 All right. Welcome, everyone, to today's Google SEO Office Hours Hangout. It's not Hangout anymore. I guess the Hangout part, we have to drop at some point. But Google SEO Office Hours, I guess. I'm John Mueller. I'm a search advocate here at Google in Switzerland. And part of what we do are these Hangout Office Hours, I guess. So awkward with the names. They keep changing. But anyway, a bunch of questions were submitted already on YouTube, which is fantastic. And it's good to see a bunch of you here as well. If any of you would like to get started with the first question, you're welcome to jump in. Hi, John. Hi. Can I start? Sure. First of all, thanks. Because the favicon problems, I don't know if you remember from a couple of months ago, I've been fixed finally. So thanks. Now it appears. OK. Thank you very much. This is the question. If I look at the statistics of the Google search console in the past six months, I see that the number of impressions has nearly doubled. So in six months, Google shows twice the time on the search console to people in internet. So I can say, OK, this is very good, because maybe Google is liking the site. But at the same time, the average position went down significantly down. And so the overall number of clicks is the same. So how can I interpret this? I mean, what is to be changed or to improve the situation? Because these are two different signals. One is positive and the other is negative. Yeah. So with any kind of changes that you see in the performance report, my recommendation is always to try to drill down and find some examples of this change in a way that is very visible. So for example, you could look in to see if there are specific pages that have changed in the ranking or in the number of impressions, or if there are specific queries where you see this change. So one thing that might be happening is that maybe just more people are searching. It might also be happening that you're showing for slightly different queries than before, where there may be more impressions for those queries, but you're not quite in that competitive range yet. So all of these things can help you to figure out a little bit better where your site is positioned at the moment and what you might be able to do to improve there. So for example, if you're being shown for more competitive queries, more things that more people are searching for, then that could be a sign that Google has recognized that you're relevant for those kind of queries. And now it's up to you to really improve the quality of your website overall so that Google says, well, this is actually a good result for this query. And that's something that ultimately you have to figure out with your business goals in mind. And think about is this the query that I think my site should be ranking for? Or maybe is Google showing it for the wrong query? And if so, how can I help Google understand my content a little bit better? So it's not just blindly focusing on the number of impressions and the number of clicks and the number of the rankings that you have there, but rather trying to figure out how can you position your business in a way that makes sense for your business, for your business goals, and brings it into the right place for users. OK, thank you very much. And it's sometimes really tricky with these kind of things because also because of things like internationalization. So that's something that we sometimes see with our own content. For example, I don't know, the Webmaster Central blog in Canada sometimes ranks for the word Google. And you can imagine then the number of impressions from Canada is suddenly really high, but the number of clicks is really low because normal people, when they search for Google, they don't want to go to the Webmaster blog. So it's something where sometimes you can do something to control it. And sometimes it's just, I don't know, the algorithms being a little bit weird. OK, thank you. Sure, thank you, John. Thank you very much. All right, any other questions before we jump into the submitted ones? Sergio, I think you're trying to say something, but I can't hear anything. He did post a question in the chat, though. Oh, in the chat. OK. All right, let me see. Ooh, OK, long question. One of my client sites has been around for more than five years and held good rankings for keywords related to hypnosis, self-help, but no real waves, and subliminal messages. As a form of therapy, we've noticed over time, especially after the medical and birth updates that these rankings dropped to the second and third page of results. Oh, man, I see you're struggling with the microphone, but I'll just read for the moment, and it's like, if you get online, no problem. The content has always been great, in comparison to the pages that now rank on page 1. I can see that the first five positions for many keywords have been awarded to medical and psychologically authoritative sites, which is great. But positions 5 to 10, there are many low-quality sites in comparison. And then there's some examples there. I think you also submitted this as a question. Let me see if I can find your question. I do see that the site is thriving in other areas and other pages that we consider more supplementary content, while the main content seems to be plummeting. My suspicion is that Google thinks the site and its authors' areas of expertise are not relevant enough to be trusted to award better rankings. Could that be the case? If so, how could we recover from it? So it's definitely possible that this is the kind of thing that is happening here, especially when it comes to, I think, medical kind of queries. It is something where our algorithms try to be a little bit more, I don't know, more critical with regards to the site and the information that we find there. So that's something where I would try to find a way, oh, I can hear something from you. Yeah, I think I've fixed it. OK. OK, cool. Cool. Yeah, I just would like to add, I didn't want to really interrupt you, but I wanted to add. I mean, I've been looking at all the queries, and it seems like the pattern indicates that the specialized sites like the psychology and all this traditional, I would say, medical institutions would dominate those rankings. Top five positions, but then let's say that the alternative ways of treating those mental illnesses are falling a bit after that, between five and 10 sometimes. So my question really is, is it worth it to keep and try going after those queries, or is it