Inside the Google Panda Update

Hi, and welcome to today's Biznology Digital Marketing webinar. I'm Chris Abraham, one of the bloggers at Biznology. Actually, my blog is out today, so please check it out. I'm the president of Social Ally, a social media outsourcing and support service. I focus on blogger outreach, blogger engagement, and internet reputation management. Today you'll be hearing from Mike Moran, the founder of the Biznology Blog, who will present Inside the Google Panda Update. But before we do, we need to recognize our sponsors. Brick Marketing, full-service SEO solutions company that increases website visitors. Marketing Pilgrim, the internet and social media marketing news blog. We cut through the bull so you don't have to. Rimm Kaufman Group. RKG is a digital marketing agency that combines superior marketing talent with sophisticated technology to increase clients' business through data-driven online marketing solutions. And Social Ally, your social media support and my home. As we wait for more attendees to join, let me review the format of our webinar. Our Biznology webinars last just 30 minutes, so you can easily fit them into your busy schedule. We record each webinar and we'll email you with that link later this week. During our speaker's presentation, you can use your GoToWebinar controls to ask a question. And actually, please ask questions. I'm sure you have them. That orange arrow opens and closes your webinar controls. If you have a question, simply type it into the box labeled Questions at any time during the event and press the Send button. I will select a few questions at the end of our webinar and pose them to Mike. While we are waiting for the last few attendees to join, I would like to remind you that the Biznology newsletter and blog are available for free at biznology.com. So if you are not already a subscriber, we hope that you will sign up soon. Or sign up now. Thanks again to all of you for spending 30 minutes with us. We know how valuable your time is, so let's introduce today's speaker. Mike Moran is the founder of the Biznology blog, a well-known expert in all things digital marketing, and Chief Strategist at Converzium, a leading social consultancy. Mike is the co-author of Search Engine Marketing Inc. and the sole author of Do It Wrong Quickly. Mike is a veteran of IBM, managing groups in IBM.com for 8 years, retiring from IBM IBM in 2008 as a distinguished engineer. So if you've ever struggled with keeping your search ranking up after the Google Panda update, this is the webinar for you. Mike, take it away.

