When Should you File a Disavow?

[JoyHawkins]

This is a question that comes up a lot. When is it a good idea to file a disavow? Is it risky because you might actually remove link juice from links that are helping you (even though they're spammy)?

Marie Haynes just published a guide on this and says you likely don't need to file a disavow for:

• Spammy sites such as low quality image aggregators that link out to millions of sites on the web automatically.
• Directories that have content scraped from the Yellow Pages, DMoz or other valid directories.
• Many sites that are labelled by a link auditing tool as “toxic”.

You should file a disavow for:

• Paid links
• Paid articles containing links
• Paying a journalist to slip a link into an article
• Paying an SEO company to publish articles on sites with which they have partnerships to provide content, especially if this is done on a large scale
• Publishing articles containing links, on sites that have very little editorial process. This includes a lot of guest posting. Not all guest posting is bad, but if you are doing it on a large scale, we believe that these links have the potential to be seen as unnatural by Google
• Offering bloggers free product to review, and getting a link in return
• Excessive use of reciprocal linking beyond what is valuable to actual users
• Creating a tool or widget that requires the site owner to link back to you in order to use it
• Large scale use of directories that were only created for SEO (i.e. no human would ever read them)
• Links embedded in the footer of html themes that you have distributed for link building
• An influx of “negative SEO attempt” links using your keywords as anchors, or using adult terms as anchors
• Any other type of link that would make you really nervous if you had to explain to the webspam team how you obtained that link

What has been your experience with disavowing?

 

[Marie Haynes]

Thanks Joy! At the risk of sounding super Canadian, I apologize for how long it took me to write that article.

I'm happy to give my thoughts if anyone has questions about the current state of the disavow tool and whether it helps to use it.

 

[JoyHawkins]

Apologizing for not being able to provide immediate free advice to all of us? Yes, very Canadian of you.

 

[JoshuaMackens]

Hey Marie,

Love the newsletter.

I agree with a few of these points but a lot of the top half just makes it sound like Google asking you to police yourself. There's no way they can tell a paid editorial from a non paid, correct? I just don't think I would disavow these even if I did pay for editorials, which is almost impossible these days anyway.

What's your opinion?

 

[Marie Haynes]

Hi Joshua, thanks for chiming in!

There's no way they can tell a paid editorial from a non paid, correct?

They most definitely do have ways to determine whether content is paid or not.

I started my career as an SEO in removing manual penalties. Google would often send 2-3 example links along with the penalty and they were quite good at finding paid links.

When I audit links, it is so easy to see patterns that imply paid linking. For example:

• If one person links to you with a keyword, that's not a big deal. If 50 people link to you using the same keyword, that smells like a pattern.
• If you've got a lot of links from sites that Google knows sell links, that's pretty easy to find.
• Some links are so easy to detect as paid. If a site wrote about Nike shoes, and linked to your website, "buy-nike-shoes-today dot com" rather than Nike itself, that smells like paid.
• If you've got multiple links that smell like SEO such as links from low quality directories, freehost blogs, article marketing on sites no one ever reads, etc. then this can cause Google to distrust all of your links.

The point that I'm trying to make, is that when manually reviewing links it is really easy, when you see links in aggregate, whether the company was paying for links.

Remember as well that if Google sees the odd paid link, they're not likely to care. It's when something is done on a big enough scale to have a strong impact on rankings that either a manual action or an algorithmic one could happen.

 

[djbaxter]

See also Google: Don't disavow bad links unless you are responsible for those links

 

[CraigJMount]

This guide is GREAT. Big fan.

@Marie Haynes- Out of curiosity, how many attempts do you make to contact the site that owns the link prior to submitting the Disavow File?

 

[Marie Haynes]

Thanks for the kind words Craig.

We only contact site owners when we're trying to remove a manual action. John Mueller has said several times in the past that when Google is looking at links algorithmically, disavowed links are just ignored, and there would be no added benefit to actually get the link removed.

Of course, if they are links that my client or their agency controls such as their own PBN, or microsites they've set up, it's best to remove those links.

But, in doing outreach, the majority of the time the emails get ignored. I had a very low success rate in getting links removed.

This is a few years old. The Penguin algorithm has changed since I wrote this, but the part on disavowing vs removing should still be quite applicable: Penguin Recovery: Should You Be Removing Links or Just Disavowing? - Search Engine Watch

 

[JoshuaMackens]

Hi Joshua, thanks for chiming in!



They most definitely do have ways to determine whether content is paid or not.

I started my career as an SEO in removing manual penalties. Google would often send 2-3 example links along with the penalty and they were quite good at finding paid links.

When I audit links, it is so easy to see patterns that imply paid linking. For example:

• If one person links to you with a keyword, that's not a big deal. If 50 people link to you using the same keyword, that smells like a pattern.
• If you've got a lot of links from sites that Google knows sell links, that's pretty easy to find.
• Some links are so easy to detect as paid. If a site wrote about Nike shoes, and linked to your website, "buy-nike-shoes-today dot com" rather than Nike itself, that smells like paid.
• If you've got multiple links that smell like SEO such as links from low quality directories, freehost blogs, article marketing on sites no one ever reads, etc. then this can cause Google to distrust all of your links.

The point that I'm trying to make, is that when manually reviewing links it is really easy, when you see links in aggregate, whether the company was paying for links.

Remember as well that if Google sees the odd paid link, they're not likely to care. It's when something is done on a big enough scale to have a strong impact on rankings that either a manual action or an algorithmic one could happen.

It sounds like you mean PBN's. I agree. A PBN site is pretty easy to tell right off the bat.

