Most of us know that when we search for a phrase or a word on Google, an algorithm kicks in to find and serve the most relevant pages. Google has always dominated searches throughout the globe and it’s mostly because of its ability to provide the most accurate results at the quickest speed possible. However, more recently Google has taken an even more bullish advancement on it’s algorithmic technology, taking online search to a whole new level. In 2015, Google released RankBrain as part of the search algorithm called Hummingbird. As we know it, Google’s RankBrain has revolutionised the Search Engine Optimisation (SEO) world forever.
“RankBrain has gone a step further than the Knowledge Graph in matching the right searches with the best results. So much so, that it is now handling a substantial percentage of all of Google’s queries.”
RankBrain Explained
Google’s RankBrain is a machine-learning artificial intelligence (AI) system. By being an AI, the dynamism means that it can equal or even exceed human intelligence at one specific thing: processing Google’s search results for ambiguous queries. Fundamentally, RankBrain teaches itself. It does so through the use of mathematical processes and an advanced understanding of real language semantics.
Every day, the RankBrain Google algorithm learns more about how and why people search. It then applies this “knowledge” to future cases, so that Google can more accurately serve the user with relevant content. It is one of the most important ranking signals used by Google – but what are signals? They are elements that help Google determine how your webpage will rank. A keyword on a webpage, for example, is a signal.
What is the Knowledge Graph?
Launched in 2012, the Knowledge Graph allowed Google to search for “things not strings”. The previous system would only look for pages that matched the spelling of a word. The Knowledge Graph, on the other hand, could look for the meaning behind the word, and its connection to other words. So for example, by searching for ‘Trump’ (even though it is only a word), the Knowledge Graph would probably bring up highly relevant themes and current topics around the word – for instance, the ‘Presidential Inauguration’ or the ‘Muslim Ban’.
The problem with the Knowledge Graph is that it is not a machine-learning AI system. This means that somewhere, it still requires human input. Someone, somewhere, had to add a database that linked the connection between things.
So, how does RankBrain differ from it?
RankBrain makes the link between searches and content on pages that do not have the exact terms that were searched for. Simply stated, it acts as a refinement tool. Isn’t that what the Knowledge Graph does? Yes, it can refine queries, but instead of having a machine come back with the results, a human has to do most of the work – unlike RankBrain. There are some automations, but the problem is that, in 2013, 15% of all daily searches were new searches. This means that they had never been searched prior, and humans had to process them. This is becoming more prevalent, as we delve further into the mobile world, in conjunction with the adoptive use of voice searches.
Google designed RankBrain to better interpret these ambiguities, these so-called “long-tail” queries, and translate them effectively. It is helping Google to better understand what each of its search queries are about, with minimal human input. It also applies what it has learnt for future queries, by connecting the dots, through similar cases.
How does RankBrain actually work?
RankBrain sees patterns between seemingly unconnected and complex searches. From this, it can then understand how they are similar, and gives the searcher the best possible results. Let’s use a simple example. When someone searches for “best flower shop in Perth”, RankBrain may associate it with an alternative but more popular search, “best Perth flower shops”. It would then serve the user the most popular results from the second search.
It uses what was learnt from past searches, giving the search results that users would like the most, and which are highly effective. Whereas humans could correctly guess 70% of the time, RankBrain has had an 80% success rate.
Moreover, it does not stop at written queries, but also evaluates and analyses oral ones. If you ask Google orally “what is the name of that song that goes ‘I’m the master of my sea'” (see image below), RankBrain helps Google apply context and gives you the exact song that you’re looking for.
So, how much should RankBrain affect your SEO strategy? Let’s take a look.
What does RankBrain mean for SEO in 2018 & beyond?
RankBrain and the future of Google
RankBrain is a machine learning system. This implies that in the future, it won’t need any input, and with a dataset, it will be able to apply its learning process. RankBrain updates will happen automatically, as it refines its own algorithm. It also analyses and learns the new jargon created by industry experts. Digital marketing is evolving so quickly for example, that every year, you will hear new terms and words. RankBrain will look at the context to learn about each new word, and give you the results that you are looking for.
It also shows the direction that Google wants to take, especially when it comes to SEO. It would be no surprise, if the Google search engine became pure AI in the near future. But don’t let RankBrain scare you, because it mostly targets non-understandable searches, and it wouldn’t make any sense to try and create pages for these long-tail keywords.
RankBrain and the future of SEO
Therefore, when it comes to RankBrain and SEO, the best thing to do is to make sure you focus on producing great content. The better the quality, the bigger the reward, and you will score higher in search results. The focus is less on keywords, and more on topics and phrases. It shows that Google’s search algorithm itself is constantly learning and becoming more humanistic. It is, on the other hand, distancing itself from being more automated and robotic.
This means that Google has now adapted its interpretative technology. It has done so, to the point that keywords are only one single factor, out of the 100’s of SEO signals that are considered. And with the surge in mobile search terms becoming more varied and conversational, search results are becoming more tailored.
The question is no longer how we find our keywords. Rather, you should focus your efforts on solving questions, with specific answers plus, the intent of the query.
