Compare the running times on a chart album from the 1990s with one from the last few years and the difference is obvious. Tracks are shorter. Intros barely exist. The first chorus arrives within the first minute, sometimes within thirty seconds.
This is a documented shift and its main cause is unusually specific: a rule about when a stream counts.
The thirty-second rule
On the major streaming platforms, a play generally registers as a stream — and therefore generates a royalty — once a listener has heard roughly thirty seconds of it.
Before that threshold, nothing. A listener who plays twenty-five seconds and skips has generated no revenue at all.
That single rule reshapes songwriting incentives completely. The critical objective becomes getting a listener past thirty seconds, which means the most compelling part of the song has to arrive early.
Hence the disappearance of the long instrumental intro. Hence songs that open on the chorus. Hence the "cold open" structure where a track begins mid-phrase with no build at all.
Why shorter is also better paid
The second effect, and it's arithmetic. Streaming pays per stream, not per minute.
A three-minute song and a six-minute song generate the same royalty per play. So if you can write two three-minute songs instead of one six-minute one, and listeners play both, you've doubled the revenue from the same amount of music.
This has produced observable behaviour: albums with more tracks, each shorter. It's also produced the practice of splitting what would once have been a single long piece into separate tracks.
Whether artists consciously reason this way varies. What's clear is that the ones who do it are rewarded, and over time that shapes what gets made regardless of individual intent.
The playlist effect
Third mechanism, arguably as important as the first two.
A large share of streaming listening happens through playlists rather than albums. Placement on a major playlist can transform a track's performance, and playlist algorithms weigh completion rate — how often listeners finish a track rather than skipping.
Shorter songs have higher completion rates almost by definition. So shorter songs perform better in the systems that determine visibility, which drives more listening, which reinforces the placement.
There's also a structural point about sequencing. A track that fits smoothly into a playlist of other tracks benefits from consistency — similar tempo, similar length, no jarring intro. Music optimised for playlist context sounds different from music optimised for album context, and the incentive now points at the former.
What's been lost
I want to be careful here, because "songs were better when they were longer" is not a defensible claim. Plenty of great music is short, and the three-minute pop single predates streaming by many decades.
What has genuinely been squeezed is a particular kind of structure: the slow build, the extended outro, the instrumental section that isn't a hook. Those require patience from a listener and the current systems don't reward asking for patience.
The album as a sequenced experience has also suffered, though that started long before streaming. A collection designed to be heard in order, with pacing across the whole, sits awkwardly in a world where most listening is shuffled.
What's been gained
Genuinely some things, and it's worth saying.
Economy of writing. A song that has to earn attention in thirty seconds tends to have less padding. A lot of 1970s album tracks are three good minutes and four indulgent ones, and nobody misses those four.
Volume and variety. Lower barriers to release mean far more music exists, from far more people, in far more places. The distribution advantages that used to determine who got heard have weakened considerably.
And the death of the filler album track. When every track has to justify itself in a playlist, the deep cut that exists purely to reach ten songs has less reason to be recorded.
The counter-movement
There's a reaction, as usual. Long-form releases, extended pieces, artists deliberately writing against the format, vinyl releases sequenced for sides.
These tend to come from established artists with audiences who'll follow them, which is the usual pattern — constraints bind hardest on people who need the systems to work for them.
What I'd note is that the constraint here isn't aesthetic taste, it's a payment threshold and a recommendation algorithm. Those are contingent things. If the royalty model changed to pay per minute listened, the incentives would flip within a couple of album cycles.
Which is a strange thought: that the average length of popular music is set less by what sounds good than by a number in a licensing agreement.
One more wrinkle
There is a related distortion worth flagging, because it affects what ends up on albums at all. Since a play counts the same regardless of length, a thirty-second interlude generates the same royalty as a four-minute song. Some artists have leaned into that quite openly, filling out releases with very short tracks, and the practice is common enough that it has a name in industry conversation.
Whether that is gaming the system or simply responding to it is a matter of framing. What it demonstrates is that the unit of payment shapes the unit of creation, which is not a new observation — the three-minute single existed because of how much music fitted on one side of a shellac disc — but it is unusual to be able to watch it happen in real time.