Playlist Malfeasance Will Create a Streaming Crisis

Streaming economics are facing a potential crisis. The problem does not lie in the market itself; after all, in Q1 2019 streaming revenue became more than half of the recorded music business and Spotify hit 100 million subscribers. Nor does it even lie in the perennial challenge of elusive operating margins. No, this particular looming crisis is both subtler and more insidious. Rather than being an inherent failing of the market, this crisis, if it transpires, will be the unintended consequence of short-sighted attempts to game the system. The root of it all is playlists.

Streaming makes casual listeners ‘more valuable’ than aficionados

Streaming took the most valuable music buyers and turned them into radio listeners. Now, as the market matures, it is taking more casual music consumers and also turning them into radio listeners. Although curated playlist penetration is still low (just 15% of streaming consumers listen regularly to curated playlists, fewer than listen to podcasts), the impact on listening over indexes.

While a lean-forward, engaged music listener may select an album or a handful of tracks to listen to and then move on, casual listeners might put on a 60-track peaceful piano playlist in the background while studying, doing housework etc. The paradox here is that casual fans have the potential to generate more streams than engaged listeners.

With casuals being the next wave of streaming adopters, their impact will increase. But despite being ‘more valuable’ they will also reduce royalties, because more streams per user means revenue gets shared between more tracks, which means lower per-stream rates. The music industry thus has an apparently oxymoronic challenge: it is not in its interest to significantly increase the amount of media consumption time it gets per user, but instead it will be better served by getting a larger number of people listening less! 

Current market trajectory points to more streams per user, which – for subscriptions, where royalties are paid as a share of revenue – means lower per-stream rates.

Playing the game

Against this growing background consumption trend, streaming services, labels, songwriters and artists are all making matters worse by gaming the system whether that be by structuring songs to work on streaming, creating Spotify friendly soundsor simply gaming playlists.

With playlists being so important for both marketing and revenue, it was inevitable that people would seek out ways to attain any possible advantage. Consequently, playlists are becoming gamed, whether that be major labels getting more than their fair share of access to the biggest playlistsor ‘fake artists’filling them out.Most recently, Humble Angel’s Kieron Donoghue identified a cynically constructed playlist called ‘Sleep & Mindfulness Thunderstorms’(all terms optimised for user searches) that contained 330 one-minute songs of “ambient noise of rain and a few thunder storms thrown in for good measure”. The one-minute track length ensures they are long enough to qualify for a royalty share, but short enough to ensure that a typical listening session will generate a vast quantity of streams, thus generating more royalties.

The twist to this story is that this playlist was created by Sony Music and the artist behind all these tracks appears to be a Sony Music artist. Crucially Sony isn’t the only one doing this, with UMG getting in on the actand Warner Music signing an algorithm.

Playlist deforestation

This sort of activity may make absolute commercial sense but is creatively bankrupt. It certainly makes record complaints about ‘fake artists’ ring less true. Just because you can do something does not mean that you should. This model works until it doesn’t. In fact, there are parallels with deforestation. A logger in the Amazon will likely not be thinking about the destructive impact on the environment he is directly contributing to. In similar manner, it is unlikely that the people creating these playlists realise that they are contributing to a market-level crisis. This is because, the more of these types of playlists that are created, the lower per-stream rates they will generate for everyone.

Well, not ‘everyone’. If overall streaming revenue rises but stream rates decline, then the companies with large catalogues of music, especially those that are also creating arsenals of playlist-filler ammunition, will still feel revenue growth. For individual artists and songwriters, however, royalty payments could actually fall.

Fixing the problem

The casual listening problem will not fix itself. In fact, despite labels worrying about declining ARPUthe only way they can keep ahead of declining streaming rates is by increasing their share of streams. That means more of this sort of playlist gaming activity, which further accentuates the problem.

There is however a simple solution: reduce per-stream rates for lean-back playlist plays.This would ensure the songs people actively seek out get better pay-outs. The demarcations between lean back and lean forward used to be elegantly simple (e.g. Pandora versus Spotify), but now curated playlists and other forms of streaming curation are supporting radio-like behaviour on the same platforms as on-demand. It is time for royalty models to catch up with this new reality.

Pandora’s Rate Ruling Reveals The Cracks In Streaming Economics

The much anticipated outcome of yesterday’s Copyright Tribunal decision was a 20% increase of Pandora’s ad supported stream rate from $0.0014 per non-interactive stream to $0.0017. The result was roughly equidistant between the two parties’ preferred rate: Pandora wanted $0.0011, SoundExchange (the body that collects the royalties on behalf of the labels) wanted $0.0025. As with any good compromise neither party will be truly happy, though on balance Pandora probably came out slightly better. Both the rate and the whole rate setting process shine a bright light on the economics of streaming, especially when contrasted against on-demand services.

