The Google Disavow Tool is one of the more mysterious and controversial tools in the world of SEO.
I remember when the tool was first introduced. It seemed like Google was finally saying, “sometimes we get it wrong.”
Search engine marketers have always focused on building links, and the announcement of the tool marked a new era of removing links.
That said, the tool is clouded in mystery. It’s unclear whether or not disavowing links has a significant impact on ranking (especially in the absence of a manual penalty).
So, is the Google Disavow Tool worth using?
That’s what we’re going to test.
What is the Google Disavow Tool?
The Google Disavow Tool is an advanced tool available in the Google Search Console. This tool allows webmasters to “disavow” bad links.
If someone builds a subpar link to your site (including yourself), you can tell Google you want nothing to do with that link. Pretty enticing, right?
Yes and no.
While the disavow tool seems like a great tool to help SEO’s eliminate bad links, it’s not without its downsides.
Why Use the Google Disavow Tool?
The appeal of the disavow tool is straightforward. It gives SEO’s the opportunity to eliminate or minimize the negative impact of bad links.
In a perfect world, this tool would allow SEO’s to hand-select which links they want to count toward their rankings.
In reality, there are a few potential downsides.
Why NOT Use the Disavow Tool?
Why isn’t every webmaster using the disavow tool to refine their backlink profiles?
In short, it’s unclear whether or not the tool is effective. More importantly, it’s unclear whether or not the tool can have negative impacts on rankings.
Here are a few reasons why I’ve been hesitant to use the tool to its fullest.
Google Says It Ignores Bad Links
Google’s algorithm has come a long way in the past decade, and Google has explicitly stated that the algorithm ignores most spammy links.
You know that webpage you recognize as spam in less than a second? Google says the algo can do that too.
While this is reassuring, I don’t fully buy it. Sure, Google may be able to recognize obvious spam sites, but that doesn’t mean that bad links can’t hurt your site. Links are still a part of Google’s ranking algorithm, and bad links can hurt your rankings.
If Google was able to identify and automatically disavow every bad link, we would all have perfect backlink profiles (which is certainly not the case).
The Subjective Nature of Links
As search engine marketers, we do our best to guess what types of links Google likes. Sure, certain “good” links are obvious. We’d all love a dofollow link from the New York Times.
With the exception of a few clear winners and losers, the quality of a link comes down to a subjective analysis (from both Google and search engine marketers).
When you disavow links, you “guess” which links Google will view as “bad.” Guess right and you may improve your rankings. Guess wrong and you may end up hurting your rankings.
Assume you have the following backlink profile (as viewed by Google):
- 100 Good Links
- 100 Bad Links
- 100 Neutral Links
You perform a link audit and remove a total of 100 links. Your GOAL is to have a backlink profile that looks like this:
- 100 Good Links
- 0 Bad Links
- 100 Neutral Links
In reality, you could have a backlink profile like this:
- 80 Good Links
- 40 Bad Links
- 80 Neutral Links
While you eliminated 60 bad links, you also took out 20 good links. Since we don’t know exactly how Google’s algorithm ranks links, this could have a negative impact on your site. Assume Google’s algorithm could automatically recognize and disavow your 100 bad links. The net effect of your link audit is a loss of 20 good links and 20 neutral links.
Point being – a link audit/disavow is a fragile, inexact process.
The Links You Don’t Disavow
This one is a bit of a “conspiracy theory” of mine….
When you disavow links, you are telling Google which links you don’t want to be associated with.
In a way, you are also telling Google which links you DO want credit for. This can be both good and bad, depending on how you acquired your links.
Assume you have the following link profile:
- 20 Good Links You Built Yourself
- 5 Good “Natural” Links (built without outreach or any other method)
- 20 Bad Links You Want to Disavow
After you disavow the bad links, you have a link profile that looks like this:
- 20 Good Links You Built Yourself
- 5 Good “Natural” Links
Notice any issue?
Your new backlink profile lays out your link-building blueprint. If there’s any footprint or commonality across links, Google can better recognize it.
Does Google work like this? Who knows…But this has always been a consideration for me.
Who Should Consider Using the Google Disavow Tool?
With all of this in mind, who should even consider using the Google Disavow tool?
This isn’t a black and white issue, so there is no perfect answer, but we can consider a few different types of sites.
