music: download vs. streaming

The Wall Street Journal had a short article on Christmas Eve about the music business.  I can’t see any way to use its contents to buy or sell a particular stock today.  But I thought it was interesting, anyway–and it has some bearing on the issue of owning vs. renting, which I think it a key way in which the internet is changing the world.

Here are the facts, obtained by the WSJ from a “major record company”:


The firm in question can identify how many individuals access its content, both by downloading albums/individual tracks from services like iTunes, and by listening on services like Pandora or Spotify.  In the latter case, the music company gets a miniscule payment from the service–much less than a penny–whenever a subscriber listens to a track.

For this music company, a downloader/CD buyer in the US generates $14 in yearly revenue.  A premium streaming service subscriber (still a very small number in the US) is worth about $16.

In Sweden, where over 60% of the population uses streaming services (3x the proportion in the US), a streamer is worth $17.75 a year; a downloader generated under $4.  Although it doesn’t explicitly say so, the Journal seems to me to be clearly suggesting that this is where the world is headed.


As you’d expect, downloading is very strong during the initial release of an album, when marketing spending is highest.  Streaming, on the other hand, usually builds slowly.

“Many” albums “eventually”–no quantification of number or time-frame–make more money from streaming than from downloads.

The sense I get is that albums fall into three groups:

–blockbusters, where initial download sales are very high and where streaming begins to contribute more than half of total revenue within a few months

–successes, where download sales are acceptable and where cumulative streaming revenue ultimately surpasses download sales–but only after a number of years.

–duds, where download sales are mostly a function of marketing hype and where both revenue streams dry up as soon as marketing spending ceases.


There’s really not enough data to be sure, but…

…as usually is the case in the physical world, renting looks like its ultimately more profitable for the owner of the property than selling, even when the transactions are done in bits and bytes

…the current mad rush to create new streaming services seems to validate the Pandora/Spotify model

…the value of the music company backlist has been severely underestimated

…the fact that the Journal many albums achieve streaming success, not “must” suggests there are lots of duds that can be the focus of cost-cutting.

Leave a Reply

%d bloggers like this: