Reconsideration Requests
Show Video

Google+ Hangouts - Office Hours - 24 October 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: All right. Welcome, everyone, to today's Google Webmaster Central office hours hangout. My name is John Mueller. I am a Webmaster Trends Analysts at Google in Switzerland, and I try to connect webmasters like you all with our engineers, and make sure that information flows in both directions, where possible. We have a bunch of people here already. When people drop out, you're welcome to try to jump in. We have the link on the event listing. There are a bunch of questions already submitted. But as usual, if one of you wants to get started, feel free to jump on in. No more questions.

AUDIENCE: I have actually one question, John. Recently, you recommended that we use RSS feeds together with sitemaps. And that we include in the RSS feed any pages that have material changes. I was wondering if you could provide a little bit more context around this. So for example, if I have a product page, if there is a new user review, is that a material context to include that page into the RSS feed?

JOHN: You could do that. Sure. Now, I think that's something where, if the content significantly changes, that's relevant. If the content, the primary content, doesn't change, then I wouldn't update it. So for example, if you have a sidebar with recent news articles and your primary content stays the same, then I wouldn't update the date. But if there are new comments added to the post, if there's new information added to that page, then I'd definitely update the date so that we can re-crawl and re-index that page again.

AUDIENCE: We definitely have the date on the sitemap, but would you also put it in the RSS feed?

JOHN: Sure. So for us--

AUDIENCE: How big of an RSS feed do you recommend we have? Like, the last 10 things, or the last 100?

JOHN: It depends on how quickly we crawl your site, and how quickly you update your site. So I'd take a look, for example, at how frequently we're currently polling your RSS feed, and try to match the number of items, so that it matches that approximately. If we're currently looking at it, maybe once a day, then make sure that you have at least one day's worth of updates in that RSS feed.

AUDIENCE: Just a quick question more on, kind of, crawling sites, that sort of thing. We've seen recently that one of the sites since the Penguin refresh kind of came out-- I don't think it's really linked to that, but they've seen kind of a drop-off in their rankings. But we've seen a massive increase in the amount of times that that website has actually been crawled in Webmaster Tools. Is there normally a sign, say, for instance, if your website is getting crawled a lot more, is there normally something that potentially could be wrong with it that's causing it? It's just weird that it's kind of coinciding when it's dropped off, in terms of the rankings.

JOHN: I wouldn't assume that there's a direct relationship between the crawl rate that we have and, kind of, how we view that site from a quality point of view. So what sometimes happens is, we crawl pages a lot that we don't actually show a lot in Search results, just because we think there are lots of changes here, and we want to make sure that we keep up with those changes, so that when we do find something that we really want to recommend and search, we'll have that as quickly as possible. So it's not that there is any kind of relationship between how we would show a site in Search and how we would crawl it. There's one exception, which is, if it's really blackout spam, where we can recognize that it doesn't make any sense at all for us to crawl this site, then that's something we might say, OK, we're not going to crawl it. But apart from that, pretty much everything, even normal manual actions, will still result in us crawling normally, as we would just when we see content updating. So the crawling part is kind of more a technical aspect, and the quality aspect, how we think that this site is relevant for rankings, is something that we do separately. So independently of the crawling.

AUDIENCE: OK. So kind of a spammy site would actually see the crawls drop off more so than actually increase, because you don't see the need to crawl it on a regular basis because it is kind of blackout, that sort of thing?

JOHN: This is something that we do only in really extreme cases. So it's not something that, if you have a few affiliate links or if you have a little something, some links that you forgot to put the nofollow on, that we'd stop crawling, or that we slow down crawling. It's really if this is just pure blackout spam where we'd say this is gibberish, it's automatically generated, really, there's absolutely no need for us to even look at this-- then we'd stop crawling that. But apart from that extreme, pretty much everything else would get crawled normally.

AUDIENCE: OK. No, that's perfect. Thank you.

