Is there a fairer way for streaming services to pay artists? It’s an oft-asked question that hangs over the music industry, yet relatively little is done to answer it on a systemic level. Certain platforms have been doing their part to fly the flag for other models. In 2021, Soundcloud announced they would be shifting to user-centric royalty distribution for a portion of their independent creator base. More recently, in September of this year, Deezer, a long-time advocate for alternative models, revealed a new blueprint for an experimental ‘artist-centric’ streaming model in partnership with Universal Music Group. And last week, payment models hit the headlines again when it was revealed that Spotify is changing its royalty model from Q1 2024 in order to combat streaming fraud and shift $1 billion in payouts towards ‘working artists’ over the next 5 years.  

In this in-depth interview, we speak to Ryan Rauscher, author of Payment Option Transparency, a research paper that outlines the wide ranging impacts of the user-centric model on music streaming income. Commissioned by Germany’s PRO MUSIK, an association for freelance musicians, and published in June 2023, the paper is an essential read for anyone looking to educate themselves on the significant impact a user-centric payment model would have on artists, songwriters, and the industry at large.

Hello Ryan, thanks for taking the time to talk to us today. To kick things off, can you tell us how the research paper into alternative payment models has been received by the industry?

RR: It’s been positive in that it’s been featured in a number of international publications such as Music Ally, Music Business Worldwide, Digital Music News, Hypebot. Overall though, I think it deserves a lot more attention. There’s some key insights in there that I think are important, and the first one is very simple, which is that an alternative payment model like user-centric can actually have a significant impact on the income of artists and songwriters. This was one of the main motivations for creating the report in the first place – to counter the narrative from previous research that a user-centric model wouldn’t actually change much. 

I knew from previous analyses that I’d conducted, that the results from other papers and studies were being interpreted incorrectly. So the takeaways that came out of them were misleading – not on purpose – but they led a lot of people to believe that there wasn’t a lot of impact from an alternative payment model. We decided to do our own research in order to show the actual impact of alternative payment models, and to put user-centric and potentially other models back on people’s radars. The most important thing is that artists, songwriters and everybody who represents them needs to be talking about alternative payment models a lot more and take this very seriously.

What led you to focus on the user-centric payment model in the report? 

RR: User-centric was really the main form of an alternative payment model that was discussed within the industry at conferences and at DSPs when we set out for this project. Starting with Deezer. But then SoundCloud implemented it as well; Tidal wanted to implement it. A lot of research was being done and it was the main alternative model that was being analysed, so it made sense for us to focus on this model. 

In the beginning, we were planning to not only look at user-centric, but at listening time as a payment criteria instead of stream count as well. We did calculations there and also used the amount of listeners as an alternative or additional criteria. We were planning on analysing the impact of different kinds of models, but we figured that user-centric is complicated enough. Again, the real message is that everybody needs to understand that the impact of other models can be significant. And we all need to start educating ourselves about various models, and collectively discuss what the pros and cons of each of them are.

Do you think that the report has moved the needle in terms of starting much-needed conversations in the industry? You said it’s put other models back on people’s radars.

RR: Well, that’s the goal. But what’s been achieved with releasing the paper is essentially having a factual basis to point to. As I said before, I did previous analyses before this paper, and I was personally convinced that payment models are one of the most, if not the most, important topic facing artists and the industry. I wanted to convince everybody of this. 

The problem, again, is that there is research out there saying it has no real impact. I realised that the first step is to actually show that there is in fact an impact, so that’s what the study does. Now the real conversations have to begin. A decent amount of people have been made aware of these results, but it should be known to a lot more people. I think there’s still just a very small number of artists, managers, people working at labels, publishers that even know how important it is.

It does feel like there’s a disconnect in terms of discourse around “fixing” streaming and actually looking at the ways to do that.

RR: Yes, well, first of all, there’s a lot of topics when we’re talking about the inefficiency or the potential unfairness in music streaming. Those discussions are located all along the value chain of music, starting with how much consumers pay for their subscription; then how much the streaming services collect for their services; how much the labels and publishers share with their artists, and other things. The payment model is really just one part of that, though I would argue the most important one. 

It’s important to clarify at what stage in the value chain the payment model sits. Then, even if you were to just focus on the payment models, there’s so many narratives and myths about what the impact would or wouldn’t be. There’s still people out there claiming that user-centric would move money from mainstream artists to niche artists, which it does not do. Then there are the claims about the impact on genres and how they are impacted differently by user-centric. But there’s no real direct correlation there.

There’s a lot of complex issues here. While we’re here, we urge people to read your Debunking Myths series on your Linked In because there are a lot of myths when it comes to things like streaming fraud, and they are addressed there. 

