In-depth Q&A with Google’s John Mueller
Hear from Digital Marketing League expert John Mueller on why he joined Google, the impact of Coronavirus on Google the company and Google search, and much, much more.
On 9-June-2020, Google’s John Mueller spoke with Textbroker’s Jochen Mebus and Zohra Belmahdi about a wide variety of topics. The session filled the final slot of the first-ever Digital Marketing League online conference, powered by Textbroker and friends.
Jochen and Zohra had some questions of their own, and also relayed questions from attendees of Digital Marketing League. As a result, John spoke about a wide variety of topics, including his own background and, of course, why Google ranks some pages well and others not as well.
We expect John to return to the second Digital Marketing League, scheduled for 23-October-2020. Subscribe to our content news to be informed as soon as registration starts.
Read the full transcript
Jochen Mebus: Welcome everybody. This is our Q&A session with John Mueller from Google and we’re going to start right now so please keep on sending questions – we’ll try to get through all of them – if not I’m sure there’s a way that we can ask John later on and he will answer.
Maybe we just start with a smooth question first. John, tell us a bit about yourself; you know, How did you get to Google, what positions did you have there, and what makes you live in Switzerland?
John Mueller: I don’t know. I’ve been in Switzerland for a really long time so it’s hard to say. I mean, I moved with my family when I was a kid. We moved to the U.S. for 10 years when I was a kid and then we moved to Switzerland and I came along and finished my school here and then went to study robotics at the university here in Zurich which is a little bit different. And during that time, I built up a small software company and as you do with software companies you start making websites as well. So I was making websites and kind of making our own website and at some point realized that it’s not just kind of normal to appear in the Google search results on page five for your company name, but actually probably makes sense to be here on page one for your company name at least. And that I think kind of pulled me into all of the things around search. We created one of the early site map generators at the time and I don’t know I just dove into all of these topics around search…tried things out, made various test websites. Some things worked well, some things worked terribly, but I also ended up spending a lot of time in the help forums chatting with other people doing things, other people who were trying things out, and from that I don’t know. Somehow I received an email from someone at Google asking if I wouldn’t mind coming along for lunch at some point, which was kind of lucky because the email was on a domain that I didn’t actually monitor the email for. But somehow it worked out you know.
Jochen: That was the one day when you looked at it, right?
John: Yeah, more or less. And that was, I think I started at Google 12-13 years ago and have mostly been doing similar things. Talking with the people who work in web search and people externally who are making websites.
Zohra Belmahdi: Okay, thanks. So, I have a question which is also a question in the chat actually. As we live in special times due to the Covid pandemic how is life at Google during this pandemic and what are the effects on the area you are working in?
John: I think one of the things that happened fairly quickly is that within Google we were quite conservative with regards to the situation, and essentially we closed the offices very quickly and everyone has started working from home since I don’t know March sometime early. And that I think for the most part that worked really well. But the whole situation around the pandemic, around coronavirus, affects people in very different ways, and I see some people have really started to excel by working at home and kind of being without distractions and being able to focus very well, and other people struggle quite a lot with kind of this general worry that the world is going to end…nobody knows what’s going to happen next. And I think like all of those reactions and everything in between is completely normal. People will respond in different ways and that generally means that overall probably things are going a little bit slower than they could be. I suspect things are catching up again a little bit now that people have settled into kind of this cycle and realized that, actually, it’s not going to go away after a few weeks. But now there are other things happening in the world where, depending on where people are located, they’re very kind of front of mind as well. Especially the events in the U.S., for example now, that’s something that are very strongly affecting some people as well.
Jochen: Okay and I think in that sense you’re a very normal company, because I guess every one of us has the same experience, working from home. Some people like it more than others. And so yeah, let’s go into the questions here in the chat, and I have to apologize we can’t ask all of them because I see a lot here, but let me randomly pick one. So, Elias is asking, I’m starting to change my website from German-only to a multilingual site. Is it better to have a separate domain for each language, have a sub-domain for a language, have it in the URL, or what’s the best advice that you can give for that?
John: So for multilingual sites the only important part is really that there are separate URLs for each language, so that you don’t use one URL and automatically swap the language versions on there, but rather that you have separate URLs. And how you set that up, on separate domains, on subdomains within the path, or within query parameters, or whatever, that’s essentially up to you. So, all of those options work well for multilingual sites. If you’re targeting individual countries then you need to have it split by some kind of a logical segment within your website. That could be a different domain, could be a sub-domain, it could be a sub-directory, but it has to be a clear part of your website. So, you can’t put like the country at the end of the URL; it has to be somewhere near the top of the URL structure.
