--- type: "Article" title: "RankBrain | topical authority content" description: "RankBrain explained: what Google confirms, how it relates words to concepts, the myths, and how to test your pages for concept gaps." resource: "https://www.laurelinlabs.com/search-systems/rankbrain" tags: ["rankbrain", "topical-authority-content", "google-machine-learning-ranking"] generated: { by: "laurelin-site-build/2026-09-02", at: "2026-09-02T12:00:00+00:00" } status: "stable" stale_after: "2027-03-02T00:00:00+00:00" --- # RankBrain RankBrain is a machine learning system Google uses to understand how words relate to concepts, so it can rank pages that are relevant to a search even when they do not contain every word in it. Google confirms it in its ranking systems guide and describes it as its first deep learning system in Search. It is not something a site can optimise for directly. What a site can do is make sure each page clearly covers the concept behind the searches it is shown for. ## What is RankBrain? RankBrain is an AI system Google uses in ranking to understand how the words in a search relate to real world concepts. It lets Google return a relevant page even when that page does not repeat the exact words the searcher typed. Google lists RankBrain as a live system in its [ranking systems guide](https://developers.google.com/search/docs/appearance/ranking-systems-guide), alongside [BERT](https://www.laurelinlabs.com/search-systems/bert) and [neural matching](https://www.laurelinlabs.com/search-systems/neural-matching). ## How does RankBrain work? Google has said little about the internals. Its public description is that RankBrain learns from seeing words on many pages which concepts they belong to, then uses that to reorder top results. In a [February 2022 post](https://blog.google/products-and-platforms/products/search/how-ai-powers-great-search-results/) Pandu Nayak gave the example of a search about the consumer at the top of a food chain being matched to pages about apex predators. Google separates the jobs of its language systems. In [March 2019 Google's Search Liaison account](https://www.seroundtable.com/google-explains-neural-matching-vs-rankbrain-27300.html) summarised it as RankBrain relating pages to concepts and neural matching relating words to searches. RankBrain works on ranking, not on finding candidate documents in the first place. ## What evidence is there that Google uses RankBrain? The evidence label is **Confirmed**. Google names RankBrain in its [ranking systems guide](https://developers.google.com/search/docs/appearance/ranking-systems-guide) and described it at length in a [2022 blog post by Pandu Nayak](https://blog.google/products-and-platforms/products/search/how-ai-powers-great-search-results/), which states it is still one of the major AI systems in Search. The launch date is less tidy than most summaries suggest. The 2022 Google post says RankBrain launched in 2015, and [Wikipedia records Google confirming it on 26 October 2015](https://en.wikipedia.org/wiki/RankBrain), while a [March 2019 Search Liaison tweet reported by Search Engine Roundtable](https://www.seroundtable.com/google-explains-neural-matching-vs-rankbrain-27300.html) says Google began using it in 2016. The likeliest reading is a 2015 launch with wider use in 2016, which matches [Search Engine Land's June 2016 report](https://searchengineland.com/faq-all-about-the-new-google-rankbrain-algorithm-234440) that it was then used for every query, but Google has not reconciled the two dates. ## What are the common myths about RankBrain? The most repeated claim is that RankBrain is the third most important ranking factor. That comes from a 2015 interview reported by Bloomberg, which [Wikipedia summarises](https://en.wikipedia.org/wiki/RankBrain) as Google saying it was third after links and content. It is a single 2015 remark, not a current Google statement, and Google has not repeated it. - Myth: you can optimise for RankBrain. Google's [Search Liaison said in 2019](https://www.seroundtable.com/google-explains-neural-matching-vs-rankbrain-27300.html) there is nothing special site owners need to do. - Confirmed: it is still in use. Google's [ranking systems guide](https://developers.google.com/search/docs/appearance/ranking-systems-guide) lists it among current systems, not retired ones. ## What should sites do about RankBrain? Since RankBrain matches concepts rather than strings, repeating a keyword adds little. The practical, testable question is whether a page actually covers the concept behind the searches Google already shows it for. Search Console shows those searches. Pull the queries a page earns impressions for, group them by the concept they express, and check the page answers each group. A cluster of related queries with impressions but poor positions often means Google associates the page with a concept the page never addresses. That is a hypothesis to test by adding the missing coverage and watching the same queries, not a guaranteed fix. ## Which Laurelin audit checks test for RankBrain? No site audit can observe RankBrain itself. These checks measure proxies: whether pages cover the concepts searchers are looking for. [Search Console queries show concepts the page never covers](https://www.laurelinlabs.com/audit/topical-content/331-search-console-queries-show-concepts-the-page-never-covers): the new check for this page, comparing the concepts in a page's queries with what the page says. [Content not matching search intent](https://www.laurelinlabs.com/audit/topical-content/168-content-not-matching-search-intent): a page answering a different need from the query cannot be rescued by concept matching. [Title doesn't match content intent](https://www.laurelinlabs.com/audit/topical-content/137-title-doesn-t-match-content-intent): a mismatched title sends mixed concept signals. [Keyword stuffing and over-optimisation](https://www.laurelinlabs.com/audit/topical-content/167-keyword-stuffing-over-optimisation): repetition of one phrase is the habit RankBrain made least useful. [No clear answer to the page's question](https://www.laurelinlabs.com/audit/topical-content/170-no-clear-answer-to-the-page-s-question): being matched to a concept is wasted if the page never resolves it. ## What are the key dates for RankBrain? - 2015-10-26: Google confirms RankBrain (Bloomberg report) ([source](https://en.wikipedia.org/wiki/RankBrain)) - 2016-06: Reported as used for every query Google handles ([source](https://searchengineland.com/faq-all-about-the-new-google-rankbrain-algorithm-234440)) - 2019-03-21: Search Liaison explains RankBrain versus neural matching ([source](https://www.seroundtable.com/google-explains-neural-matching-vs-rankbrain-27300.html)) - 2022-02-03: Google describes RankBrain as still a major AI system in Search ([source](https://blog.google/products-and-platforms/products/search/how-ai-powers-great-search-results/)) ## FAQ ### Is RankBrain still used by Google? Yes. Google lists RankBrain as a current system in its ranking systems guide, and a 2022 Google post called it one of the major AI systems powering Search. ### Is RankBrain the third most important ranking factor? That claim comes from a single 2015 interview reported by Bloomberg. Google has not repeated it, so it should not be treated as a current fact. ### Can I optimise for RankBrain? Not directly. Google has said there is nothing special to do; the useful work is making sure each page covers the concept behind the searches it appears for. Hub: [Topical Content](/topics/topical-content.md)