Pandora’s semi-interactive radio service operates under statutory rates in the US that are set by the Copyright Royalty Board for a few years at a time, with inflation baked in. This means a continual rise in rates (see figure). It also gives Pandora a degree of certainty over its mid term future but prevents record labels from negotiating for better rates (publishers however are able to strike direct deals with Pandora). Spotify, and other on-demand streaming services, negotiates deals directly with multiple record labels, publishers and rights bodies. Deals typically come up for renewal every couple of years, involve large upfront payments and Minimum Revenue Guarantees (MRGs). They also run the risk of core product features being threatened in renegotiations – as we saw with the labels’ dalliance with killing off freemium this time last year.

The most significant difference between the models is how the per stream rate works. For on-demand services a royalty pot as a % of revenue is determined. This is then divided between rights holders based on plays in a given period and allocated on a per stream rate basis. Thus royalty payments remain a comparatively constant share of revenue, assuming of course that the service hits the MRG targets – if it doesn’t the share increases, often above 100% of revenue. This model also implies a clear ceiling to the potential profit an on-demand service can earn. By contrast Pandora pays out on a (largely) pure per stream basis. The direct consequence of this is that Pandora is able to increase it revenue per play faster than its rights cost per play which in turn creates the potential to grow margin (see next figure).

Between 2009 and 2014 Pandora’s content acquisition costs per listener hour increased by 27% from $17.52 to $22.29. This reflects both the CRB set rate as well as deals with rights bodies and publishers. But over the same period Pandora’s revenue per listener hour increased by 114% from $21.48 to $45.97. Now clearly, an increase in revenue per hour does not inherently mean increased profitability, or even profitability at all. Indeed, Pandora’s continued losses have been a perennial bugbear for investors. But Pandora has chosen to invest its increased revenue to grow its business, building out regional ad sales teams and making acquisitions such as Next Big Sound, Ticket Fly and Rdio. In short, Pandora could have been profitable for some time now if it had so chosen. Instead it is chasing a bigger prize, namely to become the single biggest revenue driver in US radio. To get big it needs to spend big.

Pandora’s Core Strength Is Being Able Increase Profitability Per User

The underlying principle is clear: while on-demand services have little meaningful way of increasing revenue per user with the current model, Pandora has more than doubled revenue per user in 6 years while rights costs have declined in relative terms. Content acquisition costs fell from a high of 82% of revenue in 2009 to 48% in 2014. That rate will increase in 2015 due to direct deals struck with publishers and the $90 million pay out for the pre-1972 works ruling. But it still remains well south of Spotify’s 70%+.

On Demand Services Have Similar Fixed Costs But Tighter Margins Because Of Royalties

While there is a clear case for semi-interactive radio rates being markedly lower than on-demand rates many of the fixed costs of both types of streaming business are the same.  Both have to commit similar amounts to product development and tech, bandwidth, data analysis, reporting marketing, customer care, management. This puts on-demand services at an operational disadvantage compared to webradio services.

If paid-for streaming services are going to become commercially sustainable there is going to need to be pricing and product innovation to both reach more mainstream users (cheaper tiers) and to drive more revenue from high value users (more expensive tiers and bolt ons). Right now there is relatively little commercial incentive for on-demand services to innovate upwards as profitability will remain largely the same. There is an opportunity for labels to offer Spotify and co a Pandora-style pure per-play license structure for all products launched above and beyond the standard 9.99 tier. This would give the services the ability to follow Pandora’s path of growing revenue per user faster than rights costs per user, thus improving commercial sustainability and allowing them to invest more in product innovation.

Rights Frameworks Need To Engender Commercial Sustainability

Pandora is one of the few stand out, independent success stories of the entire history of digital music. It has become one of the world’s biggest music services despite being largely constrained to the US, it has built a commercially viable model and it has delivered a big return for investors via its IPO. Only Last.FM, Beatport and Beats Music can genuinely lay claim to having delivered big returns for their investors. There are many mitigating factors, but the unique licensing structure Pandora operates under is the single most important one. Do songwriters and labels feel that they’re getting short changed? Absolutely. But it is in the interest of every music industry stakeholder that the economics of digital music are structured in a way that enables standalone companies like Pandora, Spotify and Deezer to thrive. Otherwise there can be no complaints when the only options left on the table are companies like Apple, Amazon, Facebook and Google whose interest in music all stems from trying to sell something else. That’s when artists and songwriters are really at risk.