What Google Search ranking systems are and how they work

Reading time : 9 minutes

Google’s algorithms work tirelessly to provide users with the most useful information, using various factors and indicators to evaluate content and create rankings.As we know, there are hundreds of ranking factors, such as the words used in the query, the relevance and usability of pages, the reliability of sources, the user’s geographic location, and his or her browsing settings. All of these processes take place virtually instantaneously and make up what Google has now dubbed “Google Search ranking systems“, which are the focus of a new official guide explaining how they work and revealing which algorithms are actually used for ranking and which have been deprecated.

Google Search ranking systems: how the automated systems work

On the page now published on Google Search Central (and at the moment not yet translated into Italian) we read that Google uses automated ranking systems that examine many factors and flag hundreds of billions of web pages and other content in our search index “to present the most relevant and useful results, all in a fraction of a second.”

This complex mechanism relies on several “ranking systems“, some of which are part of Google’s main ranking systems, which are the underlying technologies that produce search results in response to queries, while others are involved in specific ranking requirements.

Google regularly improves these systems through rigorous testing and evaluation and provides “notification of updates to our ranking systems when they might be useful to content creators and others,” always with the ultimate goal of fulfilling the mission “to organize information globally and make it universally accessible and useful.”

Google search, the main factors

Given the vast amount of information available, it would be virtually impossible to find what we’re looking for on the Web without an organizing tool: that is what Google’s ranking systems do, which are designed precisely to sort “hundreds of billions of web pages and other content into the search index to provide useful and relevant results in a split second.”

As mentioned, these algorithms are based on an (extensive) set of factors that also vary in weight and importance depending on the type of search-for example, the date of publication of content plays a stronger role in answering queries related to topical issues than queries about dictionary definitions-but they all fall into five broad categories of major factors that determine the results of a query:

  • Meaning. That is, the intent of the search, with language models trying to understand how the few words entered in the search box match the most useful content available.
  • Relevance. Systems subsequently analyze content to assess whether it contains information relevant to the search query (e.g., whether it includes the same keywords as the query on the page, in headers, or in the body of the text), using aggregated, anonymized data on interactions to verify that the page has other relevant content beyond just keywords.
  • Quality. Google’s systems then prioritize content that seems most useful by identifying indicators that help identify content that emphasizes expertise, authority, and trustworthiness, i.e., E-E-A-T metrics.
  • Usability. In usability analysis, content deemed most accessible by users may also perform best, with ratings on aspects such as ease of viewing from mobile devices or speed of loading.
  • Context. Information such as location, previous search history, and Search settings allow Google to ensure that the results shown to a user match what is most useful and relevant to them at that moment.

The clarification: ranking systems and system updates

The article with which Danny Sullivan introduces Google Search ranking systems also dwells on an important lexical distinction: in contrast to what has been done so far, in fact, Google has decided to differentiate the use of the words “systems” and “update” to avoid confusion, especially when subsequent enhancement updates intervene.

Therefore, the system identifies a ranking algorithm, while the word update will be used only for subsequent improvements to that process. More precisely, a system is constantly running in the background, while update refers to a one-time change to ranking systems.

Concretely, then, Google admitted that the naming Page Experience Update or Helpful Content Update is incorrect, because it then makes it complicated to name and understand updates (which become something like “update of updates”), and so it anticipated that in the future it will use the wording “system” for algorithmic innovations and “update” only for updates to the respective systems.

Which Google ranking systems are currently active

Google’s new guide then goes into detail about the currently active and functioning Search ranking systems, listing them in alphabetical order.

