Reconsideration Requests
Show Video

Google+ Hangouts - Office Hours - 25 August 2014

Direct link to this YouTube Video »

Key Questions Below

All questions have Show Video links that will fast forward to the appropriate place in the video.
Transcript Of The Office Hours Hangout
Click on any line of text to go to that point in the video

JOHN MUELLER: OK. Welcome, everyone, to today's Google Webmaster Central Office Hours Hangouts. My name is John Meuller. I am a Webmaster Trends Analyst at Google Switzerland, and I'm here to help you guys with any web search, webmaster related questions you might have about Google and everything that's associated with that. As always, does one of you want to get started and ask the first question, before we start with the Q&A?

BARUCH: Is it true that we have to verify all websites in order to get all the backlinks?

JOHN MUELLER: To get the information about the backlinks, yeah.


JOHN MUELLER: They still count. They're still there. But in Webmaster Tools, we show it, I believe, per hosting and per protocol. So if you have some links to the www versions, some links to a non-www version, then that's something you might want to verify separately, so that you have that information all together.

BARUCH: So I can get all the backlinks, just from verifying one website. If I want to get the exact amount of backlinks, I need to verify a couple of websites, right? In order to get all the data?

JOHN MUELLER: So we do this by the URL that we have it indexed by. And if you have your website consistently indexed with one version of the host name and protocol, then we'd show all of that information there. But if some of URLs are indexed with www, some without www, then that's something we have to keep track of separately. And we'd show you the links, the crawl errors, those kind of things, separately for those different parts of your site, or the different variations of the same content.

BARUCH: OK. Thank you.

JOHN MUELLER: So if you have things like canonicalization set up properly, the rel=canonical 301 redirect all to one version, then that's what we're going to focus on. But if you don't have it set up like that, then maybe we'll have some in one version, some in the other version.

BARUCH: OK. Thank you.

JOHN MUELLER: All right. We have this question by Barry Schwartz. When is Penguin 3 launching? [LAUGHTER] I don't really have anything new to announce at this point, so I can't really tell you anything special, Barry. Sorry. But I'm sure you'll hear from us or from the general webmaster community when things change. I get different site links when changing the browser. Does Google personalize site links? If yes, what's the effect Google site links with personalizing? As far as I know, site links aren't personalized, but they are generally set up in a way that they're different per language. So if you have one browser set to English, or your Google cookies set to English, and you open an incognito window that-- I don't know-- has a different language, by default, then that's something where you'd see a difference in the site links. Because if you're searching for your company name, and your whole settings are set for Germany, then we'd probably show you different information than if we see your settings are set to English. So I don't know if that's the effect that you're seeing there. That seems like it might be what you're actually doing. Barry asks, if you're watching live, you can't submit questions via the Q&A. I thought you can. No? Oh, well. OK.

BARUCH: We can't see you, as well.

JOHN MUELLER: What? You can't see--

BARUCH: We can't see you. No. We can only see Googlebot.


BARRY: I can see John.

MALE SPEAKER: Oh, now we can see.

JOSH: I see you just fine, John.

JOHN MUELLER: OK. I was just about to make a funny face, if you guys can't see me. That's probably not the best thing. OK.

KEVIN: Hey, John? I had a question.


KEVIN: So I was talking to Matt [INAUDIBLE] at SMX West. And I posed a question to him that he was intrigued by. If you have a franchise or dealer network, like Pizza Hut, and they are primarily concerned in geo-related terms, the duplicate content issue kicks in. And we've tested this. We had a grouping of sites at 235 top level domains. So these are not sitting on a manufacturer subdirectory or such. My question to him is, if that company has figured out exactly what would represent their brand very well, would convert very well, why would they be forced to have to create false stuff on a page to get around the duplicate content issue? Shouldn't the team be able to see that, when a geo-qualified query is put in place, that the duplicate issue should not play? So if I look for a pizza in Sarasota, Florida, then I look for pizza in Fort Wayne, Indiana, shouldn't you guys be able to figure out that it doesn't matter that the content is different on each page? Or the content is the same on both page, because it is absolutely relevant for that market. Now I can tell you what Matt said, but I'm interested if that ever made it to the team.

JOHN MUELLER: OK. It'd be interesting to see what's exactly happening there. So if you have specific URLs that you can send me, that would be useful. But in general, there are two places where we filter duplicate content. And depending on how much of the content is different on these pages and how much is kind of the same, or the way the website as a whole is structured, the network of sites is structured, it might fall into one or the other category. So the first thing we do is, when we index these pages, when we crawl them and see that they're exactly the same, or it looks like these are essentially exactly the same pages, we might fold them together into one URL and index them like that. If we can tell that they're actually different pages, that these are essentially meant to be indexed separately, even if there's only location, or the phone number, or something like that has changed on these pages, then we'll try to index them separately. And at the time of the search results when someone searches for something, we'll try to filter there. So if someone is searching just for pizza, we'll pick one of these pages that are probably, essentially, all generic for that term. If someone is searching for pizza and a specific phone number or a specific location, we'll try to find the page that's matching most specifically within this group of pages and just show that one and filter the other ones out. So essentially--

KEVIN: Unfortunately-- go ahead.

