October 2026
There is something fundamentally wrong with the way music streaming royalties are distributed. At least, that's the state of the music industry discourse in October 2026.
The villain is the pro-rata model, also known as the market-centric payment system. Under this model, streaming platforms pool revenues from subscriptions and advertising and distribute them to rights holders based on their share of total streams.
The main criticism is that the biggest artists, who already dominate streaming charts, capture a disproportionate share of the money. Even from subscribers who never listen to them.
The proposed solution sounds almost too obvious: user-centric royalties. Instead of pooling everyone's money, each listener's contribution would be allocated exclusively to the music they actually consume.
What's not to like?
The argument against pro-rata usually comes down to three things.
First, fairness. Why should my subscription help finance artists I don't listen to? If I'm paying for music, shouldn't the money follow my personal preferences?
Second, the distribution of wealth. Pro-rata is widely criticized for disproportionately benefiting the biggest artists, whose enormous streaming numbers allow them to capture the largest shares of the royalty pool. User-centric, its supporters argue, could redirect some of that money toward smaller artists with dedicated audiences.
Third, fraud. Under pro-rata, artificially generated streams can increase an artist's share of the overall royalty pool, potentially taking money away from legitimate artists.
These arguments have attracted some prominent supporters. Deezer publicly championed user-centric royalties in 2019, while SoundCloud went further and introduced its version, called fan-powered royalties, in 2021. Both presented the approach as a way to make royalty payments more closely reflect fans' preferences and improve compensation for independent artists.
But there's a problem with this argument.
The first claim is about what we consider fair. The other two are about what actually happens to the money.
And those are very different questions.
The debate has generated a surprisingly substantial body of research.
Researchers have examined user-centric royalties using real streaming data, theoretical economic models and simulations. Studies have been conducted in Denmark, Finland, France and other markets, with some of the most important contributions coming from Roskilde University, Digital Media Finland and France's Centre national de la musique (CNM).
The results are interesting, although not nearly as straightforward as the public debate might suggest.
A 2017 study by Digital Media Finland, commissioned by the Finnish Musicians' Union and published in 2018, compared pro-rata and user-centric distribution using Finnish streaming data. It found that user-centric allocation could redistribute revenues away from the most popular repertoire toward less popular artists. The study suggested that the biggest beneficiaries would not necessarily be the smallest artists, but those somewhere between the most popular and least popular recordings.
In 2020, researcher Rasmus Rex Pedersen reviewed four empirical studies in A Meta Study of User-Centric Distribution for Music Streaming. The review found that the results varied between countries and datasets, but generally suggested some redistribution away from the most popular repertoire. Importantly, these studies were not unanimous about which artists would benefit.
Then came one of the most comprehensive investigations.
In 2021, France's CNM, working with Deloitte and using data supplied by Spotify and Deezer, published a study comparing the two models.
Its findings appeared to support part of the argument for user-centric royalties. The study estimated that the top ten most-streamed artists would collectively lose around 17% of their revenues under user-centric distribution, while the top 10,000 would lose approximately 0.4%. It also suggested that rock and pop would benefit at the expense of rap and hip-hop, and that older catalogue music could gain relative to newer releases.
But there was an important qualification.
For artists ranked below the 10,000th most-streamed artist, the average financial benefit was estimated at no more than a few euros per year.
That's not exactly the revolution in artist compensation that some people might have imagined. But it also raises another question: perhaps chasing a few additional euros for the long tail of artists is counterproductive, and the debate about redistribution should focus primarily on the smaller group of large and mid-tier artists, for whom the financial difference could actually be meaningful.
That doesn't make the principle irrelevant. But it does raise the question of whether the industry is focusing on the problems with the greatest economic impact.
More recently, a 2024 study published in Information Economics and Policy analyzed 890 million streams from 154,505 Deezer subscribers in France. Its findings provided a more detailed picture of who could win and lose under user-centric distribution.
Superstars would lose approximately 6% of their revenues, while rap and hip-hop would lose 13.2%. Pop would gain around 10%, rock 7%, classical music 19.7% and jazz 10.1%. Recordings more than ten years old would gain 5.1%, while new releases would lose approximately 3%.
At first glance, this seems to confirm the argument that user-centric royalties reduce the dominance of superstars.