RR: Thank you. The complexity of the streaming issue also comes from the definition of the problem, because that’s something that hasn’t really been talked about enough. If you want to fix something because you’re saying something’s potentially wrong and needs fixing, well, you need to clarify and define what actually needs to be fixed, and what the actual problem is. 

I focused the first couple of articles in my Linked In series on what the actual problem with our current payment model is. That’s the first step to really being able to get your head around whether something like user-centric or artist-centric would be potentially favourable in comparison to today’s model.

You mention first steps – are there any simple solutions that would go towards starting to fix the problem? For example raising subscription prices for DSPs.

RR: Higher prices, if consumers are willing to pay them, are good for music, for all the creatives, for everyone in the industry and they are long overdue. But we have to be clear that subscription prices are on a completely different step in the value chain from the payment model. The best way to describe this is that you can pour more into the revenue pool by increasing prices and having consumers pay more. But if you have a problem with the distribution of the revenue and your diagnosis is that you’re allocating in the wrong way, that will lead to money going into the wrong places and not the right places. If that’s your diagnosis, you can pour double the revenue in but you still have the same problem. So in a way, while it’s good that there’s more revenue, more revenue also amplifies the problem in money distribution. You need to solve or improve both. I would argue that the current pro rata model is so flawed that this is even more urgent than putting more money into the pool.

So no quick fixes here then… 

RR: I think we need to fix the payment model, but it’s going to be a longer process. There are developments happening right now and there will be new, or at least tests of new models emerging. But I think one of the most important steps that can and should happen is to pull artists and artist managements and labels into the discussion and educate them. The best way to do that is for the streaming services to provide them with information around what would happen to their revenue under a model like user-centric, artist-centric and so on. Otherwise there is almost no transparency around the actual impact for artists. And I feel like that’s something that needs to happen so that everyone can have an informed discussion. Then it will take some time to draw conclusions and reach the compromises that we always need in order to implement something on that scale. The complicated thing about changing the payment model is the decision making and getting everyone on board. I think conceptually, if you read the paper that we put out, it’s not that complex to understand that this is important and should be changed. And you could, if everybody agreed, implement a new model quite quickly.

We touched on Deezer earlier – what are your thoughts on the Universal x Deezer deal? Is it a step in the right direction?

RR: Yes and no. In my view, the artist-centric model is still essentially the pro rata model that we have today. It doesn’t look at the user level or factor in the user’s individual streaming behaviour; it still looks at stream count for streams above 30 seconds, so it’s still a pro rata model, just with some additional rules applied. 

(Update: After this interview, Deezer issued a press release stating that the “artist-centric” model does include a user-centric element, albeit in limited fashion – with pro-rata still being the dominant payment logic for artist-centric.)

I think the step in the right direction is where they distinguish between different forms of consumption. So in this case, they’ve called it a search boost or active boost. I think that’s a good step, to distinguish between something like active and passive consumption, or lean forward vs. lean back, and factoring in what they call a boost factor in terms of artist-centric. One of the models we looked at in the beginning, which we ultimately decided not to publish in our paper, was exactly that, looking at a pro rata model and active versus passive consumption. So that’s a good step, and from what I’ve seen in terms of responses, most people seem to agree.

So we’re clear, what are the distinctions between active, passive, lean forward, lean back?

RR: Well, firstly, I haven’t seen a definition for this particular artist-centric model of how they distinguish what they’re looking at. But generally, depending on the streaming platform, they measure consumption in different ways. Spotify, for example, has their categories by source of stream which could be a playlist, an album, radio or search. A very simplistic example would be: if you listen to playlists from other users or third parties, then that is passive. If you listen to your own playlist, that’s active. If you listen from your library, that’s active. If you search for an album, that’s active. And then maybe radio is passive.

But in general, the way streaming services are categorising or measuring the consumption is not designed to distinguish between active and passive. If we decide to do that in the future, to allocate revenue based on active versus passive, I think there’s an incentive for streaming services to measure things that help distinguish between the two forms of consumption. So that’s the active boost which I agree with others that we need to figure out how to best do that. But I think it’s a good step. 

What are some of those added rules within ‘artist-centric’ that are applied to the pro-rata model?  

RR: One is distinguishing between professional artists and non-professional artists by looking at two metrics: the amount of streams that an artist has and the amount of listeners. I think there are a number of problems with this, e.g. that these thresholds are absolute numbers and not relative ones. Absolute numbers are always a bit weird to handle because on Spotify, for example, there’s a lot more consumption happening there and more users than on Deezer. 

