On the weekend the New York Times had an article from Clyde Haberman, a brief history of musicians’ earnings in the digital age. It included this (now oft-reported case):
… thus far, the people who create the music on which others build their fortunes often receive mere rivulets of reward. Not everyone is a Beyoncé or a Taylor Swift (who has removed her entire oeuvre from Spotify to keep it behind a pay wall). Many more musicians are like Zoe Keating, a cellist from Northern California who described her situation in detail last year. Over a six-month period, Ms. Keating’s songs had been played on Pandora more than 1.5 million times; that earned her all of $1,652.74. She had 131,000 plays on Spotify in 2012. She took home $547.71, or less than half a penny per play.
So, does Spotify pay well? The reporter thinks it’s obvious, since 131,000 is a ‘big’ number, but ‘half a penny’ is a ‘small’ number.
But can’t we do better? I’m not a music reporter, but it seems to me that Spotify’s payments are not that out of line with earnings from radio or CD sales.
First, consider radio. I don’t have detailed numbers on radio listenership, but this piece from Bryan Farrish suggests that at any one second the top radio station in New York has about 157,000 listeners (and, one assumes, top stations in other markets have smaller numbers). But that means – and I’m working in ballpark terms – Ms Keating had as many people hear a song of hers on Spotify in 2012 as if she had had one play on New York’s top radio station. Would one play on New York’s top radio station net over $500 in payments to the musician?
Second, let’s go to CD sales. Here’s a low-tech guide to royalties. Suppose I buy a CD with 12 songs, by a singer-songwriter (thus I combine payments to one artist). This would net – ballpark – about $1 for the writer/publisher part of the deal, and about $1 for the recording artist part of the deal. So, around $2, or about 17 cents per song. If I listen to the CD more than 34 times (which is certainly true of many of my CD’s), the royalty rate is in Spotify territory on a per-listen basis (it’s higher than Spotify if I listen fewer than 34 times, less than Spotify if I listen more).
So, again, does Spotify pay well? I’m honestly asking – I don’t know the answer. But it seems to me from these back-of-the-envelope guesses that on a payment-per-person-listening-to-a-song basis it is not wildly out of line with the payments to musicians from either radio play or CD sales. I don’t know the answer precisely because of reporting that thinks the Spotify numbers, with no context, say all that needs to be said. And if concerns about ‘winner-take-all’ are your concern, I think we also need to compare the Gini curves for all sorts of recorded music earnings.
If I’ve missed something, or you know of a good piece that answers this question, do let us know!
Time Magazine has an article with some interesting numbers here:
http://business.time.com/2013/12/03/heres-how-much-money-top-musicians-are-making-on-spotify/
It lists the top 10 songs on Spotify in the week after Thanksgiving in 2013. Using the higher estimates listed, Spotify paid a total of $5,186,000 in royalties for these songs since they appeared. And yet Spotify says it paid over $500 million in royalties in 2013. The top ten songs thus received only about 1% of the total payments. This would not support the winner take all argument.
Unfortunately, there isn’t enough data to make clear conclusions. (And the data isn’t exactly scientific.) Older, popular groups might have a host of albums on Spotify and be making a lot of money from many songs even if they aren’t in the top 10. And the top ten songs might be surrounded by other popular songs on the albums which also make a lot of money. Their total sums could be much higher than for the top ten songs. Expanding beyond the top ten songs, the large majority of the income might still be given to a small number of artists. The article notes that niche rock groups are making much less than classic groups, but it says its data is calculated on the back of a napkin.
It should also be noted that after the publishers are paid, the actual artists only get about %10 of what’s left. That would be about $500,000 for the artists of the top ten songs, or an average of $50k for each. Without data, we can only imagine how it goes down from there. Why isn’t such basic data available?
Composers supply the inventory that companies like Pandora sell. There is no way the inventory should be valued at one thousand times less than its retail price and no way for composers to sustain supplying the inventory at that rate. Do you think that those companies would be in business if the percentages were reversed, i.e.: with the companies receiving $1600 after facilitating 1.5 million plays? The owners of those companies are billionaires while unfairly exploiting the composers.
As we parse royalty numbers for musicians, we ignore entirely the lack of any payment to most writers on the web. It has become taken for granted that bloggers do their work for free and that online reprints pay nothing (zero, zip). The notion that micropayments be collected from readers goes nowhere. Advertising pays little on most sites. I’m glad musicians are trying to get paid for their content, but wish they’d extend their concerns to writing (some of which boosts their profiles and presumably sales).