Jochen: Okay, cool, thank you and actually Zohra I think the questions we thought of are partly also those that the audience is asking, because my next question was actually this one. So, let’s take another one from the audience then…pretty long question here from Tom, but I’m trying to read it quickly. I have a question about internal link juice. We’re planning to have a mega menu in the header of our shop that would contain quite a lot of internal links to our product pages. Their consultancy agency gave them the tip to implement PRG patterns to master links with formulas so Google does not follow the links and they have to save up crawl budget and optimize internal link flow. Would you consider this a bad practice or is there use in implementing this?
John: I think with mega menus in general what tends to happen is the structure of a site gets very flat, and the one thing I would watch out for there is that it becomes too flat in that we can’t recognize which part of the site belongs together anymore. So there’s something to be said for reducing the height of kind of the crawl of a website so that we can crawl a little bit more streamlined, but at the same time not flattening it out completely…so finding kind of a reasonably flat and long site structure. I don’t know, it’s weird to describe, that’s kind of what you should be aiming for. And usually using something like PRG patterns, they require some work on the server side to be handled properly. I feel that’s a lot of unnecessary overhead that in the end makes it possible for things to go wrong in weird ways that you don’t understand, and I generally would recommend more using things like maybe just the no follow links for links that you don’t want to have kind of followed from a crawling point of view. The issue with kind of adding all of this extra complexity with the PRG structure just means that things can break in ways that only Googlebot will notice that you might not notice when you access the site yourself. Whereas with clear no follow links you can use any tool to crawl your site, you can recognize those links, and, say, kind of understand that things are still happening normally.
Zohra: Okay another question, maybe a bit linked, what about the disabled tool? Some people at Google say that it is useful only when you have a manual action, and other Googlers say that it can work in other situations with no manual action. So, is it possible to have a clear position there?
John: A clear answer? I don’t know…it’s like who are these people who are not saying what I say? No, so the practical use of this disavow tool is really for cases where you have a manual action on links, but it might be that this manual action on links is something that is going to happen in the future. So that’s kind of the other side of the story in that if, when you look at your site’s links, if it looks like anyone from the web spam team, if they saw your site’s links, they would say, “We need to take a manual action maybe next week,” then that’s something where you would probably want to use a disavow tool to help clean that up ahead of time. Kind of the third option there that I sometimes recommend the disavow tool for is in situations where you’re just really worried. Where, when you look at your website, maybe you’re not like a really advanced SEO person, and you look at your website, you see, “Oh there are a bunch of crazy links that maybe someone in the past has built for my website. I don’t know if the SEO built it for my website, or some competitor, or maybe some random spammer built these links.” And if you’re losing sleep over those links then using the disavow tool is a good way just to make sure that you’re sure that Google is not going to take those into account…kind of a way to just make things clear so that you don’t have to worry about that at all. So those are kind of the three options: when you do have a manual action, when the manual action is just around the corner, or when you’re really kind of lost and you don’t want, you really want to make sure that these links are not taken into account, and you can’t really judge whether or not they’re bad enough to trigger a manual action or not.
Zohra: Okay, that was a clear answer.
Jochen: I can do the next one again from the audience. Ulf is asking, “What do I do when I sell products from a brand when the manufacturer insists on using only these official brand texts as far as duplicate content is concerned?” That’s a great question which we at Textbroker also get quite often.
John: I think that’s perfectly fine. So, when we look at a website or a page in general with regards to duplicate content, we look at it on different levels. On the one hand, the most obvious one is when the whole page is a one-to-one a match of an existing other page, then that’s something where from a technical point of view we can just say, “These are exactly the same pages, we can fold them together, and treat them as duplicate content,” which is not a bad thing. The other probably more common aspect is when individual parts of a page are duplicated across maybe other pages as well, which is where maybe the description or the title or some of the images are exactly the same as on other pages. In a case like that we will still index these pages individually, and we will try to rank them individually based on other information that we have about that page. So that could be other information directly on the page which might be a review maybe from an expert from your site. It could be reviews from users. It could be comments from users. It could be related products that you also sell. It could be something as simple as your address or your phone number, where if someone is looking for this product locally then your page is the one that has the address on it that is local; therefore, that will be the one that we show. So just because the description is exactly the same as on other pages doesn’t mean that that page will be less valuable. You need to add value to other parts of your page as well.
Jochen: But still any risk would be much lower if you had an original description on each page, right?
