Yesterday, YouTube published its first every Copyright Transparency Report. With that step, it follows in the footsteps of Google, Facebook, Reddit and other major industry players in publishing such a report.
However, YouTube’s first report comes out in a very different context than the others. While it’s interesting to see the amount and kinds of copyright notices other providers see, YouTube has faced far more copyright criticism than those companies.
For years, it has been accused of having a sloppy, easy-to-abuse system that results in fraudulent takedowns, excessively strong-handed enforcement and still misses a large amount of pirated content that’s available on the site.
In fact, just days before the Copyright Transparency report came out, prominent Twitch Streamer Ludwig, who made the jump to YouTube in a major defection, had one of his first streams on the platform halted over copyright issues.
So, what is in YouTube’s inaugural Copyright Transparency Report? It looks specifically at the first six months of this year, starting on January 1 and ending June 30. The 13-page PDF provides, which you can download here, provides a wealth of statistics about YouTube’s handling of copyright issues.
To that end, here are seven key takeaways from the report itself.
1: The Sheer Volume
According to the report, YouTube handled more than 729 million copyright claims during the period of the report. Given that the report covers exactly 180 days, that represents over 4 million claims per day or more than 166 thousand claims per hour.
That is a volume of notices that is completely impossible to handle by humans, even with an army, and that volume plays into many of the more controversial decisions YouTube has made when it comes to copyright.
2: The Focus on Automation
According to YouTube, they receive copyright notices from four different sources. Two of those sources, Content ID and Copyright Match, are automated and the other two, the Public Webform and the Enterprise Webform, are not. However, of the notices they’ve received, more than 99.3% come from automated sources.
In fact, Content ID alone makes up 99.08% of all copyright claims and handled more than 722 million during the report’s time period.
3: The Issue of Access
Of the four means users can file copyright notices with YouTube, only one is available to the public. That is the Public Webform. Both Content ID and the Enterprise Webform are limited to just a small number of rightsholders (under 10,000 in both case) and are only used by a few thousand.
In short, that 99% of notices are filed by just 4,893 different rightsholders that both have access and use Content ID.
That said, Copyright Match stands out as something of a middle ground. It’s an automated tool. According to the report, over 2 million rightsholders have access to it and that includes all members of YouTube’s Partner Program and all channels that have successfully filed a copyright complaint.
However, as we noted at the launch of Copyright Match, it’s a very limited version of Content ID that lacks many of the core features of the service.
4: The Issue of Wrongful Claims
One point that YouTube drives home is that, according to them, is that there is no serious issue with wrongful claims (especially ones they don’t catch). However, the vast majority of the wrongful claims they get come through the Public Webform.
According to YouTube, over 8% of the claims they receive through that form are abuse, people deliberately trying to misuse the system, and another 6% are invalid, either because they misunderstood the law or filled out the form incorrectly.
By contrast, the Copyright Match Tool, the second most broadly available tool, only has an abuse percentage of 0.14% and an invalid request rate of 4.2%.
However, despite the higher rates of abuse, the vast majority of content reported by the Public Webform was pulled down, with 83.8% of the content being removed.
5: Very Few Counter Notifications
YouTube makes it much easier than most hosts to file a counternotice to a copyright claim. However, in spite of that, very few do.
The Public Webform sees the highest rate of counternotices, with some 5.2% of claims receiving one. However, the Enterprise Webform has just a 1.9% counternotice rate and the Copyright Match system has just a 1.3% rate.
But as low as these rates are, they’re actually very high among other kinds of hosts. In general. Part of this is because YouTube makes it much easier to file a notice, but a much more copyright-aware culture on the site.
6: Even Lower Content ID Dispute Rates
Counter-notifications to DMCA notices carry a significant legal weight with them. Content ID disputes, on the other hand, do not. Despite that, only 0.5% of all content ID claims were disputed.
While this still represented nearly 3.7 million disputes, it is a remarkably small percentage given the criticism that Content ID has received over the years and the ease with which users can dispute those claims.
7: The Resolution of Content ID Claims
What is perhaps most interesting is the outcome of those 3.7 million Content ID disputes. When they are first disputed, they go back to the rightsholder. At that phase, a full 60% are resolved in favor of the uploader, meaning that the claimant either withdrew the claim or let it expire after 30 days.
Though I find it dubious to claim a 30-day wait for a resolution is “in favor” of the uploader, it shows that claimants don’t simply rubber stamp reject appeals as widely believed.
If a claimant rejects the appeal and the uploader files a second appeal, the claimant then must submit a full DMCA takedown notice. According to YouTube, that happened a mere 38,864 times during that time frame. However, even among those 38,000 notices, which had been disputed twice previously, only 8.9% of cases saw a counternotice and less than 1% of those counter notices involved a lawsuit.
One of the interesting things about this report is the way that very large numbers have changed how its reported. With so many copyright notices, even small percentages lead to thousands or even millions of incidents.
So, when headlines read “Millions of YouTube Videos Hit By False Copyright Notices” the headlines are factually correct but only telling half the story.
Honestly, this report doesn’t really tell us much that we didn’t already know. YouTube receives an astronomical number of copyright claims, the vast majority are automated and very few result in disputes. However, because they receive so many, even the small percentage of disputes results in a large number of them.
In my mind, this report doesn’t make YouTube look any better or worse, it just reaffirms what we all have long suspected.