Music and Streaming, Part I: Strategy
The business fundamentals behind DSP's and what this means for Music.
It’s great to be back! There’s a special promo this May, a year’s subscription to The Radio Edit for £30. Sign up while you can - further articles in this series about music streaming will be for paid subscribers only. Get the promo here.
At the end of the last decade, digital streaming appeared to be music’s salvation. Today’s view of streaming goes from ambivalent to antagonistic. From streaming’s prioritisation of old music at the expense of new, to the economics of royalty payments, so-called “digital service providers” - DSP’s - are referenced as enemies as often as they are as friends.
To address this ambivalence, in this article and those that follow, I examine the fundamentals of DSP’s – both in their business strategy and in how they attract and retain users. Part explainer and part analysis, I aim to answer a question posed some time ago in the London Review of Books article, What Does Spotify Want? - and to give food for thought for those of us who care about the medium and artform of music about streaming’s place in our culture.
What is a DSP?
The DSP’s we use today are internet technology platforms that serve their customers a conglomeration of different media. In these articles we’re mostly talking about subscription services that have a basis in music (but offer other audio and video content). While iTunes or other download stores are, strictly speaking, DSP’s, they’re not covered in this article.
Product
In Internet and app-based businesses, product refers to the consumer technology offering. It’s the basic set of features and functionalities on offer. For DSP’s, this tends to be split by customer type. Those on a trial get a limited set of features; often a shuffle-based experience with reduced search functionality and no ability to download content for offline playback. This is designed to differentiate the premium experience for subscribers: search, skips, ability to build libraries and create sharable playlists, offline playback etc. The idea is that the friction in the free experience encourages conversion to premium.
DSP product is largely undifferentiated; whatever app you choose, the experience is more or less the same.
Business Strategy
Streaming services exist to deliver different media to end users, monetising the delivery through subscriptions or advertising.
Some services are offered by larger corporations and may have a dual function – not only for the business unit producing them to be commercially successful (as defined by the parent company) but also for the DSP to support another part of the business strategy (e.g. proprietary devices, Prime, distributing owned content, etc.)
Other DSP’s seek profitability and free cash flow from a standalone product– and have business strategies designed to achieve these goals. Spotify is the most successful and best known incumbent that isn't part of a bigger corporation.
Many surviving smaller DSP’s burn investor cash as they attempt to grow at speed so that they can generate profit in the future. They’re in survival mode, trying to balance investments with what may be structural unprofitability and an insecure cashflow trajectory. You’ll often see these services “pivot” their strategy as new investors come in, put new teams in place and look for different ways to differentiate and achieve scale.
To grow, subscription services need to acquire new customers, through trial, repeat engagement and conversion; and to retain existing customers by driving repeat visit and providing a good customer experience. The more often a customer uses the service, the more likely they’ll convert from a free trial or the more they’ll feel like they get value for money and carry on with the subscription. The commercial strategy determines the marketing strategy, which is all about generating these two activities, which support the obsession of consumer tech firms: scale.
Cost Structure
Streaming apps have high fixed costs: developers, server space, other staff, etc. They may also have some fixed content costs, in the form of advances or of minimum revenue guarantees. It’s expensive to launch a streaming service, mostly because the music licensing deals are costly; barriers to entry are high. The cost they’re liable for, of the music they deliver to customers, is dealt with via a revenue share model. Under this model, subscription revenues are pooled and shared between record labels, the streaming services, and music publishers in that order of magnitude. Labels and publishers are then compensated on a pro-rata basis. If Ed Sheeran’s music drives 5% of streams, Ed Sheeran’s record label gets 5% of the share due to record labels.
Other content formats, like podcasts, are compensated differently – some via their own advertising deals, some via advertising sold by the DSP against their content. Audiobooks are paid out of incremental revenues subscribers pay for their content, or out of bundled subscription revenues – which can mean less money for record labels.
Streaming services don’t compensate artists directly. They pay record labels, who compensate artists based on the royalty share agreed in the contracts between artist and label.
There’s debate about the extent to which artists have benefited enough from the rapid growth of music streaming; the extent to which labels get a reasonable share of subscription revenues, or whether the framework treats all artists equally are also moot points. Spotify is under scrutiny for both issues. The share of revenues they pay out to music rights holders has declined significantly in the last 10 years (Spotify paid out 40% less per subscriber in 2024 than it did in 2014). Spotify has negotiated better deals due to its scale, but it has also reduced content costs by offering different audio formats as well as music that falls outside of standard licensing deals (as described in Liz Pelly’s book ). A Spotify user personally, my recent interactions with the app do make me wonder if the customer experience is optimised to serve me content that’s cheaper.
