Answers/Original research/What Music Platforms Actually Pay Per Play — Seven Years of My Own Royalty Data

Original research

What Music Platforms Actually Pay Per Play — Seven Years of My Own Royalty Data

By Boboy Watson·Last verified August 31, 2026·Original research

Everybody argues about per-stream rates using numbers they read somewhere. I exported my entire distributor history instead — every royalty line since my first release — and calculated it.

45,894 royalty lines in the export. 677,639 verified plays. $3,081.30. June 2019 to July 2026. 48 platforms, 150 songs, 178 countries.

There are another 41.3 million plays in that export that I am not counting, and I will show you exactly why before the end of this page. Everything above this line, and every rate below it, is built from the plays I can stand behind.

Here is what it says.

Method, and its limits

I exported the complete unfiltered bank-details report from my distributor and aggregated it. Three decisions matter for reading the numbers:

Rates are normalised for split percentage. On a track where I hold 50%, my earnings are half the royalty. Using raw earnings would understate what the platform actually paid, so per-play rates divide out my share. Totals are reported as actually received.

Play types are kept separate. A download, a lyric display, a radio spin and an on-demand stream are different units. Averaging them together produces a headline that collapses under scrutiny, so they are categorised and reported apart.

Small samples are excluded. Anything under 500 plays is noise and is left out of the published tables.

What this is not: these are net amounts reported to me by one distributor, not gross platform payouts. It is one catalog — Hawaii reggae and hip-hop, US-weighted — so my country mix is not yours. And critically, this is recording-side money only. Nothing here includes performance royalties, mechanical royalties, or SoundExchange. Those are three separate pots that never appear on a distributor statement at all.

Every platform, ranked

The top of that table and the bottom of it are separated by a factor of about 2,900.

The finding that reorganised how I work

Group those platforms by what kind of play they are, and the picture gets brutal:

18.66% of my verified plays produced 0.47% of my money.

Sixty-three percent of my plays — the on-demand audio streams — produced seventy-four percent of it. Social and short-form, nearly a fifth of everything I can verify, produced fourteen dollars and fifty-four cents across seven years.

I spent years chasing the number in the first column. Every viral Facebook video, every big view count, every screenshot of a play counter. It was the wrong column. The money was in the row I was barely paying attention to.

The 41 million plays I am not counting

The first version of this page led with a different number. It said 98.7% of my plays came from social video and produced 1.6% of the money, and the number doing the work in that sentence was 41,302,112 Facebook plays for $36.14. My most-played song had 29.96 million of them.

Then I looked at where those plays were.

My real audience — across every Spotify stream I have ever had — is 73% United States, then New Zealand, the UK and Germany. A song with thirty million genuine plays does not have 771 of them on Spotify, and a beat made in Puna does not go viral in Argentina while nobody in Hawaiʻi notices. Both songs spiked in a single month — 14.5 million in May 2020, 6.3 million in August 2019 — and decayed like a trend, not a catalog. Outside the "Facebook" store, 99.99% of those plays do not exist.

What happened is the same mechanism I describe on the Splice and Content ID page, running in the other direction. Facebook's Rights Manager fingerprints audio. Both of those songs are built on non-exclusive loops. When somebody else's video used the same loop and went around Spanish-speaking Facebook, the fingerprint matched my registered recording, and the plays — and thirty-six dollars — were attributed to me. They are real royalty lines. They are not plays of my music.

So I have taken the "Facebook" store out of every rate, share and headline figure on this site, and I report it separately: 41,302,112 plays, $36.14, 6,562 royalty lines, excluded. The plays on my own Facebook posts — the Social Media Pack line, 70,270 plays and $1.79 — stay in, because those are mine.

The verified numbers are smaller and the argument survives them intact. To put the same catalog side by side: Apple Music gave me $1,110 from 175,162 plays. All of social and short-form combined gave me $14.54 from 126,426. Same songs, same years.

Why a Facebook play is not a Spotify stream

I want to be fair about what these numbers mean, because the comparison is not clean and pretending otherwise would be dishonest.

A Facebook "play" is mostly my song appearing inside somebody else's video, autoplaying in a feed, often muted, usually for a few seconds, and almost never chosen. A Spotify stream is a person deliberately playing my song for at least thirty seconds. Those are genuinely different events, and it is reasonable that they are not worth the same.

So the low rate is not straightforwardly a scandal. But two things remain true regardless:

The strategic conclusion does not change. If you are an artist deciding where to put your effort, the ratio is the ratio. Feed volume does not convert into money at any scale you can realistically reach. To match my Apple Music earnings on Facebook rates, I would need roughly 1.27 billion plays.

The scale is not intuitive to anyone. Nobody looking at 41 million plays guesses $36. That gap between what a number feels like and what it pays is precisely what makes creators optimise for the wrong thing for years, and it is why I published this.

The same platform pays wildly different rates by country

Where your listeners are matters as much as which platform they use. Spotify, my catalog, 1,000+ play markets only:

A US Spotify stream paid me about 47 times what an Indian one did. Same platform, same song, same month. Rates track local subscription pricing, so a viral moment in a low-ARPU market can add enormous play counts and almost no revenue.

Rates over seven years

The story people tell is that per-stream rates are collapsing. In my data that is true for one platform and false for the others.

Spotify and Apple Music are essentially flat. YouTube Premium is flat and highest. YouTube's ad-supported tier fell about 24% across five years — the only genuine decline in my catalog.

The 2023 Spotify dip is a mix effect, not a rate cut: that was my highest-volume year in lower-paying markets.

One more thing the data says about downloads

Eighty-seven iTunes downloads paid me $141.84. That is $1.63 per sale — around 500 times what a Spotify stream pays.

Downloads are a rounding error in volume and 7% of my lifetime revenue. It is worth remembering that selling something to a person who wants it still outperforms renting attention by orders of magnitude.

What I actually changed

• Stopped treating view counts as progress. A million feed plays is worth about eighty-eight cents.
• Started optimising for saves and playlist adds, not reach — the on-demand row is where the money is.
• Stopped paying to boost content in low-ARPU markets. Volume there flatters the dashboard and does nothing for the bank.
• Went and did the registration paperwork. Everything above is one of four royalty streams. Finding out that 73% of my recording money came from 1% of my plays made me go check what I was missing on the other three.

That last point is the one I would press hardest. This entire study measures a single pot. If you have not registered with a PRO as both writer and publisher, with the MLC, and with SoundExchange, then the money in this analysis is the only money you are collecting at all.

I am an artist reporting my own numbers, not a lawyer or an accountant. This is one catalog in one genre with a US-weighted audience — treat the ratios as instructive and the absolute figures as mine, not yours.