Knowledge base v2.13.0 · Community edition · data through 2 October 2026
Topic · The index and its signals
Google said spam is stopped at indexing: index selection loads most spam signals and keeps documents detected as spam out of the index, and Google's spam policies say violating sites may rank lower or not appear at all. Google has used statistical models against spam for over 20 years and now uses more and more AI; it said SpamBrain is built on Gemini and fine-tuned for spam, which is not in its documentation. Figures said on stage differ from the record: Google's 2021 webspam report gives 2018 as SpamBrain's launch year (2022 was mentioned on stage), and its 2022 report's '5 times more spam sites' compares 2022 with 2021, with 200 times more than at launch, rather than comparing SpamBrain with older algorithms. Google said its testing shows more than 99% of visits from Search are spam-free, a figure in its 2022 webspam report. Google said spam metadata such as white-on-white text is stored with a page's tokens (not in its docs), and its scaled content abuse policy covers automatically translated scraped content. Day 3 put the scale in numbers: Google discovers about 40 billion spammy pages a day, the figure in its webspam report for 2020 and its How Search Works page (a Day 3 slide headed 'In 2023' repeated it, and the quality talk said tens of billions), and its spam systems remove a very large amount before results are served. Google said AI has changed how it builds spam updates, letting it evaluate more candidates, catch new types of spam and launch faster; four spam updates had shipped in 2026 by 2 October, the September 2026 one still rolling out. Google repeated its aim of keeping more than 99% of results spam-free, which its webspam reports measure as visits from Search. Keyword stuffing remains a policy violation, and Google counts AI slop, mass-produced LLM content, as scaled content abuse that it is working hard to remove. Day 1's second recording added that the crawl scheduler very likely deprioritises URLs or sites known to be historically spammy. On Day 2 the non-Latin-script talk cited the October 2023 spam update, whose post says it improved coverage in many languages, naming Turkish, Vietnamese, Indonesian, Hindi and Chinese, particularly against cloaking, hacked, auto-generated and scraped spam. Author’s view: the post names neither Persian nor Arabic and does not mention link spam, so it does not show that the bought links the presenter saw working in those languages were addressed.
Things in this topic 6
Counts are claims that name the thing. All things
What to do
Google's crawl scheduler very likely deprioritises a URL when the URL or its site is known to be historically spammy.
Speaker Gary IllyesIn Day 1, 14:05 · How crawling worksEvidence transcript
Google's guide to A/B testing says never to show one set of URLs to Googlebot and a different set to humans: that is cloaking, which is against Google's spam policies whether or not a test is running.
Publisher Google Search CentralAnnotates Day 1, 16:00 · How Google thinks about crawl budget
Used byrequirement DEV-PRF-01
Google also stores spam metadata with the tokens of a page, for example that text was white on a white background, so ranking can use that information.
Speaker Gary IllyesIn Day 2, 11:30 · Understanding what's on a pageEvidence transcript
Used byrequirement DEV-HTM-06
The presenter of the non-Latin-script talk said Google announced in October 2023 that it had improved its spam-detection coverage for languages: the spam policy was global, but the coverage improvement was language-specific.
Speaker a second community speakerIn Day 2, 14:30 · Lightning session G: InternationalisationEvidence transcript
Spam stops a document at indexing: when Google detects that a document is spam, it does not let it into the index.
“If we detect that something is spam, then we're not letting it into the index.”
Speaker GoogleIn Day 2, 15:30 · Calculating (some) signalsEvidence transcript
Google uses more and more AI to detect spam, and Google's testing shows that this AI-based detection is highly accurate.
Speaker GoogleIn Day 2, 15:30 · Calculating (some) signalsEvidence transcript
SpamBrain, Google's AI-based spam detection system, is nowadays built on Gemini and fine-tuned specifically for finding spam.
“built on top of, well, nowadays, Gemini, and fine-tuned to that specific purpose of finding spam”
Speaker GoogleIn Day 2, 15:30 · Calculating (some) signalsEvidence transcript
Used byglossary term SpamBrain
SpamBrain is central to Google's spam-fighting efforts and has been improved many times since its launch.
Speaker GoogleIn Day 2, 15:30 · Calculating (some) signalsEvidence transcript
SpamBrain was said to detect 5 times more spam sites than the spam algorithms Google had launched before it.
Speaker GoogleIn Day 2, 15:30 · Calculating (some) signalsEvidence transcript
Google's testing shows that, thanks to SpamBrain, more than 99% of visits from Search are now spam-free.
Speaker GoogleIn Day 2, 15:30 · Calculating (some) signalsEvidence transcript
Index selection loads most of Google's spam signals and acts as a gatekeeper so that no spam enters search results.