JOHN: All right. Let's take a look at some of these questions here. "Is each algorithm, like Penguin and Panda, an iteration of their previous release? So a website would not be affected by Panda 2.0, 2.1, 2.3, et cetera, they would just be affected by the current Panda. Or does each iteration focus on different things?" Essentially, these updates are independent of each other. So it's not something where we'd say, if your website was previously affected, then it will always be affected. Or if it wasn't previously affected, it will never be affected. It's, essentially, this new update that we might do at one point is the current status, and we turned the old stuff off. So each of these are essentially independent. And sometimes we do change the criteria. Sometimes we do change the algorithms that compile this data. So that's something where one website might be exactly the same, but at a later point, our algorithms look at it and say, based on my new understanding of how we can recognize quality, this website might be a lot better, or it might be worse than it was in the past. So that can change over time. "Are e-commerce sites treated differently regarding content? My site has a lot of boilerplate text, plus only slight variations in product-- means similar descriptions, so pages aren't that unique. Would there be some kind of a site-wide penalty?" We don't have a duplicate content penalty. It's not that we would demote a site for having a lot of duplicate content. Sometimes, that's essentially just how it works. I have a slide from one of the last sessions, where we had something similar. Let's see if I can find it. So this is-- let me just pin this. This is a very colorful view of web pages, where these are essentially different web pages that have the same block of text. This is symbolized by the green color here. This means that, for instance, the same product is on different websites, or the same product description is used across the same website multiple times. We understand that this is the same content block. We understand that there is unique content on the top and the bottom here, which means that these pages are unique. There's value in keeping them separately. What happens in Search is, if someone searches for something that's in this, kind of, generic block that's shared across these pages, we'll just pick one of these pages, and show that in Search. So this is, essentially, one of these three pages. It kind of depends on which one we would find more relevant, or maybe personalisation comes into play a bit. But we'd just pick one of these three pages to show. Whereas, if you searched for part of the generic text as well as something unique to one of these pages-- so in this case, something from the red page-- then we'd try to pick that red page. It's not that we'd say this green text is duplicate, therefore the page is bad. But we'd say, this has unique aspects of this content that mean that it's relevant, depending on how the user is searching. So it's not that we would penalize a site for having this kind of duplicate content within the site, or across different sites. But it's just the case that we'd try to filter out the duplicates and try to just show the most relevant one. And that's, essentially, not a sign that anything is broken. It's not something that you would need to manually tweak, or fix in any way. It's essentially just how the web works. And we understand that there's a lot of duplicate content out there. And we try to make that work. So it's more a technical problem on our side than anything you would need to do on your side. "Is the Penguin now on an automatic monthly roll out? If not, is the next one coming soon, or a long time away? Is this an algorithm update or just a data refresh?" This is, essentially, an update that's currently rolling out over a couple of weeks. So I wouldn't assume that the current state is the final state. We're still kind of rolling that out, step by step. And the goal is to have it update a little bit faster, but I can't pinpoint any specific update frequency where I'd say this happens every month on the third Monday, or something like that. We're working out how we can roll this out, and how we can make this a little bit of faster update cycle. So the goal is to make this process a little bit faster, a little bit easier for those who are actually working on the websites to fix these issues. But I don't have anything that I can guarantee at the moment, and say, it'll be happening on these dates. I think you'll see it going forward, as things update. People tend to notice this and talk about it a lot. "Using exact match anchor text like 'Boots - London' in navigation or content in headers on a site can be bad for SEO when a site has unique content and keyword density is natural?" No, not necessarily. Essentially, try to use the anchor text within your site that makes sense for your users-- that they understand. And if that's an exact match anchor text, that's fine. If it's something that has more descriptive anchor text, that's fine. If you have different anchor texts depending on how you navigate to that element, for instance, if you have a product and a category page, and it's also shown in the News section on the sidebar, then those probably have different anchor texts. And that's fine, too. So it's not the case that you would need to artificially modify your site to be optimizing anchor text within your pages. And keyword density, in general, is something I wouldn't focus on. Make sure your content is written in a natural way. Humans, when they view your website, they're not going to count the number of occurrences of each individual word. And search engines have kind of moved on from there over the years as well. So they're not going to be swayed by someone who just has the same keyword on their page 20 times because they think that this, kind of, helps search engines understand what this page is about. Essentially, we look at the content. We try to understand it, as we would with normal text. And if we see that things like keyword stuffing are happening on a page, then we'll try to ignore that, and just focus on the rest of the page that we can find.

AUDIENCE: Hi, John. Can I ask a question?

JOHN: Sure.

AUDIENCE: I just wanted to confirm that-- we have a couple of keywords which come in the ranking, maybe on second and third page on one day, and on another day they got [INAUDIBLE]. So what could be the possible reason behind it? I mean, one time they are on the first, second-- or sometimes even first page. And again, on the next day, they are nowhere in the search results. So what could be the possible reasons for that?

JOHN: It's hard to say. One aspect that I think always comes into play is things like personalizations, geo targeting, those kind of elements, where some users might search for your website and they'll find it on the first page because it's very relevant for them. But other users, maybe in other locations, they search for your website and it's not so relevant for them, so it shows up on maybe page three or page four, even. So that's one aspect that always comes into play. I think, kind of, moving away from the idea that rankings are always going to be the same regardless of how you look at your website, that's something that a lot of people still need to do. So there's a lot of personalization in ranking. There's a lot of geospecific information in rankings. So different people looking at different times from different locations are bound to have different rankings that they see. And that's a part of the reason why we focus on the average composition in Webmaster Tools, so that you see what people actually saw, and what the average was there. And if you want to drill down, maybe you just select a smaller time frame to look at that specifically. But it's not something where I'd focus on one specific ranking number, and say, oh, I have to aim for ranking number three and then everything will be good, because nobody will always see ranking number three for your website for those queries. It's always fluctuating.

AUDIENCE: That is fine. But let's say we are measuring in some tool, and we are looking on with the average ranking. So some keywords, let's say they are in between 20 to 30. We can understand that it can move [INAUDIBLE], so on and so forth. But one keyword is coming within the 20 and 30 one day, and next days it is going out of hundreds. So it is something which is very, very variable. And so what could be the possibilities?

