 All right. Welcome, everyone, to today's Google Webmaster Central office I was staying at. My name is John Mueller. I'm a webmaster trends analyst at Google, usually from Switzerland, today from Mountain View with a bunch of guests that came over to join us. Would you like to introduce yourself? Sure. Start over here. Hey, guys. I'm John Vantyne. I'm a director of SEO at Truecar here in San Francisco. And I'm stoked to be on the other side of the camera for once today. Hi, guys. My name is Chris. And I'm the senior SEO manager for Instabage in San Francisco. And I'm excited to be here for the Hangout. Hello, my name is Massimo Paolini. I am the CEO of Spectrum Group Online. It's a digital marketing agency. And I'm very pleased to be here as well. Hi, my name is Misha. I work at ZipRecruiter. We're a job site with millions of pages where employers can post jobs and job seekers can look for jobs. And I'm happy to be here as well. And I'm Matthew Williams. And I work at Cloudflare, leading the charge in content and SEO. Cool. Fantastic. Nice mix of people. Most of you, I guess, are from San Francisco? Came down? Yeah. Cool. It's all about 20 minutes. OK. Cool. All right. So maybe we can start. Do any of you who are live here want to introduce yourselves as well? While we're introducing people? Yeah, I can. Hello. I'm Nikol Minkov. I'm a CEO at SEO Agency at Serpact from Bulgaria. We are a local SEO agency. And this is. Hello. My name is Dylan, but short with DDo. I'm the head of SEO of Serpact, a local agency here in Bulgaria. And while here, it's 2 AM in the morning. Yeah. You guys are crazy. We love SEO. SEO is our passion. Hello. My name is Saidu. I am a CEO manager of FieldClicks. I'm from Sydney. Wow. Cool. That's pretty far. What time is it in Sydney? It is 11. Oh, that's reasonable. That's so great to me. I'm in Adelaide, which means it's 10.45. So not quite as far. I'm an SEO consultant, and I also help out with the Google Web master forums, which John runs. Cool. Very day ahead of us in the future. Hello. My name is Alexander. I'm a technical deal at about you in Germany here. It's one o'clock in the morning. Crazy. I try to do them in European daytime, too. So OK. That's good. So OK. That's good. All right. Shall we start off with a question from one of you, maybe? Sure. I mean, I guess mine might be sort of a quicker, simple one, but I had a question that you, I think you've touched on in the past. Someone asked about file names for sitemaps, and you said the file name, or I think you said the file name shouldn't be that important. It shouldn't matter. But we're seeing a situation now where we're generating sitemaps daily over 1,000 individual XML sitemaps for this massive website. And the file names change every day because the date is included in the file name structure. So could that be problematic? Since Googlebot may not even see all these file names every day, they might not know which ones they crawled, which ones they didn't, and we're moving it over to the next day. The sitemap index file references them each day. So Googlebot would know where to find these, but should we try and keep a static file name rather than having dynamic ones each day? Yeah. I generally tend towards having a static sitemap file or mostly so that in little URLs try to remain in the same sitemap files. OK. So unless you're generating new content every day, that amount of new content, then I try to keep the same sitemap files. Well, we are. It's vehicle inventory, so it's highly ephemeral. OK. Well, if the URLs are changing every day, if you have new content every day, then of course you have to generate new sitemap files over time. But if you have existing content in existing sitemap files, then I try to keep those as stable as possible just so that we can re-crawl those sitemap files as quickly as possible to really kind of keep up with the changes rather than spending time crawling sitemap files that don't exist anymore. I see what you're saying. OK. Can you say content? Do you mean the content on the pages themselves or the URLs in the other sitemaps? The URLs in the sitemaps. Yeah. Good point. Yeah. Cool. All right. All right. I'll jump in. Go for it. So here's a question for you about CDNs in general. So there's this back and forth, like I mentioned before you started recording about how with CDNs, some people go on and say, hey, Cloudflare is damaging my SEO because I've moved on to a CDN and now it has a shared IP address. Do CDNs in general damage SEO or are there circumstances in which a CDN can be detrimental and should be avoided? Good point. From our point of view, we try to treat CDNs more as kind of a hosting platform and in general you can host your website wherever you want. So if you use this platform or that platform that doesn't really matter for us. What does matter for us though is the amount that we can crawl from kind of a hosting platform and generally when we see that a website shifts the hosting that they have to some other platform we'll be a bit more conservative with how much we can crawl from there. So we'll usually kind of step back a bit and say, oh, we're not sure how much load the server setup will actually support. Especially if it's a bigger website that we've been crawling really hard and we'll probably kind of step back a bit, not crawl as hard and then over the next couple of days or weeks we'll kind of ramp that up again to what we think is a reasonable load that will not cause any problems. So that's one thing that sites sometimes see when they shift to a CDN or when they shift between different hosting platforms in general. But that's with regards to crawling and crawling is of course kind of necessary that we can actually index the content but if it's content that's already indexed then we don't need to re-crawl it all the time to actually show it in the search results so it's not that your site would rank lower because you moved to a CDN or because we're crawling slower, it's just we're trying to concentrate the crawling at the moment and the URLs that we can pick up new, those are the ones that we can index of course and if there are things that we can't get to because the website is so big and we're not sure how much we can actually crawl from the website then that's something where you might see it. But that's totally temporary and that's also something where if you have a really large and dynamic website where you start seeing this when you move to a CDN or a different hosting platform there's a form in the help center that you can fill out I think report a problem with Googlebot where you can say well I moved to a new infrastructure you can crawl me with I don't know 10QPS and then the Googlebot engineers will take that and say okay we'll like twist the dial here and kind of make this catch up a little bit faster. That's great information. And you can do that the other way too if we're crawling way too much and I don't know your hosting platform is set up in a way that it just provides all this capacity and you just have to pay for it and you're like oh wait Google is like killing my wallet then you can also go there and say hey I need you guys to crawl a lot less and what really helps is if you give us concrete numbers what I'm okay with rather than just saying well you're crawling too much or you're not crawling enough because then the engineers are like okay maybe like 10% more or like 100% more, what do you want? Great job. And a related question if say your images on the CDN and the CDN is a separate domain to your own domain would that have a negative on the association of the