Music and Streaming, Part II: Marketing
In the post-music streaming era, do artists and labels give too much away?
It’s taken a while to get this one out, so I’ve extended the promotion for new subscribers: get 40% off a year’s subscription. Redeem by August 30th. You can also claim an extended 30-day free trial in the same time period to get behind the paywall on this article.
This is part II of a series about music streaming. Part I, about DSP strategy, can be read here.
The Funnel is the term that streaming apps use to describe how they acquire new users. It’s a term inherited from subscription businesses that rely on a free trial model to grow subscriptions, be they gym memberships or wine clubs. In the funnel model of user growth, once aware of the service (a.k.a consideration), listeners are thrown in at the top to try it out (trial ); a smaller number try it again (repeat); and a smaller number still convert to a paid subscription (conversion).
Getting the maximum amount of users in at the top improves the overall number at the bottom. DSP’s maximise the width of the funnel by offering free trials of varying lengths and by offering free tiers. Often monetised through advertising, free tiers offer complete music catalogues but put friction in the process of discovery and playback (e.g. limiting the listener’s ability to choose tracks) or limitations in the tech experience. Free tiers can power acquisition at scale. The problem with free tiers is twofold: first, the value of the music that’s offered is largely offset for the chance of monetisation down the line; second, by offering music for free, they can perpetuate a perception among consumers that music has little value.
The arguments for free tiers won out in the aftershock of internet piracy in the 2000’s. Users might be getting music for free, but there was a path to remuneration and wasn’t that better than the alternative? Spotify still makes this argument, updated so that the free tier helps recorded music to monetise in parts of the world where streaming is relatively new. In a news release put out to accompany stats around their contribution to the music business (see this Decode article from January ‘25) Spotify exec David Kaefer explained that “We offer an ad-supported free tier, while some services don’t…more than 60% of Premium subscribers were once free tier users. Bringing in users who don’t expect to pay for music, and deepening their engagement, means they’re more inclined to become subscribers in the future…”
Yet record labels now apply increased pressure to change the dynamics of free tiers – for example, Sony chair Rob Stringer told investors that “ “Music should not be ‘free’ or a ‘cheap bargain’ still after a decade of such positive value for…proof of concept in mature territories.” You can understand Stringer’s argument: Sony continues to acquire and exploit the music of artists that are household brands, from Pink Floyd to Queen. They’ve paid big money to acquire these rights: what does it do for the artist’s brand if people can consume it for free?
In other industry segments, where businesses compete for potential customers at scale, developing, marketing and advertising the business’ brand has a positive effect on consumer growth. This work helps to define the product or business’ unique values and differentiate them from competitors. But all the DSP’s have technology and features that are broadly analogous, and the catalogue is more or less identical. In this context, arguments about the value of marketing the brand are oppositional: on the one hand a complete waste of time, since the product is the same; on the other hand, the only lever available to create points of difference in the minds of potential users.
As the product landscape has evolved, subtle differences between the different players have emerged. Spotify accentuates innovation in technology, especially algorithmic playback and AI, and the multiple media it makes available to users. Apple is a pure-play music service that espouses the old-fashioned tastemaker model espoused by Zane Low, a DJ who rose to fame in the MTV and Radio era. Amazon’s music offering focuses on voice activation and plays on the parent brand’s value and convenience attributes. Yet, after some forays into these investments in the last decade, mass advertising campaigns have not been pursued by DSP’s in mature markets, most probably as the cost/benefit doesn’t stack up where streaming penetration is close to maxing out.
In this landscape where brand attributes are undifferentiated through a lack of advertising, differences in product matter – which makes free tiers all the more compelling as a marketing lever. The downside? Giving stuff away for nothing trains consumers to believe it should be free on the one hand while valuing the channel that gives it to them this way on the other.
