Exclusive Q&A with Google’s Gary Illyes at BrightonSEO 2017

Harry Dodd
Sep 15, 2017
SEO

Excitingly, Google’s very own Gary Illyes (Webmaster Trends Analyst) made a special guest appearance at September 17’s Brighton SEO. It’s not often that you get to be in the room for a live Google Q&A, so we’ve captured the whole thing on video for you to see! From mobile first indexing to Gary’s favourite SEO weather report, there’s so much to learn from this exclusive interview.

brightonseo

Video Transcript

Gary Illyes: No, actually. [turns to walk back off stage]

Interviewer: I told him that if Barry Schwartz doesn’t have at least 20 blog posts about this then I didn’t do my job.

Gary Illyes: Well he will have 20 blog posts anyway.

Interviewer: I will have everyone note that he is wearing a Harry Potter outfit (inaudible).

Gary Illyes: Gryffindor

Interviewer: So for people who don’t know you, why don’t you tell them a bit about yourself?

Gary Illyes: I’m Gary. [waves]

Interviewer: What do you do? And they want you to pronounce your name, unless you want me to attempt it?

Gary Illyes: I’m Gary from Google, that’ my last name. And I say that because my last name is Hungarian and it’s very interesting to pronounce so I prefer just Gary. I’m a trends analyst at Google and right now I’m trying to figure out how to sit in these chairs. When I am actually working then I was a ranking engineer for about three years. I was working with (inaudible) organisation and company (inaudible) by then. I also work in pretty much every corner that we have in search, in the search tank. What else? I [sic] internally I’m educating people, mainly search engineers, how Google really works. We have a few hours long session on that every month. And now I’m doing mostly public speaking on behalf of Google Search, Web Search.

Interviewer: [Raises arm]

Gary Illyes: And what?

Interviewer: Can I attempt to pronounce your name? It’s Illyes. Close?

Gary Illyes: That’s not that bad [smiling]

Interviewer: Okay, lets dive in. I’ve had a lot of people ask me about the first click free programme in the search results. [Laughs] I’m doing my job. Where sites like Wall Street Journal and New York Times, sites that are behind a paywall in exchange for being ranked well in the search results they have to give searchers a couple of articles free over a certain period of time. Now apparently, they were doing away with this first click free programme so that (inaudible) a lot of rumour, so that sites can still rank, yet when people go to click through to them they hit a paywall. So what is changing with that?

Gary Illyes: Well I cannot tell you yet what is changing, but the leads are at least partially true, especially in the sense that this year in 2017 they were thinking a lot about the first link free programme and how can we work better with the publishers and how can we ensure that users are getting what they are looking for, also in a way that is search engine friendly. So we’re working with Financial Times and New York Times for example on [sic] we were experimenting with them and their teams on new solutions, but unfortunately I don’t have news for you today. I can just confirm that yes we are working on new things and as we get closer to something finished then we are going to talk more about it.

Interviewer: Is there any chance we’re going to see something like a subscription tag in the search results, similar to how they have it in Google News where it alerts you. [pause] That’s a cheeky look right there.

Lets talk about Fred. So.

Gary Illyes: Who?

Interviewer: You are the person that created Fred. So Fred is basically an algo that.

Gary Illyes: It’s not one algo, it’s all the algos.

Interviewer: So basically you can confirm it’s not a single algo it’s just like a whole umbrella of a whole bunch of different changes and updates that everyone has just kind of put under this umbrella of Fred.

Gary Illyes: Right, so the story behind Fred is that basically I’m an asshole on Twitter. And I’m also very sarcastic which is usually a very bad combination. And Barry Schwartz, because who else, was asking me about some update that we did to the search algorithm. And I don’t know if you know but in average we do three or two to three updates to the search algorithm, ranking algorithm every single day. So usually our response to Barry is that sure it’s very likely there was an update. But that day I felt even more sarcastic than I actually am and I had to tell him that. Oh he was begging me practically for a name for the algorithm or update because he likes panda or penguin and what’s the new one. Pork, owl, shit like that. And I just told him that, you know what from now on every single update that we make unless we say otherwise will be called Fred, every single one of them.

