Google Core Updates

For search marketers, knowing the dates of these Google updates can help explain changes in rankings and organic website traffic and ultimately improve search engine optimization. Here are the major algorithmic changes that have had the biggest impact on search.

Link analysis and PageRank

Launched: Used when Google first launched; foundational paper published in 1998
Roll outs: PageRank has evolved substantially and remains part of core ranking systems. This is an ongoing system, not a fixed public update schedule.
Goal: Use links between pages as evidence when finding useful search results.
Hazards: Buying or automating links to manipulate rankings
Treating a third-party DA or DR score as Google PageRank
Important pages with no useful internal links
How it works: Example: other cycling pages link to a bicycle-repair guide. Google can analyze those connections to help understand the guide and its usefulness for a repair query. PageRank is one evolved core link-analysis system; a raw backlink count is not the whole ranking algorithm.
How to stay safe: Check that relevant category pages and guides link to the repair page and that their links reach the intended URL. Stop manipulative link campaigns; qualify genuine paid placements so they do not pass ranking credit.
Sources: Google ranking systems guide: link analysis and PageRank Google Research: Brin and Page, The Anatomy of a Large-Scale Hypertextual Web Search Engine (1998)Google Search spam policies: link spam

Bigdaddy infrastructure

Launched: 4 January 2006 (public preview feedback post; exact initial deployment unverified)
Roll outs: The January 4 post said the preview was live at two data centers and estimated a later transition. That estimate is not a verified completion date.
Goal: Provide search infrastructure that supports better canonical URL handling.
Hazards: The same content appearing at inconsistent www and non-www URLs
Conflicting redirect destinations or duplicate URLs
Treating an infrastructure preview as a broad core ranking penalty
How it works: Example: the same product appears at www.example.com/item and example.com/item, with inconsistent redirects. Bigdaddy prepared infrastructure for better canonicalization and related URL handling. Matt Cutts explicitly distinguished it from an algorithm update or data refresh; most changes were under the hood.
How to stay safe: Choose a preferred version of the product URL. Check alternate versions, redirect destinations, and internal links for consistency; investigate the indexed canonical separately from a ranking change.
Sources: Matt Cutts: Feedback on Bigdaddy data center, 4 January 2006

Caffeine

Launched: June 2010
Roll outs: A continuous indexing architecture announced in June 2010, rather than a periodic content-quality filter.
Goal: Add newly found or changed pages to the index without waiting for a large index refresh.
Hazards: A new article has no crawlable link from another page
Robots rules or server errors prevent Googlebot from fetching an updated page
How it works: Example: a newly published match report can enter the index sooner instead of waiting for a large index refresh. Caffeine introduced continuous indexing to add fresh material faster. It was an infrastructure change, not a content penalty or a broad core update.
How to stay safe: After publishing a new page or a substantial update, inspect that exact URL in Search Console. Check whether Google has discovered, crawled, and indexed it; a new page is not guaranteed to be indexed or rank.
Sources: Google: new Caffeine index (2010)

Google Panda

Launch date: February 24, 2011
Roll outs: Historically monthly; became part of Google's core ranking systems in 2015.
Goal: De-rank sites with low-quality content
Hazards: Duplicate content, Plagiarism, Thin content, User-generated spam, Keyword stuffing, Poor user experience
How it works: Google Panda was designed to down-rank sites with low-quality, spammy, or thin content. Initially, Panda was a filter rather than a part of Google's core algorithm, but in 2015, it was officially incorporated into the ranking algorithm. Becoming part of core does not establish a real-time schedule or guarantee faster filtering or recovery.
How to stay safe: Check for duplicate content across your site
Check for plagiarism externally
Identify thin content in your site
Audit your site for keyword stuffing
Sources: Google: finding more high-quality sites in Search (2011)Google: guidance on high-quality sites (2011)Google guide to ranking systems

Freshness ranking improvement

Launched: 3 November 2011 (ranking improvement announced)
Roll outs: A significant ranking improvement announced on 3 November 2011. Google currently describes multiple freshness systems, rather than a single named recurring rollout.
Goal: Show current information when the query needs a current answer.
Hazards: Last season’s scores or event details on a query seeking the latest result
Old model information in a current buying comparison
Changing dates without checking whether the facts need updating
How it works: Example: a week-old football score can be wrong for a current-score query, while an older tomato-sauce recipe may still answer a recipe query. The November 2011 improvement changed how ranking assessed these different freshness needs. It built on Caffeine’s faster indexing but was a separate ranking change.
How to stay safe: For scores, recurring events, or product comparisons, verify the current result, event year, models, and specifications. Update changed facts and identify the period covered; do not rewrite a still-correct evergreen answer merely to make its date newer.
Sources: Official Google Blog: Giving you fresher, more recent search results Google ranking systems guide: freshness systems

Google Penguin

Launched: April 24, 2012
Roll outs: Reported historical dates: May 25, 2012 | Oct 5, 2012 | May 22, 2013 | Oct 4, 2013 | Oct 17, 2014 | September 27, 2016 | October 6, 2016. Google announced real-time core integration on September 23, 2016.
Goal: De-rank sites with spammy, manipulative link profiles
Hazards: Links coming from poor quality, "spammy" sites
Links coming from sites created purely for SEO link building (PBNs)
Links coming from topically irrelevant sites
Links with overly optimized anchor text
Paid links
How it works: Google Penguin originally aimed to identify and down-rank sites with unnatural link profiles, deemed to be spamming the search results by using manipulative link tactics. Since September 2016, Penguin is part of Google's core ranking algorithm and operates in real time after recrawling and reindexing. The change made it more granular: spam links can lose ranking credit rather than causing a blanket penalty for the entire site. Recovery is not guaranteed to be faster.
How to stay safe: Monitor link profile growth
Get rid of harmful links
Check for penalty risks
Sources: Google: another step to reward high-quality sites (2012)Google: Penguin becomes part of core (2016)Google Search Console: disavow guidance

