Okay, it’s perhaps unpopular at the moment, but I’ll say it right up front: Apples-to-Apples, on a per listener basis, streaming pays songwriters much better than traditional broadcast royalties.
Even the suggestion of discussing music streaming with a bunch of songwriters and musicians is a good way to polarize a group. In the grand scheme of things – music industry-wise – there’s been nothing like digital streaming services like Pandora or Spotify.
Streaming services are in some nebulous intersection of a Venn diagram where there’s music programming/curation, downloads, rentals, social media and concert promotion. The services assume pieces of each of these models, but not enough to be able to draw direct parallels.
This post was inspired by an “old” article (December 2014), but surfaced on my Facebook feed a couple of days ago. The Business Insider published a post entitled Pharrell Made Only $2,700 In Songwriter Royalties From 43 Million Plays Of ‘Happy’ On Pandora. The title of the article implying that Pharrell is getting an extremely bad deal.
Now, I’ll admit, 43,000,000 plays of a song certainly is impressive! But $2,700 for 43 Million plays is just fractions upon fractions of a penny per play (roughly $0.0000628 per play)??? The knee-jerk outcry from those following the story on the Internet was predictable: “Artists should be fairly compensated for their work!”
It seems to be almost fashionable to speak out against streaming services especially since a few big name artists (Taylor Swift and Björk for example) have recently denounced streaming services and removed their music from the on-line catalogues.
Are the rates unfair and is the artist being gouged? I don’t know. As I mentioned above, the music streaming model is unique and new ground to be broken.
Streaming is not the same as buying a CD or digital download – you don’t own a copy of the music (or more accurately, the licence to that music). I’ve heard streaming be compared to a rental service, but there again, in the midst of some similarities, there are differences that stem from the purely digital format of the product.
The third parallel I see drawn is between streaming and radio. This is evident in the number of articles (in addition to the Pharrell piece above) that quote the per-stream rate that songwriters and publishers get paid on their statements. The minuscule rates that streaming services pay are typically put up against broadcast rates that performance rights organizations (PRO’s) such as SOCAN, BMI, ASCAP and others pay. The PRO rates usually come out as amounts that can be calculated with just two decimal places – as opposed the eight or nine decimal places required for per-play streaming royalty rates.
Apples and Kumquats
That’s really not the end of the story.
The number of listeners that a single radio spin hits is magnitudes more than a single play on a streaming service. Spotify and Pandora are typically consumed by individuals with each streamed play being received by one person. The intent of a song being played on the radio is to be broadcast and heard by literally thousands of people at the same time.
A quick google search of US pop radio audience revealed that 56 million people listen to commercial pop-music radio per day. I posted the BI article on my Facebook page with the anecdote that 43 million streams of “Happy” equates to roughly less than one spin of the song on terrestrial commercial pop radio.
To paraphrase the question in my post:
What is the PRO rate broken down to a per-listener level so that streaming and radio rates can be compared
on a more equal (apples to apples) level?
The Rabbit Hole
Needless to say, finding royalty statements and figures on radio audiences is not an easy task. I’ll freely admit that there are an abundance of caveats to my calculations and assumptions. I tried to be as conservative as I could as I was dealing with information that was private and non-transparent.
My first task was to find some royalty statements where a song has been played on traditional radio and streaming services over the same time period.
Lo and behold, a songwriter by the name of David Lowrey (Camper Van Beethoven and Cracker) did just that. I found a post of his that included his royalty statements where he shows how much he got paid for plays of his song “Low” (performed by Cracker) by various radio and streaming services. Seeing as his statement was the only publicly available example I could find that showed a full set of royalty data for a song, his earnings are going to be the straw fodder for the straw-man that I’m about to build.
The song “Low” was streamed over 1,000,000 times on Pandora and netted him less than $17. Granted his songwriters share of the song is 40%, so in reality, the 1 million streams netted the songwriters just over $42. Over the same period of time, “Low” was played 18,797 times over terrestrial radio resulting in a total songwriter payout of $3,434.45.
