Google Ups the Ante – Say Hello to RankBrain

Home / Google / Google Ups the Ante – Say Hello to RankBrain

I’ve never been so excited about change at Google as I am today!  While many search engine optimization firm owners are probably losing sleep right now I am not.

News broke yesterday of a mysterious new ranking factor called “RankBrain”.  I did some more digging and found out a whole bunch of stuff about RankBrain that I’m going to share with you here.

First, a bit of background.

Google has struggled for a long time in being able to try and interpret what some searches mean.  For example, if you were to search for “who is the president’s wife” Google may have, in the past, been unable to correctly return the result because there was no direct correlation between president = Barack Obama therefore president’s wife = Michelle Obama.  However, over the past few years Google has learned how to program their algorithms to make some “guesses” based on relationships.

These guesses, however, were still influenced by humans building relationships that the algorithms could learn. In the above example, Google’s algorithms did not know the relationship between president’s wife = Michelle Obama until a human programmed it.

RankBrain moves Google’s abilities in new directions.  This new system can learn and make guesses based on relationships it has learned about in the past.   In fact it’s gotten so good at matching queries to results that in just a few short months it’s moved up to become the third most important ranking factor in Google!

So what are the top two?  Well Google won’t say, but I have my own opinion on that which I will share later.

But just what is RankBrain?  Is it a replacement for what Google has been doing?

The answer, I believe, is no.  It’s not a replacement – it is an addition.  Hummingbird, which rolled out in 2013, was Google’s revamp of its core ranking algorithms.  When you talk about how results are displayed on a search results page, you are talking about how Hummingbird displays results on a search results page.

Hummingbird is comprised of many different “sub-algorithms” if you can call them that.  Panda, Penguin, Pigeon, Mobile and more are all different parts of the main hummingbird algorithm designed to do things such as fight spam, filter out low quality pages, and return mobile friendly results.  There are others as well, but most people are aware of these ones.

But within these and other parts of hummingbird are also ranking factors – things links text on the page, anchor text, links, domain age and authority, PageRank and so on.  All these factors, combined with those algorithms form Hummingbird.

RankBrain is part of this second group – the ranking factors, or signals.  And, as mentioned above, it has gotten so smart since the beginning of this year that it’s now the third most important factor.

How Does RankBrain Work?

Up until RankBrain, relationships between terms had to be manually programmed in.  There was no way for the system to understand that Apple could be both a company and a fruit.

RankBrain is now able to intuit relationships based on previous relationships it is aware of.  So, as in the example above, it knows that Barack Obama is the President of the United States and his wife is Michelle, therefore it can intuit that Michelle Obama is the wife of the President of the United States without having actual knowledge of this fact.

But it goes beyond that.  It can learn and guess based on past relationships between words or phrases it already knows about.

For example, this blog post posted on the Google Open Source blog back in 2013, suggests that the system “can learn the concept of capital cities, just by reading lots of news articles — with no human supervision”.

Think about that for a second.  A system that can form relationships where none previously existed within it’s current knowledgebase.

Why RankBrain?

I think the reason for this is because so many Google searches are for long tail searches, or searches that Google has never seen before.  Approximately ½ a billion searches per day in Google are for phrases it’s never seen.  So in order to improve search results Google has had to have human intervention to help the ranking algorithm improve those results.

RankBrain is able to remove that human intervention and determine results on its own by analyzing previous similar queries and then return appropriate results based on that analysis.  In other words, a system that can learn with little to no human intervention.

It must be incredibly accurate to become the third most important ranking factor in just a few months.  In fact, according to Bloomberg, Google asked the same engineers who build the algorithms to analyze results, as well as also got RankBrain to analyze the same results and found that “while the humans guessed correctly 70 percent of the time, RankBrain had an 80 percent success rate.”

Where Do We Go From Here?

Because it is so new and no one has really done an analysis of the system to find out what kind of “factor” it is, it is hard to say what impact, if any, this will have on SEO.

Personally I don’t see a huge impact on SEO as an industry.  In fact, I think it could help because now we will have a much better chance of ranking for those long-tail searches than before.

Why do I think that?

Remember when I said that RankBrain was the third most important ranking factor?  Well I still believe that links and content are number 1 and 2.  Therefore the sites with the most quality links and best quality content will generally rank at the top of the search results.

Therefore, those same links and content will help with those complex long-tail searches as well because they are still ranking factor #1 and ranking factor #2.

Plus, from my understanding RankBrain is, for now, targeting those long tail searches.  Since we SEO’s play in the high search volume market RankBrain likely doesn’t kick in and, therefore, doesn’t impact rankings.

Of course as time goes on I’m sure RankBrain will expand its foothold – either impacting a broader swath of phrases and not just long tail, and could even move into other areas such as knowledge graph improvements.

We are still months or years from that but I do think RankBrain will revolutionize not only the way we search, but the way Google handles those searches.

Are you concerned how RankBrain will affect your current search engine marketing?  If so click here to contact us and receive a free, no obligation analysis

Related Posts