But the study also found that the least popular artists could lose revenue, while the principal beneficiaries would be artists in the middle of the popularity distribution.
In other words, the alternative model could benefit established mid-tier artists while making life even harder for those at the very bottom. It could also redistribute income away from emerging rap artists toward older recordings and more established genres.
The research was showing that user-centric royalties could change the distribution of money. It wasn't establishing that the change would necessarily be better for every small artist.
Here's where things become uncomfortable.
Even if user-centric royalties sound intuitively and ethically fairer, some research supports this mathematically. But that doesn't mean they will produce the financial outcomes their supporters expect.
We might introduce a new system, celebrate that fans' money finally goes to the artists they listen to, and discover that some of the artists we wanted to help are actually making less.
It's a little like changing the tax system to fund environmental initiatives according to individual taxpayers' preferences. Everyone feels better because their money supports causes they personally care about.
But imagine the result is that subsidies for electric vehicles decline and more people buy petrol cars.
The new system might feel fairer to the individual taxpayer while producing the opposite of the intended environmental outcome.
The analogy isn't perfect, but the distinction matters: a fairer-looking allocation mechanism from the listener's perspective doesn't necessarily produce a fairer distribution of outcomes for artists.
And there is research suggesting that this is more than a hypothetical possibility.
Glenn McDonald, Spotify's former Data Alchemist and author of You Have Not Yet Heard Your Favourite Song, approached the question from a different angle.
Instead of starting with the assumption that user-centric payments would benefit smaller artists, he looked at the relationship between listening intensity and artist popularity.
Under pro-rata, heavy listeners have more influence over royalty allocation because they generate more streams. Someone who listens to 1,000 songs a month effectively has twenty times the influence of someone who listens to 50, assuming all streams are counted equally. Under user-centric, both subscribers would allocate approximately the same amount of money if they paid the same subscription fee.
But why should they necessarily have equal influence?
The user-centric approach resembles the old economics of buying CDs. Nobody cared whether you listened to an album once or a thousand times. The artist was paid when you bought it. Streaming changed that relationship by making consumption measurable, so why should we automatically preserve the old logic? And if heavy listeners generate substantially more consumption, should they perhaps pay more? That would raise an entirely different debate about subscription pricing, although neither royalty model requires consumption-based tariffs.
More importantly, the different weighting of heavy listeners creates a potentially counterintuitive outcome for artists.
The critical question is: what kind of music do heavy listeners consume?
If people who listen to 1,000 songs a month mostly listen to superstars, pro-rata gives those superstars disproportionate influence.
But what if the opposite is true?
McDonald reports that, when he examined this relationship at Spotify, heavier listeners tended to listen to less popular artists. His reported measure of relative artist popularity was approximately 0.83.
That suggests a rather inconvenient possibility.
Because heavy listeners contribute disproportionately to the stream count, pro-rata may already direct more money toward less popular music than a user-centric system would.
And perhaps this helps explain another curious feature of Spotify's business. Why does it continue to accommodate the enormous volume of DIY music uploaded through distributors like DistroKid, despite the copyright disputes and outright fraudulent releases that can come with such an open ecosystem? Could it be that fans of smaller, independent artists are also among the most engaged listeners, and therefore more willing to pay for Premium subscriptions?
That would make the long tail valuable to Spotify for reasons extending well beyond the royalties those artists generate. Of course, this is a hypothesis, not evidence of Spotify's actual motivations. And legitimate DIY artists shouldn't be confused with the bad actors exploiting the same distribution infrastructure.
The very mechanism critics describe as unfair could be redistributing revenue toward smaller artists.
To be clear, McDonald's observation doesn't prove that every smaller artist benefits from pro-rata, or that the effect amounts to a 17% increase in their royalties. The underlying data and calculation aren't publicly available for independent replication.
But it challenges the central intuition behind much of the user-centric movement.
Perhaps heavy listeners aren't just listening to more music. Perhaps they're also discovering more music.
And perhaps pro-rata rewards that discovery.
This is probably the most interesting limitation of McDonald's argument.
His finding describes listening behavior at Spotify during a particular period. But why should we assume that the relationship between listening intensity and artist popularity remains constant over time, or that it applies equally to other platforms?
