Central Perk: SEOmoz Daily SEO Blog

vineri, 28 ianuarie 2011

SEOmoz Daily SEO Blog

How Organized Spam is Taking Control of Google's Search Results

Posted: 28 Jan 2011 12:26 AM PST

Posted by invseo

This post was originally in YOUmoz, and was promoted to the main blog because it provides great value and interest to our community. The author's views are entirely his or her own and may not reflect the views of SEOmoz, Inc.

In the past few months, I have been watching a very unsettling trend unfold in very competitive ecommerce search results on Google. It appears that huge amounts of money are put into place to systematically and successfully manipulate highly competitive search terms in order to sell fake merchandise of almost every bigger brand out there. Some of these sites even solely exist to steal people's money and don't ship anything at all.

Rest assured that I am not talking about some people doing good linkbuilding or about people buying a lot of links. These operations I talk about are much, much bigger and in all cases almost certainly run by criminal organizations of some sort. They not only greatly affect US search results but are also very present in at least UK, France and Germany.

In this article, I will show you several examples of where Google’s search is absolutely broken (and by broken, I mean that 10 out of 10 page one search results are entirely fraud). I will also show you exactly how these rankings are achieved and take a look at what the impact on consumer’s may very possibly be. Last but not least, I’ll try to help you recognize these kinds of websites so you can avoid them as they become increasingly difficult to identify.

Exhibit A (“nfl jerseys”)

Let’s get started with [nfl jerseys] as our first keyword to be examined. If you take a look at the US search results on google.com with personal search disabled (add &pws=0 to any search URL), you will get a list of websites which claim to sell said wear and merchandise at a significantly discounted price. Such huge discounts can be found on pretty much any of these fake shops, many ranging up to 75% in “savings”.

Here’s the search engine results page as of 01/04/2011:

NFL Jerseys SERP

(Note: I’m not trying to “out” any particular site, so I removed any domain names in question from the screenshots)

As you can easily see, all of these sites feature ridiculous keyword stuffing in their root page titles as well as the term “jersey” within their domain name. This is both very common among them. Result #5 even contains Chinese letters.

As of this writing, the entire first results page is composed of fraudulent websites. In other words, Google’s organic results have become entirely useless for this search phrase.

Exhibit B (“pandora jewelry”)

Next, [pandora jewelry], also a very popular and well-respected brand. Positions 3, 5, 7, 8, 9 and 10 are fraud, which results in a 60% share of useless results.

Pandora SERP

Exhibit C (“thomas sabo”)

Looking at [thomas sabo] SERPs, they feel like a déjà-vu. Another jewelry brand, another wave of artificially boosted shops shipping either replica ware or just nothing at all: results 3, 4, 5, 6, 8, 9 and 10 should not be listed there at all in the first place (70%).

Thomas Sabo SERP

All of these sites try to appear as legit and official as possible.

See for yourself

Before diving into the details, I urge you to take a look at Google’s results for these queries yourself. Try searching for other brands, too. Almost every popular brand is affected by this growing issue.

How they do it

Now that I’ve shown you how seriously broken Google is, let’s take a look at why Google is ranking these sites so well. Since these are no legit shops after all, it’s obvious that there is no kind of “branding bonus” or boost through actual social media activity at hand. There’s only one thing that leads to these rankings. You’ve guessed it: keyworded anchor-text heavy links.

The interesting question is though: where do these sites get their (anchor-text rich) links?

I have taken a look at many of these sites and found out that their link profiles are basically comprised of two kinds of links: automated forum and blog spam along with some hacked websites.

Let’s take a look at the anchor text variation of result #1 for “nfl jerseys”:

Anchor Text Variation "nfl jerseys"

This site also has a page authority of 64 and a domain authority of 57, according to Open Site Explorer.

Google's best guess for "pandora jewelry" looks similar:

Anchor Text Variation "pandora jewelry"

And the #1 "thomas sabo" result:

Anchor Text Variation "thomas sabo"

You might be surprised to see plain-old forum spam work this well, but let me get one thing straight: it’s not like Google is not penalizing or de-indexing any of these sites. I see them come and go on a daily basis (although some actually seem to stick for weeks or even months).

However, these people (or rather organizations) push such huge amounts of these sites into the web that Google - obviously - is having quite a hard time catching up.