images with your own domain or anything like that? No, that should be completely fine so that's something if you had it set up like that that's the way it is, that's something that I think makes a lot of sense for websites that have big images or big static files in general to put those on a CDN where you can serve kind of static content a lot faster the only tricky part is migrating to a setup like that so if you have images on your own domain and you want to migrate them to a CDN to a different subdomain or a different domain completely then since images tend to be re-crawled a lot less frequently it can take a bit of time for that actually to happen so that's one thing with images that's a little bit different than with web content in that we crawl images frequently than web content and for images we always need the image and the landing page to be indexed so that can happen if you shift to a CDN or shift to a different host name for all of this content it takes a certain amount of time to actually catch up but afterwards that's totally fine to have it set up like that so it's yeah I think that's a good thing all right so I'm trying to get the questions that were submitted working I think I'm close but maybe we'll just take some more questions from you all until we've got that set up yeah so I wanted to talk about the upcoming mobile first indexing and mobile speed update that Google just announced so we're undertaking a big migration right now to new stack and we're wondering which technologies which should be adopting that would help us on mobile rankings and indexing and specifically we're looking into AMP and PWA so my more specific question is like do those affect indexing and ranking in Google's mobile algorithm yeah that's a complicated set of questions like a few words so in general when it comes to mobile first indexing what helps us a lot is if you have a responsive design where you have one URL which serves the same content because then when we shift to the mobile first index where we index the mobile version or the version that mobile phones would see then nothing really changes responsive design is the same HTML same URLs you don't have to change any of the links you don't have to worry about rel canonical all of those cross links they all just continue working so that's kind of I think one set up that I'd recommend if you currently have separate mobile URLs then that's one direction I'd certainly head you could potentially take that a step further and say I will move completely to AMP for example and say my mobile version is actually an AMP page and that's something that we've seen some sites do you can do that on a per page basis too so you can say well the mobile version for this part of my website is on AMP and other templates are normal of HTML when you have it set up like that then you have all of the advantages of kind of this inherently fast platform with AMP also with the caching that's available if it's on AMP that's an option there if you have AMP set up separately to your mobile to your normal website then of course you have those two URLs that we have to crawl and we have to index and recognize that they're connected sorry for the question about that so for the speed component if you had AMP as your canonical page would the speed be based just off of the AMP page then if AMP is canonical then everything is based on the AMP page so that's one option to kind of look into to see maybe that makes sense and if it's not so if we have a regular page yeah with the regular page and the AMP page what happens there is we index the regular page as the indexable content and we use whatever mobile version users would see which would probably be the AMP version as a basis for the speed so that's one thing where of course AMP pages often are very fast by default you can make slower AMP pages non-AMP pages so it's not that you need to be on AMP to kind of have this speed effect but it's an option there pwa I think is taking things to a completely another level it's not something where I'd say you have kind of a small tweak in that the HTML is just in a different format but rather often you have a completely set up which is a JavaScript based framework that provides all of the functionality of PWAs but that also means that there's a lot more complexity involved when it comes to kind of having an indexable site where you have to think about does Googlebot need to actually render the JavaScript do I need other search engines to render the JavaScript do I need maybe a pre-rendered version for some kind of services and other versions use the rendered version it's something where there's a lot more complexity involved so that's something where if you have an engineering team that can deal with this kind of complexity from JavaScript side then that's definitely an option to make a really fast site but it's not like at the same order of magnitude when it comes to work as AMP where you might say well I'll just change my HTML templates to and it'll just work with a PWA there's a lot more complexity involved would you say going into 2018 and major trends in SEO would you say that AMP is one of the more important things that most sites should be focused on if they have limited resources and they want to focus on improving one particular area you can I think that's definitely an option that the nice thing about AMP is that it's a really fast framework so it's something that offers a lot of kind of fancy optimizations by default with like the lazy loading stuff with all of the caching that's enabled the embedability it offers a lot of really neat functionality that you can kind of work with you can of course do all of that yourself with your own framework you can set something up yourself but it's a bit harder so I think there are a lot of opportunities there it's not that you need to do AMP in order to be visible in search you can do anything to be visible in search but it's a really fast framework there are lots of people working on that to expand that at the moment so I think that's probably one way where you can get a lot of value for little effort I have a question about PWA so is the crawler like looking to see if a page means all the requirements for PWA or should we be thinking about the different components and how they help things like site speed and other factors it's not that we would treat a PWA differently at least at the moment we'd say this is a PWA where we'll do these things it's more PWA has a set of functionality that they use to define that it's actually a PWA but that's more almost like a branding of those functionalities and from a crawling and indexing point of view we basically just look at that and say oh it's a JavaScript based site and we need to use rendering to view the content so we'll do that like we would with any other JavaScript framework hey John could you touch on dynamic serving how do they optimize their site and cut off about 200-300K using dynamic serving versus responsive but obviously Google says they prefer responsive so do we go back and then make it responsive or should we stick with what we've done I think if you set up dynamic serving properly that's fine as well what is important for us which I suspect some sites have done differently in the past is that all of the kind of metadata and everything is available on the mobile version as well so if you're using dynamic serving and you're stripping the page down into something that's purely the text and the CSS that's required to render on a mobile phone then things like structured data might have been stripped out so that's one