JOHN: That's possible, too. And there are lots of different reasons why that might happen. There could be things like, we are not sure if this is something that is very new, for example. If this is a page on a topic that suddenly comes up in the news, and everyone is searching for it, and we think your page has a lot of useful information on that, we might show it a little bit higher. But it's not something where I would say there is a technical problem on your side that you need to solve there. That's essentially how our ranking algorithms work. And if there were a technical problem, you would definitely see that in Webmaster Tools. And you would more likely see that your page disappears completely from Search, if we recognize, like, a nofollow-- a noindex, or a 404, or if your server doesn't respond at all-- we wouldn't be fluctuating that up and down. We'd have it either ranking, or not ranking at all. So it's not that there's a technical problem on your side. It's not that the content on your page is in any way bad, or not really visible. Usually, that's just a sign that our algorithms are fluctuating with the rankings.

AUDIENCE: OK. Thank you.

JOHN: And sometimes, it's a sign that you're kind of on the good side. Because if your site keeps coming into the first page of the Search results, then our algorithms are thinking, well, maybe this is really relevant for the first page, or maybe not. They're kind of right on the edge. So that means, sometimes you just need a little bit of a push to make your website a little bit better, so that the algorithms can clearly say, yes, this really belongs on the first page. I don't have to, kind of, reconsider it from time to time.

AUDIENCE: OK. Thank you.

JOHN: Sure.

AUDIENCE: Hey, I've got a question. Can I ask you a question?

JOHN: Sure.

AUDIENCE: Well, we received a penalty on May 15, 2012, for a content penalty for our forums, known as a technically-- forums are technically not [INAUDIBLE]. OK, after almost one and 1/2 year, the penalty was lifted. It was automatic-- the penalty was lifted automatically. And almost now, it's been one year. We cannot see any changes in the site, like, it was a site-wide penalty to technically [INAUDIBLE] forums. We were one of the last [INAUDIBLE] logistic technology forums. So as of now, we cannot see any updates. We have even uploaded a file on this [INAUDIBLE], but there's nothing-- we cannot see any changes in the traffic. I don't know what's wrong. We really require your help to, like, you know, make it understand how we can gain the traffic again, because I somehow feel that, OK, we are still in the sandbox with Google. I'm not really sure, like, do we have a sandbox, still? Things like this. Because I can see the traffic pattern on the site. It goes high, again it comes down. It will be back-- for a week it will be up, and the next week it will be the same. Like where I was previously.

JOHN: So if the manual action is no longer in Webmaster Tools, then there is nothing manual holding your site back. Then that's really just the way our algorithms are looking at your site. And sometimes it happens that a manual action is lifted, but, actually, the site continues to rank where it used to, kind of, during that manual action, just because the algorithms have changed over the years. When it comes to forums, I think it's something where it's really hard to work on the quality of the content in general sometimes, where you have a lot of user-generated content. And if all of this content is always indexed and searched, then Google sometimes has difficulties understanding how high-quality the website is overall. I don't know your website in specific, so I am just generalizing. For instance, we would see a lot of low-quality posts in a forum. We would index those low-quality pages. And we'd also see a lot of really high-quality posts, with good discussions, good information on those pages. And our algorithms would be kind of stuck in a situation with, well, there's a lot of low-quality content here, but there's also a lot of high-quality content here. So how should we evaluate the site overall? And usually, what happens is, our algorithms kind of find some middle ground, and say, well, overall, it's kind of OK, but not really great. And what you'd need to do to, kind of, move a step forward, is really try to find a way to analyze the quality of your content, and to make sure that the high-quality content is indexed and that the lower-quality content doesn't get indexed by default. So things you could do, for instance, is new posts by new posters who haven't been in the forum before. Maybe they're noindexed by default. So people can find them if they search within your website, but they don't get indexed in Search. Or, threads that don't have any answers, or threads that don't have any authoritative answers, maybe noindex those by default. And, kind of, look at your site, in whatever aspects that you think are relevant for your website, and try to come up with a way to automatically recognize which pages are high quality, which pages are lower quality, so that the pages that do get indexed are really the high quality ones. And again, I don't know your site specifically, so I can't say how far that is really relevant in your case. But especially with forums, especially with user-generated content sites, that's something that I'd really try to focus on as much as possible.

AUDIENCE: OK. And one more thing. Last time when we had a word, we got in [INAUDIBLE] forums. You said you're going to take a look on it, but I'm still waiting for your update. You said you will help me out to, to get back-- bring back my rankings. Because previously, I think, before a penalty, I used to have nearly 6 million unique visitors on my site. And as of now, today, it's a half year, I think I could say 600K. 600 unique visitors-- 600K unique visitors to my site. It's almost a 90% drop on my traffic. [INTERPOSING VOICES]

JOHN: I just see that there is really a strong aspect from the quality side of things with your site at the moment. And that's something that you can work on with your site specifically, to try to understand where the high quality parts are of your site, and how you can recognize those automatically. And I think with forums, that's something that's sometimes easier to do. But it's really a strong aspect of how we're viewing your site at the moment. Really from the quality point of view.

