Right now, all across the world, thousands of SEOs are busy working on link profiles, canonicalization, robots.txt, meta tags, H1 optimization, and a host of other SEO tasks. But every one of these SEOs is going to experience a major shift in the industry.
This major SEO shift won’t show up on Moz’s algorithm tracker. It probably won’t impact search rankings. And it may not make headlines like the mobile update did.
What is this gigantic and mysterious force?
It’s machine learning.
What is Machine Learning?
Machine learning is the science of self-programming computers. According to WhatIs.com:
Machine learning is a type of artificial intelligence (AI) that provides computers with the ability to learn without being explicitly programmed. Machine learning focuses on the development of computer programs that can teach themselves to grow and change when exposed to new data.
So, machine learning is the science of getting computers to act “without being explicitly programmed”.
But there is still a lot of programming behind getting computers to program themselves. Computers and their applications are being designed for change and adaptability. Programmers want their applications to respond to more than just intentional input.
Applications should be able to respond to external stimuli, big data, crowdsourced social activity, and a legion of other sources.
What are some applications of machine learning?
Machine learning is showing up in more areas than we might realize. Right now, on a roadway somewhere, a self-driving car is making right-hand turns (using its signal), braking (gently), and pausing for pedestrians (considerately).
The self-driving car is a prime example of machine learning put into motion, literally.
The self-driving car is heavily programmed, but not by itself. A small army of very intelligent people have spent a long time creating virtual reality maps, developing vehicular adaptations, and forming a new industry in order to help the car drive safely and reliably.
Yet, at the same time, the car does program itself.
The job of the [car’s] software is to figure out how the world is different from that expectation.
As the car figures it out, it adapts.
But machine learning isn’t usually as sexy as most of its applications. Right now, in a Facebook newsfeed near you, machine learning is controlling what you do, see, and interact with.
During Facebook’s tumultuous 2011, they switched from an algorithm called EdgeRank to a more complicated one. The machine learning of Facebook’s advanced newsfeed algorithm tries to individualize your Facebook experience based on what it thinks you want. You teach it with every linger, look, click, query, and interaction that you perform.
Here’s how Time explains it:
To ensure that those 300 posts are more interesting than all the rest, Facebook says it uses thousands of factors to determine what shows up in any individual user’s feed. The biggest influences are pretty obvious. How close you are to a person is an increasingly important metric, as judged by how often you like their posts, write on their Timeline, click through their photos or talk with them on Messenger, Facebook’s chat service. The post-type is also a big factor, as Facebook hopes to show more links to people who click lots of links, more videos to people who watch lots of videos and so forth. The algorithm also assumes that content that has attracted a lot of engagement has wide appeal and will place it in more people’s feeds.
When you clicked on Joe’s picture, or searched for “Joe B—” in your search bar, the machine-learning algorithm picked up on it. Tomorrow morning, when you open up Facebook on your phone, guess who’s updated profile picture will top your newsfeed?
It’s not quite that starkly cause-and-effect, but the principle remains true. Facebook’s newsfeed algorithm operates on machine learning.
Machine learning is expanding everywhere. Although Google’s machine learning is focused on search improvement, they also advance machine learning in a whole breadth of applications.
To say that machine learning is changing SEO is a bit anachronistic. Why? Because machine learning is already a major part of SEO.
In fact, it has been for a long time.
Right now, however, it’s growing in importance.
How is Machine Learning Changing SEO?
Like the other disparate niches and sub-niches of computer science and marketing, machine learning doesn’t seem to have much to do with SEO.
SEO, however, defies boundaries. What was once the realm of keyword-stuffing, basement-dwelling geeks is now the domain of leading executives, global headline news, and industry-disrupting trends. SEO is democratized and omnipresent. Besides, it has this uncanny ability to worm its way into just about every industry and discipline.
Machine learning has already changed SEO in massive and irreversible ways. Right now, machine learning is the third most important factor in Google’s algorithm: RankBrain.
Rank Brain is Google’s own algorithmic learning system that adapts to queries based on interpretation rather than direct input. It’s like Hummingbird, but more adaptive.
Currently, Rank Brain is the third most important signal in the Google algorithm. As Bloomberg reported:
RankBrain is one of the “hundreds” of signals that go into an algorithm that determines what results appear on a Google search page and where they are ranked, Corrado said. In the few months it has been deployed, RankBrain has become the third-most important signal contributing to the result of a search query, he said.
But it’s hard to quantify RankBrain’s impact to percentages and fractions. Like any artificial intelligence system, RankBrain adapts to input and responds to forces on an automatic basis.
The Verge explained that Rank Brain is used to “answer ambiguous questions,” but that’s only a small part of what Rank Brain does and is. Until recently, Google’s algorithm operated on signals — more than 200 of them. With Rank Brain, it operates by learning.
Here’s how Rand Fishkin traces out the new algorithmic learning process:
At present, we’re only seeing an emerging trend, not a seismic force. But bigger change is coming. As Greg Corrado, a senior research scientist with Google, said, “It’s gonna get weird.”
What Should SEOs Do About Machine Learning?
Machine learning has been lurking in the background of SEO for a long time. Now, we’re finally seeing it grow into a larger and more explicit force.
Expect Search to Change Gradually
SEO has experienced enormous change in the past few years. To call any of it “gradual” seems like a joke. However, we are now entering an arena of greater adaptation and less abrupt changes. Google’s modus operandi has been to change the algorithm suddenly and without warning.
The result of the change was that it wrecked businesses and ruined rankings. With machine learning at the helm of the algorithm, however, change will happen on an adaptive basis. No longer will engineers conceive of a new iteration, code it up, and roll it out. Instead, the artificial intelligence will drive the change by response and process.
Keep Optimizing for Humans
At the crux of the issue behind machine learning is this: Humans reign. It’s tempting to see artificial intelligence as some massive force that will develop a mind of its own and crush our civilization under its electronic dominion.
That is the stuff of science fiction. It’s called machine learning because the machines learn, not merely from abstract environmental forces, but from the behavior of humans.
To be specific, Google’s RankBrain is constantly learning from the massive influx of searches input by humans.
In a Quora discussion on the subject, one respondent wrote this:
The search ranking system has a very subjective goal – user happiness. CTR, query volume, etc., are very inexact metrics for this goal, especially on the fringes (i.e. query terms that are low-volume/volatile). While much of the decisioning can be automated, there are still lots of decisions that need human intuition.
That “human intuition” is provided by the human search queries and being learned by RankBrain.
When we look up from the complexity of machine learning, we’re struck with an obvious fact. We just need to keep optimizing for people.
Forget the algorithm! It’s going to learn, to adapt, to morph, and to improve, but it’s going to do so based on what it learns from humans. So, as long as you optimize for humans, you’re going to be okay.
Optimizing for humans is a notoriously slippery concept, however. Here are some suggestions for SEOs:
- Work in concert with UXDs (user experience designers) and IAs (information architects).
- Conduct user testing on application interfaces and websites.
- Work hard at creating engaging content.
- Track user interaction with your website (queries, CTRs, dwell time, etc.) in order to tweak it for optimal experience.
Machine learning is a powerful force in SEO, and it’s only going to get bigger.
Google has conclusively proven that machine learning is ruling and will rule the coming changes to the algorithm. There is no cause for panic, nor is there any reason to drastically change what we’re already doing, as long as we’re working hard to optimize for humans.
People are the most important factor in search. They always have been, and they always will be. As long as we’re creating content and experiences that connect deeply with humans, our search engine optimization efforts will be successful.
How do you see machine learning changing SEO?