Free songs may be a valid exchange for attracting users who will subsequently subscribe, which will remunerate the artists and labels who made the songs in the first place. But artists also implicitly endorse DSP’s through the sharing of social posts that feature DSP branding. In the absence of brand advertising, having a logo in an artist’s social feeds has premium value, not least because it is a highly credible type of social influence. It’s for this reason that, despite the de-prioritisation of brand advertising, DSP’s book premium billboard spots, like NYC’s Time Square. The “big look” this presents enables elite artists to share photos of the billboards on their feeds, capturing bragging rights. Distributed on social platforms to that artist’s millions of followers, this amplifies the DSP’s brand for the cost of the outdoor media only, saving them significant sums of money in social advertising.
This activity skirts the edge of endorsement, a specific form of IP license where the DSP benefits from the brand value inherent in the artist’s name and image. Endorsement normally requires bespoke agreements, but since these types of social posts are about the artist’s music most of the time, they sit outside of this framework (it’s a paradox of streaming that digital ads featuring artists actually don’t perform particularly well in driving acquisition, since, as noted in the last article, fandom is highly partisan). The platforms offer seemingly reciprocal activity in exchange both inside and outside of their app: placements in the app browse environment, merchandising in “carousels” (the banner placements at the top of the app screen that rotates), paid marketing to amplify some of those artist social posts mentioned above, etc.
This focus on cosmetic placements is a hang-up from the transactional era that I described in The Record Label Crisis – a harking back to physical stores, with window displays, front-of-store racking, and other types of in-store merchandising. It’s part of record label’s ingrained culture to believe that visibility correlates to discoverability.
Check the confusing array of visual placements on any DSP’s app – the matrix of different visual displays, the confusing patchwork of podcasts, music and video – does any of it compel you to start your listening journey there? Consider also what’s known as the visual inventory “below-the-fold” – the bit you have scroll down to get to – when do you ever get that far? If you could see it, the data would show you that the probability of a click decreases exponentially the further down you scroll. Consider also that voice-activated speakers are commonplace in many homes in mature streaming markets and speculate as to the share of total streams from app as opposed to voice commands that follows. One could estimate then that this visual merchandising has a low probability of success as a trigger for playback and low inherent value.
The real secret sauce for engendering plays is the listening environments designed to drive long playback (session length is a leading indicator of user conversion and retention). At one point, playlists were the main vehicles for listening when the user didn’t have a specific piece of music in mind. Curated by musical experts, often former radio programmers, this was essentially the radio model on steroids. Record labels could influence and apply leverage, and as some of these playlists had recognisable brands and large, stable audiences, they were effective at presenting new music.
The development of social media advertising formats created a brief window for record labels to maximise this exposure as mainstream media still had some potency pre-Covid - placement on a big radio station or TV performances would generate demand which that advertising could then fulfil by driving traffic to streaming services for playback (again, an updated version of the old radio-to-retail demand cycle labels could get their heads around). Yet in 2025, demand generation and fulfilment no longer happen in different places, so spending money on social ads has little impact, especially as songs are surfaced via algorithms optimised for engagement.
This automated playback not only allows for listening designed to drive user retention but also removes human curators and eliminates their role as intermediaries. Asked about the wisdom of playback decisions, DSP’s can point to the omniscient logic of the machine-learning algorithm. Rather than try influence curators, record labels can influence playback either through their deals, if they wield enough bargaining power, insisting on prioritisation in automated playback; or without it, by paying for prioritisation through mechanisms such as Discovery Mode.
Some of this might be thrown into the pot in exchange for social support. It’s more likely that a DSP would promise, say, a top 5 spot in the tracklist of a playlist than placement in something automated, as the latter might break the playback metrics the technologists managing the algorithm are judged by. The playlist would please the label due to the “optics” – another visible thing they can show artist management to prove they’re doing their job.
The implicit partnership between DSP, musician and label made sense when they were collaborating on building a new paradigm for music listening – one that powered rapid growth and took the value of recorded music almost back to where it had been at its peak in the late nineties. Streaming is not the same now as it was ten years ago – it’s a multi-media technology, one where it’s clear from their strategic moves that DSP’s commitment to music as an art form is fluid in pursuit of their long-term fiduciary objectives. They’re no longer in it for the music, basically.
Should labels, then, deliver value through free content and the co-branding of their artist images, in the same way they did years ago?