Skip to content

The Artificial Intelligence Changing the Face of Google Search

About 15% of the queries Google process every day are new. That means that nobody has ever searched them before. That’s a lot of brand new data! So how does RankBrain work and how does this new data get processed?

Despite going live in April 2015, the Google RankBrain algorithm wasn’t introduced to the world until October later that year. The announcement of RankBrain’s release was made via a Bloomberg news story:

“RankBrain uses artificial intelligence to embed vast amounts of written language into mathematical entities – called vectors – that the computer can understand. If RankBrain sees a word or phrase it isn’t familiar with, the machine can make a guess as to what words or phrases might have a similar meaning and filter the result accordingly, making it more effective at handling never-before-seen search queries.”

The objective of RankBrain is to help Google provide more accurate results and an overall better search experience for users, helping to satisfy their search queries.

What is RankBrain?

Rank Brain is an artificial intelligence (AI) programme that is used to help Google process search queries. It embeds significant amounts of written language into mathematical entities (otherwise referred to as vectors), that the computer can understand.

If RankBrain comes across a word it doesn’t recognise, it can make a guess as to words that have a similar meaning, filtering those results accordingly. Quite simply, RankBrain is capable of learning a significant amount, recognising any new patterns, and then revising the search engine results page based on its new-found knowledge. RankBrain is the only live Artificial Intelligence that Google uses in its search results and it is helping improve how we do SEO in so many ways!

How Does RankBrain Work? In Simple Terms!

Without confusing you with all the technical language, RankBrain works like this: words are typed into Google and RankBrain assigns these mathematical addresses, retrieving words based on your query that are the ‘best fit’. If RankBrain does not recognise a query, it finds the next closest thing.

Over time, RankBrain refines its results based on the interaction of users and matches the search user intent with the related results. Just as a human would process information it is given, RankBrain does the same!

Is RankBrain Important?


RankBrain used to only be responsible for 15% of all online searches but has since been involved in almost all queries entered into Google, becoming one of the most important signals used to determine search engine rankings.

How Does RankBrain Affect Your SEO?

The first step of effective SEO is defining your keywords. This means searching for keywords relevant to your business. It is this searching for relevant keywords that RankBrain will most affect as it is constantly learning and changing. As a result, your SEO campaign is more adaptable to changes in the search engine rankings landscape.

RankBrain significantly affects the Google search results, discovering the content that is most relevant and pushing it to the top. And this is something all search marketers should be thinking about because it can have quite an impact.

What’s more, because of the intricate nature of RankBrain and all its collated information, SEO strategies must be constantly improving. If you were writing an article about Mercury (the planet) and you flooded your article with the word “mercury” without including the word “planet”, your article may not rank well. With RankBrain it’s not just about keywords, you must also include other relevant words around that.

That’s why at Bamboo Nine, if we’re writing an article on blocked drains, we work with various alternatives of that keyword to help it rank. For example: blocked drain, blocked drains, drain blockage, blocked, blockages, drains, etc., This gives us the chance to rank highly via multiple search terms.

What is the Hummingbird Algorithm and How Does It Relate to RankBrain?

From the moment Hummingbird was launched by Google in 2013 it has had a significant impact on the way searches are made. Before the launch of Hummingbird, Google search queries were often strings of words and results that may not have been relevant to the initial query. An easy example is; when users searched the word “apple”, Google may have responded to their queries with a selection of fruit, rather than the product. Hummingbird transformed these irrelevant results in 3 significant ways:

  • It improved Google’s ability to interpret conversational speech, meaning that users could type questions and Google could understand it in ways it never could before.
  • It looked at the intentions behind a search, so that even if you searched an incorrect word, Hummingbird could guess what you mean – delivering relevant results!
  • It uses location targeting, allowing the delivery of relevant results local to the user. For example: if you were to search for a French restaurant and you lived in Brighton, instead of providing restaurants options in Paris, Hummingbird would give you results of French restaurants in your city.

As you can probably tell, the launch of Hummingbird significantly changed some of the ways Google worked – improving many search engine results.
However, it still its flaws.

And That’s Where RankBrain Comes In!

One of the main jobs RankBrain has is to interpret queries that it does not understand – something Hummingbird could not do. It’s constantly updating, gathering more information every second so that it can help provide relevant search engine results for users.

For example: with Hummingbird, a user may have typed in “apple” and been presented with the fruit because Hummingbird may not have recognised “apple” is also a product.

RankBrain does understand this – even in the absence of context!

With RankBrain, you can type in the word “apple” and it produces all relevant results by taking into account your previous search queries and the most popular results of various users with the same query. If RankBrain is not confronted with the context of a search query, it may deliver a mixture of relevant results (so, for the word “apple” it may deliver both the fruit and the technology to be in with a higher chance of providing what’s relevant for the user).

Basically, as RankBrain adds to its knowledge and understanding, its ability to deliver relevant results to users improves.

How Does RankBrain Learn?

I’m going to explain this using an example. Say a user types “pink apple” into Google. Without any context, RankBrain will provide multiple results for the user; the fruit, the technology, and the perfume.

RankBrain then learns about “pink apple” as several other users have been clicking on links about iPhones. RankBrain takes this information and decides that people searching for “pink apple” are more likely to be looking for the iPhone than the fruit.

Over time the results begin to change as RankBrain learns what’s relevant to users and what isn’t. And while there still may be a mix of results, there will be numerous results (such as iPhones) that dominate the page. Eventually, it may only display the results that are relevant to you.

The Link Between Visitor Behaviour Online and RankBrain

RankBrain looks at visitor behaviour on a webpage, whether it’s crawlable and whether they receive the information they need. If they do get everything they need from your article, RankBrain ranks your article highly because it is providing a useful service to its users. So, many people have stated that RankBrain takes into account the users behaviour on a site/or webpage, ranking your content for its usefulness to the individual.

So, if you write an article on how to unblock a toilet and a user lands on your page and is presented with a solution, RankBrain will rank your article highly. What’s more, if that article has a link to professional toilet unblocking services, it encourages crawlability and conversions – encouraging RankBrain to boost your content up the rankings because it is useful, provides a solution and is encouraging customer conversions.

The Final Thoughts We Have on RankBrain…For Now

Through their own experiments, Google discovered that RankBrain was 10% more accurate than Google’s own engineers when it came to predicting the best page results. RankBrain does an exceptional job of interpreting what users type in (through analysing their search history and the patterns of previous users) and providing the most relevant results. RankBrain is an incredibly powerful algorithm that is constantly learning and developing to provide Google searches with the best results.