Heard about RankBrain? It’s not exactly a new SEO concept (as it was seen making waves as early as mid-2015), but it’s one that’s gaining popularity and garnering a lot of attention from SEO experts in recent months. Why? It’s not just something Google’s testing anymore. It’s a major ranking factor for one of the world’s top search engines. And it’s here to stay.
For most of us, the concept of machine learning elicits complex images of technology, expensive, bulky machinery, even robots. But in today’s mobile, connected age, that’s not always the case.
I did some research, and machine learning is a lot more accessible than we may realize.
Familiar examples of machine learning:
- Rideshare services like Uber
- Robot vacuum cleaners
- Spam filters in your email
- Mobile check deposit capabilities with your bank
- Social media facial recognition and tagging recommendations
- Product recommendations related to recent searches from your favorite online retail store
In the simplest terms, machine learning is when a computer teaches itself how to do something over time, rather than being pre-programmed to perform the task from the get go. While it may start out with a base of knowledge, it slowly builds on what it knows, sort of like artificial intelligence.
So why am I talking about machine learning? It’s one of the latest ways technological advancements are disrupting your business’s ability to get found online.
How You Get Found Online
A lot of factors help you get found online, which can be both good and bad. On the plus side, you have nearly endless opportunities to help your business out. On the other end of the spectrum, you could be missing out if you’re not taking advantage of every digital opportunity available to you.
Ask yourself this: When someone searches for your business via the most popular search engines like Google, Bing and Yahoo, are you showing up? There may be several things impacting your search engine optimization (SEO) both positively and negatively.
Factors impacting your SEO:
- Online business listings
- Social media presence
And one we so often forget about…
Search engines like Google’s ever-changing algorithms.
Explaining Google’s RankBrain
Google’s main algorithm is called Hummingbird. This is the algorithm they overhaul regularly as time passes, to make sure they’re staying hip with what users are searching for. Google has more than 200 add-ons, pieces and parts that make up and plug into Hummingbird, like PageRank which focus on the concept of backlinks (but that’s an SEO topic for another time).
The latest factor affecting Google’s search algorithms is RankBrain. Like I mentioned, Google regularly updates and adjusts its algorithm with newer, better versions. But RankBrain isn’t that.
Instead, it’s what Google is calling it’s “third top ranking signal” falling only behind content and links. Like PageRank, it’s not its own algorithm. Instead, it’s one of the largest influencers for how Google’s algorithm works.
In prior years, Google’s engineers hand-coded 100% of the algorithm themselves. They’d rack their brains thinking up different ways the search engine could better the results it delivers to consumers, test them, then implement the changes that had the most positive impacts on user experience.
With Google’s RankBrain however, the search engine can now use machine learning to tweak the algorithm without direct input from Google’s still-brilliant engineers. Fun fact: Google compared RankBrain’s capabilities against their own engineers, and they estimated that RankBrain actually performed 10% better in delivering accurate results, on average.
How RankBrain Works
RankBrain is built to interpret search terms and phrases like a human would, instead of like a machine or piece of code would. It does this by trying to create relationships between keywords Google has never seen before and keywords Google is already familiar with. According to Backlinko, “Google RankBrain goes beyond simple keyword-matching. It turns your search term into concepts… and tries to find pages that cover that concept.”
Once it’s done the behind-the-scenes work, RankBrain starts analyzing how searchers are interacting with the results it suggests. It looks at things like:
- Percentage of clicks through to its recommended pages (click-through rate)
- The amount of time spent on those recommended pages
- Where users go from there (Some call this bounces, but let’s take it a step further and talk about pogo-sticking really quickly. No, I’m not trying to be cute. This is a real thing.)
More on pogo-sticking: This is when you enter a search term and click on a result. But instead of staying on the page to read more, you notice the result isn’t quite what you’re looking for. So you hit the back button and try the next link down the list. Maybe that one doesn’t quite suit your fancy either, so back you go again…and so on and so forth until you reach a page that actually works for you, and you hang out a while. This back and forth is called pogo-sticking in the SEO world. Search engines measure this activity, and they use it to punish the pages users “pogo” (bounce) away from and boost the pages visitors spend longer periods of time digesting.
Anyway, signals like the ones mentioned above help Google (fueled by RankBrain) determine if its search results were valuable to the user or not. If they were valuable, RankBrain pats itself on the back and gives the pages a rankings boost for the search terms it was analyzing. If it deems its recommendations weren’t valuable, it sulks a bit, and then it drops the pages down further in its results.
Making Friends with RankBrain
Now that you know how RankBrain works, it’s on you (or your SEO team, if you have one) to adjust your web strategy accordingly.
There are a few ways you can adjust your SEO work that’s unique to RankBrain. That’s not to say these techniques will only work for Google. Luckily, most search engines adjust their algorithms to incorporate similar advancements over time. So this advice should help you out in more than one way.
Improve your SEO with RankBrain:
- Move from long tail keyword research to medium tail keyword research. Long tail keywords helped us home in on very specific search phrases. But if RankBrain can understand concepts and make connections on it’s own, we can take a step back and write about broader concepts again. Think 3-word terms instead of 5- or 6-word phrases.
- Be flashier with how you name your web pages, blog posts and even images and videos. More enticing titles get more clicks than generic ones do. And more clicks are an easy win with RankBrain. Use enticing words like “Proven” or “Exclusive”.
- Don’t ignore meta descriptions. When you build pages on your website, your website builder will give you a chance to describe what’s on each page and each page element. While not required, this description shows up below the title in search results, so it impacts user perception of your content.
- Don’t forget to incorporate keywords you’re trying to rank for in your titles and their meta descriptions.
- Examine page performance on your website, and revamp or remove pages that don’t perform well.
- Trust Google. We’re not the only ones with advice on SEO. Google tries their best to help us out now and again. And according to them, this blog post of theirs from 2013 is more relevant than ever when it comes to understanding RankBrain and how they’re using machine learning to connect search terms to results. Fair warning: They get really in the weeds about vectors, models, and other math-y terms. So don’t expect a light, easy read.