something that maybe it's never going to happen, and maybe we should just do something else? I would never say that it's never going to change, because the web is very dynamic, and these things do change. But I think just leaving it as it is and hoping that it magically changes on its own, I think that's generally a bad strategy. So on the one hand, I think shifting the focus a little bit and trying to find queries where it makes more sense to show these sites. That might be an option. The other idea might be really to find a way to really highlight why your site is really the one that should be shown for some of these queries. And I realize that's kind of tricky in the sense that my usual advice is to say, well, you should make sure that your site should be ranked number one by far. And it sounds like in your position, you're saying, well, the number one rankings are kind of OK, and we agree with those. We would just like to be kind of on the first page as well. And I think that's kind of a tricky balance. So one thing I would recommend though is to look at what some of the other people in the SEO area around EAT have been writing about. There are lots of really good, I think, case studies or examples of sites that worked hard to improve their expertise, authority, and trust in terms of how they present their content and in terms of how they create the content, how they have things like author profiles, all of this. That might be something where you could look into. So I don't know all of the names offhand. And I feel if I mentioned some of them, then the other ones that I forget will be upset. But there are some really good SEOs that have been working in this area, have been working with medical sites on this topic. And I would try to search out their content and look at some of those examples. I can't guarantee that if you improve your site in that regard, that suddenly our algorithms will say, oh, this is really fantastic site. But it sounds like you're touching on the medical area there, and it definitely makes sense, at least for users, to make sure that you have all of those signals as well, that you're really saying, well, this is not something that, I don't know, some kid dreamt up in his basement, but rather we spent a long time researching this. And here is the research. And here are the people behind it and presenting it in a way that is trustworthy for users as well. And then sometimes we can pick that up for searches as well. Thank you. Thank you, John. Sure. OK. Hey, John. Hi. Go for it, sure. OK, my question is regarding mobile-fest indexing and drinking. So for example, there's some website that is on mobile-fest indexing. So from Google's point of view, I think Google will use a mobile version's content and ranking signals for rankings on desktop and while both. Is that right? Yes. Yes. OK. Now, the second question is if the website is on mobile-fest indexing, but the majority of traffic is coming from desktop, because their product is a desktop-only product. It's a software application that runs on desktop only. So the majority of traffic is coming from desktop. So in that case, will Google consider the desktop content for rankings, or it will still consider their mobile version for ranking? It would only use the mobile version for ranking. So when we shift to mobile-fest indexing or if we've already shifted that site to mobile-fest indexing, then we will only use the mobile version for indexing and for ranking. So in a case like that, one thing you can do is if it's really primarily a desktop site, like for example, if it's a software company that only sells desktop software, then you can just make a desktop site and don't have kind of a simplified mobile version for that. But it doesn't matter where the traffic is coming from or what type of users are searching for the site. If it's in mobile-fest indexing, we only use the mobile version for indexing. Sure. Thank you. That was a confusion, basically, because most of the people on desktop say you were confused if Google is considering desktop signals or not. Thank you. Sure. John, a quick follow-up on that. I know we talked last time about a classified site. I'm working with them, the fact that they're still on desktop and one of the reasons it wasn't migrated yet was because a difference is in content for pages with very high impressions. So I checked a bit into that, and I noticed pages with the highest impressions are usually category pages for classified ads. And those change very fast. I mean, it's one of the biggest classified ads in Romania. So the pages themselves change a lot in a very short amount of time. So I was wondering, when Google checks whether the desktop version is similar to the mobile versions in terms of content, is there any delay? Does it check it at this exact same time, or is there some time passing between the individual tests? I mean, there's always a delay, because you can't really do everything at exactly the same time. But in general, we should be able to deal with that. So on news sites, we kind of have the same problem, that there are always new news articles. And my understanding is we should be able to deal with that. What I have seen on not on classified sites in particular, but sometimes on e-commerce sites, is when you look at category pages, sometimes the mobile category page has a lower number of articles on it. So you might have, on desktop, you have kind of like this matrix of content, where you have, I don't know, maybe 50 items on the desktop category page. And on mobile, you just have 10, because you have this one row, or one column of content. So that's something where we could kind of pick that up on. And I think we have that in the last blog post on mobile-first indexing as well. We kind of touched on that subject. I think, in general, in a case like this, I wouldn't necessarily worry about it if you're sure that this is how we want to present it. And if we index the mobile version, it's like we have all of the content. It's just a different number of items. And it's all the same. Then in a case like that, I would just leave it be. It's not that you're going to rank better or higher with mobile-first indexing. At some point, we'll just shift the site over and say, oh, we've waited long enough. And if you're happy with the mobile content and the desktop content, then it's like, well, nothing will change. So that should be fine. Right. Makes sense. In this particular case, the number of individual listings is the