Late in 2015, Google sent the search engine optimization community for a bit of a loop when it announced that it had been working on something called “RankBrain” that would be added to its overall Hummingbird SEO algorithm. The reason this announcement came as something of a surprise to the SEO community is that RankBrain redefined what it meant to process websites and achieve rankings. Unlike the other, earlier components of the Hummingbird algorithm, RankBrain is a machine learning mechanism that can, in theory, “teach” itself about websites on the fly. Over time, this could have important impacts on SEO and best practices.
So, How Does RankBrain Work?
It’s important to understand that RankBrain is capable of making logical decisions on behalf of users based on what it already knows, as well as learning new things that will inform which results are displayed by Google in the future. This is the definition of “machine learning.” In practice, that means RankBrain lets the search algorithm “think” about what to present to a user.
One of the best examples touted by Google engineers during the launch period of the system concerned a search for measurements. A user in the United States would search for how many tablespoons were in a cup, and they would be presented with American measurements to answer this query. In Australia, an identical search would lead to return the slightly different Australian measurements to answer this question instead.
RankBrain has learned that measurements differ in each country, and this has impacted search results by presenting more regionally accurate information the first time around. In fact, tests by Google engineers show that RankBrain provides the contextually appropriate result 80 percent of the time, while its own engineering team could only predict and provide the contextually appropriate result about 70 percent of the time.
Is This a Major Update of the Algorithm?
Not necessarily. RankBrain is another component of the Hummingbird ranking algorithm, which already contains and interprets hundreds of signals as it determines where, why, and how to rank a given website for a given keyword search. Among these hundreds of signals, Google notes that RB is currently the third most important. It’s certainly something to consider and keep in mind, but the addition of RB to the Hummingbird algorithm shouldn’t immediately stir up fear or lead to a major upheaval in search result rankings anytime soon.
It’s impossible to quantify the change to a website’s search engine performance based on one additional factor in an already complex algorithm, but there’s something to be said for properly targeted keywords. The best way to stay above the fray, and avoid frantic chaos with each new SEO technology unveiled by Google, is to adopt best practices. Whether it’s RankBrain or another algorithm change, keep in mind a few fundamentals:
- Utilize a series of well-defined keywords that the target audience will actively use when searching for that website’s content or products.
- Don’t target customers that are outside the website’s content area, region, or basic interests. This may work temporarily, but algorithm and machine learning developments will eventually destroy such an ineffective SEO implementation.
- Keep a website’s content authoritative, keyword-relevant, and within best practice guidelines for SEO content. This will minimize the impact of ongoing evolution in the Hummingbird algorithm.
Machine Learning: An Evolution, Not a Revolution
In terms of Google’s Hummingbird algorithm, the machine learning that RankBrain does is currently evolutionary, not revolutionary. RankBrain is one of hundreds of signals used to identify and rank pages based on keyword relevance, and it’s currently third in the list of algorithm tools used to rank websites. Awareness, but not panic, is key to understanding and abiding by the terms of machine learning in this context.