Google Pirate

Launched: Aug 2012
Roll outs: Oct 2014
Goal: De-rank sites with copyright infringement reports
Hazards: Pirated content & High volume of copyright infringement reports
How it works: Google's Pirate Update was designed to prevent sites that have received numerous valid copyright infringement reports, as filed through Google’s DMCA system, from ranking well in Google search. Early descriptions of affected sites focused on relatively big and well-known websites that made pirated content (such as movies, music, or books) available to visitors for free, particularly torrent sites. New sites containing pirated material can still appear; a demotion system does not establish that every newly created infringing page has already been identified. Historical descriptions of the filter discussed periodic updates: previously affected sites could escape after improvements, newly identified sites could be caught, and false positives could be released. This is not a published current refresh schedule or a guaranteed route to recovery.
How to stay safe: Don't distribute anyone's content in your websites without the copyright owner's permission.
Sources: Google: copyright notices as a ranking signal (2012)

Exact match domain system (EMD)

Launched: 28 September 2012 (upcoming change announced; exact deployment date unverified)
Roll outs: The September 2012 announcement described an upcoming change. Google’s current documentation describes an ongoing exact match domain system; no fixed refresh cadence is established.
Goal: Limit excessive ranking credit from a domain name that matches a search query.
Hazards: Thin pages hosted on a keyword-rich domain
Buying a domain because its words supposedly guarantee rankings
Assuming every exact-match domain violates a policy
How it works: Example: best-plumber-in-london.example is a hypothetical domain, not proof that its pages are the best plumbing result. Google considers domain words but adjusts excessive credit for exact matches. The September 2012 announcement concerned low-quality exact-match domains, not a ban on all keyword domains.
How to stay safe: Inspect what the pages actually provide: for the plumbing example, clear service details, genuine business information, and answers to customer questions. Fix missing substance; keep or change the domain for a business reason rather than expecting the name alone to determine rankings.
Sources: Matt Cutts: announcement of upcoming low-quality EMD change Google ranking systems guide: exact match domain system

Spam-query update (often called Payday Loan)

Launched: 11 June 2013 (spam-query update started)
Roll outs: A new ranking update for spam-heavy queries began on 11 June 2013. Its original launch announcement gives no completion date.
Goal: Reduce spam in heavily abused query categories, including payday-loan searches.
Hazards: Unauthorized loan pages injected into a website
Cloaking or link schemes used to promote spam pages
Assuming legitimate lending content is penalized simply for its subject
How it works: Example: a compromised website gains unauthorized loan pages supported by spam links. Those are spam practices to investigate, not evidence that every lending page is prohibited. Matt Cutts announced a spam-query ranking update on 11 June 2013; Google’s later corporate explanation said the same spam criteria applied to lending sites and other websites.
How to stay safe: Check unexpected loan URLs or queries on your domain, inspect whether pages were added without permission, and fix any security breach. Stop link schemes you control. Assess legitimate lending pages against the same spam rules as other pages, separately from Ads eligibility.
Sources: Matt Cutts: original spam-query launch announcement Google Inc.: response to UK CMA provisional findings, 4 July 2014, page 3 and footnote 1 Google Search spam policies

Google Hummingbird

Launched: August 2013
Roll outs: —
Goal: Produce more relevant search results by better understanding the meaning behind queries
Hazards: Keyword stuffing, low-quality content, Exact-match keyword targeting
How it works: Hummingbird helps Google better interpret search queries and provide results that match searcher intent (as opposed to the individual terms within the query). While keywords continue to be important, Hummingbird makes it possible for a page to rank for a query even if it doesn’t contain the exact words the searcher entered. This concerns the meaning of the full query rather than exact-word matching. As an illustration, instructions about replacing a punctured tire can answer a question about changing a flat tire; this is not a documented LSI keyword formula.
How to stay safe: Expand your keyword research
Discover the language your audience uses
Ditch exact-match, think concepts
Sources: Google guide to ranking systems

Google Pigeon

Launched: July 24, 2014 (US; industry-reported date)
Roll outs: December 22, 2014 (UK, Canada, Australia; industry-reported dates)
Goal: Provide high quality & relevant local search results
Hazards: Poorly optimized pages
Improper setup of a Google My Business page
NAP inconsistency
Lack of citations in local directories (if relevant)
How it works: Pigeon, an industry name for a 2014 local-search change, dramatically altered the results Google returns for queries in which the searcher's location plays a part. Industry reporting described Pigeon as creating closer ties between the local algorithm and the core algorithm, while current local ranking guidance identifies relevance, distance, and prominence. This update also uses location and distance as a key factor in ranking the results.
How to stay safe: Optimize your pages properly
Set up a Google My Business page
Make sure your NAP is consistent across your local listings
Get featured in relevant local directories
Sources: Google Business Profile: local ranking factors

