rankbrain

Google RankBrain – Using AI To Improve Search Technology

If you haven’t heard of RankBrain before now then here is a brief insight as to what it is and how it works.

RankBrain itself is a machine-learning system which is based on artificial intelligence and it is currently being used by Google to help refine the search results delivered to end users.

With machine learning, the computer program that processes search queries actually teaches itself how to improve upon the results it provides in relation to what a user is actually looking for based on the search terms they enter.

However, although RankBrain is starting to play a main role in the way that Google is improving its online search services – it is only one piece of the puzzle as the main search algorithm uses a number of different processes to find the most relevant pages according to a user’s needs.

Up until 2013, the main search algorithm used by Google didn’t even have a name – but since then Google have decided to call it Hummingbird. As well as RankBrain, the main Hummingbird algorithm also contains a number of other algorithms which have been designed to filter and improve search results in different ways.

Here is a brief list of the ones we know about and what they do:

  • Pirate – an algorithm designed to remove pages which infringe on copyright from the search results.
  • Panda, Payday and Penguin – these remove pages which are considered to contain spam.
  • PageRank – which improves a pages position based on quality external links.
  • Pigeon – designed to improve local search based on a user’s whereabouts.
  • Top Heavy – included in order to remove or demote pages with too many ads.
  • Mobile Friendly – which improves the position of mobile-friendly sites.

In addition to algorithms, Google also uses “signals” as a means of ranking the pages of a website more effectively and this is where things get somewhat complicated!

At the time of writing, there are over 200 major signals that are currently used to determine how well a page will rank for a particular search term and out of all these signals – RankBrain has become the third most important.

However, up until a few months ago Google kept everyone guessing about the first two signals although they have recently been revealed to be high quality content and good quality links.

So , How Does RankBrain Improve Search Results?

how RankBrain affects search resultsRankBrain makes use of artificial intelligence, or machine learning, as a means of finding relevant content for a user who has submitted a search query that doesn’t quite match what they are actually looking for – based on the wording used in the page content. It is a means of connecting the dots, making an intelligent guess and then delivering the right page where long term search queries have been entered which Google hasn’t seen before.

It is estimated that Google receives almost half a billion queries each day that are completely new to Hummingbird and RankBrain was designed to make sense of these searches by trying to work out what a user is interested in whilst providing them with the most appropriate pages based on complex patterns of word association.

How Was RankBrain Rolled Out?

Google RankBrrain roll-out - How did it happen?Up until October 2015, RankBrain was only used in order to process search queries that Google had never previously seen – which is around 15% of the total amount of daily searches made. Move forward to June this year (2016) and RankBrain is now used for every search query Google handles.
However, although it is now public knowledge that RankBrain is the third most important ranking signal used in order to classify a page based on its content, it is not clear how or if it is used to assess the quality of a given page. Of course, the main purpose of RankBrain is to deliver search results which are relevant to a user – so as long as you provide relevant content that is well-written, fresh and informative, this new ranking signal should have no negative influence on your search positions whatsoever.