Google RankBrain: What Is it, Why Does it Matter in 2019?
Google RankBrain was in the news again throughout 2018.
Now that 2019 is here, let’s take a closer look at it moving forward.
Studying RankBrain can help you improve your SEO in the new year, particularly with voice search exploding in our the digital marketing industry.
What Is Google RankBrain?
Google RankBrain is a core component of Google’s search algorithm.
It plays a role in helping Google with searches that do not have a clear answer.
Let’s take a step back and talk about the algorithm update itself. Google’s current is algorithm is Hummingbird. It launched in 2013. It included some updates that made a bit impact in the way searches are made.
Hummingbird is sometimes described as having moved Google “from strings to things.” What does that mean? Before Hummingbird, Google looked at strings of words. Sometimes, the results might not be relevant.
An example might be a search for the word “apple.” If you searched it hoping to find an Apple product, you might have been disappointed by results about fruit.
Hummingbird Google algorithm updates hanged that by doing three things:
It improved Google’s ability to interpret conversational speech. Complex search strings are not needed. Now, it can reply to questions the way you might ask when talking.
It looked at human intent. Even if you search using a word that isn’t correct, Hummingbird can guess what you mean. Then, it delivers results based on your question.
It improved local results. By combining intent with location targeting, it delivers relevant results. If you search for a French restaurant, it will not give you results for Paris. Instead, it offers results for your city.
Hummingbird marked a significant change in the way Google worked. It improved search engine results, but, it still had some flaws. RankBrain addresses those.
RankBrain uses machine-learning algorithms to interpret unknown search strings. Its job is to help Google interpret queries it does not understand.
The “apple” search I referenced before is an example. With Hummingbird, Google knew that an apple was a thing. Instead of searching strings of words, it started with a base of knowledge. However, it could not understand that “apple” might also be a brand name.
RankBrain does understand. Even in the absence of context, it might deliver a mix of results. Some still refer to fruit, but others involve Apple computers.
Google RankBrain is not AI in the way we see it in movies. Machine-learning is a more accurate description. The “machine” of Google’s algorithm learns as it works. As it adds knowledge, its ability to deliver relevant results improves.
The History of The Google RankBrain Algorithm
Google RankBrain and Hummingbird are the latest in a long line of search algorithms. In the early days of Google, weight was given to keyword usage above all. That is where the practice of keyword stuffing came from. Remember Google Page Rank?
RankBrain represents a refinement of what Google is already doing. It is not the only component of the search algorithm we know about. Others include:
Each of these components has a different role. Penguin and Panda fight spam. Mobile-friendly rewards sites that are mobile responsive. Pigeon improves local search results, while Pirate fights copyright infringement.
As Google analyzed Hummingbird and found weaknesses, it realized the need for a contextual tool. That is where RankBrain comes in.
How Google RankBrain Works
When we first heard about RankBrain, it was in a Bloomberg article. Typically, Google announces algorithm changes only when they are significant.
RankBrain first launched in early 2015, but we did not learn about it until October of that year. Google search engineers said it played a role in 15% of all searches. That might not sound like much, but Google processes approximately 4.4 billion searches per day.
What made those 4.4 million searches different?
According to Google, they were searches that were new.
In other words, search terms that had never been used before. Google needed a way to interpret them, and that is how RankBrain was born.
RankBrain is used on all Google searches.
However, that does not mean it has a significant impact on all of them. Here is an example how RankBrain works.
In a very specific search, RankBrain will not affect the results. For example, consider this query:
Who is the US Secretary of State?
That is a simple question with a simple answer. Nothing in the way of context is needed. The results are likely to be close to what they would have been before RankBrain.
With more complex queries, RankBrain is essential. It helps add context. It works with Hummingbird’s capabilities to deliver the most relevant result rankings. For example, a query like this might require RankBrain’s input. This comes from the Bloomberg article.
What’s the title of the consumer at the highest level of a food chain?
That is an awkward query. It is easy to see why Google might need some context for it. The word “consumer” usually refers to someone who buys something. However, “food chain” is a term from biology. It refers to a hierarchy of animals and plants.
The person who typed this query wants to know the word or phrase“predator,” or perhaps “apex predator.” On its own, Hummingbird cannot deduce the searcher’s intent. With RankBrain’s help, it can.
RankBrain takes search queries and creates word vectors, mathematical representations of the search. It then searches by a vector, taking linguistics and other metrics into consideration.
It is important to note that RankBrain is not Natural Language Processing (NLP.) NLP is a form of AI that allows humans to have natural conversations with computers. It is considered the Holy Grail of search.
Another way to think of it is that RankBrain is in the background of Hummingbird.
It always provides context, but sometimes the context has little impact on the results.
How Google Uses Machine Learning in its Search Algorithms
Let’s step back to talk about machine learning strategies in general. This is a topic that may be confusing for some.
Machine learning is what happens when a computer learns by experience.
There are two types:
Supervised learning occurs when the machine uses a classified set of data to learn. It can then apply what it learns to sets of unknown data in the future.
Unsupervised learning occurs when the machine uses unclassified data to learn.
Some machines may use a blend of supervised and unsupervised learning. In all cases, they can use their experience to improve their future performance. Constant learning helps programmers because the machines can learn without help.
You know the basics of how RankBrain works, but how does it learn?
Example of the RankBrain algorithm in action
Let’s go back to the “pink apple” search example.
Someone who searched “pink apple” might be looking for data about fruit.
Alternatively, they might be looking for a pink Apple iPhone to buy. Or a photo of a pink apple. Or a perfume brand named pink apple.
If you examine the results for this search, the findings are interesting and ever-changing.
