 OK. All right. Welcome, everyone, to today's Webmaster Central Office Hours Hangouts. My name is John Mueller. I'm a Webmaster Trends Analyst here at Google in Switzerland. And part of what we do are these Office Hours Hangouts, where we get to talk with webmasters and publishers, SEOs, people like the ones here in the Hangout, and try to answer questions that have come up over the time. As always, if any of you want to get started with a question, feel free to jump on in now. I can just jump over the quick question I was asked morning to just get your ideas on it. So you've got a client where we'll have a handful of phrases that rank and fluctuate wildly. So kind of in position two, and then out the top 10 pages, say. But then we'll have another group of phrases that rank consistently. Now, the difference between them would be, say, plural and non-plural. So say, for example, the plural is plural, as ranking quite wildly. And then the non-plural is ranking consistently as the same position at a higher position. Now, obviously, more or less, they are the same phrase. And this has been going on for months. And I just wanted to get your ideas on it, because I would have thought, OK, if it was Google testing certain search results, it wouldn't go on for months as it is. So really, my question is, why have we got a group of phrases that fluctuate wildly to a group that don't change at all? Yet they are more or less the same thing? I think on the one hand, the more or less the same thing is tricky, because when we look at the phrases, we might think they're pretty much the same thing, like you mentioned, like plurals and singular form. When our algorithms look at it, they just see a bunch of words, and they try to figure out what those words might mean. And for them, it's not obvious or not not a given that a plural is the same thing as a singular, that people mean the same thing, that they're kind of synonyms or not. So that aspect is something that you're bound to see some differences. And sometimes over time, our algorithms kind of sped up and say, oh, well, it's all the same. We just combine those and show the same search results, essentially, for the plural form or the singular form. And for some cases, we show different search results, because we've seen that maybe users react differently to this. Maybe it's something that our algorithms have picked up over time. And that's kind of a similar problem to what people have in non-English languages or languages where you have characters that are not like in the normal ASCII range, where sometimes users search with maybe an umlaut, or sometimes they search without an umlaut. And in those cases as well, when we look at those words, we say, well, the same thing. But our algorithms look at it and say, well, they're different characters. Maybe someone means something different. And sometimes over time, we learn that they're the same thing. Sometimes over time, we see that people actually mean something subtly different. So that kind of difference, I think, is completely normal. The other part where things are wildly fluctuating for part of a website or part of the queries that pull up the website, that can be normal as well. And usually what I've seen that's more a sign that our algorithms are kind of on the edge and not completely convinced that your site is exactly the right one for those specific queries. So that might be kind of a hint at if you work to improve the quality of your website, just a little bit, you'll kind of get right above that edge and things will be a little bit more stable. But it's also, I guess, a general sign that even if you make it just above that edge there, then things can change over time. And you end up kind of bouncing up and down again as algorithms readjust. So kind of not just taking a small step above that edge, but rather taking a pretty big step across that edge, really makes sure that things are a little bit stable. So when you say improving the quality of the site, what would you recommend? I mean, this site will be worked on by a team of SES or one SES in particular who understands the relevance of content and the relevance of tile tax and things like that, that really understands Google in a way that probably most searches don't. So when they are at a loss, what would you suggest? If we've got all the relevant content, all the relevant tile tax, and everything's really been done, in order to try and make that as consistent as possible in terms of the ranking, what would you suggest once you've, I suppose, when you've done the content aspect of it? Is there anything more that we can do? So what I usually recommend in cases like that is to think about doing a user study and to really getting a bunch of people who aren't associated with your website, to run through some common tasks that you think are relevant for your website, try them out on your website, on competitor's websites, and really get as honest feedback as possible from these users. And I found every time I watch user studies, it's really insightful because there are these small things where in the back of my mind, maybe this is a problem, and then you see users struggle with it because they don't understand what the text means or what they need to do as a next step on a specific page. And those are the kind of things that are sometimes really insightful and really help you to figure out, okay, this is something that we need to solve in a slightly different way. Technically we've done it correct, but it's something, maybe we need to look at it in a different approach. Okay, that gives us something else to go up, I think, with the user study. Okay, well, we'll give that a go then and see if there's more that we can do. Thank you for your answer, John, appreciate it. All right, okay, let me run through some of the submitted questions. Some of them I thought were pretty insightful, so let's see what we come up with. The first one here is actually pretty much the same thing. What you mentioned, I see some of my best pages fluctuate in search periodically. One month they rank, the next month they rank a lot worse. What could be the issue? Again, this is something where our algorithms are probably pretty much on the edge and they're not completely convinced what they should be doing with your website. So the better you can convince the algorithms, the more likely things will stabilize, hopefully on a bit of a higher, higher level. We have a large number of pages related to US cities on a .com site that ranked first for a given query on google.ca, google.co.uk and other international domains, yet don't appear at all for the same query on google.com. The site ranks at google.com for other queries. We also have Canadian city pages. The majority of our links are from Canadian sites. What are some possible reasons for this parity in query results between google.com and other domains? Our theories are, one, a penalty specific to this query. Two, Google considering us a Canadian site. What's the most likely cause? The next steps to resolve that. So first