Deep Dive Into's Penguin 2.0 Penalty

Share this articleShare on FacebookShare on LinkedInTweet about this on TwitterShare on Google+Pin on Pinterest

[notification style=”warning” font_size=”13px” closeable=”true”]

Don’t forget to signup for our notification list to get informed on other Case Studies on Google Penguin updates.

This analysis was created using our Superhero Plan and the new Competitive Link Detox tool extensively. This professional plan allows you to perform analysis for your or your competitor’s sites. For small businesses we now also have the new LRT starter plan available.

Check out our new Link Detox Pro plans.

Don’t miss our next LRT Associate Training. Get one of the last spots if you’re up for professional training.

Google Penguin 3.0 launched

penguin 2.0 case study

Learn from the Google Penalty of

I am very proud to present our 9th case study deep dive into the Google Penguin 2.0 penalty written by our newest LinkResearchTools certified professional Hannes Hauser

Enjoy & Learn!

Christoph C. Cemper

Table Of Contents

1. Introduction to Google Penguin 2.0 Update

Google Penguin’s 2.0 update in mid-May 2013 caused many significant drops in rankings of numerous websites, even very popular and big websites. The most famous victims of this update are shown in the table below:

Google’s Penguin Update 2.0: Loser Analysis from Searchmetrics (23/04/2013)

It shows that generally sites with thin content and those that were/are especially weak with untrusted links were hit by this update.

We will focus on and see if the selected domain will continue to have the same thin-links problem.

A detailed look at Searchmetrics SEO Visibility shows a massive drop in visibility between 19th and 26th of May 2013:

1.1 Negative Changes in Keyword Ranks

It is useful to look at some keywords that allowed for good ranking and significant search volume in keywords in the past that then lost their ranking and caused such a drop in SEO visibility.

1.2 Six example keywords for negative ranking changes that hurt:


Keyword Search Volume (exact match in in USA)
the colbert report 40.500
oddities 12.100
hot in cleveland 27.100
the middle 49.500
tosh.0 90.500
sons of anarchy 368.000


1.2.1 Drop for "The Colbert Report"

1.2.2 Drop for "Oddities"

1.2.3 Drop for "Hot in Cleveland"

1.2.4 Drop for "The Middle

1.2.5 Drop for "Tosh.0"

1.2.6 Drop for "Sons of Anarchy"

2.0 Penalty Reasons Analysis

2.1 Too intensive money keywords linkbuilding?

It is a well-known Google action to down rank those sites which have a high percentage of backlinks with money keywords since 2008. The Penguin 2.0 update focused on this unnatural linking practice or over-optimization as well, so it is useful to see whether this problem was the reason for the drop in keyword-rankings described before.

For this purpose, LRT BLP report delivering the whole backlink profile of a site fits the most. As this report will be used for analysing other hypotheses, we enable all available metrics and 5x Link Boost (delivering maximum of 50000 links). We also enable the so-called site wide link filter in order to increase validity of our report, (search engines do not count each site wide link separately), and skip site wide links after three links per domain are found:

In just a few minutes our report is ready and we can start our analysis. We will come back to the first Power*Trust graphic later on, but for now, we will concentrate on the second part dedicated to anchor texts. In order to see the percentage of anchor texts per keyword, we have to choose "by count" on the "Anchor Text" Tab:

The data we see is actually far from suspicious, since a lot of the anchors are different variations of brand, urls, and other natural keywords like "click here".

So far no particular money keywords that were pushed too much are evident, but it is important to delve deeper into this by checking the anchor texts diversity from another angle and returning to the first graph in our BLP Report. Click on "More" situated in the right-hand upper corner and choose "Keyword":

In order to ensure we obtain reliable results, we have to classify and divide our anchor texts into four types: brand, compound (usually brand+money keyword), money and other.

This of course does take a lot of time to begin with, but you only have to do it once for a particular site and then this classification can be used in all reports.

Here is what the classification procedure looks like:

We have the following spread on completion:

Don’t forget to click on the "Relative" tab beneath the diagram because we want to operate with relative percentage figures.

We can see that the backlink profile of has the highest percentage of money keywords anchors. Of course it is natural to ask why we could not see this on the graph before. The reason for this is that the first graph gives us an overview of single anchor texts and in this case, we see the total results for all the keywords grouped by their types.

Still, it is doubtful that 45% of the money keywords in the backlink portfolio were the only reason was penalised by Google. More likely is that it was a mix of numerous such reasons which we shall attempt to find.

2.2 Country Popularity

The next step is to check whether the backlink portfolio of looks unnatural according to country popularity. This is very simple to check by clicking on the "Country Popularity" tab above the graph.