Streaming isn't always an expression of active musical preference. Listening can be driven by background playlists, algorithmic recommendations, mood-based consumption and habitual repetition.
Someone who actively explores underground jazz, experimental electronic music and independent hip-hop is one kind of heavy listener.
Someone who leaves an algorithmically generated playlist running throughout the working day is another.
Both might generate hundreds of streams, but their listening behavior could be fundamentally different.
And this is where streaming algorithms become relevant.
Imagine two songs within the same genre that are equally likely to keep a listener engaged. One comes from a major artist with an established audience, the other from an emerging musician.
If both perform equally well on engagement metrics, what determines which one the algorithm recommends?
We shouldn't automatically assume that platforms prefer the more popular song. Recommendation systems optimize multiple objectives, and popularity is not necessarily the deciding factor.
But if algorithms systematically favor already popular repertoire, passive listening could gradually shift the relationship McDonald observed.
There is also the question of catalogue music. Older recordings have accumulated years of listening data, while new releases need to build those signals. This could potentially favor catalogue in some recommendation environments, although it doesn't establish that algorithms generally prefer older music.
Interestingly, the 2024 Deezer study found that user-centric allocation itself would favor recordings more than ten years old over new releases, regardless of any assumed algorithmic bias.
Under pro-rata, these patterns would matter enormously.
The system would amplify whatever music heavy listeners consume, regardless of whether those streams reflect active discovery or passive consumption.
This makes McDonald's finding a hypothesis that should be tested repeatedly, rather than a permanent characteristic of streaming economics.
There's another reason to be careful about generalizing these results: Spotify, Deezer, SoundCloud, Apple Music and YouTube aren't necessarily the same kind of listening environment.
SoundCloud has historically played an important role in independent music, remixes, DJ culture and emerging scenes. Spotify combines active music discovery with large-scale playlist consumption and algorithmic recommendations. Deezer operates with its own audience composition, recommendation systems and market footprint.
Apple Music introduces another variable: listening through Siri and voice commands. Asking a smart speaker to play a particular artist is one thing. Asking it to play something relaxing, popular or suitable for dinner is another. In the second case, the listener delegates the choice of music to the platform, potentially changing which artists receive streams and how those streams are distributed.
YouTube is different again. Music consumption there is mixed with video discovery, recommendations, user-generated content and advertising-supported viewing. A listener might deliberately search for a song, discover it through a music video or encounter it in someone else's content. Those behaviors, and the revenue generated by them, don't necessarily resemble a conventional paid music subscription.
Even if all these platforms applied exactly the same royalty model, they could produce very different redistribution outcomes.
And in reality, they don't even use the same models.
SoundCloud introduced fan-powered royalties in April 2021 for eligible artists monetizing directly through its platform. Under this system, payments are linked to individual fans' listening activity and the subscription or advertising revenue associated with those fans.
The company subsequently reported that more than 56% of eligible artists earned more than they would have under pro-rata. In 2022, SoundCloud also claimed that participating independent artists earned approximately 60% more on average under fan-powered royalties.
Those numbers sound impressive, but they leave some important questions unanswered. What happened to the remaining 44%? How were the gains distributed between smaller and larger artists? And what if the biggest beneficiaries were already the most successful artists within SoundCloud's independent community?
Without a detailed breakdown of winners and losers by artist popularity and income, these figures don't establish that fan-powered royalties actually reduce inequality. They show that many eligible artists benefited, but not necessarily the ones the system was supposed to help most.
Deezer, meanwhile, took a different route.
After initially advocating user-centric royalties, it introduced an artist-centric payment system in 2023, developed with Universal Music Group.
Rather than distributing each subscription exclusively according to the subscriber's listening, Deezer introduced additional weighting for artists meeting certain engagement thresholds, as well as limits on excessive streaming and measures against fraudulent or functional audio.
The thresholds themselves are interesting. To qualify for additional weighting, an artist must generate at least 1,000 monthly streams from 500 unique listeners. That hardly sounds like a system designed specifically to help the smallest artists. It looks at least partly like an attempt to distinguish established audiences from low-quality or potentially manipulated consumption.
