The per-stream rate is a myth: how Spotify actually pays

Ask three people in the music business what a stream is worth and you will get three different numbers, each defended with total confidence. All three are true, and all three are useless on their own. The per-stream rate is not a fixed price. It is an average, dressed up as a fact, and it conceals more than it reveals. For anyone pricing income from recorded or published rights, the discipline is not to find the "right" per-stream rate. It is to model the mix that produces it.

Start with the mechanism. A digital service provider does not pay a fixed sum per play. It pools the subscription and advertising revenue attributable to a market and a period, then allocates that pool across rights holders according to their share of total streams in that pool, net of the platform's own commission. The blended rate that falls out the other end is therefore a function of several moving parts at once: the subscription price in that territory, the proportion of listening on free or ad-supported tiers versus paid tiers, the mix of individual, family and student plans, the currency and its movement against the reporting currency, and whether the payout mechanism is pro rata or user-centric. Two catalogues with identical stream counts in a given month can produce materially different income if their listener bases sit in different markets, on different plan types, or on different platforms.

Why the average conceals the mechanism

This is why a single headline "per-stream rate" quoted in the press, or worse, applied as a constant multiplier to a stream projection, is noise rather than signal. It flattens a genuinely complex allocation process into one number, and that number will not travel. A catalogue weighted towards higher-ARPU markets and paid subscription tiers will carry a materially different blended rate to one concentrated in markets with lower subscription pricing or a higher share of ad-supported listening. Applying the wrong blend, even with an authoritative-sounding rate, produces a forecast that looks precise and is wrong.

Modelling the mix, not the myth

The more durable approach, and the one that holds up under lender or investor scrutiny, is to build the income model bottom-up from the mix rather than top-down from an assumed rate. That means disaggregating historical performance by platform, by territory, by subscription versus ad-supported listening, and by the specific distribution or label agreement terms that determine what share of gross reaches the rights holder before it ever reaches an artist or a catalogue owner. It means testing sensitivity to plausible shifts in that mix: a move in the listener base towards markets with different pricing, a change in the balance between paid and free-tier consumption, or a renegotiation of platform-level payout mechanics. None of this produces a single tidy number. It produces a range, with the drivers of that range made explicit.

That is the difference between a headline and an analysis. The per-stream rate was never the number that mattered. The mix behind it always was.