HTTPS ranking signal

Launched: August 2014
Roll outs: A lightweight ranking signal introduced in August 2014, rather than a named broad core rollout.
Goal: Give secure HTTPS pages a small ranking signal while keeping the page's answer relevant to the search.
Hazards: HTTP URLs still return separate copies of HTTPS pages
Redirect loops or canonical tags point to the old HTTP version
Images or scripts still load over HTTP
How it works: HTTPS gives a page an encrypted connection. Google introduced it as a lightweight ranking signal, initially affecting fewer than one percent of global queries and carrying less weight than content quality. Moving a weak page from HTTP to HTTPS does not give it a top position.
How to stay safe: Open representative HTTP URLs and confirm that they redirect to the intended HTTPS pages. Check that canonical tags, images, and scripts use the intended secure URLs; repair incorrect destinations or mixed URL references.
Sources: Google: HTTPS ranking signal (2014)

Google Rank Brain

Launched: 2015
Roll outs: —
Goal: Deliver better search results based on relevance & machine learning
Hazards: Lack of query-specific relevance features, shallow content, poor User eXperience
How it works: Rank Brain is one of Google’s AI ranking systems. It is a machine learning system that helps Google understand the meaning behind queries, and serve best-matching search results in response to those queries. At the time of its public introduction, Rank Brain was reported as Google’s third most important ranking factor. While we don’t know the ins and outs of Rank Brain, Google describes it as understanding how words relate to concepts. This helps a page match a query even when it does not contain every exact word the searcher used.
How to stay safe: Maximize user experience, keep an eye on your pages user experience factors in Google Analytics, particularly Bounce Rate and Session Duration. Use those metrics to identify audience problems; Google does not document them as RankBrain ranking inputs.
Sources: Google: how AI powers search results (2022)

Google Mobile Friendly

Launched: April 21, 2015
Roll outs: —
Goal: Give mobile friendly pages a ranking boost in mobile SERPs, and de-rank pages that aren't optimized for mobile
Hazards: Lack of a mobile version of the page
Improper viewport configuration
Illegible content
How it works: Google’s Mobile Update (aka Mobilegeddon) introduced mobile-friendliness as a ranking signal for individual pages in mobile search. Pages not optimized for mobile could rank lower; mobile-friendliness did not guarantee top positions or automatically remove other pages from the SERPs. The announcement covered mobile searches in all languages worldwide and described a significant impact in search results. The goal was to make it easier for users to get relevant, high quality search results optimized for their devices.
How to stay safe: Go mobile and focus on speed and usability. There are a few mobile website configurations to choose from, but Google's recommendation is responsive design. Check the viewport, text readability, and tap targets on a phone. Passing usability checks does not guarantee higher positions in mobile SERPs.
Sources: Google: mobile-friendly rollout (2015)

Google Possum

Launched: September 1, 2016 (industry-reported date)
Roll outs: —
Goal: Deliver better, more diverse results based on the searcher's location and the business address
Hazards: Sharing a physical address with a similar business, Competitors whose business address is closer to the searcher's location
How it works: The changes reported under the industry name Possum were associated with local results varying more depending on the searcher’s location, the closer you are to a business’s address, the more likely you are to see it among local results. Possum also resulted in greater variety among results ranking for very similar queries (even close variations now produce different results). Somewhat paradoxically, industry reports also associated Possum with some businesses outside the physical city area appearing more often in local results. This was not a guarantee for every business. Additionally, industry reports described filtering among similar businesses sharing an address. A shared address alone does not establish an automatic ranking penalty.
How to stay safe: Expand your keyword list and do location-specific rank tracking. Use the query and location comparisons to decide whether additional keyword targets are relevant; the reported volatility does not establish a fixed increase every business needs.
Sources: Google Business Profile: local ranking factors

Mobile-first indexing

Launched: 4 November 2016 (experiments announced); completion announced 31 October 2023
Roll outs: Experiments were announced on 4 November 2016; completion was announced on 31 October 2023, with a small set of sites that did not work on mobile noted as exceptions.
Goal: Index the content Googlebot finds on the mobile version of a page.
Hazards: Product specifications, article text, or internal links appear on desktop but are absent on mobile
The mobile page has different robots directives or omits structured data
Images or other mobile resources are blocked from crawling
How it works: Example: a desktop article contains a full buying guide, but its mobile version shows only a short teaser. Mobile-first indexing makes that missing mobile content matter because Google primarily uses the mobile version for indexing. Experiments began in 2016; on 31 October 2023, Google announced completion, while noting a small set of sites that did not work on mobile.
How to stay safe: Compare the same URL on mobile and desktop: main text, images, structured data, metadata, and crawl access. Restore important content missing from mobile and remove mobile-specific crawl restrictions.
Sources: Google: mobile-first indexing experiments (2016)Google: mobile-first indexing completion (2023)

Intrusive interstitials signal

Launched: 10 January 2017
Roll outs: Announced on 23 August 2016; the launch note confirmed the signal started on 10 January 2017.
Goal: Let someone arriving from mobile search read the answer without first dismissing a large overlay.
Hazards: A full-screen newsletter or advertising popup covers the article on arrival
An interstitial page must be dismissed before the answer loads
Necessary legal notices are enlarged into an unnecessary obstacle
How it works: Example: a reader taps a mobile search result and must close a full-screen advert before reaching the article. Pages that obstructed this transition could rank less well. Google distinguished these overlays from responsible legal notices, private-content login dialogs, and small dismissible banners.
How to stay safe: Check the page as a reader arriving from mobile search. Remove or resize overlays that hide the main answer; keep necessary legal notices and private-content access controls proportionate.
Sources: Google: accessible mobile content and interstitials

