Spam policies for Google web search

Author: Google | Published: 2026-08-28 | Source: article

Summary

Google’s web-search spam policies define techniques that deceive users or manipulate Search — including manipulating generative AI responses. They apply to all web results, including Google’s own properties. Detection is automated plus, as needed, human review that can yield a manual action. Violating sites may rank lower or not appear. Reports feed scalable detection; Google may act against unlisted spam types.

The catalog: cloaking; doorway abuse (sites/pages created to rank for specific similar queries that funnel users through unhelpful intermediates — including multiple city/region pages that dump into one page, and substantially similar pages closer to SERPs than a browseable hierarchy); expired-domain abuse; hacked content; hidden text/link abuse (accordions, tabs, screen-reader text are fine); keyword stuffing (unnatural lists, including “blocks of text that list cities and regions that a web page is trying to rank for”); link spam (buying/selling ranking links, excessive exchanges, automated link building, advertorials with optimized anchors, footer/widget/forum signature spam; paid/sponsor links are allowed if rel="nofollow" or rel="sponsored"); machine-generated traffic (including scraping SERPs for rank-checking without permission); malicious practices; misleading functionality; scaled content abuse; scraping; site reputation abuse (third-party content parked on a host’s ranking signals; EEA handling differs — separate categorization rather than manual action); sneaky redirects; thin affiliation; user-generated spam; plus legal/personal-info demotions, policy circumvention, and scam/fraud.

Scaled content abuse is the section this ingest exists to capture: many pages generated primarily to manipulate rankings and not help users, typically large amounts of unoriginal content with little or no value, however it is created. Named examples: using generative AI (or similar tools) to generate many pages without adding value; scraping feeds/SERPs/other content (including synonymizing, translating, obfuscation) to generate many pages with little value; stitching pages without adding value; creating multiple sites to hide scale; many pages that make little sense but contain keywords. Hosts should exclude such content from Search.

Scraping is republishing others’ content without original value (or even citation), slight synonym/automated rewrites, unimproved feeds, or embed-compilation sites. Duplicate same-site content is not listed as spam here — that lives on canonicalization.

Key Claims

Spam includes manipulating generative AI responses, not only classic blue-link ranking. (official)
Scaled content abuse = many pages whose primary purpose is ranking manipulation, little/no user value, regardless of production method. (official)
Using generative AI to generate many pages without adding value is an explicit scaled-content-abuse example. (official)
Keyword stuffing includes city/region lists a page is trying to rank for. (official)
Doorway abuse includes multiple city/region URLs that funnel to one page, and near-duplicate doorways instead of a real hierarchy. (official)
Link buying/selling for ranking is spam unless qualified nofollow/sponsored. (official)
Scraping others’ content without substantial added value is spam. (official)
Same-site duplication is not in this spam list; see What is canonicalization. (official, by omission + cross-doc)
Site-reputation abuse parks third-party pages on a host’s signals; EEA vs. rest-of-world enforcement differs. (official)

Notable quotes

“Scaled content abuse is when many pages are generated for the primary purpose of manipulating search rankings and not helping users. This abusive practice is typically focused on creating large amounts of unoriginal content that provides little to no value to users, no matter how it’s created.”

Why citable: Purpose-and-value test, production-method-agnostic. The official line against “but a human clicked generate.”

“Using generative AI tools or other similar tools to generate many pages without adding value for users”

Why citable: Named AI-at-scale example. Direct tension with Core 30 / 8-Pass AI Writing Pipeline volume pipelines.

“Blocks of text that list cities and regions that a web page is trying to rank for”

Why citable: Official example of keyword stuffing that maps onto thin service-area / city-list footers.

Connections

Entities mentioned: Google, Caleb Ulku, James Dooley Concepts referenced: Core 30, 8-Pass AI Writing Pipeline, Duplicate Content (for Service-Area Pages), Programmatic SEO, GEO / AEO (Getting Recommended by AI), Local Link Building & Authority, 250 Authority Protocol · scaled content abuse (NEW: scaled-content-abuse)

Contradictions / Tensions

Tension with Core 30 AI-at-scale. Caleb builds ~30 (or more) pages with an AI pipeline, one per GBP category/service. Official spam line is not “AI pages are banned” (see Google Search's guidance about AI-generated content); it is many pages, no added value, primary purpose = rankings. If each Core 30 page is genuinely local, first-hand, and useful, it can pass. If it is a templated keyword page with swapped city/service names, it matches this definition. Do not treat Core 30 as automatically spam or automatically safe.
Tension with Dooley duplicate location pages. Duplicate Content (for Service-Area Pages) as same-site copies is not a spam violation (What is canonicalization). Doorways (city pages that exist to rank and funnel) and city-list stuffing are. Dooley’s “copy the best page across cities” sits in the gap: not spam-by-duplication, possibly doorway-like if the pages are not a real browseable service, possibly scaled-content-abuse if unoriginal at scale. Name all three poles; do not pick one.
Link buying (official spam unless sponsored/nofollow) vs. Caleb’s ~$35 “not AI slop” links and chamber dues. Editorial/sponsorship with qualification can be legitimate; paying for ranking credit is the banned case. Treat unpaid-for-ranking marketplace links as policy-risk.
Scraping / synonymizing vs. “spin it unique so it isn’t duplicate.” Official: thin rewrite of others’ content is still scraping. Aligns with Dooley/Joy that spun unique can be worse than honest copies of your own page — different issue than copying other sites.

Notes

Last updated 2026-08-28 UTC. Pair with people-first content and canonicalization so “spam” is not used as a synonym for “duplicate.” Primary source for a future scaled-content-abuse concept page.