Mega Authority Sites (No)
Mega Authority sites like New York Times and Healthline are unlikely to need the disavow tool. These sites have millions of backlinks (both good and bad), but the sites’ authority should be enough for Google to recognize that the websites are not engaged in link schemes. Not only would the disavowal process be a nightmare, but these sites have a LOT of strong links that should outweigh any bad links.
New Sites (No)
Newer sites have no reason to use the tool, because – you guessed it – they are new. It is unlikely that they have been around long enough to have a toxic backlink profile (assuming the domain is new). Of course, there are a few exceptions, but for the most part, these sites should focus more on building links than disavowing them.
Small Sites (No)
If you have a smaller site, as measured by the amount of content and monthly visitors, you probably don’t need to disavow links just yet. For example, if your site has <5,000 monthly unique visitors and traffic is stable or increasing, there’s no reason to start disavowing links. It’s unlikely that your backlink profile is holding you back. These sites should focus on content improvements, on-site SEO, performance improvements, and link building.
Growing Sites (No)
If your organic traffic is steadily increasing every month, you should not disavow links just yet. Don’t fix what isn’t broken.
If you notice a lot of toxic backlinks that you believe may hurt your site in the future, you may be able to justify a link audit, but don’t get too hasty with it.
Penalized Sites (Yes)
If your site has a link penalty, a link audit and disavowal make sense. You have nothing to lose. Google has penalized you for bad links and you should absolutely start cleaning up your link profile.
Declining Sites (Maybe)
If your site experiences a noticeable decline in search traffic (i.e. 20-50%+), you may consider doing a link audit. There are two other considerations here:
- What is your monthly traffic?
- Why is your traffic declining?
The first question speaks to authority. Did Google like your site to begin with? For example, if you have 1,000 visitors/month and you drop to 500/month, that is a sharp percentage drop, but you didn’t have much traffic to begin with.
The second question will help you properly diagnose the issue. Is your organic traffic drop related to links? You cannot answer this question with 100% certainty, but you can do a comprehensive audit. For example, Google recently announced another algorithm update for review sites. If you run a review site and your traffic dropped, it’s more likely that you need to improve the quality of your content than your backlink profile.
Stagnating Sites (Maybe)
Lastly, there are stagnating sites – sites where traffic fluctuates in a +/- 10% range for a long period of time despite active SEO efforts.
For example, if your traffic stays at 10,000 visitors/month for 6 months despite adding new content, improving existing content, optimizing Core Web Vitals, and other SEO efforts, you may want to take a closer look at your backlink profile.
In these cases, it can be helpful to dive deeper into the stagnation. High-level pageview metrics don’t tell the full story. Are certain pages declining while others are growing? Is a certain type of content performing better than another type?
If your site is stagnant, it makes sense to consider all SEO factors before performing a link audit.
Google Disavow Links Case Study
Now we get to the case study component of this post. My goal is to test the impact of link disavowals on sites that fall under the “maybe” category listed above.
The disavow tool has always been clouded in uncertainty for me. There are some good case studies out there, but the community seems to lack consensus.
I’ve yet to determine how this tool fits into my SEO strategy. Sure, I’ve dabbled in it. I have disavowed hundreds of links that are clearly spam, but I have never committed to it fully.
Should link auditing/disavowing be a regular part of the SEO process?
I have never taken the time to do a full link audit to refine my backlink profile to this degree.
And so the case study begins….
Important Note: This is NOT a single-variable case study. I wanted to test the impact of the disavow tool across multiple sites and I do not plan to pause other optimizations. If I believe that other significant optimizations have an impact, I will mention it in the case study updates.
The Disavow Process
By no means am I a link audit master, but I have been building links for almost a decade. I can differentiate between a good and bad link (as I see it).
In order to disavow links on a site with thousands of referring domains, you need to rely on both automated tools and manual analysis. In the past, I have used the SEMRush backlink analyzer, but I will be using Link Research Tools for this audit. If you are going to do a FULL link audit, I think it makes sense to use a tool that will give you the full picture (or as close to it as possible).