We can see that has links from many countries, but most of them come from the USA, which looks more than natural for a US website.

Maybe used link networks for its link building? If this is the case, then we are sure to notice that a large percentage of the links are from very few IPs.

2.3 IP Popularity

Check the IP popularity by clicking on "More" and then on IP Popularity:

The results here actually do not look so bad. A maximum of 3.9% of the links are from the same IP and from two others at 2%. This is not obvious evidence of link network use.

Please notice that this check happens on an IP based level, and every link network builder not completely out of his mind would spread his sites at least over various IPs, but maybe in the same Class-C network. Therefore we check the spread over Class-Networks as well:

Here we have similar results with maximum of 4,2% links coming from the same Class-C IP, what does not give a definitive answer whether there is any link network involved.

2.4 Follow / Nofollow Links

To look at the link status diversity in the backlink portfolio of, just click the "Link Status" tab:

The ratio between follow and nofollow links looks fairly natural, but we will check it once again in the next part when we compare with competitors.

2.5 Links with [LinkNotFound] status

Another interesting point we can see in the diagram is the high percentage of links with [LinkNotFound] status. This means that around 24% of all links were recently deleted. But why, what are those links, and who deleted them?

In order to answer these questions we have to export all links with [LinkNotFound] status and analyse them separately. In order to export them you have just to click on “[LinkNotFound]” in the column “Link Status” and you get the following table:

Now we simply click on „Export filtered data as CSV and get the file with all required URLs. We will have a deeper look at these URLs in our DTOX section of the case study (Point 2.8.1)

2.6 Weak links

By using the BLP once again, we will see what it can tell us about backlink power and trust of For this purpose, Christoph Cemper created a dedicated metric called ‘Power*Trust Metric’ that determines the quality of a particular page according to its strength and trustworthiness.

Some variations of this metric exist, for example Power*Trust dom metric for measuring Power and Trust on a domain level vs. page level. This means this metric determines the quality, especially the trustworthiness, of a whole domain and not a particular page.

Click first on the Power*Trust tab above the graph:

We can see that most of the links on are from very low quality pages (38% of total links from pages having Power*Trust equal to 0).

That might be one of the reasons for the Penguin 2.0 update penalty. It is important to check the second metric, namely Power*Trust dom metric, in order to see whether the backlink source domains are also as weak or whether its only particular subpages. Click on the "Power*Trust dom" tab and you will see the following graph:

This time it looks much better for; the quality of linking domains is spread naturally and is fairly good. Therefore, we assume that if was penalised by google for low-quality backlinking pages, there should have been another reason or reasons as well.

At this stage we already have two suspicious issues:

  1. a high percentage of money keywords as anchor texts and
  2. a high percentage of backlinks from pages with zero Power*Trust.

2.7 Competition Analysis

2.7.1 Defining competitors

In order to look at the previously described issues from another perspective, it is effective to compare with its competitors. But how do we find out who the direct competitors are? There are many ways in which to do this, but we decided to use the SEM Rush tool ( on this occasion.

Type into the search bar and choose US:


You should see the following screen:


From this, we can now see the top 5 organic competitors of

We will take four of these competitors. We are excluding as this has the least in common with keywords.

2.7.2 Quick Domain Compare

We return back to Christoph Cemper’s LinkResearchTools and start to analyse the competition with a QDC (Quick Domain Compare) Report. It helps to compare up to five domains based on the most important SEO metrics within seconds, so we enter the urls of competitors we found using SEM Rush and choose "Analyse Domains" as comparison mode:

From this screen, there does not seem to be anything too suspicious about, but if we look at two particular figures in more detail, it might become more interesting: "Backlinks to Domain" and at "Domain Popularity" (i.e. the number of linking root domains). These two figures might seem ok if we evaluate them separately, but the ratio between these two metrics gives us a so-called Sitewide Ratio or SWR. This Sitewide Ratio shows us to what extent the sitewide backlinks are present in the backlink portfolio of a website. The best way to analyse and compare the SWR for all competitors and would be by running a Juice Report. In order to do this, click on the button marked below the first column.

2.7.3 Juice Report

All the Domains are transferred to the report automatically and you do not have to type them in again. We will make a "Detail Analysis" and choose only "Link Ratio Metrics" to evaluate. This includes all kinds of sitewide ratios as well as a deeplink ratio, which will be also very interesting to analyse.


Our Juice report is ready in less than a minute and here are the results:

Houston, we have a problem! Look at this SWR ratio!

