What is RankBrain?

What is RankBrain? (And how can it help you Rank #1)

From the very beginning, Google has been in a quest for improving search. Before there was Google, search was provided by companies that indexed the internet and provided a list of sites and pages which may or may not have been what the user wanted. Google’s page rank provided users with relevant results. It was able to measure a website’s popularity with valid criteria. Google further solidified their position by providing results which fit the user’s needs.

One of the reasons behind Google’s success is it uses both on-page and off-page signals to rank a website. Additionally, Google has kept on refining its search algorithms to include various methods and signals. They have been experimenting with what works, using real-time data to validate their research.

The research and development of the search algorithm is a continuous process. It takes into account various factors or signals which affect the search results. Google aims to provide results that are relevant to the user. For instance, when a user searches for “pepperoni pizza” the results would show nearby pizza parlors, and not include pizza parlors in other states or countries. For more relevance, additional information is provided, like address, phone number, a map, customer reviews, and others.

Google incorporates these new algorithm revisions on an almost continuous basis. However, with all of the data that Google collects, the level of complexity increases. Every day, 15% of all searches are for new terms. These new searches are not yet in the company’s search index. With the growing number of searches, the volume of new terms also increases. The search giant has decided that to keep in step with the growth, it has to make use of automated resources.

Google is well-known for its use of machine learning and AI. It is this expertise that is behind their development of AI to develop and improve new search algorithms. Released in 2015, it is called RankBrain.

RankBrain Overview

RankBrain is considered as the third most important ranking signal after content and links. It is also an algorithm and AI. More importantly, it is not an on-page or off-page signal.

RankBrain is considered as the third most important ranking signal after content and links. It is also an algorithm and AI. More importantly, it is not an on-page or off-page signal.RankBrain is an AI algorithm that reads through big data, improves on the search algorithm and tests these improvements on live real-time searches. It uses indicators and analytics to validate the algorithms is introduced. If the indicators are positive, these new codes stay in the algorithm, otherwise, they are removed from the live search programs.

Before RankBrain, all Google algorithms were coded and tested by hand. However, manual coding could not be continued over the long term because of the continuing increase in websites and searches. Between January 2008 and January 2018, the number of websites increased from 185 million to 1.85 billion, a growth of more than 10 times. With the number of websites growing faster manually coding and testing search algorithms is going to be a strain on the company and Google decided on a different approach to the problem.

The process of including algorithms is straightforward, repeatable and modular. It also includes studying user behavior from millions of data points that come in every day. The total number of daily searches worldwide is estimated to be between 7 to 10 billion search queries. For different cases the user behavior is consistent. The user enters a search term or keyword and then goes through the individual search results one by one. If the website is not what the user wants, he goes back to the results page, and clicks on the next one on the list, and so on. For Google, the aim is to provide the right result without the need of going back to the SERP. This is the expected behavior for a successful search.

Machine Learning, AI and Google Rankbrain

For the results page to fit the search request, Google has to understand what the search keyword meant. This involves a better understanding of the intentions as well as the context of the search. To do this, Google relies on RankBrain to learn the nuances of the search keyword.

Searches keywords change as the users become more familiar with how to use the search engines. In so doing, more users search for keywords that are no longer exact matches. Before, a search like “black car safe gun” would result in exact matches for “black car”, “black car safe”, “car safe gun”, “car safe”, “safe gun”, and others. The old Google search would match the keyword with those it finds in web pages. It would not look for context, and the user would have to go through a lot of the search results before it finds a page that suits his needs.

Today’s searches also rely on context and association. A search for a “gun safe” would also lead to “gun racks” and other related keywords. When there is a new keyword, instead of parsing the keyword and searching for individual words, the algorithm would look for look-alike or related items that it has already searched for.

The power of RankBrain is that it can be used on big data and it can pull out information, relationships, and events. It can develop these relationships because the words or terms become associated with each other. Things like “hammer” usually appear close to “carpentry” and “nails.” RankBrain uses these types of relationships to extend what it knows and projects into those items which are not known. New searches are parsed according to their association or proximity to other words and terms in web pages. RankBrain would discover that in a significant number of instances when a hammer is mentioned in a web page not only do nails and carpentry are also mentioned, but others like “do-it-yourself”, “DIY projects”, and others are not far behind.

The word association also extends to not only proximity but also relationships between words and their functionality. This means that it learns that “hammer” is to “nails” as “screwdriver” is to “screws.” Following the analogy, other significant relationships can also be pulled out, in more complex relationships.

The above association and understanding of concepts are used to reach the end of the output, that of satisfying the user. RankBrain understands that it has found a match when the user stops using the SERP and starts reading the web page. The longer the user is reading the website, the better the match appears. If the page association works, the page’s rank increases for that keyword.

