On… Deezer nuzzling up to Universal’s ‘artist-centric’ model, a potential mass deletion, and a controversial tweak

MBW Reacts is a series of analytical commentaries from Music Business Worldwide written in response to major recent entertainment events or news stories. MBW Reacts is supported by JKBX, a technology platform that offers consumers access to music royalties as an asset class.

Yesterday evening in the music rights business, all anyone seemed to be talking about was Elliot Grainge and his new deal with Warner Music Group.

Just over 12 hours later, Grainge Jr. was swiftly nudged off the top spot of the music industry news agenda… by his father.

Sir Lucian Grainge was front and center earlier today (September 6) for the announcement that Deezer is fully adopting Universal Music Group‘s preferred ‘artist-centric’ royalty payout model.

This model will be rolled out on Deezer from next month in France; Deezer then plans to bring it to the rest of the world in 2024.

Today’s announcement was notable not only for Deezer officially canoodling with UMG’s preferred model, but because, for the first time – following months of rumors – we got a solid idea of what ‘user-centric’ actually is, and may become.

To summarize, as we stand, Deezer’s version of the ‘artist-centric’ royalty model involves three key elements:

1. Artists who attract over 1,000 listens a month (from over 500 unique listeners) on Deezer will be paid (or, at least, their label/distributor will be paid) double the royalty rate of artists who do not;
2. This royalty rate will double again if a play of said artist’s music has been actively searched for by listeners vs. being algorithmically served to them. (Upshot: Meeting the criteria of (1) and (2) here will see artists getting FOUR TIMES the money, per stream, as artists who meet neither.);
3. Deezer says it plans to “replace non-artist noise content” on its platform with its own Deezer-made “content in the functional music space”. Deezer then won’t pay itself any royalties when users play these tracks – bulking up the overall royalty pool for everyone else. 

In addition, though not strictly part of the royalty model changes, Deezer promised today to take a harsher line on streaming fraud, including the continuing refinement of a “stricter, proprietary fraud detection system”.

As ever with this kind of announcement, the bigger story isn’t what was confirmed today – it’s where it might lead in the months ahead.

That’s as true for Universal Music Group and its rivals (Sony Music Group, Warner Music Group et al), as it is Deezer and its rivals (Spotify, TikTok Music, YouTube Music et al).

Here are a few early thoughts on where this could all go – and some big questions it leaves hanging over the future direction of the music streaming business…

1) Deezer says it will ‘replace non-artist noise content’. Wait. Replace as in… DELETE millions of tracks?!

This has arguably not caused half as much buzz as it deserves: We may be about to see a deliberate mass deletion of tracks from Deezer’s catalog.

At least, that was my read of things in Universal x Deezer’s announcement today.

Before I get ahead of myself on this, let’s walk through exactly what Deezer and UMG announced today in terms of “non-artist noise” – and exactly how they announced it.

Verbatim, from the official joint press release: “Deezer’s intention is to replace non-artist noise content on the platform with its own content in the functional music space, which will not be accounted for in the royalty pool.”

Semantically speaking, you can’t really ‘replace’ something unless, in the process, that thing ceases to be.

Ergo: there seems to be a fair chance that what Deezer deems (and, indeed, what Universal deems) to be “non-artist noise content” will, at some stage soon, be completely eradicated from Deezer.

This would represent a stinging fusillade in the major music companies’ streaming-era battle against the market-share-eating popularity of ‘fake artists’, ‘functional content’, and, yes, just plain ‘noise’.

That would be true even if “non-artist noise content”, in Deezer’s words, referred simply to non-musical content, i.e. white noise, gamma waves for study etc.

Yet I think it may go even further than that.

It also seems possible – though Deezer hasn’t overtly confirmed it – that “non-artist noise” here actually also refers to at least some musical content.

I’m talking “functional music” content; often-cheaply-made production tracks with a specific application in mind (i.e. ‘music for concentration’, ‘music for sleep’ etc.).

The stuff that Spotify has pushed users towards. The stuff that often carries a ‘fake artist’ name and which – coincidentally! – tends to cost streaming services a significant amount less in royalties than major record company content.

If I’m correct, following its mass deletion, Deezer intends to “replace” this gigantic ocean of “noise” and “functional music” with a much slimmer library of Deezer-owned content, almost certainly developed using AI tools.

You’ll still be able to search for “white noise”, “music for concentration” etc. on Deezer. But you’ll be served, exclusively, the platform’s own-brand audio.

Importantly, that own-brand audio won’t be monetized by Deezer in any way.

How many tracks are we talking about potentially being deleted here?

Millions… if not tens of millions. Deezer said today, with more than a little grumpiness, that its catalog had “exploded” in the past two years, from 90 million tracks to over 200 million today.

Only Deezer knows precisely how much of that content is defined, internally, as “noise” – and how much of it will still be available on its platform in a couple of months.