I’m not going to breakdown my exact process for disavowing links, but I will be using a combination of automation and manual analysis to identify the following bad links:
- Clear Spam
- Links on Penalized Sites
- Links on Deindexed Sites
- Links with Bad Anchors
- “Obvious” Links from Link-Building Efforts
There are dozens of quantitative metrics I can use for this analysis, but I will be relying on a few main ones:
- DTOXRISK (Proprietary metric in Link Research Tools)
- Ranking Keywords (# of Keywords Ranking in Google)
- Page/Domain Power*Trust (Proprietary metric in Link Research Tools – similar to DA/DR)
- Link Anchors
- TLD’s of Referring Domains
These metrics will be used to identify potential candidates to disavow. I will go through each one manually to ensure there are no false positives. I will also dig deeper into some of the “questionable” links using tools like ahrefs and SEMrush.
From there, I will disavow both individual URLs and domains (mostly domains).
Once I have my new disavow file, I will upload it, sit back, and track the results. I will submit the links to a link indexer (One Hour Indexing) to speed up the recrawling of the pages (which could otherwise take months).
I will be performing link audits for four sites, some of which have had links disavowed in the past. Ahrefs metrics are mentioned for simplicity, but links were pulled from 25+ sources during the audit in Link Research Tools.Here are the stats:
- Site 1 – 5k/52.3k ahrefs RD/Backlinks (627 Domains Previously Disavowed)
- Site 2 – 5.6k/36.6k ahrefs RD/Backlinks (566 Domains Previously Disavowed)
- Site 3 – 3.5k/15k ahrefs RD/Backlinks (0 Links Previously Disavowed)
- Site 4 – 970/2.6k ahrefs RD/Backlinks (42 Domains Previously Disavowed)
The previous disavowals come from a very casual link audit in SEMRush, whereas the new disavow lists are the product of in-depth audits in Link Research Tools that take ~3-5 hours per site.
- Site 1 – 3146 Domains Disavowed, 37 URLs Disavowed (vs. 627 Previous)
- Site 2 – 3921 Domains Disavowed (vs. 566 Previous)
- Site 3 – 1360 Domains Disavowed (vs. 0 Previous)
- Site 4 – 1875 Domains Disavowed (vs. 42 Previous)
Starting Traffic & Notes
All disavow files have been uploaded on December 30, 2021. Below are the organic traffic metrics to start. Keep checking this section for live updates to the traffic charts, and the following section (“Disavowal Case Study Updates”) for general insights and takeaways.
Site #1 is one of the higher traffic sites in this study. This site experienced a small drop in traffic in Nov/Dec 2021, followed by a sharp increase in the last week of December 2021. Normally, I wouldn’t run a case study on a site with higher traffic, but the site experienced a gradual decline in organic traffic over the past few months. The recent spike made me hesitant to disavow links, but I will be proceeding as planned.
Update March 3, 2022
The traffic popped right before I uploaded a disavow file, only to trend back down to the previous level. I am leaving this disavow file untouched for now.
Update April 7, 2022
On April 7, 2022, I replaced the new disavow file with the previous (627 domains disavowed vs. 3146 Domains and 37 URLs in the new one).
I ran this test on site 2 in March without any negative repercussions, so I will be testing it with site #1 now.
Site #2 is the oldest of the sites in the test. traffic has been relatively stable for the past few months, but is well below its peak from earlier this year.
Update March 3, 2022
Traffic dropped on this one after the disavow upload. I will be replacing my new disavow file from the case study with the previous disavow file to see if there is any impact.
Site #3 is one of the newer sites. Link building initiatives have been limited so there was no previous disavow file. The organic traffic followed a similar pattern as Site 1, but at a smaller scale.
Update March 3, 2022
There was no positive impact here. The impact is slightly negative. I will be removing the disavow file completely from this site.
Site #4 is a unique case. This site has never gained much traction, despite the fact it has over 100 quality posts dating as far back as 2019. This was a last-minute addition to the disavow case study. While I don’t believe the site is penalized, there isn’t much to lose from disavowing links since the site already doesn’t get much traffic.
Update March 3, 2022
There was no positive impact here, but the site never really had much traction to begin with. This helps me eliminate the possibility that the site was under-performing due to bad links. I will be removing the disavow file completely.
Disavowal Case Study Updates
This is a live case study. I will continue to post updates regularly. Leave a comment if you have any questions or if there is information you are interested in seeing.