John: I mean risk is hard to say because it’s not that you will get penalized for this, but what can happen in the search results is if someone is looking for exactly a piece of text that is in this description that’s shared across a lot of different pages, then what will happen is, or what might happen, is that the snippet that we would show in the search results is exactly the same for these multiple pages. And in a case like that we will fold those into one search result entry. So if the query is such that it’s really unclear which of these pages would be the one that we would show, and the snippet would be exactly the same, then we might show yours, we might show one of the other ones. So that’s kind of, I guess kind of the risk that you’re alluding to; whereas, if you write your own description then probably that snippet will be slightly different for your pages and we wouldn’t fold them together. But I assume there’s also very strong aspects of all of the other parts of the page that flow into the ranking side of things, so it’s not just that one snippet and folding things together.
Jochen: Okay, cool. Zohra, you want to take the next one?
Zohra: Yeah, a quick question about the new search console. Will the international reports, hreflang country, be transferred into the new search console?
John: I hope so. I don’t know offhand exactly what is planned there, but we tried to remove things that we know we don’t need in the new search console, and the rest are things that we’re still moving over step by step. For the most part, the team is trying to look at the reports and think about ways to achieve the same goal in a more streamlined way to make it easier for users, rather than just to copy things over one to one. So that’s kind of sometimes where it takes a little bit longer, because we just don’t want to kind of like copy this UI into the same setup on a new tool, but rather rethink what it is that we’re actually trying to help people with, and how can we help them to do that better?
Zohra: Okay, thanks.
Jochen: Okay, here’s another one: What are the important elements in order to improve the SEO of a Google chrome extension?
John: I don’t know. That’s a good question!
Jochen: Congratulations! You asked him a question that he doesn’t really have an answer for
John: Yeah, I don’t know how the Google chrome extension pages are shown in the search results. My guess is they’re just, I don’t know, shown in the chrome directory or chrome extension directory and shown from there. But I guess what you could also do, of course, is host informational pages about the extension on your own website. And in a case like that you have full control over what you host on your website and how you present that, so that that might be an option. And in the worst case, or I guess in a good case, you would have the official chrome extension page ranking somewhere in search and your website ranking somewhere in search. So, you kind of have those two individual places where you can show information about your chrome extension.
Jochen: Okay, cool, thanks. Zohra?
Zohra: There are too many questions. I hope you haven’t picked it. Just tell me if I’m doing a question you have asked. We are building a single page application and going to rely on dynamic rendering serving correct status codes to the bots. Should we care about the real users, especially in the context of cloaking?
John: Should we care about the real users – of course! I mean, I don’t know, I don’t know exactly what they’re referring to, but of course you need to make sure that your website performs the same way for users as well in that what we index is essentially what we can refer to, and where we can send users to, and where we can kind of trust that users see the same content. So, of course, I think you should care about your users in that regard. It might be that you’re talking about something slightly different with regards to maybe dynamic rendering where you’re also doing server-side rendering for users to speed things up. That’s a totally legitimate technique as well, but essentially it should always be the case that what we index for your website is that what users would see when they go.
Jochen: Great. This is a little longer story here from somebody who I think has some bad experiences, so Fabian, I’ll try to summarize this. Fabian has the issue that competitors of his are publishing stolen and simply translated content on their website and don’t get penalized even though the texts are error-prone, not readable for the normal reader. So, the question really is, How does Google deal with stolen and simply translated content?
John: It’s hard to say, because it’s tricky to understand what stolen content is, for example. I think the other part with regards to translated content, when it comes to translations, we see them as unique pieces of content. So if you take an English page on your website, you publish the exact same concept in French then from our point of view that’s a unique page. It’s not something where we would say, “Well, this is the same topic, just a different language kind of thing.” It’s a unique page. So, if you take your content and you publish it on another website in a different language, it’s a new piece of content, it’s a new page. And similarly, if someone takes your content that belongs to you and they translate it and post it on their website, from our point of view, when we look at that, we see, well these are different words, this is a different piece of content – we can index it individually and show that in search. So just because something is translated doesn’t necessarily mean that we would say it’s duplicate content, we need to remove it from the web. It might be that it’s low quality content. It might be something where when users go there, they say, “This is terrible,” and nobody refers to it; therefore, it doesn’t get any links and doesn’t show kind of highly in search. But if you search for it explicitly, then maybe you will still see it. So that’s something where essentially the approach that you would take as a site owner there is to try to find a solution yourself with that other website, which could be finding a legal approach and saying like, “This is my copyrighted content you must take it down.” It could also be going there and saying, “Well, they spent a lot of time to translate this content, maybe they like the content, maybe I can work together with them, maybe there is kind of an agreement that we can find where they translate the content and make it available in other languages and we kind of share the results from that somehow.” But ultimately it’s not something where we would step in, even from a web spam point of view, and say, “Well, this is just translated content – we will call it spam because it’s an official translation.” We might step in because it’s low quality content. We might step in because it’s gibberish. We might step in because it’s scraping various websites. But just because it’s a translation doesn’t necessarily mean that it’s bad by default.
