The Spotify Disruption
Spotify is in a tough business.
Unlike most software companies, its core product, music, has marginal cost. While a stream is free from technical reproduction and distribution costs, the company sends about 70 cents of each dollar it makes to the music labels, publishers, and performance rights groups.
The remaining 30 cents are left to fund Spotify’s operations, and it spends all of that (and more), leaving it consistently unprofitable.
But despite these inherent constraints, the company has prospered and grown its audience—streaming all of the world’s music for “free” (ad-supported) or for a low monthly fee is an incredibly compelling product offering.
And, to be fair, the company is aware that the music business is inherently low-profit. Its intense diversification into podcasts betrays an urgent need to lower the costs of content and attain leverage over the music companies.
Yet, despite these efforts, I’d argue that Spotify’s prospects are in more danger than even they realize. Because, putting aside the inherent gross margin issues of music, the company faces two existential risks. The first centers on distribution—how customers find and revisit music they love. The second centers on the durability of their freemium business model.
Problem 1: How People Find and Rediscover Music
Today, most non-radio music listening takes place in apps—mostly on mobile devices, but to some degree on smart speakers and personal computers. I’d argue the emotional decision process for what kind of music to listen to breaks down into four basic desire paths:
I’d like to listen to exactly this song—“Blank Space” by Taylor Swift
I’d like to listen to an artist I like—Taylor Swift
I’d like to listen to a genre or vibe of music that fits my mood right now—holiday music, classic rock, 2000s hip-hop
I’d like to listen to music I will like, genre-agnostic—a playlist I made of my favorites, or Spotify Discover weekly
On a mobile device today, the process for acting on desires 1-4 is fairly straightforward.
Open music app of your choice -> Search for title/artist, or browse to appropriate playlist -> Press play
For desires 1-2, going directly to a song or artist, there’s virtually no differentiation between Spotify, Apple Music, Google/YouTube Music, Amazon Music, et al. They all have all the tracks, and roughly equivalent discovery interfaces.
For desire 3, listening to a genre-driven playlist, the services are fairly close—some might prefer Apple’s curated playlists, or Spotify’s more algorithmic offerings, but the product is still basically commoditized.
Desire 4, “play me some music I will like, genre-agnostic”, can be satisfied two ways. The first is via self-curated playlists, in which case the service in which a user has set up these playlists has some “lock-in”, as there’s friction in leaving those playlists behind. The second way to satisfy this need is via algorithmic discovery of music similar to my tastes. Spotify was truly a pioneer here with its algorithmic “Discover Weekly” playlist, which is still the best similarity-driven new music discovery system in the market.
But the fact that other services are as good or better than Spotify on 3.5 of 4 key user needs illustrates why Apple and YouTube are seen as “good enough” by huge swaths of customers who never fell into Spotify.
So, Spotify’s core product is increasingly commoditized. That’s not a huge issue as long as they continue to outperform their competitors on new user acquisition, which seems likely as long as Apple, Google, and Amazon continue to treat streaming music as a secondary, tertiary, or quaternary (?) priority. The motivated newcomer usually beats the C-team from the big, slow incumbent, particularly if the incumbents have few marketing ideas besides “include this service as a throw-in/default with other services”.
But I submit that Spotify has a much bigger problem than it might initially seem, because the interface for finding and rediscovering music is changing.
Today, most users open an app on a smartphone, tablet, or computer to listen to music. Tomorrow, music will be largely summoned to life via voice control.
Where do people listen to music? For the most part, at home, in their cars, and on headphones. In all three of these locations, it’s hard not to see touch- and click-driven interfaces being surpassed by voice-controlled ones.
At home, it’s more convenient to ask your speaker for music than to dig your phone out of the couch cushions. In a car, it’s more convenient (and safe) to ask a smart assistant to play music than to poke at your smartphone. And, on headphones, it’s more convenient to ask Siri or Google for music than to pull your phone out of your pocket.
Voice technology today doesn’t work perfectly in all of these contexts, but it will soon enough. This presents a huge issue for Spotify. In a world where you speak your musical desires and a smart assistant satisfies them, the company that controls the assistant controls demand for music. Siri, Google, and Alexa become the routers for an overwhelming percentage of music requests. (This is already happening on Alexa devices, where Amazon Music "eats" a huge percentage of music requests.)
“But wait!”, you say. “Users will still ask for Spotify, because it has their favorite playlists.” I submit that this does not matter for three of the four key music desires above: a specific song, artist, or genre playlist. These desires are satisfied by all the platforms, and the convenience and speed of voice control will make it annoying to specifically summon Spotify in order to play a specific song. Why say “Alexa, open Spotify”, followed by “Play Blank Space”, when you can just say “Alexa, play Blank Space?” Spotify will still have its superior algorithmic playlists, but that advantage will erode over time as other services learn more about their customers’ listening habits.
