Key Takeaways

  • Google documents 17 active ranking systems in 2026, from language systems like BERT and RankBrain to spam detection, and lists four more (Hummingbird, Panda, Penguin and the helpful content system) as retired into its core ranking systems.
  • Most of the 17 exist to understand what a searcher means or to judge quality and trust, so a page that answers one question clearly, shows original work and earns genuine links is rewarded by several systems at once.
  • Page experience is not a ranking system of its own. Google says Core Web Vitals are used by its ranking systems, and a good score means the largest element loads within 2.5 seconds, interactions respond within 200 milliseconds and layout shift stays under 0.1.
  • Google confirmed five core updates and five spam updates between March 2025 and October 2026, and the November 2026 core update began rolling out on November 10.
  • AI Overviews and AI Mode have no additional requirements: a page that is indexed and eligible to show a snippet can be cited, which makes strong traditional SEO the foundation of AI search visibility too.

Google’s ranking systems are the automated programs that decide which pages appear for a search and in what order. There is no single algorithm. Google’s own guide names 17 systems working together, each with a specific job, and it lists four more it has retired. When a client’s traffic moves after an update, the first step in an SEO site audit is working out which of those systems the change points to, because the response to a spam demotion looks nothing like the response to a page that simply answers the question less completely than a competitor.

I first published this guide in November 2024, and this November 2026 version rebuilds it against Google’s current documentation: every system Google lists today, the four it has retired, the Core Web Vitals thresholds it uses now, every confirmed update since March 2025, and how AI Overviews and AI Mode fit in. After more than 15 years of SEO work at Paul Teitelman SEO Consulting, I still read Google’s own documentation before anyone’s commentary on it, and that is the approach this guide takes.

How Does Google Decide What Ranks?

Google explains its approach in How Search Works, where it groups everything into five key factors. Every ranking system in this guide serves one or more of them.

  • Meaning. Working out the intent behind the words, including spelling, synonyms and whether the search calls for fresh results.
  • Relevance. Whether a page contains information that answers the query, starting with whether it uses the same words and going on to how well it covers the topic.
  • Quality. Which of the relevant pages seem most helpful, judged through signals of expertise, authoritativeness and trustworthiness.
  • Usability. When other signals are roughly equal, the page people will find easier to use may perform better.
  • Context. Location, language and settings shape which results are most useful to a particular person.

All 17 Google Ranking Systems at a Glance

Here is every active system in Google’s ranking systems guide, grouped by the job it does, with what each one means for your site.

Google's 17 ranking systems grouped into four jobs: understanding the search, judging quality and trust, shaping the results page, and fighting spam, with the four retired systems listed below

Ranking systemWhat it doesWhat it means for your site
BERTUnderstands how word order and small words like “for” and “to” change meaningWrite the way your customers speak
RankBrainRelates words to concepts, so a page can rank without the exact query wordsCover the topic, not just one keyword
Neural matchingMatches the concepts in a query to the concepts on a pageExplain ideas fully in plain language
MUMA language model used for specific tasks, not general rankingNo action needed for ranking
Passage rankingJudges individual sections of a page on their own meritsGive each question a clear heading and a direct answer
Link analysis and PageRankUses how pages link to each other to judge topic and valueEarn links a real reader would follow
Reliable informationSurfaces authoritative pages and demotes low-quality contentShow who wrote it and why they can be trusted
Original contentFavours original reporting over pages that only cite itPublish your own data, cases and photos
ReviewsRewards reviews with insightful analysis and original researchShow real experience with what you review
FreshnessShows newer content where a search expects itUpdate time-sensitive pages when the facts change
DeduplicationShows the most relevant of many near-identical resultsOne strong page per topic beats several similar ones
Site diversityGenerally shows no more than two listings per site in the top resultsMerge pages that compete with each other
Exact match domainLimits the credit given to domains built to match a searchBuild a brand name, not a keyword domain
Local newsIdentifies local news sources for news featuresMainly relevant to local publishers
Crisis informationShows hotlines and trusted alerts in personal crises and disastersNo action needed for most businesses
Spam detection (SpamBrain)Finds content and behaviour that break the spam policiesStay well inside the spam policies
Removal-based demotionDemotes sites with high volumes of valid legal or personal information removalsRespect copyright and privacy

The first group works on the query itself, turning the words someone types or says into an understanding of what they actually want.

BERT

Introduced in 2019, BERT (Bidirectional Encoder Representations from Transformers) helps Google understand how combinations of words express different meanings and intent. It reads a query as a whole, so small words that change the meaning, like “to” or “for”, are taken into account rather than skipped. As we said in our post on BERT’s launch, there is nothing to optimize for BERT beyond writing the way your customers speak.