same. But again, due to the freshness of the new ads, the actual content changes like in minutes. So if we're talking about minutes, in terms of delay, it's already different content, especially for those very high-impression categories. Yeah, I could imagine that. Maybe that throws us off. Yeah. As you mentioned, it doesn't really matter anyway, because whether it's on desktop or in shift on mobile is just going to be the same content. I mean, new content anyway. Yeah. Yeah. Yeah, that's an interesting one, though. I mean, I think with mobile-first indexing, it's always kind of a case of if your site is not shifted over, do you have a problem or not? And as long as you can determine that actually our site is OK and it's just Google that is a bit confused, then I would just leave it be and let it kind of take its time. Right. I mean, if it was an internal linking problem, that would actually be more of a concern, because then ranking signals might not be transferred correctly and so forth. But in this case, it's not the case. Yeah. Cool. Thanks. Hey, John. Hi. Good question. So I've been working with a news media publisher. And they did a recently in early August. They did a redesign. And after two to three weeks, their traffic dropped by almost 40%. So they are quite big in terms of traffic and in terms of keyword rankings and authority. And the only thing that changed during this, let's call it light redesign, was some fonts and a lot in terms of layer changes. So one thing, and we've tried different things to kind of investigate why this job happened, worked on pay speed issues that were there before, but maybe could have been the reason. We knew that some internal linking might have changed, but that doesn't totally explain why the drop happened side-wide, not in just some specific articles and other things. But one thing that we are suspecting that might be the case is a page layout penalty. So here's what happened. When the article starts, there is a video there. And sometimes in that video, a net place. Not all the times, but in some cases. When we are auditing the pages through Lighthouse, in some cases, it shows that there is no ads. In some cases, there is. So we are suspecting that that is causing the Google to see that there are a lot of, like where is this ad on mobile? That's a bolder fault that is considerably taking space. And that was before the redesign at the same place. But our suspect is that because of the Light redesign, Google totally re-evaluated the whole site. And now is thinking that this shouldn't be there and it's not a bad user experience for the users. It's a bad user experience. And for that reason, we are suspecting that this might be the reason for the drop because we're struggling to find one reason to explain the side-wide traffic drop of a both-forwarded site. Yeah. I don't think a redesign would trigger a re-evaluation in terms of understanding the page layout. So that seems like something that would be unlikely. It's hard to say with what you're saying about the redesign. It sounds like that wouldn't be playing a role with kind of how we rank things there. But sometimes there are subtle effects that aren't visible at first. So for example, if you were to shift to a JavaScript framework rather than a static HTML site, then it could look very similar, but from a technical point of view, it would be very different. Which we didn't in our case. Yeah, so one item less to worry about. I mean, not that JavaScript is bad or problematic. It's just like these kind of changes are pretty big. I don't know. Offhand, I think if you have videos on the pages and they're towards the top of the page, then that should be something that we would be able to recognize as kind of like a video landing page rather than to assume that the video is an ad. So just to clarify it, because maybe I didn't give enough insights. So basically it's a news site and the pages are a news template. So basically it is the standard style that you see in all like news media sites and nothing, let's say, very different from that. In terms of the video is like you've seen those videos when you go to some news site that they are on top, not necessarily super relevant to that article, but they are on the same topic. Like let's say there is an article about Google launch and the new Pixel phone and then we are just talking about the Pixel line, not the specific new phone that was launched. And sometimes on that video, an ad place. Some other thing that we've been working on is the CLS. So one thing that we've noticed compared to other publishers is that we had these shifts happening in our site because for some images and some ads, the space wasn't reserved and it would kind of push the pages down a bit when the ad would load or like sometimes the ad would be a medium rectangular and sometimes another format would show and this wasn't always predictable. We were addressing that, but again, we are not very sure if that could have been the reason for a 40% drop and that issue has been before the site redesign. So again, it's very tricky to pinpoint what the exact reason could have been. I don't think the CLS metric itself would be the cause because we don't use Core Web Vitals as a ranking factor yet. So just kind of the CLS side itself wouldn't be an issue. It might be that there's some indirect issue that users are confused and they start hating your website because of this shift and then over time you see an effect, but that wouldn't be an immediate effect. Yes, and it's a very light redesign meaning that some users, it's just small improvements on making some things clearer and cleaner in general. So we double checked a lot of things. It's not tracking issue because you're seeing the same things on Search Console and HREFS is reporting. And one interesting thing is that our top three rankings, which we had a lot drop, they are staying on the first page, but we are just not ranking on the top three rankings. And this has had quite some effect on the traffic that we're getting. Yeah, so I don't know your website, so it's really hard to say, but my general feeling here is that this is less of a technical issue from the redesign and more a matter of our algorithms just re-evaluating the overall site quality. So that's something where it's, especially when you're looking at the top queries for a page or for a site, then you might see that those are still ranking really well because we think your pages are really relevant for those