Spotify and other DSP’s have reportedly signed up to “Streaming 2.0”, an evolution of the revenue share model. Under this model, artists who generate less than 1000 streams a month do not generate any income for their labels from the royalty pool. It’s ostensibly a measure to limit the problem of an ever-filling bucket of music. The following article on BIG suggests it has the effect of creating a barrier for entry to new artists, but really it’s a barrier to label investment in those artists. The new model doesn’t stop anyone uploading content, but it disincentivises labels from taking risks – if they sign developing artists with no streaming history, then it becomes even harder to know when thresholds might be met and the investment might start to pay back - much better to wait and see until the thresholds have been crossed. As demonstrated by analysts including Hunter Giles and David Price, the new model has increased the share of the royalty pool paid to the most-streamed artists at the expense of those who stream the least. If this was down to some external demand factors, we might say fair enough, but DSP’s are not simple demand-fulfilment environments, they promote music as well as allow it to be played. The challenge is, as the BIG article notes, not only has major label leverage driven the change, but it can also mean better placement, exposure and prioritisation for their artists.
Content: most popular music streaming services in UK, US and Europe offer full catalogues of repertoire and like the product, are largely undifferentiated when it comes to music selection. This contrasts with video streaming services, where the model is built around exclusivity. TV series are episodic, with repeat visits to get to the end of the story; movies are windowed, where streaming selection is a punctuation when they launch on these apps. This creates a business case for exclusivity as it attracts, or helps retain, subscribers. Music consumption isn’t the same – as an auditory medium it can sit in the background as well as the foreground of our attention. We listen in bursts then move on to the next record – and with so much music available, we’d be likely to listen to an old album of a band we love if we couldn't hear new music if it was teased in episodes. The acquisition and retention effects of exclusive music are not as potent.
In this decade we’ve entered an era of streaming beyond music – podcasts, video, livestreams, educational content, etc. The different nature of these media allows services to experiment with exclusivity models (podcasts can be episodic and/or drive repeat visits). Spotify tried to build a suite of exclusives to help growth but eventually shuttered this experiment and Amazon’s investment into Wondery, a podcast production house, has recently undergone a correction. The episodic nature of narrative podcast series, or the “appointment to listen” of your favourite commentator in politics/football/the history of queer spaces or whatever, drives the repeat engagement the app is looking for; there’s no real need to do this exclusively now streaming growth has matured to the point that the best thing to do is hang on to the listeners you have (by offering everything) rather than go and chase new ones (via exclusives). These additional formats can also be used to differentiate the premium tier from the free one.
This move into new content formats was a consequence of the growth of the so-called “attention economy” – the competition for consumer mindshare linked to mobile phone consumption, from podcasts to audiobooks, livestreams to short form video. Given the necessity to keep a DSP listener on the app, to ensure successful conversion and retention, these services needed to offer incremental media so that users didn’t migrate elsewhere. Podcasts are also a cheap content format – with revenue driven by advertising and a vast amount of potential listening hours. If listening to media other than music could be encouraged, then not only could listening be diverted to lower-cost formats than music, the leverage of the music repertoire owners could be decreased and better deals could be driven by the DSP’s.
Distribution: content tech needs distribution – the apps need to be easily accessible for users to try. DSP’s that are part of a larger corporation tend to have distribution baked in – for example, Apple Music is immediately accessible to anyone buying an IPhone. When Amazon Music was growing rapidly at the end of the last decade, Jeff Bezos said in an interview with Billboard that they benefited from the “twin freight trains of Prime and Alexa”. Deezer’s early growth was powered by a bundle offer with Orange as they both shared investment from Vivendi. Distribution is the other leg of DSP’s marketing tripod, alongside acquisition and retention.
Spotify’s market leadership in developed streaming economies is despite their lack of an owned of a distribution channel, although Spotify’s first-mover advantage and the development of brand equity has made it a more desirable product than that of its competitors (which made it attractive to phone companies looking to offer a bundled streaming service). Arguably, for Spotify, brand trumps distribution.
What kind of company is a DSP? Spotify evolved out of a file-sharing start up. Apple Music was birthed from their purchase of a start-up called Mog. Amazon Music was built on the infrastructure of their MP3 business. At the start, they were tech companies. In the high-growth period towards the end of the last decade, these services had staff and leaders who’d come from parts of the music industry, driving a degree of culture shift. It felt like these firms were becoming less like tech firms and more like music companies.
Since the end of Covid, the prioritisation of additional media like audiobooks, podcasts and video, as well as of automation over curation, has seen the cultural pendulum swing back. DSP’s are unashamedly what they’ve always been: tech firms. They seek scale at pace, customer lifetime value and free cashflow to re-invest or return to investors.
They’re not in it to profile great music or to advance the artform. That doesn’t mean those things aren’t of interest to them, or that they won’t develop ways to reach those outcomes, but it is not the primary reason for the existence of these businesses. Rather than hope they’ll help us reach these outcomes for music culture, we need to encourage that outcome - which will come if we can help generate alternatives that not only offer different choices for music consumption, but drive the right competitive pressures to ensure music’s role is properly considered.
We’ll consider that further in the articles that follow, which will be themed around how music is presented, and how music is marketed, on DSP’s.