By Quick Summary
- A free SEO tool was taken down by Google based on a false DMCA claim.
- It appears to be possible to file hundreds of legitmate DMCA claims and bury one or two illegitimate claims along with them that ultimately results in all of the URLs being removed.
- Publishers appear to be using automated copyright removal tools to build requests to send to Google and not reviewing them for accuracy prior to submission.
- Google appears to be allowing as few as 75 characters of matching text to result in a DMCA takedown, which may result in quotes and other fair use content getting documents removed (including possibly this article soon).
- Google Search Console does not appear to be reporting these takdown requests / actions correctly.
- DMCA takedown requests of a Scroll to Text Fragment URL do not appear to impact the base canonical URL.
- There is no real way to automatically detect DMCA takedowns of your content.
It is a negative SEO nightmare once thought to be impossible.
An important page on your website suddenly loses all traffic from Google’s search engine, the URL has disappeared entirely from the index it seems. You login to Google Search Console and run a test of the URL which reports back that everything is fine that the URL is being crawled and indexed. Frantic and desperate for answers you check Lumen Database as a last resort and see that someone has reported your URL as infringing on their copyrighted content and Google has removed your URL seemingly without even looking to see that the two documents are far from the same.
This nightmare has for years been considered unthinkable and nearly impossible. While erroneous DMCA takedowns have occurred infrequently they were usually met with stiff penalties. DMCA complaints are also commonly human reviewed for accuracy and authenticity. The most brazen of blackhat SEOs do not care to attempt false DMCA takedowns because of their incredibly rare effectiveness and because of the legal issues filing enough fake takedowns to make a difference would ultimately create. Trust in Google’s vetting of DMCA removal requests, the public records of such filings, and the severe problems created by filling large amounts of fake requests have led most in the industry to feel safe and secure about how the system works – that it won’t work against them.
That is until just yesterday when the head of technical SEO at Merkle, Max Prin, took to Twitter to ask Google’s John Mueller about the takedown notice he had found for a tool on his website TechnicalSEO.com.
According to Max, after noticing a sudden and total loss of traffic from Google’s search engine to the Fetch & Render tool on TechnicalSEO.com he dug around going through possible causes until as a last resort he found an entry in the Lumen Database requesting the removal of a page on his website TechnicalSEO.com due to infringement of the website ThinkMobiles.com.
Digging In To The DMCA Takedown
Here is the document the DMCA takedown claims the Fetch & Render tool stole copyright protected works from: https://thinkmobiles.com/blog/popular-types-of-apps/
And here is the Fetch & Render tool on TechnicalSEO.com: https://technicalseo.com/tools/fetch-render/
As you can see the two are utterly and completely different. ThinkMobiles.com’s claim is completely without merit.
It isn’t just Max’s website either, this one copyright claim includes 84 unique URLs across 75 unique domains including big name sites such as Google Books, Medium, and Quora. All of the URLs listed in the complaint have been taken down by Google. Well, that’s not entirely true, I did find one URL that somehow was spared the axe by Google.
This little guy: https://entiresoftware.com/mobile-app-development/
That appears to be because the DMCA complaint was for a Scroll to Text Fragment URL instead: https://entiresoftware.com/mobile-app-development/#:~:text=Hybrid%20multi%2Dplatform%20apps%20are,cost%20maintenance%20and%20smooth%20updates.&text=On%20the%20other%20hand%2C%20hybrid,to%20native%20apps%20for%20instance.
Max, like any other SEO would do in this position, questioned if this was a negative SEO attack on his technical SEO tool. The idea for this type of negative SEO attack goes like this:
1. Create content on your site similar or replicated from another site
2. File a DMCA complaint to Google asking for a takedown of the original content
3. When Google grants the takedown request your ranking position improves by at least one ranking position or higher for related keywords.
Implement this process enough times, avoid getting counter-notices from the targeted parties, and eventually your website’s or document’s SEO will improve due to the removed content – possibly even being able to rank number one for valuable queries.
I dug in to see if this was actually the case reviewing possible motives for a negative SEO attack, potential gains, and the other content ThinkMobiles requested for removal.
My research showed this issue may not be the negative SEO attack it might seem at first glance. Some of the listed URLs in the complaint do appear to be documents that have scraped this website’s original content.
For example this document was first pubished in 2018, a full 2 years after the ThinkMobiles article was published in 2016: https://www.illusiongroups.com/blog/news/what-are-the-popular-types-and-categories-of-apps/
And this article was published in 2019: https://insomniacs.in/blog/what-are-the-different-types-of-mobile-apps/
Both of these articles appear to be 100% copycats of the ThinkMobiles article from 2016.