JOHN MUELLER: Yeah. We should be probably catching these a little bit better, so that we're indexing them separately. But it sounds like you're saying we're indexing them as one page and dropping them way in the beginning.

KEVIN: Yeah. So we had a test where we did this across 235 top level domains. We had moved them from a directory listing style, which was obviously performing very poorly because of thin content. And it wasn't relevant to the market when, basically, it was just a couple of terms on the page. So we moved them to 235 unique top level domains. We let that go for awhile. Some of the pages performed OK for geo-targeted queries. But the rest of them would not, because they-- they were indexed, but they were 100 listings search results back. When we deduped the content to around 30% to 40%-- which we would have rather not had to do, because the message was exactly what the client wanted-- and we put some variability onto the page that was outside of just geo-keyword and the company brand name, those kind of standard variables, then all of a sudden, they start to rank very highly. And so the quandary there, as I said to Matt, is we're not trying to trick you guys. We want to give you exactly-- in other words, Matt had said, well, why don't you just put one piece of content on that front page, like who the general manager is of each location? And I said, so basically, what you're telling me is you want me to present something different to you guys that we haven't found to convert better, because you need that so that you can discover this. So anyway, he took it down. And I wondered if that made it to the team? If you guys are working anything more on that? Or if you can help us try to stay above board with you guys? Because we don't want to have to write-- in the case of an upcoming one, we're going to have to write 2,000 different pieces of content.

JOHN MUELLER: Yeah. I probably need to look at the specifics there. I know this is something that the team is always working on. Finding ways to filter them at the right time is always a kind of a tricky thing. And at the same time, it's tricky sometimes in that some of these cases start to look like doorway pages where, essentially, it looks like-- for example, if you have an online service, and you essentially just want to advertise that to all the cities and neighborhoods, you create separate domain names for your online service and then the city name dot com. And all of those pages are essentially the same. And we need to treat them as one page, instead of as all of these separate pages. So it's tricky to find the right balance there. So that's something where I'd love to look at the specifics and see the real problem that you're seeing there, so that we can discuss the current issue with the team. [INTERPOSING VOICES] They're looking for examples. They'd love to see what's happening, where they're getting it right, where they're getting it wrong.

KEVIN: Right. We just want to make sure that the end client-- as you guys say that you want to do, when somebody is searching in Sarasota, Florida, they don't want a corporate page to come up. They want something that has local relevance and in local context. And we're trying to provide that. And we're actually doing quite well for it, but it's very labor-intensive to do that because your algorithms are not picking up when we only want it to be geolocated.


KEVIN: So I'd love to talk more about that, yeah.

JOHN MUELLER: Yeah. Feel free to send me the information in, maybe, a private Google+ thread. And I can see what we can do there.

BARUCH: You're not using Google Maps for each location, though?

KEVIN: Yeah. We use all of the IYP and so forth. But we're not just talking about showing up well in the Google local, we also want them to show up very well for organic local listings because, obviously, there's a lot of variability in the Map Pack and the 3-Pack and the 7-Pack, et cetera. And that business should score well when a geolocated term is used.

BARUCH: All right.

JOHN MUELLER: Yeah. I'd love to look at the example, so we can ticket that to a team.

KEVIN: Awesome. Thank you, John.

JOHN MUELLER: Thanks. All right. More from the questions here. The [INAUDIBLE] option in Webmaster Tools hides the specified URL errors. But does it have any effect on crawling or indexing? No. This is actually only for the UI, only to make it easier for you to go through your crawl layer. So it doesn't change anything for crawling and indexing. It doesn't re-crawl the URLs. It doesn't change the error status. It's really only in the Webmaster Tools UI. How important is private/public WHOIS information as a ranking signal, especially in a spammy niche such as casinos, pharmaceuticals, payday loans? We don't take that into account at all, as far as I know. So this is essentially something up to you. If you want to hide your WHOIS information, you can do that. If you want to make it public, that's fine as well. That's not something we'd look at as a ranking signal. In case your domain is penalized by an algorithm, and you've done all the possible adjustments to make it good, but your website is still demoted after many months, is then moving to a new domain to make a fresh, new start a reasonable solution? Sometimes. Sometimes, that can make sense. Sometimes, it's worth working on your old domain. It really depends on your situation, and not something where I'd say there is a specific time frame you should wait for or a specific set of changes that you need to make. Its really up to you. Also, the effort involved in moving to a new domain or starting fresh on a new domain is essentially personalized. It's not something where we would know how much is involved in doing that kind of a change. So this is more of a business and a marketing decision that you have to make at some point. Sometimes, it makes sense. Sometimes it makes sense to hold onto the old domain and maybe find a different solution. [INAUDIBLE] I get quite a lot of complaints about the Google dance. Do you have any comment about this, whether Google assessing Penguin 3 by now? At the moment, I don't have anything special to announce. I know that engineers are always working on improving search, so it's normal to always see some changes in search. But it's not something specific that we have to announce at the moment.

BARUCH: [INAUDIBLE] changes, right?