thing just to kind of be aware of with a responsive design you don't have to worry about that it's like all the same thing but with dynamic serving you kind of need to test for that and use the mobile version to test for the structured data to test that the images are all in there the videos are properly embedded all of that as well okay yeah I got it and on that too do you think and maybe this is more of a CDN question but you know if we have two versions kind of then the cache only gets hit if they access mobile or they access desktop versus responsive where either when they hit it or cache it do you think that impacts load time or rankings from load time in that way I don't know you'd have to test it okay I'm pretty sure that the common CDNs they can deal with the usual dynamic serving setups right yeah I mean the big key is the more you can cache the better, the more you don't have to go back to the origin to fulfill the request the faster your response rate that's going to affect you different amounts based on how large the final sizes are okay John yes I have a question if we have faster mobile responsive version of our website do we need an AMP version for our website do you need an AMP version you don't need an AMP version so it's not that it's a requirement or anything but it's a way of having a really fast page so kind of the setup is really useful but AMP is also a lot more than just the fast page it's kind of allowing this embedability that makes it possible for us to pre-render the pages when people come from search so that when they click on the result it's already loaded it's already available and similarly that apps like I believe Twitter for example they also use AMP pages and they cache them as well that they're able to pull in this content quite a lot faster than if it's just a fast mobile page so AMP is kind of the most visible part is definitely the speed aspect but there's a lot more than just speed that comes to and is there anything like this that AMP version is only important for the blog type of website or is it for all type of website like for e-commerce website if we have product page with a lot of images should we create an AMP page for those pages or we only focus on the blog post created pages you can do AMP on a per page so even if you have different parts of your website that have different suitability for AMP I'd say you could still use AMP on some pages and maybe not use it on others so that's one thing to keep in mind it's not that you have to invest and say I will switch everything to AMP or nothing to AMP you can try it out and see what works what doesn't work there are a lot of things being done especially for e-commerce with regards to AMP so I suspect that maybe in the beginning it was very easily suitable for blog or kind of new static content but there are lots of other options that are coming out that are making things really interesting for AMP as well so if at the moment you're looking at it and you're like I need this functionality and it's not available then maybe it's worth either submitting an issue to GitHub and say I need this and then I can switch to AMP or maybe just wait a couple of months and see what has been happening there I know the folks that are working on AMP it's an open source project they're moving extremely quickly so looking across different teams at Google they're one of the groups that are really roving high speed adding lots of functionality and making sure that this is useful for everyone and one more question the AMP page for the website the AMP page has both have the same content but it always has a separate URL with a domain name like something like that so do we need to create a separate site for AMP page or we can put it with the same site you can put it with the same site file that's perfectly fine yeah I have a question about pagination Google has guidelines for SEO friendly pagination and the relationship between conicals indicating the relationship between all these pages in the series but if we see a lot of pages in the paginated series being indexed and sometimes receiving search traffic I've seen some sites I know Reddit has no follow links placed on their pages deeper so page one Google can crawl they don't let them in the subsequent pages would you say there's an advantage to doing it one way like Google decide and sometimes send someone to page 100 of product listings or do you think it's best to maybe cut it off and just focus your authority on that first page in the series I haven't looked into the details of where we do it like this and where we do it like that I think offhand I try to treat that as something that depends on the website in the sense that a lot of sites will have different categories but actually the content itself will be available on multiple places so if you have I don't know a shoe store then maybe you have like under men's shoes you have this really long list but actually they're also listed under running shoes or under these kind of shoes or blue shoes red shoes or whatever and if you have all of those different categories then we don't need to go to page 100 because we'll find links to all of the content maybe in the first couple of pages so that kind of depends with regards to sites that are more kind of news oriented where there's like one long listing that lists everything then that's something where I don't know maybe it depends maybe it makes sense to go deeper maybe it makes sense to kind of take that structure and say okay I'll create an archive section of my website and just link to it from there and that's a bit harder to say for us an issue that we had when we did have page index is that we have search results and sometimes it's jobs so sometimes they come in and out and so page 10 would be index and then two days later there is no page 10 because we don't have that many jobs for that particular job title so then the user hits a page with nothing a 404 or whatever so we decided to remove page 19 and pages from it yeah I mean automotive is ephemeral in a similar way where inventory is coming and going so same issue there that makes sense so how are you all using AMP are you we tried it out we have it on our blog for now we've been exploring search result pages but haven't implemented that yet we use WordPress so our sites on WordPress we installed we've had it since it's improving all the time and it happens automatically so we have to do it absolutely yeah the plugin is great we're going to start playing with it we're going to see if we can tweak it and get more of our branding into it and what not so cool I'd say we're susceptible to improvement on our AMP front it's one of the things that we're dedicating our time to so you have AMP pages but they're not that great yet on the main marketing site I don't think we have AMP pages I think we have them on the blog as well but it's one of those things that it's a matter of picking and choosing and it sounds like this is a thing to pick for us we have it on the blog we tried it this summer for our ecommerce type pages search results listings pages, sort of like ecommerce and we sort of hacked it together so we didn't see we actually saw those pages lost some traffic and so we pulled it back but we're going to try again thinking that we're going to actually do a full rollout this time so we're doing it again that's what we're at that's interesting to say that because at AMP page we have the AMP on blog the blog on AMP we've actually found some of the pages where the user metrics aren't as strong as it was on the mobile response page the AMP on page is down