of all, I can't imagine that there's any kind of a manual action specific to a query here in a case like this. If there were any manual action on your website, you would see that in the search results. So any kind of web spam manual action, not in search results, in Search Console, of course. Any web spam manual action you would see in Search Console and you'd be able to respond to that. And say, I can fix this problem, I'll fix this problem and have it reviewed again. The biggest thing that I sometimes see in cases like this is that a site will kind of look at this in the sense that they rank on one version of Google or one country domain. Therefore they should rank the same on all others or on lyca.com for the US. And in general, the competition is very different across these different country versions. So that's something where it would be completely normal for us to have very different rankings in these different country versions of Google. So a site might be perfectly fine or really good in rankings in Google Canada because maybe there are fewer other sites that are competing for these queries on Google Canada. But because the competition is so much stronger in the US, maybe it will be ranking very different. So from that point of view, it's kind of normal that you would see this kind of difference. It's not something that I would kind of call out as a bug on our side or a bug on your side, something that you're doing explicitly wrong, that can be kind of the normal situation. With regards to geo-targeting, that's something that you can do if you explicitly want to primarily target one specific country. So for a case like that, you could use geo-targeting. If on the other hand, your website is really global and the content is meant to rank in Canada and the UK and in other international versions as well, then probably geo-targeting is not really the thing that you'd need to focus on. One thing you might consider is, does it make sense to split your site up into individual country versions? So maybe you have a specific set of content that makes a lot of sense for Canadian users, but not so much for US users. And similarly, you have content that makes sense for US users, but not so much for Canadian users. Then if you can split that into sub-domains or sub-directories on a .com, for example, then you can use a geo-targeting setting in Search Console to tell us that this is for Canada, this is for the US. And with that geo-targeting setting, we can realize that this content is specifically relevant for the US and we can rank that a little bit higher for queries within the US. So that might be an option to look at. It kind of depends on your website and what you're trying to do. If you're really aiming to be something global and all of your content is for everyone, then maybe geo-targeting isn't the right approach there. I'd like to ask about pagination and the canonical tag. Should you point to canonical tag? When paginating products on a website, the webshop site to the first page or to the page that you're on? This is a really common question that we get. I think there's a blog post on pagination out there. We also have information in our help center about this. The short version is if you set the canonical to a specific page, you can assume that the non-canonical versions of those pages are not being used for indexing. So in a case like this where you have paginated content, if the other pages are things that you really know aren't necessary to be indexed and they don't link to additional content on your website that is meant to be indexed, then maybe that's fine to say, well, I'll just canonicalize everything to the first page of my paginated set. And that first page of the paginated set will be the one that I want to have shown in search. On the other hand, if your content on page two and page three is relevant, you want that to be seen in search or it links to other parts of your website that don't have a lot of links otherwise, then you need to make sure that those individual paginated pages are also crawlable and indexable on their own so that we can follow those links and index that content. Two questions. I have two web pages like example.it and example.it slash index.php. Are they considered duplicate content? Possibly. So if they show the same content, they're duplicate content. In practice, what happens here is if we see exactly the same content, we'll be able to recognize that fairly quickly. We'll just fold those pages into one and index those as one version in the search results. Probably in a case like this, as a version that looks a little bit cleaner, like, so in this case, just the root URL instead of slash index.php. So certainly duplicate content, you could help us with the rel canonical if you have a preference there. In general, this is not something that causes any problems. When I use an auto-referential canonical for these two pages, do I have to write in the href attribute example.it slash index.php or just example.it? This is really up to you. Kind of like I mentioned, you can help us in the canonical decision and you get to provide us this input. For canonicalization, we use several factors. So not just the rel canonical, but also things like internal links, sitemaps URLs, hreflang links, all of that kind of comes together for us and based on these factors, we make a decision and say, well, this one looks like the best canonical for this page. So we'll go for this one. So the clear you can give us the canonical decision that you want to have made, the more likely we'll be able to use that. We're having issues with the messy site migration, thousands of old non-existent pages are lingering in the index seven months later. What should we do? And I think it goes on into, like we set up redirects and they didn't take effect that quickly. So we use change of address and then we went back and we made those pages live again and put a rel canonical on them and now we changed it to noindex on those pages and submitted them again. What could we do there? So I guess first off, the situation where you redirect from an old site to a new site and you do a site query for the old site and you still see content, that's completely normal. That's not a sign that anything is broken from our point of view. And if you explicitly look for old content, we'll try to show it to you even if we have an index under the new URL already. So that's kind of confusing where essentially our algorithms are just trying to help the user to find the content that they're looking for. So in a case like this, I would have just set up those redirects and let that work its way out. The change that you made here where you made those old pages live again, I think that's a part that probably threw us off completely because we thought, oh, there's a site move but actually some content is still there. So it's not really a site move. It's kind of like you're splitting a website into two separate domains. And that makes it a lot harder for us to actually process. So what I would recommend doing there is going