By comparing to other competitors, has 5 to 10 times higher SWR than all other evaluated sites. This looks very suspicious and was probably one of the main reasons for the Google penalty. But we should continue our research to ensure that we do not miss any other possible reasons for why the Google penalty was imposed.

The deeplink ratio or DLR of looks fine: it has more than 92% deeplinks (links to sub pages) of the total backlinks amount.

2.7.4 Competitor Landscape Analyzer

We continue our comparison between and its competitors, by using another type of report, CLA (Competitor Landscape Analyzer). We start with a detailed analysis report, put the urls of competitors together with, check and analyse the domain backlinks, switch on the maximum link boost in order to get maximum amount of links, activate the sitewide links filter and select main metrics (see the screen):

We will go through the already mentioned anomalies of one-by-one again using the outcome of the CLA report, so we can prove or disprove our hypothesis of possible penalty reasons by looking at competitors’ figures.

We begin by starting with a comparison of Power*Trust metric (tab Power*Trust):

We clearly see that has the most links from very low-quality pages with Power*Trust equal to zero. This graph confirms our hypothesis that those backlinks might have been one of a couple reasons for the google penalty.

We continue with the Power*Trust for the domain metric (tab Power*Trust dom):

As we have already seen in the BLP report, has fairly good linking domains. The next step will be to compare keyword types. For this purpose, we classified all the competitors’ keywords as we did previously with the keywords from

As you can see, although has a lot anchor texts with money keywords, the values it has per keyword type are almost equal to the average values of all competitors, which rather neglects our hypothesis that money keywords anchors were one of the reasons for the Google penalty.

In fact, we can see here, that 41% Money Keyword links are “natural” for this niche, and only this niche. In other niches that may be already way too high – that’s where the real value of the Competitive Landscape Analyzer becomes obvious – you understand the rules for the market segment you are analysing.

We would now like to look at whether has links from infected or unrelated websites. Click on "More" again and then choose "Theme"

This seems to be a very natural spread, along with the average value of competitors.

Our next step is to check whether has a natural ratio between follow & nofollow links compared to the one of competitors:

Looks pretty interesting, tvrage has much more nofollow links comparing to competitors, what means that 22% of nofollow links are not common amount in this particular industry, but from the other side we could argue that such ratio might look more natural.

Let’s have a short look at the domains where nofollow links are located:

The values of linking domains look very good, so we would not consider these 22% of nofollow links as something suspicious.

Our last evaluation in CLA report is to see how the deeplink ratio (tab "Deeplink Ratio") corresponds to the ratios of its competitors:

This evidence does not look particularly suspicious.

2.8 Link Detox

As a next step we use the latest tool, created by LinkResearchTools developers, "Link Detox".

We type in our URL and choose the classic mode in order to check the existing backlinks. In the latest version of “Link Detox Tool” you can also choose the theme of a domain in order to get more precise results. But in case you are not sure about it just choose the option “Automatically Detect”. This option works for all domain themes except gambling, adult and pharm; for these three themes you must make a choice.

The outcome we get shows that has a moderate risk with a risk score of 285. If we look a little bit further down, we also see that it has 61% of backlinks marked as healthy, which is relative good, but 4 links are marked toxic and therefore might be very harmful. Additionally, almost 40% of links are classified as suspicious.

Toxic and suspicious links:

As we can see from the screen, 4 links were chosen as toxic and have "deadly risk" according to the rule TOX1 (see column "Rule"). What does this rule tells us?

An example of a de-indexed website would be

This does not look too bad to begin with, although, there is now money anchor text, but there are many links in the blogroll on the right-hand side of the page. It seems as if they have the same pattern, so we pick one more to check:

This is exactly the same pattern with exactly the same outgoing links.

As all of these sites are located on a subdomain, all of these links have the same IP, which is very unnatural practice. If we calculate the number of these blogs in the blogroll we get 48; 48 blogs with the same schema and links. We could argue that this linkwheel was another important reason for the Penguin 2.0 update penalty.

It is possible to get rid of these harmful links even if you do not have an opportunity to delete them. There is one very nice feature in the DTOX report that helps you to do this. We just have to export the filtered data as Google Disavow Links by clicking on this menu above the table:

You should get a popup saying that you have to review the selected links carefully as some of them might be recognised as toxic by mistake. After you click on the "Yes, I understand and agree" button, the file download should begin automatically and you should get the following message:

You can just click on the link and go directly to the Google Disavow Tool and submit toxic links there. By doing so, you are suggesting that google does not take those links into consideration for determining the quality of your website. If you wish to disavow a sitewide link or any multiple number of links coming from the same domain, instead of having multiple urls on separate lines from the same domain, replace them instead with 1 line using the following format: and this will disavow all links coming from

2.8.1 [LinkNotFound] Links Detox

Let’s return shortly to our BLP report from point 2.5 and check the Quality of those 24% links that have [LinkNotFound] status. Surprisingly we can do it also with Link Detox tool.

