The Google algorithm is being continually tweaked so that it can be improved because they recognize that it only takes a small improvement for a competitor to take over their entire market. So far, Google has been able to stay ahead because they consistently serve the best results and Rank Brain is just one component that helps them to achieve this goal.
Although they release multiple updates each week, huge updates like Penguin, Panda and now Rank Brain, only occur once every few years and they change the way that we search and how sites are ranked. In the case of Rank Brain, it’s estimated to impact up to 15% of all searches, which means that in all likelihood it affects your traffic.
What is Rank Brain?
Rank Brain was introduced as early as 2015 and has been continually tweaked as it has been running. Rank Brain is an advanced machine learning algorithm that has been labeled by some as artificial intelligence. Machine learning takes results and figures out which is best, before continually testing without the need for human intervention, allowing for autonomous improvement.
Its job is to figure out what people are looking for when they search and then to match those results up with the correct queries so that everybody gets what they are looking for as quickly as possible. In the age of mobile and voice-activated assistants, simple and quick results are more important than ever, hence the Hummingbird and now Rank Brain updates.
How Does Rank Brain Work and Impact Results?
While there have been previous updates aimed to adjust the way that Google understands a searchers intent, up until now these have been manually coded. Rank Brain, on the other hand, uses machine learning techniques so that once it is programmed, it can run by itself and continually improve both itself and the overall algorithm.
The problem that Google faces is that an estimated 15% of all searches have never been entered into their system before and that makes it hard for them to serve the correct results. Rank Brain exists to try and decipher what is meant by these new searches by seeing patterns and connections between seemingly unconnected searches to understand how they are similar.
The key to this to look at the meaning behind the search via the use of natural language processing which looks at a variety of factors, not just the individual words that are typed. Once the algorithm can do this, it can start to link searches together to recognize that although they are different queries, they want the same results.
While a human undertaking this task would be quickly overwhelmed by the billions of searches each day, the machine learning algorithm takes it in its stride. Each day it becomes smarter because once it has linked multiple queries together, it’s far easier to add to the same ‘group’ of searches.
Things, not Strings
In an article outlining the Rank Brain update, they used the phrase ‘things not strings’ to explain how the algorithm tries to look for the thing you are searching for not the string you used. One example of this could be the search “how old is the wife of President Obama,” rather than looking for that string it understands that you want the age of Michelle Obama.
By looking for things, not strings, the algorithm can serve the precise result rather than bringing back pages that use those words in that order, which are likely to be less relevant. Plus, once Rank Brain knows that this what you were looking for it can improve the algorithm so that next time somebody searches for something related to his wife they can return better results.
Does Rank Brain Remove the Need for Keyword Research?
With all of this in mind, it’s easy to jump to the conclusion that Google has removed the need for keyword research. After all, Google isn’t using keywords to return these results, right? Well, that’s not precisely the case.
Rank Brain currently seems only to be used for queries that have never been searched for, and therefore for which Google has little to no data. For common queries, Rank Brain isn’t used, and in this case, they rely more heavily on page content. But even for new searches Rank Brain needs to read the keywords on a page to figure out which results to return.
Rank Brain isn’t specifically about keywords, it’s about better to understand search queries, and while this might have a slight impact on the importance of the on-page use of keywords, it certainly doesn’t remove the need for keyword research.