“There’s a lot of content getting uploaded to our servers every week… that we have to pay for; there is a cost to having a never-ending growing catalog.”

Jeronimo Folgueira, speaking in March

Deezer also confirmed today that it will be actively stifling the growth of its hosted catalog in future.

“As part of the artist-centric model, Deezer intends to apply a stricter provider policy to ensure quality and a better user experience,” the company said. “This includes steps to limit non-artist noise content.”

Will it also involve a widespread cull of existing “noise” and “functional music” that’s previously been uploaded to Deezer?

We shall see.

We already know Deezer has one direct economic incentive for slashing its catalog size: the associated digital hosting costs.

In March, Deezer CEO, Jeronimo Folgueira, outlined this to analysts on an earnings call, stating: “There’s a lot of content getting uploaded to our platform every week… it puts a lot of content in our servers that we have to pay for. There is a cost to having a never-ending growing catalog.”

2) If Deezer’s ‘artist-centric’ math adds up to more than 100%, we may see tweaks – and potential tensions

You’d be hard-pressed to find a music rightsholder in the world who didn’t applaud one specific element of Universal x Deezer’s announcement today: A royalty boost for artists whose tracks are actively searched for on streaming service, as opposed to being served up by an algorithm.

This principle, the idea of ‘lean back’ listening being less valuable – some may say more ‘radio-like’ – than users pressing play with their human digits, is something I’ve heard supported by pretty much every corner of the music rights business. (Even by those who usually feud!)

But the other standout element of UMG x Deezer’s new ‘royalty boost’ formula – the one that doubles the per-play royalty paid out to artists with over 1,000 streams and over 500 monthly listeners – may, eventually, prove more controversial.

Not because it’s an unpopular idea amongst the blockbuster business: In March, Warner Music Group’s Robert Kyncl espoused the virtues of a similar “multiplier” calculation model for superstar acts.

However, call it a gut feeling, but I suspect the threshold for artists missing out on a royalty ‘multiplier’ may not remain stuck at the 1,000-stream/500 listeners count.

I suspect, as time goes on, those numbers may climb higher – as will the accompanying industry tension.

Here’s why.

It’s very likely that the majority of streams on Deezer today are of music from artists who have more than 1,000 monthly plays, from more than 500 unique monthly listeners.

I can say this with some confidence: Deezer itself confirmed today that just 3% of ‘uploaders’ (aka artists) on its platform generate 98% of its total streams.

(Technically, Deezer said that 97% of ‘uploaders’ on its platform only shared 2% of its total streams… but the reverse therefore must also be true.)

We also know that the three major record companies enjoy a cumulative market share (by revenue) of the digital recorded music market of over 70% today… and that these are companies that don’t typically sign artists with less than 1,000 monthly streams to their name. (Even on Deezer.)

We also know that, according to a Competitions and Markets Authority study in the UK – citing data provided by streaming platforms themselves – some 60-80% of plays on Spotify in the UK in 2021 took place outside of the playlist environment entirely.

(That’s ‘user-curated’ plus ‘non-playlist’ in the table below – i.e. Users finding tracks themselves, and playing them, themselves.)

With this in mind, it’s possible that the majority of stream royalty payouts, under the new ‘artist-centric’ formula, would be ‘doubled’ on Deezer as a result of its new model change.

It’s also likely that at least a significant minority of these payouts would additionally receive a further ‘doubling’ multiplier (i.e. X4 overall) due to them being ‘actively’ played, rather than ‘passively’ served by the algorithm.

Whichever way you cut it, such an extensive volume of royalty multiplication may place strain on the limits of the streaming royalty ‘pie’ at Deezer.

To put it more simply: If the majority of stream payouts are being multiplied on Deezer, doesn’t it then follow that the total payout in a given month could add up to more than 100% of the royalties actually generated on the platform?

In order to offset this more-than-100% possibility (i.e. to find the additional royalty capital it needs to satisfy all rightsholders’ bills), Deezer plans to boost the size of its available royalty pot via two methods, as mentioned earlier:

• (i) Doing a better job of stamping out streaming fraud; and
• (ii) Stopping the flow of royalty money to “noise content”. (Which may or may not include human artists posing as fake artists on “functional music” playlists.)

But will these two moves really be enough to create the additional royalty capital required to offset the ‘artist-centric’ multipliers?

Deezer today said today that tracks that it considers “noise” currently represent only 2% of total plays on its platform.

Across the entire music streaming world, eradicating these “noise” tracks from royalty payments would be a nice bit of business for the major record companies: 2% of global streaming revenues last year, according to IFPI data, would amount to over USD $250 million annually.

But would this eradication of “noise” music – with 2% of total plays today from a given service’s royalty pool – make up the additional royalty money needed to satisfy a 2X and/or 4X royalty payout for a substantial amount of plays on its platform?