Interviewer: So now we’re in a perpetual state of Freds?

Gary Illyes: Correct. Basically every single update that we make is a Fred. I don’t like, or I was sarcastic because I don’t like that people are focussing on this. Every single update that we make is around quality of the site or general quality, perceived quality of the site, content and the links or whatever. All these are in the Webmaster Guidelines. When there’s something that is not in line with our Webmaster Guidelines when we change something that, or we change an algorithm that modifies the Webmaster Guidelines pretty much. Then we update the Webmaster Guidelines as well. Or we publish something like a Penguin algorithm or work with journalists like you to publish, throw them something like they did with panda.

Interviewer: So for all these one to two updates a day, when webmasters go on and see their rankings go up or down, how many of those changes are actually actionable? Can webmasters actually take something away from that or is just under the generic and for the quality of your site?

Gary Illyes: I would say that for the vast majority, and I’m talking about probably over 95%, 98% of the launches is not actionable for the webmasters. And that’s because we may change for example which keywords from the page we pick up because we see, let’s say that people in a certain region put up the content differently and we want to adapt to that. And then we test that in English as well of course, but if the change internally, when we pass the update internally. If the change is miniscule then we are going to launch it anyway, if we cannot improve for example (inaudible).

Basically, if you publish high quality content that is highly cited on the internet and I’m not talking about just links but also mentions on social network and people talking about your branding, crap like that. [audience laughter] Then, I shouldn’t have said that right? Then you are doing great. And fluctuations will always happen to your traffic. We can’t help that, it would be really weird if there wasn’t fluctuation because that would mean we don’t change, we don’t improve our search results anymore.

Interviewer: So a lot of people use the SEO weather reports to gauge [gary acknowledges] what Google is doing. Obviously you know what’s going on really from Google, how often is what the weather reports when they say it’s 110 degrees or it’s highly volatile today. How accurate is that from the back end of what you guys are actually doing. I know that we’ve had false positives from Wikipedia when they went secure and all the rank, weather trackers just went absolutely insane, but people were missing the context that, hey this massive site that ranks really well for a gazillion things changed all the URLs.

Gary Illyes: Right. I would say most of the time they get it right, but as I said, we make lots of updates, two to three per day. So if I write a script that generates a graph based on random data pretty much I will have a spike every day and that will be also inaccurate. Basically because we are making a change every single day. When, there were a couple of cases when, I only look at the Moz thing, whatever it’s called.

Interviewer: Mozcast.

Gary Illyes: Mozcast, yeah. I’m just bad with names I don’t have anything against Moz [raises hands] and usually they get it right. But the scale is very different than what we see most of the time. I have a very specific example on my mind where internally the change was way below 1% which was a standard slightly [inaudible] launch. And Mozcast reported that as insane, like everything is burning and [inaudible] already melted into glass. That was not right, basically we did not see that internally and we were just confused about what the heck are they reporting. But then other times they get it right. I don’t know how we deal with that, because you can very safely just assume that we made an update. But sure, one tool is better than the other, I like Moz personally.

Interviewer: Recently at a Google event for I believe it was Google Autos they were talking about Rankbrain and machine learning and they said that they are using Clickstream data, click through, bounce rate, stuff like that in order to rank. Were they specifically talking about the Rankbrain learning or were they actually talking about, hey we’re using click through rates and stuff on live search results.

Gary Illyes: I was trying to get in touch with the person who said that. We have quite a long thread with PR and ranking leads and with [inaudible] on trying to figure out what he was trying to say. I don’t actually know what he was trying to say. We already said that in Rankbrain, uses historical search data to make predictions on what works better for searchers and otherwise I will stick to what I said before that we use clicks and upclicks for very specific things. One of them being personalisation, like for example if you search for Python it may take us a few clicks to figure out if you mean the programming language or the snake or the bird or whatever it is, but after a while after we see that you consistently click on the birds or snakes or whatever.

Interviewer: It’s got to figure it out.