And here is what we do:

We choose “Review mode” so the tool does not check existing backlinks but only provided by us URLs. Then we just upload a .csv file that we received when we exported the links from BLP report.

We can clearly see that deleted links were pretty risky and “toxic” We may argue that those 24% of links or at least some of them were deleted by administrators after the site got hit by Penguin 2.0. Update

2.9 Competitive Link Detox

Now we would like to use the latest tool developed by Christoph Cemper – Competitive Link Detox Tool. Competitive Link Detox lets you analyze up to 10 competitors and your own site for toxic and suspicious link ratios.

This helps you understand the toxic and suspicious link ratios of your competition.

You can use the result to find strong and healthy links of your competition that you don’t have – and go after them.

This tool is very easy to use; we just have to fill in the URL of and our already defined 4 main competitors:

And here is what you get after report is ready:

As you can see has actually second lowest risk score, but let’s dig deeper and look at Detox Rules diversity (just click on “Link Detox Rule” above the graph:

Here we can see that has the most suspicious links among all competitors according to rules SUSP14 & SUSP4.

SUSP14 links are links from domains having no Google Page Rank, but Google PageRank™ is more and more inaccurate and many webmasters don´t trust that metric anymore. So we would not consider these links as the penalty reason, but the website administrator should review them manually. We would rather concentrate on links classified according to SUSP4 rule.

If we look at examples of links classified as suspicious according to SUSP4 rule we probably see some already mentioned domains, but let’s check:

As you can see most of our link wheel domains from are listed here as well.

This proved the hypothesis that the links from penalized websites were one of the reason for Penguin 2.0 Update drop

2.10 Possible penalty reasons discovered thus far?

  1. 38% of total backlinks are from pages having a Power*Trust metric equal to zero
  2. An amazingly high Sitewide-Ratio (SWR) of 1566 (the ratio between domain-wide links and the number of linking domains to the domain) in comparison to the competitors (5 to 10 times more)
  3. Toxic links in the backlink portfolio
  1. Linkwheel with 48 links created on
  2. 24% links that were harmful are already deleted

Are these reasons enough though to get such a drop in SEO visibility and keyword rankings? Maybe, but intuition tells us that there must have been also other factors also involved.

It could be that Google penalised not only the main domain of, but also some of its subdomains and as a consequence, the whole website was down ranked?

3.0 Penalised subdomain?

In order to start analysing the subdomains of, we need to find them first.
You would think that this is an easy task, but it is not.

3.1 Finding subdomains

We recommend using the Wolfram Alpha tool, located at

Now click on "subdomains" in the middle section and you will see a table like this:

Wolfram Alpha helped us to find four subdomains of the website. While, and get less than 1% of daily traffic, the subdomain gets around 4.5%. Therefore we’ll focus on the subdomain and analyse it further.

3.2 Subdomain DTOX

We put together a DTOX report of to see whether we can find some proof of toxic links pointing to the subdomain.

The result is impressive: the DTOX report contains more than 5000 links.

Moreover, this time we have about a 3 times higher risk score of 858 – and only 23% healthy links. About 80% of the links pointing to are either toxic or suspicious, so this is definitely not a good sign.

It is beneficial for this research to see what those toxic links are:

From this, we can see a very interesting pattern or SEO strategy that used but was not particularly successful. This strategy is image linkbuilding. In the "Anchor Text" column there are only image links with money keyword "alt" tags. It is a well-known fact that google ranking algorithm evaluates image links too and perceives "alt" tags as anchor texts. However, what we can see here is a huge amount of low-quality linkbuilding on free blog services. Alt tags are, in 90% of cases, the same as the last url folder, which looks completely unnatural.
In order to check how many links come from each of the free blog services, type in each subdomain in the search bar of the first column.

We start with domain: has image links with money keyword alt tags on 102 subdomains of – a huge amount.

We continue further with the domain:

This time we have links on 64 subdomains of, again, a huge amount.

There are exactly 80 free blogs with links to on domain.

The last main domain with many subdomains that we see among the toxic links is

Here we have even more links, namely 93 from different subdomains of

It is useful to check a few examples of such links and have a look at one free blog per main domain.