Thanks a lot, Chris. So some of you may be asking yourself, what is Google Panda? It sounds like this nice little fluffy thing, right? So it's a lovely little bear. But the truth is that the Panda Search Ranking Algorithm Update was actually named after the engineer at Google who came up with the idea. And so many of you are familiar with the idea that search ranking algorithms change. So Google is making changes all the time. So is Bing and other search engines. And so about 18 months ago is when Google first announced its Panda update. And so you might say, well, why are we talking about it now? Well, Panda is actually a different kind of update than just about everything else Google's ever done. It has become more and more important over these last 18 months. And unlike almost all of the algorithm updates that Google has ever done, Panda Keeps changing. So it wasn't just one algorithm change done 18 months ago, and now we move on and see other changes. There are changes that happen all the time based on this change to Google Panda. And so what we want to do is to first understand what is it that these things are— what kinds of things are happening, and what can we do about it? So first, let's think about what is a search ranking algorithm. Because you're not really going to care whether there are changes to the search ranking algorithm if you don't even understand what the algorithm does. So the first thing to understand is that there are actually 2 ranking algorithms. There's the ranking algorithm for paid search and the one for organic search, or SEO, search engine optimization. So Panda is only affecting the SEO. It's only affecting the organic search ranking algorithm. And so Panda is an algorithm change that helps decide what becomes the number one organic search result. But it has no effect at all on paid search. So whatever you've been doing in paid search, you can keep doing it. But for organic search, you want to understand how Panda has changed the game. Now, one of the reasons that search engines change their ranking algorithms is that they think they have a better way of ordering the results. So the ranking algorithm is the thing that decides in which order do all of the search results show up. Because it's interesting to see a search results screen that says, okay, here's number 1 of 10 million. If they just are applied in random order, that's not going to help a searcher very much. They want to know that the number one item is the one that's really the best for them to click on. And so that's what the search ranking algorithm does. And periodically, Google and other search engines change their ranking algorithms because they've come up with something that they think makes the results better. They think the number one result will be better. But there's another reason that they change the ranking algorithm. And that is because there are some people out there that will do just about anything in order to get their page on the first page of the search results, in the top 10. And the reason that they do that is because studies show us that almost 90% of the clicks in search happen on results that are shown on the first page. So only 10% of the people have gone to the second page and click on something. And so that shows you how high stakes it is for people to get their content on the first page. And so you have some people who will do just about anything to make that happen. And so those are usually called black hat SEO experts. So they are doing things that violate the terms of service, is what it's called. So the rules of what you're allowed to do and not allowed to do in search marketing. And so these people, they fool Google. They fool the searchers. They are very good at what they do. They spend all day, every day trying to get an edge. And you are probably not one of these people. And if you are, I have no idea why you decided to spend a half hour listening to me, because there's nothing you're going to learn here today. Instead, you're probably the kind of person that tries to avoid these kinds of techniques. You're not actually looking to trick Google. You're just trying to get your content to rank better in the search engine. And so those people are doing what's called spamming. So they're spamming the search engine. They're violating the terms of service. And if Google catches them, they're going to take action. They're going to get their stuff out of there. And Panda is actually an example of how they're taking action. And we'll talk about that. But this isn't only about spamming. It's not just about people trying to play tricks and games and all sorts of tricky techniques. What's really happening is that Google is trying to change the game so that you're not out there trying to figure out what those techniques are. So what are the things that will make Google happy? So instead of looking at the terms of service, the rules that you're allowed to play by, when you're trying to get your content ranking as a search marketer, and then trying to read those like there's some kind of legal document that you want to get every edge about. What instead I want to urge you to do is to look at what will work for all 3 parties to a search transaction. So every time there's a search, you are involved as the search marketer, Google or another search engine is involved, and the searchers are involved. And so What is it that works for all 3 of them? If you do things that work for all 3 parties in the transaction, rather than taking the approach that the black hat spammers take, which is they're trying to do something that helps them as a search marketer, but they don't really care if it helps Google or the search engines. Those are the kinds of things that Google wants to stamp out. And so what kinds of things do those people do? So they do something that is referred to as reverse engineering the algorithm. So what does that mean? What it means is they try and figure out what today's Google algorithm is looking for, and then they try and provide it by any means necessary. So, and this is stuff that unfortunately even those of us who are not spammers, who aren't black hat SEO experts, we fall into this kind of thinking. So we say, well, Google wants keywords on the page, so let's put lots of keywords on the page. They want them in the title, okay, let's put lots of them in the title. They want to see links come to our site. And so let's go out there and we'll get links. So we'll ask people for links. We'll trade links. We'll pay for links. Any way we can get links is good. But that isn't actually what Google wants. It could be that at some point in the process, that is how the algorithm is working. That's what the algorithm is rewarding, that ranking algorithm we talked about. But what Google really wants is really high-quality search results. They want the searchers to find what they're looking for. And that ranking algorithm is only a means to that end. And so what's really going on here is that if what you do is you stare down all of today's components of the search algorithm, and you can figure out somehow what kinds of things the algorithm is rewarding and penalizing, and you try and do all those things, but you're not actually creating high-quality content, which becomes high-quality search results, then what you're waiting for is for Google to change the algorithm so that whatever you're doing doesn't work anymore. But if instead you're creating high-quality content, then what's going to happen over time is that Google is going to tend to change the algorithm in ways that will reward what you're doing. And so what is it that Panda is doing in the mix of all of this around the search algorithm? So what's really happened is Panda's added a new factor to that search ranking algorithm. So instead of the search engine only looking at keywords on pages, or only looking at links to pages as a way of deciding that a page is high quality, What's starting to happen is that Panda has now added another piece to the algorithm, which is human beings rating the search results. So if there was some way for spammers to fool the search engine by buying links or by tricking the search engine into thinking that these are legitimate sites sending links over to these other sites, and all sorts of things that might cause Google to think it's a high-quality site because they see all these links coming in, but it's actually a garbage site. It's actually a site that's not that good at all. Maybe a site that's full of ads or one that is trying to get people to buy something and has very low-quality content. And most of the people actually abandon. But a few of them buy something. And it's just like spam email. You only need a few people to do it in order for it to be valuable. So if that's the kind of content that's somehow worming its way up to the top of the search engines, Then what you need to understand is that that's what this is going to ferret out. So Google is hiring these people who now have a way of rating the search results. So they're deciding, is it a good design? Did the page load rapidly? Did they think the content was high quality? Was this a good answer for the search? And so sites that get a good ranking from the human— a good rating from the human raters, are going to get bumped up in the search results, and other sites are going to get knocked down. Now, what happens after that is that you suddenly see all sorts of things going on. But one of the things that you