AUDIENCE: OK. And one more thing I've noticed in the Search results. It's like, when we post anything new on our website, our content get indexed in next couple of minutes, I could say. And when you search the same article on Google, it's showing me it was indexed one day ago. I think it's some sort of a bug with Google, I'm not very sure. Because if the content is posted in the next one, it's showing me-- when I'm searching the same article on Google, it's showing me that it was indexed one day ago. I don't know what kind of an issue is this. Basically, figuring out a lot of it, Google, like, you know, so I've just noticed a couple of things with Google, like-- what's happening?

JOHN: That sounds like some kind of a time zone calculation problem. I don't know.

AUDIENCE: Time zone?

JOHN: I don't know. You have the dates on your articles as well, right?


JOHN: On your posts and everything?

AUDIENCE: Yes. Absolutely, yes.

JOHN: But let me check with the engineering teams on that. But I think that the date itself is, essentially, just something that we show in the Search results. It's not something that we'd say, this is a primary ranking factor. But if you can maybe post a query in the Search results, or a screen shot of something, or send me that, then I can definitely take that to the Dates Team to figure out.

AUDIENCE: Can I get your email address?

JOHN: You can just send it to my Google+ account.

AUDIENCE: Google+ account?

JOHN: Yeah.

AUDIENCE: OK. I'll do that.


AUDIENCE: Yeah, thank you.

AUDIENCE: Hey, John. Can I ask a question?

JOHN: Sure.

AUDIENCE: So I just sent the URL here on the chart, and I'll send it one more. So essentially, these two URLs is-- one is for student credit cards, the other one is for Visa student credit cards. And both of those terms are things that people search for. In the past, we would even have category pages that were more refined. Like, for example, low-interest credit cards for people with excellent credit. So like a three combo, if you will. If you look at the Search results on Google for Visa student credit cards, with the exception of the first result, everything else seems like Google is not understanding that the person is searching for Visa student cards. They're searching for-- they think it's just student cards. Now, having said that, the overlap between products between the two pages is very significant, because Visa is the biggest credit card network than MasterCard. And for the student space, American Express and Discover hardly have any products. So my point is that those two pages have a significant overlap in terms of the results. From a user standpoint, I strongly believe that it's much better to send them-- if they're searching for Visa student cards, to send them to a page that has only Visa student cards, and not MasterCard cards. But I'm wondering whether this is an example that may be tripping Panda, and says, oh, yeah, the content is looking very duplicative here. Because we have 130 category pages in total.

JOHN: OK. And the texts are essentially very similar across these category pages?

AUDIENCE: No. No, the text is custom.

JOHN: Credit card types?

AUDIENCE: Yeah, the text is custom written, but you have the filters, and a lot of overlap on the results.

JOHN: OK. I'd have to take a look at this. I don't know. In general, like I mentioned before, it's not that having overlap from the category text or the product text would be a problem, as long as we can really recognize what specific aspects you're focusing on, on these pages. And I don't know if we'd recognize, like, the Visa versus other types of credit cards aspect here strongly enough, but I can definitely take a look at that with the engineers here to see that we're picking it up correctly.

AUDIENCE: What would be helpful is to give some advice of whether we should try to tailor pages very closely to the user intent.

JOHN: Yeah.

AUDIENCE: When the only differentiator is the Visa, the algorithms are still not strong enough to pick that small-- because it's a small differentiator, but a major one, in terms of user intent. And maybe we end up getting penalized for trying to tailor so closely.

JOHN: No. You don't get penalized for having this kind of duplicate content. So I wouldn't worry about that aspect. So I think what I would just try to focus on is whether or not this page would be able to stand by itself-- these different variations of these pages would stand by themselves-- or if you're essentially focusing on one very, very specific aspect that people aren't necessarily specifically looking for. Or where you're essentially just providing, like, a combination of aspects that, let's say, doesn't actually provide any real unique value on those page. So I don't know those pages specifically. But for instance, if you had, let's say, shoes, or something like that, then you'd probably want to create categories for the different types of shoes. Maybe product pages for the individual models of shoes. But you wouldn't create individual pages for every size and color variation. You would say, well, this shoe is also available in these five colors, or in these seven sizes. So that's something you would kind of mix into the specific product page, but you wouldn't create separate pages for that. I don't know how far that maps to the credit card area. That's not really one where I'd know in detail how much people are actually looking for this specific niche, or if this is something where you just say, well, this keyword is searched for 1,000 times, and we'd like to show up for all of them. It's kind of like trying to find the balance between having real value on these pages and still targeting something specific.

AUDIENCE: So they definitely have search volume, those category pages. Where would you draw the line? At 500? Or is there a number where you would say it's worth having its own page, otherwise people can-- because you can obviously go to the student page and then use the filters and isolate only the Visa cards.