In some way, and this is my personal opinion, this might be related to the Caffeine update - Google is now crawling and ranking sites a lot faster than ever before, but it appears overall search quality has suffered dramatically in the past 6 months or so.

Furthermore, link placements on hacked websites are very difficult to spot algorithmically. Granted, many of those links are not visible to the human eye and that should raise some flags since Google is capable of rendering any page, but overall it’s not comparable to catching automated posts on tens of thousands of web forums.

What really should have set Google's alarm off, though, are the link growth patterns. Let's take a look at the "nfl jerseys" top 3:

"nfl jerseys" link growth

Two of these sites started spamming back in April 2010 and are still ranking in January 2011. Go figure.

Same goes for "pandora jewelry" and "thomas sabo":

"pandora jewelry" link growth

"thomas sabo" link growth

You get the picture.

What Google needs to do about it

Rand talked about it already, and his advice is instantly applicable to this issue: Google needs to greatly lower the value of keyword-rich anchor texts.

Think about it: if Google had not at all taken anchor text into account for these sites, none of them would probably rank anywhere near the top 10 results. Their links come from very different sources, and almost none of those sources is even remotely related to what their pretending to be selling.

As long as anchor text links outrank links from actually related websites, this is not going away anytime soon. Same goes for exact match keyword domains, by the way.

I do realize that anchor text is very important, but its abuse has reached a point where it’s no longer a ranking signal to be trusted as much as it currently is. Heck, I've actually seen websites rank #3 for these terms with one single sentence on the page: "seized by Department of Homeland Security".

What it means for SEO

Google has a serious problem, and I’m sure that they have been working on it relentlessly for quite some time now.

What it means for SEO is that whatever is working for the sites mentioned in this article - it will probably stop working soon. I would not be surprised to see Google shift even more ranking signal power from anchor-text heavy links to relevant social media “chatter”. I have a feeling that it’s gaining more traction as we speak.

Of course, tweets and status updates can be spammed, bought and faked, too. But at least it will buy Google some time.

This fight is never over nor ever "won" by anyone. Ever.

How to identify these sites as a consumer

Since I don’t want any of you to order from these guys and receive either fake goods or nothing at all, here’s some advice to identify them:

Most of these sites:

offer unrealistic discounts (>=50% are pretty much everywhere)
have no actual postal address
supply only a contact form or
supply only a GMail/Hotmail email address to contact them
feature way too many “trusted logos” in their footer
are written in poor English

Considering that most of the sites I talked about earlier already ranked well while all the holiday shopping took place, I can only imagine the damage done to thousands of families and individuals.

Please be cautious and remember that if a deal sounds too good to be true, it very probably is.

Since this is my first article for SEOmoz, please let me know in the comments if you liked this article and give me a “thumbs up” if you did. In case you’ve even been affected by this kind of fraud personally, I’d love to hear from you, too.

- Rouven Balci, SEO at Toms Gutscheine

Do you like this post? Yes No

Interlinking Your Sites - Whiteboard Friday

Posted: 27 Jan 2011 01:00 PM PST

Posted by caseyhen

Hello mozzers! Wow, my first post as a Moz staffer and it's not even something I wrote or recorded, I promise the training wheels will come off soon. This week Rand dives into a topic that many of you have asked about in our Pro Q&A, “Interlinking my websites, can I do it?” As an SEO you can almost guarantee that you will be asked; “Why can’t I just buy 100 domains and link them back to my website with amazing anchor text and then conquer the world?” Rand give some helpful hints on when it is appropriate to link to your other sites and when it’s not such a great idea. After watching, please feel free to voice your option on the subject and how you deal with interlinking your websites.

Also at the end of this weeks transcription is the re-recorded version of last weeks live Whiteboard Friday which was not recorded due to a glitch. Please feel free to head back to that post and make any additional comments about the video.

View statistics for this video

Embed video

Video Transcription

This week on Whiteboard Friday we're going to be talking about interlinking your site. This is a particularly relevant, I think interesting topic for a lot of people who are thinking about like, I own several websites, or I have relationships with a lot of websites. I control a bunch of sites. Maybe they are sites of clients or partners. A lot of times they are just sites that you have owned and registered yourself or that your company has. I hear a lot of this question and introspection around like, "Hmm, how much can I interlink? Can I interlink between all the pages on all the sites to help boost my page rank and send good anchor text and all these kinds of things. It seems like owning my own websites I can control all of this stuff. It would be really good for my SEO." But there are some dangers here. The search engines don't want to be manipulated by someone who owns 50 or 100 or 1,000 different websites. So, you need to be careful. There are some best practices to follow along these lines. I'll give you some simple examples, but you can extrapolate these out to more complex situations or to the specific situation you might be facing.

