Skip to Content
Malcolm Slade, Head of Technical SEO

The author

Malcolm Slade

Head of Technical SEO

Since their inception, disavow files have been a hot topic within the SEO industry. Questions are constantly being asked of me around their impact on the sites listed, how they are reviewed, should they be used just in-case and, over the last few months in particular I have spoken to several webmasters who are terrified at the thought of being included in a disavow file.

At Epiphany we have debated, tested (as much as possible) and tried to formulate our stance on disavow files. We are a pretty open company and I’m always up for a good debate if one comes about, so I’m going to document our answers to some of the questions around disavow files.

First a disclaimer. I’m not Matt Cutts. I don’t currently and have never worked for Google. The following is what we feel at this stage in time is likely to be correct, feel free to ignore it, debate it etc.

 

Question 1: Do humans read disavow files?

Very doubtful. Although simple in structure, it is likely a case that this is to allow for quicker processing rather than human reading. It is possible (due to the presence of a leading #) that Google might pull comments into some form of review document when auditing against a manual penalty.

Question 2: What happens to the sites listed?

Google loves data. It provides free Internet access, email, document hosting and more solely with the intent of gathering anonymous user data from which it can learn.

With regard to disavow files, Google is getting loads of data from us on what links we consider to be bad with the intent of understanding what we as humans consider a bad link to be so that this can be integrated into some form of algorithm (presumably Penguin).

As with any human sourced data, the disavow data comes with loads of noise (garbage data) as most people wouldn’t know a good link from a bad link. I myself have seen Google, Bing and YouTube disavowed at root domain level. This level of noise means that the only useable data is large spikes.

Basically if your site is disavowed in a single disavow file, it is unlikely to have any effect on anything other than the link evaluation for the site the disavow file is for. If however your site is disavowed in 2000 disavow files (number added for illustration purposes only, I don’t work for Google) you might expect Google to have a look.

To get a scale of the issue around noise and spikes, the guys over at rmoov created a nice little tool / project called DISDIT. It will let you see if they have seen your site in any of the disavow data they have amassed.

Question 3: Should I edit my disavow file as links are removed from the Internet?

The way I look at this is to gauge where the most risk lies.

  • If I have a domain in my disavow file that doesn’t link to me, will it harm me? No way, links change all the time and disavow files can be stale within minutes of being submitted let alone processed.
  • Will having a site in my disavow file harm the site? This is Question 2 and the answer is - only if it is in a large number of other disavow files.
  • If I take a link out of my disavow file once it has been removed will it harm me? Well maybe. What if Google hasn’t re-cached the page hosting the link?

I would advise leaving domains in your disavow until a significant time since their removal has passed. Even then I would check they have been de-indexed as a final precaution.

Question 4: Should I submit a disavow file even without any form of penalty?

At the risk of being stoned by the tin foil hat brigade, I say yes. I would definitely recommend auditing your profile and disavowing the link of little to no value. Chances are that these links are having no benefit and are actually acting against you with regard to Google Penguin by devaluing some of the good links.

At some point the quantity of data Google amasses will allow it to iterate the Penguin algorithm. In honesty I reckon it already has a few times but I have no doubt that something big is coming in 2014!

Question 5: Can you disavow out of a Penguin problem?

I had the pleasure of sitting in on a panel a short while ago at Brighton SEO where this very subject was discussed. The answer that was alluded to was maybe, airing towards no.

That fact is that nobody really knows as it is really difficult to test due to the many tiers of activity that are involved.

  • Was the disavow file processed prior to the Penguin update?
  • Has Google re-indexed all of the pages you have removed links from?
  • Has Google re-indexed loads of pages that your links disappeared from ages ago?
  • Has Google found your new links?
  • Has Penguin evolved?

Everyone was in agreement that a disavow file helps with Penguin but the jury was well and truly out on whether a disavow file in isolation can remove a Penguin problem.

Bonus comment!

Completely out of context but one interesting discussion point that came up in the panel mentioned above was around negative SEO.

The gist was that Penguin works by balancing Good vs Bad to provide a value. If that value is below a threshold you are affected by Penguin. With this in mind the effectiveness of any negative SEO strategy is based on the proximity of the target site to that threshold.

A truly awesome site with loads of amazing backlinks is unlikely to be phased by anything less than 500,000 bad links. A site however that already has 500,000 bad links working against its good links might only require a nudge to put it over the edge.

One comment was “It takes a lot of rubbish to outweigh a small amount of good.” So based on that alone the best way to avoid negative SEO would be to get a few genuinely good links!  

Conclusion

There is no hiding from the disavow file, nor should you. Hopefully the answers above will ease some of your concerns and allow you to focus on the more important stuff like producing great content and being the best you can be.

As always, please feel free to comment, argue, agree or disagree with me. I will aim to reply in the comments as this post has already exceeded 1000 words and it was supposed to be a “quick insight”.

Good luck out there!
Malcolm Slade (@seomalc)