I am posting live updates in the section above. I will use this section to post general takeaways and insights.
December 30, 2021
Uploaded disavow files for 4 sites. Each site has a unique backlink profile and varying degrees of traffic.
March 3, 2022
We are 2 full months into this case study. Some SEO’s may argue that’s too early to gauge the effectiveness of this case study, but I think it’s enough time to take away some general insights. While SEO is a slow process, the industry is fast-paced. Google releases updates regularly and most SEOs are constantly running tests to improve traffic. If you can’t gauge the effectiveness of a significant test in two months, it’s hard to accurately determine the impacts of any efforts (unless you run true “single-variable” tests, which would be nearly impossible to do on a live site with a business behind it).
Onto some insights…
One thing is certain – uploading disavow files did not have a positive effect for any of the 4 sites.
At best, traffic stayed about the same. At worst, traffic dropped off a bit. While the impact on traffic may not be directly attributable to the disavow files alone, I think there is enough evidence to suggest that uploading a disavow file is unlikely to have a positive effect on these sites (at least not in the first 2 months).
One thing I find interesting is that there is SEO software built specifically for link audits (which encourage uploading disavow files). Link Research Tools is built entirely around this process and it has a sticker price that starts at $499/month. SEMRush includes a Backlink Audit feature and sends me weekly emails about how my “toxic backlink profiles” are putting my site at risk.
Why are SEO software companies encouraging webmasters to disavow links?
For Link Research Tools, I get it. The service is built specifically for building and maintaining a clean backlink profile. This is their business. Whether or not this process is effective is a different story.
On the flip side, you have SEMRush which offers a suite of SEO tools. I would speculate that the Backlink Audit isn’t even close to the most popular. So, why push customers to do a backlink audit when it could negate the positive impact that the other tools in the suite produce? For example, I could use the SEMRush Site Audit (which I believe is incredibly useful) to boost my rankings only to negate that impact by disavowing links.
There’s an argument to be made that the effectiveness of a link audit comes down to the link audit process itself (which is true), but most of these tools report a large percentage of links as being toxic. I’ve yet to come across a link audit tool that simply recommends removing a few highly toxic, suspicious links. In my experience, these Link Research Tools marked 50-80% of a site’s backlink profile as being toxic and SEMRush Marked 15-30%. These are not sites that have been subject to negative SEO or black hat link-building campaigns. The web is filled with spam and most websites have more “spammy” links than good ones (seriously, If you wanted to choose which backlinks you actually wanted to keep, how many would you select?).
Google has already come out and said they can identify and discount most of the spammy links that these platforms tell you to disavow. This could mean that a link audit on a well-performing site has a higher risk of disavowing a good link than it does a harmful link.
Webmasters are responsible for their own actions and should hold themselves accountable for every decision they make, but it does seem strange that SEO tools would actively promote a practice that is marginally effective at best and harmful at worst.
The results thus far give us the opportunity to expand the case study by testing a new process.
What happens when you remove a disavow file or revert changes?
Once again, there are many theories here, the most interesting being that once Google disavows a link, it will lose some of its power even if it is removed from the disavow file.
Time to put theory into practice and continue the live case study. I will be testing two things:
- Reverting back to an older version of the disavow file (Site 2)
- Removing a disavow file entirely (Sites 3 and 4)
For Site 2, which experienced a small drop in traffic, I will be reverting back to the previous disavow file which contained 566 domains vs. 3921 in the new one.
For sites 3 and 4, I will be removing the disavow files completely.
Site 3 will have 0 domains disavowed (the original # before the case study), compared to 1360 domains in the case study disavow file.
Site 4 will have 0 domains disavowed, compared to the case study disavow file which contained 1875 domains, and the pre-case study file which contained 42 domains.
These tests should help the case study insights come full circle. In the next few months we will discover one of a few things:
- Uploading a disavow file causes irreparable damage (i.e. removing the file or reverting changes either has a negative or neutral effect on performance relative to pre-case study site traffic).
- Disavow files can have dynamic impacts on performance (i.e. performance fluctuates relative to the uploaded disavow file or lack thereof).
- Removing a disavow file has a positive impact on a site (i.e. removing/reverting a file has a positive impact relative to either pre or post case study performance).
I will report back in a few months.