Israeli computer scientist Yuval Shavitt has developed an algorithm for the web that predicts which songs will become chart-topping hits. It monitors p2p file-sharing sites like Gnutella to see which songs are being passed around most frequently. Mr Shavitt has already begun working with record labels and says they've had a 30-50% success rate in identifying which early recording will become commercial successes.
I think this is a pretty good idea. For one thing it might help get more innovative songs airplay because risk-averse labels have more confidence in their potential. Maybe you've noticed, but in recent years labels have relied on traditional market research to choose releases and that means they pass over songs that are risky, raw and creative. This, unfortunately, means we get to listen to Nickleback over and over again on the radio.
And perhaps a bigger positive about this software is that it might eventually kill labels off altogether. Instead of licensing this to labels, Mr. Shavitt should let the public use it (subscription or ad-based revenue model) so we can know which songs are the most-shared and most popular. Then, who needs a label at all?
I think this is a pretty good idea. For one thing it might help get more innovative songs airplay because risk-averse labels have more confidence in their potential. Maybe you've noticed, but in recent years labels have relied on traditional market research to choose releases and that means they pass over songs that are risky, raw and creative. This, unfortunately, means we get to listen to Nickleback over and over again on the radio.
And perhaps a bigger positive about this software is that it might eventually kill labels off altogether. Instead of licensing this to labels, Mr. Shavitt should let the public use it (subscription or ad-based revenue model) so we can know which songs are the most-shared and most popular. Then, who needs a label at all?