back to the original state where you redirect all pages to your new site and let us crawl all of those pages. Don't block them with robots text. Don't put no index on those pages. Let us crawl those pages, let us see the redirects and then let that kind of settle down and search on its own. So I kind of avoid the situation where you have separate content still live on the old domain. Then an interesting question about spaces and case in URLs. I think this is something that comes up, I don't know, pretty regularly and it's always something where from a technical point of view, it's less of an issue for us but there's sometimes just practical issues that make this a lot harder than it needs to be. So the question I believe goes on with regards to spaces in a URL where you encode them with like percent 20 or sometimes you use a plus symbol. From our point of view, I believe that the plus symbol and the percent 20 and just the normal space in the URL are essentially equivalent in terms of they all map to a space. So that's something from our point of view, we treat them all the same, you don't have to worry about that. The thing I kind of watch out for in a case like this is whenever you have spaces or special characters in a URL, then that makes it a little bit harder for people to actually link to your content. So if someone is copying and pasting a URL from the browser into, I don't know, Twitter or Facebook or somewhere else, then those other services have to be able to parse that URL and figure out, oh, the space is not actually a sign that the URL is finished but rather it belongs to the URL, therefore I should make this link for the whole URL. So that's kind of from a practical point of view is something I would watch out for. A lot of people opt to replacing spaces and other special characters like commas or colons or semicolons with a dash. It makes it a little bit easy to read. It's something that from our point of view, Dell helps us to understand the words in the URL if you're kind of worried about that. And it makes it easy for people to share those URLs. So that's kind of what I would recommend doing there. From a technical point of view, if those links have spaces in them and we see them as clean links, then that continues to work. It's not that it would cause issues by itself in search, it's just that you're making it unnecessarily hard. It's something where if people want to link to your content, you should make it as easy as possible for them to actually do that. I can't get an answer on the webmaster forums for this question is.com.au, we treat it differently than .net.au. No, we treat them the same. They are country code top of the domains. So we essentially see them as content that's geo-targeted for Australia. You wouldn't be able to set the country targeting in Search Console because they're already targeted. And essentially they're the same. It's similar for .org or .net or any of the other essentially top level domains. It's not that we would say one top level domain is better than the other one. It's totally up to you. Yesterday I put a meta tag from Google webmaster on my blog, then I changed my template and the meta tag was gone and I have to put the meta tag back in again. Does Google keep crawling my blog even though I have this meta tag back and forth, essentially? Yes, Google will crawl and index your website regardless of any meta tag from Search Console that you might have on your website. So as long as we can crawl it from a technical point of view, it's not blocked by robots text. As long as you don't have a no index meta tag on these pages, then we'll try to index this content. So that's not something where being verified in Search Console is a requirement for search. A lot of sites aren't verified in Search Console and they do perfectly fine in Search as well. The thing kind of to watch out for is if this is a very new website, then it's normal for our algorithms to be a bit cautious and say, well, I don't know how much we should crawl here. I don't know how quickly we should crawl this website. I don't know where we should be showing it in the rankings in the search results because everything is just so new here. So from that point of view, if you're just getting started with a website, I would assume that it's going to take a couple of months for things to kind of settle down into more of a stable state. And that's independently of the Search Console meta tag. So it's good that you're watching out for this and it's good that you notice that it was gone when you changed the template, but it's probably not your primary priority when you're starting out with a new website. In Search Console under SiteMaps, it says 542 pages index, but under issues it shows 1,838 warnings. It states SiteMaps contain URL which are blocked by robot sex, but I have no URLs blocked by my robot sex. What could this be? So it's hard to say without looking at your website specifically. So my guess is in Search Console, you can find out more about the individual URLs that we're seeing as being blocked by robots text. And based on those sample URLs, you can probably figure out what is happening there. So sometimes we see sites submitting URLs that are within like the admin part of their website, it shouldn't be indexed, maybe they're blocked by robots text there. Sometimes we see other things where maybe a robots text file used to block a lot of URLs, but has changed to not blocking as many. And sometimes that takes a couple of days to actually be refreshed in the SiteMaps report. So those are kind of the higher level things I would look at there, but you should be able to find out a bit more what's actually happening there in Search Console. Can using Bing Search Console reduce my blog's page rank? Or I guess general ranking in Google? No. I mean, I don't know exactly what the Bing Webmaster tools do, but I'm assuming that you're using these to understand how your site is doing in Bing and probably finding technical issues on your website that you can solve in general and solving those kind of technical issues are good things for any search engine. So just because you're using a different tool to recognize and solve problems on your website, that wouldn't be a bad thing. That would actually be a good thing for your website, right? So feel free to use whichever tools work best for you, whichever tools bring you the information that you need to recognize technical problems, other kind of problems on your website and improving those will generally improve things for all search engines. If I'm trying to rank an Arabic keyword with having the URL containing the keyword in Arabic help, what if I'm using the English term instead of the Arabic one? Would that help the ranking for the Arabic keyword at all? So we do use keywords and URLs, but it's a very, very small factor. It's not the case that putting a keyword into a URL will probably won't even visibly change your site's ranking for those keywords. So that's kind of up to you. This kind of goes back to the way early question on spaces