How marketers are fighting back against Google's RankBrain and its disruption of the SEO industry

SEO is one of the most volatile aspects of digital marketing, but as marketers, we are obligated to follow the wave. Is there a way to go against it?

Google RankBrain search SEO
Thinkstock

When I look at internal documentation on why projects typically fail, I often find a similar answer: “the project failed due to requirements always changing.” For a CIO, these words are excruciating to hear, since you spend significant budget to provide your team with technology to mitigate these factors.

Of course, a carefully chosen technology stack, paired with well-planned requirements, is supposed to result in a successful outcome. There are many professions that have to get it right the first time, with little to no margin for error, such as civil engineering. Such fields place a huge emphasis on the planning stage, so that the execution stage will be smooth sailing.

Over the years, marketing teams have adopted this structured planning approach as well. Marketing teams aren’t usually given the luxury of do-overs; in fact, very little room is afforded to reverse the negative effects often associated with a failed marketing campaign.

So, when a new digital marketing medium known as search engine optimization (SEO) became popular in the mid 2000s, marketers planned their campaigns using the same slow, careful planning process.

Traditional SEO mechanisms fit perfectly with this approach. Google, knowing that its major weakness was exploitation of its algorithms, worked tirelessly to ensure that the process of restructuring its search engine results was as mysterious and slow as possible. To this day, it takes months before teams can verify whether or not a website change had a positive or negative effect on that website’s traffic.

But that all changed in late 2015. As reported by Bloomberg, Google’s new machine learning algorithm, known as RankBrain, was rolled out, and it presented a massive change to marketers. Instead of a static set of known algorithms, RankBrain machine-learned the best weightings of those algorithms, based on each keyword search, and then fed that back into its core system to adjust itself accordingly to those new weightings. And it did this daily. That meant Google’s search results could behave differently from one keyword to the next, and what worked for one keyword, niche, brand or industry may not work for the next.

Marcus Tober, founder and CTO of Searchmetrics, summed it up nicely in a recent quote in Search Engine Land, saying, “Google now looks at hundreds of ranking factors. RankBrain uses machine learning to combine many factors into one, which means factors are weighted differently for each query. That means it’s very likely that even Google’s engineers don’t know the exact composition of their highly complex algorithm.”

In short, RankBrain invalidated marketers’ old methodology of carefully planning a set of strategic moves based on a static understanding of Google’s algorithms, and we’re now in a brand new “wild west” era of SEO; something we haven’t seen since before the launch of Google’s Penguin and Panda algorithms dating back to early 2011.

For marketers, this has been a hard pill to swallow. All the experiences they’ve built up over the years, understanding models for controlling projects and executing those plans, are no longer true for SEO. The magical benefits of structured planning are hard to let go.

Uncomfortable truths

Google’s RankBrain is the ultimate “customer changing its requirements” nightmare for structured planning. How can your CMO plan and budget a 12-month project based on criteria that is guaranteed to have already changed, one month in? How do you mitigate risk and exert control in an unpredictable world?

Today’s digital marketers have had to come to grips with some uncomfortable truths:

  1. Using Google as a feedback mechanism is measured in months, not days. After each implementation, you won’t be able to get instant feedback on where you stand. It’s like deploying code, and then having to wait a few months to know if there are any bugs. For CIOs and CMOs, there is a tremendous amount of risk associated with the length of the current testing cycle. You may have the latest and greatest tools, but if your team can’t course correct fast enough, the best tools in the world won’t help.
  2. Google is a black box. You can’t know, for sure, what effect any changes are going to have. The granularity of the data you can get from Google stops at ranking positions for each keyword. There’s no way to find out why search results rank the way they do, or how much more optimization is needed to change ranking positions.
  3. The CMO may not be the right person for the job. As discussed above, structured planning methodologies do not work in unpredictable environments. This one is insidious, since your CMO is likely to defend his/her stance with past examples, to the very end. It may be time for the CIO to own SEO.

However, new technology is attempting to plug the hole created by RankBrain in an interesting way; by machine-learning its machine-learning algorithm in order to allow iterative implementation website changes designed to improve SEO without as much uncertainty.

Beating Google at its own game

Back in 2006, a small, relatively unknown team, led by Carnegie Mellon alumni Scott Stouffer, and backed by former Yahoo! Search Executive Larry Cornett, correctly foresaw these events transpiring. As reported by Search Engine Land, Stouffer, who was at that time an experienced software developer, realized that “eventually machine learning would be deployed within Google’s scoring process. Once that happened, we knew that the algorithmic filters would no longer be a static set of SEO rules. The search engine would be smart enough to adjust itself based on machine learning what worked best for users in the past. So we created Market Brew, which essentially serves to ‘machine learn the machine learner.’”

Market Brew eventually became the team’s commercial product after years of research and development. They developed a standardized model, based off of years of iterative regression testing, which behaved like a “generic” search engine.

When they paired this with an artificial intelligence technique called Particle Swarm Optimization, as reported by TechCrunch, this “self-calibrating” generic search engine model could adapt itself to the target search engine environment, essentially taking on the behavior and characteristics of that environment.

Once calibrated, the generic model became a Google-like statistical model, where SEO teams could make changes on their site (or test site), and confirm the effects of those changes in a very rapid testing cycle.

This has the potential to, again, change the entire landscape of SEO, primarily for the following reasons:

  1. Google will no longer be black box. SEO teams have had to rely on data directly from Google. Whatever enterprise SEO tool they used, ultimately it was based on scraping Google’s (black box) search results for data. Now, these teams could have the opportunity to build a statistical model, and dive deeper into the search engine model, to fully understand the interrelating mechanisms that caused differences in rankings.
  2. The rapid (iterative) testing cycle for SEO has been born. Marketers have had to wait months (for Google to process) before learning of the true SEO results of their website changes. However, if a reliable model can be referenced, marketers can test them against any changes, and know with a high degree of statistical certainty whether the changes will have a net positive or net negative effect. Of course, the teams can then confirm the predictions a few months later when Google reflects those changes in its ranking algorithm.

Conclusion

Conventional SEO tools have long predicted the effectiveness of changes made to a website by comparing those changes to known best practices. However, best practices became a moving target when Google’s RankBrain algorithm launched. We’re starting to see the emergence of technologies aimed at hitting this moving target with reliable accuracy, and it’s an intriguing development in the SEO and digital marketing industry.

Take a look at your current planning methodologies. Is your CMO planning your SEO campaign in 12-month cycles? That won’t work anymore. Does your digital marketing team have the ability to tell you what their changes are statistically likely to do? Can they measure the likely outcome at any time? If the answer is no, pay attention to the development of this new technology; it has exciting potential to give power back to marketers that they lost to Google years ago.

This article is published as part of the IDG Contributor Network. Want to Join?

SUBSCRIBE! Get the best of CIO delivered to your email inbox.