Optimize for rank brain by studying Google’s processes in the search results.
RankBrain knows that all above possibilities exist.
In the absence of context, it returns a mix of fruit, perfume, and iPhone related results.
What happens if several searchers use the term “pink apple” and click on links about iPhones? RankBrain learns that searchers who use that term are more likely to want links to computers.
The mix of results might change. It would not necessarily eliminate all fruit-based results. After all, some searchers might be looking for that. What it does is provide a new mix of results. It might be half fruit-related, and half computer-related.
Over time, a list of results might change dramatically. If a new search cropped up, googles rank brain might present an array of results. If users consistently chose links on one topic, RankBrain would notice and learn. Eventually, the SERP might display only those results users were most likely to select.
RankBrain may also affect how Google ranks search results. It learns which results are most relevant and pushes them to the top. RankBrain’s SEO impact is something search marketers should all be thinking about.
Here is a simple example I came across the other day that made me think of Google’s RankBrain. I was trying to remember the name of a movie so I described it in search, and there it was.
Machine Learning VS. Artificial Intelligence
Tracking changing technological strategies can be a challenge. That is especially true if you are not someone who has experience with technology.
Google classifies RankBrain as an artificial intelligence system but also calls it machine learning. Is there a difference?
The simple answer is, not really. In science fiction, AI is presented as something nearly human. We see it in movies like The Terminator. Machines in those films are sophisticated to the point where they are barely distinguishable from people.
Google’s RankBrain is not like that. In the real world, AI and machine learning are used interchangeably. A computer that learns has artificial intelligence.
The word “artificial” means only that there is artifice behind the learning. When people learn, we do it intuitively. We do not need to be told how to do it.
Machines must be created with the ability to learn if their creators want them to learn. There’s a design to them. RankBrain is not a nearly-human robot. It is part of Google’s core algorithm, and that is all. At least for now.
How Important is Google RankBrain?
RankBrain is used in all searches, but how much weight does it have? Google plays close to the vest when it comes to revealing details of their algorithms. However, they made an exception with RankBrain.
According to Google, RankBrain is the third most important element in the search algorithm. Google mentioned their top three factors in a Q & A from March of 2016. They are:
RankBrain is important, especially considering there are hundreds of ranking signals that impact Googles RankBrain.
There’s no question about that. But what, if anything, can be done to improve your search rank now that RankBrain is in the picture?
How to Optimize Your website for 2019
One of the most common questions I get is, “How to use rank brain to better optimize my site?”
It is a necessary question, but the answer may surprise you.
Anybody who knows about SEO strategy knows that the importance of keywords has diminished. Links and content are still the most important ranking algorithm factors. Your content marketing plan should be on point.
Is there anything you can do to account for RankBrain?
The short answer is to provide detailed quality topical content for humans that provide a positive user experience. Forget about targeting a single word, think logical keyword grouping and topics. Provide answers to questions associated with the topic.
RankBrain’s most significant impact occurs when a new web search query is used. The only way to optimize your site would be to guess what those new searches might be.
One of the things we take for granted with Google is that they do not talk about search engine optimization often. They are not in the practice of telling people how to optimize. However, Google’s Gary Illyes gave a rare interview. He was asked about optimization best practices.
Here’s part of what he said:
“Try to write content that sounds human. If you try to write like a machine, then RankBrain will just get confused and probably just pushes you back.”
Surprising, right? If you want to optimize for RankBrain, which uses AI learning, you need to make sure not to sound like a computer.
If you think about it, it makes sense the direction Google is headed with Rank Brain.
Google’s goal for years now has been to encourage quality content.
Its algorithms might be machines, but its users are not. They want the user experience to be as positive as possible.
One specific thing that Illyes recommended is reading the content on your website aloud. This is a trick many writers use. It helps them catch awkward or unnatural phrasing.
With RankBrain, reading aloud can help you determine if your content sounds natural. Ideally, it should be conversational. The more like regular human speech your content is, the more likely it is that Google will reward you with a high search rank.
Another way to look at it is this. RankBrain is always learning, and so it is always changing. Trying to precisely optimize for it is a fool’s errand. It’s akin to trying to hit a moving target. Your time is best spent polishing your content, so it’s relevant to users.
What is Google Saying About Rank Brain?
When researching the latest Google updates I always like to search for what the leaders of Google search are saying about it.
Of course, they are usually tight-lipped. But with a little digging, information is only a search away as displayed in these two examples.
“Basically, it’s a ranking factor. It’s part of machine learning. It’s something that tries to identify patterns and bucket data. It looks at data about past searches and based on what worked well for those searches, it will try to predict what will work best for a certain query. This works best for long-tail queries and queries we’ve never seen.
An example might be “can I beat Mario Bros without using a walkthrough.” Without RankBrain, we give interesting results that don’t meet my needs. But with RankBrain, we can give results that satisfy my question.
RankBrain will understand better what results from work for queries. It’ll understand that certain stop words should not be dropped. Sometimes the word “with” is dropped from a query, but RankBrain will understand that we need to keep it.”
Webmaster Trends Analyst at Google at SMC advanced conference
Final Thoughts: RankBrain and Optimizing For Google
There’s no question that RankBrain will play a role in your SEO rankings in 2019 and moving forward. If your site is already full of quality, rich, topical content, you are well on the right path.
Of course, the one thing we know about Google is that nothing stays the same for long. The chances are good that RankBrain will be updated on a regular basis, if not in real time along with the other ranking signals and learning algorithms Google employs.
Your goal should be to keep your content relevant for search intent, conversational, and up-to-date. If you do that, you will give your website the best opportunity to please the ever-changing Google algorithm.