and the bounce rate is higher but obviously we feel like there's just no going back you just have to go forward and optimize as best for AMP but are you seeing anything where there's some type of metrics where it could potentially be not as strong as before so for the most part the issues I see there are that it's hard to track the metrics when it comes to AMP because depending on the way users go to AMP pages it might be on a cache somewhere it might be hitting one set of analytic setups you can compare that to an analytic setup in a completely different way on the rest of the site so it's really tricky to measure that properly also when someone goes from a cached AMP page to your actual blog by clicking around on other pages how do you track that? is a new visitor on the blog is that a bounce on the AMP page so that's I think that's one aspect that's a bit tricky where in talking with people they mention these things and they don't know themselves what exactly is being tracked and probably that's something that we can improve from the AMP side to see try to find a way to create I guess more coherent metrics for these pages regardless of where they're actually located and loaded but I suspect that's not trivial to do that would be super useful just for the record I've heard of early adopters that removed AMP and then realized subsequently that they weren't tracking it properly maybe it was more effective than we thought it was yeah that could be us I'm talking about AMP will the new Google search console include the ability to filter pages by AMP? I don't know should it? What would you like to see there? well so when I when I look at a CEO results I look at pages more than a keyword so I look at a page, how is that page doing then I look at what's on the page what is Google matching to it and here I'm seeing that there are several AMP pages being served and how they're performing if I wanted to only look at that that would be how I use the filter how are they then I could see my data aggregated for AMP so how is the position of those pages changing and whatnot okay that sounds interesting I don't know what the plans are there so in Search Console there's the feedback link and the team is actively following pretty much everything that's submitted there and tracking that so if there's something that you'd like to have in the new Search Console make sure you submit it there also if there are things that you really like in there they're always happy to hear that too let's get to that so I have two more generic questions one of them is directly related to WordPress is there a difference to Google between a page and a post no didn't fix that so even the category I mean now that people argue with me about the category and the fact that they can have tags and all that perfect the last question has to do with updating a content on a blog or on a page I feel that if we update the date to the current one because we made changes it's misleading because it doesn't have the original date of when the article was published yet if we update it with current information that date doesn't reflect that so what's the best practice I think that really depends on what you like to do so I know that that's sometimes tricky in the sense that sometimes you just have one date on a page and Google has to figure out what that date means but for the most part we try to figure out like what the significance of this date is and does it make sense to show it to the user, does it make sense to treat this as new content or updated content so we just see it as a date that was updated I was looking at some results attached to a page specifically and another wish is for the search console is the ability to open a page directly from it I find a trick around it I find a workaround but so oh yeah so when we looked at the page and the results our result was a middle of the page and I started looking at the sites up top first of all there's one site that had three listings on one page the main diversity not happening as much or that can happen, that makes sense interesting, okay and all of those articles have newer dates than our article which makes sense it's fresher content therefore they would be up higher so I don't know that article we do need to adjust it we need to refresh it I think that's totally up to you totally up to you so I think the ranking of fresher content higher that really depends on the query and what we think the user expects sometimes they're looking for reference material and it doesn't make sense to just show something that's fresher sometimes they're looking for something new and updated then maybe it does make sense to show that a little bit more and kind of those expectations they do change over time so that's something you particularly see when some event happens in I don't know an earthquake in California hopefully not this week but it can happen where people when they search for California now maybe they want the Wikipedia page that was written I don't know a couple of years ago essentially but if they search for it after the earthquake then they probably want the news, they want something new and that's relevant to that event that recently happened so these things they definitely shift over time so dovetailing on that so if you is it better than because you want like updated content let's say there's a DDoS attack then what is DDoS might not be as important as how are we dealing with this particular set of circumstances does Google then see the domain is generally more reputable around that topic if it seems to follow trends providing quality content in a particular direction as events unfold I suspect that that can help but it's not that we have kind of this I don't know concept of this domain is really awesome and we should kind of I don't know see it as something that's valuable in the future as well but rather we try to look at it at a single point in time so if you've been following the trends for the last five years and you suddenly stop then that doesn't mean that your website is still relevant for something new that happens but it's certainly something that collects value over time kind of like you have a good website that people refer to all the time then that does build a little bit of trust and kind of expectations from users as well so that's I think that's a useful thing to do but it's not the case that you can just say well I was really good for a couple of years now I can just coast and it'll still rank number one so in our in our case we have these pages like the search result pages where you know again it's jobs they come in and out all the time and so the page is basically changing every every few minutes let's say so we're not like trying to submit a new sitemap to google every time you know the list of jobs on the page changes but sometimes we'll make changes like the title tag and the description or we'll make a change that temple itself so is there a way to like sort of notify google that it is like an actual significant change to the page rather than just like regular content updates if you will I would use a sitemap file for that so the last modification date in the sitemap file that's really useful for us especially if you've been using that as a way to tell us about significant changes on a page so that's really something that we pick up for crawling and trying to recognize bigger changes on a site or on a per page basis what's problematic for us with a sitemap file is if you tell us like every page has changed all the time my website is hosted with PHP and the sidebar changes every time you pull up a new page