Does Disavowing Links Work in 2019? 

Is there any benefit to using Google’s disavow tool if you do not have a manual action? In this article, we will discuss some case studies of sites that appear to have benefitted from disavowing a large number of links previously made for SEO reasons. We will also share some cases in which site owners saw drops in traffic that we feel are related to not filing a disavow. 

If your site has been actively involved in link building, you may want to consider reviewing these links thoroughly. Ask yourself whether the links you were getting truly were recommendations for your site. Or, were the majority of links built, primarily there for the purpose of improving Google rankings? There is a possibility that this history of building links to try and improve rankings could be causing your site to be considered less trustworthy by Google. If so, filing a thorough disavow could help remove this distrust and your site could see improved rankings.

What is the disavow tool?

Google’s disavow tool is something that site owners can use to ask Google to ignore some links that are pointing to your website. Each link that points to a site carries signals with it. Some of those signals can be good. For example, if an authoritative and trustworthy site recommends your content via a followed link, that usually sends Google good signals. But, other links can send Google bad signals if they are links that contradict Google’s recommendations. 

If you have enough low quality link signals, this can theoretically cause the site to have difficulty ranking at its full potential on Google searches. In this article we will discuss what Google has recommended in regards to disavowing. We’ll give some evidence to show that unnatural links still can cause problems for a site algorithmically, even in the age of Penguin 4.0. You will read some case studies describing sites that have filed a disavow and have subsequently seen improvements. You’ll also see some sites that did not file a disavow, against our recommendations, and subsequently have seen losses in traffic.

Our goal is to help you determine whether your site could benefit from filing a disavow. We are excited about some of the results that we have been seeing recently after disavowing. While we can’t promise that every site that disavows will see improvements, it is our opinion that if a site has a history of making links primarily for the sake of improving rankings, those links can be hindering your ability to rank. Disavowing those links can help.

Just starting to learn about the disavow tool?

If you are new to understanding the disavow tool we would recommend the following resources:

Your Start-to-Finish Guide to using Google’s disavow tool - This post is almost five years old now, but still quite relevant and helpful.

Disavowing in 2019 and beyond - Written in February of 2019, this article describes in great detail our thoughts on why unnatural links can hurt a site algorithmically even in the age of Penguin 4.0.

Google: Disavowing Bad Links May Help Google Algorithmically Trust Your Links - Information from Search Engine Round Table on disavowing in 2019

7 Things You May Not Know About Google’s Disavow Tool - Did you know there is a size limit to the disavow file? This document contains some interesting things you may not know about using this tool.

Do we need to disavow since Penguin 4.0? Doesn’t Google just ignore unnatural links now?

In September of 2016, when Google came out with version Penguin 4.0, the general consensus amongst SEOs was that unnatural links could now only hurt a site if the site received a manual action from Google’s webspam team. Several Google employees made statements on social media to indicate that Penguin would no longer be demoting spam algorithmically. Rather, they felt confident in their abilities to ignore links that were made just for the purpose of manipulating Google rankings.

Gary Illyes on devaluing spam

For a while, following the release of Penguin 4.0, MHC only recommended filing a disavow for sites that had received a manual action for unnatural links. We felt that Google would simply ignore unnatural links in their algorithmic calculations. 

It turns out this is not completely correct.

In February of 2019, Google’s John Mueller shared some interesting news with us, telling us that there can be a benefit for some sites to disavow unnatural links, even if they did not have a manual action.

When asked whether unnatural links could hurt a site algorithmically, John said:

john mueller smiling

“That can definitely be the case. So it’s something where our algorithms when we look at it and they see, oh, there are a bunch of really bad links here. Then maybe they’ll be a bit more cautious with regards to the links in general for the website. So if you clean that up, then the algorithms look at it and say, oh, there’s– there’s kind of– it’s OK. It’s not bad.” 

But wait? Google told us that they are ignoring unnatural links, right? 

Actually, what Google has told us is that the Penguin algorithm ignores unnatural links. But...there are likely other algorithms that Google can use to assess link quality. This is what we currently believe at MHC in regards to how Google handles links:

  • Ultra spammy links like comment spam, forum link blasts, spammy image directory links, etc. are likely completely removed from all of Google’s link calculations.
  • If a site has a small number of links that don’t line up with Google’s guidelines re link schemes, those links are likely ignored by Google’s link quality algorithms. 
  • If a site has enough links that don’t line up with Google’s guidelines, then it can trigger some sort of distrust which has the potential to strongly impact your ability to rank for YMYL queries.