Google Fred

Launched: March 8, 2017 (industry-reported date)
Roll outs: —
Goal: Filter out low quality search results whose sole purpose is generating ad and affiliate revenue
Hazards: Low-value, affiliate-heavy & ad-centered content
How it works: Fred is an industry nickname associated with reported quality changes in 2017. Industry reports about affected sites described content sites (mostly blogs) with low-quality articles on a wide variety of topics that appear to be created mostly for the purpose of generating ad or affiliate revenue. Advertising or affiliate links alone do not establish a violation.
How to stay safe: Review Google Search Quality Guidelines and watch out for thin content. If you show ads, make sure the pages they are found on are high-quality and offer relevant, ample information.
Sources: Google: guidance on high-quality sites (2011)

Neural matching

Launched: 2018
Roll outs: Introduced in 2018 as a retrieval system that works alongside other Search systems, rather than a scheduled standalone update.
Goal: Find a page about the concept a person searched for, even when the query and page use different words.
Hazards: A page repeats the query but never explains the subject
A product or service page leaves its purpose and use case unclear
How it works: Example from Google: a search for “insights how to manage a green” can lead to advice about managing a green personality type. Neural matching connects concepts across the full query and page, helping retrieve relevant documents beyond literal keyword overlap. It works alongside other Search systems.
How to stay safe: Check whether the page names and explains the actual subject behind the query. For the personality example, identify the green personality type and give concrete management advice; repeating the unclear phrase alone does not answer it.
Sources: Google: how AI powers search results (2022)

SpamBrain

Launched: 2018 (launch year confirmed by Google)
Roll outs: Continuously improved since launch. The December 2022 link-spam update was a later use of SpamBrain, not its original launch.
Goal: Keep spam and malicious behavior out of Search results.
Hazards: Hacked or automatically generated spam
Manipulative practices that violate Google's spam policies
How it works: SpamBrain uses AI to recognize spam patterns, including hacked sites and gibberish pages. For example, a business site injected with gambling pages can become a spam source even though its owner did not publish them.
How to stay safe: Inspect unexpected indexed URLs and pages you did not publish. Remove injected pages and fix the security weakness that allowed them to appear; review deliberate spam practices separately.
Sources: Google: SpamBrain launched in 2018 Google: December 2022 link-spam update

Broad core updates

Launched: Varies by rollout; calendar below covers 2018–2026
Roll outs: Google releases broad core updates several times a year. Use the calendar below for individual rollout dates and their precision.
March 2018 broad core update
March 2018 · exact start/end unverified
Google confirmed that a broad core update had been released the previous week.
Google SearchLiaison announcement
April 2018 broad core update
16 April 2018 · end date unverified
Google confirmed the launch afterward and referred back to its March broad-core guidance.
Google SearchLiaison announcement
August 2018 broad core update
August 2018 · exact start/end unverified
Google confirmed another broad core update and retained its earlier broad-core guidance.
Google SearchLiaison announcement
March 2019 core update
March 2019 · exact start/end unverified
Google adopted the explicit month-and-year name March 2019 Core Update.
Google SearchLiaison naming confirmation Google SearchLiaison rollout announcement
June 2019 core update
3 June 2019 · end date unverified
Pre-announced the previous day, then explicitly confirmed live on 3 June.
Google SearchLiaison live confirmation Google SearchLiaison pre-announcement
September 2019 core update
24 September 2019 · end date unverified
A named broad core update announced for release later the same day.
Google SearchLiaison announcement
January 2020 core update
13–16 January 2020 (US/Pacific)
A confirmed broad core ranking update with rollout dates recorded by Google.
Google Search Status Dashboard
May 2020 core update
4–18 May 2020 (US/Pacific)
A confirmed broad core ranking update with rollout dates recorded by Google.
Google Search Status Dashboard
December 2020 core update
3–16 December 2020 (US/Pacific)
A confirmed broad core ranking update with rollout dates recorded by Google.
Google Search Status Dashboard
June 2021 core update
2–12 June 2021 (US/Pacific)
A confirmed broad core ranking update with rollout dates recorded by Google.
Google Search Status Dashboard
July 2021 core update
1–12 July 2021 (US/Pacific)
A confirmed broad core ranking update with rollout dates recorded by Google.
Google Search Status Dashboard
November 2021 core update
17–30 November 2021 (US/Pacific)
A confirmed broad core ranking update with rollout dates recorded by Google.
Google Search Status Dashboard
May 2022 core update
25 May–9 June 2022 (US/Pacific)
A confirmed broad core ranking update with rollout dates recorded by Google.
Google Search Status Dashboard
September 2022 core update
12–26 September 2022 (US/Pacific)
A confirmed broad core ranking update with rollout dates recorded by Google.
Google Search Status Dashboard
March 2023 core update
15–28 March 2023 (US/Pacific)
A confirmed broad core ranking update with rollout dates recorded by Google.
Google Search Status Dashboard
August 2023 core update
22 August–7 September 2023 (US/Pacific)
A confirmed broad core ranking update with rollout dates recorded by Google.
Google Search Status Dashboard
October 2023 core update
5–19 October 2023 (US/Pacific)
A confirmed broad core ranking update with rollout dates recorded by Google.
Google Search Status Dashboard
November 2023 core update
2–28 November 2023 (US/Pacific)
A confirmed broad core ranking update with rollout dates recorded by Google.
Google Search Status Dashboard
March 2024 core update
5 March–19 April 2024 (US/Pacific)
Changed multiple core ranking systems and how helpful content is assessed. The rollout finished on 19 April; Google announced completion on 26 April.
Google Search Status Dashboard — March 2024 core update Google Search guidance March 2024 core and spam announcement
August 2024 core update
15 August–3 September 2024 (US/Pacific)
Aimed to surface useful, original content from a range of sites, including small or independent publishers, and better recognize site improvements.
Google Search Status Dashboard — August 2024 core update Google Search guidance August 2024 core announcement
November 2024 core update
11 November–5 December 2024 (US/Pacific)
A broad refresh of Google's core ranking systems, with the same people-first content guidance as other core updates.
Google Search Status Dashboard — November 2024 core update Google Search guidance
December 2024 core update
12–18 December 2024 (US/Pacific)
A second core rollout shortly after November's update, with improvements to Google's core ranking systems.
Google Search Status Dashboard — December 2024 core update Google Search guidance
March 2025 core update
13–27 March 2025 (US/Pacific)
A broad core ranking update focused on improving the relevance and quality of search results.
Google Search Status Dashboard — March 2025 core update Google Search guidance
June 2025 core update
30 June–17 July 2025 (US/Pacific)
A broad core ranking update that rolled out from late June into mid-July.
Google Search Status Dashboard — June 2025 core update Google Search guidance
December 2025 core update
11–29 December 2025 (US/Pacific)
A broad core ranking update that completed on 29 December.
Google Search Status Dashboard — December 2025 core update Google Search guidance
March 2026 core update
27 March–8 April 2026 (US/Pacific)
A broad core ranking update, separate from the spam update released three days earlier.
Google Search Status Dashboard — March 2026 core update Google Search guidance
May 2026 core update
21 May–2 June 2026 (US/Pacific)
The latest announced general Search core update as of 1 November 2026; its rollout finished on 2 June.
Google Search Status Dashboard — May 2026 core update Google Search guidance
Goal: Reassess which pages answer searches best as websites, available information, and people's needs change.
Hazards: A comparison page omits the prices, differences, or limitations readers need to choose
Instructions skip steps that competing pages explain
A traffic drop is called a penalty before anyone checks the affected queries or technical changes
How it works: Google changes several core ranking systems rather than publishing a separate rule for every affected website. A page can move down while other pages become better answers, without a spam violation. For example, when reviewing a falling comparison page, check whether it only repeats product descriptions while competing pages explain the differences and show evidence. That comparison is an investigation step, not proof of a particular undisclosed ranking factor.
How to stay safe: Wait at least one full week after completion. In Search Console's Search results report, compare a stable week after the rollout with a week before it began; inspect the pages and queries that lost clicks or positions.
Open the affected pages and the competing results. Add missing steps, comparisons, sources, or first-hand evidence where those answer the reader's question. Check recent redirects, indexing changes, and outages separately; do not promise recovery from an edit.
Sources: Google Search Central: core updates guidance Google Search Status Dashboard — May 2026 core update

