July 7, 2016
Google Now Using RankBrain for All Queries
“Machine learning is a core, transformative way by which we’re rethinking how we’re doing everything.” –Google CEO Sundar Pichai 1
Google’s machine learning artificial intelligence system, RankBrain, has gone from processing fifteen percent of searches last year to one hundred percent of searches currently. Active since April 2015, RankBrain is now considered the third most important ranking signal after links and content. 2 Don’t worry digital marketers: you don’t have to become machine-learning ninjas now that Google is using the technology to process every single search query—though its engineers may well be expected to.
Jeff Dean, Senior Fellow in the research group in charge of the company’s machine intelligence deep learning team, Google Brain, noted that while ‘Rankbrain is involved in every query,’ it doesn’t affect all rankings—but ‘a lot’ of them.
What Is Machine Learning?
We’ve increasingly heard the phrase in the world of SEO, but what is machine learning? A form of artificial intelligence, machine learning occurs when engineers use algorithms and data to “‘teach software to accomplish” tasks on its own rather than relying on them to continuously program it. Compared to older methods of programming dependent on thousands of lines of static code, machine learning is a much more dynamic model requiring frequent modification.
Google has been experimenting with machine learning for years, incorporating the technology into products like Gmail’s Smart Reply before discovering in 2014 it could help its search engine return more relevant results based on user engagement. RankBrain went live the following year.
How Could RankBrain Influence SERPs?
RankBrain could help Google’s search engine uncover user intent for unique queries. As of 2013, fifteen percent of the three billion searches the company processed every day were completely novel: no one had searched for them before—at least not on this particular search engine.
Google’s search refinement tools prior to RankBrain included stemming and synonym lists and databases describing related phenomena for search engines, all of which, at least initially, required human labor. RankBrain is designed to identify patterns in complex and seemingly unrelated queries, infer similarities, and continuously build on this knowledge to help it interpret future rare or unprecedented searches. 3
Once the intent of a search is uncovered, RankBrain may associate it with a more popular query Google already has user data for and return results for that query instead; for example, when you search for “best tacos in Long Beach,” RankBrain might show results for “best Mexican food in Long Beach.”
RankBrain could help Google provide relevant, high quality results for the large chunk of queries its search engine has never seen before.
How Does RankBrain Impact You?
Since the engineers at Google are still experimenting with RankBrain, SEOs and content marketers are unlikely to see definitive guidelines anytime soon.
While Google’s Senior Vice President of Search, John Giannandrea, believes machine learning will transform everything from vehicles to medicine to, well, humanity, RankBrain is ultimately designed, for now, to reward excellent content. The recommendation for digital marketers echoes that of years past: focus on creating outstanding content.
Additional Sources:
1 – https://backchannel.com/how-google-is-remaking-itself-as-a-machine-learning-first-company-ada63defcb70#.bby98xiwg
2 – http://searchengineland.com/google-loves-rankbrain-uses-for-every-search-252526
3 – http://searchengineland.com/faq-all-about-the-new-google-rankbrain-algorithm-234440
