How Google Finds Unnatural Links to Your Website

Exposing a Link Selling Network in a Matter of Minutes

As we all know by now, a big part of the major blog link networks were de-indexed by Google shortly before the penguin update. I‘ll show you how you can uncover a big link network in a few minutes (without joining the network). We’ll use just the information that’s displayed on a networks front page and results from reports we got from Link Research Tools.

Why would you want to expose a popular link network?  Well, we’ve all heard about the emails being sent out by Google for “unnatural” links.  Google has told us not to worry, don’t panic, and that they are just being transparent. Do you really want to set back and relax? My suggestion is to be aware of your own link neighborhood and find out if you are associated with any spammy “unnatural” link networks.

Neighborhood Watch

Many years ago my mother told me two things that are also true in the digital world. I’m sure she wasn’t meaning to, but she did a good job at describing co-citation on the internet.

Stay out of bad neighborhoods!
You can get away with spending some time in bad neighborhoods. But if you spend a lot of time in bad neighborhoods, you are more likely to run into problems.

You are known by the company you keep.
Pick your friends carefully! You are not only known for who you are and what you do. You are judged even more by the people you surround yourself with.

It is very simple. If you have bad co-citations with a majority of backlinks from a spammy link network, Google will see the pattern very easily, and you will be penalized sooner or later.

Where do we start?

With all the noise about the downfall of the most popular networks, I decided to choose a popular network that hasn’t received much attention during this mass de-indexation.
This network has been around for a while and publicly boasts of having tens of thousands of customers, and thousands of blogs in a recent auction.

The front page has a section devoted to “success stories”, that’s where I found the seed that lead us to find the rest of the network. A quick look at the traffic stats reveals that this website tanked in rankings between March and April. Please note we didn’t want to out that network here fully public, but will give you specifics in the full report we provide for a tweet here.

How it’s done

We are going to take advantage of the fact that a link network leaves a massive network footprint that can be found by doing some forward and backwards link analysis.

Here is what we are going to do.

• Review the backlink profile of our seed site and look for 5-10 suspicious sites. These suspicious sites will be our suspect “link sellers”.
• Run the suspected “link sellers” (spam network sites) through Link Juice Thief, a common forward link analysis tool, and look for “link buyers”.
• Take the list of “link buyers” and run it through the Common Backlinks Tool.
• The Common Backlinks Tool will return a big list of “link sellers”.

Backlink Profiler – BLP

I ran the target ‘seed’ website through the Backlink Profiler tool to see if there were any suspicious links (Actually, I was counting on it).

This website had quite a few links. I decided to arrange the results by the number of indexed pages of the linking domain and do a manual inspection of the biggest sides, just like the quality reviewers from Google do. I picked 10 domains that had quite a lot of indexed pages. These pages were stuffed with spun content and off topic money keywords. You can get a detailed list of the links in the full report we provide for a tweet here.

Link Juice Thief – LJT

My next step was to find out if these websites had any outbound links in common. We can safely assume almost any link coming from one of these websites is there because an SEO arranged for it. I’m not saying that the owners of the websites knew about the spammy link practices. However, they are now stuck with these links.

I ran the websites through Link Juice Thief, a common outbound link analyzer, and returned a list of 11 potential “link buyers”. This list contains both legit companies and affiliate websites.

Common Backlinks Tool – CBLT

Ok – So we found some websites that we suspect are building backlinks from the same link network. Now what?

The next step was to check what links these websites have in common. I ran the list through the Common Backlinks Tool and found 815 backlinking domains! I certainly didn’t manually inspect every single one of them. But, the ones I did inspect were filled with spun text and money links.

Love letters from Google

In January – February of 2012, Google sent out 700,000 messages to webmasters. There was confusion about why the messages were sent out. Matt Cutts tried to clear things up by stating these 700K messages were not ALL concerning “unnatural links” (only 3%).

Nevertheless, it appears the confusion sparked a new interest for Google. Now that SEOs are starting to calm from fear of the Penguin, the Google Search Quality Team is busy sending out a new batch of messages directed toward the use of “unnatural” links! You have to love the “transparency” campaign being continuously promoted by Google.

With cute names like Panda, Penguin, or Google Search Quality Team, do we have anything to fear?   Well… If you live in the right neighborhood, then you are probably ok.  But if you’re in a bad neighborhood, it is better to disassociate yourself ASAP.

How Google can find Millions of “unnatural” Links

Out of the 700,000 messages Google sent in Jan-Feb, we were told 3% (21,000) were for unnatural links.  On July 19th, 2012, Google started sending another round of unnatural link warnings to countless webmasters. Just like the first round, this has caused a stir in the SEO community.

For arguments sake, let us assume that Google is sending out approximately 10,000 emails a month for “unnatural” links.

With an available tool and just one run we uncovered a link network and found 815 potential “link selling” websites. This was done by choosing 1 target ‘seed’, then comparing the link profile of 10 websites that are suspected “link sellers”.  Looking at the results, we could clearly see that these “link sellers” belonged to a bad neighborhood.

Of course, we don’t have exact numbers. However, we do know that the Google Search Quality Team is not a one man show or running just one report. Moreover, as much as a large corporation tells you that they are being “transparent”, it just means that there is a lot more going on in the background.

So, let’s go with these numbers…

We assume that they are sending out approx. 10,000 “unnatural link” emails every month and each of those websites would report back another 10 “bad websites”. Some more, some less, some not. Ten should be an average here, because frankly the really big link buyers buy hundreds and thousands of links, not ten.

Based on our case study here that would result in

10,000 x  815 potential “link selling”sites they find.

That’s over 8 Million potential Link Selling Websites every month!

Conclusion

With only a few websites known to play against the rules Google can find a lot more.

But not even Google can avoid false positives. Even SEOmoz received a notification about bad linking habits. This could be that it was because Rand offered up his corporate website as a target of a negative SEO experimen on the Traffic Planet forums. Regardless, we know that Google is going to continue sending out the messages and cleaning up spammy networks.

Today more than ever, I feel the need to point out the importance of natural links. You saw how easily we exposed a popular link network that has weak, non-trusted, and “unnatural” links.

I hope you take my mother’s advice and decide to stay out of bad neighborhoods. It’s a recipe for long term success both in SEO and in your private life.

I did not want to reveal the exposed network or the specific websites we were talking about here. But then to back all this up and encourage discussion you do need specifics. That’s why we decided to document the whole research process in a 12 page report plus all the excel files and link lists that you can download on our site for a tweet.

Final Conclusion – Link Detox or Headache?

We know the whole concept of forward and backward network analysis is not popularly known and this might sound a bit “over your head”. Although all the research can be done with existing tools we understand that not every webmaster has the time to dive into this. That’s why our new product is launching in the next days – Link Detox – and it promises to be the Alka-Seltzer for your unnatural link warning headache.