'reaching' 237 million people is a lot different than 237 million people actually listening to your song. Old media like radio have a long history of making these stats as inflated as possible to keep ad rates high. Things change drastically when you can accurately count exactly how many people are tuned in at any given moment.
Listeners, like web site visitors, can only be reasoned about statistically. This is even more true in this century than it was in the last century. If a service (say Spotify or Pandora or whatever) has a population P then the probability that they request a performance of a given piece might be f(P). Popularity might be measured by the number of unique people who request a performance g(P) and quality might be measured as the mode of the number of times a unique person requested a performance.
The remuneration rate, like the 'average selling price' for a commodity, is the median price of the performance across all requests.
My original point is that the 'size' of the music market, in terms of the number of times a song is played, is very very large. There are probably close to a billion 'plays' of songs on any given day in the US. I'm sure Pete Warden could come up with a way to figure out an exact number, I base that guess on half the population, 150M people, and 6 songs per person per day median, so 900M plays.
If the ASP, to artists, is $132/32000 (not saying it is, just its the only real number we've got in this discussion) that would make a daily sweep of like $3.7M. So if, as an artist you were the top song that day you might take home 1% of that or $37,000.
My overall point is that the music industry that is being birthed out of the ashes of the old one, looks a lot different. It's more 'on demand', it's more relational, and it's less understood. Consequently there are a lot of people trying to mold it to something they can control and get the money out of, and that reality distortion makes seeing the real forces driving it harder.
So when you reason about the music business and someone says "Gee the artist only sees one tenth of a cent each time spotify plays their song.", that sounds horrible! A song, for only 0.1 cent? Gee I'd have to listen to the same song all day just so they could buy a cup of coffee at Peet's. But if that song is reasonably popular and it's being played 10 million times a day, that means the artist is making $10,000 a DAY on just one song. That sounds awesome! The difference between these two is all that probability / market size reasoning that is sometimes glossed over.