JOHN MUELLER: We make a lot of changes. Yeah. I mean, it's hard to quantify that, because there is so many also shared systems and algorithms at Google, that you can't really say this is just specific to search, and it doesn't affect anything else. Or this is just specific to-- I don't know--- Google Drive and not using URLs. So we make a lot of changes. We have a lot of engineers working on search to keep things moving forward. Manual action revoked for unnatural incoming links. Site-wide, the rankings are still low. If we create a new site without 301 redirecting, can we use the old content? Is the content penalized as well? Essentially, if you set up a new website, and you don't redirect, then that's something we try to see as a separate website. If the new website looks exactly the same as the old website, then it starts to look to us like you're just moving to different domain, but you forgot to set up the redirect. So that's something where, from a technical point of view, if it looks like you're just getting canonicalization wrong, we'll say, oh, well, we'd like to help you with this and treat it as one website, which would also involve forwarding those spammy signals there. So as much as possible, I'd recommend, if you're going to go to a different domain because of problems that you have with the old one, really set up a new website so that we don't see that as one website and try to forward everything there. We uploaded a blog by mistake to the wrong site which has been indexed. So what's the best way to transfer the content to the correct site without risking duplication? Shall we simply remove the one and add to the other? Or do we need to do a 301 redirect? It kind of depends on the website and what you did there exactly. If you just uploaded it to the wrong domain and you fix it a little bit later, then chances are you can just delete the old one that you uploaded incorrectly and put the new one up. If you've left it like this for awhile, then I'd definitely set up redirects so that we can forward all of those signals to your preferred version that you do want to have indexed. But even without redirects, we'll probably figure this out, if it was changed within a reasonable time after your initial upload. Google announced that HTTPS is a ranking signal. A lightweight signal for the moment that's SSL on the checkout page. Is that enough for the intention to make all site pages HTTPS? What's the best practice? Essentially, we see this on a per page basis. So when we do the rankings, we look to see if it's indexed as HTTPS and treat it appropriately. So if you're just putting this on those checkout pages, then those will be the pages that we'd treat like this in the search results. So from that point of view, our recommendation would be, really, to move the whole site to HTTPS, if you can. I realize it's easier said than done. And for some sites, it's very involved to actually move everything to HTTPS. So this is, like you mentioned, a very lightweight signal at the moment. It's not something where we'd say you need to do this immediately, otherwise, your site will disappear from the internet. It's a recommendation that we have, but you can do that, step-by-step, whenever you have time.

BARRY: John, can I ask a question?


BARRY: So HTTPS, so a lot of people, obviously, did the migration. Some URLs stay secure. But there's some mismatched content errors and stuff like that. Obviously, you have to be secure. You validate that it actually has a valid certificate. You validate that the page is secure. But do you also validate that there are mismatch errors on there? I mean, how far along do you go where it's not getting that tiny ranking boost?

JOHN MUELLER: At the moment, we look at just if the URL is indexed as HTTPS or not. We use the content mismatch issues to determine canonicalization. So if you have the same content on both HTTP and HTTPS, and you don't have a redirect set up in either way, and we can tell that the HTTPS version isn't perfect in that you have these cross-site issues with HTTP, HTTPS, then we'll fall towards the HTTP version for canonicalization. But if you set up a redirect, then we'll probably just follow that redirect and index it as HTTPS.

BARRY: Even if that has mismatch errors, it will still get that ranking boost?

JOHN MUELLER: Even if it has some mismatch errors at the moment, yes. It's possible that that will change at some point. It's also possible that, maybe, we'll be a little bit stricter with regards to which certificates work for this, in the sense that, if you have a really old certificate with a lower bit count key, then that's something that we might say, well, it's not really secure. We should probably treat it a little bit differently. But I think, at least for the current time, for the midground, it's probably just going to remain focused on the HTTPS part. I think that's a really big stake that a lot of sites still have to make, so it's something where we don't want to further complicate things unnecessarily.

BARRY: All right. Thanks.

JOHN MUELLER: Do shopping comparison sites have a low priority, as the content is largely aggregated from different sites? It's not the case that we treat these shopping comparison sites differently at any rate. It's more the case that we just don't find that much useful and compelling content on a lot of these sites. So it's not that, by definition, we'd say these are bad sites. But it's just that a lot of sites take a bunch of feeds and compare them and publish them one-to-one, which is essentially auto-generated content. Nothing unique and valuable is added there. So that's kind of what I'd focus on there, not the type of site, but rather just making sure that, regardless of the type of site you have, you really have significant, unique and compelling content of your own in there. I'm moving from page HTML to without the .html. As I'm moving to WordPress, will a 301 redirect be enough for the rankings to flow through? Yes. A redirect would definitely be the right thing to do there. There is a tiny amount of page rank that's kind of lost with a redirect. But if you've set up these redirects, that will essentially forward all the signals, or most of the signals that we had from the old page to the new page. Not having .html at the end is fine. You can have .html there as well. It's not that we have any preference either way. Sometimes, you can also set up a new CMS when you move to use the exact URLs as you had before. So I'm sure you could set up WordPress to have .html at the end if you wanted to. And if you can do that, then that's something I generally prefer doing. Then, that way, you don't have any of these problems associated with a move to different URLs within your website while we have to re-crawl everything, re-index everything, re-understand the site structure, the way the links are between the pages. We can essentially, if you keep the same URLs, we can just keep running the way that we have your site now.