JOHN: I don't think there's any hard line, where we'd say, this is the number that you need to focus on. So that's not really something I can help with there. I think what you're always balancing is having really, really strong pages for these products, versus having, kind of, medium strength pages for a lot of different products. There are situations where people might find a more generic page a little bit more useful. And if you had one kind of really, really strong generic page, you could probably get just as much traffic as if you had hundreds of variations of pages that each get a little tiny amount of traffic. So that's something where you have to balance having something really strong and generic versus having something that's seen as kind of lightweight and kind of spread out across a lot of variations. And I think, looking at these pages, it's probably somewhere on the line where it could make sense, but maybe it doesn't make sense that much. It really depends on, I guess, how strong your website is in general in these areas. So I wouldn't say this is really something you should absolutely not do, in your case. But I would also, kind of, shy away from saying this is something that you have to refine even more, or where these refinements as they are now is the absolute best thing that you need to do. I think this is something where you might even try to do something like AB testing, and think about which aspects really work well, and which aspects are essentially spreading it too thin.

AUDIENCE: But you wouldn't look at these areas as areas that may bring a penalty?


AUDIENCE: Or maybe hit by Panda?

JOHN: I don't think that this is something that we'd say would bring a penalty. I think if you were to spread it out even more, then I could see the Web Spam Team maybe looking at this and saying, well, these are all doorway pages. Because, essentially, they're the same, and just variations of keywords. I think, at the moment, you're probably at the level where this is fine, from the manual point of view. But spreading it out more would definitely kind of go into the doorway page area, where the Web Spam Team might look at this and say, well, there are millions of keyword variations indexed here. And they're nonsensical keyword variations indexed here. So I don't know, student credit cards for senior citizens, or something like that. That happens sometimes. We see that happening from time to time where these pages are just automatically generated. But if these are variations that make sense, if these are pages that are not automatically generated, then I wouldn't see that as web spam. I wouldn't see that as something where Panda would say, this looks bad. I think this is something where you would probably want to test with users, to see how far they understand your site, how far they trust your site, how far they feel that this is providing relevant information, and not just boilerplate text that's put together in different variations.

AUDIENCE: I think the results are a little different. You're only looking at Visa cards. Whereas, on the other page, you're looking at all the cards.

JOHN: Yeah.

AUDIENCE: And the user said they are looking for the Visa college card.

JOHN: But this is something where I'd maybe also do, like, an offline study with people. People who aren't associated with your website. Give them a specific task to complete on your website, and maybe on some of your strongest competitors, and see how they're able to complete that task. Ask them the questions from the Panda blog post, what was it, maybe two or three years ago? And get their feedback directly, like that. I think that's something that would work really well with your website. And would probably give you some information on where you need to focus on more, if these are things that, maybe from the design, or from the layout, or from the usability, you need to work on. Or maybe there are specific niches that you're not paying attention to. So that's something where I think an offline user study would really provide some value.

AUDIENCE: Thank you.

AUDIENCE: Hi, John. I have a question.

JOHN: Sure.

AUDIENCE: So we have a global Selects page [INAUDIBLE] domain. And we have approximately 20 subdomains, which are transactional. We keep seeing the Selects page appear on the transactional sites and site links [INAUDIBLE]. We don't want it there. We tried to demote it to [INAUDIBLE], to no avail. We tried, literally, everything we could think of to break the link between the transactional domains or its global Selects page, but it still persists to be in the site links. I just want-- is that actually a fault? Do you know of anything that we can do to actually remove it [INAUDIBLE]?

JOHN: So the demotion of cyclings is something that, like the name says, it's a demotion. It's not that we would remove it completely from site links. The thing to keep in mind there is also that these site links are put together based on the query that people are searching for. So it might be that one page has some site links shown for one query, but it actually has completely different site links shown for other queries. So we try to match that. But there is no way to completely remove a URL from the site links, apart from putting a noindex on it. A noindex is probably a bit too big of a hammer for your case, right?

AUDIENCE: Yeah. We want the Selects page to run in its own right, when people are searching for the brand, or the company, if you like. It could certainly be dealt, once we took care of it, in transactional subdomains, and that's what confusing, is because we're searching all our subdomain in France, for example. And then we have a global page appear in the site links, which clearly isn't the subdomain. It's--

JOHN: OK. So what you could do there is also use the hreflang markup, to let us know about the different language and country variations that you have. Specifically if you mentioned, like, a subdomain for France and a subdomain for a global website, basically, then that's something where the hreflang markup could really help us. Where we'd be able to say, well, this is the global page, and someone is searching for the French page, therefore, it doesn't make sense. Or, this is the English page, and someone is searching for a French page. Therefore, we shouldn't put this as a site link.

AUDIENCE: OK. That's good advice.

JOHN: All right. Let's go through some of these other questions here.

AUDIENCE: Can I ask you a last question?

JOHN: All right. One quick and last one.

AUDIENCE: I'm coming to Switzerland. Is it possible to meet you?

JOHN: Maybe. You can send me a note on my Google+ account. I travel a lot, so I'm not always here. But if you're here, we can see. Sure.