so therefore every page has changed every second and we look at that and we're like well if everything is new nothing is new let's ignore that completely but if you give us reasonable information say well this is something that actually changed then that's useful to have so I guess in our case when we do make a change it's like across all the pages at once so it would be like let's say hundreds of thousands of pages let's say they move on to a new template or make some big change but it's on all of them that would still be okay that's fine but just not like all the time so do you use priority as well? No we don't use priority at all we don't think so yeah we announced that in the beginning we thought that might be useful to get some kind of input from the webmaster on what they think which one of the pages are important but it's something that's very tricky everything is important or they take some plugin to actually provide the priority information and then the plugin is like well these pages are like this these pages are like that and in that case we can figure that out just as well as a plugin so that hasn't turned out to be as useful as we thought initially so not to change the subject completely but internationalization we're at this position where we're now breaking into these international sites like I mentioned before we started recording we have cloudflare.co.uk or whatever these different sites are and we want to make sure that as we we localize we have the right canonicalization factor so we don't damage our primary domain by siphoning off the page rank is there steps we can take are there best practices to implementing that globally? I don't know so the tricky part there is there are lots of things that you can do so you can go to help center and has all of the information for localization the hreflang the geo targeting setup there are lots of options that you can do but figuring out which ones of these are relevant in your case so that you just don't go overboard and say I will do everything and make a version for every country in every language because then you do run into that situation where you're like diluting yourself more than you're providing value in individual areas so finding that balance is something that you almost have to do intuitively almost to figure out where can I really provide significant value in a niche that I haven't covered so far which makes sense to kind of separate out so if you're talking about the UK and the US then maybe it makes sense to treat those as one version just say well this is English it's like English content is fine for everyone but maybe it's the case that UK users have very specific expectations with regards to CDN maybe I don't know with privacy or whatever kind of laws they have in the UK at the moment maybe they have different expectations than people in the US have and then it does make sense to kind of separate that out what if you're like dealing with content duplication let's say you have like the Spanish version or you have variety of different things but you have images that are the same you have some assets the same and then you have 11 domains that are duplicating content can that be penalized or seen as you know flag by panda or something like that as superfluous or is that generally understood that you'll have different languages and do some elements that are duplicate usually that's fine but it's again kind of the situation where you have to be careful that you're not diluting everything too much especially if you're kind of taking content that was really strong in the US market in English and you're like saying okay I'll make different versions of that for different countries also in English then suddenly kind of the value of your English content is kind of spread across all of these different versions and if you were ranking in English in the US on the basis of one really strong page and suddenly you have a bunch of pages that are not quite as strong anymore and that might kind of shift things around that other people are able to rank instead of you so that's something kind of to keep in mind and usually what I'd recommend if people at a conference or someone asked me I'd recommend finding some consultant who has experience with this because this is something that you don't do often it's not something that you can easily tweak afterwards so finding someone who has experience working with large international sites and kind of where it makes sense to kind of invest time in and where it makes sense to kind of step back okay maybe one version of I don't know German content is fine we don't need to do it in different country versions finding that balance is something that sometimes a bit tricky and having someone who has done that before is really useful all right let me try to go through a few of the questions that were submitted so that they weren't submitted for nothing can Google understand arrows in PDFs to show useful information so I looked at the screenshot that was there as well and from our point of view we try to extract the text from a PDF and use that for ranking of the PDF itself but we don't kind of try to extract a kind of structure of where words are connected and what diagrams look like so at most we might pull that out as an image and say this is an image within the PDF but it's not that we try to understand that this piece of text has an arrow pointing to another piece of text we're a business with a number of different branches and we use schema markup when we look at similar companies who have marked up theirs everyone's a slightly different and even though they all seem to go through the testing tool can they all be correct how does Google deal with this I think it's something where we kind of have to look at the options that are available with schema.org markup and the options that we support for search and to pick the ones that actually work for you so sometimes there are different ways of doing the same thing sometimes there is schema.org markup that we kind of like accept but we don't use it for anything within search and obviously the validation tool will say oh this is fine but if we don't use it for search then you're spending time doing things that don't really have an effect so I'd recommend looking at the search developer documentation and picking from those types that are in there which ones actually make sense for your individual pages when you say used do you mean things that result in a rich result because isn't there some merit to marking up things that don't visually manifest in the search results so there's definitely some value in providing schema.org in general for us to help us to better understand the pages but if you're limited with the amount of time that you have available you don't just take all of your pages and mark up all of the entities you kind of have to focus on what actually provides value for you so that's one thing where we don't mind if people mark up more on their pages than we actually show directly in the search results but it's not the case that you will see a real clear reaction based on that. Some people try to look ahead and say well this mark up might be really useful for voice based assistants for example and they might say well I suspect Google will start using this kind of mark up maybe next month maybe in a half a year and they'll start implementing that ahead of time so I think that's something you could definitely bet on these things but it's kind of hit and miss maybe we will maybe we won't so if you have schema mark up is it one of those things where generally can be an indicator