We spoke in our article on the importance of scientific consensus in medical sites  about how important trust is in Google’s algorithms. We do believe that Google has many factors that they multiply together in some form, to get a quality score for each website. We also believe that one, or perhaps several of those factors deal with the level of trust that Google has in your site. If one of those trust factors is lacking, then this could result in your quality score being multiplied by a decimal, which has the potential to greatly reduce your score. 

The actual calculations Google performs are likely more complicated than simply multiplying scores together. If you would like to read more about the types of calculations that could be used in determining trust, we would encourage you to read this Stanford paper published in 2004 called Combating Web Spam with TrustRank. While TrustRank is not an actual Google patent, it shows the types of calculations that could be done to determine trust for a page on the web.

Threshold Trust Property

The point in writing about this is not to explain exactly how Google determines trust, but rather to show that it can be done programmatically. We believe that there are many factors that can contribute to Google’s assessment of trust. We have outlined several described in the Quality Raters’ Guidelines in this article on the September 27 core Google update. One of the most important components of trust that Google can measure is link quality. 

August 22-29, 2019 (and likely September 13) were possibly big link related Google updates

Each week, the team at MHC does what we call, “Algo checks”. Each of our auditors review the analytics for every client they have had, past and present, and comments on any significant changes in Google organic traffic trajectories. We look at seasonality and other factors to try and determine whether there is an obvious reason for traffic to go either up or down. 

A pattern that we noticed recently was that quite a few of our clients were seeing increases in traffic, and a small number were seeing decreases, all between the time of August 22-29, 2019. The interesting thing that we noted though was that each of the sites that saw a change significant enough for us to report on had some type of issue with unnatural links.

We believe that Google made changes to their algorithm at this time in how they assess link trust. While there may be other elements of trust that were reassessed by Google at this time, if you were hit at this time, we would strongly recommend having a thorough look at the quality of your link profile. Near the end of this article you will read our recommendations on what types of links to look for and how to determine whether you may benefit from filing a disavow.

Could this link related change be related to Google’s new rel=nofollow news?

Google told us recently that they are making some changes to how they handle nofollowed links. You can read more here on our take on this. Some have speculated that the ranking changes we saw with the August 22-29 updates could be related to this change. We do not think this is the case.

It is our understanding that the main change Google made is that they now can choose whether or not to pass PageRank through a link that is nofollowed. This may mean that some sites that have a large number of nofollowed links from truly authoritative sites could start to see some benefit from those links. However, it really shouldn’t cause sites to see significant drops, unless perhaps a competitor benefited from big gains.

You will see in the case studies below that some sites that we identified as having unnatural links, that did not file a disavow saw some significant declines in this time period.

Disavow case studies

Medical information site

This site is a YMYL site that provides users with important health information. The site had not seen drops in conjunction with core updates, but has struggled to rank for quite some time now.

Disavow work done

The MHC team audited this site’s links and identified over 700 domains which contained links that we felt Google would consider unnatural links. This site had a large number of links that we classified as “article for links”. Many of these links were what many SEO’s sell under the guise of content marketing. While we believe that many aspects of content marketing are good, if you are sending content to publishers with the main intention of gathering links that pass PageRank, there is a strong possibility that these links are doing more harm than good.

Here is a section of our link auditing spreadsheet for this site. Yes...we still audit links manually. The “CLASS” on every link below has been manually decided by an MHC auditor. It takes forever, but in our opinion, it is by far the best way to be accurate.

MHC link audit sheet

We also disavowed some spammy links that every site gets. While these are unlikely to cause ranking suppressions, there is no harm in throwing them in a disavow file.

For this site, we recommended disavowing over 50% of their link profile. The site owner agreed despite the fact that they had spent a large amount of time and money in building those links. 

Their traffic increases are spectacular:

Increase example graph

It is often hard to prove that filing a disavow was the sole reason for a site to see traffic increases. We believe that in the above case, the disavow truly was what propelled this site to better rankings. While we did recommend a number of site quality improvements, the site owner has not yet implemented those changes. The only change made to this site was filing an extremely thorough and aggressive disavow. 

