Opinion: Viacom vs. YouTube, the numbers
This article was originally on a blog post platform and may be missing photos, graphics or links. See About archive blog posts.
Ousted Merrill Lynch analyst Henry Blodget, who now writes the intriguing Internet Outsider blog, offered a must-read post today that tries to replace the speculation about what people watch on YouTube with something that approaches science. The post summarizes a report by Vidmeter Inc. that traces the source of the most popular videos on YouTube over a four-month period in 2006-07. Judging by the take-down notices YouTube received from Viacom, NBC and other copyright owners, Vidmeter concluded that a little more than 9% of the 6,725 videos tracked were posted improperly, while nearly 91% had the copyright owner’s authorization -- possibly because they were truly user-generated videos. Significantly, the 9% that were removed accounted for less than 6% of the views generated by the 6,725 videos tracked. (Viacom had a disproportionately large share of the views, clocking in at 2%.) In sum, Blodget wrote, the study undermined Hollywood’s argument that people flock to YouTube to watch bootlegged clips from the major studios and networks. Instead, it suggested that ‘traditional media videos make up only a small percentage of YouTube views.’
The study’s authors caution readers against extrapolating the sample data to estimate the total amount of unauthorized videos on YouTube. There are some obvious problems with the methodology: for example, it doesn’t distinguish between authorized posts and unauthorized ones that the copyright owners don’t try to remove. And the most popular videos on YouTube may not be representative of the total catalog; the study counted only 72 videos from Viacom that were removed, but the conglomerate sent take-down notices for 100,000. Still, the small number of popular YouTube files that apparently came from major content providers like Viacom bolsters YouTube’s argument that it hasn’t built its business on the backs of Hollywood.