Mobile Speed Update

Launched: 9 July 2018 (rollout); announced 17 January 2018
Roll outs: Announced on 17 January 2018; the rollout note was added on 9 July 2018.
Goal: Use speed as a mobile ranking consideration for pages that deliver the slowest experience.
Hazards: Visitors wait for large images or scripts before they can read the main content
A team treats a high lab score as proof the page must rank above a more relevant answer
How it works: Example: a mobile page that takes a long time to display its answer can be one of the slow experiences this signal addressed. The Speed Update targeted the slowest pages and a small share of queries. Strongly relevant content could still rank despite poor speed.
How to stay safe: Run PageSpeed Insights or Lighthouse on the exact mobile URL to identify what delays loading. Fix the largest reported delay and retest that page; one lab score is not a ranking guarantee.
Sources: Google: page speed in mobile search (2018)

Site diversity system

Launched: 6 June 2019 (announced; exact rollout start not established here)
Roll outs: Google confirmed this change separately from the June 2019 core update. It remains a named system in the current ranking guide.
Goal: Give searchers a wider range of sites in prominent results.
Hazards: Relying on several pages from one site to occupy the same top results
Treating fewer displayed listings as proof of a quality penalty
How it works: Google generally shows no more than two listings from one site in top results, with exceptions. For example, if several pages from one shop match a query, some may be omitted so other sites can appear. Subdomains usually count with their root site.
How to stay safe: Record which pages appear for the target query and related specific queries. Check indexing separately before concluding that a missing third listing was penalized; a subdomain does not automatically provide another set of slots.
Sources: Google SearchLiaison: site-diversity announcement Google SearchLiaison: two-listing rule and exceptions Google SearchLiaison: separate from June core update Google ranking guide: site diversity

Original reporting improvements

Launched: 12 September 2019 (announcement; changes had already been made)
Roll outs: Google described recent global changes and revised rater guidance. It did not identify a single rollout start or completion date in that announcement.
Goal: Make significant original reporting easier to find and keep it prominent longer.
Hazards: Rewriting other reporting without new information
Confusing the newest summary with the original reporting
How it works: If one publisher interviews witnesses or uncovers new records and others summarize that story, Google aims to make the original reporting prominent for longer. Rater feedback evaluates the systems; a rater does not manually assign that article its ranking.
How to stay safe: Compare your story with the material it cites. Add your own verified interview, documents, data, or analysis where it serves the story, and clearly credit reporting you used.
Sources: Google: elevating original reporting (2019)