Service area business

This site is another interesting case study because all that they contracted us to do is audit their links and file a disavow. We have not advised on additional quality improvements for this website.

We recommended disavowing about 800 domains which represented almost half of their link profile. A large number of these were links that we classified again, as “article for links”.  These were links that were in articles that their former SEO company had shopped out to websites. While some would try to argue that these were editorial mentions, in reality these articles would not have been written or published if it weren’t for the links they contained. We also disavowed a number of low quality directory links and spammy links as well just to be safe. 

This site started to see improvements within a couple of months of filing a disavow. They see improvements with most core quality updates, and other updates that we at MHC think were link related.

disavow filed increase example

Informational site

This site had us do a site quality review for them. We noted in the link section of their review that there were a large number of links from guest posts that likely Google would see as unnatural. We did not do a full link by link audit for this site as they had a list of the links that they had created from guest posting. This site filed their own disavow, and saw increases starting August 22, 2019.

informational site disavow

We have several other sites in our profile that are showing nice increases following filing a disavow. We haven’t shared them all as several are sites for which we have signed an NDA and also, quite a few have implemented many changes to their site that could help improve Google’s assessment of quality. In those cases it can be challenging to say that the increases seen are completely caused by the disavow work.

Sites that saw drops 

The most interesting part of our investigation was when we looked at sites that we had recommended disavowing a large number of domains, who did not file a disavow. In our site quality reviews, what we do is spend several hours reviewing a site’s link profile to determine whether a link by link audit and subsequent disavow would be likely to help them.

For the site listed below, the client provided us with a list of links made for them by their PR firm. It was our opinion that the majority of these, if not all, would be seen by Google as paid, unnatural links, created to manipulate Google rankings. The site owner elected not to disavow these links. They saw significant drops starting August 22, 2019, the date on which we feel Google made an update to how they assess link quality:

no disavow example

The next site is a lead gen site. It has some significant quality issues. We have just recently begun to work with this site and the site owner will hopefully be instituting some of our advised changes soon. This site has a large number of keyword anchored links that appeared to us to be unnatural. We are in the process of auditing this site’s links and have not yet filed a disavow.

no disavow filed example two

The next site is another one that had unnatural links but did not file a disavow. The links were mostly self made links in articles. While this site’s drop is not dramatic, we are anticipating we can reverse it by doing a full link audit and filing a disavow.

no disavow filed example three

Again, we have other examples (that we do not have permission to share) that show similar patterns where we have recommended filing a disavow and they did not and subsequently saw drops in Google organic traffic around August 22-29 or September 13, 2019.

What can we conclude after looking at the above sites?

We recognize that we are making some important recommendations based on results that we are seeing with just a small handful of sites. Our hope, now that more site owners are being convinced that disavow work can still help, is that we will have more exciting case studies to share with you in the future. For now though, we really do feel that there is enough evidence to support the possibility that Google has made some changes to how they assess link quality. 

We are not the only ones seeing benefits to disavowing. 

Stephen Kenwright posted this traffic graph of a site that added 500 domains to their disavow and is now seeing nice changes.

And here is a nice success story posted by TJ Robertson.

FAQ about disavowing in 2019

How can I tell whether filing a disavow could help me?

You will hear different answers to this question depending on who you ask. We do not recommend using an automated tool to tell you whether you have unnatural links pointing to your site. John Mueller has said in the past that if a tool can tell you which links to disavow, Google is likely ignoring those.

If you have either paid an SEO company in the past to build links for you, or have built your own links, the quality of those links should be assessed. Ask yourself, honestly, the following questions:

  • If the links in this post were nofollowed, would there still be an obvious benefit to our company?
  • Would someone reading this post truly believe that the website that is linking to us is recommending our content, our business, or our authors?
  • Would I feel comfortable showing the webspam team how I went about acquiring links?

If you’ve got a little bit of a gut check after reading those bullet points, we would urge you to consider having your links audited. If you are interested in having an MHC team member’s opinion on whether your links are likely to be seen as unnatural, our link overview plan, may be of help. 

Deciding on disavowing links that you have spent months, or perhaps years, and a lot of money creating, is not a task to take lightly. You may also want to cross reference your Google organic traffic across our list of known and suspected algorithm updates to see if you have drops that happened in conjunction with a suspected link related update. If so, this is more evidence to support the possibility that self made links are causing Google to distrust your site and suppress your rankings.