queries, but if the general shift is a little bit downwards and some was like, well, overall we think your site is a little bit less, I don't know, less important from a quality point of view. So it's like overall shifted a little bit down, but the top ones are still very relevant, so we still show those. So just from hearing these things, it sounds like it's something in that regard. And it's really hard for me to say if I'd be able to figure out in more specifically if I had the site and we're able to use my tools, my feeling is that this is just an overall shift in how we evaluate the quality of the site. Yeah. May I share the site with you in private on Twitter? Just so you can take a look. Yeah. Or I mean, you can also leave it here in the chat. I pick up the chat afterwards and look through the URLs that are submitted there. So if you have like a comment and just drop it in the chat and I can pick it up here too. Or that way. I'm happy to take a look, but if it's just like a general shift in the way that we understand the quality of your site, then I don't think I'd even have anything specific to say because I wouldn't be able to say, oh, on this page you need to improve this. It's just, well, overall our algorithms are a little bit more critical with regards to news and maybe like overall things could be improved there. Yeah. It's the only thing that kind of had made this very tricky for us is that it was immediately after the light redesign. So it could be a coincidence, but we couldn't base. Yeah. Yep. Yeah. I'll share that with you. Thanks a lot. Yeah. I think it's always tricky with these things that happen at the same time, but we make so many changes. It's really hard to kind of like avoid some coincidences sometimes. Yep. Yep. Thank you. Okay. Let me run through some of the submitted questions so that we don't lose track of those and we'll definitely have more time towards the end to discuss more other topics. Follow up to the last Hangouts question on Naked URL links without anchor text. So Naked URL is essentially just when someone uses the URL as the anchor text instead of a text or word for the anchor text. Doesn't mean that without the anchor text, there's no value in that link. Like this great website is linking to another site with just the URL. Would you pass some of the greatness to the linked site even though there's no anchor text? Yes, there is absolutely value in a link without anchor text, but of course we just don't have as much context. So you could imagine a situation where even internally within your website you just use the link itself without an anchor text and of course we would be able to crawl your website. Of course we would be able to figure out which of these pages are important, but we'd lose a little bit of value or a little bit of context from that link. So it's a bit like making a page with a text file instead of an HTML file where you can format things and specify headings and titles and things like that. So you could do that, but we lose a bit of information. I updated my content at the end of last week and Google only picked up the updated content yesterday. I've never seen such a delay. Usually it takes at maximum a few hours to update content. Why does it currently take so long? Will it be back to normal soon? So we don't guarantee any specific time with regards to crawling and indexing and depending on the site and where things change within the site, sometimes we can pick up those changes within a couple of minutes, sometimes hours or days and sometimes it takes months to pick up those changes. So that's something where from our point of view it's not that there is a clear minimum response time for indexing, but rather depending on the pages that you change, sometimes it takes longer, sometimes it's very quick. For example, if we recognize this is a page that is very critical to our site where things are changing very frequently, then probably we will pick those changes up fairly quickly. On the other hand, if we realize this is actually a page that has been the same for the last 10 years and you make a change there, then probably it will take a couple of months for us to realize that actually you changed this page. So from that point of view, there's a wide range there. What you can do to make it so that we pick up these changes a little bit faster is to let us know about the changes. So you can do that with a sitemap file. That's very common. Most CMS systems, if you're using something like WordPress or Blogger or whatever, they will automatically generate sitemap files or feeds for you. And then it's just a matter of you kind of submitting that feed to Search Console. And then once that feed is submitted, then all of the updates automatically go to Google. So that's kind of the fastest way to get things automatically get picked up. If there's something really critical on your site that is changing and you really want to make sure Google picks it up as quickly as possible, then you can also use the Inspect URL feature in Search Console and submit the change there. I would really only use this for kind of exceptional purposes. So if you're just changing text naturally, if you're adding a few new articles, then there's no need to use any of the submit URL features. But if there's something really important and urgent that you need to change, then that might be an option. If we submitted an incomplete sitemap, does it affect the overall search performance of the website? All web development work is done by another agency, and we already sent recommendations to replace the old sitemap with a new one. So until this is done, will this affect our search campaign at the moment? Probably not. So the reason I'm saying probably not is if your sitemap file is missing the pages that you care about, kind of the new and updated pages, then, of course, we need a little bit longer to actually find all of those new and updated pages. So that's kind of the potential downside. On the other hand, a sitemap file only helps us to crawl a little bit better. So it's not the case that our crawling would only focus on the sitemap file, and we would crawl kind of the, or we would kind of suspend the normal crawling of the website and only focus on the sitemap. That's not the case. It's really the case that we crawl your website normally anyway, and then the sitemap file helps us to crawl a little bit better. So if you make changes on your website and they're picked up through the normal crawl, then that's