Understanding native language to deliver more target search results has always been a goal for Google Search. In 2015, Google released RankBrain, their first big step toward achieving this goal. Just last week, they made, "the biggest leap forward in five years" toward better understanding the native language with the release of BERT.
BERT is designed to better understand the nuances and context of words in searches and better match those queries with more relevant results in search and featured snippets. Watch the video below to learn more.
Hey, what's up and welcome to Hack My Growth. In this episode we're going to be discussing BERT, Google's new and improved AI ranking algorithm. All right, let's go. Are you looking to grow your business but you're not sure where to start? That's where we come in.
Thanks for checking out this video. If this is your first time watching or maybe you've been watching a while and you haven't yet hit subscribe, please do so now. We would love to have you join our community. We create new content each and every week to help you get more out of your digital marketing activities. Don't forget to hit the alert button that way you know each and every time a new piece of content is published.
Like I said in the intro, this video is going to be about BERT and it's Google's new AI enabled ranking algorithm. This is something that was released in October of 2019 between ... I think, 25th is when the Search Liaison started putting out some information on Twitter about it. And this is something that Google has been working on for quite some time. About a year ago they talked about this open source project known as BERT and it stands for Bi-directional Encoder Representation for Transformers. Now if you're hearing that you're probably going, "Great, I can no longer do SEO because I have no idea what that means." Completely understand it. Sometimes this stuff when it gets into the nitty gritty and to the machine learning, and the algorithms, and the AI involved it can really be overwhelming. But take a deep breath, exhale because it's not really as complicated as it may sound from the marketer standpoint. You don't need to understand everything that's going on under the hood to know what Google is doing here.
So this is something that we need to pay attention to for a couple of reasons. One, Google has called this the biggest leap forward in five years. This is something they've been investing in. As you know RankBrain was something that came out a number of years ago, a lot of people freaked out about that. And this is not replacing RankBrain, as you can see, but it will work together with RankBrain. So right now it's only impacting about 10% of the queries. It's in the US, it's English speaking queries only. It will be rolling out to more as time begins to evolve and Google shifts, and change things. Now this is something that they'd been investing a lot in and it's impacting not just native search, so the blue links, but also the rich snippets.
Now Google's goal has always been to understand the context of the query. Now if you remember where you were in SEO maybe in 2009 or 2010 it's a lot different than it was back then than it is today. Back then we talked a lot about keywords and keyword placement, and how Google was really honing in on one specific term, and then again they were still trying to figure out the context but really only had language. Now what this does, this bi-directional encoder, it's a form of what's known as NLP or natural language processing. That is the process of a computer trying to understand how language works. Like I said, back in 2010 they would look at a keyword. You could do something like bird watching, right? This is a an example, and you would type in bird watching. Google would look at the term bird, it would look at watching, it would give you results that had something to do with both but it could be a number of different things that the users are really looking for. Is it a bird watching club? Is it a bird watching park? Is it a bird watching adventure? Is it something else that's called bird watching that has nothing to do with the actual watching of birds?
There's a lot of things going on there so in order to understand that computers had been been put through this machine learning process of natural language processing which helps them to better understand language and how people are using language.
With BERT, what they've done through all of these studies, it's been enabled to them to increase the accuracy by two percentage points. That's a huge deal when it comes to computing power and understanding what's happening when people are asking Google these questions, and trying to learn things online. Now there's a couple of really cool examples that they gave, that Google gave, with a test that they've done of where it's helped really improve the accuracy of what the user was looking for to help them get to the right piece of information.
A lot of SEOs are right now going, "How do we optimize for this? What should we do? How do we make sure that our sites are BERT ready?" You can't do anything to be honest. It's a machine learning algorithm. It's taking data and it's trying to understand the context of that data based on the learnings that it's been given so you can't optimize for BERT. What you can do though is continue to optimize for end users. And this is something if SEOs, content marketers, digital marketers would really grasp instead of chasing trends and chasing algorithms, you would actually do a lot better. Know what your user is looking for and make sure you give them the comprehensive, accurate, step-by-step details so that it fulfills their expectations, and they take that next step.
Instead of focusing on, "How can I break down and natural language processing that Google is using in order to ..." You're not going to be able to do that on your own or really ever unless you've gotten the type of quantum computing power that they're using, and the things that they're doing to really understand this. But I'm going to give you a little bit of a breakdown to how this works and then how we can translate that back.
So right here I've got a query, a simple query, "How to optimize for search." The subject of this query is about optimization. So optimize, that's what you're trying to do, you're trying to optimize. These green lines here on top where Google will take fragments like, "How to", "Search" and then, "Optimizing," put it together, right? So how to optimize. All right, that's what's happening. And then what do they want to optimize? They want optimize search. Now you could get a lot of interesting queries for that. You may get some very accurate queries for that because we know a lot about search and Google does as well. But as you notice that you're having these uni directional agreements or you need directional connections which means they're just going one way. So, "How to optimize," and then, "Search," optimize search.
Now if you're into NLP and you know a lot about machine learning this is a very oversimplified example. Feel free to comment and tell me how wrong I am below, but as I understand it and how the process works and the courses, and the classes that I've taken, this is just a simple breakdown. We're seeing these one way connections. Now what BERT is able to do is build these multi or bi-directional connections between all of these terms. So we're understanding, "How to," we're understanding the, "Optimize," we're understanding, "Search." But we're also understanding this word here, "For." There's things like for and to that are used in the English language that have a ton of weight, and they can change how that question is being asked, and the results that it should see. So if we did, "How to optimize to search," that would be the user, how I can do searching better, right? If I had, "How to optimize for search," this is how do I optimize for something. That little word changes the meaning of that sentence quite substantially so with this new BERT algorithm they're able to better understand those connections. Again this goes back to what can we do about it? Know what your users are asking.
This is where you can use tools like Search Query or any of your other SEO ranking tools that are out there, right? We use SEMrush all the time. It allows you to see what people are searching for, what are the questions they're asking, look at what Google is showing for those results and then make sure that you're meeting or exceeding the user's expectation. That they're giving them the exact answers that ... to their questions that they're asking online. All of this stuff is very interesting. It's something that we should pay attention to. You should go and look to see if you've been impacted. Right now, like I said, only about 10% of queries have been impacted by this and Google doesn't always use BERT. If they feel like this algorithm is going to help them give a better answer to a specific query, it applies the algorithm in the same way that RankBrain works. They have all of these tools in the back end that they're trying to use to give better results to the end user.
Focus on the users, understand that computers are trying to understand how languages are used, understand how the questions are being asked, the types of questions your users are asking, and then give them the results. How can you fulfill that solution?
Search is being moved more ... Honestly, digital marketing is pushing more towards the user. What does the user need? And this is something Google has always been after. This is something that every platform that's been marketing is after. What do people want to see? Great, that's what I'm going to give them. Any media platform. These are the things you can start to think about, understanding how this all works. You don't need to get deep into ML and AI unless you're interested. Google's got a number of resources. I'll give you some links in the video that you can read to help educate yourself on this.
If you've got any questions, please comment below. We'd love to help you walk through this process but in the long run, take a deep breath. It's another algorithm. We don't need to force ourselves to do anything. We just need to continue to answer the questions and give our end users exactly what they're looking for, and that's how you're going to win, not just in search, but really win in business. All right. Thanks so much for watching and until next time, Happy Marketing.
We are a participant in the Amazon Services LLC Associates Program, an affiliate advertising program designed to provide a means for us to earn fees by linking to Amazon.com and affiliated sites.
Headquartered in Melbourne, FL