I mean more along the lines of a website that people actually visit and has use. They write high quality articles. If theoretically you could pay for placement on such a site, how would Google be able to tell without the website writing "sponsored content"?

 

[Marie Haynes]

I don't know the exact answer to this question, but I can tell you that it is usually easy for us to determine on manual review whether there are a lot of self made links. I am positive that Google's engineers can algorithmically determine this as well.

On a manual review, here are the types of things that make us suspicious that a site has been overdoing link building in the form of guest posting:

-Similar use of author bios across many sites
-Lots of articles with keyword anchors where the link doesn't make sense. For example, if a site references Nike shoes, and links out with the anchor "Nike shoes" to any site other than Nike, it's probably a self made link. A couple like that won't hurt, but a bunch looks like a pattern.
-Lots of links from sites that advertise guest posting, or that they will accept content for publishing and in return get a link.
-Links on sites that Google has flagged as potential sellers of links.
-Lots of links on sites that have never sent any traffic

There's a patent that Bill Slawski wrote about (https://gofishdigital.com/did-the-groundhog-update-just-take-place-at-google/) that mentions that if a particular resource has a lot of links that are never clicked on, it could be considered a spam resource. This is likely considered amongst many other factors.

Gary Illyes said at Pubcon Austin that Google knows which parts of authoritative websites to trust and which to ignore. He said that they know to ignore links from the contributor's section of Forbes.

I suppose it is in theory, possible that you can get paid links that evade these patterns. Again, if you get a couple of paid links that Google doesn't detect as paid, that could help. But, if you cross the line to the point where Google starts to question the quality of your links and whether they are paid/self made or whether they are true recommendations of your content or brand, then you may find that even your good links are treated with less trust and don't count for much.

 

[JoshuaMackens]

I don't know the exact answer to this question, but I can tell you that it is usually easy for us to determine on manual review whether there are a lot of self made links. I am positive that Google's engineers can algorithmically determine this as well.

On a manual review, here are the types of things that make us suspicious that a site has been overdoing link building in the form of guest posting:

-Similar use of author bios across many sites
-Lots of articles with keyword anchors where the link doesn't make sense. For example, if a site references Nike shoes, and links out with the anchor "Nike shoes" to any site other than Nike, it's probably a self made link. A couple like that won't hurt, but a bunch looks like a pattern.
-Lots of links from sites that advertise guest posting, or that they will accept content for publishing and in return get a link.
-Links on sites that Google has flagged as potential sellers of links.
-Lots of links on sites that have never sent any traffic

There's a patent that Bill Slawski wrote about (https://gofishdigital.com/did-the-groundhog-update-just-take-place-at-google/) that mentions that if a particular resource has a lot of links that are never clicked on, it could be considered a spam resource. This is likely considered amongst many other factors.

Gary Illyes said at Pubcon Austin that Google knows which parts of authoritative websites to trust and which to ignore. He said that they know to ignore links from the contributor's section of Forbes.

I suppose it is in theory, possible that you can get paid links that evade these patterns. Again, if you get a couple of paid links that Google doesn't detect as paid, that could help. But, if you cross the line to the point where Google starts to question the quality of your links and whether they are paid/self made or whether they are true recommendations of your content or brand, then you may find that even your good links are treated with less trust and don't count for much.

Thanks for all the info Mary! You're a true resource here!

 

[adammaxum]

I would only disavow if..

1. the anchor text profile was ruined by a majority of spammy websites

or

2. there was clearly a penalty of some kind involved.

I also disagree with disavowing most of these link types below...I've seen many top ranking websites using a variety of these strategies and if they disavowed their link profile they would cripple their rankings.

Of course, every link is its own discussion based on a variety of factors.

-----

• Paid links - how do you know if a link has been paid for unless it's blatantly obvious? The majority of links on sites are likely paid for in some shape or form all the way up tp the biggest publishers online.
• Paid articles containing links -- how do you know if the article was paid for? How does Google assess whether an article was paid for is the more important question.
• Paying a journalist to slip a link into an article -- as long as the journalist is posting on a high quality website relevant to your business, I see no issue here. This happens all the time whether through direct payment or influence or connection.
• Paying an SEO company to publish articles on sites with which they have partnerships to provide content, especially if this is done on a large scale -- most times the SEO company is posting content on crap blogs so this is a bad strategy, but if the SEO has relationships with high quality publishers and there is no direct connection between the two parties, I wouldn't disavow this type of link.
• Publishing articles containing links, on sites that have very little editorial process. This includes a lot of guest posting. Not all guest posting is bad, but if you are doing it on a large scale, we believe that these links have the potential to be seen as unnatural by Google Again, hard to distinguish how much editorial process is involved unless you actually run through the process. It varies so widely that disavowing these types of links would be incredibly time consuming and honestly, as long as the website has a proper quality score, there is very little risk.
• Offering bloggers free product to review, and getting a link in return This is common practice...as long as the link is relevant to the bloggers audience I see no reason to disavow this type of link.

 

[PaulSteinbrueck]

This article in Search Engine Journal reports what Google recently said about disavow.

And last time I looked, I couldn't find the disavow tool in Google Search Console anymore. Does anyone else see it?

 

[CraigJMount]

Google Search Console

 

[PaulSteinbrueck]

Google Search Console

Yes, it still exists in the old GSC. What I meant is I don't see it anywhere in the new GSC. That seems to imply they don't really want us to use it.

 

[MirageLimo]

We disavowed back in 2014 and it probably helped. We had hired some seo's that might not have been the most reputable. Then the new seo found their links from sites that were hacked, etc. and disavowed them. Nowadays we hear Google knows which links to ignore. Disavowing only alerts them to new ones they may have missed.