Link Removal - Link Analysis Steps:
Ever since Google stared sending out warnings to webmasters, the release of the Penguin algorithm update, and the introduction of the infamous Disavow tool, many SEO’s and business owners who were doing great, suddenly had to scramble to recover. It became very clear that Google was not going to allow any link schemes and networks interfere with their search results. “Spammy” or lower quality sites that existed just to basically house low-quality content were de-indexed or lost a lot of traffic overall. Links coming from these sites suddenly were worthless, at best, and harmful to your rankings, at worst.
These updates have helped improve results and have greatly reduced spam and lower-quality content. Our team at DigitalNext compiled a checklist that we use, not only on how to choose a site to build a link, but how to do a link audit.
The important thing to note is that this criteria should not be taken literally and it does require a dose of common sense, along with the ability to think out of the box. We recognized early on, through internal debates, that we could not automate this process. A human touch was needed to ensure we did not remove a good link or exclude a good and relevant site to build a link on.
We are sharing this criteria to not only help webmasters, but to also get the opinions of SEO’s out there to further improve our processes.
1. Is the site related or not? The common thought process is that if it's not related then you should disavow or not build a link from a hypothetical site. For example, a website that focuses on “puppy training” should not be selected to get a link for someone who sells “digital microscopes.” However, one might argue by saying that a blog post will be written on digital microscopes being used to spot fleas on a dog. Remember to think and be human. Would you build this link for your website or will it pass a Google manual reviewer?
2. The page has more than 200 outgoing links and all the external links are do-follow. This could be a sign of a spammy blog, directory or other website. There are exceptions to this rule (i.e. webmaster made a mistake or did not know) but this is usually a deal breaker for us.
3. Sites that generate automatic links are also marked as bad. Have you ever gone to a site and noticed that a bunch of random words are highlighted with ads and other internal links? Sites that use these kinds of techniques usually add no value to the end user.
See this example:
Here we have links for:
http://rugstorene.co.uk/ on all below mentioned pages, mr-search.de/tag_available/. mr-search.de/tag_budget/, mr-search.de/tag_contemporary/, mr-search.de/tag_hall, mr-search.de/tag_high/, mr-search.de/tag_including/, mr-search.de/tag_indian/, mr-search.de/tag_large/, mr-search.de/tag_leather/, mr-search.de/tag_modern/, mr-search.de/tag_natural/, mr-search.de/tag_offering/, mr-search.de/tag_oriental/, mr-search.de/tag_quality/, mr-search.de/tag_rugs/, mr-search.de/tag_rugstore/, mr-search.de/tag_selection/, mr-search.de/tag_shaggy/, mr-search.de/tag_sheepskin/, mr-search.de/tag_specialist/, mr-search.de/tag_store/, mr-search.de/tag_traditional/, mr-search.de/tag_wool.
4. Domains that have the following words: SEO, sex, porn, poker, blackjack. Online gambling and SEO sites were built like crazy but they were really just useless sites. Although there are some good sites that rank for the above phrases (i.e. Craigslist ranks for sex keywords and it’s a relevant and authoritive site) usually the sites are of low quality. One additional test we like to run, aside from the ones listed here, is to use semrush.com to see if the site in question ranks for any other keywords just to be sure.
5. Domain’s Theme looks suspicious : Avoid any site your going to build a link to seems like it may have malicious intents. Some examples are torrent sites (downloading movies / audio), live video streaming sites (movies that are copywrighted, television shows).
6. The link comes from a page that has more than 10 backlinks from 1 C-Class IP address. This could mean that it is a link coming from a possible spammy or unnatural link network. Now, there are exceptions to this rule but overall sites like this will require a manual review by our team. Tools that check for this are tools.seochat.com/tools/class-c-checker-tool/ and www.ip-report.com/.
This can again be detected using the above-mentioned tool or link detox.
Links coming from 192.99.35
7. The detected links are coming from the same domain name registrant as other linking domains. This means that the links are possibly from a link network. Once again, this is something that triggers a manual review. We obviously want to avoid any link networks as Google hunts these down with ferocity. The only tool that we know of (as of today’s article date) that can easily check this is http://www.linkdetox.com/.
