RankBrain is a relatively new AI-driven upgrade to Google’s search algorithm introduced in October 2015. Through AI and machine learning, RankBrain helps Google to process search results with the goal of providing the most relevant search results to its users.
Why is Google RankBrain Algorithm such a big deal these days?
Experts predict that with the prevalence of smartphones and smart appliances, voice search will quickly become a norm, and half of all online searches will be voice-based by 2020. This means that queries are more likely to be in the question form, with longer, more complex structure and maybe even a more conversational tone.
As you will see in the text below, RankBrain was designed to handle such online search. To be honest, Google’s algorithms introduced earlier would not be capable to deal with voice search, because this type of queries requires understanding context and users’ intents.
There is another major reason for introducing such an advanced search mechanism. Even though Google’s search engine is the most popular one in the world, handling countless queries on a daily basis, a research shows that 15% of the incoming queries are new. In other words, Google has never before processed 15% of what people enter into its search bar.
Now it becomes even more transparent why would it introduce an AI-based search algorithm, doesn’t it? An AI-based algorithm could better interpret the 15% of unfamiliar queries.
Anyway, as we all know, all Google wants is to give its users the best possible results, even if they search for something unique.
And this is where Google RankBrain truly shines.
We’ve prepared a comprehensive guide to help you figure out RankBrain – what it means for your current SEO efforts, and what you need to do to optimize your website so that it has a more significant impact on search results.
5 Things You Need To Know About Google RankBrain
In a debate with Eric Enge of Stone Temple, Google’s Gary Illyes revealed that RankBrain works with “layered neural networks,” trying to predict the best search results based on historical data and through deep machine learning.
While Google RankBrain is merely just a component of the entire search algorithm, it does have a very specific function – to determine user’s intent.
Roughly speaking, the story goes as follows. First, RankBrain figures out exactly what a query is about. Then it draws from similar queries in the past and chooses among common search signals the most relevant ones for that particular query that will determine the value of websites. After this, it helps to rank and display the webpages in the search results.
Let’s illustrate this with a very basic example. We will later explain this process in more details.
Imagine it’s Sunday afternoon, and you enter “NFL scores” into Google’s search bar. RankBrain will first figure out that you want to see the scores of the NFL matches.
Then, considering it’s Sunday and that matches are currently being played, RankBrain will figure out from this context that your intent is to follow the scores in real time. Therefore, it will arrange the search results so that you will first see websites that offer live scores feature.
It’s some pretty cutting-edge stuff. According to Illyes, “it works really, really well with queries that we’ve never seen before.”
Here are five things you need to know about RankBrain.
RankBrain vs. Google’s Engineers
Back when Google was still testing machine learning programs and software in its labs, an early prototype of RankBrain was put against a group of Google’s engineers to determine the best page for a given search. Guess what? RankBrain outperformed these crack code commandos by 10%!
No, it wasn’t a fluke. With proof in hand that RankBrain actually works, Google rolled it out in 2015, working alongside with the core algorithm to deliver even better search results.
It’s Now The Third Most-Important Factor For Ranking Websites
In just a relatively short time, RankBrain became the third most important ranking factor out of Google’s 200+ list of ranking signals that determine a website’s position on search results. It’s just under “backlink profile” and “quality of content” factors.
Of course, that’s a pretty simple way of looking at things, given that many subfactors fall within “links” and “content.”
It will suffice to say, however, that Google sees the contributions of RankBrain as pretty important. It even goes as far as letting RankBrain to decide which other signals should carry more weight given a particular query.
Voice Search Is The Future of SEO
As the virtual assistants market continues to heat up, experts predict that by 2020, 50% of all searches will be powered by voice. Even as early as 2016, Google’s CEO, Sundar Pichai, revealed in a keynote that 20% of the queries Google processes are voice searches.
Also, the growth of personal virtual assistants is seen across all demographics, with young and old users alike embracing the new technology.
That said, optimizing your website for voice search is a must.
Much of Google RankBrain’s Learning is Done Offline
RankBrain observes pretty much everything users do while conducting online searches. This data is later used, among other things, for determining the quality of the websites they visit.
However, before it actually makes any adjustments to search engine results pages (SERPs), it does a series of A-B split-tests to validate its own findings. During this process, it “asks” itself, “If I bump this page higher up on search results because of its great CTR, will more users appreciate it?”
Data from these constant comparisons and validations is processed offline, with the analysis being used to tweak search engine results bit by bit. These adjustments go live the next time someone runs a similar search.
This cycle of offline learning and testing is what makes RankBrain valuable to the future of Google search.
