RankBrain: SEO friend or foe?

Humans recognize patterns well, and the humans who work at Google have recognized that keywords, backlinks, domain ages, title tags, and meta descriptions are all great factors that can be used to sort and rank websites.

Yet recognizing such patterns requires gathering a whole lot of data from which to learn, something humans aren’t so great at.

bluebrain-4

Machines, conversely, are great at gathering data, but they’re not so good at recognizing patterns, figuring out how those patterns fit into the big picture, and understanding what they mean. Basically, they can’t read the writing on the wall the way humans can.

So computers and people need to work together in order to perform the complicated process of information gathering and analysis that lies underneath a search engine’s results. Wouldn’t it be great if somehow, the two could simply be fused together?

Enter RankBrain, Google’s machine-learning artificial intelligence that processes a “very large fraction” of search results each day, the AI that not only gathers data, but also sees the patterns therein.

Hi, RankBrain. what are you?

RankBrain has not usurped PageRank’s throne. Rather, it’s a part of the bigger algorithm that takes search queries, interprets what the user is searching for, and figures out how to submit that request in new ways.

robothand

Take the search term “Matt Cutts,” for example. A few years ago, Google and other search engines would have looked for pages matching that exact term. Nowadays, Google’s algorithm understands who “Matt Cutts” is a lot better.

Searching “Matt Cutts” might also grab pages and information that match “Google spam team,” “SEO,” or even “cookie guy.” (If you don’t know why that result would come up for Matt Cutts, look it up. It’s a funny story.)

In short, Google can now do a better job of recognizing the relationships between words — what they mean, what concepts they reference, and why they’ve been strung together to make a search term. RankBrain deepens this understanding.

According to Bloomberg, it “uses artificial intelligence to embed vast amounts of written language into mathematical entities — called vectors — that the computer can understand. If RankBrain sees a word or phrase it isn’t familiar with, the machine can make a guess as to what words or phrases might have a similar meaning and filter the result accordingly, making it more effective at handling never-before-seen search queries.”

Think of RankBrain like a clerk at Bed Bath and Beyond. If you walk into the store and say that you’re looking for “that plastic doohickey that squeezes the juice from a lemon without getting any seeds in the vinaigrette,” the clerk — or the AI — will synthesize that string of words and respond, “Oh, a citrus squeezer? Those are in Aisle 11; follow me.”

RankBrain? Don’t you mean SkyNet? Or Agent Smith?

Nope. RankBrain is not sentient. Saying that RankBrain “learns” doesn’t mean that you could sit down with it, show it the basic principles of algebra, and set it loose to solve equations for X.

Instead, it is fed quantities of historical searches and their results, and extrapolates this information to make predictions about future searches.

robotchess-4

If these predictions are deemed to hit the mark with relative accuracy, then the humans who engineer RankBrain release its latest version. To continue the math analogy, it’s somewhat like a student who takes multiple practice SATs to ensure she can score well before sitting down to the real thing.

And RankBrain has proven that it can score really well.

Google’s search engineers, who craft the algorithms that underpin the search software, were asked to take a look at some search results and guess what they thought would be ranked on top. They did pretty well, choosing the correct ones 70% of the time. RankBrain, however, was right 80% of the time.

In fact, Google found that turning off RankBrain “would be as damaging to users as forgetting to serve half the pages on Wikipedia,” Greg Corrado, a senior research scientist at Google, told Bloomberg.

Okay, but what will RankBrain do to SEO? Is it going to kick over our sandcastles?

Considering the fact that Corrado told Bloomberg that RankBrain is the “third-most important signal contributing to the result of a search query,” it’s logical to say that it can and does affect SEO.

Although Google began rolling out RankBrain a few months ago, it hasn’t had the same effect that Panda or Penguin did.

If anything, RankBrain has probably helped SEO by pulling in more relevant search factors. Where a local pizzeria may not have ranked as well in some local results pages before, it may now be earning a top slot, thanks to RankBrain’s intuition.

For now, RankBrain is less like HAL 9000 or the Terminator T-800, and more like Johnny 5 or Wall-E: a friend, not an enemy.

My (thoroughly human) prediction is that smart SEO agencies will begin to experiment, particularly when it comes to small businesses and local search, to find out if they can pull back the curtain on RankBrain and learn what makes it tick — thereby increasing their clients’ chances of movin’ on up in SERPs.

Related reading

Super food diet selection in wooden bowls. High in antioxidants, vitamins, minerals and anthocyanins.
cma feature
Search Console Search Analytics
i_fought_the_law
Simple Share Buttons