perfectly fine. The sitemap file doesn't change anything with regards to ranking. So if you have an older sitemap file, then that doesn't mean that your pages will rank in any way differently. It's really only about this crawling part, where we might want to pick up changes a little bit faster if you make specific changes on your site. And if you have an old sitemap file, then that doesn't get picked up. One thing I'm kind of worried about here with this question is that you're telling the agency to use a newer sitemap file, where really what you should be doing with a website like this, or in general with a website, is to have your sitemap file generated automatically. So instead of kind of manually replacing the sitemap file, you should make sure that you have a system in place that automatically generates it all the time. Just so that any time you make changes within your website, then those changes get picked up automatically. And there's no kind of manual step involved in getting that updated. So that's kind of my recommendation. One thing you can do, depending on your setup, especially if you're at a company where you have different departments working on different parts of your website and you have kind of a marketing department that does the website and then maybe a tech part that does something different, you can host your sitemap file somewhere else. So in your robots.txt file, you can specify a location of your sitemap file. And that can be somewhere else. It could be even on a different domain. So if you have access to the back end of your server and know when things are changed, then you could put your sitemap file on a separate domain, just for sitemaps, for example, and use that as a way of submitting always live sitemap files, even if the content itself is something that takes longer to be updated. So that might be an option there. How to write a canonical tag and use a sitemap for a multilingual website? Wow, so many sitemap questions. So cool. So the canonical tag, it's not really a tag. It's a link element that you place into the head of the page, the web page itself. So it's not something that you would put into the sitemap file, but rather it needs to be in the HTML page itself. And that's something that needs to be in a specific format, in a specific part of the file so that we can process that and trust it. So that's like depending on how you create your web pages, you might need to look into that in particular. For a multilingual website, you can use the hreflang annotations between different language versions. And these different language versions you can put either into the head of your page, like with the rel canonical or into a sitemap file. And that's, I guess, a little bit different from the rel canonical in that the hreflang annotations can be in either one. With regards to the mix of canonical and hreflang, the important part is that all the individual language versions of your pages should be canonical to themselves. So the canonical tells us which of your pages you prefer to have indexed. And if you say, for example, the English version is my canonical for the French version, then we may say, well, then we don't need to process the French version. We will only index the English version. And usually that's not what you want. Usually you'd like to have all of the different language versions indexed individually. So lots of different answers there. I don't know which of these might help you there with your question. It's a little bit vague. But hopefully that helps refine things a little bit. I've been trying to feature my website on Google News. I successfully submitted my website in Google News three or four months back. Is it mandatory to add a news sitemap in order to feature it in that section? Any other recommendations would also be helpful. So I don't know too much about Google News with regards to how to get things submitted there and into the Google News side of things. So it can really help you there. I have heard from other people that things are a little bit backed up with regards to getting new websites into Google News. So maybe that's something where you'll need to be a bit more patient. With regards to the news sitemap, I don't know for sure. So I do know it's something that we strongly recommend to use a news sitemap, because especially on news sites, it's really critical that we pick up the news content as quickly as possible. So that's something kind of to keep in mind. But we do have a lot of this documented in the news publisher help centers. So I would strongly recommend going there. And I believe there's also a news publisher help forum where you can ask more specific questions on these kind of things. But also, like I mentioned in the beginning there, I have heard from people externally that kind of getting new sites into Google News is a lot harder now, or it takes a lot longer time. So kind of, yeah, I don't know, just to set expectations. How breadcrumbs help SEO? Does a breadcrumb schema show any rich result in search? Should we include the home page as a first position in the breadcrumb schema? Should we add the current page as the last element of the breadcrumb schema? So I think you almost answered your first question. How breadcrumbs help SEO? In general, breadcrumbs, especially when you're talking about the breadcrumb structured data, they don't change anything for SEO. It's not that your pages rank any differently. But rather that we would show them differently in the search results, why don't we understand the breadcrumb markup? So we could show kind of that breadcrumb trail in the search results as a rich result. It wouldn't change anything from ranking. It just makes it a little bit easier for people to kind of jump in at the place where they would like to be. With regards to the home page and the current page in the breadcrumb schema, I'm not 100% sure. But my understanding is that we made it a little bit clear in the documentation that this is optional. So you can include your home page and the final page. You don't necessarily have to. We definitely understand where your home page is. We definitely understand where your final page is. So we can interpolate a little bit from there. The one place where breadcrumbs could have an effect on SEO is less around structured data. But if you use them to actually create links on the page. And in a case like that, the effect is essentially that you're cross-linking these different pages. So if you have a website that has multiple category levels, for example, then you're linking from one product to maybe the subcategory and to the higher level category in the breadcrumb in the HTML on the page. And oftentimes, that's a good thing. For users, they can navigate and find the category that they want. For crawling, it definitely helps us as well. So that's something that kind of makes sense. But if you already have this HTML and you're wondering should I add the structured data or not, then that's really just a display change. And with regards to the display, we also try to figure out which breadcrumbs to show automatically. So it might be, depending on your website, that we're already figuring out which breadcrumbs to show. You can see this when you search for your own content. I'm running a local business website. There are a few competitors. My question is for a long time, my web and some of my competitor websites are working very well against some competitors who made a website on WordPress. But now, suddenly, for the past two to three months, all of the webs made on WordPress are on top. And the most frustrating thing is that their entire content is copy and paste. Their page speed, PA and DA scores are worse. Just a quick side note, PA and DA are not metrics that Google uses. These are from third party tools, but they can be useful sometimes to compare things. There's nothing which I'm able to find that they have better than my website, except their backlinks. They have 80% more backlinks than mine. And all those three to four webs are made on WordPress and ranking well, and they're only one to two years old. So the question, are backlinks still this much important for web ranking? So I think a few things worth mentioning here. On the one hand, the age factor is not necessarily something where we'd say, well, older sites deserve to rank higher, or newer sites deserve to rank higher. Sometimes new content is very relevant. We will show that visibly. Sometimes older content is. Sometimes we show a newer domain. Sometimes we show an older domain. So it's definitely not the case that you need an old domain name to rank well. The other question that I kind of read between the lines is, is WordPress better, or is a custom CMS better? Is another CMS better? And from our point of view, we don't care which CMS you use. So it doesn't matter to us if a site is using WordPress. It doesn't matter to us if a site is using Wix, or Blogger, or any of these other systems. Essentially, we look at the HTML pages that are generated, and all of these systems have worked really hard to make reasonable HTML pages. I think that's one of the really cool things about the web in, I don't know, the last 10 or so years, in that if you're using any of the common setups to create web pages, then chances are, kind of by default, things will work reasonably well with regards to search, and you don't need to do anything custom to make them work even better. So from that point of view, if WordPress works for you, and even if you're not creating a blog, but rather maybe a company website, or in shop even, then feel free to keep using that. Or if some other system works for you, then that's also totally fine. So from that point of view, I think whether or not a site is on WordPress or not does not play a role. With regards to links, we do use links as a factor in some of our algorithms, but we use a lot of other things. And links are probably not the one that I would say is the most critical item here. It's really hard to say much about this specific case, because there is not a lot of detail here. So this is something where I'd be tempted to say it would be useful to go maybe to the Webmaster Help Forum so that others can take a look to see if there's something specific here. But the shift from position three to position one or the other way around is something that can happen for a lot of different reasons. And just because a website has some things that are worse doesn't mean that it will automatically rank lower than other ones. So a really common case that comes up in the forums and when talking with people is, for example, maybe a site has hidden content on it somewhere. And people will come to us and say, oh, this website is ranking above mine, but it has hidden content. And your Webmaster Guidelines say hidden content is bad. Therefore, you should remove that website from search. And from our point of view, we might recognize other things that are good here. And we might even recognize that there's hidden content there. But if we can recognize there's hidden content there, then we can also ignore it. So just because there are some aspects of a site that are worse than yours doesn't mean that it will always be ranking lower. Maybe there are other things that are actually pre-reasonable. Or what might also be happening is that these sites in the search results are actually very similar or very similar with regards to how they fulfill the user's need. And then it's something where if we were to take this to our search quality team and say, well, these sites are all very similar, but the number three is the one that really wants to be number one, then they'll tell us, well, if they're so similar, then there's no reason for us to change anything with regards to ranking here. And the way that you kind of work around this is by making sure that your website is by far the most relevant one for these queries. So make it something so that if we were to take this case to the search ranking team, then we could go and say, well, we have a bug in our systems because what we're showing in the top results here is clearly a bad result. And what we're showing at number three or number four is clearly the one that we should be showing here. And in a case like that, then the ranking team will be able to look at that and say, yeah, we need to make some changes to better understand the unique things that are happening here. But if all of these are kind of equivalent and they're kind of doing the same thing, then the ranking team will say, well, we could spend a couple of months working on tweaking the ranking for this site. And maybe 50 people will see this improvement, which is OK. Or we could spend a couple of months improving something for a lot of other sites or a big mass of users. And probably they'll focus on the bigger issues. So I don't know. I guess, in short, my recommendation would be to maybe go to the Webmaster Help forums and get some input from other people, kind of more, maybe