Google RankBrain has had a massive impact on how search queries are analyzed.
But what exactly is it and how does it work?
We’re always looking for new ways to optimize our content and rank higher on Google. Knowing the basics of RankBrain and how it handles search queries can help better understand how to optimize your content.
What is RankBrain?
There are many parts to the algorithm, some of which Google has announced (like PageRank, which assigns pages with ranking “scores”). RankBrain is the part of Google’s algorithm that focuses on natural language processing. It combines machine learning and AI to understand the intent of a user’s search term.
When RankBrain was first released, it handled about 15% of all searches. It was designed to handle longer search terms (or longtail keywords) and searches that Google had never seen before. Now, RankBrain is used to analyze almost every search.
In short, RankBrain is a machine learning system that determines the intent of search queries to display only the most relevant results.
How Does RankBrain Work?
While we’ve been given an idea of how RankBrain works, no one knows exactly how it handles search terms from start to finish. With a system this complex, there are many well-kept secrets that happen behind the scenes.
Here’s what we do know:
The old algorithm used to simply see search terms as words and characters. This severely limited how well it could understand the context and display the most relevant results.
That all changed with the introduction of RankBrain. Rather than looking at strings of characters, it sees each search term as an object, or entity. (Want to see how that looks? Try it out for yourself with Google’s Natural Language API demo.)
So when someone searches a term or phrase, RankBrain extracts “entities” from the text and determines what each entity is (person, place, organization, consumer product, etc.). Each entity has a unique ID.
RankBrain also takes other factors into account, when available—particularly things like location, date, and other information that the searcher may not intend for it to use. (Hence, searching incognito can be pretty useful if you don’t want it using location and pre-existing search data.)
So how does a simplified version of this process look?
- Google analyzes a query and breaks it down into entities.
- It determines how relevant each entity is to the overall search term, as well as how they’re related.
- It tries to determine the intent in order to more accurately rank content. If it can’t, it may pull other data available to it, such as location data and pre-existing search data, and uses this to display the most relevant search results.
Obviously, this is really simplified and there’s a lot more going on behind the scenes. However, this gives us an idea of how RankBrain analyzes search terms and goes beyond trying to match strings of characters to an “exact match.”
How to Optimize for RankBrain
According to Google, you don’t.
Their answer is typically the same—by optimizing for “users,” you’re also optimizing for the search engines since the end goal is always to “improve user experience.”
But we know that doesn’t always pan out and until some crazy advancements are made with Google’s algorithm and it can flawlessly understand search intent, there has to be some balance.
In the meantime, we just have to work on trying to understand how Google understands search terms. (Which can be incredibly confusing sometimes.)
Here are some things that can help.
Review Search Results
While people might use terms like “topic” and “niche” interchangeably, Google clearly sees them as different things and doesn’t completely understand how people use the terms. So while you can use these in similar ways, it might affect which term Google most strongly associates with your content.
It’s important to understand what search engines see as closely related words and use variations of those to help strengthen and focus your content. (That’s not to say you should turn to keyword stuffing—but that you should avoid lots of instances of language and terms that could be confusing to search engines.)
You can get an idea of what it considers “related” by taking a look at the search results. Similar words appear in bold, even if they aren’t in your query.
In this example, the query was “bold text,” but you can see that “letters” is in bold as well:
Not every instance of “letters” is highlighted in the search results. When it’s determined that the text is referring to the kind of letters that you would write to someone, those are left out of bold/highlighted text.
So when you’re researching your keywords, you’ll want to try and get an idea of how Google associates certain terms.
Understand Search Intent and Who the Searcher Is
Search intent can be tricky, but it’s a critical piece of your keyword research.
You want to go beyond understanding the usage and meaning of a term and truly understand the intent of a search term. What kind of audience would search that term? Do they already have some knowledge of the topic they’re searching for, or are they beginners?
This is where things can get a little hairy, especially with low-volume, “niche” keywords.
Take this for example:
While these search results for “freelance writer websites” are a huge improvement when I ran this search about a month ago, the snippet indicates that there’s still some confusion on what searchers are trying to find.
Previously, these search results were riddled with results for lists of freelancing sites (some of which had nothing to do with freelance writers and were only for designers).
Now, every one of the results shows examples of sites that belong to freelance writers, where you can hire a writer. (Not where writers can find a job.) The snippet is the only thing that indicates that searchers might be freelance writers that are looking for work.
So, if you were to write an article that aimed to rank for “freelance writer website,” you would want to write about how to build a website that gets writing clients, rather than listing a bunch of sites where writers can find work.
You also want to understand who is searching for your content. In this case, it’s probably going to be a mix of beginner writers who want to know how to build their online presence, or intermediate writers looking to go from their day job with a firm to freelancing on their own.
Remember the NLP tool we talked about earlier?
That’s going to come in handy if you want to get an idea of how Google understands your content. It won’t look pretty, but you can get a complete breakdown of your content. (And it might take a little longer, but you can feed this thing entire documents.)
If you really want to optimize for RankBrain, you’ll want to learn a little more about natural language processing and how Google uses it. You don’t have to go too in-depth, but at least play around with this tool and understand how it makes associations.
Optimization is Important, but Prioritize Your Readers
While it is important to consider search engines and how they see your content, your audience always comes first.
If your content suffers due to optimizations made purely for search engines, you’ve taken a wrong turn. While you might be optimizing for one part of the algorithm, there are other things that affect your performance, like site speed, mobile optimization, how much time readers spend on your page, how quickly they bounce, and more.
If you had to pick one thing to write for, make sure it’s your audience and not the machine. Your audience is what supports your business, not Google. It can help get your content in front of people, but hitting rank one shouldn’t necessarily be the end goal.
Write content that appeals to your audience, in a language they can easily understand. If you really make a connection, they’ll often share your content and get it in front of even more people. This is a lot more valuable than getting your site up a few spots by the end of the week.
Don’t forget your real audience!