8. The detected links are coming from the same IP address as other linking domains. This means that the links are possibly from a link network. Tools that check for this are tools.seochat.com/tools/class-c-checker-tool/ and www.ip-report.com/
In the below case, we are getting links through 3 sites which have the same IP - 220.127.116.11
So we just check 192.99.35
9. Manually analyze the link growth of every domain. If the last 4 months of links to the website are significantly less than in the last 12 and 24 months, it may mean that the domain is abandoned by the current owner. (Many expired domain link networks are, especially after the Google Penguin / Panda updates).
10. The domain name has the same Google Analytics code as other linking domains. If multiple domains have the same Google IDs, then it is a strong signal of a possible link network. Once again, this requires a human reviewer. This can be checked using http://www.linkdetox.com/. Its reports give us a parameter 'GA-ID'. We know of no other tool.
11. The domain name has the same Google Adsense Publisher ID as other linking domains. If multiple domains have the same Google IDs, then it is a strong signal of a possible link network. This can be checked using http://www.linkdetox.com/ its reports give us a parameter 'Adsense-ID'. We know of no other tool.
12. The links are coming from common web link directory footprints. These were often just setup to artificially inflate link popularities and/or sell links from. These can even be Article sites.
Here are some examples or red flags:
Powered by PHP Link Directory, Powered by PHPLD, Powered by WSN Links, powered by PHP Weby, Powered by cpLinks, Powered by cpDynaLinks, Directory Script by PHP Link Directory, Powered by BosDirectory, Powered by Link manager LinkMan, Powered by Gossamer Links, Powered by K-Links, Powered by In-Link, Powered by eSyndiCat Directory Software, Powered by: qlWebDS Pro, Powered by Article Dashboard, ArticleMs Skin Powered by ArticleMS from ArticleTrader
13. Auto-spun posts in articles and newspaper advertorials. This is difficult to detect sometimes but we use Copyscape.com to help us. Many times, when reading the content, it will just not read right (native speakers are best for this), and we trust our common sense for these decisions.
14. We ensure the links actually rank for keyword phrases using 2 separate tools to ensure accuracy. We use SEMRush.com, Spyfu.com or Rankwatch.com to check this data. It stands to reason that if a website has been around for a long time and has many pages but ranks for nothing, it’s because Google probably has the site penalized. This, once again, needs to be taken with a grain of salt. It could be that business has not done any SEO but is a great website. A site that does not rank requires more digging though.
15. Domain not cached. If a website does not rank for its title tag, then it’s a big red flag and requires further review.
16. Link coming from Article Directory and Domain PR is <=1. Article directories were great several years ago for link building but not so anymore. We typically avoid article directories unless they are really authoritive.
17. Link coming from Web directory and Domain PR is <=1. Directories were also great for link building before but no longer. We stay away from most directories unless they are authoritative.
18. Link from a Bookmarking website with Pr <=1. I’ve personally never liked social bookmarking sites. We avoid these and use them sparingly and only if they are very authoritative.
19. Link of a Buying Keyword as Anchor text is more than 5%
If we take a site: http://www.roid-shop.com, and check the backlink anchor phrases through Ahrefs.com. The maximum link building has been done on a buying keyword.
20. Link is coming from a Domain whose age is more than 3 yrs and Domain PR = N/A or 0. This is once again a red flag but does require further digging.
21. Links through Multiple Anchor Text with the same URL on a referring page and also getting multiple links from the same page.
Targeting the same page with multiple anchor text in the author bio section through anchors 'buy cannabis seeds' and 'cheap cannabis seeds.'
From the author bio section, 3 links have been given on anchor text 'best online auction sites', 'classic auction' and 'best reverse auction sites'.
Please feel free to share you thoughts on what can be added to this list or if you suggest any other tools. If you have any questions, please let me know as well. Also, remember that with all of these points, the human element must be present when making decisions.