Zohra: Okay, thanks. A question again about the link to the crisis of the corona. In some markets, the coronavirus outbreaks has completely changed the landscape, such as in the event sector. Users are searching for online events rather than physical…however, when things go back to normal, is there likely to be a lag in the search results, and will we need to wait for something major like a core algorithm update to see things go back to normal?
John: I don’t know what will happen when things go back to normal. I hope that happens at some point. I mean, it will at some point certainly happen. My feeling is that when things go back to normal it won’t be from one day to the next like we had with the lockdown, but rather something that’s gradual, that takes place over a certain period of time, maybe a couple of months…probably different in individual countries. And for changes like that, that’s something that we can probably pick up through normal crawling and indexing, where we recrawl content on your website, we understand these pages are new and updated, and we can reflect that in search based on those changes that you have there. So it’ll be less like from one day to the next we have to switch everything to physical events, but rather some things might still be online because they were planned that way and then new things start appearing better…more in-person events and over time that will shift again.
Zohra: Okay, just another question which is linked to this one so would be difficult to predict. I can see that you cannot predict everything but the question is, Will the Covid crisis change the date for the end of the mobile first index project when all sites should be in this index in September as announced?
John: We’re still discussing that. I was just discussing that with the team this morning and I don’t think we have an absolute answer just yet. I think, what some of the things that make it a little bit tricky, is on the one hand, we still see sites that are improving with regards to mobile first indexing. We also see a lot of sites that are not responding at all to the emails that we send them in search console, or that are not even verified in search console, where I assume these are sites that are not ready for mobile first indexing, but also which have been more or less abandoned websites that people have created over the years, and they just leave them running because it’s easier to keep them running. And those are things that we probably don’t need to expect to shift over at a certain period of time, and that’s something where we’re still a little bit discussing what we need to do there: if we should just shift some of those sites over and see what happens, if we need to kind of find ways to help people move over a little bit more. I was hoping we would get a lot of feedback on this when I mentioned it in one of the Google search news videos just to see, are people really struggling with the timeline here, or is it something where most sites are moved over and people are happy and kind of okay with that, and I haven’t really received much feedback at all with regards to this. So my feeling is probably we can keep that date, but we’re probably going to decide in the next couple of weeks just to kind of have a clear answer.
Jochen: Yeah, no answer sometimes is also an answer. Next one here, What is the importance of content on a website to rank on Google? Does this importance become more or less important with respect to the other factors? Why is a website with excellent content, technically excellent, like speed, canonical, etc. with a natural link profile punished by EAT updates?
John: I think it’s really hard to say what specific website it is, but I think for websites overall there are lots of different factors that play a role. On the one hand, the content itself is something that really needs to be high quality and relevant for the user, but sometimes the content is written in a way that is technically correct but is still wrong, for example, and then our algorithms should ideally be able to recognize that somehow. And sometimes the content is nice content but it’s presented in a way that it’s really unusable for users, maybe in a way that it’s really hard to read, or maybe it’s presented in a way as if it’s a really old website kind of thing where people have trouble trusting that content perhaps. All of these things play together and overall we need to find a balance with regards to how we rank those pages in search. And that’s something, I don’t know, someone recently asked me with regards to like the speed metrics as well, is that going to be more important or less important than maybe some of the other aspects with regards to ranking? It kind of falls into the same bucket of things where we have all of these different aspects with regards to a website, and for the individual queries we need to figure out which of these aspects are the critical ones. If someone is searching specifically for your website, if they’re searching for your company name, then we should be able to figure out that your website is the right one to show even if it’s otherwise a terrible website, it was really slow and I don’t know written in a way that is mostly gibberish, but if we can tell it’s your official website and people are searching for your website, we should show that. And that’s something where, for other topics, maybe if someone is searching for a theme that is also on your website, then that makes sense to take all of these aspects more into account. But it’s really the case that there’s no kind of like clear ranking factor that is above everything else. It really depends on what the user is searching for, how they’re searching, and based on that we need to figure out which factors are the right ones to emphasize here.
Jochen: Maybe adding a question: so if that person had asked that question now sent you an email with the particular website case, could he hope to get an answer for his specific case, or is that something you generally don’t do?