RankBrain

RankBrain arrived in 2015 as the first deep learning system Google deployed in Search, and it is still on Google’s list of active systems in 2026. It helps Google relate words to concepts, which means a page can rank for a search even when it does not contain the exact words used. The lesson for a business is to cover a topic the way an expert would explain it, rather than repeating one keyword.

Neural Matching

Added in 2018, neural matching connects the concepts in a query with the concepts on a page. It is how Google can match a vague, conversational search to a page that answers it in different words. Pages that explain their subject clearly and completely give this system the most to work with.

MUM

MUM (Multitask Unified Model) is an AI model that can both understand and generate language. Google is specific that MUM is not used for general ranking: it powers particular applications, such as improving searches for COVID-19 vaccine information and the callouts shown with featured snippets. There is nothing a site needs to do for MUM.

Passage Ranking

Announced in 2020, passage ranking lets Google judge an individual section of a long page on its own merits, so a strong answer halfway down a long guide can rank for a specific question. That is the case for clear headings with a direct answer under each one, which is also how AI answers tend to pull information from a page.

The Systems That Judge Quality and Trust

The second group decides which of the relevant pages deserve to rank, using links, authority, originality and expertise.

Google uses how pages link to one another to understand what they are about and which might be most helpful. PageRank, developed by Google’s founders at Stanford in the late 1990s, is one of the core systems in that work, and Google says it has evolved a lot since then. Links still matter, and they carry the most weight when they come from relevant sites a real reader might click through from, which is how we approach link building for clients.

Reliable Information Systems

These systems surface the most authoritative pages, demote low-quality content and elevate quality journalism. When Google is not confident about the results for a fast-moving topic, they can show a content advisory instead. For a business, authority is built in the open: a named author, real credentials, cited sources and a reputation that other sites confirm.

Original Content Systems

Google’s original content systems help original content, including original reporting, rank above pages that merely cite it. They also read canonical markup to work out which page is the primary version. Publishing your own data, case studies and photographs is the most direct way to benefit, and it is a theme that runs through every quality system Google describes.

Reviews System

The reviews system, renamed from the product reviews system when Google widened it beyond products, rewards reviews with insightful analysis and original research, written by experts or enthusiasts who know the subject well. If your site reviews products, services or suppliers, show the evidence that you have used them: your own photos, measurements and a clear verdict.

The Systems That Shape the Results Page

The third group decides which of the qualifying pages actually appear, how fresh they need to be, and how many can come from one source.

Freshness Systems

Freshness systems show newer content for searches where people expect it, such as a recently released film or a breaking event. Not every search needs fresh results, so an update only helps when the facts on the page have actually changed. When they have, update the page, its date and its details together, which is exactly what this guide has done.

Deduplication Systems

A search can match thousands of near-identical pages, so deduplication systems show only the most relevant of them. They also stop a page shown in a featured snippet from being repeated lower on the first page. Several thin pages on one topic end up competing with each other, which is why one strong page usually wins more than five similar ones.

Site Diversity System

Google generally shows no more than two listings from the same site in its top results, and treats subdomains as part of the main domain for this purpose, although it can show more when they are especially relevant. This is another reason to consolidate overlapping pages rather than publish variations of the same article.

Exact Match Domain System

This system stops Google from giving too much credit to domain names built to match a search exactly, such as a domain that strings together a service, a city and the word “best”. A domain works best as a brand, and brands earn their rankings through the quality systems above.

Local News Systems

Local news systems identify and surface local news sources in features like Top stories and Local news, and they matter mainly to local publishers rather than to businesses.

Crisis Information Systems

When a search suggests a personal crisis, such as one related to suicide, sexual assault or poison ingestion, these systems show hotlines and content from trusted organizations. During natural disasters and other emergencies, SOS Alerts show updates from local, national and international authorities.

The Systems That Fight Spam and Abuse

The last group protects the results from manipulation, and it is the group that matters most for anyone producing content at scale.

Spam Detection Systems and SpamBrain

Google’s spam detection systems, including the AI-based SpamBrain, find content and behaviour that break its spam policies and either demote it or remove it from results. Spam updates are improvements Google makes to these systems. The policies to know in 2026 include scaled content abuse, where many pages are generated mainly to manipulate rankings rather than help people, site reputation abuse, where third-party content is published on a host site mainly to borrow its ranking signals, and expired domain abuse. All three were added to the policies in March 2024, alongside long-standing rules on link spam, cloaking and doorway pages. For any team using AI to produce content, scaled content abuse is the policy to understand: Google’s concern is pages made to rank rather than to help, however they are produced.