BERT in Search

Launched: 25 October 2019 (announcement)
Roll outs: Announced on 25 October 2019. Initial ranking coverage was English US, while featured snippets had wider multilingual coverage; further expansion was planned.
Goal: Interpret the relationship between words, including words such as 'to' and 'from' that can reverse a query's meaning.
Hazards: Copy blurs who an instruction applies to or the direction of a process
A heading uses a keyword string while the paragraph answers a different question
How it works: Example from Google: a query about a Brazilian traveler going to the US needs different visa information from a US citizen going to Brazil. BERT uses the context around words such as “to” to preserve that direction. The 2019 initial ranking announcement covered about one in ten English US searches, with further expansion planned.
How to stay safe: Read the complete query, including words such as “to”, “for”, and “not”. Check that the answer addresses the correct direction, person, or condition; a visa page must distinguish the traveler’s nationality from the destination. There is no special BERT keyword density to target.
Sources: Google: BERT and query understanding (2019)

Featured snippet deduplication

Launched: 22 January 2020 (public confirmation)
Roll outs: Google's spokesperson confirmed the change on 22 January; SearchLiaison explained it on 23 January. Deduplication is now an ongoing system behavior.
Goal: Remove repeated appearances of the same page and make results easier to use.
Hazards: Counting a featured snippet and its repeated listing as two independent ranking wins
Assuming a removed duplicate listing means the page was deindexed
How it works: If a page supplies the featured snippet for a query, that same URL is not repeated as another ordinary listing on the first results page. Losing the second appearance does not mean the page left Google's index.
How to stay safe: Record the featured-snippet URL and ordinary results for the query. Compare that page's Search Console clicks and impressions before concluding that the changed layout harmed its visibility.
Sources: Danny Sullivan: deduplication confirmation (22 January)Google SearchLiaison: deduplication explanation (23 January)

Passage ranking

Launched: 10 February 2021 (initial US-English rollout, US/Pacific)
Roll outs: Announced on 15 October 2020; the initial US-English launch was confirmed on 11 February 2021 as having happened the previous afternoon. Wider rollout was planned.
Goal: Find useful answers inside pages that cover more than one subject.
Hazards: Burying the answer in unclear text
Assuming every passage becomes a separately indexed page
How it works: A long appliance guide may contain a short section explaining why a washing-machine filter leaks. Google can recognize that section as relevant to the specific question and return the guide's page; the section does not become a separately indexed URL.
How to stay safe: Find the passage that answers the query. Give it a descriptive heading and a direct explanation, keeping the surrounding instructions needed to understand it; do not create a separate page solely to obtain passage ranking.
Sources: Google: passage-understanding announcement (2020)Google SearchLiaison: initial passage-ranking launch

Product reviews update

Launched: 8 April 2021
Roll outs: Launched in April 2021; expanded beyond products in April 2023. In November 2023, Google announced regular, ongoing improvements without periodic rollout notifications.
Goal: Distinguish a review with research, comparisons, and reasons for its verdict from a rewritten product description.
Hazards: A 'review' copies manufacturer specifications and adds an affiliate link
A 'best' list gives no reason to choose one option over another
Performance or testing claims have no supporting method or evidence
How it works: Example: a laptop review that rewrites manufacturer specifications gives readers little original evidence for choosing a machine. The system rewards original research and insightful analysis. It launched in 2021, initially in English; April 2023 expanded coverage beyond products to services and other subjects. In November 2023, Google announced regular, ongoing improvements and the end of periodic rollout notifications.
How to stay safe: Show first-hand tests or observations, such as measured laptop battery life, and explain the test conditions. Compare specific benefits and drawbacks, identify who each option suits, and separate supplied specifications from your own findings.
Sources: Google: product reviews update (2021)Google Search Status: April 2023 reviews expansion Google: November 2023 update Q&A and ongoing reviews improvements

MUM

Launched: May 2021 (announcement)
Roll outs: Announced in May 2021. Google's February 2022 explanation described specialized uses and distinguished them from general ranking systems at that time.
Goal: Support particular Search tasks, including recognizing vaccine-name variants and improving featured-snippet callouts in Google's announced uses.
Hazards: A team rewrites pages to chase a supposed 'MUM score' that Google has not documented
A specialized AI feature announcement is mistaken for a broad core rollout
How it works: Google used MUM for specialized applications such as improving searches for COVID-19 vaccine information. It announced MUM in 2021. Its February 2022 explanation explicitly distinguished those uses from the general ranking work of RankBrain, neural matching, and BERT at that time.
How to stay safe: When investigating a traffic change, first identify the affected surface: web-search results or a specialized Search feature. Verify the documented change for that surface before replacing headings or adding keyword variants because of the MUM announcement.
Sources: Google: how AI powers search results (2022)

Page experience rollout

Launched: June–August 2021 (announced rollout window)
Roll outs: The mobile rollout was scheduled for June–August 2021; this range is the announced plan rather than an exact completion record.
Goal: Account for page experience, including loading, responsiveness, and visual stability, alongside the content's relevance.
Hazards: An advertisement loads above a button and moves it while someone tries to tap
The page displays content but freezes when someone interacts
A passing performance score is used to excuse an incomplete answer
How it works: Example: an article jumps down as an image loads, causing a reader to tap the wrong link. The mobile rollout combined aspects of page experience alongside other ranking considerations. Its 2021 announcement projected a June–August rollout; this is the announced plan. Relevance remained crucial, and good scores did not guarantee top positions.
How to stay safe: Run PageSpeed Insights on the affected URL and inspect concrete failures such as content shifting while images load. Fix the flagged cause, then retest loading and tapping on mobile. Check current Core Web Vitals; a score alone does not prove ranking recovery.
Sources: Google: page experience rollout plans (2021)