AUDIENCE: I'll come after 10th of November. So I will meet you when I'll be there in Switzerland.

JOHN: Yeah. Probably, just don't come to Switzerland just for visiting here, because-- it's a nice office, but it's not that nice. There are lots of other things to do in Switzerland as well.

AUDIENCE: Fair enough. Not a problem.

JOHN: Send me a note. OK. "What do we have to modify in Webmaster Tools when you change from HTTP to HTTPS?" You don't have to modify anything. But I'd make sure that you have both variations indexed-- or listed in Webmaster Tools, so that they're verified there. So that you have the information about both of those variations separately. Instead of modifying things in Webmaster Tools, you'd essentially just set up the redirect, set up the rel=canonical. Maybe set up HSTS, if that's something that you're aware of as being relevant in your case. But you don't have to change anything specific in Webmaster Tools.


JOHN: Yes.

AUDIENCE: John, have you seen any other-- I asked Barry last week, in the chat, whether he'd seen the secure stuff come back. And he said he'd seen no recovery, the same as we hadn't when we moved to secure. And we've now unwound it because it killed half our traffic. Have you seen any other anecdotal stuff, or people mentioning that? Or any update from your team as to why that might have happened?

JOHN: We've seen maybe two or three other cases that we passed on to the engineers. But we also looked at the general data. And in general, it seems to work. So for most of the sites, that works fine. But the specific examples where it's not working are really useful for the team to figure out what's going wrong, what we should be doing differently. So in your case, since you already wound it back, it's something that's harder for the team to look at, because they don't see the current status.


JOHN: But I wouldn't roll it forward again just to break it again.

AUDIENCE: No, we're not going to make any other--

JOHN: OK. Good. But if anyone runs across this kind of situation where you move to HTTPS, you don't change anything else on your site, and suddenly you see a significant drop in traffic or impressions, that's something we'd love to hear about. "My online website has a Googlebot problem, as it doesn't show the index.php page as cached, but a random page." That sounds like something where we are kind of getting confused as to which page we should actually index for this page there. And usually, that's a sign that something with canonicalization isn't working as it should be. So I'd work to make sure that the rel=canonical is working properly. That you're not doing anything fancy with the URL parameter settings in Webmaster Tools, so that we can crawl these different variations so that we can index them separately. But it sounds like something with the canonicalization isn't working as it should be for your site. "Large quantities of view state code force the actual text content of the page further down in the HTML. Can view state block prevent search engines from indexing page content? And what would be the recommended solution?" So I think most search engines can index HTML pages that are really large in the meantime, which is several megabytes large. And I doubt you'd have a view state block that's that big. So that's more something where you'd probably see from the usability side of things a little bit of an effect, in that, if people with slow connections have to download a big HTML page to get your content because you have this big view state block in there, then that's something that they might see. But it's not something that would block search engines from actually seeing the rest of your content. And view state blocks are things that have been around for five, 10 years now, so it's something that search engines have to deal with. It's not something I'd say is going to cause significant problems. "Is it enough to have a rel=nofollow to an affiliate link, or should they be blocked by robots.txt also, and redirected and nofollowed? Which is better?" I'd just use the rel=nofollow there. We essentially dropped the links with the rel=nofollow from our link graph, and you don't need to obfuscate it. You don't need to hide it with the robots.txt. You don't need to block it with an extra redirect. A rel=nofollow is just as fine there. The problem with the using robots.txt and redirects and things like that is that it adds an extra layer of complexity to your site, to those links. And it's something that can go wrong. And something where you could accidentally be blocking normal links from passing page rank, or where you could be sending users to a redirect that doesn't actually reach the right target. And all of those things are just, like, extra complexity on your site. And if you can reduce the complexity, I think that's always a good idea. So I'd just use a nofollow on those kind of links. And that kind of solves it for that. Let's see if we have some other questions here with more votes. I'll just-- +1. All right. "How do you get into the queue to ask a question?" So I post a link to the Live Hangout to the Events page when I open it up, and usually it fills up pretty quickly. So for the next one, if you want to join in live, just refresh the Events page regularly, and make sure that you find that link early. And then you could jump in live. For the question and answer, you submitted it in the right place, so that's the right place already. "I've been trying for months to get my website to rank with good content and meta tags. I've read all the white-headed advice out there, but nothing works. Is there anything obviously wrong with my site?" Let me just push it in here and see what happens. And it says, no, there's nothing obviously wrong with your website. So I think I'd just double check to make sure that, from a technical point of view, everything is working right. And just continue working on your website, with more and better content. So it's not that there's anything specific holding you back. You, essentially, are just ranking where you'd normally be ranking. "What's Google's point of view regarding lazy loading images for below the fold content? Is Googlebot able to trigger lazy loading scripts to view all the images on a page for SEO purposes?" This is a tricky thing. So on the one hand, you can test this with Fetch as Google in Webmaster Tools. You can check the regular view to kind of see what we're able to pick up. It only shows the first page load, though. So if there are images that are below the fold, or that are lower on your page, and they're only loaded when a user actually scrolls down to that part of the page, then I could imagine those are things that Googlebot might miss out on. Because we're not going to scroll to all possible locations of the page just to see if anything else changes on this page. So if there are images that you need to have indexed for Image Search and you have them on your page, I'd make sure that they're either embedded directly with a normal image tag, so that we can load them directly and use them as a normal landing page for Image Search, or that they're really high up on the page, so that it's also clear to us that we should load this right away. And also clear to the user, when they search and go from Image Search to your page, that this is actually the relevant landing page for that image. So from that point of view, if you want to have those images indexed, I'd just make sure that they load on the first page view. If you don't care if they're indexed-- where you'd say, well, if users come here with those images, that's fine. If they don't come here with those images, that's also not that bad-- then, maybe keeping them below the fold and lazy loading them is a possibility as well. Chances are we won't be indexing those for Image Search, though. We might index them separately, but it's not something that we index directly with that specific landing page. What you could do in cases like that is have specific image landing pages. So Wikipedia does this, for example, in that they have images embedded in the individual articles. But they also have landing pages for these individual images that are linked from the main article. And if we can follow that link to that image landing page, then we could index that image landing page, and show that one in Search. "My question relates to [INAUDIBLE]. We get a lot of traffic from them daily, and would like to know how to filter exclude them." So this is, I think, a third-party website, some tool that crawls or scrapes these websites to get some information. And I've seen someone from their site post on our forums and say you can block it with the robots.txt, or you can contact them directly as well. There are some forum threads out there that have more advice on how to, kind of, block that traffic, if that's something that you find problematic. Now from our point of view, what happens outside of the Googlebot traffic on your website is essentially up to you. And if you want to block other bots, that's essentially your decision, not something that we decide, or we'd guide you on.

