Hasn't this been tried before?
  • (a) Sort of.
  • (b) The same question was probably asked of the Wright brothers. Like them, we've solved some problems that kept other inventors from getting off the ground.
What very few people seem to realize is how badly web sites have failed up to now at identifying good material by undiscovered musicians and other content creators.
Any number of sites have offered unsigned bands a common, discoverable place to try to sell their songs. But when it comes to identifying the best ones, these sites have simply trusted that somehow, they'll bubble up to the top as the most "popular"—whether in terms of sales, plays, "favorites," star ratings, or whatever.
In actuality, this has rarely happened. When they're first launched, sites for independent music tend to draw so little traffic that just a few sales or "likes" can make a song a popularity standout. This environment is more nurturing to bands with a lot of friends than those with a lot of talent.
The arrangement is also considerably less than ideal for the general public. Visitors assume the most popular songs are the best ones, and click on them first. After sampling maybe five or six of these without hearing anything they especially like, they conclude (erroneously) that nothing on the site is very good, and leave without coming back. Everyone loses.
And because the site fails to catch on, traffic stays low, which continues to favor the bands with friends over the ones with talent. The vicious cycle repeats.
To break it, the first step is to recognize that the best songs can never be purchased or "favorited" enough to become popular unless people first hear them.