“index selection acts as a gatekeeper, ensuring that no spam enters our search results”
Speaker GoogleIn Day 2, 15:40 · Deciding what goes in the index?Evidence transcript
Google's spam policies define scaled content abuse as generating many pages mainly to manipulate rankings, with little or no value to users, no matter how they are created, and list automated translating of scraped content among the examples.
Publisher Google Search CentralAnnotates Day 2, 14:15 · Focusing on Internationalisation and Localisation
Used byrequirement DEV-INT-10glossary term Scaled content abuse
Google's October 2023 spam update post (4 October 2023) says the update improved coverage in many languages and spam types, cleaning up spam reported in Turkish, Vietnamese, Indonesian, Hindi, Chinese and other languages, particularly cloaking, hacked, auto-generated and scraped spam.
Publisher Search Central blog (4 October 2023)Annotates Day 2, 14:30 · Lightning session G: Internationalisation
Google's spam policies say that sites violating them may rank lower in results or not appear in results at all.
Publisher Google Search CentralAnnotates Day 2, 15:30 · Calculating (some) signals
Used byrequirement DEV-HTM-06
Google's 2021 webspam report says SpamBrain, its AI-based spam-prevention system, was launched in 2018 and has been continuously improved since.
Publisher Search Central blog (21 April 2022)Annotates Day 2, 15:30 · Calculating (some) signals
Used byglossary term SpamBrain
Google's 2022 webspam report says SpamBrain detected 5 times more spam sites in 2022 than in 2021, and 200 times more than when it first launched.
Publisher Search Central blog (11 April 2023)Annotates Day 2, 15:30 · Calculating (some) signals
Hidden text is recorded at token level: Google stores spam metadata such as white-on-white text with the tokens, so leftover hidden keyword blocks are a liability, not neutral clutter, and should be removed.
Author Ibrahim AnjroAnnotates Day 2, 11:30 · Understanding what's on a page
Used byrequirement DEV-HTM-06
Treat machine translation as a first draft: have a native speaker review it and adapt dates, calendars and units before publishing, because unreviewed bulk translation that adds little value can also fall under Google's scaled content abuse policy.
Author Ibrahim AnjroAnnotates Day 2, 14:15 · Focusing on Internationalisation and Localisation
Used byrequirement DEV-INT-10
The October 2023 spam update cited in the non-Latin-script talk names neither Persian nor Arabic (only 'other languages') and lists cloaking, hacked, auto-generated and scraped spam, not link spam, so it does not show that the bought links the presenter saw working in Persian or Arabic search were addressed.
Author Ibrahim AnjroAnnotates Day 2, 14:30 · Lightning session G: Internationalisation
The SpamBrain launch year mentioned on stage, with hesitation, was 2022, which does not match Google's documented 2018; 2022 is the year of the improvements described in Google's 2022 webspam report, so cite 2018 as the launch year.
Author Ibrahim AnjroAnnotates Day 2, 15:30 · Calculating (some) signals
The '5 times more spam sites' figure said on stage matches Google's 2022 webspam report, but the report compares 2022 with 2021 (and gives 200 times since launch), not SpamBrain with earlier algorithms; quote the documented comparison.
Author Ibrahim AnjroAnnotates Day 2, 15:30 · Calculating (some) signals
The lower selection bar in under-served languages is an opening for content written for that market, not for machine translation at scale: index selection also applies spam signals, and Google's spam policies count generating many pages from scraped content through automated transformations such as translating, with little value for users, as scaled content abuse.
Author Ibrahim AnjroAnnotates Day 2, 15:40 · Deciding what goes in the index?
Used byrequirement DEV-INT-10
Google's slide, headed 'In 2023, there were...', said 40 billion spammy pages are detected every day.
Speaker GoogleIn Day 3, 12:00 · What are quality updatesEvidence slide photo
Google estimated that a spam update affects sites within 1-2 days of the rollout, 1-2 weeks for continuous processing (said on stage as 2 weeks on average) and months for batch refreshes.
Speaker Gary IllyesIn Day 3, 15:45 · How long does it take to..?Evidence slide photo, transcript
A community speaker recalled that around 2009 writers bent titles and phrases to include a keyword because their CMS counted keyword density, and said that hitting the ratio really worked at the time.
Speaker Community speakersIn Day 3, 10:45 · Lightning session K: Facets of qualityEvidence transcript
Google's quality talk named the spam systems among Google's ranking systems and said site owners probably interact a lot with the freshness systems, the reviews system and other ranking systems.
Speaker GoogleIn Day 3, 11:15 · How Google thinks about QualityEvidence slide photo, transcript
Google's quality talk asked what would happen to the web ecosystem if everyone used AI to flood the web with surface-level content like the example it had just shown.
Speaker GoogleIn Day 3, 11:15 · How Google thinks about QualityEvidence transcript
Google's quality talk said Google discovers tens of billions of spam pages every day.
