bcc:

Can we listen our way out? 5/24/2021

In my cursory following of the work of the Union of Musicians and Allied Workers, something has been nagging me: are the Justice at Spotify demands sufficient? Specifically, can the changes in payments demanded by UMAW make a substantive impact on the incomes of artists?

Spotify currently distributes payments proportionately, based on a pool of revenue allocated to artists, with is then distributed based on relative listening volume (and probably some specific contract terms for certain artists and labels).¹ The way it shakes out is that most artists—certainly independent artists—earn around 1/3rd of a cent per stream. UMAW is calling on Spotify to pay at least 1 cent per stream, and to make pricing “user-centric,” which would mean that the revenue pool allocated to artists is only constrained by the amount of listening that users can do in a given month.² On the one hand, a penny per stream is triple the current rate; on the other hand, it’s only an additional 2/3rds of a cent. Does it actually matter?

To get a sense of what it might mean for an artist and a fan of an artist, I wanted to compare how Spotify listener activity and the royalties it would generate under a PPS model stacks up against a more traditional transaction (buying a record, cassette, T-shirt, etc.). A couple of friends agreed to request their Spotify data and send it to me so I could aggregate it and compare the pricing implications.³ Here’s a summary of the data I received from each person:

Based on what I know about them, Listener 1 is probably a typical listener and Listener 2 is a more intensive listener—by which I mean to say that Listener 2 is on the high end of the distribution in terms of # plays/year. If PPS were implemented, Listener 1 would have generated $82.17 for the artists they listen to, while Listener 2 would have generated $143.34.

For the individual artist, my estimate is that buying an album on Bandcamp or at a show nets around $4-5.⁴ The good thing about these transactions is that the artist gets the money right away, regardless of when or if the fan listens to the record. Whereas for a 10 song album, a fan would have to listen through 40-50 times to provide the artist a similar amount in royalty payments on Spotify via a PPS model.⁵

One potential upside for the streaming version is that theoretically there is no ceiling (under PPS) to how much an artist could earn from a listener in a year, and those earnings don’t have to be tied to releasing new music that the fan wants to buy. To earn $10 in two years from one fan in the traditional model, the artist has to sell 2 records, which probably means they have to write, record, and release two records in the same period. Under PPS, the artist could release only one album (or EP, or whatever), and if the fan really likes it, they could earn $4-5/year in perpetuity.

So the question seems to come down to: in a year, how many artists would get $4-5 “from a listener” under the PPS model, based on listening habits?

The answer seems to be: not many, and those who do are unlikely to be the working-class, independent musicians UMAW is advocating for.⁶ To illustrate, below are each listener’s top 25 artists by play count, along with total minutes listened to each, and the estimated PPS Revenue for the year.

 

For Listener 1—who I’m considering a typical or median listener—, no artist in their top 25 received enough plays to earn $4-5 under PPS. Listener 2 only had two artists break that threshold (and again, I suspect they are a rarer listener in terms of # plays). The other noticeable thing is that, in both cases, most of the artists in each listener’s top 25 are already quite popular. Listener 2 had some exceptions (e.g. Droning Muffler and Dig Nitty—both great), but by and large the artists in these lists are already well-established. For most Spotify users, independent working artists are likely to fall within the long tail of the distribution of artists, which means any one person’s listening habits will only provide these artists a few cents of revenue per year.⁷ In short, a PPS payment model is not a sufficient replacement for prior modes of individual fans buying music from individual artists that they like.

Of course, this argument is based on what I’m inferring from a couple of listeners’ data. They are essentially toy examples, but I think they are reasonable ones. I don’t have data from the artist side of things. So I can’t say for certain whether, for a small working artist, the breadth of casual or incidental listeners is enough to generate a reasonable amount of revenue at a penny per stream (the benefits of “scale,” as the companies like to say). Does breadth of access to many listeners compensate for the fact that most people don't listen to small artists obsessively enough to simulate a record sale? And does it work a penny per stream but not at 1/3rd of a penny per stream? Certainly PPS is better than the status quo, and maybe that’s the point. But I wonder how much it changes an artist’s assessment of the long term viability of pursuing their art. Suffice to say, I’m skeptical that it changes much.

Another argument might be that, in a given year, royalties form any one listener isn’t very much, but over time it accrues to larger lifetime earnings from each fan. That may be true, but then the question is whether money now is more valuable than money in the future. If you’re a working artist, trying to earn enough to make your next record, pay your bills, etc., it seems to me that PPS isn’t getting you there.⁸

All that said, what’s a better direction?