$42 for almost 1.16 million streams versus $3,434.45 for 18,797 spins on radio? The numbers seem to illustrate that streaming services are ripping off the artist: allowing listeners to listen to music with virtually no compensation to the artist.
Or is it?
A single play on radio reaches literally thousands of people while a single stream generally reaches one person at a time. It really begged the question: What was the actual audience size that heard those 18,797 radio plays? And furthermore, how does that $3,434.45 translate to a per-listener rate?
This is where it kind of gets “fun”. Actual and accurate audience numbers of radio audiences are either closely guarded secrets or not really known to any level of accuracy. I suspect a bit of both.
I found a couple of sites (1, 2) that shed some light on radio station rankings and audience numbers for Vancouver. I settled on numbers that represent Vancouver (where I live) as a reference-point. The population of Vancouver is just over 600,000 (Greater Vancouver came in just shy of 2.5 million people) in 2013. In the grand scheme of things, when compared to cities in the US, Vancouver’s size would result in it just making it to the top-30 biggest cities by population in the US – a decent “average” city.
As for satellite radio audience sizes, after much searching, I found an obscure article from 2008 (before the merger of Sirius and XM) that gave listenership for a few programs. Without anything more recent, I thought that the sound of “Low” would probably fit in the “Pulse (’90s)” show (330,700 listeners in 2008).
Scribbles on a Napkin
I’ll be the first to admit that there are probably much more accurate numbers for these calculations, but assumptions were made and documented (below).
For David’s song “Low”, as per his published royalty statements,
The initial analysis seems to point to the fact that getting your song played on satellite radio is a freakin’ awesome deal and Spotify / YouTube / Pandora just plain suck.
However… After plugging in some (assumption ladened) audience numbers, the figures get adjusted to a play-per-listener playing field:
Terrestrial Radio Audience Size: I chose 102.7 “The Peak” as a representative radio station in Vancouver. It plays “adult contemporary” music – programming that I felt fit with the sound of “Low”. Compared to other stations in the city, the listenership of “The Peak” lands in the bottom half of the ratings lists – which in turn give the estimate a level of conservatism. The radio stations’ site combined with 2012 PPM estimates seem to suggest that the daily listener audience is somewhere around 70,000 people. I made the assumption that at any instant when a song is played (assuming it makes prime-time rotation), there may be roughly 10,000 people listening.
SiriusXM Satellite Audience Size: Again, I chose to err on the side of conservatism – sticking with 2008 pre-merger audience numbers even though SiriusXM subscriber growth has increased steadily over the past four years (~19 Million in 2010 to ~27 Million in 2014).
Streaming Audience Size: Although most streams can be assumed to be delivered to an individual, the reality is, a portion of those devices could be connected to speaker systems where there are more than one listener in attendance. I made the assumption that 50 percent of the time, there may be at least two people listening to a stream and used 1.5 as the audience size.
With the (again, assumption ladened) audience numbers inserted, in the analysis, satellite radio seems to be the worst deal for songwriter royalties, followed by terrestrial radio. Surprisingly, Spotify seems to be the highest per-listener songwriter royalty rate.
Other than comparing unadjusted royalty rates being a gross mis-representation of relative compensation, I don’t know if I can arrive at any solid conclusion as to whether streaming is beneficial to the songwriter or not. The numbers seem to suggest that on a per-listener basis, streaming is a better deal for songwriter performance royalties. That being said, depending on the source you reference, streaming constitutes anywhere between 10% and 15% of music consumption methods. Radio is still, by far, the dominant method of music consumption and discovery.
The question still remains though – What is streaming and is the pricing / compensation model valid?
I don’t know.
I’ve just compared streaming to radio by attempting to “equalize” the metrics. Streaming shares a lot of similarities with radio, but it is also, for the most part, controlled by the listener. So, how does the benefit of random-access control translate into prices and compensation? Are they optimal for the market or will they / should they shift?
It’s going to be interesting to see how this all shakes out.