in the URL. I'd recommend making it as easy as possible for your users to recognize what your content is and how to link to it. And if your users primarily speak Arabic and they recognize Arabic words in the URL a lot better than English words, then use Arabic words. Match what you're doing to what you think users actually could profit from. Please update the SEO starter guide. I think I heard this in the last hangout as well. So I know the team is working on something there. How does Google treat faceted page links from category sub-category pages like black formal shoes, black leather formal shoes? Is it a good idea to keep them or point the canonical to the main page like formal shoes? Again, this is kind of a similar question to the paginated series. It really depends on what you want to have indexed and which content you have on those individual pages. If the content on kind of the higher level category pages is all you want to have indexed and links to the individual products that you want to have indexed as well, then maybe that's fine. On the other hand, if some products are only linked on the more sub-category type pages, then maybe it makes sense to keep those in the index as well. So that kind of depends on your website, how big your website is, how much content you have, how you organize that content. And it's less something where from our point of view, I would say you always need to do it like this or you always need to do it like that. A lot of the questions that we get are kind of in this group of things where I almost have to say it depends. And it's not that I'm trying to be lazy and not answer your question, but it's really something that you kind of build up the experience over time. And you know your website best. You know your user's best. And you kind of have to watch out for what works best for users and focus a bit on that. In duplicate content pages without a canonical link, I mentioned Google's fold all the value to one page. If yes, then that means the ranking of the canonical tag won't improve even after adding the canonical tag. Yes, that's correct. So if we're already folding things together into one URL, which is very common when we see duplicate content, then nothing really changes if you additionally add a rel canonical there. However, with the rel canonical, you can kind of guide us to which URL you want to use as a canonical page. So maybe you have two URLs and one is a nice URL that you think is really awesome when that's shown in search. And the other URL is a bit of a messy one. Maybe it has the keywords in the wrong order or something like that. Then if you tell us about this with the rel canonical and with the other signals, then we're more likely to actually pick the URL that you want. So it's not that the ranking will significantly change, but it's more that we'll kind of follow your lead and say, well, this guy sounds like he really knows what he's talking about. All these signals add up and point at the same page. So maybe we should just do what he says. When the switch to mobile first indexing occurs, if a site has a responsive design and AMP, will Google use the AMP responsive version or potentially both for ranking? So at the moment, the plan is if there is a mobile version of a site, then we will use that for indexing and ranking. That means if you have a responsive design, then the responsive version is the mobile version. That's the one we'll use. There are setups that you can do to make the AMP version the mobile version of a site. But it's kind of tricky to set up. And I haven't seen anyone in practice do that, where you have a desktop only site and an AMP page. And you treat AMP page kind of as an M dots version of your site. So technically, you could do it like that. And if you do do it like that, then we will use AMP version because that's the mobile version of your page. But in practice, most of the time, you have a mobile version already, and then you add an AMP as well. So that wouldn't change anything there. If a site links to its internal pages with tracking parameters, will it affect the crawling and indexing? Yes, it does. So if we need to crawl the internal tracking parameter URLs, then we will crawl a lot of URLs that maybe aren't as useful for search, that all lead to the same content. So in a case like that, what I would recommend doing is maybe using the URL parameter handling tool to let us know that these tracking parameters are irrelevant and that we can ignore them. So that helps us kind of. Other things you can do is try to avoid using internal tracking parameters within your internal URLs so that we don't have to kind of make this mental shift and say, well, it's linking to this page today, but maybe those parameters are irrelevant and we can kind of crawl it without the parameters so that we don't have to worry about which URL version we actually have to crawl. I created a new website two months ago and it's showing on Moz that I have a 11 out of 17 spam score. Is this normal? I have to be honest, I have no idea what a spam score on Moz is. So I have no idea what they use to calculate there, so I can't really comment on whether or not that's normal. You'd probably have to ask them. They're a really bunch of smart people and they make some really neat tools. So I would assume that they're looking at something for your website and maybe these are things that are actionable for you. Maybe these are things that are less actionable for you. So I believe you're also the website that's in Arabic and I know sometimes SEO tools have trouble understanding how things work together in non-English languages and especially if you're talking about a script that's the other way around and you don't have the traditional word breaks within your content like English text would have, then sometimes these SEO tools get a bit confused. So that's something I kind of take with a grain of salt but I would definitely check with the folks at Moz and kind of ask them what they think there because I really have no idea what they're testing. Google recommends creating pages with quality content. I wanted to know what can be done on product category pages or product listing pages which will improve the quality of the page. I really have no answer for you there that you put this meta tag on a page and then it's high quality content. So that's something where you kind of have to work with the content yourself and with the people that are actually using your website, maybe doing a user study as well, to figure out what you need to do on those specific pages. What reason would pages not be cached? Lots of reasons. It's hard to say. So the no archive meta tag is probably the most obvious version but for us the cached page is not something that we see critical for search. So that's not something where we would say see that as a problem if individual URLs didn't have a cached page. We tried to provide this for users because sometimes it's really useful but it's not something that I would assume means anything specific if it's not cached. Why isn't