John: Sometimes. Sometimes we can help. So it’s something where there, you know, I don’t know, it’s like always, “it depends,” but how do I say it in a different way? It’s something where if I can look at the website and the queries that they send me and clearly recognize that our algorithms are doing something wrong, then that’s something I would just forward to the internal ranking teams and tell them, “Look for this particular query…we’re not showing this website as the first result and that’s clearly wrong. You can objectively look at those search results and say that that is the wrong result that we’re showing. This one will be the right one.” And that’s something where I wouldn’t necessarily send feedback back to the person who created the website, who sent me the email, saying, “We’re going to fix this,” because I can’t promise that we’ll fix it. But I can bring it to the right people so that they can make a call themselves and figure out what they need to do over time. Sometimes there are cases where I look at the website internally and I see, oh, there’s this really obvious technical thing that they’re doing wrong and they just don’t see it because we don’t show it on search console or whatever. And that’s something where I would be able to go back to the person who sent me the email and say, “Hey, this is something that you can do on your side. Maybe it does, maybe it doesn’t change your rankings directly, but it is something that has a significant impact that you can change.”
Jochen: Okay, cool, so, maybe in summary, maybe all of you guys out there who feel you’ve been treated unjustly, maybe you feel encouraged to send John an email.
John: I mean what really helps me with these cases is when people can send me kind of general queries where it’s really obvious that we’re doing it wrong. If you send me this like five word query where it’s like, “My website should be ranking number one,” like two people search for that in the last year and they’re probably you, then that’s not something where I think I can pass that on to the engineering team and tell them to do something. But if it’s a short query and it’s really obvious we’re doing it wrong, that’s like the perfect thing to forward to the teams.
Jochen: Okay, cool. Zohra, do you want to take the next one?
Zohra: Yes, someone has experienced a tricky thing. So, he said, “We have some competitors that basically copied our site structure and content. With the Google maker update they gained a lot of visibility while our own site remained pretty much the same. So, our site still has a lot of high-quality articles that haven’t been updated a lot because they are still on point. They rank pretty high. So, I mean the competitors, they rank pretty high. What can we do to improve our visibility?”
John: Now I think in a case like that, if they’re really copying your content, then the first step I would try to do there is maybe find a way to take legal action directly. Often there are ways that you can do that. It depends a bit on what exactly was copied and where you’re located, where the other site is located, but a lot of times you can use the DMCA process to take those other pages out and just clearly say, “This is my content, this should not be published on Google.” So that’s I think the first step that I would take in a case like this, because that’s really, I don’t know, something that you can control yourself where you can step in and kind of make that call. If someone is just taking your website as inspiration and essentially created a new website which is kind of similar to yours but has their own content, then that’s a lot harder. It’s a lot trickier for us to say we need to take action or we don’t need to take action. Maybe they were inspired by your website, maybe they kind of copied things directly, but if it’s not a case of something where, from a legal point of view, we would need to take action through a DMCA complaint, then it’s essentially a new website that just looks like your website.
Jochen: All right. There’s one on single page applications here. It’s Michel who I had the presentation with this morning – Hi, Michel – so he has a few clients that have thousands of SPA pages, Google indexes them. He’s asking, “When is it the right moment to start implementing an SSR version of that site? That’s quite some development time…is that investment really worth it when you have a smaller page like less than 100 pages, or do you recommend this more for larger, more complex sites? What’s your advice there?”
John: Yeah, so we’re really good with rendering I think, and we’re getting better and better at it. So, for a lot of these single page apps, if we’re already keeping up with the content that you’re providing, I think we’re in a pretty good place with regards to Google specifically. With regards to other search engines, that’s a different story. With regards to maybe social media sites, like if you’re seeing that your users are sharing your content on Twitter or on WhatsApp or wherever, then maybe it makes sense to do something server-side for that specifically. I think the other aspect where it can make sense to have something server-side rendered is, I hope I’m remembering this correctly, things around, for example, Google Shopping. So, for example, if you have an e-commerce site that you’re building up within your single page app framework, then it might be that the Google product search crawlers have trouble picking things up properly because they’re not able to render all of the pages all of the time. But otherwise, if you just care about Google web search, then probably we’ll be able to keep up with a reasonably sized site if we have to render it ourselves.
Jochen: Okay, cool, Zohra?
Zohra: I pick a question from Jasmine about the categorization. So the question, “Is it possible for Google to mis-categorize a website, for example, categorizing a service-based site as a blog based on the amounts of and type of content living on the site? How could we take steps to get categorized correctly?
