Pete Cashmore (Mashable!) has an excellent post on how easy it is to cheat YouTube ranking system.
It seems that by simply refreshing your video page every few seconds/minutes, YouTube logs it as a new view.
I got to admit I'm surprised they don't even do a simple IP check (although this is also something you can overcome with some basic free applications) to make sure it's not the same person who watch the video again and again.
When looking at this you got to think again on the 100 million videos per day figure that the GooTube deal was based upon.
This all thing brings me back to two important concepts I would love to see emerge:
- One statistical platform to rule them all - When you look at Hitwisw and comscore (not to mention Alexa) data, you often find some large discrepancies. How better it would be if all the major sites would have used on objective independent statistical platform?
Of course I don't see this happen any time soon. When there is misinformation, there is room to say you are much better than you really are. This kind of platform is in the best interest of the advertisers but not necessarily of the publishers (besides the ones in first place of course). - Better ranking algorithm - Why rank the top videos just by number of views? Why not combine a few criteria such as views, number of talkbacks, comments, spread of views among different users, locations and over time and even more in order to create the rank? Simply put it, why not use a similar concept to the Google Page Rank in order to rank site content?
If a content rank will be determined by multiple criteria, it will be much harder to game the system. Especially if each criteria will get a different importance factor.