I thought it would be interesting to get some kind of historical fix on whether this is really an unfair rate per song, and if so by how much. I'm also curious as to whether "per play" is really the right metric: if I sold you a CD and it sucked, you'd almost never play it so that rate might seem high... but be unimportant because you'd sell so few CDs. How does that translate to the Spotify model?

Maybe this is obvious to others? But not to me.