more objective input, and really to think about what you could be doing on your site rather than to focus on what your competitors are doing. Can fragment identifiers be used to optimize for rich features snippets? I don't think so. So fragment identifiers are these URLs with the hash or the number sign in them. And they're generally generated as links on a page, and they jump to a specific part on that page. So that's kind of the traditional thing with regards to fragment identifiers. And essentially, you're on the same page. You're just going to different parts of that page. And when it comes to indexing, we essentially drop all of those fragments. We ignore them completely. We see them as a link to the same page. And then if it's a link to the same page, there's nothing we need to do there. And that wouldn't change anything which with regards to rich features or the featured snippet aspect there. Sometimes we do pick up the fragments with regards to making a cleaner snippet in the search results. But that's really something where we understand this is a specific part of the page. And we can link to that part of the page directly. And we will just link to it like that. So you often see this for Wikipedia pages which use these fragment identifiers quite regularly. The other place where these fragments are used sometimes is on JavaScript sites. And for us, this is really problematic because, like I mentioned, we drop the fragment for indexing. And if the content on the JavaScript site is only loaded when the fragment is included, then it's very likely that we will not be able to index that content. There's a very, very small number of sites where we do use the fragment identifier for indexing. And that's essentially a very small number of cases where in the early days of JavaScript indexing, we thought this would be the only way to pick that up. And in the meantime, we realized that we shouldn't be doing this because it just causes so much trouble. So for the most part, we drop those for indexing completely. While Google rejects AdSense applications, just say specifically what went wrong with that application, that's the best update you can do for sake of your customers. I don't know how AdSense is handled here. And I imagine they get a lot of applications for things that are not relevant to be shown. But I have no idea. You need to maybe add that to the AdSense help form. I have a Hindi word in my domain, which is a language. Will it affect my articles written in English? So we do use some signals from words in a URL, but it's very, very small. And especially if we can pick up the content on the page, then we can essentially ignore the words in the URL. So if you have, I don't know, a Hindi word in your domain name and your content is in English, then that feels perfectly fine from my point of view. It is very common to also have international websites where you have maybe a domain name in French or in German or in English. And the content itself is in other languages as well. And that's totally fine. So that's not something I would worry about. Is it recommended to use keywords as it is the content to get the best results on Google? Does the crawler use fundamentals of AI to make combination of keywords that are closely related to the content and then rank them? We want to know how a crawler picks up keywords to be ranked on Google. So I don't know. There's lots of ways that we pick up keywords. But I think in general, we essentially look at the content of the page. Way in the early days, the keywords meta tag was a thing. But essentially, the content on the page is really what users see. So that's where we try to understand which words we would show this query for. And we do try to figure out combinations that are more like synonyms or that are equivalent with regards to keywords on a page. Sometimes we figure out which things are acronyms or a singular and plural. And we try to understand, is this page relevant for both of these versions? Or maybe it's just relevant for one of those versions. But this is something that is quite complex. There was a video that we put out, I think, the beginning of this year from the webmaster conference we did in Mountain View from Paul Har, which was, I think, a really interesting session. And he goes into a lot of these aspects with regards to keywords and when we understand things are similar or equivalent, when we understand that things are different. So I think some of the examples that he had where you have a page that is about New York and a user is searching for York, should that page about New York also be ranking. And of course, we should be understanding that New York are two words, but they belong together. And York is a different location that should be ranking individually. And all of these things are really kind of unique problems and interesting to look at. So I would definitely take a look at that video. I think the only short answer, if you really want something short here, is you don't need to put all of the variations of all of your keywords on your pages. If we understand your page is about a specific topic and has some of those keywords on there, then we can understand the rest itself. So you don't need to put all synonyms on your pages. You don't need to kind of do this SEO thing where you include all of the typo versions of your keywords on the same page. We can figure that out. OK, wow. We're already at time. Maybe I'll just open things up for more questions from some of you, and then I'll pause the recording so that it doesn't get too long. And then maybe we can continue a little bit kind of off the recording afterwards, too. Hey, John. Hi. I did post a question. So we have the, I understand Google Mobile Indexing is going through AMP sites first for ranking, but it's an e-commerce website. So not the entire website is AMP. I want to say about 50% of it. So Google is first trying to index AMP. At some point, it's hitting a regular desktop site link. But the desktop is responsive, hence mobile friendly. So it's now also grabbing all the mobile or the rest of the sites. From a regular page, I'm linking the AMP pages with the link rel AMP HTML and from the HTML doing AMP. So I'm kind of tagging them each other. But this ideally would be duplicate for mobile versus AMP, right, or no. So when it comes to the paired AMP setup, I think that's what you have there, like the normal HTML and then the AMP page and then the linking between the two. We would use the normal