Helpful content update

Launched: 25 August 2022 (rollout); announced 18 August 2022
Roll outs: Launched on 25 August 2022. The September 2023 classifier update ran on 14–28 September (US/Pacific); helpfulness assessment became part of core systems in March 2024.
Goal: Reduce results created mainly to capture search traffic when they leave the reader needing to search again for an answer.
Hazards: Dozens of articles summarize other sites without adding reporting, experience, or a usable answer
A page promises instructions or a comparison but never supplies the steps or differences
A site publishes unrelated trending subjects solely to attract search visits
How it works: Example: a gardening site publishes tax-refund and smartphone-deal pages solely to chase search traffic, without knowledge that helps those readers. The people-first system assessed this kind of search-first content. Its September 2023 classifier update ran on 14–28 September (US/Pacific). Since March 2024, helpfulness assessment is part of core ranking systems.
How to stay safe: For pages created to chase traffic, check who can verify their claims and what the intended reader can actually do after reading. Replace empty summaries with verified steps, first-hand examples, or evidence for that audience; remove pages that offer no useful answer.
Sources: Google: helpful content update announcement (2022)Google Search Status: August 2022 helpful content rollout Google guide to ranking systems: helpful content Google Search Status: September 2023 helpful content rollout

December 2022 link spam update

Launched: 14 December 2022
Roll outs: An all-language rollout began on 14 December 2022, using SpamBrain to neutralize unnatural links.
Goal: Stop paid links and networks created to pass ranking credit from supplying that credit.
Hazards: Buying links to manipulate rankings
Sites created to pass outgoing ranking credit
Depending on benefits from links Google has neutralized
How it works: Example: a site buys links from a network built to pass ranking credit. SpamBrain can detect sites buying links and sites created to supply outgoing links, then neutralize the unnatural links’ effect. The December 2022 update covered all languages; ranking credit from those links could be lost.
How to stay safe: Check invoices, paid article placements, and link-building campaigns for links created to manipulate rankings. Stop those campaigns and assess any manual action separately. Do not forecast recovery from links whose former ranking credit has been neutralized.
Sources: Google: December 2022 link spam update

March 2024 spam policies (Policy)

Launched: Announced 5 March 2024; the site reputation abuse policy took effect on 5 May 2024.
Roll outs: Three policy changes, separate from the March core and spam rollouts. Site reputation rules changed again in November 2024 and August 2026.
Goal: Stop expired-domain exploitation, mass-produced search bait, and third-party pages exploiting a host’s established ranking signals.
Hazards: An expired school domain is rebuilt with low-value casino pages to exploit its old signals.
Thousands of copied or lightly rewritten pages target searches without adding useful information.
A publisher leases a coupon section mainly so third-party pages can exploit its ranking signals.
How it works: The policies target abuse, not the tool or business model alone. Reusing an old domain for a useful site, AI-assisted writing, affiliate links, and third-party content are not automatically violations. Read historical site reputation wording alongside the later changes.
How to stay safe: Sample repurposed-domain pages and large content batches; remove or rebuild pages made mainly to manipulate rankings.
Check whether third-party sections serve existing readers rather than merely borrowing host signals. Apply current regional rules.
Sources: Google: March 2024 update and new spam policies Google: current spam policies

Recent spam updates

Launched: Varies by rollout; recent calendar below covers 2024–2026
Roll outs: Automated spam detection runs continuously. Google announces notable spam updates; the calendar below records those rollouts separately from broad core updates.
March 2024 spam update
5–20 March 2024 (US/Pacific)
Rolled out alongside the March core update as Google announced new spam policies. The site reputation abuse policy had a later effective date of 5 May.
Google Search Status Dashboard — March 2024 spam update Google Search guidance March 2024 core and spam announcement
June 2024 spam update
20–27 June 2024 (US/Pacific)
A separate update to Google's spam-detection systems.
Google Search Status Dashboard — June 2024 spam update Google Search guidance
December 2024 spam update
19–26 December 2024 (US/Pacific)
A global spam update covering all languages, beginning the day after the December core rollout finished.
Google Search Status Dashboard — December 2024 spam update Google Search guidance
August 2025 spam update
26 August–22 September 2025 (US/Pacific)
A global spam update covering all languages, with a rollout extending into September.
Google Search Status Dashboard — August 2025 spam update Google Search guidance
March 2026 spam update
24–25 March 2026 (US/Pacific)
A global, all-language spam update that finished before the March core rollout began.
Google Search Status Dashboard — March 2026 spam update Google Search guidance
June 2026 spam update
24–26 June 2026 (US/Pacific)
A global spam update covering all languages, separate from May's core update.
Google Search Status Dashboard — June 2026 spam update Google Search guidance
August 2026 spam update
18–21 August 2026 (US/Pacific)
A global spam update covering all languages, completed in August.
Google Search Status Dashboard — August 2026 spam update Google Search guidance
September 2026 spam update
24 September–8 October 2026 (US/Pacific)
A global spam update covering all languages, completed on 8 October.
Google Search Status Dashboard — September 2026 spam update Google Search guidance
Goal: Detect practices such as cloaking, link schemes, hacked pages, and mass-produced content made to manipulate rankings.
Hazards: To manipulate rankings, Googlebot is shown a useful guide while visitors are deceptively shown unrelated advertising content
Paid links or link networks are used to pass ranking credit
Hacked pages or hidden links are injected into the site
How it works: Spam updates improve Google's automated spam-detection systems, including SpamBrain. Violating sites may rank lower or disappear from results. Link-spam updates can remove ranking benefits from unnatural links, and those lost benefits cannot be regained from the same neutralized links.
How to stay safe: Check Search Console's Manual actions and Security issues reports. Remove injected pages or links, stop cloaking, and remove the specific spam practice you identify; a third-party 'toxic' score alone does not establish a violation.
If Search Console shows a manual action, fix the stated issue and follow its reconsideration process. Automated spam changes have no equivalent notice or guaranteed reassessment deadline.
Sources: Google Search Central: spam updates guidance Google Search Status Dashboard — September 2026 spam update