Google excluding the backlink factor and ranking only on basis of quality content and user signals as there's just too much spam with backlinks? There is also a lot of really useful signals from backlinks. So it's not the case that it's pure spam. So from that point of view, I can definitely see why it would make sense for our quality engineers to say, well, we still get a lot of value out of backlinks. And what we do, especially with regards to links, is we ignore a lot of links where we say, well, these are links that are obviously spammy. We will ignore them. We won't take them into account and we'll try to focus on the links that we think are still kind of relevant. So personally, I wouldn't see backlinks going away as a part of search. But obviously, these things evolve over time. And the nice part about search when we look at ranking is that there's just so many different signals that come together. So just because a competitor has a link from this website doesn't mean that you need to get that link, too. We could be ranking your site just as high or maybe even higher based on other things that we know about your website that aren't involving your links. So it's something where you definitely don't need to just copy and paste and do exactly the same thing as your competitors are doing. You can kind of leapfrog them by focusing on other parts of your website. Let's see. If I change the host country for a .com site with geotarget set as unlisted, does it have any effect on ranking? I change the .com site from US hosting due to issues we're posting in Australia. And ranking has kind of dropped to page 3 over a period of a couple of years, dependent on the search. So in general, yes, you can set it to unlisted. What happens there is kind of a subtle difference in that if you don't set a geotargeting for .com site or for a generic top level domain, then we will try to figure that out for you. We'll try to figure out, oh, this is probably targeting users in Australia. We will just automatically geotarget for Australia. Whereas if you set it to unlisted, then you're telling us you don't want us to make any guesses about this. You don't want to target any specific country. So that's sometimes useful for websites that are really, truly international, where you really, really don't have any specific country that you're primarily focusing on. And it's usually pretty rare that sites that they're geotargeting to unlisted. And I know of maybe two or three cases where we've ever told a website that they might see improvements by changing that to unlisted. And usually that involves things like moving to a domain name that was very strongly tied to a specific country. And now you want to be international. But it's really, really rare that I look at a site from a forum thread or another escalation, say they need to set this to unlisted. So I wouldn't assume that that's necessarily the case in your situation. And in particular, when you say over a period of a couple of years you've seen a drop in rankings, then from my point of view, that suggests that this is a completely natural change in rankings. So if you're looking at something that takes a couple of years to change in rankings, then probably this is just how things have settled down and has nothing to do with the more technical things on your website that you're doing there. So probably you'd want to focus more on things like the overall quality of your website and really kind of take things up to a next level from that point of view. The cached copy won't work if your site isn't Angular and you're not using pre-rendering, says Vishal. Yes or no. So the cached page that we show is the raw HTML that we receive from your server when we request that page. So if you're using JavaScript to generate your content, then we will link to those JavaScript files from that HTML page. But a lot of times due to kind of browser cross-origin protections, JavaScript won't work in exactly the same way if we fetch it from one URL with regards to fetching it from your website directly. So what probably happens in cases like that is depending on the JavaScript that you use on your website, if it can't run for security reasons within the browser on a different domain, then it won't work. So it's not necessarily the case that the cache breaks when you use JavaScript. It's just from a technical browser point of view, it's on a different origin. And if you pull content from another origin in there, then there are certain kind of protections in play on the browser that prevent that JavaScript from doing exactly the same thing as it did before. And again, that's not something that would be indicative of what we actually use for indexing. The cache page is really just something that we do for users. It's not something that we say is reflective of what we index. When we can render a page that's using JavaScript and pull in all of that content for indexing, we will use the content from the JavaScript there, even though we show a fairly empty cache page. All right. Let's see. Bunch of questions left. That's good. When I use SEO tools, most of these tools give me a bad readability score. How is Google really seeing my page? And how can I get a better readability score? Or does Google not care? So again, I believe your website is the one that was in Arabic. So if you use SEO tools that are focusing on English content, then I would assume that it gets totally confused when it sees kind of the right to left content in Arabic and kind of the different word patterns where you don't have spaces in between the individual words and the words don't map to specific English words. So I would take these SEO tools with a big amount of salt. And especially if you're thinking about things like readability, which I think is great, then I would work more with actual people who understand Arabic and get their input on this rather than just taking a random score that some SEO tool pulls together. I know SEO tools are getting a lot better, and international websites is a really big topic. So some tools might be better at this than others, but especially if it's a tool that's really focusing on English content, you kind of need to be careful and understand what the tool does before taking any action based on the numbers that it pulls out. I have pages marked with the rel next and previous and the rel canonical, but I still see them in the HTML improvements section in Search Console. What's up with that? So we show things in the HTML improvement section when we've indexed them like that. And even though you have rel next and rel previous and rel canonical set on there, we will probably index that page first in the first step for the way that it is. And then in a second step, we'll process the canonicals and kind of the links and things like that. So if you look at that list and you say, oh, I know what is happening here. I already have solved this with a rel canonical there, then that's fine. This list of HTML improvements is a suggestion for you to kind of help find issues that you might want to