  1. BERT. Short for Bidirectional Encoder Representations from Transformers, BERT allows Googe to understand how combinations of words can express different meanings and intents.
  2. Crisis information systems. Google has developed systems to provide specific sets of useful and timely information during times of crisis, whether in personal crisis situations (when people search for information with queries related to suicide, sexual assault, poison ingestion gender-based violence or drug addiction Google shows hotlines and content from trusted organizations) and for more general crises (such as SOS alerts during times of natural disasters or widespread crisis situations such as floods, fires, earthquakes, hurricanes and other disasters, whereby Google shows updates from local, national or international authorities with emergency phone numbers and websites, maps, translations of useful phrases, donation opportunities and more).
  3. Deduplication systems. Google’s search systems aim to avoid posting duplicate or near-duplicate Web pages: Google searches can find thousands or even millions of matching Web pages, which can sometimes be very similar to each other, and algorithms show only the most relevant results to avoid unnecessary duplication. Deduplication also occurs with featured snippets: if the positioned result of a web page is elevated to become a featured snippet, it will not be repeated a second time on the first results page.
  4. Exact match domain system. Google’s algorithms “consider words in domain names as one of many factors in determining whether content is relevant to a search query,” but this specific algorithm ensures that it does not give too much credit to content hosted on domains “designed to match certain queries exactly” – for example, creating a domain name containing the words “the best places to eat lunch” in the hope that all those words in the domain name will push the content up in the rankings is useless.
  5. Freshness systems. Google has several “query deserves freshness” systems designed to show more up-to-date content for queries where freshness is needed and expected. For example, if someone is looking for information about a newly released movie, they will probably want recent reviews rather than articles older than when production began; or, normally a search for “earthquake” might bring up material on preparedness and resources, but if an earthquake has occurred recently, articles with more recent news and content might appear.
  6. Helpful content system. So far known as HCU, as mentioned above, this is a system designed to ensure that people see original and useful content “written by people, for people” in SERPs, rather than content created primarily to get traffic from search engines.
  7. Link analysis systems and PageRank. Google has several systems that understand how pages link to each other to determine what pages are about and which ones might be most useful in response to a query. Among these is PageRank, one of the main ranking systems used when Google was first launched.Although how PageRank works has evolved a lot since then, it continues to be part of the search engine’s main ranking systems.
  8. Local news systems. Algorithms that work to identify and surface local news sources if they are relevant to the query, for example through our “Main News” and “Local News” features.
  9. MUM. Short for Multitask Unified Model, MUM is an artificial intelligence system that can understand and generate language. It is not currently used for general ranking in Search, but rather for some specific applications, for example, to improve searches for COVID-19 vaccine information and to improve callouts of featured snippets appearing in SERPs.
  10. Neural matching. Neural matching is an artificial intelligence system that Google uses to understand representations of concepts in queries and pages and match them together.
  11. Original content systems. They are used to ensure that Google shows original content prominently in search results, including original reports, before pages that simply quote them; . this includes support for special canonical markup that creators can use to help Google better understand what the main page is if a page has been duplicated in multiple places.
  12. Removal-based demotion systems. Google has rules that allow removal of certain types of content: if a site receives a high volume of valid requests to remove content, this is used as a signal to provide better results (and the site is demoted in searches). Google specifically distinguishes between legal removals (demotion signals for copyright infringement or defamation claims, counterfeit goods, and court-ordered removals) and personal information removals (demotion of sites that engage in retaliatory exploitative removal practices or for doxxing content and automatic protections designed to prevent explicit non-consensual personal images from ranking high in responses to queries involving names).
  13. Passage ranking system. This is an artificial intelligence system used by Google to identify individual sections or “passages” of a web page to better understand how relevant a page is to a search query.
  14. Reviews system. An algorithm designed to reward high-quality product reviews, content that provides in-depth analysis and original research, written by experts or enthusiasts who know the topic well.
  15. RankBrain. Is an artificial intelligence system that helps Google understand how words are related to concepts; RankBrain allows Google to return relevant results even if they do not contain all the exact words used in a query, understanding that the content is related to other words and concepts.
  16. Reliable information systems. Google has multiple systems for displaying reliable information, such as bringing up more authoritative pages, rewarding quality journalism, and degrading low-quality content; if reliable information is missing, the systems automatically display content alerts on fast-moving topics or signal that Google does not trust that much the overall quality of the results available for search, suggesting to the user how to search in ways that might lead to more useful results.
  17. Site diversity system. This algorithm prevents Google from showing more than two web page results from the same site in the top positions, to prevent a single site from dominating all the top results. However, Google may still show more than two results in cases where the systems determine that it is particularly relevant to do so for a particular search.
  18. Spam detection systems. They deal with content and behaviors that violate Google’s anti-spam rules; the Internet still has huge amounts of spam that, if left unmanaged, would prevent the most useful and relevant results from being shown, which is why a number of spam detection systems, including SpamBrain, handle content and behaviors that violate anti-spam rules and are constantly being updated to keep up with the latest ways in which the spam threat is evolving.