Speaker GoogleIn Day 3, 11:15 · How Google thinks about QualityEvidence transcript
Google's quality talk said a big portion of the new pages Google discovers every day is spam, adding that the speaker did not know of Google ever publishing the percentage.
Speaker GoogleIn Day 3, 11:15 · How Google thinks about QualityEvidence transcript
Google's quality talk said spam systems are one of the ways Google protects search quality, removing a very large amount of spam before results are served.
Speaker GoogleIn Day 3, 11:15 · How Google thinks about QualityEvidence transcript
Google's quality talk said Google had released four spam updates in 2026 by the time of the talk on 2 October, with one still rolling out.
Speaker GoogleIn Day 3, 11:15 · How Google thinks about QualityEvidence transcript
Google's quality talk said its spam updates mostly target spam pages, so most site owners in the audience were probably not affected.
Speaker GoogleIn Day 3, 11:15 · How Google thinks about QualityEvidence transcript
Used byglossary term Spam update
Google's quality talk said evaluating the quality of the web at scale is messy and hard, and that Google's work on it is never done.
Speaker GoogleIn Day 3, 11:15 · How Google thinks about QualityEvidence transcript
Google's quality talk said AI has fundamentally changed how Google builds spam updates, letting it evaluate many more candidates and drastically increasing its velocity, so it launches faster with more impact.
Speaker GoogleIn Day 3, 11:15 · How Google thinks about QualityEvidence transcript
Google's quality talk said AI lets Google catch new types of spam and catch more of it.
Speaker GoogleIn Day 3, 11:15 · How Google thinks about QualityEvidence transcript
Google's quality talk said Google reads spam reports and uses them to decide what kinds of spam to tackle next.
Speaker GoogleIn Day 3, 11:15 · How Google thinks about QualityEvidence transcript
Google aims to keep more than 99% of search results free from spam and said it already achieves this, thanks to advances in AI.
Speaker GoogleIn Day 3, 12:00 · What are quality updatesEvidence transcript
Scaled content abuse is becoming a problem again: in the early 2000s pages were churned out with Perl or PHP scripts, and now the same is done with LLMs.
Speaker Gary IllyesIn Day 3, 16:00 · Wrapping all up: AI, Search, and making sense of everything.Evidence transcript
Used byrequirement DEV-SPM-02
The cheaper tokens become, the more AI slop is created, and Google counts AI slop as scaled content abuse.
Speaker Gary IllyesIn Day 3, 16:00 · Wrapping all up: AI, Search, and making sense of everything.Evidence transcript
Used byrequirement DEV-SPM-02glossary term AI slop
Google said it is going to great lengths to get AI slop out of its search results.
Speaker Gary IllyesIn Day 3, 16:00 · Wrapping all up: AI, Search, and making sense of everything.Evidence transcript
Used byglossary term AI slop
Google's spam policies define keyword stuffing as filling a web page with keywords or numbers to manipulate rankings in Google Search, for example repeating the same words or phrases so often that it sounds unnatural.
“filling a web page with keywords or numbers in an attempt to manipulate rankings in Google Search results”
Publisher Google Search CentralAnnotates Day 3, 10:45 · Lightning session K: Facets of quality
Google's ranking systems guide lists Penguin as a retired system: announced in 2012 to combat link spam, it was integrated into the core ranking systems in 2016.
Publisher Google Search CentralAnnotates Day 3, 10:45 · Lightning session K: Facets of quality
Google's Search Status Dashboard says the September 2026 spam update began on 24 September 2026, applies globally to all languages and may take up to two weeks to roll out.
Publisher Google Search Status Dashboard (Ranking incident history), Google Search Status Dashboard (September 2026 spam update)Annotates Day 3, 11:15 · How Google thinks about Quality
Since April 2026 Google's spam report page says Google may use a report to take manual action, and that the submission text is sent to the site owner, so reports containing personally identifying information are not processed.
Publisher Google Search CentralAnnotates Day 3, 11:15 · How Google thinks about Quality
Google's webspam report for 2020, published in April 2021, says Google discovers 40 billion spammy pages every day.
Publisher Search Central blog (29 April 2021)Annotates Day 3, 12:00 · What are quality updates
Google's spam updates page says its automated spam detection systems run constantly, and a notable improvement to them, such as to the AI-based SpamBrain system, is called a spam update and listed with Google's ranking updates.
Publisher Google Search CentralAnnotates Day 3, 12:00 · What are quality updates
Used byglossary term Spam update
Do not treat keyword-density or word-count scores from CMS plugins as editorial targets: Google's spam policies treat filling a page with keywords to manipulate rankings as keyword stuffing, and Google says it has no preferred word count.