focus on. It's not a sign that we count these against your website when it comes to search. Let's see. If a website URL path is updated across all pages, for example, slash 2 was removed from the URL, how could that affect the rankings with a delay of one week? So in general, whenever you update the URL structure of your website, whether it's like adding a subdirectory or removing.html or removing.php and the endings of the URLs, then that's something that means for us that all of the old URLs that we know about are not really valid anymore and we have to completely re-understand your website again. So what usually happens there is we start seeing a bunch of redirects from the old versions of the URLs to the new ones. So I assume you're doing redirects. If not, you definitely need to do that. But again, we'd see kind of all of these redirects. We'd see the changes in the internal linking within the website. And then we kind of have to figure out which of these pages map to the old ones and then kind of figure out how that whole structure works together again. So that's something where I would definitely expect to see some amount of fluctuation over a longer period of time until we can really understand your website properly again. So obviously, you've done this already. It's not something I would roll back. But in general, when you're looking at a website and you're going, oh, this would be nicer without HTML or without that HTML, or this would be nicer with all of the URLs in uppercase or with dashes instead of spaces, then that's something I would, at first step, kind of collect and think about what other changes in my URL structure do I want to do for my website in the foreseeable future and try to make all of those changes at the same time so that you don't have these fluctuations every time you make a small change. So if you change your spaces to dashes and then a few months down the road, you say, oh, I can remove that HTML as well, then every time you're making these changes in the URL structure of your website, that means all of the URLs on your website are changing. That means you're seeing fluctuations again. So that's something you usually want to try to avoid doing too often and instead kind of plan ahead and try to figure out what you want to have in the end. I have a couple of websites. For some reason, Google keeps blocking our website, mentioning that it's not safe to visit. However, we have the whole website scanned and there is nothing in our files. Google sends us to review the website. We check it. Everything seems clear and the review goes through and then it comes back again. What's up with this? So I looked at a few of the websites that you mentioned there. And from what I can see is that you have occasional problems with phishing content on your website. And the phishing URLs that I saw on our site were URLs within some of the common content on your website. So things like include directories or JavaScript files that you're sharing across different plugins, perhaps. And usually, for me, that means someone was able to manipulate your website, hack into it, and place these phishing files on your website and is able to do that fairly quickly and take those out again. So often they will come, place some phishing landing pages on your website, use those for email phishing, collect email addresses and passwords from users until your website is blocked, and then they remove those files so that you as a webmaster, when you look at that, you're like, my website is fine. I don't know what the problem is. Why is Google complaining? But actually, there are some vulnerabilities probably still in place that allow the hacker to kind of go back and say, well, it's not blocked anymore. I will just add those phishing pages again. So I would definitely look at the URLs that we send out in those notifications, the URLs that are specific to the phishing content, and think about maybe which plugins have access there. Think about what you can do in general to make sure that your website stays on top of all updates across all plugins. Maybe do a general security review where you also check your server logs to see what is actually happening on your website so that you can make sure that there is really nobody out there who's kind of hacking into your site and placing this content there. And sometimes this is really tricky. These hackers can be really, really fancy with how they place this content so that you, as a webmaster, when you look at those URLs, you might see a normal 404 page. But if someone comes in via email, they would see the normal phishing page that's asking them for username and password. So that's something where sometimes it also makes sense to get professional help from a security company that knows what to watch out for. And oftentimes it's a matter of cleaning things up properly, locking things down properly, and putting good processes in place to kind of prevent this from happening in the future. Let's see. Does Google read EXIF data in images such as Geo? We optimize our images, but it strips EXIF and wonder if we should keep the EXIF data. No. As far as I know, we don't use any of the EXIF data in images for image search. So if you want to strip that out like the location, that's perfectly fine from my point of view. I think there is one exception where we try to pull out some like a title of the image or the name of the image from the EXIF files. But things like Geo information or the camera information, the camera settings, all of that we definitely don't use for image search, image ranking. So what you can do if some of this content is important for you that users actually recognize, and you think users might be searching for that specifically, I would pull that out and put that on the landing page itself, maybe near the image so that we can understand that it belongs to the image. And that helps us to figure out what the best landing page, the best images for some of these queries. Let's see, some kind of long question. How long does it take for Google to de-index a website? I have a new site that was indexed in a week. And after that, I set the robots text to disallow. And what could be the problem there? So if you set the robots text to disallow everything in your website, then what happens there is not that we remove your website from search, but rather that we remove the content that we knew about on your pages from search. So the snippet and the general kind of on-page information that you might have on a page we wouldn't be able to take from your website. But we might keep your URL in our index based on other factors that we know. So if there are links pointing at your URL and maybe those links have an anchor text and all of that kind of tells us, well, this URL is probably about this specific topic. Maybe we can rank it for that topic, even though we don't know at all what is on this page. So if you want to have a page or website removed from search, what you would need to do is let us crawl those URLs and let us see that there is actually nothing there