ROBB: John?


ROBB: Can you just expand upon what the thinking is behind losing any rank at all for a 301? If the site owner is aware and conscious that, OK, I'm moving here, telling everyone by doing the right code that it's moving here, Google picks it up as it's moved here. What's the thinking behind any loss at all?

JOHN MUELLER: I'd have to double-check. [LAUGHTER] I'm not sure, offhand, what we had there. I know we had some spam issues where sites would essentially play the 301 game and just cycle things around. And for those kind of issues, it can make sense to treat them more like a very strong link, rather than taking everything one-to-one and forward it on. So I know that's something that used to be a problem. I'm not sure if that's the complete answer to your question there.

ROBB: Now that it's moving here, Google picks it up as it's moved here. What's the thinking behind any loss at all?

JOHN MUELLER: Oops. Someone has an echo. I'd have to [INAUDIBLE]. Yeah. So I'd have to double-check for a complete answer there.

ROBB: OK. Thanks. Do you think it's negligible, though?

JOHN MUELLER: Yeah. Yeah. I mean, for normal moves, it's not something where you'd see that in any way. Usually, it's also one of the reasons why we recommend telling people to update their links, if they have links to the old versions, as well. So that's just to kind of keep things sanitized and keep things moving forward, but it's not something where I'd say you would see any noticeable drop in the search results by doing this once or twice. Of course, if you're doing a row of these and cycling through a bunch of domains, then that's something where you'll probably see a small or visible drop. But if you're just moving from one domain to another, it's technically there. It's worth mentioning. But it's not something where you'd really have to worry about that.


FERNANDO: Hello, John?


FERNANDO: Hello. Can I make you a question?


FERNANDO: Sometime I go in a previous Hangout. We talked about how a good amount of good and natural backlinks are not the only factor to achieve a top 10 position in very competitive queries, as an example. We work on a higher website. And it started with a brand new domain three years ago, or so. But you talk about doing something remarkable and different than the other websites to be on this top 10 position. And now we are doing that. We are working on very good content with a little bit of insider information useful for the traveler. We are creating all of this, especially here in Spain, for this kind of content. But this content is not obtaining very good positions. And I would like to ask you how much time does it take to have all the backlink power you know, or the recognition for Google for this new content on the overall website to have a very good position I think it deserves, with the backlinks we have and the content we are creating?

JOHN MUELLER: It's really hard to say. I don't think there's any specific time frame where I'd say, after this time, you should see these changes. From a technical point of view, it can take quite some time to crawl and index all of this new content, which could be, depending on the size of your website, in the order of weeks, or a couple of months. In addition, what you'll see is the quality algorithms have to understand the new signals about your website, the way you're writing, the content you have there. That's something where I'd imagine-- I don't know, I'm pulling a number out of my head-- maybe a month or two as well there. And then we have to update those algorithms. So if you're making significant content changes on your website, and your website is of a reasonable size, I'd guess maybe around a half a year or so for you to see those changes visible in search. So it's not something you'd see from one day to the next, also not within weeks. It's more of a matter of months.

FERNANDO: Yeah. Thank you, very much. It's something that we see. It's a very competitive field of the field we're working. The time is very slow. We're a little startup. And we've got crazy with this, because we want things faster. But we understand a lot the quality and the responsibility of Google to bring only good websites in top 10.

JOHN MUELLER: Yeah. It's really tricky when you're going from a-- I don't remember your website in detail. But when you're going from a website that's kind of iffy from the quality to something that's really good, it's really something where our algorithms are sometimes a bit reluctant in saying, OK, well, this page is good. Therefore, the whole website must be good as well. We kind of need a strong confirmation from various signals to really trust that information and say, well, this whole website changed significantly. We should treat it very differently in search, going forward. So that's something that just takes quite a bit of time.

FERNANDO: Thank you, very much, John.

JOHN MUELLER: All right. A client side has multiple backlinks from article farms, but ranks well for many terms. Should I remove them, as a rule? And will this improve the search results further, or still hinder the already good results? That's a good question. So essentially, kind of what you're asking there is are-- I'm going to assume that these articles farm links are unnatural links that you placed there, or that you had someone place in their articles. And if that's the case, essentially, what you're asking is you-- a new client comes to you and has a lot of unnatural backlinks, should you be proactive and clean that up? Or should you leave them there and hope Google doesn't notice? And from my point of view, the answer is obviously cleaning that up is always a good idea, so that you don't have any potential issues going forward. So maybe they haven't been picked up manually from our Webspam team, but if you leave there, that's always kind of a potential risk that you run into going forward with a site like that. Or maybe the algorithms will pick that up and say, well, these are all bad links. And I don't know how to treat the rest of the links, so maybe we should be careful with this site. On the other hand, if these are essentially just random links that happen to be an article farms, but these are things that you placed unnaturally or that you had some SEO place for this website, then that's something where you can make a judgment call there and say, these are kind of OK, these are not too problematic, or these are just things I don't want to be associated with. And I don't want to run the risk of them being seen as being unnatural. And we'll just disavow them, maybe, on our domain level, to kind of make things a bit easier. So essentially, if you know that they're unnatural backlinks, I'd definitely clean that up, going forward, and be proactive about that. If you're not sure about that, then that's kind of the judgment call you can make. With the disavow tool, you have the ability to clean that up without actually changing those sites. So that might be a possibility there. But that's more something where you have to make a judgment call, as you often do with a client who comes to you with issues that they might not tell you about completely.

