The latest major update by Google (Penguin) did a hatchet job on sites participating in link networks. As a result, many webmasters are hunting down websites that might appear to be a part of a link network and severing ties. However, those sites themselves could be a link network by connecting and optimizing anchor text links between each of the sites.
Keep in mind there are many sites that have optimized anchor text pushing link juice to their other domains, but have not been penalized. People link between their sites. It’s natural. It happens. However, Google’s dial is way high for link networks, and optimizing links between your sites could make you appear to be a network.
But before you freak out and do a full-blown investigation to figure out why your rankings took a nose dive, read Dr. Pete’s post. It’s critical to not overreact.
If you do suspect you’ve been smacked by Penguin, below is one of many processes of finding your own link network and how all of your sites are inter-connected.
For this exercise, you’ll need either a subscription to SEOMoz’s OpenSiteExplorer or MajesticSEO and Screaming Frog. The example uses MajesticSEO because its index is larger, but it is not necessary.
Let’s begin with even finding all of the sites you or your client might own:
1. Plug main each domain you own into ReverseInternet.com:
2. Look at the other sites on the same IP and nameserver and that share the same Google Analytics and AdSense code:
3. Once you have a list of all sites that share the same IP, nameserver, Google Analytics, and AdSense codes, find the ones you or your client owns. Keep on drilling down to create a comprehensive list of all sites potentially owned by either you or your client. Keep in mind other sites on the same IP or nameservers or have the same Google Analytics and AdSense codes might not be owned by the same person.
Now you need to see how they’re connected!
The next step is to pull the backlinks from MajesticSEO.com for all of the sites, regardless of whether you think you own them into one Excel sheet. If the site does not have links pointing to it, use MajesticSEO’s historic index.
We want to see how the sites have received links from the same company currently and historically. This view helps us understand how Google might have seen them as interconnected and have perceived the sites as a link network. (Note: If you have less than five small sites, you may be able to cut ahead and start with using Screaming Frog.)
1. Log into MajesticSEO (yes, you need to sign up).
2. Explore the domain:
3. If the site does not have current backlinks, use the historic index to see how it was once getting links:
4. Create a report:
5. Create a domain-level report so you can see as many links pointing to your site as possible:
6. Go to your reports page:
7. Select the report you just made and click the download:
8. Now prepare your download:
9. Go to your downloads page:
10. Now you can FINALLY download your report, unzip it, and open in Excel.
11. Repeat this process until you have all of your links in one Excel sheet, stacked on top of each other and only one header.
You could have 1,000,000,000 plus links in your spreadsheet and performing calculations on 1,000,000,000 is without a doubt going to break Excel and leave you incredibly frustrated. Additionally, we only need if and how the sites are connected, not how many links pointing to them.
1. Once you have all of the websites the company owns, we need to clean the URLs so we can remove duplicate links (from source to target URLs).
2. Copy and paste the target and source URLs into new columns as seen in this example:
3. Select the clean target and source URL columns by selecting the source column holding shift then select the target column so it looks like this:
4. Use the find / replace function to remove http://, https://, www. and every directory by finding finding /* and replacing with nothing:
and
5. Now you should have two columns that have a clean target and source domains:
Next you need to remove duplicates. Select all of your data and go to the data tab in Excel, then select the remove duplicates button:
6. Unselect all columns then only select your clean target and source columns:
Now remove duplicates.
7. Now we need to see which sources are linking to our target URLs to find how we’re interlinking. Use the countif() function, using the clean target URL as the range and clean source URL as the criteria:
8. Let’s filter our countif() statements by everything over 0:
Often times it’s a lot easier to see how we’re interlinking instead of looking at a bunch of rows. This will make it easier to spot clusters of sites interlinking.
1. Take the clean target and source URLs from the above spreadsheet and paste them into a Google Docs spreadsheet:
2. Let’s go to Google Docs and create a Google Fusion Table:
3. Select the spreadsheet you just made:
4. Now you should have your spreadsheet in Google Fusion Tables. Go to the experiment menu and select the network graph option:
5. VOILÀ! Finally, you can see how all of your sites are interlinked:
While we can use MajesticSEO or OpenSiteExplorer to find links between our sites, some of these links might have changed, and we want to find the most updated links in our network, so let’s use Screaming Frog.
Why not use Screaming Frog to begin with? Well, if you have 500 sites, running Screaming Frog on all of them could take quite a while so we would use MajesticSEO or OpenSiteExplorer to knock through them more quickly (especially, if you’re talking about very large sites). If, however, you have just four or five sites (as found by ReverseInternet.com), you can forgo the above steps and jump into Screaming Frog.
1. Go to the Screaming Frog application and run a report on each domain:
2. After the report has finished, go to the External tab, then Export > All Links:
3. Now we need to copy and clean the source and destination URLs to see the exact places and anchor text we’re using to link between our sites:
4. Just like before, we need to see where our target URL is also the source URL. This will help pinpoint where we’re interlinking. Just as before, use the countif() function.
5. Let’s filter by countif() cells over 0, then look at the anchor text to see if we might be over-optimizing:
6. Now, let’s remove those over-optimized links if necessary!
You might be saying, “It’s completely reasonable to link between your sites. The GAP does it!” Guess what buddy? You’re not the GAP. Chances are, if you’ve been in link networks and have one of your own, you’re under Google’s microscope. Before you fire off that reconsideration request, you need to do EVERYTHING that could potentially be against their guidelines. Having a bunch of sites interlinking with optimized anchor text is one of them.
Okay, okay, there are legitimate reasons for linking between your sites. One is your other site offers a different product in another country. It would be a bad user experience to remove the links for people that want your products in their country. You might want to no-follow those links so there is no question that you’re not “inflating your pagerank.”
At the end of the day, you want to do your homework to make sure you’re covering all bases before filing a reconsideration request. Removing a few links, in my opinion, isn’t enough. But then again, I’m an Excel data nerd. 🙂
Feel free to hit me up on Twitter for any questions: @ethanlyon and Google+ or at SEER Interactive.
