The Economics of Streaming

Spotify has over 600 million users. Of those, roughly 250 million pay for a subscription. The other 350 million listen for free, interrupted by ads, and generate a meaningfully different kind of revenue — one that flows through the system in a completely different way. Understanding that difference is the first step to understanding why the streaming economy works the way it does for everyone inside it, including the artists who fund it.

Two Pools, Not One

The royalty system on major streaming platforms is not a pay-per-stream model in the way most people imagine it. There is no fixed rate — no agreed fee that a platform pays each time a song is played. Instead, platforms operate what is called a pro-rata pool system.

Here is the basic mechanics. Each month, a platform takes all the money it has collected — from subscriptions, from advertising revenue — and puts it into a single pool. It then looks at the total number of streams across the entire platform and calculates what percentage of those streams each artist, song or rights-holder was responsible for. That percentage determines their share of the pool.

This has a profound structural consequence. Your royalty is not determined solely by how many streams you receive. It is determined by how many streams you receive relative to everyone else on the platform. If Drake releases an album that consumes a quarter of all global streams for a month, the pool shrinks for everyone else. You are not competing against a fixed price. You are competing against the entire catalogue.

The two revenue streams — subscription and advertising — typically feed separate pools, though the precise mechanics vary by platform and licensing agreement. This matters because ad-supported listening generates significantly less per stream than premium listening. A user paying £10.99 a month contributes a known, stable amount to the pool. A free-tier user generates revenue based on how many ads were sold against their listening session, what those ads were worth in that market, at that time of day, in that advertising climate. That number fluctuates constantly, and it is almost always lower.

What Platforms Actually Pay

Platforms do not pay artists directly. They pay rights-holders, which is a category that typically includes record labels, distributors, and publishing administrators. The platform licenses the right to stream recordings and compositions separately — the master recording and the underlying composition are distinct in copyright law, and both require licensing.

When a stream occurs, the platform pays out to whoever holds the master recording rights (usually the label or the artist’s distributor, if they’re independent) and separately to whoever administers the publishing rights for the composition (the melody and lyrics). These are governed by different agreements and different royalty rates, which is why a songwriter who does not perform their own music can receive streaming income from recordings they never made.

What reaches an artist from the master side depends entirely on their contract with their label or distributor. A major label deal might see an artist receive 15–25% of the label’s receipts after recoupment of advances. A distribution deal with a service like DistroKid or TuneCore typically passes through 80–100% but charges an upfront fee or takes a commission. An artist who owns their masters and has a direct distribution arrangement keeps the most. The platform’s payout rate is only one part of the calculation.

Industry estimates for per-stream payouts consistently land somewhere between $0.003 and $0.005 on Spotify, though this figure is an average across all agreements and pool calculations — it is not a rate that exists anywhere in a contract. Apple Music’s model, which weighted more toward its fully paid subscriber base historically, has tended to produce higher per-stream averages. Tidal and Amazon Music Unlimited, with smaller but predominantly premium user bases, have historically generated higher per-stream figures as well, though they reach fewer listeners.

The Playlist as Infrastructure

Algorithmically, streaming platforms do not think of playlists as a cultural curation exercise. They think of them as retention tools. The platform’s commercial interest is in keeping users listening for as long as possible — longer listening means more subscription renewals, more ad impressions, more data. Playlists are the primary mechanism for extending session length.

Editorial playlists — the ones curated by human teams at platforms like Spotify’s New Music Friday or Apple Music’s equivalents — carry enormous weight precisely because they are trusted by users who have learned that the recommendations are reliable. Getting added to a prominent editorial playlist can shift stream counts by orders of magnitude in a week.

But the algorithmic playlists have become more commercially significant over time, because they scale. Spotify’s Discover Weekly, Release Radar, and the personalised radio features are generated by machine learning systems that model listening behaviour across the entire user base. These systems look at patterns: what users with similar listening histories play next, how long they listen before skipping, whether they add a track to their own library, whether they return to it. A song that triggers positive engagement signals — completions, saves, playlist adds — gets surfaced to more users with similar profiles.

This creates a feedback loop that heavily favours certain structural qualities in recordings. Songs that hold listeners through the first thirty seconds — the threshold at which most platforms count a stream — are algorithmically advantaged. Songs that build slowly, that require context, that reward repeat listening over immediate engagement, face structural headwinds in a system optimising for retention signals.

