The Subscription Penalty

I spent months asking independent artists what bothered them about music distribution. I expected to hear about AI policies, payout speed, metadata headaches. Those came up. But the answer I heard most often, in different words from different people, was that they resented paying every year for music that was just sitting there.

For a while I filed that under complaints about price. It isn’t about price. It’s about structure.

Rent on a finished thing

Most distribution runs on subscriptions. You pay an annual fee and your catalog stays live on Spotify, Apple Music, and everywhere else. Miss a payment and your releases can come down. A few services offer a “keep your music up forever” tier, but the default shape of the industry is rent.

A song is a finished asset. Once it has been delivered to the stores, the cost of keeping it there is close to zero. The file sits in a database. Nothing about it requires annual labor from the distributor. The artist pays annually anyway, whether the song earned ten thousand dollars that year or nothing at all.

If you release constantly, you never feel it. The fee spreads across a dozen releases and the math works out. Subscriptions are genuinely a good deal for high-volume releasers.

Most artists are not high-volume releasers. Most put out a few songs and then life happens. A day job, a kid, a stretch where nothing comes, a pivot into something else entirely. The music doesn’t stop existing and the subscription doesn’t stop billing, so every year the artist has to decide whether their old work is worth paying to keep alive.

That’s the subscription penalty. It isn’t a penalty on failure. It lands on inactivity, which is where most creative careers spend most of their time.

Why the model persists

Distributors didn’t land on subscriptions out of malice. Recurring revenue is predictable, investors want to see it, and it smooths out a business where most releases earn very little. From the distributor’s side of the table the model makes complete sense.

It’s worth noticing whose risk it manages, though. A subscription moves the uncertainty of the music business off the platform and onto the artist. The distributor gets paid whether or not the record performs. The artist pays either way too, and the artist is also the one who absorbed the cost of making it.

Artists have gotten used to this enough that they talk about their “distro bill” the way they talk about a phone bill. But a phone company is providing service every month. A distributor delivered a file once.

What the interviews actually turned up

When we ran customer discovery for ONCE, I expected the conversation to be about AI. Most of our releases come from AI-assisted creators and the noise around them is loud, so I assumed the pain would be about acceptance, takedowns, platform policy.

It mostly wasn’t. Their problem was everybody else’s problem, only sharper. AI-assisted creators tend to release in bursts: a prolific month, then a quiet quarter. A subscription serves that pattern badly. They buy a year of capacity and use it in two months.

That changed how I think about the market. I had been sorting artists by human versus AI. The split that actually predicts whether this model works for you is how evenly you release. Subscriptions assume a steady output that most artists have never had.

The alternative isn’t complicated

Charge once, per release, at the moment the work is made and the artist wants it out in the world. Distribute it. Let the artist keep their royalties. If they never release again, they owe nothing and the music stays up.

That puts the incentives back in a sane order. The distributor earns when the artist creates, so its growth depends on artists actually making things rather than on renewal dates.

The obvious next question is where the money comes from instead, and the honest answer is that it can’t come out of a cut of the royalties. A permanent percentage is the same rent collected a different way. That’s a longer argument than this post, but the short version is that the fee has to be the whole fee.

I’m building ONCE on this model, so I’m not a neutral party here. The model isn’t really the interesting part anyway. What’s interesting is how long the industry has charged rent on finished work without anyone calling it that, and how quickly artists recognize the alternative when they see it.

If you’ve ever opened a renewal email and done the math on your own catalog, none of this is news to you.