Would it do so even when combined with a harsher crackdown on music streaming fraud?

If not, how will Deezer make up the monetary surplus it needs under ‘artist-centric’ to cover to pay all of their rightsholders what they’re due?

One obvious option: Tweaking the ‘artist-centric’ formula as time goes on to reduce its expenditure.

Aka: increasing the minimum listener/stream count threshold artists will be required to surpass in order to achieve a royalty multiplier.

If said tweaking takes place, perhaps artists will end up requiring a minimum of 3,000 monthly listeners to earn a double-royalty rate. Perhaps they’ll require 10,000.

Smartly, Deezer and UMG have already given themselves the bandwidth to carry out such tweaking if required.

In their announcement today, the two companies clarified that they may make “data-based adjustments to optimize [the artist-centric model’s] performance” in future.

If this ‘tweaking’ scenario does play out, raising that minimum plays/listeners threshold, expect certain industry tensions to rise – and not just amongst ‘independent distribution/services’ companies like Believe/TuneCore.

Sony Music is an interesting case, for one, because it spent $430 million buying independent artist services platform AWAL (plus Kobalt Neighboring Rights) in 2021.

What happens if some of AWAL‘s developing artists fall beneath the royalty-multiplier minimum threshold?

Bottom line: Michael Nash, Chief Digital Officer and EVP of Universal Music Group, told the Financial Times today that Deezer’s switch to the new ‘artist-centric’ model would be “revenue positive” for UMG.

If, as you’d expect, the same outcome also applies to the other two major music companies (should they license the model), Deezer will clearly have to balance its finite monthly royalty pool – somewhere, somehow – with an equal “revenue-negative” impact.

3) Will the other majors sign up? (And Wherefore art, user-centric?)

Neither Sony Music nor Warner Music has so far – publicly, anyway – agreed to license Deezer’s ‘artist-centric’ model (as devised with UMG).

As things stand, the model has only officially been licensed by UMG, and will clearly act as a valuable testing ground for Sir Lucian Grainge and co. (Especially if Universal is looking to perfect the model in the ‘real world’, before pushing it towards Spotify etc. in future licensing deal negotiations.)

The jury’s out on whether Sony Music will embrace the model on Deezer (see AWAL point, for one thing, above).

Yet with Warner Music Group, Deezer may find a more sympathetic ear – and a more willing licensing partner.

According to Cofsem data published by Euronext, under Deezer’s current ownership structure – despite the streamer’s publicly-listed status – Access Industries continues to be a significant shareholder in the platform.

Access Industries, of course, is itself the majority owner of Warner Music Group.

Not just that, but both Universal Music Group and Warner Music Group (as well as Access) helped to finance a EUR €135 million ‘PIPE’ (private investment into public equity) fund for Deezer shortly before it went public last year.

Consequently, it appears likely that both Warner Music Group and Universal Music Group remain minority shareholders in Deezer. (Sony Music did not participate in the ‘PIPE’ fund.)

Speaking of Warner Music Group… you may remember that last year WMG announced its own experimental royalty-model partnership, with SoundCloud.

That licensing agreement centered on the ‘user-centric’ royalty model, which sees the royalty-bearing proportion of each individual subscriber’s monthly payment apportioned only to the artists/tracks that specific individual has listened to.

Of the plus-points put forward by supporters of ‘user-centric’ royalties (or in SoundCloud parlance, ‘fan-powered royalties’), one is particularly compelling: User-centric naturally neuters many of the commercial incentives of streaming fraudsters that use premium accounts to siphon money out of the ‘pro-rata’ system.

This is a very hot subject right now – following news yesterday (September 5) that multiple gang members (and/or ex-gang members) in Sweden have confirmed they’ve used such a technique to launder money on Spotify in the past.

Spotify protests that it clamps down hard on ‘premium’ streaming fraud. Naysayers suggest Spotify (and co) will never be as incentivized to do so as music rightsholders might hope (partly because SPOT catching and banning premium streaming fraudsters… also means SPOT reducing its own subscriber numbers – one of the company’s most important investor KPIs).

Regardless, today’s announcement from UMG and Deezer contained no nod to ‘user-centric’ (UC) economics at all.

On the matter of fraud, UMG clearly prefers to trust in Deezer’s anti-fraud technology, rather than a sweeping UC solution.

Today’s announcement, therefore, didn’t just clarify what Universal’s ‘artist-centric’ royalty model is, and what it may become.

It was also as clear an indication as we’ve seen to date that Universal Music Group is no fan of ‘user-centric’ – and is determined to move forward without it.

JKBX (pronounced “Jukebox”) unlocks shared value from things people love by offering consumers access to music as an asset class — it calls them Royalty Shares. In short: JKBX makes it possible for you to invest in music the same way you invest in stocks and other securities.Music Business Worldwide