BARUCH: John? Can I ask you about Webmaster Tools?


BARUCH: So for instance, let's say there's a product, a brand that has sales figures, right? And then, at the same time, you have another speaker that's called blah, blah, blah, but they don't want to add it to the website because it's a whole different product that people know it is like a different website. Is it OK to add it in the same Webmaster Tools account for the speaker brand?

JOHN MUELLER: Sure. I don't see any problem with that. The Webmaster Tools is essentially mostly a technical tool that makes it easy for you to spot these issues. It's not something where we'd say, you have to sort them out by logical groups, or anything like that. So that's essentially a tool for you to use. And how you use that, how you split those sites into different accounts is up to you.

BARUCH: OK. Now is this for the data to flow, because different people want their business places separated?

JOHN MUELLER: That's essentially up to you. You can do that either way.

BARUCH: Thank you.

JOSH: Hey, John?


JOSH: I was just wondering. Most people think that Panda is strictly an on-site algorithm, that, if you're hit by Panda, all you have to worry about is your own site. I was wondering how much do we have to worry about bad links or bad references in relation to Google's quality algorithms. And if we have to worry about that, will a disavow file help? Or should we be deleting?

JOHN MUELLER: Panda is essentially a quality algorithm that looks at the content on your site. So it's not something where we take links into account at all. So that's not something where I'd even worry about the disavow file there. Usually, we also try to keep these algorithms separate, in the sense that they're not interacting. So if, for instance, Panda were to take some of the same signals as Penguin, then that would make it harder to debug on our side what exactly is happening with the website with specific signals. So as much as possible, we try to separate these, so that you're really only looking at effects from one specific algorithm. So that's not something where you'd need to disavow. If you see problematic links, of course, you can clean that up. But it's not something where the Panda or the quality algorithm would be looking at these links.

JOSH: Right. It's not going to happen with Panda. What about internal links? Wasn't it a tactic in 2008 or so that people would go into their blog and they would link every exact match query to an internal page, because they thought that would somehow boost their ranking? Or maybe it did back then. Is this a quality issue that we should be looking at?

JOHN MUELLER: By itself, it's not a matter of the page rank being passed or specific anchors being used. But the way you internally link your website can, of course, affect the quality of your website as perceived. So if every second word on your website is underlined and links to a specific topic or a tag page, then that's something that makes it really hard to read this page and to understand it appropriately. And it makes it look kind of like a low-quality or almost spammy type page. So that's something where it's not the matter that these links are there, or that they're passing page rank or not. It used to be that people would no-follow all of the internal links to hoard the page rank on specific pages. So that's not so much of a problem from a technical side. But from a practical side, it can make the page look a little bit lower quality or higher quality, depending on how the internal structure is set up. And the same applies to the template of the site, to the way that you design your website. All of these things add up and, in the end, can give the impression that this is a really high-quality site, or it looks kind of like low-quality. The webmaster didn't really care about this website. They just installed a bunch of plug-ins to link every other word, those kind of things.

JOSH: Right. OK, thanks, John, very much.

JOHN MUELLER: All right. Site is a gov site. Went from position 2 to 21 overnight, and then stayed there for over six months. See zero movement up or down. The site that replaced it is, but spam. No webmaster penalty and other penalty. What would you do? Should I give you a site? It'd be interesting to look at the site, yes. I mean, there can be lots of different reasons for changes like this where, maybe, the overall quality of the website looks kind of bad and is more problematic. Even if it's an official government site doesn't necessarily make it the best website of its kind, or the best website to return for these kind of queries. So I wouldn't focus so much on the fact that it's a government website, but rather just look at it like you would any other website and see what you can do to make sure that everything is working the best that it can work there. So that's something where I'd take a step back and look at the site overall, rather than focus on specifics like this is a gov site, and the other one is kind of spammy. But rather focus on your website and see what you can do to improve it there. If a site has too many duplicate pages in the millions, moving those pages into subdomains based on category and blocking them helps SEO, for example, blocked, allowed with original content. Does this still hurt the main domain? Essentially, I'm not sure what you mean about duplicate pages. If they're really duplicate in the sense that they're one-to-one copies-- and we'll generally figure that out and just ignore those pages-- if these are essentially pages that are auto-generated and look almost exactly the same, maybe search results pages, maybe aggregated content, then that's something that will look bad, regardless of where you put it and regardless of whether or not you have them blocked by the robots.txt. So of course, content behind robots.txt we wouldn't be able to crawl and index like that. We'd index it by the URL. But it's still kind of looks bad. So what I'd do there, if it's really a case that these pages are low-quality aggregated content, is try to null index them as much as possible, so that they really fall out completely. Maybe even remove them completely from your website, so that even users don't run across them. Because if you think that these are kind of low-quality duplicate pages, then maybe your users will feel the same way. With regards to just moving them to a subdomain, our quality algorithm does look at the site overall. And if it looks like these subdomains are essentially part of the main website, we'll treat the whole thing as one website. And if it looks like the overall quality of your whole website is fairly low, then we'll treat the pages on your site appropriately. So that's not something where I'd just shuffle those URLs around on the website. I'd really try to find a more permanent solution for this kind of low-quality content, assuming it is low-quality content that's just aggregated and looks very similar, very duplicative. If this is the content that is essentially really one-to-one copies, then that's something maybe it's worth blocking them in the robots.txt. Maybe it's worth really finding a way to prevent these duplicative URLs from even being discovered.

