiliate pages and more, plus how to check and recover.
8 min read
Google’s spam policies list the practices that can get pages demoted or removed from Google Search. Most of them, such as malware, cloaking and hacked content, are a security or engineering concern. A handful are things content teams can drift into while doing normal work: publishing at scale, templated location pages, partner content, affiliate pages and link building. This guide explains how the policies are enforced, which ones a content team can actually break, where the line sits between normal practice and abuse, a self-check for programmes at risk, and what to do if you get a manual action.
The spam policies are the floor. The bar above it, what makes content good rather than merely allowed, is in Google’s helpful content guidance: a working rubric.
How the policies are enforced
- Automated systems first. Google’s ranking systems guide describes spam detection systems, including SpamBrain, that are updated regularly. Their effects arrive without notice, often around announced spam updates.
- Human review and manual actions. A reviewer can apply a manual action, which appears in Search Console’s Manual Actions report and comes with a message. It can affect part of a site or all of it.
- The list is not exhaustive. The policy page says Google may act on practices it does not list, and that moving violating content to a new subdomain, folder or site to escape action is itself a violation (policy circumvention).
- They cover AI answers in Search too. Google’s documentation changelog records that, from May 2026, the policies also apply to generative AI responses in Google Search.
Which policies concern a content team
The policy page currently lists these practices. The risk column is our judgment of how likely a typical content or SEO team is to cross the line in normal work, not a measure of how serious each one is.
| Policy | Content team risk | Where it usually comes from |
|---|---|---|
| Scaled content abuse | High | Programmatic pages, mass AI drafting, auto-translation |
| Site reputation policy | High, for publishers | Partner, coupon, affiliate or white-label sections on a strong domain |
| Doorway abuse | High | City, region or keyword-variant landing pages |
| Thin affiliation | Medium | Affiliate pages built from merchant descriptions |
| Link spam | Medium | Paid or exchanged links, guest posts placed for links |
| Expired domain abuse | Medium | Buying an old domain to host unrelated content |
| Keyword stuffing | Low today | Old pages, location lists, footers full of keywords |
| Scraping | Low | Republishing others’ content without adding anything |
| User-generated spam | Low to medium | Unmoderated comments, forums, profiles |
| Hidden text and link abuse | Low | Text hidden to rank, not tabs or accordions that users can open |
| Cloaking, sneaky redirects, misleading functionality, malicious practices, hacked content, machine-generated traffic | Not usually a content decision | Engineering, security, ads and third-party scripts |
The page also lists other grounds for demotion or removal: legal removals, personal information removals, policy circumvention, and scam and fraud.
Scaled content abuse
Google defines it as generating many pages mainly to manipulate search rankings rather than help users, however the content is made: by AI, by people, or both. Its examples include using generative AI to produce pages without adding value, scraping or lightly rewriting feeds and search results, stitching content from several pages together, spreading content across sites to hide the scale, and keyword pages that make no sense to a reader.
The test is purpose and value, not volume or tooling. Plenty of legitimate sites publish thousands of templated pages. The difference looks like this:
| Practice | Normal | Abuse |
|---|---|---|
| Templated pages | Each page carries data a user wants and cannot get from the next page: a price, a schedule, a specification | The same paragraph with a name swapped in, so the only difference is the keyword |
| AI-assisted drafting | A person who knows the topic shapes, checks and adds to every page; facts are verified | Drafts published in bulk with light edits, on topics nobody on the team knows |
| Translation | Localized and reviewed for each market, with local details | Machine-translated at scale and published unreviewed |
| Topic coverage | Pages your audience needs, in areas you know | A page for every keyword variant with search volume |
Google’s guidance on generative AI content is consistent with this: using AI is fine, using it to mass-produce pages without adding value is not. For how to scale without crossing the line, see how to scale content production without sacrificing quality.
Site reputation policy
This policy, called site reputation abuse in older articles, covers third-party content published on a site mainly to benefit from the host site’s ranking signals: for example coupon, payday-loan or product-review sections run by a partner on a news or education site. Google’s November 2024 clarification says it applies regardless of whether the host is involved in or oversees the content, which closed the white-label and licensing arrangements many publishers relied on. The same post says freelance content is not a violation in itself.
The policy page lists factors Google’s reviewers weigh, such as whether the content is presented and held to the same standard as the rest of the site, and lists what is not a violation: wire services and press releases, syndication, user-generated content and forums, columns and opinion, clearly labelled advertorials, and properly qualified affiliate links. Moving such a section to a subdomain or another folder does not help; Google treats that as circumvention.