In fact if we dig just a little we can find more and more of the URLs in the complaint that do appear to legitmately infringe upon ThinkMobiles copyrighted works.
While it is is possible that ThinkMobiles submitted hundreds of legitimate complaints to hide one illegitimate complaint, that seems like a lot of effort, especially when there is a more sensible explanation.
If This Isn’t A Negative SEO Attack Then What Is It?
While we can’t rule out that ThinkMobiles really wants to take down Merkle’s TechnicalSEO.com Fetch & Render tool for some reason, it doesn’t quite fit. Mostly because ThinkMobiles stands nothing to gain from this takedown as they do not offer a Fetch & Render tool of their own, documentation on this type of tool, or appear to have any stake in such a tool or other similar SEO software.
From examining the full complaint in the Lumen Database it appears clear to me that the takedown of dozens of URLs was requested based on some poorly done automation instead of being some nefarious deed.
The full complaint from ThinkMobiles to Google dated January 22, 2021 includes 10 separate claims. 9 out of those 10 claims are based on short phrases/sentences only 75 to 112 characters in length.
For example the design agency Digital Cotton used the following phrase on one of their pages: “Hybrid multi-platform apps are fast and relatively easy to develop – a clear advantage.”
A review of the page this is on shows it does not appear to be a ripoff of the ThinkMobiles 2016 article and likely should not be removed under DMCA. You can see that page here: https://www.digitalcotton.com/thestack
The two Google Books URLs removed are for the textbook “Log On To Computers – 8” by Madhubun Books. The passages about Hybrid Apps in the book and ThinkMobiles article seem incredibly similar, but matching it for only the phrase “They are built using multi-platform web technologies (for example HTML5, CSS and Javascript).” is akin to taking down a dictionary page over a definition.
There is something definitely wrong going on here by both Google’s handling of these requests and by ThinkMobiles.com’s process in generating the takedown requests.
Since January of this year they have filed at least 100 DMCA takedown requests (and possibly as many as 377) with 10 claims in each one. The copyright claims appear to be aimed at scraper sites stealing their content and republishing it on junky domains or would-be PBN blogs, so the intention at least on the surface appears pure. However, many of the claims submitted in the complaints to Google contain this line “No infringing URLs were submitted.”
That line makes it definitely feel like an automated system is going through each article on their website, breaking down parts of that article (I was able to find up to 3 sentences in some claims), looking for duplicates of this written content on the web, and adding those URLs to a copyright complaint claim for Google to takedown.
Another SEO, Craig Harkins, chimed in to say this exact issue has happened to his employer (InterContinental Hotels Group) at least once in the past.
I’m still not entirely sure how TechnicalSEO.com got wrapped up in this as nothing on their page incldues content from the ThinkMobiles article or the claimed exact phrase “Hybrid multi-platform apps are fast and relatively easy to develop – a clear advantage.” The tool does display page content/code after a fetch is done and it is possible that Google’s cache of the tool included this at somepoint or that whatever automation was used somehow ran a query on the tool and then saw the content making it think it was infringing.
Google’s Search Liasion Danny Sullivan has said he will be sharing this incident internally due to “enough weirdness” and recommend using a counter-claim if you ever find yourself in this situation.
Google Search Console Not Reporting DMCA Takedown Requests / Removals
This event does bring to light another issue to be aware of. Google’s John Mueller replied to Max stating that he believes GSC should be sending alerts about DMCA takedowns and that if that isn’t working it is probably a bug.
If someone was reporting your content as being copied from your website then how would you know if Google is not telling you?
There do not appear to be any notification tools that use the Lumen Database API and their API terms state that it is to be used for research purposes only, meaning we likely will not see notification system of any kind based on the Lumen Database any time soon.
Bigger brands do tend to get hit with DMCA requests more frequently and if single sentence matches as in this case are enough for a Google DMCA takedown then media websites, large blogs, recipe bloggers, etc… might also want to keep a close eye on their analytics or Lumen Database for signs of possible takedowns. Performing a search and viewing the top 10 pages of results on Lumen Database is free for the public, but after that you’ll need to apply for a “research account key” to see more of the content in the database.
It is my sincere doubt that this becomes more normal or gets serious usage as a negative SEO tactic, and in the future if you see what Max did do not consider this a remote possibility until it is the only remaining explanation.