Ever come across a website in the search results and thought by yourself:
How can I rank for this?
We have known for a while that Google is developing hyper quickly. The past few years they have been developing themselves in particular in the field of language:
They are not only showing results matching your exact search query, but Google also understands their grammatical conjugations, plurals and lets you score for synonyms of a search query.
Powerful work from Google
So far we, mere mortals, can understand Google’s algorithms and behavior. Seeing a grammatical conjugation or a synonym popping up in the search results is something that fits in our own thought pattern.
At one point maybe you thought to yourself:
”Gosh, clever search engine”
Google however has been taking leaps recently. They are now using Machine Learning to understand, rate and rank your website in the search results.
What is Machine Learning?
Machine Learning means that computers are programmed in a way they can teach themselves new stuff and are able to improve themselves.
Artificial Intelligence has been studied for years for all kinds of applications (mobile phones, cars, robots, utilities on a computer, etc…) and we see the results of these studies everywhere around us (think about Siri on an iPhone, automatic screen brightness, assistance robots in care, etc…).
No wonder that Google as well started with it to see how they can take their own search engine algorithm to the next level by using Machine Learning. Google is now not only looking which words are on your website, but is also beginning to understand what these words and articles mean.
By understanding the meaning of a text, they are able to show their ‘customers’ even better and more relevant search results (no matter which words you exactly use on your website).
This new ‘Context‘ (vs. Content) era has been announced for years with the rolling out of their ‘Hummingbird‘ update, but with the developments we now see with Machine Learning they will be able to take this to whole new levels very quickly.
Back to the basics of SEO
Before I dive into the story about machine learning and Google, I want to take a moment to show you how Google used to rank your website for years:
Google has pointed out hundreds of factors that are of influence to the ranking of your website in the search engines. This basically means your website is split up into hundreds of pieces.
Then all of these factors are incorporated into a Google algorithm. This algorithm is developed by Google engineers based on these hundreds of ranking factors.
Each of these factors has it’s own value in the eyes of Google, and by combining all of these factors, Google is able to rate your website and compile the search results.
Optimize your pages to get higher rankings in Google?
Check Our Über SEO Checklist >>>
Next Level SEO: Rankbrain
Google is now beginning to use Machine Learning in their search engine algorithms. This piece of technology they are using and developing is called Rankbrain.
Rankbrain works differently from the way Google started out ranking websites in the search results. Rankbrain does not look at the hundreds of predefined factors to then rate them using an algorithm.
Google Rankbrain works as follows: Employees from Google have determined the ”Training Data”. These data contain the best and most popular websites and search results, the ones that are doing well in Google.
Then, rankbrains goes out to take a look at these websites. Based on those websites, Rankbrain builds it’s own Machine Learning Algorithms. Even Google’s software engineers don’t know how exactly this algorithm is developed or what Rankbrain is “thinking”. This algorithm could just be containing the basic factors once predefined by humans, but there’s no way to be sure.
Rankbrain now looks at a website as a whole, instead of splitting it up into hundreds of different factors connected to each other. Your whole website and what is happening around it is now being screened by Rankbrain’s algorithm at once, and based on how it’s rated your website will get it’s rank in the search results.
It is important to know that there is absolutely no human influence involved in this search algorithm. Everything is done automatically and happens based on Machine Learning technology. Rankbrain gets Training Data, and then builds it’s own algorithm to rate and rank websites. This is then evaluated and the process is repeated.
So Rankbrain is continuously getting better in rating valuable websites and compiling the search results.
What are the important factors to score in the future?
It’s not easy to put a finger on that one, but you can bet it will be a full assessment of your page, including all external influences.
You could then think of, inter alia, the following:
- The variation in your content (multimedia)
- Quality of your content
- Is your content answering questions
- Time on a page
- (Relevant) backlinks
- Behavior of a visitor
- The degree of positivity / negativity in news about your business
- Engagement (shares, but also comments on your pages)
- SEOcial value (mentions of your company on social media, likes, shares etc.)
Rankbrain will however look at your website and pages as a whole. Will your content and the meaning of your website or page be the very best and most valuable to a visitor? Then Rankbrain will certainly take this into account when rating your website.
A rule of thumb you can apply in order to accomplish this (this applies to everything you do to your website) is the following:
Are you offering quality on your website? Are you helping out your visitors? Are you inspiring them and giving them something they can work with, or are you keeping them interested? Then you are doing a good job!
Or are you writing lengthy articles, so you might score better in Google? 😉
My advice is to do the latter. Because when you are valuable to your visitor, you can rest assured you will be the same to Google.
Because of the recent developments and future developments to follow, you will have to spend less time on hard SEO, but more on creating value.
The structure (technique and readability) of your website will remain an important aspect off course, but if you are working with a good website system or a professional website design company you will not have to worry about that.
How is Google doing this again?
I can imagine you thinking: “How is Google doing this again? They already know so much, and now they have machines improving themselves?”
That’s right 😉 all to make sure they present their users (customers) with only the very best results. The technique Google is using to do so, they didn’t develop. Keeping in mind the motto “better buy good, then build bad” they purchased the company ‘Deepmind’ for half a million Dollars in 2014.
Deepmind was no search engine giant or something like that, but developed machines that were improving themselves to get better at a game. How that worked can be seen here:
What’s cool to see is that the computer not only “learns” to catch the ball every time, but is also learning itself a strategy to play out the level as quickly as possible. As you can see, we are really talking about the so called Machine Intelligence.
At this moment Rankbrains rating is third most important factor in the current Google algorithm.
I think it is safe to say that this will become the most important factor in the very near future, and given the speed of developments and growth in the field of Machine Learning I even dare to say that Rankbrain will in time replace the entire search algorithm.
Exponential growth of Artificial Intelligence
Because it is not humans, but machines now “improving” machines, the development in the field of Machine Learning speeds up. But what are the consequences of all of this?
The intelligence of a computer at this time is in different areas comparable with that of a chimpanzee, and that to people is often interesting or fun:
But in essence there is not a whole lot of difference between the intelligence of a chimpanzee, a human, and a super smart person like Einstein.
Because the computers are able to continuously improve themselves, it will also be the same in terms of intelligence, so no increasing linear line, but a line that grows exponentially:
You can therefore assume that it is not very long before it will take computers to outsmart humans (experts expect that this issue will be reached by 2020).
Cute or Scary?
And what impact will this entail? We humans in fact look very different to the world then machines:
Source: Things that Make us Smart – Donald Norman
Famous persons from the IT industry, like Bill Gates and Andy Rubin (Android), have said to even be a little scared of these developments.
Machines have no moral or feelings (unless it had been programmed), and there’s no way to know how far they will go to reach a preprogrammed target (would it be possible that a small mistake in coding make computers taking over the world? 😉 ).
Personally this doesn’t scare me, I think these developments will produce a lot of beautiful new things in the future. We as people and entrepeneurs will be relieved from ‘tedious’ tasks, and will be even better supported in getting in touch with other like-minded people.
Google has again made a nice step forward in getting there with Rank Brain.
What do you think of the new developments at Google? Will you go to work to make your site even more valuable, or are you going to prepare for a hostile takeover of the machines in our world? 😉