An interesting phenomenon in Google Video is the resonance of popular videos. Once a video becomes popular it is show in the top 100 and there it has an even much larger chance of being seen by a large number of members which makes it even more and more popular. Due to this resonance effect many videos once they reach the top 100 tend to stick right there for long, not because they are actually better than new videos but because they have managed to reach the top 100 list and therefore get a boost in popularity just by being among the top 100 list. It's a circular and unfair thing.
Google seems to calculate the popularity of a video by counting the number of users who have clicked on it to view it. Perhaps they have a more complex algorithm for determining popularity including how many people viewed the video till the end as well as other criteria. Yet still the resonance effect dramatically accelerates some videos and keeps them at the top much longer than they deserve.
To counter this effect, Google introduced the Movers & Shakers video list which lists videos that are gaining popularity quickly even if their popularity has not reached a high level, but what counts here is speed of increase in popularity and not the popularity rank itself. This has helped a lot in reducing the resonance effect by giving a higher chance for new videos to climb up the popularity ladder and shake down old videos which had clung for long to the top 100 list due to the resonance effect.
Google then introduced yet another feature to help reduce the resonance effect. If a user is signed in with his Google account when visiting Google Video, his popular list is customized to reflect his own interests depending on his search and video viewing history. This again helped reduce the resonance effect.
The final push that Google has made till now to reduce the resonance effect even more was to introduce Recommendations which again rely on each users video search history yet rely less on the popularity of a video and more on the users own interests.
Although social networks and community driven content, which are said to be part of the buzz word Web 2.0, are expected to give even more democracy to the Internet by letting people decide what is good and what is bad, what is popular and what is not, yet it looks like Google should also work on developing a more robust algorithm to determine the popularity or the 'niceness' of a video on not just rely on simple calculations of how many users have viewed a video.
Perhaps Google should consider adding a feature similar to the fascinating interestingness feature found at Flickr which is able to automatically discover marvelous photos using a complex algorithm.