might ask yourself is, well, look, even Google doesn't have enough money to pay human raters to look at every blessed search and all these different sites that come up. I mean, that's crazy. How could they possibly pay for that? And the answer is they can't. So it would seem like this is not a very useful thing to add to the ranking algorithm, because you can only rate a small number of sites. But Google's smarter than we are. And they actually are using a technology called machine learning that allows them to apply ratings to sites that they cannot afford to have a human being actually go in and rate. So how does that work? Well, machine learning— is a technique that looks for patterns, so features. So you might be able to identify that these 2 people in the picture are brothers because you can kind of look at them, they look a little bit alike, they seem like they might have a resemblance to each other, but not everybody might be able to recognize that pattern. And certainly just handing this picture to a computer, the computer couldn't identify something like that. But the computer is actually very good at identifying patterns in text. And so if your site seems to look like sites that are getting low ratings from the human raters, then you're going to get knocked down in the search result. Whereas if your site seems like it matches the same kind of patterns that a high-rated site got, Then you're going to see that your site is going up in the ratings, according to— in the rankings, according to the Panda algorithm. And so what's really going on here is that the computer can now identify which sites seem to look like other sites. So they have something in common. They have common features that make them look the same as another site. And so even though Google can only afford to pay human raters to rate a small number of the sites, they're rating enough that they can use machine learning to detect patterns of how they're similar to each other. And so if they've rated a site that they think is similar to yours, then they're going to assign that same kind of quality rating to your site. So now what you might be asking yourself is, so what are they looking for? What are the kinds of things that it might mean when you say it looks like low quality. And the answer is, we don't know. Now, the reason we don't know is twofold. One is that Google doesn't tell us. And they have a reason for not telling us. They're not going to want to tell us because they don't want the spammers to suddenly find out what are the kinds of things they need to do to look like high-quality sites without them actually creating high-quality sites. But the other reason, that is, in some sense, Google doesn't even know. And the reason for that is the same reason that when you looked at the 2 brothers in that picture, that you could say to yourself, well, I see some kind of resemblance. But you can't actually explain what it is you see that causes you to think that they look like each other. And so the same thing is true when it comes to these patterns. The computer doesn't explain why it says these sites are similar to each other. It just has picked out certain patterns that make it think that things are similar to each other. So it's possible that sites that have longer titles are more likely to be spam, or that sites that have higher ratio of pictures, they have many, many more pictures on the site than the average site, Maybe those are low-quality sites because they tend to be full of ads. Or maybe sites that have lots of common runs of words. So there's like 2 paragraphs that are on this page, and those 2 paragraphs are on 30 other pages on the web exactly word for word. And maybe those are sites that have been put together by robots that are out scraping sites and just dumping pages out there. We don't know. So I made all these things up. We have no idea what the— that the patterns actually are. But here's what we do know. When we see what's going on, we can see that there are certain kinds of sites that used to rank highly that now aren't anymore. And so content farms— these are sites that have been put together by computers that are scraping content from one place or another, older content. So for some kinds of search queries, it's not useful for you to have content that's old. It's like if somebody says, They're looking for digital camera reviews. They don't want to see a review page from a year ago. Other sites that are loaded with ads seem to be getting ranked lower. And that's because people tend not to like them. And also vertical search sites. So if you search for something, and then the first thing you do when you click on the page is it does another search, say in a shopping search engine, those are tending to get knocked down. But interestingly, not the Google Sites. Those ones seem to still hang in there. And now, as we talked to you about earlier, this isn't just bad luck. This isn't just some terrible thing that happened to people where they were doing all the right things, and now all of a sudden the ranking algorithm changed and they're screwed. Now, for some people, it is bad luck. Because some people really might have high-quality sites, and the algorithm isn't working that well. And that can always happen with technology, that there's some unfairness. But for a lot of the sites, this is really a good idea. Because they really didn't have high-quality content. And that content is now getting knocked down. And other, better content is coming up. And this goes back to what we said before. There's 3 parties to everything that's done in search. There's you, the search marketer. There's the search engines. And there are the searchers. So every time you do something that's really trying to appease the algorithm, when you're saying, well, this is what the algorithm is looking for, so that's what I'm going to do, then you're actually taking a step that's a dangerous step for you. Because when the algorithm changes to be a better reflection of what searchers are really looking for, your content is more likely to get hit. The other thing that we've noticed is that if all you look at is the rankings, which of your pages are number 1, or how many of them went from number 4 to number 9, If those are the things you're looking at, one of the things that some people have noticed is that sites that actually lost a lot of traffic because their rankings went down, instead of having the number 2 result, now they have the number 9 result, so they're getting a lot fewer people clicking through to their site, they actually didn't lose much sales. And so what might be happening is Google might be improving its ability to send qualified leads. They might be starting to understand Not just which sites get clicked on and that people go to, but which ones actually convert. And so they're changing which— and so if you had a low-converting keyword, it could be that that is part of what's going on with their understanding whether something's high quality or not. And so you might ask yourself, so Panda happened 18 months ago. So is it over? Is everything okay now? So if I got through the last 18 months, do I not have to worry anymore? Well, yes and no. So one of the things that we talked about at the top is that Panda actually keeps changing. It wasn't just one algorithm change. What happens is that there are constant changes in the human ratings. There are constant changes in which features the machine learning is picking out. And so whether it's people looking at ratings that are coming from social media, or whether it's things that fit directly into Panda, There are all sorts of other ways of assessing quality. And those things are changing all the time. And all of them are designed to understand quality better than merely looking at link patterns did in the old days. Now, the other thing that's true is if Panda is working, and it's been going for 18 months now, and Google hasn't backed off from it, at some point Microsoft Bing and maybe other search engines around the world might start implementing a similar thing. Because this isn't rocket science, this technology. It's stuff that people do know how to do. They just never applied it this way before. So what we really want to do as we sum up is to say, it doesn't make sense for you to do this because you're being a good scout. You don't do this because you're trying to get into heaven. What we're telling you to do is the more you focus on producing content that people actually like, that searchers think answers their questions, that really is the best answer, for a particular search keyword, that's what more and more the search engines are trying to reward. And they're spending enormous amount of time doing that. But it's also good for you because it's less risky. There aren't as many ups and downs. If you really have quality content, that every time the search engines change their ranking algorithm, they're still going to have your content up there if they've done a good job with the change. And so if you constantly focus on quality content and a quality experience for the searcher, over the long haul, success is going to follow. So that's all I had for today. So Chris, I'm sure we got some questions. I'm hoping that people will come back next time when we talk about using social media for B2B sales. We skip August and we'll come back in September. But Chris, why don't you run us through our sponsors again, and at that point we can open up for questions.