I’m intrigued by the Catalytic Sound project, written about semi-recently in Pitchfork. At the time, artists in the collective were earning something like $46/month—nothing huge, but that was with under 150 subscribers. Even growing the subscriber base to 1-2000 people could start to provide meaningful, predictable income for these artists.

I can imagine a future where “record labels” evolve towards something more like this collective model. They could operate more like a community, sharing fan revenue more evenly, helping to ensure sustainability for all the artists involved. Label-collectives could work together to build support for each other, through featuring artists from each other’s rosters, doing special collaborative releases, and bundling subscriptions.

Over time, a label or group of labels could chip in to develop an open-source framework that essentially creates a miniature streaming service for a given label (something like Ghost, but for music). It would allow labels to upload/release albums, curate and contextualize their releases, collect subscriber payments, and distribute proceeds across label members. This framework could be deployed on a label’s own servers, or cloud infrastructure that they purchase. Listeners may start out having multiple apps for the labels they support, but eventually the framework could incorporate APIs that allow a listener to access music from all of the label-collectives they subscribe to in a single application.

The fact that this seems to be working for a group of experimental musicians makes me particularly optimistic that it’s a system that can work more broadly. For artists operating on the creative fringe, or writing longer compositions, stream-based royalties isn’t going to work, and it’s unlikely that any immediately obvious, “works at scale” solution is going to be viable for them. By leaning into a more collective, community-focused framework, more artists have a chance of receiving real, material support for their creative projects.⁹

I don’t think Spotify is going anywhere, so in some sense I agree it is worth seeing whether it’s possible to improve their payment structure for artists, to increase their transparency, to fight their emotional engineering. But I also feel a bit pessimistic about the odds of success, or what success under the terms defined would really mean. I’m more curious and excited about what we might do if we imagine that services like Spotify didn’t exist.

¹ One of UMAWs demands is to make these special agreements more transparent, which I’d love to see. But it seems unlikely.

² It occurs to me that maybe a goal in all of this is to demonstrate that Spotify is actually incapable as a business of supporting a fair payment model for artists. Maybe without a cap on artist royalty payments, they go under.

³ I was a Google Play user until they shut down, then switched to…*checks notes*…Napster after reading somewhere that they offer 2 cents per stream (Which makes me wonder why we don’t try to move people to the platform that pays more, but as we’re about to see, even 2 cents a stream doesn’t seem game-changing to me).

⁴ This will be higher or lower depending on how artists price their stuff, or whether it’s a digital download or a cassette or a record, but this seems like a good average price range. People in bands who have sold stuff more successfully than I have can correct me, but I don’t think it’s much less than $3 or much more than $7, and within that range I think my arguments hold.

⁵ This is assuming a best-case scenario, where each listen counts. It seems plausible to me that Spotify has fraud algorithms that would actually punish this type of obsessive listening in terms of royalty payments.

⁶ I just noticed as I was posting this that Spotify in some ways corroborates this with their own data, through their recently-created "Loud & Clear" website

⁷ A tangential thread is that per-stream royalties turns listening into a kind of work. Listening to music generates money for the artists I like, and so my playing their songs supports them. No one thinks about it this way, probably because no one should, but each song play becomes some kind of micro labor/consumption, which then means there is always a little bit of capitalism going on when we listen on these platforms.

To be more specific, there is capitalism not just by having a Spotify subscription, but in the act of listening. Whereas I could say that going to a show or buying a record has some transactional quality to it, but once the initial transaction is complete (at the door or at the merch table), our listening can become something else, decoupled from the financial transaction that previously took place. In a space where every stream counts (whether it’s a penny or 1/3rd of a penny), that can never quite be true.

⁸ Maybe there’s some model out there where an artist can initially earn a higher amount in the first years of a release’s life, by constraining its availability to certain channels or direct purchases (I seem to remember Taylor Swift doing this with one of her albums). This would help to cover expenses and get them to the next record. After a certain period, PPS takes over and basically acts as some kind of long term earnings account for back catalog items. In a way, this is what is happening for older artists, who appear in both of my example listener’s top 25. Whatever Bruce Springsteen records or Brian Eno records my friends are listening to likely came out years ago, and now these artists are getting a steady trickle of revenue from people’s ongoing listening.

The idea seems interesting, but again I’m unsure how to make it viable on a service like Spotify, especially for working artists near the start of their careers.

⁹ I think it can also encourage label-collectives to take on a more diverse range of artists, since they will want subscribers to feel that they have access to a decent range of music that changes over time. There could be upstart artists who release digital-only first, giving them a chance to earn money and gain new listeners. It could reduce the need to consider the ability of an artist to tour when adding them to the roster.