AUDIENCE: Thanks, John. Was just a matter of optimizing the loading speed of the pages, with the help of lazy loading.

JOHN: OK. Yeah. I think it definitely makes sense for a lot of pages. But especially with images, it's something you just have to keep in mind if they're loaded when a user scrolls down to a specific location. And chances are Googlebot won't be doing that.

AUDIENCE: Yep, I get it. Thank you.

JOHN: "Why is it important to have rel=next and rel=prev for paginated article? If Google brings the most relevant page of the article in the Search results, why not just have a single rel story to link on all pages to the first page to show an association?" Good point. I don't know how we would compare that to other variations. I know this was, at least, a significant problem with how we were indexing really long articles. Because it was hard for us to understand which ones were tied in together in which way. So oftentimes you'll see a link on the bottom going to the next five articles. So it's not clear that there is really a straightforward way of stringing these articles together, and saying, this is the big article and this is the right order. So that's something that might be missing with a single link just to the Start page. But essentially, this is something that helps us to understand your content better. It's not a requirement. If you have long articles split up on multiple pages and you don't use rel=next and rel=prev, then we'll try to figure it out.

AUDIENCE: John, quick question.

JOHN: Sure.

AUDIENCE: On the user-generated content, let's say a customer has written a review about the product in all caps. It's a legitimate review, it's just in all caps, for example. Is that a problem from a quality standpoint?

JOHN: That should be fine. I don't see a problem with that.


JOHN: I think it's something you might want to look at with other people on your website in general. But from our point of view, we don't care if you use caps, or if you use Comic Sans as your font, or if you use pink and purple rainbows. It's essentially something between you and your users.


AUDIENCE: Hi, John. Can I ask you a question, please?

JOHN: Sure.

AUDIENCE: My question is that I was working previously on a server for a website with .com. And I was presently redirected to .ae, which is, according to the page, an update. A locational update. Generally, we need to target in Arab countries. And we had built links with some directories, et cetera, said a .com website previously. Can I get a link from same website which is redirecting to this .ae website from that directory [INAUDIBLE]?

JOHN: If that's your local variation of your page, I think that's generally fine. I'd just be careful with regards to getting links and just, like, submitting them to directories, and, kind of, artificially creating links to these pages. Because that's something that our web spam algorithms would try to pick up on. So if the links to your page are essentially just links that you placed yourself on various directories, or on friend's site, or you're exchanging them, or you're buying them from somewhere, then that's something that our web spam algorithms would try to pick up on. So creating a separate local version of your site is perfectly fine. If you're targeting a local market and you have a local domain name, that's fantastic. But I wouldn't, kind of, play the game with the links, and just say, well, this is a new variation of my site. Therefore, I'll submit it to a thousand directories and get links that way. Because that's something that could end up being counterproductive, where our algorithms would say, well, we don't trust any of the links to this site because there's so many manually created links out there. We don't really know which ones are actually the useful ones.

AUDIENCE: I would like to know how to resubmit, redirect all that with .com to .ae. How to tell our users that we are redirector-like?

JOHN: If you have a redirect setup, like a 301 redirect, that's perfectly fine. That's not something that you need to do additionally. If you're moving your whole website from one domain to another one, you can use the change of address tool in Webmaster Tools, which lets you say, this is my old domain. This is my new domain. I am moving everything over one-to-one to the new domain. And that helps us to pass all of the signals that we have to your new domain.