The Top SEO Ranking Factors

SEO marketers and researchers have been studying ranking factors to determine what they can do to improve their website search results ranks. Part of their research includes experimenting with their web pages, content, backlinks, promotions and keywords. They also sift through every press release and document from Google regarding their search algorithms, as well as interviews, tweets and blog posts by prominent Google personnel. According to various researchers, Google uses more than 200 factors in ranking a web page for a search.

Among all the ranking signals, Google recently announced that RankBrain was the third most important. This is a significant announcement in more ways than one because RankBrain was only introduced in 2015. It is one of the latest ranking factors used by Google. Marketers did not even consider RankBrain, or the way it works, in the years before it was introduced.

At the top of the ranking factor, the list is content. Content is king. People search the internet for content. It only makes sense that the quality of the content should be at the top of the list. The second most important is links. Links show Google what websites and pages are credible and are worth reading. RankBrain, the third most important ranking factor connects content and links and helps verify the level of authority of the website, based on the audience reaction.

The Biggest change in SEO Algorithm since Penguin

Before RankBrain, Google’s search engine algorithm was based on on-page and off-page ranking factors. RankBrain is an algorithm and is neither an on-page nor an off-page ranking factor. Instead, it helps create a correlation between content, links and user satisfaction. With RankBrain, online marketers have to study how their audience and visitors would react to their content. Online marketers then have to adapt their approach to meet their visitor needs.

On-page and off-page factors are things that a website administrator or owner can directly manipulate or help along. On-page factors include the content, keywords, headers, images, alt-text, outgoing links, and loading time, among others.

Off-page factors include backlinks, social media mentions, likes, and shares. A lively discussion about the webpage invites more attention online and counts towards social media factors.

Before the introduction of RankBrain, the main indicator of a page’s content was the number of links and visitors. RankBrain has gone deeper by analyzing visitor behavior. The tools are already there on the website with Google Analytics.

Google Analytics is a tool that gathers data about the traffic to and from your website. You have to include a snippet of code on your webpage for it to work. This serves as a tracking device to determine visitor behavior. Among other things, it knows what page or website referred to the visitor. For RankBrain, the visitors from Google search are included in its ranking algorithm.

Another important improvement that RankBrain brings to search results is “intent”. It tries to understand what the user wants. There are many ways that RankBrain does this. For services or establishments, it finds those results which are located near the user. A search for “pineapple pizza” would show restaurants and fast-food joints that offer pizza with pineapple, as well as recipes for “Hawaiian pizza”. With the use of big data on the user, RankBrain relates the current search with a recent history of articles read and searches. When a person views videos about mixing cocktails, searches for “alcohol from sugar cane” the search would show pages about rum, and other alcoholic beverages, instead of alcohol production for fuel.

RankBrain goes beyond the most popular search results according to on-page and off-page SEO efforts. It tries to provide more useful and focused information on the search.

UX Signals used by RankBrain

RankBrain upvotes a particular page for a specific keyword. It also downvotes a page when users don’t like it. To decide on what to do with a website or page, RankBrain seems to be using the following signals:

  • Click-Through-Rate (CRT). This consists of the organic click-throughs from the SERP to the website.
  • Dwell Time. This is a measure of the length of time the user browsed the website. The longer that the user stays on the website, the higher the rank it gets.
  • Bounce Rate. The bounce rate is the opposite of Dwell Time. If the user spends less than 2 seconds on the website before returning to the SERP, then it is penalized for the keyword.
  • Pogo-sticking. This is when the user clicks on a SERP, returns to the SERP almost immediately, and then clicks on the next search result.

Intuitively, it is possible to know if a user likes a page suggestion from the SERP or not. If the user reads the page and stops going back to the SERP, then the page was useful. That’s the reason for the above signals. These are indicators of satisfaction of the SERP entries.

The above UX signals follow a flow of action that shows interest in the page. The user clicks on a link on the SERP, sending him to the page. If the content fits his requirements, he stays on the page and reads through it and the page gets upvoted. If the user doesn’t think the page is what he was looking for, he goes back to the SERP, leaving the page which gets downvoted by Google. The user repeats the process clicking on a new link to see if it fits his search.

Making use of RankBrain to your advantage

Google has mentioned that RankBrain is the third most important ranking signal. However, since it is more of an algorithm and is not an on-page signal, it needs some planning to use. It measures the behavior of users when they search for a keyword. The longer that the user stays on the page after a search, the more RankBrain upvotes are given to the page.