Should we be disavowing all guest posting and content marketing links?

The answer to this depends on scale, intent, and the quality of your links. Here are some general guidelines we use, but keep in mind that every case is individual:

  • If there are guest posting and content marketing posts that contain keyword anchored links to your own site, then these are top priority to disavow.
  • If you have a large number of links, (whether or not they are keyword anchored,) that come from a relationship that your SEO company has with publishing companies, consider disavowing. It is likely not challenging for Google to detect which publishers are regularly trading links for content.
  • Look at the traffic to pages that you have done this kind of link building for. Did traffic increase shortly after getting these links? If yes, then they could be considered as good by Google. If you got a large number of guest posting/content marketing links to a particular piece of content and you did not see increases in ranking/traffic for those pages, then most likely Google is detecting these as unnatural links. At best, they are likely ignoring those links. At worst, they could contribute to a lower assessment of trust for your site and impede your ability to rank.

Please note that we are not saying that all guest posting and content marketing is bad. Guest posting can be a great way to make a name for yourself or your company. When I, Marie, guest post on an authoritative SEO site, this is good for business. There are so many benefits outside of the link that we might receive. 

Content marketing, done correctly, should be done with a primary goal of raising awareness for your company and driving business. We find that many businesses say, “We publish these articles so that we can spread the word about how amazing this company is,” when in reality, the main purpose for publishing many articles is to build links, and often on a large scale.

How long does it take to see improvements in traffic after filing a disavow?

In our experience, if a site is going to benefit from filing a disavow, it can take anywhere from a few weeks to a few months to see the uplift. What we tend to see is that the site will suddenly see an increase that corresponds with a quality update. Our theory is that link quality is a component of site trust. With quality updates, Google likely reassesses trust scores. 

It is also important to note, that we have not seen any cases that we can recall, where disavowing spammy links alone resulted in increases in traffic. The sites that we are seeing nice gains in are ones that were extremely aggressive in finding and disavowing links that were purposely made in an attempt to improve Google rankings. 

It should also be noted that the sites that we saw improvements for after disavowing, were ones that also had good, natural links pointing to their site, alongside of the self made, unnatural ones.

Should you use an automated tool to help you do a link audit?

We do think that some of the automated link auditing tools are great for helping to locate your links, and organize them in a way that allows you to do a manual audit. However, we do not recommend relying at all on an automated tool’s determination of whether or not a link is unnatural. Remember, we quoted John Mueller above as saying, “if some random tool can tell you which ones to disavow, they're probably already being ignored.”

The only sites that we have seen make nice improvements following disavow work are those for which we have been extremely thorough in assessing each link manually and deciding whether it is one that Google would likely see as unnatural. In our experience, the tools are all decently good at finding ultra-spammy links, but often struggle to find patterns of the types of self made links that can cause Google to distrust your site.

Want to hire us to review your links?

At MHC, we are quite proud of the work we do in regards to disavowing. Marie and her team have removed hundreds of manual actions and have seen many of Google’s examples of unnatural links given in response to reconsideration requests. Over the years, we were one of the few companies to publish case studies on Penguin recoveries. Since 2012, we have pored over every word Google employees and Google documents have said on the subject of understanding how Google deals with unnatural links. 

Here is more information on MHC’s link audit and link overview plans.

Our goal is always to help. We offer two services for link auditing. The most comprehensive is a full link by link audit. We gather links from many sources, create a spreadsheet containing one link from every domain, and then our team manually reviews each one. We make assessments and look for unnatural linking patterns. We then share our findings with you so that together, we can make the final decision on what to disavow. We will then file your disavow for you.

If you are not certain whether you need a link audit, you can order our link overview service. For this, an MHC team member will spend several hours spot checking links and looking for obvious patterns of unnatural linking. We’ll cross reference this against your disavow to see if you are correctly disavowing these links. 

In many cases, when we do a link overview, our advice to the client is that further disavow work is unlikely to help. If we do feel that filing a disavow could help your site, we can give you a quote to have us do a thorough link audit and disavow.

Your thoughts?

Have you had success with disavowing? If so, what types of links did you disavow? When did you see the uplift in traffic? We would love to hear your thoughts below.