One regional change: following discussions with the European Commission, Google stopped applying site reputation manual actions to results in the European Economic Area from 30 August 2026, according to its announcement. Outside the EEA nothing changed. Treat this as an enforcement detail, not permission: the content is the same problem everywhere.
Doorways, expired domains and thin affiliation
- Doorway abuse is many similar pages that each target a variation of a search and funnel visitors to one destination. Location pages are the classic case. A page per office with its own address, team, hours and local details is useful. A page per town you could serve, with only the town name changed, is a doorway.
- Expired domain abuse is buying an expired domain and using its history to rank content that has little to do with the original site and little value. Acquiring a domain and redirecting it to the closest equivalent content on your own site is a different case; publishing unrelated content on it is the risk.
- Thin affiliation is affiliate pages that copy merchant descriptions or reviews without adding anything of your own. Affiliate pages with original testing, comparisons and opinions are fine; Google asks that the links be qualified, for example with
rel="sponsored", as described in Qualify your outbound links.
Links, stuffing and the rest
- Link spam includes buying or exchanging links that pass ranking credit, and guest posts or articles placed mainly for their links. Paid and sponsored links are fine when qualified with sponsored or nofollow.
- Keyword stuffing is filling a page with keywords or numbers out of context, such as blocks of city names. Old pages and footers are the usual place to find it.
- Scraping is republishing others’ content, even reworded, without adding value. Quoting with attribution and commentary is not scraping.
- User-generated spam is spam left by visitors in comments, forums or profiles. Moderate it, or qualify user links with ugc or nofollow.
- Hidden text means text hidden to influence ranking. Content in tabs or accordions that users can open is not hidden in this sense.
A self-check for content programmes
Run this for any programme that produces many similar pages, or any section built with or by a third party:
Is this programme at risk?
Would the pages exist if search engines did not?
Does each page carry something a user cannot get from its siblings?
Does someone who knows the topic check every page before it goes live?
Is any section written or controlled by a third party?
Are paid, affiliate and user links qualified?
Search ConsoleIf you get a manual action
- Read the Manual Actions report. It names the issue and whether it affects the whole site or part of it, often with example URLs.
- Fix every affected page, not a sample. Google says fixing only some pages does not lift the action for any of them.
- Make sure Google can crawl the fixed pages: not blocked by robots.txt, a login or noindex, unless removal is the fix.
- Request a review in the same report. Explain the problem, what you changed and how you will prevent it, with evidence. Reviews take from several days to weeks, and longer for link issues. Do not resubmit while a request is pending.
- For site reputation actions, Google’s November 2024 post says noindexing the content is not enough on its own: you also need to request reconsideration, and you should not redirect the old URLs to a new location.
Algorithmic demotions come with no report and no reconsideration. The fix is the same in substance: remove or improve what violates the policies, then wait for Google’s systems to reassess the site, which can take months.
Where Morrison fits
Morrison’s content analyzer scores every page in your site index on eight dimensions taken from Google’s helpful-content and spam guidance, including a Spam score and an Originality score. On a section of templated pages, export the URL grid and sort it by Originality: a long run of low scores across near-identical pages is what to look for. Chat can then read a sample of the pages side by side to confirm. It does not tell you whether Google has taken action; that is in Search Console’s Manual Actions report. See the content analyzer and duplicate content detection.
Key takeaways
- Most policies are not about content, but a few are: scaled content, site reputation, doorways, thin affiliation and link spam.
- The test is purpose and value, not whether AI or templates were used.
- Third-party sections are the publisher’s risk, even with first-party oversight; moving them is circumvention.
- Self-check programmes before scaling them, not after a spam update.
- Manual actions need every affected page fixed and a documented reconsideration request.
Sources and further reading
- 1Spam policies for Google web searchGoogle Search Central
- 2A guide to Google Search ranking systems (spam detection systems)Google Search Central
- 3What web creators should know about our March 2024 core update and new spam policiesGoogle Search Central Blog
- 4Updating our site reputation abuse policyGoogle Search Central Blog
- 5Update to the site reputation policy in the EEAGoogle Search Central Blog
- 6Manual Actions reportSearch Console Help
- 7Latest documentation updatesGoogle Search Central
- 8Qualify your outbound links to GoogleGoogle Search Central
- 9Creating helpful, reliable, people-first contentGoogle Search Central
- 10Google Search's guidance on using generative AI content on your websiteGoogle Search Central
CEO, Morrison
Ulrich is CEO of Morrison and founded Bonzer in 2017, growing it into one of Scandinavia's leading SEO agencies with 900+ clients across Copenhagen, Oslo, and Stockholm. At Morrison he leads strategy, operations and go-to-market, bringing years of hands-on SEO and content work to the platform side of the business.