By 2025, Deezer reported that approximately 85% of its recording royalty pool was covered by the artist-centric system. The company has promoted the approach as improving remuneration for genuine artists and reducing fraud, but detailed independent evidence showing how different artist segments have benefited remains limited.
If SoundCloud reports improvements for certain independent artists while Deezer promotes its artist-centric approach, we cannot simply combine those results and conclude that user-centric royalties work.
They are different systems, applied to different audiences, with different rules.
The underlying consumption patterns might be just as important as the allocation formula.
Fraud prevention is one of the strongest arguments in favor of user-centric royalties.
Under pro-rata, a fraudulent stream can increase the recipient's share of the overall royalty pool, potentially diverting money generated by legitimate subscribers.
Under user-centric, the revenue attributable to an account is generally limited by that account's own economic contribution.
Consider someone operating a bot account that generates 100,000 streams for a single artist.
Under pro-rata, those streams could influence the allocation of the entire platform's royalty pool.
Under user-centric, a paid account could generally direct only its own allocable subscription revenue to that artist.
That makes certain forms of fraud much less attractive.
But does that make user-centric immune to fraud?
Not necessarily.
For a paid subscriber, the economic contribution is relatively predictable. Advertising-supported listening is more complicated because revenue depends on factors such as advertising impressions, geography and advertiser demand.
If a fraudulent account generates additional legitimate billable advertising impressions, it could potentially increase the revenue associated with its listening activity even under user-centric distribution. Whether those impressions would actually be accepted and paid for by advertisers is another question.
This is particularly relevant because ad-supported accounts can be attractive for fraudulent activity: they generally do not require the same recurring subscription payments as premium accounts. However, there isn't sufficient public evidence to conclude that most streaming fraud occurs on ad-supported tiers.
And the scale of fraud is real.
In January 2026, Deezer reported that approximately 8% of all streams across its catalogue in 2025 were fraudulent, while up to 85% of streams on fully AI-generated music were identified as fraudulent. The company said it demonetized detected fraudulent streams and excluded fully AI-generated music from algorithmic recommendations.
These figures are striking, but they don't demonstrate that Deezer's artist-centric allocation model itself reduced fraud. The company also uses detection and enforcement systems, and the effect of those measures cannot easily be separated from the royalty formula.
SoundCloud similarly argues that fan-powered royalties reduce incentives for bot manipulation, but publicly available independent evidence quantifying the reduction attributable specifically to user-centric allocation remains limited.
There is another potential complication.
Imagine a small artist with a few extremely dedicated fans who listen to the same songs repeatedly. Under user-centric, those fans' payments would be directed primarily to that artist, which is exactly what the model is supposed to achieve.
But concentrated, repetitive listening can also resemble some patterns associated with artificial streaming.
If platforms respond by aggressively flagging unusual activity, could legitimate niche artists and their most loyal fans become collateral damage?
This is a potential risk, not a demonstrated consequence of user-centric royalties. But it raises the question of whether stronger fraud enforcement might create new problems for the very artists the system is intended to help.
Recent theoretical research also challenges the assumption that pro-rata is inherently defenseless against fraud. In his 2026 paper On Click-Fraud Under Pro-Rata Revenue Sharing Rule, economist Hao Yu modeled conditions under which fraudulent streaming would not be economically attractive even under pro-rata. The result depends on assumptions about the costs and effectiveness of manipulation, but it suggests that fraud prevention is not determined by the royalty allocation formula alone.
Perhaps the solution to streaming fraud is better fraud prevention, rather than a complete redesign of royalty distribution.
After years of research, public debate and actual experimentation, the answer remains surprisingly complicated.
User-centric royalties have an obvious appeal: listeners can feel that their money directly supports the artists they love.
That is a legitimate argument for changing the system.
But it is different from claiming that user-centric royalties will necessarily increase incomes for smaller artists, reduce inequality or solve streaming fraud.
The available evidence suggests that outcomes depend on listening intensity, audience composition, artist popularity, platform design and the specific allocation rules.
Some artists could benefit. Others could lose.
And the artists who lose might be precisely the independent and niche musicians the reform was supposed to support.
Which leaves me with a question.
If user-centric royalties aren't necessarily better for smaller artists, and under some conditions could actually make them worse off, is making listeners feel better about where their money goes a good enough reason to change the system?