Thanks, Mike. I'm sure our attendees have a much stronger idea of how the Panda update affects all of our search results. But you didn't answer every question. So I am going to say a bunch of words to give our listeners and watchers some time to ask some questions. I've got several good questions from our— while I do have one question teed up from our audience, I'd like to get a few more. And while I'm doing that, I'd like to remind our audience that it's not too late to ask their own questions. please, by typing into the questions box in the GoToWebinar control. But before we get to the questions, I'd like to thank the sponsors again. Thank you to Brick Marketing, full-service SEO solutions company that increases website visitors. Thanks to Marketing Pilgrim, the internet and social media marketing news blog. We cut through the bull so you don't have to. Thank you to Rimm Kaufman Group, RKG is a digital marketing agency that combines superior marketing talent with sophisticated technology to increase clients' business through data-driven online marketing solutions. And thank you to my very own Social Ally, your social media support. Now on to the questions. Let's see, as of right now— oh good, smashing. There are 2 so far. So let me ask you the first one. Do you have any idea, Mike, how small the human subset that is rating pages is?

I don't. What I do know is that there have been many spottings on the web of advertisements where Google has been advertising for human raters. And so we know that they've recruited for this. And we know that they've hired people to do that. And there have even been a couple of— things that I've seen where people who have done the job of rating have actually talked about it a little bit. And so we know that it's going on. And Google has been public about the fact that this is how the new Panda algorithm works. But I have not personally seen any explanation of how small the subset is. What you can know about it, though, is that because they've been doing it now for 18 months, Over time, they are building up more and more of what in machine learning they refer to as training data. So training data says, this is the absolute correct answer that a human being gave us. So even though raters probably aren't perfect, they are probably a lot better at detecting quality content than the old computer algorithms were. So we know that over time they grow more and more and more data. And we also know that what people think of as high-quality content does change over time, but it changes rather slowly. So they don't have to have a big investment in human raters in order to build up an enormous amount of data that can really help. And a lot of the decision of how much data they need depends on how good a job they think that they're doing of analyzing those patterns that the machine learning tells them. If they think that they've got a really, really good handle on the patterns, then they don't need as much training data. But if on the other hand they think the patterns aren't working as well as they would like them to, then they need to collect more training data because the more you get, the more accurate machine learning becomes.