We start with and look at

We see a simple page on a site with numerous, and probably automatic, generated short-articles with a completely unrelated YouTube video at the beginning and few photos, whereas the second photo links to and has an alt tag "bobby sherman mp3s"

The second example of a blog on domain would be:

Completely the same structure and even the same video.

An example from domain will look like this:

Again, the same video and image linking to

Another interesting fact is that all the pages we reviewed before were created in September 2010, meaning at some point in time received an enormous amount of new links from a very limited number of IPs.

We think it would be helpful to look at two examples of blogs from BlogSpot domain:

Here we have a slightly different pattern; there is no video present but one of the images links to the money keyword alt tag – to – again.

Another BlogSpot example would be


Again, we can see the similarities in pattern to the first BlogSpot example with its numerous pictures. In this case, the last one links to the subdomain

We assume that this kind of image Linkbuilding was one of the main reasons for the penalty that was imposed on

4.0 Summary of Findings and Recommendations

4.1 Summary

Possible problems that we found that could have led to the Penguin 2.0 penalty of

  • 38% of total backlinks are from pages having Power*Trust metric equal to zero
  • An amazingly high SWR of 1566 (ratio between domain wide links and the number of linking domains to the domain) in comparison to the competitors (5 to 10 times more)
  • 2% toxic links in backlink portfolio
  • Link wheel with 48 links created on
  • A massive spammy Blackhat and probably automated image link building activities to subdomain

4.2 Recommendations

In order to get do away with the Google penalty, there are some necessary actions that need to be taken:

  1. In order to decrease the amount of links coming from low-quality pages, might either delete a few existing ones manually or if this is not possible, submit them into disavow tool
  2. In order to decrease the high SWR, should delete some site wide links and focus more on content links from good quality pages
  3. In order to get rid of toxic links, it is necessary to export and submit them to the Google disavow tool as described before. Especially in the case of toxic subdomains links, we advise that the site-wide disavow with "" parameter be used.

Although we would not claim that poor link building was done by management of alone, it should now take care to implement the mentioned recommendations and perform another evaluation straight afterwards.

As the next step should do a clean-up, disavow what cannot be cleaned up and then submit a Google reconsideration request, because otherwise there’s little hope for recovery.

This case study was written by Hannes Hauser, CLRTP and was reviewed and approved by Christoph C. Cemper for publishing as Certification work for the Certified LRT Professional level.

A word from Christoph C. Cemper

Certified LRT Professional This analysis was conducted and post written by our LRT Associate Hannes Hauser, CLRTP.

Hannes showed proficiency in doing a backlink profile audit using the LinkResearchTools and pointed out critical issues to fix. Therefore I’m happy to certify Hannes as the latest Certfied LRT Professional by approving and publishing his research on our site.

This is Hannes’s the next step towards the Certified LRT Xpert level which is pre-requisite for the Certified LRT Agency certification. Both will qualify him to receive consulting leads from us. Our goal is to provide our community and clients a high quality service, and our certified experts are key to that.

I look forward to future work by Hannes on his way. I can recommend Hannes to work with you whenever you get a chance!

Michael Marshall, CLRTP

What do you think about this Penalty Analysis?

Let us know!

[notification style=”warning” font_size=”13px” closeable=”true”]

This analysis was created using our Superhero Plan and the new Competitive Link Detox tool extensively. This professional plan allows you to perform analysis for your or your competitor’s sites. For small businesses we now also have the new LRT starter plan available.

Check out our new Link Detox Pro plans.

Don’t miss our next LRT Associate Training. Get one of the last spots if you’re up for professional training.

Don’t forget to signup for our notification list to get informed on other Case Studies on Google Penguin updates.


Hannes Hauser

Hannes Hauser

Senior SEO Manager at E-Quadrat Communications GmbH
Hannes has been in the online marketing business since 2004, with the continuous focus on Search Engine Optimization for international clients. He currently works as a Senior SEO Manager at E-Quadrat Communications GmbH. based in Vienna and provides creative SEO solutions for numerous well-known market players.
Hannes Hauser
Hannes Hauser

Latest posts by Hannes Hauser (see all)




  1. Michael Levin on July 24, 2013 at 8:43 am

    Good Job, Hannes!

  2. Andri Yarusman on July 24, 2013 at 11:34 am

    Awesome … wonderful SEO report on Google Penguin 2.0 penalty reasons, i can learns so many things from this report … eventhough i still need to learn some terms used here … Big thanks bro!

  3. Vibhu Satpaul on July 29, 2013 at 11:20 am

    This was really helpful didnt know a subdomain can cause such a huge issue for sites.
    Good analysis. Would keep this in mind.

Leave a Comment