Removal-Based Demotion Systems

Sites that receive a high volume of valid copyright removal requests can have their other content demoted, and so can sites with exploitative removal practices for personal information. Most businesses never encounter this system, and respecting copyright and privacy keeps it that way.

The Retired Systems: Where Panda, Penguin, Hummingbird and Helpful Content Went

Google also keeps a short list of systems it has retired. In Google’s documentation, retired means a system was folded into the core ranking systems or superseded as those systems evolved, so the ideas behind each one still shape results today.

Timeline of Google's four retired ranking systems: Hummingbird in 2013, Panda in 2015, Penguin in 2016 and the helpful content system in March 2024, each folded into the core ranking systems

  • Hummingbird (2013). A major improvement to Google’s overall ranking systems made in August 2013, after which the systems kept evolving.
  • Panda (into the core in 2015). Panda judged content quality, and its work became part of the core ranking systems in 2015.
  • Penguin (into the core in 2016). Penguin targeted link spam and was made part of the core ranking system in 2016.
  • Helpful content (into the core in March 2024). The helpful content system judged whether content was made for people or for search engines, and it became part of the core ranking systems in March 2024. Our take on the helpful content update still applies, because the standard stayed when the separate system went.

Page Experience and Core Web Vitals: A Signal, Not a System

The earlier version of this guide, like our coverage of the page experience update at the time, treated page experience as a ranking system. Google is clearer about it today: “There is no single signal.” Its core ranking systems look at a variety of signals that line up with a good page experience, and Core Web Vitals are the part Google confirms its ranking systems use.

Core Web VitalWhat it measuresGood score
Largest Contentful Paint (LCP)Loading: when the largest element on the page appearsWithin 2.5 seconds
Interaction to Next Paint (INP)Responsiveness: how quickly the page reacts to a tap, click or key pressUnder 200 milliseconds
Cumulative Layout Shift (CLS)Visual stability: how much the layout moves while it loadsUnder 0.1

Interaction to Next Paint replaced First Input Delay as a Core Web Vital in March 2024. Google is also candid that good scores do not guarantee top rankings, and that page experience aspects beyond Core Web Vitals do not directly help a page rank higher. Fast, stable pages earn their keep with visitors first: when our own site moved to a code-based build, the largest element on its mobile homepage went from 4.7 seconds to 2.1. If your site is slow, page speed optimization is usually the quickest technical win available.

Core Updates and Spam Updates: How the Systems Change

Google changes its systems constantly, but it confirms two main kinds of change on its Search Status Dashboard. A core update is a broad change to the core ranking systems that shifts how pages are assessed across the web, while a spam update improves the spam detection systems. Here is every confirmed ranking update since March 2025.

Timeline of every confirmed Google ranking update from March 2025 to November 2026: six core updates, five spam updates and one Discover update, with the length of each rollout

UpdateStartedRollout
March 2025 core updateMarch 13, 2025About 14 days
June 2025 core updateJune 30, 2025About 17 days
August 2025 spam updateAugust 26, 2025About 27 days
December 2025 core updateDecember 11, 2025About 18 days
February 2026 Discover updateFebruary 5, 2026About 22 days
March 2026 spam updateMarch 24, 2026Under a day
March 2026 core updateMarch 27, 2026About 12 days
May 2026 core updateMay 21, 2026About 12 days
June 2026 spam updateJune 24, 2026About 2 days
August 2026 spam updateAugust 18, 2026About 3 days
September 2026 spam updateSeptember 24, 2026About 14 days
November 2026 core updateNovember 10, 2026Still rolling out at the time of writing

Two habits make updates far easier to read. First, wait until Google marks a rollout complete before judging it, because rankings move throughout. Second, compare the same number of days before and after in Search Console, by page and by query, rather than reacting to a single day. Our recap of Google’s 2023 algorithm updates shows how much a single year of updates can reshape a market.

How Do AI Overviews and AI Mode Fit In?

The biggest change since this guide first ran is that many searches now open with an AI answer. Google’s documentation on AI features is direct: “There are no additional requirements to appear in AI Overviews or AI Mode.” A page needs to be indexed and eligible to show in Search with a snippet, and the same SEO fundamentals apply: crawlable pages, internal links, important content in text and structured data that matches what is visible on the page.