that we can index. So that can be by returning a 404 if the website doesn't exist. That can be by having a no index on those pages if you want the website to exist, but not the index. So those are kind of the main options there. If you need to remove a website urgently from search, which sometimes happens, you can use Search Console for that. So you'd have to verify ownership of that website first in Search Console. That also might mean removing the disallow so that we can see the verification file or the verification meta tag. And when you've verified it in Search Console, you can do a site removal request. And then for a certain period of time, I believe it's 90 days, we will remove your whole website from search pretty much on the same date. So that's probably the fastest way if you really need to have that website removed from search. Our company is Chocolate Supplier. Our blog has hundreds of blog posts with a few being technical chocolate information. As a result, these few posts don't provide a value to the wide audience, but rather to a niche audience. Does this mean the quality of these specific blog pages are low since we have a high balance rate due to the audience not understanding the content? Should we remove those pages? Would that help improve our website? So I think this is something that a lot of sites kind of find challenging. And from our point of view, you can definitely leave those pages there. But what I would recommend doing there is trying to work out which audience you want to target with your website in general. And if that audience includes people who love seeing these technical posts, then I would definitely keep those. Whereas if the audience that you want to target is really just a consumer audience and you know they don't care about how fast you stirred the chocolate or which temperature and all of that stuff, then maybe it makes sense to put those on no index so that they don't kind of bubble up for people who are just searching for, I don't know, cute chocolate candies or something like that. But it really depends on you. It's not something where I would say there are a few people reading this and a lot of people reading this. Therefore, I need to remove this kind of smaller audience content. Sometimes that smaller audience is actually exactly who you want to target or an important part of your audience anyway, so I would keep that. So I wouldn't blindly look at things like bounce rate and analytics or kind of time on site for individual pages, but rather kind of think about what you want to do with your website and what you want to put on your website to handle that. All right, have a bit of time left. I think I have this room for a little bit longer, so we can keep going a bit. What else is on your mind? John, can I just say if I had a quick couple of questions, please? Sure. So first question. We've got a phrase that is commonly used as a commercial term, which would be a brand. So for example, Vacu, so Vacu, so even. How would you feel about getting recognized as a brand's name, because at the minute it's too general and too top level, but it's not? Because we want to get it in the knowledge graph and we want to get it to return site links as well. So what would be the best course of action to get it written as a brand? Probably outside of Google. So usually what happens there is we try to pick up signals across the board, whatever we find across the web. And it's not the case that you can just mark up around a word and claim it's a brand and then suddenly we will treat it differently. It's more the situation where we see that something is obviously a brand based on how users search for it, how users react around that content, and based on all of this bigger picture thing, then we might say, oh, if people are searching for this word, that means they want to have information about the brand, which means we should be showing maybe knowledge graph entry or Wikipedia page, all of those things. So there is no technical thing that from our point of view, you should just put around your words on their page. It's more a matter of doing all of the hard work that's associated with building a brand in real life. Would the Google My Business help with that in any way, or if it's such a good Google My Business? People are searching for that, and you have a local entry for that. Why not? I mean, these are all small things where you're kind of doing the brand building work where you're making sure that you're present when people are searching for you and you're showing your name when people are searching for your general type of content. And all of these are small things that kind of add up and help to change people's minds about your business, about what you're offering there. So looking at all of these small things, it maybe makes sense to pull together lists, or probably what I would do in a case like this is go and look at some marketing textbooks and kind of see what people are doing offline when it comes to brand building and see which of that applies to your specific case. OK, cool. And the second question then, what punctuation is ignored when calling a website to rank for, say, nonsensical terms? So for example, the phrase would be serviced offices London. Now, obviously, that doesn't make sense by itself. But if we're to write it into a sentence, say, thinking about serviced offices, comma, London-based businesses, my, et cetera. So that's obviously in a sentence where we've got a comma in between. It will in the middle of the key phrase that we're looking to target. Problem that we've got there is we've got competitors who are outranking us for this phrase just by stuffing it in their text. Now, we don't want to follow suit. We do want it to grammatically make sense. So what would you recommend in that instance? Are commas somehow, in terms of proximity of words, is that increasing the proximity between the words or what's the crack there? I would try to keep it as natural as possible. So if that means you have commas between those words, go for it. So that's something where I wouldn't worry too much about individual punctuation and how that works. But at the same time, if it's a phrase that doesn't really make sense, then probably people aren't really searching for that either. So that's something you kind of have to keep in mind, is I have this list of 5 million keywords and phrases that I want to target. Which one of those are actually things that are relevant for my business? So some people might have searched for this weird combination of words, but does that mean that it's really a critical aspect of my online presence or not? Yeah, I think it's certainly got a search form behind it. So people are searching for it, but they're being kind of lazy and quick as searches are these days. That's why they're just quickly serving services in London and then just see what kind of pops up. But obviously, to make it as relevant, yeah, it's a tricky one really. Because we don't want it to us. It looks as though Google is kind of rewarding them for being a bit