Gary Illyes: Search results will figure out that yes this person wants results about the animal. Experiments. When we do experiments we take a certain percentage of our traffic and we measure how users interact with an update and there we are, we use short clicks or long clicks to figure out if it’s better or if results got better or worse for specific queries.

Interviewer: Are they using machine learning language or machine learning in other parts of the algo right now or is it still just Rankbrain.

Gary Illyes: I mean, machine learning.

Interviewer: It’s like the new sexy thing in SEO right now.

Gary Illyes: But it’s a tool. It’s a generic tool that you can use for many things. And of course we are using for many things like we have brilliant machine learning engineers who can implement crazy crazy things and find new patterns that humans couldn’t before. So we are using it in many places, but it’s not a tool that you can use for everything. If you have to decide for example, what the result for 1+1 is you are not going to use machine learning for that. If you have an algorithm that can already figure that out then you are going to continue to use that algorithm because it’s simple and effective. You don’t want to spend engineering hours on that. But yeah, we are using machine learning more and more, not just in Google search but pretty much every single product that we have from Translate to Photos to Google Picture, whatever.

Interviewer: Let’s talk about links.

Gary Illyes: Oh links.

Interviewer: I’ve got more questions about links than anything else. Is Penguin still evolving or is it just at a constant state of being and just trucking along doing it’s thing.

Gary Illyes: I haven’t heard about Penguin for a very long time and I’m very glad about that. I know that we made a few fine tuning to Penguin. We made Penguin real time about a year ago I think and after that we said that we’d just drop it and let it do it’s thing because its more frankly better. We don’t have to do refreshes and shit like that.

Interview: Will there be a 5.0 penguin or?

Gary Illyes: After that we haven’t touched it.

Interviewer: No, it’s just going to keep doing it’s thing.

Gary Illyes: Basically, like I said we did some fine tuning to make sure it’s catching the right things, but otherwise it was like a done deal.

Interviewer: Okay. Current votes on the disavow rule. Should site owners still be disavowing links or should they be trusting that Google can handle it for the majority of situations. Unless they’ve obviously been going out and doing bad things.

Gary Illyes: If it makes you feel better, then sure. [inaudible] of the disavow tool is that the disavow tool is extremely powerful. Basically, I was looking at people who disavowed links from CNN and from the telegraph and whatever because they just didn’t know why would, or why did they get that link. That was bat shit stupid.

Interviewer: And people are disavowing links like this, it’s kind of crazy.

Gary Illyes: So if you know what you are doing, sure do it. The good thing is that you can also experiment with it. Basically you can disavow a whole domain for example and see if it causes anything, then you can just remove the domain and when we crawl your pages that link from that domain you’ll just get back all of your, what is called link authority or shit. Something like that, basically you get.

Interviewer: Link juice.

Gary Illyes: Not link juice, there is something fancier that’s used over here. Yes it’s link authority, because there is domain authority. We don’t have domain authority by the way, but sure.

Interviewer: Have there been more linking manual actions being given out since Penguin turned into this state of discounting.

Gary Illyes: I don’t know.

Interviewer: The link selling market has exploded obviously since Penguin came out because there’s, it’s being discounted. There’s a lot of people saying yeah we’ll get you all these links from forums and all these various publications, some better than others. And because, we don’t have to worry about Google because Google’s just going to ignore the bad ones and we keep all the good ones. What do you say about that?

Gary Illyes: If you are thinking about link building by buying links then yeah go ahead and buy links from forums for example. That would be great for your business because you lose like $3,000 or something like that and you will not benefit from that. Another thing is that you do have to keep in mind, that we still have manual actions for links and I’m actually very happy when I see that some bigger site has been hit with a manual action for buying links. Mainly because we are really clear on a lot of this in our guidelines and if you go against that then, if you actively trying to manipulate our search results, you deserve that manual action. And, unfortunately, it’s very easy to get rid of that manual action because we have the disavow tool for example, which I don’t necessarily like. But then I’m not in a position to make or for example remove this tool. I would do it, but maybe because I’m an ass.

Interviewer: So talking about this then, is there still any value in doing link audits. Like there’s some people that say you need to do full link audits every month. Other people are like meh, let Google handle it. What do you think about whether you should spend the time doing it or not.