Discovery and What It Actually Means

Platforms market themselves to artists partly on the promise of discovery — the idea that a great song, once uploaded, can find its audience without the gatekeeping structures of radio, retail or traditional media. This is partly true and partly a compression of something more complicated.

Algorithmic discovery does surface music to genuinely new audiences. Artists like Billie Eilish, Omar Apollo and Phoebe Bridgers all benefited from streaming-era algorithmic amplification before they had mainstream label support. But the algorithm does not operate in a vacuum — it amplifies existing signals. A song with no streams, no saves, no playlist adds, no social activity feeding into it has no signals to amplify. The discovery system is better understood as a multiplier than a starting mechanism.

Pitching a track to editorial playlists — which Spotify allows directly through Spotify for Artists ahead of release — is one way to seed those initial signals. Getting a track placed in a sync deal (television, film, advertising) creates a spike that algorithms read as a signal of cultural relevance. Social media activity that drives searches and direct streams feeds back into the platform’s understanding of a track’s momentum. Discovery, in algorithmic terms, means: the system noticed that something is happening here and decided to show it to more people. The something has to start somewhere outside the algorithm.

Payola’s Modern Form

Spotify introduced a feature called Marquee — later expanded into a broader promotional suite — that allows rights-holders to pay to have tracks promoted to users who have already shown interest in an artist. This is targeted advertising within the platform’s own ecosystem, and it is disclosed to users. It is not payola in the traditional sense.

What exists in a greyer space is the practice of playlist placement-for-payment, which has operated through third-party playlist curators for years. Independent playlist curators with large followings have accepted payment for placement, creating a market outside the platform’s official structures. Spotify’s terms of service prohibit paying for streams but have navigated the playlist curator economy carefully. The practice muddies the signal quality that the algorithm relies on, which is why platforms have incentives to suppress it even when enforcement is inconsistent.

The more structurally significant version of this problem is the existence of content farms — entities that upload enormous volumes of low-cost, algorithmically optimised content (often ambient, lo-fi, or functional music) designed to capture pool share with minimal creative investment. Because the pool system rewards relative share rather than absolute value, capturing even a fraction of a percent of total streams across millions of tracks represents real money. Platforms have made efforts to identify and remove this content, but the incentive to produce it is a direct consequence of how the pool model is designed.

The Label Minimum Guarantee Problem

Major labels negotiated their early licensing agreements with streaming platforms from a position of power — they controlled the catalogues the platforms needed to be viable. Part of those negotiations historically included minimum guarantee payments and equity stakes. Spotify’s pre-IPO equity was partially distributed to major labels as part of its licensing agreements, meaning that when Spotify went public in 2018, Universal Music Group, Sony Music and Warner Music Group all held shares that generated significant returns.

This created a structural alignment between the major labels and the platform’s commercial success that does not exist for independent artists or small labels. The labels have incentive to see the platform grow. Whether that growth is good for the broader ecosystem of artists who feed the platform is a separate question.

The minimum guarantee structure also means that if the pool falls below the guaranteed level in a given period, the platform makes up the difference. This insulates major label income from downturns in a way that purely pool-dependent independent artists are not insulated. The floor exists for some participants and not others.

What the Numbers Require

To earn the equivalent of a US federal minimum wage annually from Spotify streams alone — roughly $15,000 — an artist would need to accumulate somewhere between three and five million streams per year at average payout rates, assuming they own or control a favourable portion of their master rights. That is not accounting for publishing income, which arrives separately and adds to the total, but it illustrates the scale at which streaming becomes economically meaningful as a primary income source.

For catalogue artists — those with decades of recordings and large back libraries — streaming has been transformative. Music that once stopped generating income when records left retail shelves now earns continuously. For emerging artists building an audience, streaming income is better understood as a signal of reach than a primary revenue mechanism. The economic return comes later, and elsewhere: from live performance, sync, merchandise, licensing, direct fan support.

The streaming economy is not broken in the sense of being dysfunctional — it functions exactly as its incentive structures predict. The question worth sitting with is whose incentives designed those structures, and what a different set of incentives might produce.

Use the Resonillator Streaming Royalty Calculator to model how stream counts, tier splits, and rights ownership translate into actual income for your music.