Hacker News Reader

Comment by emilecantin | Hacker News Reader

I think a better payout structure for Spotify would be to go per-user. For example, for each subscriber a given amount from their subscription plan is earmarked for payouts. Let's say about 4$ per user. We divide that amount by the number of streams the user had that month, and we simply pay that amount to the artist.

Examples:

- I only listened to one song this month? The full 4$ go to that artist.

- I listened to 2 albums by 2 artists this month, on repeat? 2$ each (assuming the same number of plays for each)

- I listened to 399 tracks from one artist, and one track for another? 3.99$ to the first and 0.01$ to the second.

- The gym next door is streaming Mariah Carey's Chrismas album on repeat? Their 4$ goes to Mariah Carey and mine still goes to the artists I enjoy.

I think the current model of per-stream across the whole platform distorts the playing field towards more popular artists that play in commercial settings.

View on HN

>I think a better payout structure for Spotify would be to go per-user. [...] - I only listened to one song this month? The full 4$ go to that artist. [...] I think the current model of per-stream across the whole platform distorts the playing field towards more popular artists

Your idea has been suggested many times but as other sibling comments already noted, Spotify didn't have the leverage in business negotiations to implement that type of payout.

In other words, Spotify needed the vast song catalogs from the Universal + Sony + Warner -- more than those Big 3 Labels needed Spotify.

Your idea has higher payouts (i.e. "more fair") for independent musicians at the expense of less payout for the Big 3 Major Labels. In contrast, the current payout system favors the Big Labels at the expense of independents. Why did this happen? Because the Big Labels have more leverage.

Some factors leading up to the Big Labels negotiating from a position of strength to get more favorable treatment from the "weaker" Spotify:

- 2000: Napster lost its case against Metallica and RIAA

- 2000: mp3.com lost its case for "users CD-ownership-verify-then-playback service"

- 2003: Big Major Labels felt they got screwed by Apple & Steve Jobs iTunes deal with the flat 99 cents pricing.

By the time of 2011 licensing negotiations, they were already winning their lawsuits against the internet "stealing" their music. They were in no mood to give Spotify a sweet deal for access to their song catalogs.

This is why the no-name artists get fractions of pennies adding up a few dollars while Taylor Swift gets millions from Spotify. This unbalanced financial arrangement was the carrot to help get the Big Labels on board. To add to this, Spotify also gave the labels part ownership stock[1]. In contrast, smaller independents like CDBaby and TuneCore didn't get offered Spotify stock. Examples like that should give an idea of how much leverage the Big Labels had.

[1] https://thehustle.co/the-economics-of-spotify/#:~:text=The%2....

> In contrast, the current payout system favors the Big Labels at the expense of independents.

Is there an actual statistic around this? The linked article you provided doesn't have it. I'm curious:

1a) What is the average per-stream rate for a song from one of the Big Labels?

1b) What is the average per-stream rate for a song from one of the Big Labels, that is then passed along to the artist after the label takes their own cut?

2) What is the per-stream rate if you self-publish your song via DistroKid instead (that then publishes it to Spotify), that goes directly to the artist?

It's not obvious to me that 1b) is larger than 2). Maybe it is -- but are there actual numbers here? Especially for "average" or "median" artists at a label, not Taylor Swift.

Not sure why you're getting downvoted, I'm wondering something similar. Anecdotally (and maybe incorrectly), I'd expect people who listen to a lot of indie music to stream more songs per month than those who listen to pop & big label music, which means that big labels would be making less money if they get paid per stream than per user? Obviously there are probably many other factors, I'd love to see the numbers.

I find the explanation easier to believe that "per stream" is a much simpler & transparent system for the artists that causes less outrage towards Spotify[0]. And to cater to big labels, Spotify just increases their specific per-stream rate (as parent's article explains).

Still, it causes issues like the botting ones in the OP. I don't know, it's really hard to make a conclusion here without the actual numbers.

[0] Imagine you get more streams but less money than your friend with the same indie deal, just because his listeners listen to less music overall. You could convince ME that it's fair from an outsider perspective... but probably not an artist who's trying to make a living.

I think your last bullet point is the one that sticks for me the most. If I only listen to independent death metal then why is my money going to Taylor Swift or Joe Rogan or whoever.