The best approach is to find topic opportunities, which means that you should make sure your website is answering the questions of your prospective customer who searches on Google. This also forces us to being more functional with our websites, as opposed to just focussing on keywords, and technical optimisation.
So, with everything stated above in mind, how can you successfully create content that will rank high in search results? Let’s investigate this process.
How can you create killer content with RankBrain in mind?
First, keep in mind that RankBrain is made specifically for long and ambiguous sentences. Do you want to test just how efficient this new algorithm is? It doesn’t matter if you type something as straight forward as “digital marketing”, or as ambiguous as “how can digital marketing help me?”, thanks to RankBrain, you will have the best results delivered to you within seconds. RankBrain turns complex sentences into helpful results. But what does that say for marketers? How should you approach writing content?
You should use the semantic SEO practices to help Google, and RankBrain, understand the context and meaning behind what you are writing about. This means that you should make sure that your site is optimised for searches by:
- Providing unique content.
- Making your site mobile friendly.
- Using keywords that associate with your business.
- Using hyperlinks to relate to other quality content.
- Making sure your website is responsive.
Most of those points are easy to understand. But let me explain a bit better about how you should write and organise your content. To get the best out of search engine, you will need to provide:
- Targeted content.
- Content that sounds natural.
- Include synonym keywords or topics.
Providing Targeted Content
The rules of traditional SEO stay the same. To rank high, you will need to create quality content, and to do so, start by asking yourself who your audience is? Who are you writing for? Are you targeting B2B or B2C? By creating different personas, you will be better able to write relevant content.
Not sure what a persona is? Think of it as the fictional image of your perfect client. Who is he? What are his interests? What will motivate him to buy? Or what will make him walk away? When establishing your persona, think about the type of problems your client may be looking to answer when searching online. Secondly, align your content with your persona’s needs. Lastly, since you know what motivates that perfect client, you will be able to use the right terms and words in your content. Let’s look at an example.
Utilise personas to create targeted content
You are a car dealership. You want to write content that will bring more traffic to your website. To do so, create your perfect client “Rob”. Once you have created Rob, you can analyse who Rob is, what motivates him to buy, and on the other hand, what factors will make him walk away. Once you have answered all of those questions, you can then write your strategic content, directly targeted to him.
RankBrain will know what Rob wants when he searches for a new car. This means that if you have the right terms, RankBrain results will match with your results.
But creating targeted content will not always suffice to get RankBrain to serve your site as a result. Especially, if your content is stuffed with keywords, and doesn’t sound natural. You can avoid this by using “you” for example, and telling your customers a story. Paint your customers a picture to help them understand. If you look at the image below, you will notice that the top ranking link makes use of this type of conversational style, that works well.
Natural Sounding Content
Over-optimising your page may have gotten you to rank higher before, but search engines are getting smarter. Nowadays, a page stuffed with keywords is more likely to get penalised, than anything else. Yes, keywords are still important to SEO, but a piece of content that flows naturally will rank much higher than one clearly written for search engines and not users.
Today, RankBrain will not necessarily match the search query with content including the exact wording of the query, especially since it deals with ambiguous searches that most likely will not have an exact match. However, it will match it with the most relevant article based on context. If you write your content with the user in mind, and answer their questions, for instance, naturally, keywords will match your user’s queries.
If I go back to the car dealership example, in the past, you could have stuffed your content with terms such as “buy new cars” and “buy used cars”. However, with RankBrain and other Google algorithms, you will want to answer user problems. “The new car you need that won’t break the bank” or “the perfect used car for off-road driving” etc. are good examples.
You will see that when you write content in a way that it makes customers feel like you’re talking to them, you will rank higher in search results. It’s all about being natural.
But while you shouldn’t overuse keywords, it doesn’t mean you should forget about them completely. However, gain points in the RankBrain book by focussing on synonym keyword and topics, rather than exact keywords.
Think about synonym keywords & topics
If you are writing content with the users in mind, you should naturally include keywords. However, make sure it doesn’t sound too repetitive and “stuffed”, and rather, use co-occurrence. What does that mean? Think of it as the frequency in which terms, or the synonyms of those terms, appear in a piece of content, as this is very important for SEO.
Again, imagine you are a car dealership. You may want to use the terms “affordable cars”, but also mention “cheapest cars” or “2018 affordable vehicle” etc.. Because RankBrain looks for context, even if the keywords are not the exact words in the query, if it is related, RankBrain will still serve your content to users.
Co-occurrence will not only help you to rank higher. It will also allow you to sound less redundant, and make your content sound more natural to users.
The bottom line is that you should target your content towards the users, while keeping RankBrain and other algorithms in mind. But, if you write to answer a particular need, you are bound to write quality content, that will eventually bring your content to the first results of search engines.
So, now that you understand what RankBrain is, and how to write content with it in mind, what do you think Google has in store for 2018? Let alone, in the years to come?
What does the future hold?
Google is clearly getting smarter about evaluating search queries. And with the RankBrain Google algorithm, it is also becoming more precise in making sure that only the best results are served to its users. So, what else has Google got planned for the future? Only time will tell, but be sure to subscribe to our blog, to stay posted!
