In October 2015, Google launches what is now known to be the first use of artificial intelligence in its ranking algorithm.

While it has been more than half a year since its release, very few information about it has been made official by Google.

A lot of people were left in the dark. What it actually is? What kind of signals does RankBrain track in order to rank websites?

Does the AI automate everything Google does and now Google engineers are left with nothing to do?

There’s a lot of myths going around so let’s find out and separate the facts from the fiction.

So What is RankBrain?

Simply put, RankBrain is an algorithm that helps Google break down and understand complex search queries.

You can say that RankBrain translates search queries from “human” language into “searchese” or language that can be easily be processed by Google.

For example:
1 – Original query = “New York Times Puzzle”
2. Translated query = “Puzzle?Crossword (New York Times)”

If you notice, in the translated query, the keyword “Crossword” is added into the mix. This is the brilliance of the AI part of RankBrain.

RankBrain utilizes artificial intelligence to find the right meaning and context behind ambiguous keywords.

In this example, RankBrain guesses that the keyword “puzzle” in this context means a crossword puzzle.

This conclusion is made by understanding that “New York Times” is a newspaper that’s popular with its crossword puzzles.

Which brings me to my next point. How does RankBrain “understand” this?

Another great feature of RankBrain is that it uses deep learning process in order to synthesize new information out of existing ones and find similarities in differently worded queries.

And RankBrain does all of this without human interference. How awesome is that?

Why Was RankBrain Created?

In order to further understand RankBrain’s purpose I think it’s a good thing to take a look at the issues the algorithm is designed to address:

• Handling complex searches – Everyday Google deals with long and complex new searches (15% of all search) that haven’t been seen before. Lacking previous history on handling these new searches, Google needs a way to relate them to previous searches to deliver the most accurate results.
• Handling ambiguity – Searches involves a lot of ambiguous keywords that can mean anything based on different context. Having a self-learning AI to process existing search histories allows Google to guess what a keyword mean more effectively.
• Improving accuracy – Moving from keyword matching to context understanding allows more accuracy in handling variants of the same keyword (singular vs plural, abbreviations, misspellings etc)

By understanding the main concern RankBrain is trying to solve, we can rule out the more outlandish speculations of what the AI system does.

How Does RankBrain Work?

You might be wondering where does RankBrain fits in Google’s grand scheme. Is it a ranking algorithm? Does it penalize bad results or something?

Not in my understanding.

I’ve found a great illustration by Damian Fanish from Quora that help explain it easily.

RankBrain works in conjunction with Google’s other ranking algorithms.

Like previously mentioned, RankBrain refines search queries and filters the search results. The filtered results are then ranked by Google’s ranking algorithm as usual and then presented to the users as a search engine results page.

As you can see, RankBrain does not replace the existing algorithms but feeds them a better filtered queries.

How Will RankBrain Affect Us?

For us marketers and SEOs, the answer is surprisingly, not much. Think of it as an optimization effort by Google in improving their processing.

That being said there are still quite a few things that RankBrain bring to the table:

• Increased tolerance for long-tail keywords – Previously, users need to “play around” with the keywords they use in searches in order to get the right result. By processing long, complex searches, RankBrain allows users to type in queries in a more natural way.
• Supporting platform for voice search – Speaking about queries in a natural way, we can’t avoid from talking about voice search. Voice search technologies like Siri, Google Now and Cortana is experiencing more and more usage. Unintentional or not, RankBrain plays a vital role in helping voice search assistants return a more accurate result out of voice searches.
• Less emphasis on long-tail keywords – Here’s one thing that make things easier for marketers. Now that we know RankBrain handles long queries and break them down to more commonly used terms, we no longer have to emphasize a lot on optimizing for long-tail keywords.
• More flexibility in keyword usage – This also opens up more room for flexibility in keyword usage. Instead of spamming the same exact match keyword, we are free to use more LSI keywords, synonyms and alternate terms to improve readership.

Will RankBrain Kill SEO?

As usual, as Google rolls out new algorithms, it’s unavoidable to get people clamoring SEO is dead yet again. Doesn’t help that this time around it’s about an AI system hell bent on taking over the world.

The answer of course, it won’t. RankBrain might change a little bit in how we do keyword research and implement them in our contents but that’s mostly it.

In another word, SEO will become even more sophisticated and we have to rely a lot more on quality rather than technical SEO.

Addressing things like semantic web, LSI keywords and content readability becomes more important than ever. Competition will become even harsher as basic SEO just won’t cut it anymore.

The Future Of RankBrain

And as usual, nothing is certain especially in fast evolving industries like SEO. For now RankBrain’s application is quite minimal to put it in simple term.

In the future, it’s very likely that RankBrain will be used in a wider application than just interpreting search queries.

Should the AI is advanced enough to actually understand all contents online, then it will be a huge breakthrough in the industry. So huge it might make a lot more technical stuff like tags and metadata obsolete.

To close it all, I’d like to make it clear that most of the information in this post is made after analyzing Google’s official announcements, a patent that’s related to RankBrain and what other experts thought about it.

Given what little is officially stated by Google and piecing up pieces of the puzzle, I might need to emphasize that this speculation of mine might not be entirely accurate.

But should more information become available I will sure to update this post to reflect any new discoveries made. Let me know in the comment section whether you love or hate RankBrain!

The Google Survival Guide:
Complete Protection From Potential Algorithm Updates

• 5 Critical Lessons To Prepare Your Website
• Patent-sourced Prediction On How Google Will Change In The Future
• Tips To Avoid The Worst-case Scenario
• Instantly Actionable Steps To Follow

Updated: 30 October 2024

About Allysa

Allysa is always eager to learn something new whenever she got some free time and that includes SEO and inbound marketing. She also has a passion for traveling and discovering the unknowns.