With per-stream pricing, it only does if you don't stream enough songs. If you listen to much more music than the average Spotify user, then all your money will go to independent death metal, plus someone else's will also go there even if they only listened to Taylor Swift and Joe Rogan. Of course, there's still the the fact that big labels have better per-stream-rates than indies, so you'd have to stream two-ish-or-whatever times the average amount.

And honestly, I don't know if that's bad for indie? I mentioned this elsewhere on this thread but anecdotally my friends who listen to a LOT of music mostly listen to indie, not big label. Maybe I'm wrong/my bubble doesn't generalize? I'd love to see numbers if anyone has them.

I don't think this matters at all.

Whether royalties are calculated per-play-per-user or per-play-per-all-users, I don't see any evidence there's any meaningful difference at all.

Most people who subscribe to Spotify listen to lots of different stuff -- that's the entire point of streaming. Virtually nobody is listening to one song a month, or 2 albums per month. They listen to playlists and radio and their 8,000 liked songs on shuffle. None of the examples you give are very realistic. (Somebody who only listens to one album is just going to buy the album, not subscribe to Spotify.)

In the end, the two models are going to average out to essentially the same.

If you don't think they do, that means you think there's some correlation between the amount of music listened to per month, and whether a person's tastes skew more mainstream or niche. Can you show that's the case, and in which direction? Because that's entirely non-obvious.

(And commercial settings like stores are a drop in the bucket of Spotify subscriptions, they just don't matter. Big places like grocery stores and department stores and fast food don't use Spotify, and coffee shops or clothing boutiques are just as likely to be playing niche stuff.)

"Rounding errors" like these always favour the larger players. It's a form of slowly extracting money that should have gone to smaller artists and instead gets funnelled into the pockets of major label artists who arguably don't need it, at the expense of smaller artists and especially niche genres that don't have the same mass appeal as pop music. That's a group that's severely lacking representation in the discussions around how to pay fairly.

Another way they're screwing smaller artists is next year they'll be requiring each song to have 1k streams _per year_ before qualifying for monetization. They estimate that'll steal another ~$40M/year from smaller artists to help line the pockets of majors. It's actually sickening to watch from the sidelines.

> "Rounding errors" like these always favour the larger players.

I don't see it -- how?

> requiring each song to have 1k streams _per year_ before qualifying for monetization

Each stream is ~$0.004. It doesn't seem unreasonable at all to require a song to hit $4 per year -- or $0.33/mo. -- to qualify. There is some level of overhead involved here, after all. These are ludicrously small amounts of money.

Also how does this "help line the pockets of majors"? It's all just going back into the same pot of revenue that gets shared the same as before. So proportionally it's helping the artists with smaller songs still over 1K streams just as much.

And even if it's $40M, that's just 0.3% of Spotify's overall revenue. Proportionally, it's pretty irrelevant.

This isn’t per artist, it’s per song, which means an artist might put out an album and only the single gets widespread play and if the few thousand plays the rest receive all fall just under 1k they’d be discounted to unpaid plays. Remember, for smaller artists every dollar counts. Most have already seen their revenues from music drop by 90-95% since Spotify ate the industry, so why should they be so willing to accept that the first 1k plays are worth $0.00 and the next are only worth $0.003 or less.

This is also compounded by their refusal to adopt a user-centric payment model where the artists you actually listened to that month are the ones your money goes to. Look at comparisons others have published and you’ll see how their current model siphons money from smaller artists towards major label artists (I’m discretely typing this in bits from our local chamber orchestra’s AGM and can’t look up facts for you atm).

With this change, it ensures that even moving to a user-centric model they’ll still be able to underpay smaller artists. It’s not so alarming on its own, but it ensures they can continue to keep the status quo even if they give in to pressure and adopt a user-centric payment model by being able to not have to pay for much of their catalog.

But the artists never made any money in the first place. Spotify allowed for the insanely long tail of no-play artists to get a platform to begin with, paying a nominal fee seems like not that big of a deal, considering the same folks would never have gotten any exposure.

Artists had plenty of thriving platforms to get their music out digitally before Spotify came along. The difference is almost every one of them paid better or let you sell your music. Spotify has accounted for artist music revenues dropping by ~90% and has done almost nothing to help smaller artists that they weren’t already getting elsewhere. They grew in popularity due to offering listeners endless playlists. Sadly, Songza was killing it in this area too and could have been a serious Spotify competitor until Google bought it and ruined it almost overnight.

