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?
One thing to focus on is that the rate per song in the CD era can't fairly be compared to the rate per song today - particularly for popular songs that have a global reach. I would agree that "per play" is only a partial metric, but would caution against trying to compare too much over time.
at the end of the day, it's art...'fair value' is a very unique construct
the main point I was making was that comparing 20 years ago to today in this case is likely not equivalent (at best I would say it's an incomplete view)
I don't really know how you quantify "number of plays" through time and attribute a "value per spin". Take John Coltrane's "Blue Train" - $10 for that album in 1957 would have given you almost 60 years of unlimited play (if you could keep the record clean!).
To a certain degree, "Buying an album" has always been a gamble. Spotify/Amazon/etc make it easier to (a) not buy albums that suck, and (b) buy more singles instead.
Maybe this is obvious to others? But not to me.