Gary Illyes: I was chatting a lot with some SEOs from big companies, in house SEOs about what they are doing and it varies by company. They have different fields of views about why they are disavowing things. Personally, I don’t think that doing the audits very often makes sense because as you said we are pretty good at ignoring links and if we see that the links are coming in organically or that we are, it’s extremely unlikely that we would hit a site with a manual action. And if your links are ignored by penguin, again you don’t need to care about that. I run my own site, which gets about, perhaps 100,000 visits per week. It’s been up for four years maybe and I don’t have a disavow file, I don’t even know who links to me.

Interviewer: Another popular link tactic that probably everybody here has received lately is link reclamation based or broken link building where, hey Jen, I see you linked to Google and this post four years ago and the links broken, but hey. Link to our crappy SEO guide that has absolutely nothing to do with this instead. So does Google see this news article by them has been out there for four years and all of a sudden it changes to some other link, does Google devalue a link based on an old articles rank being changed. Or is there actual value for people to go out and do this work to build broken links?

Gary Illyes: I think there is value in fixing your broken links. [inaudible] If you link to some random crappy article that doesn’t have anything to do with your article then we are pretty much just going to ignore that link because the context doesn’t match and in fact that’s Penguin material, Penguin food. I think that if it makes sense to have that link, then sure. If it doesn’t make sense then ignore their request.

Interviewer: Mobile first index.

Gary Illyes: Oh goody.

Interviewer: Another one of your favourite topics. Anything new to report like people are like, is there an update on when it’s happening?

Gary Illyes: No. We are going to talk more about that. I vote the next few weeks. Word of warning, I was wrong about Penguin or when Penguin is launching for about half a year. I gave three or four predictions of when it’s going to launch. It doesn’t depend on me. The reason I say that we are going to talk about it in the next few weeks is that we are preparing, I’m preparing a blog post about some updates around it. And if everything goes right then, and our experiments show that we can move to the next stage then we are going to publish it and get some working ideas about mobile first indexing. At this point I don’t think that you have to worry about it too much. Especially if you have a responsive site. If you don’t have a responsive site, then you have to think about it.

Interviewer: Yeah, Google just had a blog post actually today, talking about if you have an m-dot site you should think about going responsive.

Gary Illyes: We published that?

Interviewer: Yes, you did.

Gary Illyes: So, if you have an m-dot or a separate mobile site then just make sure that the content you want to rank for is actually there on the mobile site as well. Pay attention to meta data, no index, hreflang, crap like that. Structured data, again, pay lots of attention to that and images media. Make sure that’s also done on the mobile site. If you have a desktop site then, or only a desktop site then well, what’s wrong with you? You probably want to create a mobile site at one point. But otherwise, you will still be in the mobile first index even with a desktop site. And you don’t have to worry that you are going to disappear in the mobile first index.

Interviewer: And just to confirm again, because people keep asking, content behind tabs on mobile. Wikipedia is a perfect example, that’s perfectly fine because that’s for UX reasons.

Gary Illyes: Yeah.

Interviewer: A lot of people have amp as their only version of the mobile page. For mobile first, will Google will Google crawl the amp version of the page?

Gary Illyes: Yep, these are easy. Let’s do this more.

Interviewer: Do people need to worry about switching rel canonicals?

Gary Illyes: No.

Interviewer: Google will sort it out for themselves going forward, like if people are adding new pages, should they have a mobile page as the canonical.

Gary Illyes: I think just semantically it would make sense, but then again we are working hard to figure that out on our own and we don’t want to ask people to swap their rel alternates with rel canonicals because it would just take like 30 years.

Interviewer: Obviously, amps been pushed through by Google for the last year or so. Right now on mobile, you know top story carousels you have to be amp. Is there any plans to expand in the regular organic where.

Gary Illyes: We have ten blue links that are amp.

Interviewer: Yeah, but pages that aren’t amp. Will there be any demotion or anything for pages that aren’t amp or will we see the top 3 spots reserved for just amp pages.

Gary Illyes: I don’t know about such plans.