of a site that's assuming the mark up is done correctly is more optimized or is it generally just it's either going to help you appear on a rich snippet or not and it's really sort of a moot point otherwise we don't use it as a quality factor so it's not that we would say this site has schema mark up therefore it's a better site but sometimes we can recognize entities better if you mark up entities on a page which is kind of a more flexible type thing where if someone gives a really long query to us and we recognize oh they're asking about this entity and not using exactly those words and we know this other page is also about this entity maybe doesn't use exactly the same words then we can still match that and that helps us to try to find more relevant results together but the most visible effect is really just the rich results in the search results John, one question from here Alright, go for it Thanks, John, like example we use amp technology one year and more for the first time we integrate amp for our both posts like all of you but after 6 months before 3-4 months ago we tried to integrate amp for our website but the problem not problem but the normally amp give us once with more slash amp we use WordPress and after the great conference in Munich about amp technology we start to migrate our website to full basic amp without slug amp now I have a question, we have a lot of index page with amp slug can we what will be the best practice in this case to redirect this amp slug to the normal slug or to keep it I would redirect it if you have a new version of that URL then I would just redirect it Okay, thanks Alright, let me see what else we have here our e-commerce site has a YouTube channel showing our products with the Panda also look at this when taking into account and valuing our website as a whole usually our quality algorithms try to focus on the site itself so if you have a separate site essentially which could be like a YouTube channel which could be a blog or Twitter account whatever then we would treat those as separate entities so just by having a good YouTube channel doesn't mean that your website will rank better of course if you have your YouTube videos embedded in your website then there's the availability of video snippets showing up which are sometimes really useful but it's not the case that we would say these different silos that you have because they're all there that we would treat all of the website as something more valuable Can you give any tips on how to rank for voice search without impacting desktop mobile rankings how should we amend our content to cater for both I don't know I think it really depends on the website and it's tricky still in the sense that it's very early days for voice search so there are different things that we're trying to do there are different things that users are trying to do to get information one thing that I saw recently is we sent out a lot of emails to I believe sites that have podcasts and recipes specifically to get them to verify within I think the Google Assistant setup that might be something that makes sense if you have this type of content that we can show within provide point users within voice interactive model another thing that sometimes comes up are the featured snippets where if we recognize that your pages have a good answer to a specific question that people have then maybe we'll pull that out as a featured snippet on desktop and mobile and maybe we'll use that as an answer on voice so that might be another thing to kind of look into but I think it's really very early and it's hard to say how that will evolve it might be that at the moment everyone's thinking about voice but the final state is more that there's this mixed mode interactive thing where you have a screen and voice and maybe that's what will actually work out better and you use schema to mark it up so that for example the recipe works with home for example I don't know how it actually works I know they emailed a bunch of sites that use the recipe markup but I don't know how that actually works I saw something last week saying about it and there was also another one that talked about doing the same and prepping content for Google help but you also don't need schema to get a rich result so it's possible that these answers are coming from non-schema sites I think for recipes it makes sense to have the markup there because there's a lot of text around it sometimes and getting the ingredients and the preparation and the metrics around the recipe like how many calories, how many servings all of that that makes sense to have in some kind of machine readable way but I don't know how that ends up being used in practice part of the problem is in Switzerland a lot of these things are not actually launched so I have a Google home at home I have an Alexa at home but it's way different than what you would probably see here interesting because I did a Google home reading the recipe and I used it and it worked ok cool hi John, quick question alright so it's about the site navigation how important is the site navigation for Google in other words I mean more complicated and descriptive navigation or something more user-centric and more simplified how Google treats these two situations for us we usually don't look into the usability as much but rather kind of the crawlability for us the more important part for us because we need to be able to start one place in the website and actually reach the rest of the website in a reasonable way so that means things like having links on the pages is really useful so one pattern that we sometimes see especially for reference sites is that the navigation will be mostly through a search box where it's like you can look up like a list of items and the search box is actually how you find the individual articles and obviously for us that's terrible because we don't know what to search for to find the content we don't have links to kind of click through but if you have links on the pages and you link between related pages you have kind of a category structure then that makes it a lot easier for us to start anywhere on the website to kind of crawl out from there and reach the whole rest so on that note in terms of URLs something I brought up earlier but we're revisiting our URL structure and we're kind of considering two options one is more like a thematic URL structure where different templates are grouped in folders by theme so let's say you have a nursing job and there's like a search results page and then a bunch of jobs are under the same folder or doing it more by templates so that a certain kind of template type is under one folder structure and then other templates and then another the reason for that would be like they have sort of different user engagement metrics like bounceware and things like that so it might make sense to kind of separate them so that it's easier for Google to measure which group of templates what would you recommend there from our point of view it doesn't matter you can do it either way that's really something where we try to see URLs primarily as an identifier and at most we might want to recognize patterns that are tricky to crawl so if we recognize there's an infinite crawl space under this I don't know, sub directory or kind of sub pattern that you want to watch out for but that's more kind of from a technical point of view and less from a quality point of view so if you have a structure that works well for you that's what I'd use I from personal point of view I try to see it more as something as what do I want to measure and how do I need