HTML page for indexing. So from that point of view, the AMP page is more supplemental for us. And we can show that when people are on mobile on appropriate devices. But for indexing, we would use the normal HTML version. OK, and is there a reason why AMP won't get traffic? Like, I see standard devices like iPhone or, I don't know, Android or something like that. I don't know why that might happen. So one thing that has to be the case is that this cross-linking has to be correct and the AMP page has to be a valid AMP page. You can use kind of the AMP tester for that. The other thing that sometimes happens is because of the way the AMP pages work, you have to do kind of the analytics there separately, especially when we show it as a page on the AMP cache. Then you can't just use the same Google Analytics set up there. You kind of have to mix those two together. And that's something that's sometimes confusing and that you look at the analytics side and it seems, oh, nobody's going to my AMP pages, but then you kind of need to add that separate AMP part to it as well. OK, for the most part, since AMP is a separate set up from development perspective, do you suggest user maintaining two versions or potentially three desktop mobile and also AMP or just kind of step away from AMP? I see a value for speed and all, but what are your original thoughts? I don't know. So I think, in general, from Google's point of view, AMP is a great way to make really fast pages. And there's some features in search that rely on AMP to work well, especially things where we need to embed kind of a page in an AMP viewer type of situation. Then we need to have an AMP page for that. If your content is not relevant for those search features, then it's more a matter of kind of like the speed side. Can you generate the same kind of speed with your normal HTML pages as you can with the AMP pages? And if so, maybe it makes sense to focus more on the regular HTML pages. But if AMP is the way that you can make your mobile site really fast, then I would definitely continue using that. OK. All right. Thanks. Sure. Adrian, a follow-up question on the issue. I've shared the domain on the chat. So you know what, for what side you're talking about. Just to clarify, I think you mentioned earlier on that you don't think that the site was re-evaluated because of this, let's call it, design refresh. Could something else that we've been using, and that has been used even before the redesign, we've been using something like tracking parameters on the URL, so to all the articles that we learn from the home page. So we know what traffic is being generated from the home page to which article. So we know which kind of sections of the home page are driving the most clicks internally. This is just to measure some things that we needed. But the pages themselves, they had the canonical implementation the right way. Could that have had some impact? Because home page is the most authoritative page of the site. And maybe those pages weren't getting as much value, but this has been implemented through redesign as well for a couple of months. That could, I don't know, depending on how you have that implemented, that could have an effect in that if from the home page you're linking to the articles with a unique URL and then from the articles, you have the canonical back to a clean URL, then is that about the setup? So it is the same URL. It just has URL parameters in the end of the whole URL. It's like UTM, but we modified them a bit. We removed that I think about two weeks ago, but it's not like we're seeing something happening yet. So one thing you can do to check that hypothesis whether those particular tag URLs are ranking in search. So looking at the performance report to see is Google focusing on the clean URLs or is Google focusing on the parameterized URLs? And if Google is focusing on the clean URLs, then we configure the canonical part out. And that's all fine. If Google has been focusing on the parameter URLs, then it seems like, well, we got confused with your site structure, and then over time, we will dilute the value because we're not sure which of these pages we should actually be associated with your home page, for example. Yeah, we check that. And it is the clean URLs that are ranking. So the canonical is working the proper way. But there are so many things that could have happened. So we're testing and trying and thinking about everything. Yeah, I can sympathize with that. If there is a big change that happened, then trying to figure out what exactly is responsible for that is something that I think anyone would do. I do think if you're seeing an overall drop like this from one day to the next, then it seems a lot more like a quality issue rather than a technical issue. If it were a technical issue, then you would usually see kind of a subtle decline over time. Whereas we reprocess things for indexing, then some things go a little bit faster. But it's really something that would take about, I don't know, a couple of weeks' time to be fully processed. And if you're seeing it from one day to the next, then it seems a lot like our algorithms are kind of classifying your site slightly differently. I see. So yeah, this could be the case because we saw a small decrease in August 30, then it accelerated dramatically on September 3. So by September 3, it started really dropping. It hasn't been a couple of weeks in the end. I mean, technical issues would be a little bit easier to figure out and clean up. So I understand focusing to make sure that all of the technical things are lined up first. But it really feels like, from a quality point of view, it might be worth getting some more input from people and seeing what could you be doing slightly differently. But I really don't know your website. So it always feels awkward to say, oh, your website's quality is bad. And it sounds like it's not bad. It's just, well, we thought it was a lot better in the past. So it's tricky. I hope that you will have a chance to just take a look at the site. And any input that you could give us would be very helpful to kind of work out what happened. Sure. I'll take a look. OK, let me pause the recording here. For those of you watching the recording, thanks for sticking around. And thank you for everyone who's submitted questions along the way. I'll still be here a little bit afterwards if any of you want to stay and chat a little bit longer. And otherwise, I wish you all a great weekend. Bye, everyone.