The A.I. Battle for SEO Supremacy

In the race for the top of the search results, many have turned to the power of artificial intelligence

If you’re reading this right now, chances are you’re playing a part in the learning process of a gigantic AI neural network. Don’t be alarmed! Unless you’re hiding behind several layers of encrypted networks, you’re like most of us: just surfing the web.

Nonetheless, the fact remains: Every online search, share, purchase and send we engage in feeds a morsel of info, byte by byte, to the perpetually hungry AI networks controlled by Google, Facebook and Amazon. And that info is sold to countless thousands of corporations looking to sell you products.

From a utilitarian perspective, what we’ve (mostly) willingly sacrificed in privacy we more than reap in convenience. We’ve replaced sales people in favor of virtual stylists, capable of algorithmically distilling our idiosyncrasies into the law of averages before supplying us with tasteful, abeit mass-appeal, wardrobe recommendations. We’ve ceded roadway companionship to watchful automatons with the contingent hope that we’ll never have to make awkward small talk with a driver again. We even let Google step in to finish our sentences. Well, regardless of how much hyperbole you choose to employ in couching the terms of the robot revolution, one thing appears certain: Artificial intelligence and machine learning will soon be eliminating a lot of jobs.

A.I. and SEO

In the world of SEO experts, this certainly looks to be the case. The evolving, self-learning network behind Google that obscures their search algorithm from rudimentary understanding gave birth to the SEO profession—a cadre of specialists who made a living teaching you how to get noticed more in Google searches. But with the advent of ever-changing formulas, updates and self-correcting tweaks that now comprise the fluctuating Google search labyrinth, we’ll soon be forced to navigate unaided by these search pros.

Those formulas weren’t always so obfuscated. Prior to 2013, Google operated on a set series of rules that determined search rankings. Rules such as don’t make your meta-description too long, include H1 tags, provide this much content, yada, yada, yada. These standards still matter, but over time new rules have been layered into the search ranking mix of variables, all manipulated and massaged into increasing complexity by second-order derivatives and other mathematical acrobatics. Consequently, instead of trying to reverse engineer search algorithms based on known rules, SEO specialists began resorting to A/B testing in order to make educated guesses about what factors the algorithm was favoring based on a per keyword basis. In other words, pure cause and effect observation and experimentation.

Specifically, after 2013 the odds were stacked against SEO professionals with the roll out of the Panda and Penguin updates. This series of AI-driven improvements aimed to make Google a more optimal search engine less prone to trickery and exploitation, and it pretty much worked.

RankBrain Arrives

Some thought this signaled the death of SEO as a profession. Yet the industry lived on, only to suffer another AI-delivered blow in 2015. That’s when Google incorporated RankBrain, a complete machine learning AI system to run their search engine metrics. In the words of Google senior research scientist, Greg Corrado:

“RankBrain uses artificial intelligence to embed vast amounts of written language into mathematical entities – called vectors…In the few months it has been deployed, RankBrain has become the third most important signal contributing to the result of search query.” 

In the face of the growing adversity brought about by AI-generated complexity, SEO experts retreated into a mixture of white-hat SEO tactics such as organic link building and content pillowing with a smattering of black hat here and there, aka PBN’s, page swapping, and in a few laggard circles, keyword stuffing. Arcane stuff. Gimmickry. Borderline necromancy.

Fighting Fire with Fire

Is there hope for a once proud profession? Some SEO gurus are preparing to fight fire with fire. Scott Stouffer, Carnegie Mellon alumni and founder and CTA of MarketBrew.com, argues that the future of SEO won’t be in retroacting, but recreating. Essentially, futurists in Stouffer’s camp believe that in order to understand Google’s search algorithm, SEO professionals need to use a machine learning model of their own to reverse engineer the algorithm from the ground up. In other words, use the machines against themselves in what is basically analogous to the plot of Terminator 2: Judgement Day.

Okay, let me try and explain it better:

If you play Super Mario Brothers enough, your brain will gradually learn the way the game engine itself operates—how fast enemies move, how high Mario can jump, and so on. And, by creating a facsimile of the game engine within your own brain and acting in accordance to its rules, you can start to outsmart those rules and win the game. If we translate this to SEO speak, if you build an AI system capable of asking Google one billion queries and learning how it responds, you can conjure up a replica algorithm to mirror Google’s. At least that seems to be the premise behind a number of SEO companies such as MarketBrew and RankScience.

If what these businesses claim is factual, then they have essentially atomized the SEO industry to its bare essentials to rebuild using a refined set of code as the foundation. Plug in your website, and with the minor human touch of software engineers to keep it running, SEO experts to twiddle and tweak when necessary, and traditional marketers to window dress it, you’ll wind up with your own mathematically optimized website. Competition in the SEO industry has started to shift more in this direction, with more investments in new AI technologies such as genetic algorithms and particle swarm optimization to better model the exact nature of Google’s RankBrain functions. Yeah, for marketers it’s that important.

So, the question becomes: Can data scientists and engineers rig an effective algorithm capable of outperforming traditional methods of gaming a system purpose-built to stymie would-be SEO wizards? Because if they can, we may see employment in the SEO industry decimated to bare bones levels, ironically enough, another victim of the robot revolution.

For more inside-the-industry know how and analysis of marketing trends, check out our Sonar Blog.