JOHN MUELLER: Any way to download search query data via Google Spreadsheets? I think we recently had a problem there, that the download via Spreadsheets was stuck, or that it basically just did a download in the browser. But essentially, if you download that data as a CSV file, you can also just drag and drop it into Google Drive. And then you'll have it as a Google Spreadsheet file again. So that's something you could do there. Since you mentioned API, one thing that we do have is a Python script that downloads the search query data, I believe, as well, what you could use to automate this process. So if you wanted to pick up your search query data on a daily basis or on a weekly basis, then that Python script should work. That Python script, I believe, is on one of our older blog posts, maybe, two or three years back. I have a .NET site with a WordPress blog in a subdirectory with PHP. Does having these two sites in different coding languages under one domain affect the ranking of my website? The WordPress site, for example, lives under the URL No. Having different programming languages on your site isn't something that would affect your site's ranking. That's not something we'd even take into account. So it's not the case that we'd say, oh, .NET is good, or .NET is bad, and we'll rank it higher or lower. If you have a mixture, it'll be bad. Essentially, this is a technical issue up to you. You can create your site in whatever method you want. And provided we can crawl and index those pages separately, that we can follow the links through your site, then we'll try to index it like that. It doesn't really matter what the technical infrastructure behind that it is. One thing to keep in mind, though, is, if you're moving from something like static HTML pages to a PHP site, if you change the URLs of your site to go from .htm to .php, then we'll see those as separate URLs. So you should set up a 301 redirect for those URLs. If you can keep the same URLs within your website, then that makes it a lot easier. But if you just have different parts of your site on different CMSes, then that's up to you. That's perfectly fine. [INAUDIBLE] to the .gov's website question. The question is more if there can be a manual placement to my website at 21, as it's been there and not moved for all the months. No, we don't have any kind of a manual penalty that move sites to a specific location. And if there were a manual penalty in place, then you would see that in the Webmaster Tools, definitely. So if you don't see anything in the Webspam Section in Webmaster Tools, then your site is, essentially, just ranking naturally at that position. And it's not something that's unnaturally being held back or tied to that specific ranking. How soon is a domain that appears in Webmaster Tools used as a ranking signal? Essentially, at the same time. So if we see a new domain that's, maybe, linking to a website, or a new domain that we pick up for crawling and indexing, as soon as we index it, we'll forward that information to the Webmaster Tools team. And they'll use that in their data. So essentially, the web search side happens almost a little bit earlier, before you see that in Webmaster Tools. Sometimes that's a day ahead of time. Sometimes the Webmaster Tools data is just updated every couple of days, or once a week, so you might see a small delay there. But essentially, as soon as it's visible in Webmaster Tools, it will at least have been seen by the web search site as well and can potentially be used in ranking there. Three days after Googlebot requested many 410 Gone resources, some of them are still indexed. How many days does Google need to de-index the resources after discovery of their 410 status? We don't actually have any fixed timeline for that. And sometimes, it can take a little bit longer. Sometimes, it's a little bit faster. Essentially, the main difference we have there is, with a 404 versus a 410, a 404 is just not found. A 404, we'll generally treat as being more of a temporary error, traditionally. So we've seen a lot of pages go to a 404 result code, and then flip back to a normal 200 result code showing content again. So we kind of have been a bit more cautious with the 404s in that we'll re-crawl the pages, maybe, a few more times to actually confirm that it's really a 404, and then drop it out of the index. Whereas, with a 410, we'll probably just crawl them once, or a maximum of twice, and say, OK, this is really gone. We'll remove it from the index. But the process, after we've confirmed that the result code is like that, depends a little bit. And it can definitely take a few days for that to be visible in search, or rather to be removed from search.

BARRY: And the 418 is more like a 404? Or a 410?

JOHN MUELLER: The 418, what's that? The teapot, right?

BARRY: Yeah.

JOHN MUELLER: We treat that as a kind of a generic 400 error, so we treat that more as a 404. So if your page goes from having content to showing a teapot, then it might take a few re-crawls for us to actually drop it out.

JOSH: That only indicates that the website in question is both short and stout.