Author Ibrahim AnjroAnnotates Day 3, 10:45 · Lightning session K: Facets of quality
The spoken 'tens of billions of spam pages' a day matches Google's How Search Works page, which says its systems find 40 billion spammy pages every day, and the later Day 3 slide '40B spammy pages detected every day' (under 'In 2023'); cite 40 billion a day.
Author Ibrahim AnjroAnnotates Day 3, 11:15 · How Google thinks about Quality
Cite 40 billion spammy pages a day as Google's published figure, from its webspam report for 2020 and its How Search Works page: the slide's 'In 2023' heading does not make it a 2023 measurement, because the 2021 and 2022 reports give no daily count.
Author Ibrahim AnjroAnnotates Day 3, 12:00 · What are quality updates
Google's webspam reports measure the 99% figure as visits from Search (2020 and 2022 reports) or searches (2021 report) that are spam-free, not as search results; cite it as more than 99% of visits from Search being spam-free.
Author Ibrahim AnjroAnnotates Day 3, 12:00 · What are quality updates
Treat machine translation as a first draft: have a native speaker review it and adapt dates, calendars and units before publishing, because unreviewed bulk translation that adds little value can also fall under Google's scaled content abuse policy.
The answer is not a claim that Google detects AI text. It says ranking favours text that reads as natural to people. The risk with AI content is scale without value, which falls under Google's scaled content abuse policy, not the tool itself.
Google uses more and more AI to detect spam, and Google's testing shows that this AI-based detection is highly accurate.
Statistical models have been used at Google for over 20 years, for catching spam and originally for the 'Did you mean' feature.
SpamBrain, Google's AI-based spam detection system, is nowadays built on Gemini and fine-tuned specifically for finding spam.
Gemini is not part of Search, but it uses crawlers for data, shares some technologies such as tokenization and deduping, and grounds on the Search index.
SpamBrain, Google's AI-based spam detection system, is nowadays built on Gemini and fine-tuned specifically for finding spam.
Statistical models have been used at Google for over 20 years, for catching spam and originally for the 'Did you mean' feature.
Google's quality talk said AI has fundamentally changed how Google builds spam updates, letting it evaluate many more candidates and drastically increasing its velocity, so it launches faster with more impact.
Google uses more and more AI to detect spam, and Google's testing shows that this AI-based detection is highly accurate.
Google's quality talk said AI lets Google catch new types of spam and catch more of it.
Google uses more and more AI to detect spam, and Google's testing shows that this AI-based detection is highly accurate.
Google's slide, headed 'In 2023, there were...', said 40 billion spammy pages are detected every day.
Google's quality talk said Google discovers tens of billions of spam pages every day.
Google's spam updates page says its automated spam detection systems run constantly, and a notable improvement to them, such as to the AI-based SpamBrain system, is called a spam update and listed with Google's ranking updates.
SpamBrain is central to Google's spam-fighting efforts and has been improved many times since its launch.
Google may never fetch a lower-quality site's sitemap again: once it figures out the site is of lower quality, it no longer wants to fetch the sitemap.
Google's crawl scheduler very likely deprioritises a URL when the URL or its site is known to be historically spammy.
Google estimated that a spam update affects sites within 1-2 days of the rollout, 1-2 weeks for continuous processing (said on stage as 2 weeks on average) and months for batch refreshes.
When spammers find loopholes that Google's systems miss, Google releases spam updates that change its systems to catch them.
Scaled content abuse is becoming a problem again: in the early 2000s pages were churned out with Perl or PHP scripts, and now the same is done with LLMs.
Reviewing the spam slide, the speaker said link spam would probably be replaced with scaled content abuse as the spam type worth talking about today.
The cheaper tokens become, the more AI slop is created, and Google counts AI slop as scaled content abuse.
Google's spam policies define scaled content abuse as generating many pages mainly to manipulate rankings, with little or no value to users, no matter how they are created, and list automated translating of scraped content among the examples.
Google's Search Quality Rater Guidelines point out that the tool used to create content is not the problem, but how it was used and what for.
Google's spam policies define scaled content abuse as generating many pages mainly to manipulate rankings, with little or no value to users, no matter how they are created, and list automated translating of scraped content among the examples.
Google aims to keep more than 99% of search results free from spam and said it already achieves this, thanks to advances in AI.
Google's testing shows that, thanks to SpamBrain, more than 99% of visits from Search are now spam-free.
Never hide text from users, such as white-on-white or off-screen keyword blocks
Rests on 3 claims, 3 of them in this topic
Publish machine-translated language versions only after human review and localisation
Rests on 5 claims, 3 of them in this topic
Do not generate large sets of low-value pages from templates, feeds or LLM output
Rests on 7 claims, 2 of them in this topic
Keep connect time and time to first byte low and stable under crawler load
Rests on 13 claims, 1 of them in this topic