> Most people who subscribe to Spotify listen to lots of different stuff -- that's the entire point of streaming. Virtually nobody is listening to one song a month, or 2 albums per month.

Considering the asks you're making, I imagine you'll have no problem substantiating this?

For a while Deezer was pushing a model similar to this, they called it User Centric Payment System: https://musically.com/2019/09/11/deezer-steps-up-its-efforts...

Here's another article on it: https://www.engadget.com/2019-09-11-deezer-royalty-system-sm...

Unfortunately, they've since changed it to a model that the labels wanted.

> I think the current model of per-stream across the whole platform distorts the playing field towards more popular artists that play in commercial settings.

Which is very much in favor of the actual people (Major labels) who needed to be convinced to license all their artists to Spotify in the first place. So not really surprising.

It does create weird incentives though. You make more money not by people listening to your music more but buy the people that do listen to you listening to other artists less.

I can already see the internet:

"Don't listen to your discover weekly or new music friday it hurts the artists you like."

And so the diss track is finally justified.

I think you can fix that incentive issue relatively easily by allowing people to choose the allocation arbitrarily after some sort of minimum tax, e.g.

The first $4 of your monthly sub is divided amongst all artists you listen to based on time spent listening, and any additional money you decide to add to your subscription is distributed to artists as you see fit. You can elect to use the same distribution algorithm as the first $4 of your subscription, use the same algorithm with different weights (e.g. your favorite artists listen time is doubled before payout is computed) or you can choose to have an arbitrary percent of it go to whatever artists you want, even if you don't listen to any of their music. Spotify already has to have a payment engine to support paying all the artists anyway, generalizing it beyond fixed subscriptions seems like an organic way to address the issue of unfair income distribution if they were interested in doing so (I don't think they are).

In the above system, the issue of "don't listen to other artists" only comes up if you don't have enough money to give to the artists you want to support, no different than the incentives of "don't buy the CD of artist A or you won't be able to afford the CD of artist B, who you like and wish to support more."

This solves one problem, in that I the consumer can be confident that my money is going to the artists I enjoy. I don't think it solves another problem, which is that you have a flat-fee being split many different ways. Most people don't listen to just one artist, they may shuffle their favorite songs or listen to a discovery playlist to find new favorite songs. Such a system isn't going to send $2 to 2 different artists per customer, it's going to send fractions of pennies to potentially dozens of different artists. Worse, the more music a person consumes, the more diluted their subscription fee becomes.

It doesn't really address the problem of paying artists per stream, it just means that the per stream rate is variable and dependent on how willing your fans are to play your music on repeat each month.

It's still probably better than the current model, but the only way I'd really be comfortable with it is if we cut out the middle-men. For example, a service owned and operated as a non-profit (or profit-capped) organization owned by an artist collective. The artists set the monthly price based on what makes sense for them to make a living, and the cut given to the service can be minimized to what it takes to run and build the service rather than what it takes to make a profit on top of the work of the artists.

While I've also always thought about this model, I'm curious as to why Spotify doesn't do it? Is it hard to calculate the revenue from a single user? Or is just in their interest that it's complicated? (Btw it should probably be listening time and not nr of tracks, or we only get 30s tracks going forward).

Also, would law of big numbers make it so that this kinda evens out like this anyways with today's system?

The question is who decided this payout structure? Is it Spotify (I clearly don't know why they would care any other way) or the big artists (I can see why they would want to be paid more as a condition of their presence on the platform, because of greed but also because of legitimate reasons).

This would incentivize artists to release more, shorter tracks.

And if that is what users want, then those tracks will get a lot of plays, and by extension payment.

Shorter tracks would also mean users have more opportunity for listening to a variety of songs/artists. A single user can only listen so much in a day (capped at 24 hours, and more likely 12 hours if they aren't doing some kind of phantom streaming). This would then ultimately dilute the payments to artists releasing a large number of shorter tracks.

Could use minutes played instead of tracks played then.

As far as I know this is more or less the payout structure Tidal uses.

This seems a better approach, but in reality I don't know if it is actually better.

If I understood their plan correctly, it was only the higher $20 tier that did it, and I think only $5-10 was actually split between your top 2 artists for a given month. They ended that program a few months ago.

I'm sure that smarter people than I will point out flaws in this model (aside from Spotify making less money lol) but I love it.

It would be split into tons of artists for most people but a guaranteed chunk of the monthly would be nice.

Reply on news.ycombinator.com