JOHN MUELLER: Yeah. Yep. It's kind of a special code. I think the thing to keep in mind, if you're setting up a teapot page like this, if people are linking to that page, then we'll be dropping those links, because we don't have anything to index there. So it's a kind of a weird situation in that you can't really build up page rank for your website with a page like that.

BARRY: Will Google consider that, if you don't do a 418? Would Google consider that a soft 418?

JOHN MUELLER: I don't think we have any algorithm specific for that at the moment. But if we see this happening a lot on the web, who knows?

BARUCH: But you guys ignore that code, right?

JOHN MUELLER: Yeah. We treat it as a 404, the 418.


ROBB: And John, if we've used the Remove URL Tool within Webmaster Tools to remove outdated content and the request comes back and said, yes, we recognize this page, we can see that it's no longer there, we've removed it, how often do you need to remove things before they're properly gone? Because I've been removing pages that we believe you're seeing as duplicate on another site for every two weeks since the beginning of June. We're now on about the sixth time we've removed them. And every time we search for them, they're still coming up.

JOHN MUELLER: I probably need to take a look at that, what specifically you're doing there. So it sounds like it's not on your site, but rather on someone else's site.

ROBB: Well, it was another site that we owned that we thought might be part of our other bigger issue. So we actually closed the whole thing down, whole site just 410. We went as far as moving the domain to another registrar or with someone else, just distancing ourselves from the whole thing. But the site, when we do site code on that site, still comes back with 20, 30 results, all of which are URLs that we've definitely removed half a dozen times. And it seems to just be like a cockroach that won't die.

JOHN MUELLER: I mean, if this is still your website in the sense that you it in Webmaster Tools, you can do a site removal request in Webmaster Tools. And that'll be honored, I believe, for six months.

ROBB: Mm. Well, it's in there, but it's not verified, because the whole thing is gone. So as soon as it returns a 410 and it has no Webmaster Tools code snippet in it, it's no longer verified within Webmaster Tools.

JOHN MUELLER: What you could do there is verify it with the DNS method, for example.


JOHN MUELLER: That's one way you can verify the website without putting any content on there. And once you have it verified, you can do a site removal request. And that's something that will take out all of the URLs within, approximately, a day. And that'll be in effect for about six months. So during that time, we should have been able to re-crawl everything and to reprocess everything, remove those URLs completely. But if you're just doing this on a per URL basis, it's sometimes a bit tricky in the sense that the URLs have to match exactly. And if you submit one URL that's currently shown, it's possible that we have another copy of the packet on a different URL elsewhere that we bubble up in that place after your removal request. So that's where you're playing Whac-A-Mole, almost, if you're taking out individual URLs. But if you do a site removal request for a site that you have verified, then that will definitely take everything out within about a day.

ROBB: OK, I'll try that. Thanks.

JOSHUA: Hey, John?


JOSHUA: It's been a couple of months since the authorship images were removed from search results. Is there any surprises about how those results went? Or did it just seem as expected? And also, I was wondering why leave in authorship images on the Google+ posts, but not keep them in on the other? So in personalized search, you still see authorship images for relevant people, and then also some page images in personalized.

JOHN MUELLER: Yeah. I know the Authorship team has been working on that to simplify things and clean things up there. I don't know what the next step will be there. So that's something where they're aware of this discrepancy there and the issues around that. And I don't know how much data we'd want to show there and where the next steps will be taking us there, but I know they're looking into things that could help there.

JOSHUA: Yeah. And there's a couple of other things where you've got-- it shows in personalized search results so-and-so plussed this, or so-and-so has shared this content. Any ideas on whether we'll be seeing more stuff like that related to Google+ in the search? Or is that going the other direction?

JOHN MUELLER: I don't have anything specific to announce at this point. It's always kind of tricky with these things, where you guys would love to see which direction we're heading, what we're doing as a next step. But we try to take these things and announce them when they're ready and when they're actually live. So I'm sure you'll see when the next steps happen, but I don't have anything specific to announce at the moment.

JOSHUA: Alrighty. Thanks.

JOHN MUELLER: And I know Barry is trying to read my face now, but sorry.

JOSHUA: Yeah, John. I noticed that they did catch up with those. Because I mentioned to you a couple of months back about there was still some I follow links in the Google+ posts that a couple of types of the hyperlinks were still follow. And before, some of them became no-follow, but it was kind of a gradual process. And it seems to be quite well-covered now in that all hyperlinks across the board are no-follow.


JOSHUA: I was wondering, OK, the no-follow, when webmasters use it, it gives a message to the search index, these are not links that-- you know, they're either paid, or they're not links that we feel we want to pass authority to. But if Google uses it in the Google+ system, what is the message that that is to give, in general? I mean, because Google can have those or ignore them. And so I was wondering. It appears that the implicit message is that we're not going to give authority to any links. I mean, is it mainly a concern about spamming and things like that, where if we just use no-follow across the board, then it's on the safe side, you know, like Wikipedia and a lot of big sites have done?