spammy and being a bit keyword stuff. I'm sure that clearly, I'm sure that isn't the case, but we don't really know what to say to the client. You've written it well, yet you're falling behind these competitors that have stuffed it. And naturally, their main reaction is, well, why can't we do the same thing? It's obviously working. It's what Google wants. And it's kind of trying to push that back a little bit. It's a good situation. I'm sure there's many situations like this. It kind of comes up. But I completely agree with what you're saying, right in a grammatical way that makes sense and makes sense to users. But we want people to be able to see that content. And if they're not accessing the website because it's not ranking as well, then that's the problem I think we face. Yeah, I mean, it's something also where you kind of have to look at what Google could be doing past just keyword matching. Because keyword matching is something that is obviously pretty old school. And we have a lot of engineers at work on improving search. And they're not going to sit there and say, oh, five words next to each other in the comma in between. Therefore, this page should rank or should not rank because of this comma. They're going to try to figure out what is it that you're trying to say with this specific query? Does London mean that London has to be the next word here? Or does London refer to a location? And if this page is clearly about this location, then maybe that's enough for us. Maybe that full phrase never actually has to be in that order on the page as long as we can see, oh, this is about service offices. And this page is clearly London focused. So this page might be a really good match. So it's kind of, I'd say, focusing on the keywords that people are actually using is certainly a good thing. But I would try to look at the bigger picture and figure out what is it that people are actually searching for and how can you cover those aspects on a page in a way that is easily recognized by users and by fancy algorithms. OK, OK. Well, I'll put those up. So one example that I saw at an SEO conference a while back, this was in German, where someone was doing something similar where they were trying to get their business plus city name to rank because they were based in the city. What they noticed is they didn't have a local phone number on their page. So essentially, our algorithms, for whatever reason, looked at this page and thought, oh, this is about this topic. But it's not clearly located in the city that the user is actually searching for. So maybe it's not actually the type of result that people are looking for. So kind of clearly covering both of those standpoints that apparently helped them quite a bit to make it clear this page is really about this topic. We really do this kind of business. And we're obviously located in this city. We have a local address. We have a local phone number. It's like everything lines up for us to say, well, duh, it's certainly in the city. And it certainly does this. Therefore, the combination of those two words, actually, this is a good match. OK, I guess there's something to look at. This wasn't a question of mine. But yeah, I'll speak to the person and see if we can find. But yeah, that gives us another way to look at it, I suppose. No, thank you for that, and I appreciate it. Sure. All right. More questions. What else is on your mind? Hey, John. Hi. Yeah, actually, we run multiple news websites. And what happens is sometimes we have different teams for both the websites. Now, what happens is one of our team covers the news first. And they both are relevant websites. So what happens is we have to give them the source and credits for that particular article. Is that OK if we give them a no-follow source link? Sure. So what I've seen from other kind of bigger news businesses where you have multiple websites covering the same topic is that they will often use a rel canonical on the other pages, say this piece of content is relevant for these three local news places because they also cover kind of regional news. But this is the one that we really, really want to have indexed. And with a rel canonical, you can kind of cover that fairly well. And you can say, well, there are three versions here, but this is the one that is most important for me. Maybe it's the one that monetizes best for you. That's, I think, a good option as well. And you were kind of telling us, combine all of these three pages into this one strong version. So instead of just putting a link to the version that you think is the original source, maybe it makes sense to say rel canonical to this version. And this is my really my preferred version, the one that I want to have indexed. OK, yeah. Thank you. All right. Yes, I can barely hear you. But go ahead. Are you going now? Yes. All right, so just wanted to know a man around this near duplicate pages, basically. We have plenty of pages which is actually on the same topic. So just wanted to know if we can actually put canonical on these two main pages or just redirect them to our main pages. So basically, you can see one of the examples of these ones, like the formal leather shoes, something like that. But there ultimately are the formal shoes. So just wondering if Google try to understand the topic itself, like the majority of the content is talked about, the formal shoes, it's just the difference in the attributes, like what the users are actually looking for. If we can optimize our main page with all these attributes by making content itself, making pages, or you can say some of those section of the page and include those words, or maybe doing the altogether and paragraph around explaining the user, basically, what they're actually getting. So is it somehow helpful? Sure. I mean, that's something you can do. If you know that these kind of sub-sub-category pages map best to a higher level category page, then using a rel canonical is fine. It's just important to keep in mind, with the rel canonical, when we follow that, we ignore the content on the other page. So if you have a page about blue shoes and you say rel canonical is my sporting shoes category, then if the sporting shoes category doesn't have any content around blue shoes, then we don't know anything about blue shoes for your website. So that's something kind of to keep in mind. If you do have all of the details covered on the main category page as well, then perfectly fine. Go for that. Thank you, John. All right. So it sounds like we were kind of covering all the questions here. Let's take a break here. I'll set up the next batch of Hangouts in two weeks again. So if there's anything on your mind, until then, feel free to jump on in there. In the meantime, feel free to use the Webmaster Help Forum as well. Contact us on Twitter if you want. And hopefully, things will work out in your way. So wish you all a great weekend. Thanks for joining in. Thanks for all of the questions that were submitted. And see you maybe next time again. Bye, everyone. Bye, John. Have a good weekend. Yeah, bye.