John: I don’t think that’s really something where you would need to do anything particularly, because it’s not, at least as far as I know, not the case that we would try to rank pages depending on the category of the pages that we think. So, things like blog posts or articles or even a category page, from our point of view, they’re html pages that are essentially the same. I think where the categorization can play a role a little bit is if it’s an e-commerce page versus a random blog post. So if you’re writing about a product and you’re writing about how you’re using it, or what else you’re doing, and you have this long blog post, then from our point of view that looks like an informational page about a product. Whereas, if it’s an e-commerce product page then that’s something where we’d like to recognize that, actually, there’s like a shopping cart here; there’s a way to buy this product; there’s a price associated with it; you can go there to buy it; we can kind of show it within the search results for people who are looking to buy a product. So that kind of categorization is something that could play a role, but that’s something where you pretty much have that in control yourself and essentially how you’re creating those pages and usually not something where we run into a conflict of understanding what this page is about. The other kind of categorization that I’ve sometimes seen is with regards to content that’s filtered by safe search and content that’s not filtered by safe search. And that’s also something where usually the difference is very obvious when you look at those pages in that you see, well, this is an adult content page; this is not an adult content page. When you look at the search results if you turn safe search on and off you see it or you don’t see it, then it’s pretty clear what is happening there. And that’s also something where usually the answer with regards to what you need to change the categorization is pretty obvious as well. It’s like, obviously you remove all of the adult content on these pages and then we can treat it as a page that doesn’t have adult content on it.
Jochen: Okay, great. Natalia is asking, taking into consideration the last update and the EAT principle, how important is still the quantity of the content (in this case, meant as the amount of words in a piece of content), in comparison to the relevancy of the user request? And does a site with thin but pertinent content have a chance to rank in the long tail keyword?
John: Sure. It’s something where, as far as I know, none of our algorithms count the words on a page. So, it’s not the case that you need to have a certain minimum number of words on a page in order for us to rank that in the search results. But obviously if you have the words that you want to rank for on a page, then that’s something we can take into account. Whereas, if those words are not on the page at all, it’s really hard for us to recognize that these are synonyms to things that you mentioned, then we can’t really show that page. But just because there is not a lot of content on a page that doesn’t mean it can’t rank. Just because there’s a lot of content like a long article on a page that doesn’t mean that it will rank, either. So ideally finding something that works well for users, something where also when users come to your pages they understand kind of what you want them to do next, which is usually the more important part, then that makes it a lot easier. So, if you’re trying to sell something, have that clear call to action so that they can find the content they were looking for as well as recognize the call to action that you have there.
Zohra: Okay, thanks. Now a question from Jesus about EMD’s (exact match domains): “I heard some years ago that Google was going to penalize EMD’s, but they are increasingly in the SERPS in transactional searches. Is Google going to take any action?” (Sorry for my English.)
John: No, it’s perfect. As far as I know we don’t have any plans to take out the EMD’s from the search results. What we did say at the time is that we’re going to take action on low quality sites that are built on EMD’s. But that’s, from what I’ve seen, it’s mostly due to these actually being low quality websites. So less the case that, “Oh, it happens to be an EMD and therefore we will not allow this website to rank,” but more that actually someone decided to spend the money on a domain name rather than on content, and then of course there’s not a lot of great content on the website. So that’s something that we try to not choose highly in search. So that’s kind of where we’re going there and sometimes people have an exact match domain and they have good content and that’s something that we should be able to show in search as well. The difficulty that I find personally with exact match domain names is that it can make it very hard to expand the scope of your website. So, if you have bestshoes.com and you want to sell, I don’t know, t-shirts, then suddenly it feels a bit weird that on bestshoes.com you go and buy your t-shirts. Whereas if you have something a bit broader which could be your brand, which could be anything else, then it’s a lot easier to expand the scope of your website without having to rethink everything that you have around your website.
Jochen: Great. We’re moving to a question about images. If a website has lots of images as part of a how-to guide, will that affect the weight of the remaining parts of the website in the index? And also, the question extends about image resolutions as well as pixel graphics…so what about that?
John: So, for web search in general the amount of images on a page doesn’t play a big role. The one place where, kind of like thinking about this question a bit, the one place where it might play a role is with regards to the rich results that we have. So, specifically, I think the how-to rich results are ones where you can display an image. Not 100% sure, I think that’s the case. So if you have images that map to these how-to rich results where you would like to have some of these steps being shown directly in the search results, then that might be a good opportunity to kind of be a little bit more visible in the search results there. With regards to images, otherwise, for web search they don’t really play a big role there, in that for web search we primarily look at the text content on a page, so the images themselves are more kind of decorative for web search. They help users understand the page a little bit more and if they understand the page better maybe they will recommend that page which is kind of good. For image search, on the other hand, the images do play a big role obviously and how they’re embedded within the page, they play a big role. So, to understand things like the alt text associated to the image, the caption under the image, the text around the image as well, all of that plays a role for image search. And also for ranking in image search, the quality of the images also plays a role. So if you have low quality images then obviously we will try to find other higher quality images to show in image search. And even if we were to show them in image search then users might not be as interested in clicking on your results because the preview is kind of low-quality image. So depending on the way that people are searching for your website, if they’re searching visually, then maybe image search is a stronger priority. If they’re only searching for web search then, apart from these how-to rich results, maybe the images are more kind of something between you and the user with regards to how they recommend your website indirectly.