RankBrain Will Never Be SkyNet
Google’s shiny new toy is AI-driven. It’s designed to be good at a very specific task – understanding natural language and how it relates to words and phrases entered into the search bar. It then uses this understanding to make adjustments to a particular search result.
And that’s it. There’s some pretty nifty machine learning going on throughout the process, but that’s the limit of RankBrain – it’s not going to wake up one day with human cognitive faculties.
So, we don’t have to worry about it gaining sentience and suddenly deciding to launch a doomsday attack on the entire human race.
SEO in the Era of RankBrain
From the very first iteration of the PageRank search engine algorithm that Larry Page and Sergey Brin came up with in 1996, experts were trying to figure out how to put their websites among the top results.
SEO as we know it today is a result of incremental adjustments made to ensure the best possible online experience for anyone looking for answers, solving a problem, making a purchase, or finding entertainment using a search engine.
The inner workings of Google’s core search algorithm is mostly a mystery, and no one outside the company knows exactly how it works.
But through trial and error, educated guess-work, insights from leading digital marketing experts, and the occasional interviews with Google’s senior officers, the SEO community has managed to flesh out a pretty good idea on how optimization ought to be done with RankBrain now being around.
Understanding How the Google RankBrain Algorithm Works
It Analyzes Keywords via User Intent
This goes beyond matching strings of text – RankBrain figures out what you mean when you type in a query.
This is what RankBrain was built for, as pointed out by Google’s Gary Illyes. Even when a user types in a bunch of keywords in a query that Google has never encountered before, RankBrain can go one step further and connect these new keywords to the ones it has seen before.
Let’s illustrate this with an example by Moz’s Rand Fishkin, which he used at one of his Whiteboard Friday sessions.
Assume you have the following five search queries:
- Best Netflix shows
- Best shows on Netflix
- What are good Netflix shows?
- Good Netflix shows?
- What to watch on Netflix?
RankBrain knows that these five queries are essentially looking for the same thing. No matter how the queries are typed, the user’s intent is, “I want to know which programs on Netflix are worth watching right now.”
RankBrain can form concepts out of keywords and find pages that best fit those concepts. So, no matter how different the words are in each of the above examples, as long as they mean the same thing, search results will show pretty much the same output – highly recommended programs or movies currently available on Netflix.
RankBrain Analyzes User Satisfaction
Taking its job a step further, RankBrain can validate its work by the way you like or dislike its search results. It can also adjust these results in the future searches.
Through user experience signals, RankBrain can determine if you’ve clicked on a search result and spent a lot of time on that page, or if you’ve hit the back button on your browser to look at another result.
Simply put, when a user finds a site down the list and spends more time on it because it has awesome, useful content, RankBrain can give this site a little boost so that it appears higher next time someone does a similar search.
Conversely, if a user merely pogo-sticks through a site (i.e., clicks on the link displayed in a SERP, has a quick look at the site, decides it’s not what he’s looking for, and then hits the ‘back’ button to find another place), RankBrain could choose to nudge this page down so that more helpful sites appear next time someone searches for the same thing.
RankBrain merely aims to please the users. By looking at a site’s organic click-through rate, dwell time, bounce rate, and pogo-sticking, RankBrain can “decide” to adjust said website’s position on search engine rankings depending on how favorably users interact with the site.
Let’s go back to the Rand Fishkin’s example above. To figure out what the search is about and which results deserve to be displayed first, RankBrain might go through the following relatively simple set of ranking signals:
- Keyword matching
- Link diversity
- Anchor texts
- Domain authority
- Relevant topics
- Content depth
Thanks to historical search data, RankBrain already knows which sorts of websites will yield the best results for this particular query, and therefore, knows which ranking signals will be more important in this case.
So, when it comes to the Netflix shows, Google knows that pages that feature the latest shows will tend to get higher CTA and dwell times than pages that feature shows from the last year.
For this reason, out of our eight ranking signals, RankBrain might decide that “Freshness” is paramount, and could bump up sites featuring the latest Netflix content.
As for the rest of the ranking signals, maybe “Engagement” might be as equally important. If a lot of people view that particular page and leave comments, for example, RankBrain might decide that this ranking factor should be also taken into account.
Underneath these two, “Relevant Topics” and “Content Depth” might be somewhat important as well, whereas “Keyword Matching,” “Link Diversity,” and “Anchor Texts” may be considered irrelevant for this particular query.
To sum up, with user satisfaction in mind, RankBrain may decide to bump up (or bring down) sites based on ranking signals it deems essential for a particular query.