AUDIENCE: John, can I bring a short question?

JOHN: Sure.

AUDIENCE: Last one. I know you've been said this a lot, and we've discussed it before, but I just want you to reiterate something. Is the number of directories within a URL a ranking factor? Does this influence in any way the rankings of a page?

JOHN: No. No.

AUDIENCE: I know, but I just want you to re-explain that. I mean, the clicks are more important than the--

JOHN: I wouldn't focus on clicks. But I think it's something where we see the URL primarily as a technical identifier. And if you put keywords in there, that's fine. If you don't want to put keywords in there, that's fine. If you use URL parameters, that's fine. It should just be an identifier for that page. And the directories are essentially just a way of structuring that information there. So it's not something where we'd say this helps your ranking or is bad for your ranking. Or being like five levels deep is worse than being three levels deep in the URL. We just have it as an identifier.

AUDIENCE: OK. Thank you.

JOHN: Sure. We're out of time. Let me just scroll through the questions to see if there's anything specific we can answer really quickly. "Webmaster Tools says that the sitemap size limit is 50 megabytes, but a recent post says 10 megabyte. What's up?" We support 50 megabytes for a sitemap file, but not everyone else supports 50 megabytes. Therefore, we currently just recommend sticking to the 10 megabyte limit, and just splitting things up a little bit more. "How long does it take for hreflang sitemaps to be crawled?" We essentially have to re-crawl the individual pages. So it's not a matter of just the sitemap. We have to crawl the pages, too. And crawling pages takes different amounts of time. Some pages we crawl every day. Other pages, every couple of months. So you're probably looking at a period of, say, three to six months for all of these pages to be re-crawled. And then we can take that into account for the international targeting session. But it's not something that happens after this three to six months. It's something that gradually happens step by step, as we re-crawl more of those pages. Let's see. Ooh, so many questions left. OK. Let's just open it up for questions from you guys. Let's see if there's anything from you all that--

AUDIENCE: John, I don't have a question, but I have a kind of request. When you post this to YouTube later, will you keep in the bit prior to everyone joining, so that I can share it with my developer?

JOHN: That automatically gets removed. So it's basically only the part where we go on record.

AUDIENCE: So no, then. They're going to have to trust me then.

JOHN: Yes. Well, you can join me in the next one.

AUDIENCE: Completion to the previous question.

AUDIENCE: I'll be here, anyway.

AUDIENCE: Is the number of clicks the fastest way to get the content-- is this inference the importance of that specific contents in rankings? I mean, it's equally the same if you get to a content on 10 clicks or on 3 clicks.

JOHN: So like from your Home page, how far you have to click?


JOHN: Let's see. To some extent, it does matter, because that's how we forward the page rank within your website. So depending on how your website is structured, if content is closer to the Home page, then we'll probably crawl it a lot faster, because we think it's more relevant. And that could theoretically fall into the ranking side of things as well. But it's not something where I'd say you artificially need to move everything three clicks from your Home page. It could be just as relevant if it's five clicks away from your Home page, or if you have to go through a form, or something like that. So I think if something is critical for your site, I'd move it higher up in your hierarchy, like closer to your Home page. But in general, that's not something that you, kind of, artificially need to do.

AUDIENCE: Thank you very much.


AUDIENCE: Yeah, John, please. I have a question for you.

JOHN: All right. One last question.

AUDIENCE: My question is that I was working for a website since one and a half year. And even my ranks wasn't even changing from third page to the front. What was the problem? I was posting blogs or [INAUDIBLE] all the stuff, like doing all that to generate backlinks and everything. But it was not pushing up my rankings from previous-- it was not going back or coming front. I don't understand why that was happening.

JOHN: I imagine what was happening is that our algorithms were picking up on these, kind of, artificial backlinks. So if you're posting blog comments with links to your site, if you're doing guest blogging with links to your site, if you're submitting to directories, those are the kind of links that, when we look at them, we say, well, the webmaster placed these. Therefore, we shouldn't really count these as being, kind of, recommendations of your website, because you're just saying, this is my website and I really like it. It doesn't tell us that this is something that other people like, other people recommend. So that's something where maybe this kind of link-building activity that you were doing there is almost counting against your site, in that the algorithms are saying, well, there's so many problematic links happening to this site that we don't really know how to trust the rest of the links to this site. So that's something where I think doing less link building and really focusing on your website, focusing on your content, making it easy for other people to recommend your content, would probably lead to more success than you, kind of, posting the links to your site separately.

AUDIENCE: OK. Thank you.

JOHN: All right. With that, let's take a break here. It's been great having you guys here again, and I wish you a great weekend. Thank you all for joining.

AUDIENCE: Thank you, John. We wish you a great weekend, too. Goodbye.

AUDIENCE: Thanks, John.

AUDIENCE: Have a good weekend.

AUDIENCE: Bye, everyone.

AUDIENCE: Bye. | Copyright 2019