So this is the first existential risk to Spotify—that the convenience of summoning music via voice will hand the power to the companies that control the voice assistants. Inconveniently, the companies with the leading voice assistants also have their own music services. Spotify’s lead in distribution and penetration on mobile will be gradually abstracted away by the simplicity of speaking musical desires out loud, and having them satisfied.
Music will become a market where, for the most part, users ask for the content they want and expect to get it without a fuss, rather than one where they have to decide which “music marketplace” to shop at first, and then search for the content second. In a world where getting to content quickly and easily is paramount, and every provider has the same content, the providers that control distribution win out.
Problem 2: How People Pay for Music
Customers “pay for” music in two ways. They either listen to interstitial advertisements between tracks, or they pay actual money to listen ad-free. Spotify has done a tremendous job converting its ad-supported users into paid ones, and has over 100 million paid subscribers (40+% of its monthly user base).
It also makes about 90% of its revenues from paid customers—ad-supported users only contribute about 8% of revenue (and are on average unprofitable). One could describe the ad-supported business as a giveaway to customers in order to continue upselling them on the paid Spotify offering.
All this is to say that the money in music is made in premium subscriptions. But the advantages of a paid subscription on Spotify (as compared to a free account) are not that significant, beyond ad removal. At present, a free user can satisfy needs 2, 3, and 4 from above—they can listen to a specific artist, a genre playlist, or a personal playlist on shuffle. The only need they cannot satisfy is to listen to a particular song on demand.
Is Spotify’s paid business sustainable, based on (1) single-track selection and (2) ad removal? I submit that it is not, as:
Most users don’t care enough to pay today
Other streaming providers can provide these benefits in a way that “seems free” to most customers
(More speculatively) the entire business model of bundled subscription music might collapse in a world where discovery is content-first, not app-first
Let’s walk through these points in order.
The first point is simple. Even with massive effort—discounts, partnerships, relentless annoying ads, etc—Spotify has not been able to convert the majority of its users to the paid tier. This indicates that a majority of customers are simply unwilling to pay, and are willing to forgo track selection and listen to ads.
The second point is probably Spotify’s most obvious strategic concern. Three of the largest companies in the world (Amazon, Apple, and Google), see Spotify’s 30% margin on its paid users and lick their chops. Amazon doesn’t need to make money on music—it just needs customers to spend more money on its online marketplace. Apple would give away music for a few more iPhone sales without question. Google sees music as a nice addendum to its ad-free YouTube subscription.
All three are happy to package music as a “seemingly free” add-on to other subscriptions. Apple is adding it to its Apple One subscription bundle, Google includes it in YouTube Premium, and while Amazon still isolates “Music Unlimited” as a paid standalone service, you have to imagine Bezos would love to roll it into Prime (and will probably get around to it).
So, Spotify is fighting a four-way battle with companies that want to make ad-free, on-demand music a throw-in as part of other subscription bundles that drive their core businesses. This issue, by itself, might be the disruption of their business (particularly in combination with the distribution issues laid out above).
But as a final thought—things might get even worse for Spotify. It’s possible, though not assured, that once artists recognize that a significant portion of music listening is driven by users invoking their names directly (e.g. “Alexa play Kanye West), they’ll want to control that relationship directly.
If I tell Alexa or Google to play Kanye, shouldn’t Kanye control that experience? Shouldn’t Kanye personally welcome me to his voice-controlled application, and offer to play my favorite album or some of his newest tracks? Shouldn’t he be able to sell me merchandise and live-show tickets with native, personalized “ads” between tracks? Shouldn’t I be able to subscribe to “Kanye Gold” for $9.99 a month to get exclusive access to music drops and behind-the-scenes videos? Shouldn’t I be able to pay even more to get a personal birthday DM from Kanye?
In a world where music is summoned to life by directly invoking track titles or artist names, isn’t the natural endgame that the artists reclaim distribution of music from the marketplaces? Sure, this experience would only satisfy needs 1 and 2 above (we’d need a solution for playlists and genre stations), but it seems logical that the music subscriptions of the future will be directly to the artists themselves.
This type of unbundling is happening in journalism—individual journalists are claiming a direct relationship with their readers via platforms like Substack, rather than relying on aggregated publishers to collect revenue and parcel it out to them parsimoniously. The analogy isn't perfect, since most of the value in music is in the label-controlled "back catalog" of older, successful tracks, so the artists would still have to work with the companies that control their masters to build direct relationships with fans.
But the labels, publishers, and PROs could still collect their 70% rake in this universe, passing along the rest to artists directly, and this would dramatically expand the revenue that artists could collect directly from fans. So the entrenched interests might support this future.
Of course, it’s possible this artist-first music universe will never happen—maybe the software platforms will squash it, or the music companies will hate it for some reason that I fail to foresee. But if it does happen…Spotify is in even deeper trouble.