But what I find most concerning is the distinction between ‘professional’ and ‘non-professional’, because there’s not really a way to do this, in my opinion. It has not been clarified so far, whether these benchmarks – 1000 streams, 500 listeners – are global benchmarks. If you need to have 1000 streams globally and 500 listeners globally, these benchmarks are more easily reached by artists that are in major music markets. So a U.S. artist can reach them more easily than, let’s say, an artist in Costa Rica with a local fan base. It may be harder for those artists to reach 500 monthly listeners. 

And I think it’s a bit worrying to potentially tell certain artists that they’re not professional. It’s important to know there’s another boost for professional artists. So if you don’t meet the 500 listeners, your streams will be essentially cut in half. Or the other way around – if there’s a U.S. artist that has 2000 or 501 listeners, their streams count twice as much. I think there’s a problem with that in a conceptual sense. Who will decide what the benchmark is? Because you’re essentially creating a case where the most powerful music company and one of the DSPs are policing this. And once you give them that authority, they could raise that threshold in the future.

My assessment is that this is not really a great solution for the actual problem. And then in addition, you have opaque rules put in place by someone where you don’t know what they will do in the future and what the motivation is behind that, because they haven’t talked about their motivation really in specific terms. 

Just to have a complete picture, there’s also a third measure that they’re implementing with artist-centric, which is that they’re saying that there is a category called ‘non-artist noise content’, where they group things like white noise and nature sounds and other things. But they haven’t released a list of what they think falls into that category. So they will remove all of that content that they deem fits into that category from Deezer and will replace it with noise and “functional music” create with AI or through other-means by Deezer themselves. Again, they haven’t clarified the motivation or how they got there. I think it’s never a good approach to distinguish between good and bad in a black and white way, because you’re never going to get to a great place when you’re trying to police different forms of audio. Also, there is a need and a demand for that sort of content. And I believe the problem is not that it exists on the platform, the problem is how the consumption is valued and measured.

So how would this type of audio play out in a user-centric model? 

RR: Right now we’re just measuring streams, so there are rain soundtracks out there that have hundreds of millions of streams and that’s why they get a lot of revenue. Under user-centric, rain sound producers would lose more than 90%, in some cases 98, 99% of their revenue because these huge amounts of streams come from very low numbers of listeners. So you could, in theory, have the user-centric model essentially achieve what they’re trying to do there without someone deciding who is part of the music club and who isn’t. And generally there are going to be businesses and potentially Spotify or whoever, that create a subscription for rain sounds and white noise for €2 or $2 a month. You’d have to service that need and put a price on it. So the streaming services put that subscription for your noise audio and the music subscription and sell it as a package at a lower price. And then you’ll end up in the same place as before. You’ll have to decide how much – let’s say $10 for music and $4 for other stuff. So you have $14, but you’re going to sell it at $12. You need to decide how much of the $12 will go towards music content, and how much will the noise audio get. Are you going to divide that by streams or something else? So you’re going to end up essentially in the same place that you started with. 

So I think, again, to summarise, and I’m sorry for going into this detail, but the boost and distinguishing active and passive is a good thing to at least try out and see what the effects are. The other rules around this are essentially still a pro-rata model, which they have coined as artist-centric. I think they set a dangerous precedent and are solutions that aren’t really equipped to address the actual problem with allocation.

The concept of the winners and losers that was mentioned in your report. It feels like in every model there are winners and losers.

RR: Yes, but I think you can do it a lot better than we’re doing now. So maybe I should talk about what I think the problem is. We’re only looking at total streams right now. So if you have a rain soundtrack that has 200 million streams, that rain soundtrack is going to get a lot more revenue than a newcomer artist whose new single has 5 million streams. And that’s how we have decided to measure what it’s worth, what the consumption is worth, what the value is that has been created for the listeners. 

But I’ve worked in data analytics at labels for a period of multiple years, and there’s streams, there’s listening time, there’s amount of listeners. You can look at every listener individually, which is user-centric and there is tons of data on the consumption. That rain soundtrack is going to have a lot less listeners than that newcomer with 5 million streams. Those 5 million streams are going to come from a lot more listeners. 

If you’re working and you’re listening to a ‘listen while you’re working’ playlist, or if you’re listening to something in your sleep or it’s to soothe you or to relax – all of that content generates a lot of streams very easily because it’s happening in the background. And if you’re only looking at streams, total streams, and basing your payout, your revenue allocation on that, that’s the actual problem, because you’re not capturing the value that it’s creating very well. 

There’s a lot more indicators, for example, listening time, listeners and so on, active, passive consumption that you can look at and you can use that for payout. And if you do that, you can apply the same rules for everybody. You can apply user-centric to every artist and producer and songwriter, and that would mean that rain sound tracks would get a lot less revenue, which is potentially more aligned with the value they actually created. So I think there are models that you can apply transparently to all artists and producers that still are more favourable to everyone in the industry than it is today.