to have a structure that allows me to measure that so if you're trying to separate things out by theme and having it different URL structure by theme makes sense for you to compare the performance of these different types of jobs then that might make sense whereas if you just have in an extreme case you just have a hash as the URL then obviously you can't use analytics and say well everything that starts with this is like compared to other ones that's a lot harder we could use content groups and analytics possibly I don't know I don't know analytics that well but it's something where I look more at other issues rather than SEO so that might be tracking, that might be usability that might be whatever makes more sense from the infrastructure point of view if you have a URL structure like trees, is there a notion that the stuff at the bottom of the tree strengthens the stuff at the top just based on the URL structure not really I sometimes get people confusing the URL structure with the navigation so you get questions like is the folder structure important my answer is typically no would that be right that you don't really care about folder structures yeah so for crawling indexing ranking it really doesn't matter but again sometimes you care about the folder structure and need to have it in a different way so that you can measure things or track things or whatever would that also be true for an eco mass shop where you have things that are organized in folders like trousers, short trousers something like that yeah that usually works fairly well there too the main thing that especially with e-commerce sites that I kind of watch out for is anything around faceted navigation that you try to put that more into URL parameters rather than into the path so with URL rewriting that makes it harder for us to recognize that this is actually part of faceted navigation and maybe we can crawl it in a more optimized way so things like faceted navigation pagination, I definitely put into URL parameters also if you have any kind of session IDs definitely put that into parameters as well rather than within the path okay may I know we have a large scale e-commerce site with mobile alternates so all the stuff that makes your life interesting and it's about data consistency we're having trouble to submit the data to google in fast enough for example with site maps we generate them and by the time google comes and sees it and crawls the pages on no index and generates errors does this have negative effects it doesn't really have negative effects that we think the rest of your site is bad but obviously we won't have indexed those pages so yeah that generates problem with hreflang text for example so you don't have an index if they're not indexed then the hreflang doesn't need to do anything there I think that's something where I would try to figure out in a case like that is what you can do to optimize crawling in general which could be to simplify the URL structure to try to kind of exclude all of the pagination and the faceted navigation parts out which could also be to think about what you can do to host on a faster platform so that crawling can be done a lot more faster so that we don't run into this situation as often that you submit something and by the time we get to it you've completely changed what this page is about so kind of improving the efficiency of crawling would be something that I try to look into there but that's probably also one of the harder things to do I think improving the efficiency of your URL structure internally that's really tricky but I think it makes a lot of sense to focus on that especially if you're a really large website because then you do see bigger effects if you can exclude certain parameters and suddenly you have one tenth of the URLs on your website then that does make a big difference there's a question generally about SEM and SEO and how the lift works together so when you spend more on SEM can you, does that also potentially affect your SEO rankings where you now have high clicks because people see it at the top they click through and then it carries some page rank or is that separated out and you don't get the lift from some of the paid search in other words what's the interplay between them we treat them as separately as possible so that's something where from our point of view it's really important that what we show in the organic search results is not affected by any kind of financial relationship that's in place so whether or not you buy Google ads that's totally up to you that shouldn't affect at all how we show sites in search so any kind of effect that you would see from people going to the ads and the ads getting high click through rate and I don't know more visibility maybe that's something that wouldn't have any effect at all on search what of course can help there is if you're starting out and you're trying to kind of bring your name out there and people see your content and they're like oh this is really fantastic service and they start recommending it because they saw it in some ads which could be anywhere, it could be on TV that's something that obviously indirectly could have an effect but that's kind of like just normal marketing it's not tied to any specific advertising that you're doing another effect that I believe happens is if a searcher sees an ad on top and then a listing either directly under or middle of the page there are some people are more inclined to click things on the organic than on the ad I have no idea so that's possible that there are usability studies like that this research in 2013 somebody internally in Google put something together more recently they had another one that came out so originally the latest one said was 50% more if you're on the top and over 90% in the middle I have no idea so you're saying if you're ranking lower you get more clicks basically it feels yeah interesting and I did find that that's pretty cool I have no idea so I have no insight at all we really try to keep that separate also internally when anyone from the ad side contacts us and says oh I have this really important client and they're not going to buy any ads anyone unless we solve this SEO problem we can't help them that's something where we really need to say well this is a different part of Google we want to make sure that organic results are really independent of any financial relationship there and we need to clearly say well they can come join the hangouts they can post in the health forum they can use any of the other public channels but we can't give them any special treatment with your own analysis do you take into account if ads are there or not for example the effect of click through rate and you might be using that to decide how good search results are so do you kind of go there's three ads above things might be different as far as I know we don't use that directly for search anyway we do use the user behavior as something when we analyze algorithms so when we have different algorithms that we're trying out or different settings for algorithms we try to figure out which one of these actually works better across millions of searches and for this kind of A B testing I don't know what they would do specifically with regards ads I suspect they try to treat them just as organically as possible and say well some of these results have ads some of them don't and we need to kind of look at the bigger picture and look at it across really like millions and billions of searches to get like a real world feeling for that that's not something that you can just look at one search results