I actually met a human rater 7 years ago in a blogging class I did at the Writers Workshop here in DC. She was going through Japanese sites. She was a brainiac here in DC. And so this isn't a brand new thing, but I guess maybe they're stepping it up. Next question, Mike. Any idea what weighing— what weighting percentage the human and machine ratings have versus the old school input like links, page title, or word count?

No. The truth is that that probably changes all the time. It changes all the time in a couple of ways. One is that, as we talked about, Google is constantly changing the weightings in its rankings. And so even though once in a while they will tell you that they're coming out with a new algorithm update, the truth is that those changes happen fairly frequently. And there are minor changes that happen all the time. So those things are constantly being adjusted. But the other thing that's true is that we have to have a lot more sophisticated thought process around the ranking algorithm, because the truth is that there isn't one ranking algorithm. There are probably hundreds of ranking algorithms. The very first thing that a search engine does when it analyzes the keywords that have been typed in by a searcher, is it tries to select what the right ranking algorithm is to use. So let me give you an example. We talked earlier about how if someone types in digital camera reviews, it's pretty important that Google not show a result from a year ago. And so the factor of freshness is extremely important for a keyword like digital camera reviews. Because digital cameras are changing all the time. And so you really need to understand that. On the other hand, if someone types in, I don't know, long-term care insurance reviews, long-term care insurance doesn't change all that much. The companies that were really good last year are probably pretty much the same this year. And so having something that's right up to the minute that was published in the last week is probably not as important Something that was published 6 months ago, a year ago, might be a very, very good resource for that keyword, even though they're both about reviews. And to take a more extreme example, if someone types in Abraham Lincoln's birthday, well, it's not really important at all that that information be up to date. He was born February 12th, 1809. And no matter how old the page is that says that, it's still correct. And it's not going to change next month, next year, or 10 years from now. And so depending on what people are searching for and what they're looking for, it could be that freshness is a very important factor in that ranking algorithm, or it could be kind of a middling factor, or maybe it's no factor at all. And so Google first tries to figure out what is the right ranking algorithm to apply to different search keywords. They kind of categorize the keywords into certain buckets that they choose a ranking algorithm for. There is likely no answer. Even if Google could tell you or would tell you what the percentage is, there probably isn't one answer. There's probably dozens or more answers of how it's applied to different kinds of ranking algorithms that they use every day. Okay.