You mentioned sounds to sleep to. When we’re talking about rain sounds not having that many individual listeners, are we implying that it might be streaming fraud? Perhaps people globally are just struggling to sleep and might actually be intentionally playing it.

RR: While that type of content is potentially most easily recreated using AI, it doesn’t necessarily have to do anything with AI. And frankly, AI doesn’t provide a lot of additional value for those producers because it’s easy to create anyway. And I don’t think there’s necessarily a lot of fraud in that space either because there’s enough demand, enough streams generated anyway right now just by a low number of consumers. So it’s legitimate behaviour; it’s just that we’re over inflating the importance of this content by only looking at the one metric that it overperforms in, which is overall streams.

This is a slight segue into a different area, but user commitment is one of the factors that in the user centric model you said would determine the impact. There were lines in the report that talk about the superfan and allude to their value. At Song Sleuth, we obviously see a huge value in the fan and the UGC that they would create. And yet we’re noticing that there are relatively few dedicated campaigns or artist campaigns that really maximise UGC and this fan relationship. Why do you think this is? 

RR: There’s a lot to say to this question so I’ll try to keep it as short as possible. First of all, SoundCloud and MIDIA Research, who published a paper on user-centric commissioned by SoundCloud, say that superfans are the most important thing when you’re operating under user-centric. 

In my view, that’s a simplification – at least when you implement the basic version of user-centric, without any additional rules. And also, catering to superfans, trying to reach them and increasing their consumption is what the industry is already doing today. Everybody’s focusing on finding their superfans and getting them to stream and activate them. Under user-centric, the one thing that will happen is, say someone who pays $10 in a subscription fee streams only 20 tracks in a month. If you can get ten of those streams, you can get half of their subscription fee from those listeners. With ten streams, you’re going to get a lot less revenue from your Superfans because they have 1000 streams or 500 streams per month, and you’re only going to get let’s say 5% of their subscription revenue. So actually, under user-centric, it’s going to be just as lucrative to cater to casual listeners beyond your superfans.

One issue that arises with the user-centric model is that, in the simplest version of it, you could have one user who has one stream in a month. If you get that stream, you get all of their money. But that would be very extreme, and I think you would need a way to avoid that. 

I’m not going to go into detail, but I think with user-centric the focus would not be on superfans as much as it is today, rather it will be, to name two examples, more on older demographics and on music that caters to a diverse audience.The industry has, for decades, been focused on young demographics and teenagers and people in their twenties and thirties. With user-centric, there would be incentive to create and market music to older demographics, because they might only have 200 streams per month, so if you can get them to stream your music, you can get a decent share of their revenue.

This could be exciting for artists that are older right now, but also for artists that are young today who could keep earning significant revenue as their audiences grow with them in age. Right now, musicians are like athletes. Your whole career is very centred on the first one or two decades of your career.

Ending on a more positive note, what other developments, technologies or companies are you kind of excited about? 

RR: I am hopeful that we’re moving away from this age where we saw UGC as a problem, and we are entering a new paradigm where we pull in that creativity from creators that love music and love to spread the message and spread good music. Pull those creators into the value chain.

What we’ve been doing is trying to find UGC and then claiming all the money for the industry and the artists, which is great in that otherwise UGC would be a huge problem for artists, but there’s no incentive really for the creators. And there’s a lot of them who know a lot about music, who could create great things with music. But if we take all the money as an industry, then professional creators are not going to use the music because they want a share of that money for their service. You see a couple of business models and even YouTube wanting to do something like creator music, where the creator and the music rights holders, artists and songwriters, share in the revenue. I think that’s a very exciting thing, because pulling creators into the value chain is just going to create a whole new business area where individuals and companies can create something with music and earn their livelihoods with music and professionalise. That will make the whole music landscape a lot more vibrant and bring music to a lot more people in very exciting ways.

Anybody who provides platforms where you can either licence music or curate playlists or introduce music to a lot of new listeners in any other shape or form and then have those creators and companies share in the revenue, is going to, I think, solve a very big problem, which is how do I get heard and introduced to a lot of people? It will potentially create a whole new class of creatives and curators that right now don’t really have a business model that they can tap into. You can’t earn money with playlists directly. There are some grey areas where people are trying to do it somehow, but you can’t really make a playlist and then automatically earn money because people are listening to it. So there’s a lot of great things that are going to happen in that space, potentially.

To read Ryan’s paper, Payment Option Transparency, visit https://paymentoptiontransparency.com/. Ryan’s excellent Linked In archive of articles on payment models can be found here.