page and say well this is like this and this is like that but it's interesting because the search team is running experiments all the time it's something that they're like when you search you're almost always in a number of experiments which might be things like small pixel shifts and the search results might be some subtle color changes sometimes it's subtle ranking differences that's something that's run all the time and it's in talking with these people and they have really surprising effects that they don't expect so looking back for example when we started with the mobile friendly label that we show in the search results we tried different variations of like saying this is mobile friendly or not mobile friendly or showing icons or different colors and we assumed that things like a green phone would be useful as a symbol rather than writing it out and it turns out people were confused about what time for people to actually realize what this meant that it's not like it will call someone on the phone but rather it's suitable for your phone and these are all things that you have to try out so I think even any other website you have to constantly be testing and trying things out even things that you assume don't have any effect at all or that you assume you know exactly what the effect will be because sometimes it is surprising so I do have another question last time I was in the meeting we talked about the fact that we were seeing more Google generated meta descriptions and I have seen some examples of some pretty hard fails there for example a technical client that has a page with a ton of specs on it we're now seeing in Google search console multiplications as a keyword that was matched to the page it doesn't matter obviously if someone clicks it they won't go there but it's obviously when I go and I do a search for that query I see the Google generator thing as just the numbers from the table and it doesn't make sense so where do I put the feedback for that you can submit it in the search results on the bottom there's a feedback link that's really useful if there's something specific especially if it's a bigger theme that you're seeing not like one search results page that is bad then that's something that could go into the help forum that could be something that you could send me directly as well usually for both of those channels it's not something that the search team would get back to you and say oh we tweaked this we found this bug or we changed this but rather they take this feedback and they try to use that to tune the algorithms for the future and that might be that they change it right away it might be that the next iteration that they do a year later they take that into account I'm fine I don't find that getting the feedback but giving it in this I think that's really useful so that keyword density question that was mentioning earlier even large enterprise platforms that give advice professionally for SEO will say you should have X number keywords of an instance on a page and that's like a hard line if it's 5,000 words if it's 100 words on the page and Matt cuts out a video from 2011 where he says that's not really how this works right but they still give this advice can you comment on in 2018 is it basically the same thing as it was in 2011 and that's not the way to look at it yeah I wouldn't look at keyword density I don't think that's a useful metric what I would look at though is if you mention your keyword at all so not necessarily one to one the exact keyword but kind of like what people are searching for that you want your page to be found for and that you at least mention that on those pages that's kind of the level that I would look at there so sometimes when we do site clinics we see websites that are really fancy really nice but you look at them and you're like what are they trying to offer and like what would they want to rank for like as a human you can't figure it out and that's the kind of situation where if you at least like say well I don't know a service provider for this then at least search engines will be able to say oh yeah a service provider and this topic that they're working on that's a good thing to match this for whereas if there's just like we're the best company for I don't know it's like in California small businesses is like okay what do you actually do and maybe on the images it's kind of clear what they do but when we look at the text if we don't have that at all that's a bit of a problem especially I think with the higher end of small businesses that's a problem because they have enough money to hire a graphic designer to actually make a website but not enough to actually figure out what this website should be ranking for and what should be placed on the website I have a question about the search console so we found that there's a huge difference between the new beta and the old one I guess we'll call it in terms of the number of pages that we have indexed it's like a 3x difference to old one being higher and that also doesn't match when we do like a cyclone on our domain so like which one of those numbers trust worthy in your opinion and how fresh is it typically that number how fresh I don't know how quickly they update that data at the moment I think it's something like weekly I don't know it might be that they're still kind of fine-tuning that so the site query numbers are definitely numbers that are more of an approximation than anything else that's something that we really tune to be an approximation so we try to get a number as quickly as possible and that can be off by a factor of 10 or a factor of 100 so it's more like well you have a lot of pages or you don't have a lot of pages the number in search console for the index status count I think that's a lot closer I'm not so sure of the differences with index status composition I think the new one does call what the difference is there would be the most exact one that I would use is a site maps index count that's something that's updated immediately that's like on a day-to-day basis we look at what is actually indexed based on what you submitted which I think is a useful way to look at it because we index a lot of pages that you don't really care about those would be counted in the graphs but that's not really something that you care about we found an infinite calendar on your site and suddenly you have a million pages of index more it's not going to bring any traffic because nobody's going to search for events in 2500 it's like quick question about fresh data this is a little bit off the topic but in Google keyword planner do you know how often the search volume data is updated there's a lot of varying opinions on this but I can't get into that because obviously new keywords are in there but how often are they refreshing data for existing keywords I have no idea I think that's run by the adwords team I figured it was a long shot I have no idea I'll keep looking I think we're way over time it's been good having you all here it's good having you all stick around until way way late so yeah it's our pleasure gentlemen you guys are crazy we just want Google and SEO and our wives will kill us yeah oh that would be a shame alright so let's take a break here thank you all for joining and I'll set up the next hangouts probably I don't know maybe tomorrow maybe next week so that if there are any questions that need to be answered as well you can add those there and as always you're welcome to post in the forum or tweet us on Twitter and try to get answers there as well alright thanks again and see you next time thank you all bye