Jochen: Maybe extending, Zohra, if I may quickly, that question for a question from me: So does that mean if we talk about videos, in order to have videos in web search, is the consequence of what you just said, that you should need a transcript or an article together with that video so that you can recognize it in web search, meaning the video itself wouldn’t get recognized?
John: Yeah. So, for web search we do show the video snippets sometimes, but the ranking itself is really done on the landing page, on the text of the landing page. So if you just take a video and you host it on your website, and you host it on an empty page as a video landing page, then we could theoretically show that empty page in the search results. We might not have a good snippet but it would be really hard for us to rank that page. So we wouldn’t go and look at the video and try to extract the words from the video or try to extract the images and understand what the images are about in the video, but we would really focus on the words on the page.
Jochen: Okay, thanks. Zohra back to you.
Zohra: About the voice search, is there a timeline for the Google voice search and should websites already implement the necessary technical data? Should the content be tailored to it already now like simple sentences and so on?
John: I don’t know. So I don’t know when we would show that in search console. I think it’s kind of a tricky situation still in that when I ask around it feels like a lot of people are not really using voice search to find websites at the moment. So that might be something where the general user side of things needs to change a little bit first, and at some point maybe it will be the case that people use voice to search for websites more and we have some kind of voice search results that we can present and report on in search console. But it feels like at the moment it’s still the case where we don’t really have a ton of data that we can show in search console which is very actionable. So that’s something where maybe the voice ecosystem overall needs to move along a little bit more before we can start showing things in search console as well.
Jochen: Great. I got two questions here about Google News. So first one is: We have competitors that are listed in Google News; they are publishing advertisement articles, paid articles, and they show up in Google News. Is that intended or is that something that should or can be reported?
John: I don’t know. So, I don’t know much about the policies with regards to Google News, so it’s hard for me to say if it should be reported or not. It feels like in general these kind of things are topics that you can definitely forward to someone at Google. So specifically for Google News publishers in the help center there’s I believe a contact form where you can contact someone from the Google News publisher team and kind of ask them questions or forward them information. So if this is something that you’re seeing as a Google News publisher then maybe it’d be useful just to pass that along and leave the decision up to whoever on Google’s side is kind of processing these reports with regards to what the next steps will be. From my point of view, I really don’t know how the policies are on the Google News side there.
Jochen: It certainly is a good way to get clarity and so probably you’re going to give the same answer to the second question which is, With the change of Google News there’s no possibility to apply anymore for Google News. The info says it’s automated. Do you have any tip or help that you can announce here so that people do get listed?
John: Yeah, same thing there. I would also use the contact form in the Google News publisher help center.
Jochen: Okay, cool, good. Zohra back to you.
Zohra: Yeah, maybe another question about the categorization, to finish the part of the categorization. You just mentioned e-commerce sites – in which category do price comparing websites, affiliate websites, fall into? Would you consider for purchase queries also price comparing websites?
John: I don’t think we have an exact answer for something like that where we would say, “This is exactly like this or exactly like that.” From a practical point of view, I do sometimes see price comparison sites ranking for more shopping-related queries, so it feels like we’re already trying to show these appropriately in the search results. But I don’t know if we’d have an exact categorization for those kind of sites that kind of fit somewhere in this vague area of informational page and kind of transactional page directly. There’s a broad spectrum there, so I don’t think we would have any kind of clear classification or name for that.
Jochen: Right. I have a question from Jesus from Spain who I know pretty well and I know what he’s talking about. So, why do we find more and more websites from other countries when we search in Google in Spanish especially from Latin America even when the hreflang is well implemented and there are other websites of interest in the country after the result from Latin America?
John: I don’t know. It sounds like, “Why is Google search broken?” I mean these are the kind of examples that I really appreciate because it is something where we can objectively go to the team and say, “A user in Spain was searching and we showed them a shop that is based in South America somewhere. That’s obviously not going to be very useful for that user.” So those kinds of examples, if you run across that, that’s something I love to forward to the team. With regards to hreflang implementations, that’s sometimes a bit tricky because sometimes the implementation can break in weird ways, and if it’s not on your website you might not see that. The other part is, with hreflang you tell us what your preferred country and language is, but that doesn’t mean that we wouldn’t show it to other countries and other languages. So if you tell us, “This is for one particular country in South America,” and you don’t have a page specifically for Spain, then we might still show that South American page to users in Spain just because it’s the closest match that we have there. But again, if you’re seeing this consistently, if you’re seeing this for generic broad queries where it’s really wrong that we show a version from a different continent to you, then I’d love to have more examples like that.
