Song intros and streaming payouts

Why did pop songs lose their long introductions the year streaming began paying by play?

Put on a hit from 1985 and you will often wait fifteen or twenty seconds before anyone sings: a drum figure, a synth wash, a guitar building up. Put on a hit from last year and the vocal is frequently there before you have finished settling into your seat.

The tempting story is that streaming did it — that a platform started paying per play and the intro was the first thing thrown overboard. Before we accept that, though, it is worth asking what exactly a payment rule can and cannot do to a piece of music. A royalty formula does not write bars. It changes what writing a particular bar is worth. So the real question is not "did streaming shorten intros" but "what, precisely, did streaming make an intro cost?"

Consider how the money actually arrives. On the major services a stream becomes royalty-bearing once the listener has heard roughly thirty seconds of the track. Sit with the shape of that rule for a moment. It is not proportional. Twenty-nine seconds pays nothing; thirty-one seconds pays the same as three full minutes. Which means the entire commercial value of a recording is decided in a window shorter than most songs' first chorus, and everything after that threshold is, from the payout's point of view, free.

Now ask what an intro does in that window. Its traditional job was to establish a groove and build anticipation for a listener who had already committed — someone who dropped a needle, bought a ticket, or waited through a radio DJ. Anticipation is a pleasure that depends on captivity. But the streaming listener is holding the skip button, has an unlimited catalogue behind it, and pays nothing extra for abandoning you. Under those conditions an instrumental build is not building anticipation; it is spending the only seconds that count on material the listener cannot yet evaluate.

So the incentive inverts. Once payment is a threshold rather than a rate, the writer's problem stops being "how do I reward attention" and becomes "how do I survive to second thirty." The strongest hook goes first, the title arrives early, and the intro shrinks to whatever is needed to establish key and tempo — sometimes to nothing at all.

Does the evidence bear that out? Partly, and the caveat is the interesting half. Hubert Leveille Gauvin's 2017 study of United States top-ten singles found average intro length falling from around twenty seconds in 1986 to about five seconds by 2015, alongside faster tempos, earlier first mention of the title, and more words per second. Notice the dates. The decline is a three-decade slope, well underway before Spotify launched in 2008 or reached scale in the United States. Whatever caused it did not begin with per-play royalties.

What would explain a long slope? An attention economy that predates streaming: radio programmers cutting intros to keep a listener from turning the dial, the skip button on a CD changer, the shuffle on an MP3 player, then the preview clip on an online store. Each of these made abandonment slightly cheaper for the listener, and each nudged the same way. Streaming did not introduce that pressure. It did something sharper: it wrote the pressure into the payment formula, so that a behaviour previously punished only in vague reputational terms was now punished in cash, per listener, and measured.

And it added a second incentive nobody had before. If revenue accrues per qualifying play rather than per album sold, then two two-minute tracks pay twice what one four-minute track does. That is a direct subsidy on brevity and on track count, and it is visible in the rise of very short songs, interludes counted as tracks, and long deluxe editions.

Two corrections keep this honest.

The first is about causation. It is very hard to separate "artists responded to the royalty rule" from "artists responded to skip data," because the same platforms deliver both, and skipping is punished twice over — no royalty, and a negative signal to the recommendation system that governs playlist placement. Playlisting is plausibly the larger force, since it decides how many listeners hear the track at all. Anyone claiming to have isolated the thirty-second rule as the cause is claiming more than the data supports.

The second is about the word "lost." Long introductions have not vanished; they have migrated to contexts where attention is already bought. Live performance still opens with an extended vamp. Film and game scores still build. Genres whose listeners arrive committed — much of metal, jazz, ambient, progressive rock — keep their intros comfortably. What changed is not a musical capability but which venue rewards it, which is exactly what an incentive argument should predict: not the disappearance of a form, but its retreat to the places where it still pays.

It is also worth noticing that this is not a story of decline. A constraint that forces the strongest idea into the first four seconds is a demanding brief, and songwriters have met it with real craft. The sonnet was a constraint too.

• Whether the same threshold explains why choruses now often arrive twice before the first verse ends.
• What happened to album sequencing once the track, not the album, became the unit of payment.
• How per-stream economics differ under user-centric versus pro-rata royalty pools.

Every term the collection defines is gathered in the glossary.