The first rule of this interlinking phenomenon is to link for humans and users first and to think about engine second. Now, the reason for this isn't because I'm telling you oh you have to be super pearly white hat and we can never do anything that would affect or impact or manipulate the search engines. I understand that you're going to want to get good anchor text. You're going to want to flow link juice from powerful pages to pages that need to be indexed that you want to rank well. That's okay so long as you think about why humans would want to pass from one page to another.

Let me give you an example. Let's say I am on this domain that I own and control. I can find a page on here where I talk about something that exists on domain two - a resource, a relevant topic, something where you can learn more information about that specific subject, whether it is commercial intent, information oriented, a cool piece of link bait, a piece of news, whatever it is that exists over here. That makes great sense to link over to, and I think it is perfectly fine and legitimate to link from those. For example, if there is some great news about how elephants at a particular zoo have been thriving in the new environment and you want to rank well for elephants at this particular zoo, let's say at the Bangalore Zoo, it's fine if you have some content over here that mentions those words, that is talking about them in a post, in an update, in a news item, something relevant, and you link over. That works great, because that way humans, who might want to learn more about this topic, can go over there and get that information. It is exactly what they want. Now engines as well will see and recognize that.

But if you do something spammy or manipulative, and this goes to rule number two, and you put something like "elephants Bangalore Zoo" in the footer of every page. This page has it. That page has it. This page has it. And they are all linking over to this one, and then domain three and domain four and domain five, they're all linking there too. That's super weird. So, putting those things in footers, having overly optimized anchor text, anchor text that just doesn't sound natural, doesn't fit with the flow of the page, has nothing to do with the content of what is on there, throwing in unrelated links, throwing in "elephants Bangalore Zoo" when this page is about where to buy pens. Just that kind of stuff is going to be confusing to humans as to why it exists, and that will mean that it might get spam reported to Google. It might be seen manually by quality raters, or it might be algorithmically detected. None of which you want to have happen. Besides which, you don't want to be scaring off your users with this manipulative, junky stuff anyway. Users are sensitive to spam and manipulation just like engines are and they'll be turned off. They'll think less of your brand and your site when they see that type of stuff. So, watch out for that.

Third, last rule, Google really knows a lot about what is happening on the Web. Not just through things like Google Analytics, but through Google Webmaster Tools, through e-mail accounts, through FeedBurner, through the Google toolbar. If they see that you appear to be trying to hide a link profile from them and link profiles are looking really similar between your different domains and there is lots of interlinking happening and the registration or hosting looks like it matches . . . . there are some SEOs certainly out there who are advanced and sophisticated enough to be able to spread their network out and have that rigid discipline about never visiting the same two sites with the Google toolbar on and making sure that no Webmaster Tools accounts are linked and all this kind of stuff that black hat operators often have to jump through these different hoops. There are ways to do it. As an ordinary marketer or ordinary operator, you know, SEOmoz-type white hat operator, you're really going to want to be authentic.

It's okay to link between sites that you own. I wouldn't be paranoid about this. Just make sure that you follow these rules and do it in a genuine, authentic way. Then Google is not going to be like, "Oh, this guy owns these sites. He's linking back and forth between these." They're going to be like, "Oh, yeah, he had some news about elephants and he points over here. Yeah, that makes total sense." Or, you know, "On his 'about page' he mentions the other sites that are also owned by the business owners of this site. That seems totally reasonable and fine." Maybe the privacy policy might have this privacy policy or these terms of use govern these five different websites and link out to all of them. That's okay. That makes total sense. But having manipulative anchor text in the footer of every page or in the side bar of every page on all these different domains, that is going to getting you in trouble. I'd watch out for that.

All right, everyone. Hopefully you've enjoyed this edition of Whiteboard Friday. Take care. We'll see you again soon.

Video transcription by SpeechPad.com

Live Whiteboard Friday From Last Week

View statistics for this video

Embed video

For the transcription of last week's video, shown above, please visit the video's post here.

Do you like this post? Yes No