Jochen: Great. Zohra, I think we have one more, right?
Zohra: Yeah. I’ll let you do that one. I wanted to ask another question.
Jochen: Sure, go for it.
Zohra: Sometimes webmasters are hit during the Google update despite they haven’t done anything obviously wrong. So, in that case which advice would you give to somebody that is hit badly and does not know what to do and what was the cause?
John: Yeah. I think that’s really hard sometimes, and I really sympathize with these kind of situations because it’s something where not everyone knows what is changing. But from our point of view, what we try to do is improve the relevance of the search results. And that could mean that we show certain pages less frequently and show other pages more frequently even though both of these pages, objectively speaking, are both okay and are both good pages. So usually what I recommend in cases like this is to try to look at it more objectively, and maybe take a step back yourself and think about what it is that Google is trying to do for some of these queries. And another thing that I try to do here is get input from other people who are not directly related to your website. So we published I think a blog post maybe last year about some of these core updates, about some of the questions you can ask yourself, some of the things that we think about when we try to improve our algorithms. And it might be useful to go through some of those questions with people who are really able to look at your website objectively, who are not directly related to your website, who can give you a little bit more feedback on where they see things maybe on other websites that are done in a better way that you could do as well. So that could include things like different kinds of content, kind of the presentation of your content, how you’re presenting that you’re an expert in this field. All of these things are aspects where it’s not one meta tag that you need to tweak to make things work again, but sometimes you just need to rethink your approach to how you present your business, your services to kind of an external audience.
Jochen: Maybe a question adding to this even though you might not answer it. There’s been a lot of kind of noise and rumor in the SEO community after the May core update. Can you disclose a little more what was the main point of that update?
John: I don’t really have anything specific to call out on that. I think what is sometimes a little bit confusing, or interesting I guess with some of these updates, is that we tend to talk with the various quality and ranking teams that are working on improving the search results. And sometimes they’re more surprised by kind of the confusion externally than they expect. So some of the changes that we launch are things where we say, “Well this is a really small change – maybe nobody will even notice that this is happening.” And then everyone goes crazy that we contact the ranking team like, “What did you do?” and they’re like, “We just made some small tweaks.” And on the other hand, sometimes the ranking team will say, “Oh, this is a big thing. It’s like so many things are shuffling around.” And then externally people are like, “It all looks the same to me.” So, it’s really kind of tricky sometimes to judge what will be seen more visibly externally versus what will be less visible externally. I mean, we do evaluate how many search results will change based on the changes that we make. So, we know a little bit about what the actual user effect will be, but whether or not that gets picked up by the SEOs is sometimes unrelated to that.
Jochen: Great. I think we’ve only got a few minutes left. There’s one important question two people are asking: Is there a way that we can share your email address because I think they want to file some complaints with you, or what’s the best way of contacting you?
John: Ideally just on Twitter and send me a note there. If there’s something more explicit that you want to send, kind of a big chunk of text, let me know on Twitter and then I can follow you and you can send me that by direct message there.
Jochen: That sounds great. Okay.
Zohra: Maybe one last question.
John: Yeah, one more, go for it.
Zohra: Maybe your last one: What is a question that you would like to get asked but that people didn’t ask you so far and then what would be the answer?
John: What I’d like that they ask. So, I think one of the things I noticed is that people ask all kinds of different questions and I really appreciate that. So it’s not so much that there’s one kind of question that’s missing to me, but what I really like is that people are just always so curious about search. And they just have very unique setups and they’re doing something that’s kind of special to them that’s not something that everyone is doing the same, and I really appreciate that. So, seeing all kinds of questions, all kinds of different questions, is more what I appreciate than people just asking the same question over and over.
Jochen: So, I hope we did not ask all the same questions that you’ve heard a million times.
John: No, this was great.
Jochen: Great, thank you very much John; also on behalf of everybody who asked the questions. Apologies to those whose questions we haven’t asked, but you just heard how you can reach John over Twitter, and so thank you very much. This was the first edition of the Digital Marketing League. A full day. Zohra and I are pretty tired now and I will probably need a glass of wine tonight, so thanks everybody for joining us. John, if we do the same thing in the U.S. do you think you would be willing to do the same thing again, at a proper time?
John: I mean, I don’t have to travel, so…
Jochen: There you go.
John: So that’s perfect, right?
Jochen: Right. We would choose a time that is convenient for Zurich-based people. So, you know we’re looking at doing a similar event in the US so also for all of you we’ll keep you informed and then you’ll be back on again, John. Thank you so much.
John: Sounds great.
Jochen: All right. Thank you. Thanks everybody – have a great day and thanks for joining Digital Marketing League.