RankBrain Vs. Past Google’s Updates
Before introducing RankBrain, Google has made a number of game-changing updates to its search algorithm.
- Panda. Rolled out in 2011, it was an adjustment to bring high-quality content to the users by promoting or demoting websites in search results, depending on the usefulness and relevance of content. Effect on SEO: sites with more helpful content were ranked higher.
- Penguin. Implemented in 2012, it was an adjustment for backlinks. It worked as a deterrent for black-hat SEO practices such as link farms, link pyramids, and similar illegal techniques. Effect on SEO: sites with high-quality backlinks rank higher.
- Hummingbird. Announced in 2013, this major update put even more emphasis on human interaction, context, and natural language. Effect on SEO: long-tail keywords now have more significance, as they help in determining the context of a search.
- Pigeon. Released in 2014, it improved the quality of local searches, tying in results to Google Maps and other services. Effect on SEO: small businesses are now more visible in search results. New rules and guidelines for local SEO are now in play.
Google is already a lot more sophisticated compared to its earlier iterations. However, by introducing machine learning to the core search algorithm, it is now even better equipped to deal with the challenges brought by voice search, multi-word queries, and more complex, long-tail key phrases.
With RankBrain, Google can discern user intent behind a particular query. This will affect your keyword strategy for sure, so let’s take a look at some general hacks for optimizing your site to match the RankBrain-powered algorithm.
4 Tips For Optimizing Your Site With Google RankBrain in Mind
First of all, let’s clarify that RankBrain motivates you to reassess and re-adjust the overall SEO strategy by providing top-notch quality to the users. Employing black-hat strategies would not only be useless but incur a penalty as well.
The Long-Tail Keyword Strategy is Now Obsolete
Rand Fishkin says that if you haven’t killed it yet, it’s time to ditch the antiquated “one keyword, one page” strategy.
Even before the RankBrain update, the focus has been shifted towards high-quality content. Basically, Google has told site owners to create content with humans on mind, not search engines.
RankBrain merely drives this point home. Since it can already group together similar keywords and key phrases, it knows that classic variations (and even common misspellings) of these terms are essentially the same thing, and will use them to show almost identical search results.
That said, optimizing for long-tail keywords is now obsolete. Even worse – it could adversely affect your site’s overall user experience.
Instead, go for medium-tail keywords, those that have just a little bit of search volume, but aren’t as competitive as the top-tier keywords. Build your page around these medium-tail keywords.
Improve CTR By Optimizing Titles and Description Tags
Neil Patel of NeilPatel.com has pointed out on several occasions that you can make your site more RankBrain-friendly by working on its click-through rate.
Assuming your web copy is already polished, the next thing you’d want to do is to come up with a compelling headline and a description tag that essentially hooks users in right away, telling them to click your website because you have the answers to their questions.
Your site’s title and meta-description will pop up on a SERP, so take the effort to make these extra catchy. After all, your site’s title and meta-description are appearing along with other results. Standing out from the crowd could grab users’ attention at least to give you a click.
Take a cue from magazines in a newsstand. Magazines with flashy and engaging titles have a better chance of being picked up by readers. The same goes for your site and search engine results.
Write With Empathy
Studies show that posts with higher emotional value not only increase the likelihood of reeling in more visitors but also tend to get more shares.
You could also try working on some power-words or action-words like “easy,” “effortless,” “how to,” “simple,” to help your headline stand out.
Instead of just having “Best Netflix Documentaries” as your headline, go for something like “7 Must-Watch Netflix Documentaries For Entrepreneurs.”
Optimize Description Tag for Better CTR
To create a good description tag, answer the following question, “Why should anyone click on your site’s link?”
Of course, you will include your target keyword and add some emotion via power-words or action-words (just like your headlines). But description tags allow you to insert a quick pitch, and you could use this to your advantage to further stand out in a page full of search results.
Brackets and Parentheses Work
One of Brian Dean’s favorite headline hacks is using brackets and parentheses. Citing a study done by HubSpot, headlines with brackets outperformed non-bracketed titles by 33%.
Going back to our “Best Netflix Documentaries” example, one way you might spin it would be “5 Awesome Netflix Documentaries (That Make You Smarter).”
Numbers Work Too
A study conducted by BuzzSumo supports the idea that numbers used on headlines improve CTR.
Optimize Content For Bounce Rate and Dwell Time
Once visitors land on your website via search results, the goal is to keep them there for as long as you can.
Well-organized, thoughtfully-crafted content is the best way to lower bounce rate and improve overall dwell time on your site.
This is particularly true in this day and age when everyone is using smartphones for conducting search. People want answers right away, so at the very least, your site should not only be geared for mobile traffic but also fast-loading and secure.