Interviewer: A new trend that’s actually popped up over the last few weeks. People are putting their amp pages that are ranking in the carousel. They just have a little snippet, a little paragraph of amp content and then to read this article click here to go through to our full experience site where you have ads, interstitials, popups and all that kind of crap. Is there going to be any issue with this, or can webmasters do this and not worry that Google’s going to be like, get out of amp.

Gary Illyes: We are aware of this trend. We are looking at what can we do about it. How can we make sure that our users will get a good experience.

Interviewer: Is there any plans for and additional amp boost?

Gary Illyes: Not that I know of.

Interviewer: Or mobile boost?

Gary Illyes: Not that I know of. I think we tried to stay clear from those kind of boosts nowadays and we prefer, we strongly prefer organic adoption, thankfully versus [inaudible] with boosts.

Interviewer: Will we say an additional HTTPS ranking boost then?

Gary Illyes: I’m not working on that. No the reason she asked is because I created that boost 3 years ago maybe, 4 years ago. It’s a tiny boost, it’s very up to.

Interviewer: It was basically a tie breaker before.

Gary Illyes: Yeah it’s a tie breaker yeah. And in the blog post we said that we may increase the boost in the future, but I’m the owner of that boost and I’m not working on it so.

Interviewer: It’s going to sit there as is. For people that are switching to HTTPS. It’s kind of, you guys are saying it takes about 3 to 4 weeks in order to see rankings recover as Google reindexes the new site. Has, as Google has been doing this more and more has this dropped down to less than that or should webmasters still be prepared for 3 to 4 weeks of ranking fluctuations and traffic changes and all that.

Gary Illyes: I work a lot on that. Basically trying to figure out the signals that are lost during the transition, during the site move. And we identified I think 3 of which we fixed 2 and the one that’s left is some obscure crap that no one cares about. I would say that if your rankings don’t stay stable with a move, then something is wrong. And not necessarily on our side. I would be very happy to take a look at moves that don’t go, HTTPS move that don’t go smoothly. Mainly because I thought that we fixed the issue with that.

Interviewer: Okay, so then we switch over from HTTP to HTTPS, like the big things are 301 redirects and sitemaps. Is there anything that people are making mistakes on other than those two things that are impacting how well they do?

Gary Illyes: Critical resources should be on HTTPS as well because then we will pick up the HTTPS more readily. If you are comfortable with the move or with the end product then consider HSTS because then browsers that support HSTS, which is pretty much all of them I think will automatically load HTTPS versus HTTP.

Interviewer: When people are doing this transition a lot of people are saying we’re going to redo our URL structure, our categories.

Gary Illyes: Yes, No, no, no [shaking head].

Interviewer: People should just do the HTTPS on it’s own and then deal with any other changes.

Gary Illyes: Yes. Yes, yes, always. If you are doing any sort of site move. Don’t change everything at the same time. If you want to do a redesign of your site then redesign the site. After that, once the ranking fluctuations have gone away because if when you redesign your site it’s, I can almost guarantee that you will see ranking fluctuations. And that is working as intended. But, don’t combine redesign, with a URL structure change or type of moving to HTTPS or crap like that because you are going to shoot yourself in the foot, and with a cannon pretty much.

Interviewer: Is there any chance that Google is going to start marking sites as insecure in the search results as we start moving forward.

Gary Illyes: I’m not aware of that such plans.

Interviewer: I can see Kelvin there, which probably means we’re right out of time.

Gary Illyes: Yes [raises clenched fists].


It was great to hear these behind the scenes insights from Gary. Hopefully he’ll be making appearances again in the near future! If you’d like to find out more about this September’s BrightonSEO, check out our roundup article here for they key takeaways from speakers.



This post was compiled by:

Harry is an SEO Account Manager at iThinkMedia with over 4 years experience working in digital. He is involved with both the management of SEO clients and the day to day work, with a widespread knowledge of SEO. Harry also specialises in big data and working with heavy Excel spreadsheets.

WHAT DO YOU iTHINK?

Leave a Reply


Your email address will not be published. Required fields are marked *


Contact Us
Type the code you see below:
captcha