5% of tube site uploaders upload half the videos, top 10% upload 2/3rds
 

We've been doing some R&D on a completely new and different way to protect people from content theft. Part of that has involved studying the behavior of people who post videos to tube sites. Here are some of the stats we found from our sample:

The top 5% of uploaders uploaded 51% of the videos, an average of 114 each
The top 10% of uploaders uploaded 66% of the videos , an average of 73 each
57% of uploaders uploaded just 1-3 videos each
The top 4.6% of users uploaded half the videos

I'm curious what percentage of site members ever upload anything to a tube.
If say 2% of members are tube uploaders, and 10% of THOSE upload 2/3rds of the videos, that would mean that only 0.2% are serious uploaders.
In other words, out of ten thousand members, only two would be problem uploaders.


Random reminder: back up your stuff now. Don't forget databases, server config, etc. Or just use Clonebox. :)

Does this surprise you?

My tubes just get a lot of people that create profiles and never upload. There are uploaders ... but its typically 1 or 2 videos and amateur in nature...

Which uploaders? The ones that live in the Phillipines that are paid to keep fresh content getting uploaded to the tubes through the FTP brokers that charge for stolen content to keep the illegal tubes fresh?

Pretty much the same ratios have been around for 12 years starting with the News Groups and on to the P2P systems.

It's about what I expected, but I'm often wrong, so me THINKING the numbers would look that is different from actually knowing what the actual numbers are.

To me it's interesting to KNOW that a site with 5,000 has, on average, only one guy causing a problem. It means you don't have to go DRM for everyone, you just take care of that one guy.

Or the uploaders that are provided passes to members areas to keep the content getting uploaded to the tube?

Many people fail to realize that there is a whole other market out there for providing stolen content to some of the tubes.

So we can extrapolate from your numbers that somewhere around 80% of the uploading is caused by 20% of the users. You know it's funny because you can actually find that same ratio in lots of places. I'd be willing to bet than in most if not all affiliate programs, 20% of the affiliates generate 80% of the sales.

We should totally document that pattern and name some sort of principle after it. Oh wait somebody already did http://en.wikipedia.org/wiki/Pareto_principle

Piracy is not as big and unbeatable as many of gfy members want to make it look(why?).

Go for those huge uploaders and you will slow down piracy a lot!

This sample was 95% user uploaded. Webmaster theft is another issue. Almost by definition it's bulk in nature.

Raymor, I think the question is how are you targeting or identifying uploaders since I thought the same as Seeric. Based on your numbers, I'd bet a lot of money those "independent uploaders" are those paid by the same tube sites, and it's not exactly a secret :2 cents:

How did you come up with these numbers?
I find it very hard to believe its possible to gauge this.
Many folks have multiple "users" if they are even real users

Until someone goes after the Tubes to prove they pay these guys nothing will change. Offering some of these uploaders $5,000 to spill the beans would be a way to make millions.

It would be interesting to do a study on how much tube content is in fact user uploaded. Clearly the tube site owners have to seed their new tube at minimum.

Those who know me know that I often use phrases like "I suspect", "I wonder if" and "it could be". I tend to avoid stating as fact something that I don't actually know. In this case, because it's an ongoing investigation I can't reveal my methods and sources, but I can say with a high degree of confidence that very few of the uploaders for these tube sites were paid. I'm not saying "I suspect that's true" - I'm pretty confident that it's true.

I have no agenda here and I have no reason to try to convince anyone of anything, so
readers may decide for themselves how much stock to put in these numbers. Those who know me may decide that if I say they weren't paid, they probably weren't. I can't go into detail about my methods, but I simply don't need to because I'm simply posting data FYI, take it or leave it.

You can have some confidence that the data is reasonably accurate because as another poster it does approximate a Pareto distribution (80/20 rule). That's a distribution frequently seen in populations, so it tells you that these results appear reasonable. Had we found that 1% of uploaders were responsible for 92% of the files, that would be non-Pareto and suggest an outside influence such as payment, or simply error. In this case it's Pareto, so reasonable to believe that the weren't a several being paid.

Thanks for putting some useful information onto this board. It's nice to see an actual study being done, and fact being reported.

For anyone who's actually gathered data from a sample and processed those statistics, it's always amusing to see the amount of hindsight bias that comes out when you release the final data. There are some laughable examples in this thread.

Again, thank you for useful information.