Google systems that are no longer active (or embedded in other tools)

The guide page also lists for historical purposes some of Google’s systems that are no longer independently active, but are now found to be embedded in subsequent or have become part of the search engine’s broader core ranking systems (which are the underlying technologies that produce search results in response to queries).

  1. Page Experience system – Introduced in 2020, Page Experience was introduced as a system that evaluated a number of criteria to determine whether a web page offers a good user experience, specifically analyzing page loading speed, level of mobile optimization, absence of intrusive interstitials, and use of HTTPS protocol for security. In situations where there are many possible matches with nearly equal level relevance, this system helps give preference to content with a better on-page experience. In April 2023 it disappeared from the list of active systems, and Google subsequently redefined Page Experience rather as “a concept to describe a set of key aspects of on-page experience that site owners can focus on,” and not a “separate ranking system.” To be precise, Danny Sullivan clarified that the Page Experience complex was never actually a ranking system, but rather “signals used by other systems”
  2. Hummingbird. Launched in August 2013, this was a major enhancement to Google’s overall ranking systems, which have since “continued to evolve, just as they had evolved before.”
  3. Mobile-friendly ranking system. This system intervenes in situations where there are many possible matches with relatively equal relevance, giving priority and preference to mobile-friendly content that has better display on mobile devices, which is more useful for people searching from smartphones and tablets. Today it is incorporated into the Page Experience system.
  4. Page speed system. Originally announced in 2018 as the “Page Update,” this algorithm would intervene in peer situations to better rank content that loaded faster for users from mobile devices. Today it has become part of the Page Experience system.
  5. Panda system. Announced in 2011 and dubbed “Panda“, this system was designed to ensure the visibility of original, high-quality content in search results. Over time it has evolved and since 2015 has become part of Google’s core ranking systems.
  6. Penguin system. Announced in 2012 and dubbed the “Penguin Update,” it was designed to combat link spam and was integrated into the core ranking systems in 2016.
  7. Secure sites system. Announced in 2014, it was an algorithm that ensured priority in rankings for HTTPS-protected sites all else being equal; according to Google, it helped encourage the growth of secure sites at a time when the use of HTTPS was still quite rare, and has since become part of the on-page experience system.

The evolution of Google – and SEO

This information is useful first of all to have a compass on what are the main systems that are currently at work to form Google’s rankings and SERPs, but also to know some interesting details about Google’s consideration of these systems and what contribution they actually make to ranking.

For example, we can see that in most cases these are tie-breaker systems, that is, they serve to break the parity of factors and conditions thereby determining which page and content should appear first. It is then curious to find out that Google still uses a system that interprets the exact match for the domain, but then concretely tells us that it is not worth investing on such a constructed domain name just for ranking purposes because it would be in vain.

More generally, however, this guide gives us practical information for our own business, starting with the lexical shift desired by the search engine-although (at least for now) we will not change old articles renaming updates to system, thus leaving the old names, also as a matter of habit.

To be sure, this is yet another sign of how much Google is changing and continuing to evolve, both in the way it presents information to users and in algorithm updates, which consequently determine an adaptation of SEO best practices as well, which must keep up with what it means to properly optimize a website today.

For example, until not so long ago the definition of relevance simply meant that a web page had to be about what the user was searching for, but today that is no longer sufficient because content must also be useful, original, and tied directly to the search intent. Google is increasingly moving away from keyword identification to an understanding of the multiple meanings inherent in search queries, and it has made it clear to creators to stop writing content focused only on keywords because it appears unnatural and forced.

The other considerable aspect is context, the setting in which something is said or done, which provides meaning to those actions or settings.Today, the context of a search query can influence the results, and Google is redefining what it means to be relevant by understanding the user’s context.