JOHN MUELLER: It's mostly a matter of trying to treat user-generated content appropriately, in the sense that we have this platform for you guys to use, for you to share content, share links, those kind of things. But at the same time, we need to take care that those links aren't passing page rank to sites that are essentially spammy or problematic. And when you write a Google+ post, and you're talking about a business, for example, it's really hard for us to tell are you talking about this in a good way or bad way. Or are you saying look at the spammer here? So essentially, we're trying to be on the safe side there and just treating those links within the user-generated content as links that we don't really know how they should be treated. So that's kind of in the sense our recommendations for websites in general that have user-generated content where, if you can't be completely sure about these things links, maybe it's best to just no-follow them, so that if someone were to try to take advantage of that system, at least those links wouldn't be passing page rank to any kind of problematic sites.

JOSHUA: OK. But there's also been the talked about frequently in the past about the we'd like to give more relevance or authority to the persons that are well-known, relevant, and authoritative persons. So those seem to be going in the same direction with each other.

JOHN MUELLER: It's hard to find a balance there. It's kind of tricky, sometimes. And this is something where, if you have comments on your blog, you probably have the same problem in the sense that, theoretically, you could be going through and saying, oh, I know this person. He's a really smart guy. I'll make sure that his links are always followed. Or maybe this other person is an expert in her field. And I'll make sure that her links are always followed. And everyone else just has a no-follow. And it's really hard to do that, even on a fairly small scale. When we're looking at something the size of Google+, it's even harder to go through there and say, well, these links have always been good, or 99% of these links have been good and useful for us. We'd like to pass page rank for all of these links from this guy that they're posting. So it's something I know the Google+ team is trying to find a balance there. And I imagine you'll continue seeing things sometimes passing page rank, sometimes not passing page rank, and kind of fluctuating there, and trying to find the right balance.

JOSHUA: All right. Thanks. That's good to hear.

JOHN MUELLER: Let's take a few more from the Q&A. And if we have enough time, we'll open things up. I'm looking to perform a full site redesign. If I keep the URLs and the content on my pages the same, should I experience any issues with rankings? My concern is I undergo a site redesign, I lose important rankings. If you significantly change your site, it's very likely that you'll see some kind of fluctuations within your site's rankings. So they could go up. They could go down. It's something that's impossible to say in advance. There are things like the internal linking structure, which might be changing for a website, which could affect how we treat those pages. There are other aspects, like recognizing which part of the page is boilerplate and ignoring that which might cause fluctuations in the short run, and yet other aspects around the quality of the site, overall, where maybe, if you do a complete redesign, we'll think, oh, this looks a lot better than it looked before. And we'll treat it appropriately in the search results. So these are all aspects that come into play. It's definitely not the case that you can be guaranteed that everything will remain the same, as long as the URL and the primary content doesn't change. So you're very likely to see fluctuations. If you do a good redesign, then you're probably likely to see fluctuations in a positive way. So that's something where you should expect some changes. All right. Let's just open it up for questions from you guys, since we just have two minutes left.

ROBB: John, do you have any news for me?

JOHN MUELLER: Nothing specific, no. But I know the engineers are working on those specific issues there.

ROBB: Before Christmas, maybe?

JOHN MUELLER: I always prefer if things change faster. Yeah.

BARUCH: [INAUDIBLE], will come.

ROBB: It's been 15 months.

JOHN MUELLER: Yeah. That's frustrating.

BARRY: And expensive.

BARUCH: I was just thinking. I know you can't tell us when Penguin's coming or Panda and all that. And Panda's an internal algorithm. But if Google lets us know this, then the webmasters can all prepare to make things better. And things can be better. This way, everybody's notified. I mean, one day, will that ever happen where, we'll announce when Penguin's coming. So get ready. Do what you've got to do. And if you missed it, then you're out. This way, it's fair for everybody. I'm just suggesting it, you know?

JOHN MUELLER: Yeah. It's always tricky with algorithms. I'd prefer, sometimes, if we could just give that information to all webmasters so that they can respond to that feedback that our algorithms are giving. But it's sometimes tricky in the sense that these algorithms aren't one-to-one mapping to anything specific on your pages. So it's not something where we could say, for example, Penguin algorithm thinks your website isn't so hot, that there's anything specific on your pages that you could change, or anything specifically we could point at, or even anything specific where we'd say, our algorithms think your site is kind of overall lower-quality.

BARUCH: I'm just saying this so nobody points fingers at you guys, you know, for the long run.


ROBB: I don't think that's ever going to not happen.

JOHN MUELLER: Yeah. I think we might see some changes in that regard, but I'm not 100% certain. And I can't guarantee anything there.

KEVIN: Cue Josh, comment about PR. Go Josh.

JOHN MUELLER: I have to run. There's someone waiting for the room here. [LAUGHTER] I'm very sorry. But I'm sure you'll be in one of the future Hangouts, and we can follow-up on that.

JOSH: Oh, I'll be there.


JOSH: Thanks a lot.

KEVIN: Thanks, John. Thanks.

JOHN MUELLER: And hope to see you guys again in one of the future Hangouts.

JOSH: Thanks, John. Have a good day.

JOHN MUELLER: You too. Bye.

JOSHUA: Thanks, John. Good show.

JOSH: What comment about PR? What was I supposed to say? Oh, you left. Damn. | Copyright 2019