One of the hottest topics of discussion in the world of search marketing is Google’s RankBrain machine learning system. Google recently confirmed that it uses RankBrain in all search queries it handles (2+ trillion per year) and that it ranks anywhere from the most important to third most important factor in ranking results for a given query. Outside of those facts, there’s very little about the system that’s set in stone, at least to those of us working outside of the walls of Google. Consequently many SEOs are asking how to optimize for it, which is where the discussion seems to spiral into a whirlwind of hype and speculation. So how do we react to the increasing presence in the search landscape without falling victim to the hype and making a bigger deal of something that really might not be a big deal?
Sensationalizing Machine Learning
Thanks to pop culture sensationalizing artificial intelligence, machine learning carries a certain stigma that leaves many wild-eyed and wondering if it will take over the world. Movies like Eagle Eye and iRobot haven’t helped public perception, but even dramatic portrayals aside, it can be hard to define. There are many definitions for machine learning available, but I think this one from SAS describes it best:
“Machine learning is a method of data analysis that automates analytical model building. Using algorithms that iteratively learn from data, machine learning allows computers to find hidden insights without being explicitly programmed where to look.”
In other words, machine learning allows a computer to use observation to draw patterns and correlations over time without a computer scientist setting up the calculations for those correlations. So, no a machine learning isn’t scheming to take over the world. It’s just finding relationships between data points really, really fast.
How RankBrain Probably Works
In the case of RankBrain, Google is most likely using the system to better understand the relationship between linguistics and intent and use it to deliver relevant content much faster than a team of humans possibly could. For example, Rank Brain could understand that someone searching “best seats available Giants” is looking for tickets closest to the field at a Giants game, rather than furniture. It also might equate the search to a voice search like “I want to buy the best tickets at the Giants game.” It would then translate these searches into something very straightforward like “lower level tickets for sale at New York Giants game on September 18th, 2016” and correlate it to results for lower level or field-side seats on sites like StubHub, Ticketmaster or some other ticket vendor.
The Truth about Google
With that said, even Google doesn’t fully know how RankBrain operates. At the end of the day, it’s important to remember that Google makes the bulk of its money through advertising. In order to sell adverting at a premium price, it needs to continue to have a popular platform among searchers. To maintain that popularity, it has to be relevant and useful. RankBrain helps improve relevancy and usefulness. Therefore, the best (and maybe only) way to optimize for it remains the same as any other ranking factor: produce relevant, useful content. Yes, the hype will remain and some crackpot SEOs will try to use it to scare their customers or to position themselves as experts on a largely misunderstood topic.