One more question. And because we're— this is from— the first question was from Tom Burke. The second was from John. And the third one is from Don Sturgill. Should a blog include a static description block of copy? If so, how should it be placed, such as a sticky post, a features box? How does this affect SEO?

So a static box is something that can be put on a blog, because what you're looking for is for to explain what the blog is about. And what I would say is a sticky post is probably not the best way to do that because that kind of gets in the way of whatever the latest information is out there. A sticky post would probably be there on top so that if people come to the blog to look at what's new today, they'd have to go a little bit further down to see that. What I tend to recommend is that it be a piece of the layout of the page. And so it's on the page somewhere, but it doesn't necessarily jump out. It's something people can read while they're reading the post if they're interested in it. And that's the way we do it on Biznology blog. But it doesn't have a tremendous impact on search optimization. And the reason is because what causes blog posts to be found are words that show up in the meat of the post. So when Google sees that, for example, on the Biznology site, every single post has the same boilerplate about what the Biznology blog is about, then no particular page on the Biznology blog becomes a really good page for any of those keywords because it's on every page. And so I wouldn't worry a whole lot about it from a search marketing point of view. But I would worry much more about it from a usability point of view, from a point of view of what's the right thing for your blog reader.

I have one quick question, Mike. In lieu of Google+ being integrated into search results, Should we or shouldn't we as singletons or business people be rushing in to create and maintain personal and/or brand accounts on Google+?

Well, I think that there's no doubt that there is search marketing value from Google+ because Google has lost its access to the Twitter feed. So it doesn't have as much ability to understand what's going on minute to minute that it used to do. And especially if you're a business that focuses a lot on things that change rapidly. So things like new models, things like stories in the news. I mean, if you are CNN or some company that always is working on things that are changing, then having up-to-the-minute information flowing through Google+ can really help you with your search results. But for other companies, it probably makes sense to Focus on what seems to bring you bang for the buck. So any of these techniques are things that are worth spending time on, but the truth is that it's a question of what you're not doing because you're working on this. So if it takes you time away from things that are more important, then probably it doesn't make a lot of sense for you to work on Google+ unless you're starting to see buildup of people being added to circles for you in Google+. If you're seeing that there are people that are starting to that are starting to follow you and look for information from you in Google+, then that's a reason to do it. The search marketing benefits will just come as a bonus for that. But if you're not seeing that, if it's not a place that you think you want to spend the time on, if you're already overloaded with what you're doing in social media on Twitter, Facebook, or other places, then it's okay for you to hold back for a little while. It's really an individual decision based on where you think the right place is for you to put your time.

Thank you, Mike. That's all the time we have for today. Thank you, Mike, for these great ideas, and thanks especially to our audience for your participation and your awesome questions. If any of you had questions that we did not have time to answer, you can email your questions to Eileen, eileen@mikemorangroup.com, and she will be sure to get them to Mike for the answer. Later this week, we'll send you all a link to the recording of this webinar to listen to again and watch and to share with others. We also invite you to mark your calendars for our next Businessology webinar, as always free, so please tell your friends. Using Social Media for B2B Sales, scheduled for 11 AM US Eastern Time on September 25th. Thanks again, and I appreciate all of your time. Have a good day.