Ask Google Anything: Gary Illyes Talks RankBrain, Panda, Penguin and More
The Ask Me Anything session with Google Search is always an SMX highlight. The audience is full of digital marketers eagerly waiting to hear what Google Webmaster Trends Analyst Gary Illyes will reveal in the AMA with Search Engine Land and Marketing Land Editor Danny Sullivan. Read on for this Googler’s statements on RankBrain, Google Assistant, Penguin, Panda and more!
— Natasha Aidinyantz (@MyCreative_UK) June 23, 2016
Google says that RankBrain is the third most important ranking factor; is it really a ranking factor or a query alignment tool?
Gary: Basically, it’s a ranking factor. It’s a supervised machine learning platform. RankBrain is the same thing, it’s learning what worked well for queries and what results are good for a certain query. It works well for long-tail queries. He gave the example query of “can I beat Mario Bros. without using a walk through”
With RankBrain, the results are more reasonable than without it.
Is it more query refinement and less ranking?
Gary: It will understand what results will work better for queries. Stop words are dropped from queries but RankBrain understands the conversational manner of search and can give results accordingly. It’s less about understanding the query and more about understanding how best to score the results.
Do we have a RankBrain score?
Gary: We don’t have a RankBrain score. The root of your question is if you can optimize for RankBrain.
RankBrain is a new, unique score. It’s enhancing our relevancy in the search results based on what you have on your pages. It makes sure the user gets the best page for the query. If you keyword stuff your content, it will almost certainly not be good for you.
How many queries is RankBrain processing?
Gary: Not sure; RankBrain will refine and learn from every query it gets.
Should we fear that it’s going to take over the world? Most of the AI movies show us that the machines will take over …
Gary: Not all of the world.
[insert audience laughter]
Is the number one ranking factor content or links? Which is the most important ranking factor?
Gary: It depends on the query, what you’re looking from and what the numbers say. I can’t give you a concrete answer because it depends on too many things.
What’s the deal with Google Assistant?
Gary: Frankly, I have no idea. I know we’re still wrapping our heads around how to experiment with this new cool idea. It’s based on machine learning. We need to know what you want, how you want it and where you want it.
How much of the algorithm is going to become AI based?
Gary: Machine learning is extremely important for us. We’re focusing on machine learning and what it can do — not just in Google Search but in all of our products. At what stage are we? I can’t tell you because I only have a vague idea. We don’t want to get to a stage where someone sends us a bad query or a query that we have bad results and we don’t want to get to the point where we don’t understand why the machine gave that result.
Do terms in the URL help in any way with rankings?
Gary: TLDs do not play a role in how we calculate relevancy for a specific result. ccTLDs to play a role in ranking, you’ll perform better in the google local. There are certain cases where we will look at it but in most cases we won’t. I’m not advocating to buy keyword rich domains; it doesn’t have a super power. My recommendation is that it can definitely help you if you’re describing your product in the URL because it helps your user.
Google said it was going to give you more data (more than the current 90 day view) in Search Console almost three years ago; when is this going to happen?
Gary: We went through a very long transition with Search Console. We’ve figured out how to we can make a longer time of data happen and is something we’re working towards a little faster (and we have the buying from management).
Let’s move on to Penguin … are we going to get a new one anytime soon?
Gary: Penguin, I will not say a date because I’ve been wrong too many times. I will not say any time frame anymore.
What’s the deal with Panda? You said it was part of the core algorithm but is stuff running through it constantly?
Gary: It is not real time but it is continuously running. We collect the data and then roll it out, refresh that data and roll it out again.
What’s our time period on each roll-out?
So it takes months to get through all the sites on the web?
At this point, unfortunately, my computer died >.< There was one more key piece of information coming out of this Google AMA, though — the official end of Google Authorship.
— MichelleRobbins (@MichelleRobbins) June 23, 2016