Explicit fake content protections (Targeted ranking)

Launched: Announced 31 July 2024; a complete rollout window was not published.
Roll outs: Targeted ranking and protection changes, separate from named broad core and spam rollouts.
Goal: Reduce searches for a person’s name exposing non-consensual fake sexual imagery.
Hazards: Hosting fake explicit images of real people without consent.
Repeatedly reposting imagery removed under Google’s explicit-fake-content policies.
How it works: Google described lowering explicit fake content and favoring reliable, non-explicit information when available. Many removals for fake explicit imagery can also reduce other content’s rankings on the hosting site. Reporting about deepfakes is different from hosting the abusive imagery.
How to stay safe: Remove abusive imagery and duplicates, handle removal requests, and prevent republication.
Cover the issue with original, non-explicit reporting. Do not assume legitimate reporting or consensual material is automatically classified as non-consensual fake imagery.
Sources: Google: explicit fake content ranking protections Google: Search documentation change history

Site reputation abuse clarification (Policy)

Launched: Policy clarification announced and effective 19 November 2024.
Roll outs: Not another November core rollout. January 2025 FAQ edits were editorial; August 2026 later changed enforcement for EEA searchers.
Goal: Address exploitative third-party sections even when a publisher approves them—for example, white-label casino pages hosted mainly for the publisher’s ranking signals.
Hazards: Treating editorial sign-off or a licensing deal as automatic proof of compliance.
Moving the same offending pages from a subfolder to a subdomain without fixing the exploitation.
How it works: The November rule said first-party oversight alone did not resolve site reputation abuse. Third-party content, freelance work, and affiliate relationships were not automatically violations. The issue was exploiting host signals; current regional enforcement also requires the August 2026 qualification.
How to stay safe: Inspect the affected section in Search Console’s Manual actions report; fix the stated issue before requesting reconsideration.
Review what third-party pages contribute for readers. Do not merely relocate violating pages or redirect URLs to recreate the exploitation.
Sources: Google: November 2024 site reputation clarification Google: current spam policies

Discover updates

Launched: 5 February 2026 (the rollout shown below)
Roll outs: The February 2026 Discover rollout ran on 5–27 February (US/Pacific), initially for English-language users in the US. Its announcement planned expansion to other countries and languages.
February 2026 Discover update
5–27 February 2026 (US/Pacific)
A Discover-specific core update, initially for English-language users in the US, emphasizing local relevance, less clickbait, and deeper original content.
Google Search Status Dashboard — February 2026 Discover update Google Search guidance
Goal: Surface more locally relevant, original, in-depth content and reduce clickbait in Discover.
Hazards: A headline promises a revelation the article never provides
The page repeats another publisher's story without additional reporting or explanation
A fall in Discover clicks is investigated only in the Web Search report
How it works: The February 2026 update changed the systems that surface articles in Discover. Google described stronger local relevance and topic-level expertise, alongside less sensational content. This is a Discover-specific update, and its traffic effects should be assessed in that context.
How to stay safe: Check that the headline matches what the article actually establishes. Add original reporting, the facts behind a claim, and clear dates when the story depends on recent events.
Use Search Console's Discover performance report to find the affected stories and dates. Compare that change separately from Web Search traffic.
Sources: Google: February 2026 Discover core update Google Search Status Dashboard — February 2026 Discover update

Back button hijacking (Policy)

Launched: Announced 13 April 2026; enforcement began 15 June 2026.
Roll outs: An explicit malicious-practices spam policy, not a June core rollout. Affected pages may face manual actions or automated demotions.
Goal: Let visitors return to their previous page when they press Back.
Hazards: After arriving from Search, a visitor presses Back but an ad script sends them to another ad page.
Deceptive history entries or intercepted navigation prevent an immediate return to the previous page.
How it works: The problem is defeating expected Back behavior, including through advertising code or included libraries. Normal navigation and legitimate history updates are different from trapping a visitor; using a browser-history API alone is not the violation.
How to stay safe: Reproduce arrival and Back navigation; identify and disable the script or ad integration that traps visitors.
Retest pages with third-party ads. If a manual action exists, fix the behavior and submit a reconsideration request in Search Console.
Sources: Google: back button hijacking policy and enforcement date Google: current malicious-practices policy

Site reputation: EEA enforcement (Policy)

Launched: Announced 28 August 2026; enforcement changed from 30 August 2026.
Roll outs: A regional policy change, not a core rollout. The distinction follows the searcher’s location, not the publisher’s location.
Goal: Address sections exploiting host signals while applying different enforcement inside the European Economic Area (EEA).
Hazards: Assuming an action affects a coupon section identically for searchers in Jakarta and Berlin.
Assuming the EEA change guarantees that a third-party section retains its host’s ranking benefit.
How it works: Outside the EEA, a site-reputation manual action can affect the relevant section’s results. For EEA searchers, that impact does not apply; the section may instead be separated and rank independently over time. Neither outcome guarantees previous rankings or makes every third-party page a violation.
How to stay safe: Read Search Console’s notice and compare affected results by searcher region. Address the issue or request reconsideration if mistaken.
Review the section’s purpose, integration, and editorial contribution. Do not assume an action outside the EEA requires noindex for EEA results; current documentation says otherwise.
Sources: Google: August 2026 site reputation enforcement change Google: current regional site reputation policy