Going one step further, the page has to be informative, authoritative and comprehensive. To do that, it has to be complete, original and unique. It has to cite other websites for the information it contains. As an authority, other websites should also link to it as reference material. Due to its comprehensive discussion of the topic, it will necessarily be a long article. These are some considerations when writing page content, or a blog entry.

Google recommends that websites become authorities and avoid spammy content. The above strategy fits perfectly with Google’s concepts. To take full advantage of RankBrain requires some changes in SEO strategy.

Use Medium Tail Keywords

One on-page strategy which was successful before was the use of long-tail keywords. With RankBrain, this has become obsolete.

Formerly, websites were filled with articles optimized for long-tail keywords. These were exact strings of words which were variations of the main keyword. These fitted exactly with the words used in search and each long-tail keyword had its article. The problem was that the long-tail keywords looked like one another, or they used the same words but arranged differently. For example, one article would be optimized for “GPS tracking devices for cars” while another would use “small vehicle GPS trackers”. Searching for both these keywords would show almost the same list of links.

Long-tail keywords intend to cover all the bases. It tries to make sure that in whatever way the keyword is entered on Google search, that the website would rank high on SERP. RankBrain understands that the long-tail keywords are the same and disregards the different permutations. This results in long-tail keywords being relegated to a small niche of results that few users even search for.

Using medium-tail keywords offers several advantages. It substitutes for the marketer’s use of long-tail keywords. It covers all the bases which permutations of long-tail keywords were supposed to cover. It also has more searches, but not a significantly larger competition.

Using medium-tail keywords that have a relatively large volume of searches with a small competition is a big advantage for a small website. If there are only a hundred websites that rank for the keyword, then it is easier to rank for that keyword. Since there are more than a thousand daily searches, then the sites on the front page can easily garner more than a hundred visitors a day from search. This strategy has been successfully used by online marketers before. But by optimizing RankBrain, the results can be better than they were before.

With only a few competitors, it is easy to be an authority site. Additionally, it would be easy to write about in-depth articles for these niches.

Boost your Dwell Time

The dwell time is the recorded period from the moment the website visitor goes to your page and the time he leaves. If the visitor was looking for something, specifically if the visitor clicks on a link from the Google search results. RankBrain correlates the dwell time for a given search result. The longer that a visitor stays on the site, the higher the page would rank for the keyword. To improve rankings it becomes necessary to keep the visitors interested in the contents of the page.

There are several ways to improve dwell time. The easiest way is to make interesting, relevant and highly informational content. The article should include as much information about the keyword. Also, the article should lengthy but not fluffy, or lacking in depth. Articles have grown longer since Google started emphasizing content. Whereas before, articles could be 200 to 300 words in length, subsequent Google algorithm rollouts suggest article lengths of more than 500 words. For more meat in the content, the articles may need to be more than 1,500 or 2,000 words in length. With more information in an engaging article, the visitor would stay longer to read the page. An offshoot of this is that the visitor would also be more likely to read the other pages on the website.

Improve CTR with Improved Brand Awareness

There is something to be said about becoming an authority. Once a website has established itself as an authority, it becomes easier to maintain the level of credibility not just for visitors but also for Google. Website owners can build on this authority and improve on it by keeping themselves visible. This can be done not only by posting more articles relevant to their niche or keywords but also by using social media to extend their reach.

Word of mouth, or in this case, a social media post which turns viral, has a heavyweight with RankBrain. A known-brand can generate more social media links, and at the same time, improve on the SERP ranking.

Re-use and Re-imagine Poor Performing Pages

Website owners have always had problems with poorly-performing pages. They cannot delete these pages because the website will lose volume. Also, removing the pages would lower the website’s index for its keywords. With no choice but to keep these pages, there is another alternative, which is to re-use and re-imagine these pages.

Old web pages usually have a short content. It is possible to re-use these short content pages, merge and rewrite them into a longer article. The article would still use the same keywords, but it would have a new title. Necessarily, the article would also have new images, alt-text, layout, and keywords. Also, the article would be submitted as a new article, which adds to the page indexing.


In hindsight, Google has been working towards RankBrain. It is a logical evolution or direction for search. Google started with a page rank algorithm trying to sort out websites based on what users wanted to find. They have used on-page factors used by website owners to determine if these are the sites which are users would want to read. Google has used backlinks to determine if the popularity of sites based on the volume of sharing and citations. RankBrain now adds the user satisfaction and experience to weigh the authority and credibility of the site.

RankBrain has been successful in providing users the information that they need. There are other ways to optimize a page to fit RankBrain. What is important is that the improvements and overall strategy take into account the nature of the visitor’s search, and to provide the information required.