Two further details matter when you plan for AI search. Both features can use what Google calls query fan-out, issuing several related searches across subtopics to find supporting pages, which rewards sites that cover a topic in depth. And traffic from these features is counted in Search Console’s Performance report under the Web search type, so it is already in your numbers. Google adds that clicks from pages with AI Overviews tend to be higher quality, with visitors more likely to spend time on the site.

Being cited is still different from ranking first, and the brands that appear most often in AI answers are usually the ones the wider web already discusses. How AI SEO differs from traditional SEO covers that gap, and an AI search audit measures where your business appears today.

What This Means for Your Site: A Practical Checklist

  1. Answer one question per page, and answer it early. BERT, neural matching and passage ranking all reward a clear question with a direct answer under a descriptive heading.
  2. Show original work. The original content, reviews and reliable information systems all favour first-hand data, real examples and named, credible authors.
  3. Earn links a real reader would click. Link analysis rewards relevance and genuine endorsement, and the spam systems watch for everything else.
  4. Meet the Core Web Vitals thresholds on mobile. Aim for the largest element within 2.5 seconds, responses under 200 milliseconds and a layout shift under 0.1.
  5. Keep one strong page per topic. Deduplication and site diversity mean overlapping pages compete with each other, so merge them and redirect the weaker addresses.
  6. Stay well inside the spam policies. Pay particular attention to scaled content abuse if your team uses AI to produce content, and to site reputation abuse if you host third-party content.
  7. Read updates in Search Console once they finish. Compare equal periods before and after, by page and by query, before changing anything.

If you would like a second pair of eyes on any of this, our on-page optimization work starts from exactly this checklist.

Paul Teitelman beside an invitation to book a free consultation and audit your site against Google's ranking systems

Frequently Asked Questions

How Many Ranking Systems Does Google Use?

Google’s own ranking systems guide lists 17 active systems in 2026, from BERT, RankBrain and neural matching to link analysis, freshness, site diversity and spam detection. It also lists four retired systems, Hummingbird, Panda, Penguin and the helpful content system, whose work now sits inside the core ranking systems. Google presents these as its notable systems rather than a complete list, and many signals feed into them, so the number is best read as a map of the main jobs rather than an exact count.

Does Google Still Use RankBrain?

Yes. RankBrain is still listed as an active system in Google’s ranking systems guide in 2026. Launched in 2015, it was the first deep learning system Google deployed in Search, and it helps Google relate words to concepts so that a page can rank even when it does not contain the exact words someone searched. There is nothing to optimize for RankBrain directly, and covering a topic thoroughly, in the language your customers actually use, gives it the most to work with.

What Happened to the Helpful Content Update?

The helpful content system became part of Google’s core ranking systems in March 2024, and Google now lists it among its retired systems. Being retired does not mean that the standard disappeared. The core systems still look for content made to help people rather than to rank, and they now apply that judgement through every core update instead of a separate one, which is why recoveries and declines since then have moved with core updates rather than with helpful content updates.

Are Core Web Vitals a Ranking Factor?

Yes, Google says Core Web Vitals are used by its ranking systems, and it recommends good scores for success in Search. A good score means the largest element loads within 2.5 seconds, the page responds to interactions in under 200 milliseconds, and the layout shift score stays under 0.1. Google also says good scores do not guarantee top rankings, so treat them as a baseline: fast, stable pages keep visitors on the page long enough for the content to do its work.

How Do You Show Up in Google’s AI Overviews?

Google says there are no additional requirements to appear in AI Overviews or AI Mode. A page needs to be indexed and eligible to show in Search with a snippet, and the usual fundamentals apply: crawlable pages, internal links, important content in text and structured data that matches the page. Beyond that, clear answers under descriptive headings, original information and a brand the wider web discusses all make a page easier to cite when Google runs several related searches to build an answer.

How Does Google Rank Local Businesses?

Google says local results are mainly based on three factors: relevance, distance and prominence. Relevance is how well a Business Profile matches the search, distance is how far the business is from the person searching or the location in the search, and prominence is how well known the business is. A complete Business Profile, accurate details across the web, steady reviews and a strong website all help, and our local SEO work is built around those three factors.

How Often Does Google Update Its Ranking Systems?

Google changes Search constantly, but it confirms the larger updates on its Search Status Dashboard. Between March 2025 and October 2026 it confirmed five core updates, five spam updates and one Discover update, and the November 2026 core update began on November 10. Core updates have recently taken about two weeks to roll out. The practical rhythm is to expect several confirmed updates a year, wait for each one to finish, and then compare equal periods in Search Console before making changes.