RankBrain can infer that the longer someone lingers on your site, the more likely it is that they’ve found what they’re looking for.
Furthermore, the way visitors click on specific links or buttons on your site is an indication of engagement or involvement. As long as they don’t bounce right away, you’re safe.
RankBrain picks up on all these signals and uses them to adjust rankings on search results.
Push Content Above The Fold
You might want to reconsider those fancy large-image headers.
Your visitors will want to get answers right away, so you should not only make an effort to bring all your awesome content up above the fold but also make sure this information is first to load whenever someone lands on your page.
Publish Long, In-Depth Content
If the idea is to keep visitors on your site for as long as possible, the best way to hook them in is by actually providing insightful and relevant content.
Google already prefers longer quality articles over a plethora of shallow 300-word pieces, and RankBrain further reinforces this preference.
Sprinkle other rich media into your content. Aside from high-resolution photos, consider using infographics, pictographs, video, sound bites, and downloadables as these all contribute to the usefulness of your page, user engagement, and most importantly, the amount of time visitors spend on your site.
Use Short Intros
Another practice to ditch is using several paragraphs to lead into your piece before getting into the gist of things. Digital marketers previously used this tactic to squeeze in an extra keyword mention or so, but RankBrain has rendered that practice obsolete.
Besides, people already know what they’re looking for, so give it to them right away. There is no need to bury answers in paragraphs of fluff text.
Articulate the main points, address them right at the beginning, and even go as far as highlighting a couple of key sentences in bold characters.
The idea is to hook your readers in as you tell them what your article is all about right away.
Break Up Content
Another turn off for many readers is long blocks of uninterrupted text.
Break up your piece into smaller, easier to digest and easier to process chunks. The strategic use of headers and subheaders help organize the information. Implement breakout areas to highlight useful, related information. Use shorter paragraphs and consider splitting long, winding sentences.
Those extra pieces of rich media (video, images, etc.) all help to make your content easier to process as well.
Other Google RankBrain Optimization Tips
Brian Dean of Backlinko has a few additional SEO tips to help you have a better performing website on search results in the era of RankBrain.
Improve CTR by Increasing Brand Awareness
When faced with two similar organic search results, users will be more likely to click on a brand they recognize, respect, and love. This is why investing in getting the name out there will pay off nicely in the long-term, especially for brands that are just starting out.
Consider putting in some resources into ad campaigns on Facebook, have a useful or informative email newsletter that your visitors can sign up for, and do some outreach on the strength of your awesome content.
Monitor Paid Ads For Keywords You Should Be Using
Check out the paid ads of your target keyword, and if you notice specific words that somehow stand out, try to include them in your meta-description as well so that you could also be found by those keywords. You need to align them alongside the main keyword.
Re-Optimize Old Content
Check back on some of your old posts from time to time and give them the much-needed update.
Some of these might already be irrelevant, in need of a more current spin, or a fresh set of examples.
Also, don’t forget that you might have had optimized old content using outdated practices. Make sure to check your on-page SEO for those old pieces as well.
Use LSI Keywords To Fill in Content Gaps
Latent Semantic Indexing (LSI) keywords are terms and phrases naturally associated with a particular topic, helping Google better understand and give context to your web pages.
Now we know that the old practice of keyword stuffing is dead. Instead, RankBrain appreciates liberal use of LSI keywords as they let it to determine if your content is sufficiently thorough and comprehensive.
Make a list of other words, terms, or phrases commonly associated with your main topic or keyword, and make sure your webpage covers as much of these as possible. There are also marketing tools that can help you build a list of LSI keywords.
In essence, RankBrain helps Google determine which signals carry more weight given a particular search query, and it does so by analyzing user’s intent.
Gone are the days when one set of fixed inputs governed every single search query; RankBrain’s system of weighing ranking signals is a real game-changer in this regard. Also, variations of the same keyword is now rendered obsolete. If they all have the same meaning, RankBrain will recognize them as such and treat them as one keyword.
User’s intent drives RankBrain, and as such, it can make decisions on behalf of the Google’s core search algorithm – deciding which of the 200+ ranking signals will be given a higher priority in any specific query, or which keywords are associated with certain phrases and terms under a particular context.
Simple changes to the SEO practices in the era of RankBrain include the following:
- ditching the antiquated “one keyword, one URL” strategy and focusing on overall user experience;
- improving click-through rate by utilizing thoughtfully crafted titles and meta descriptions;
- improving bounce rate by adding a wider selection of helpful, relevant media along with written content.