Music
Playlists replaced the radio programmer, and discovery started behaving differently as a result
For most listeners, something else now decides what comes next. The shift changed not only what gets heard but what kind of record makes sense to release.

This works through playlist discovery in the order the parts actually depend on each other.
The short version
- A playlist is a context, and a record placed in the wrong one is heard against unhelpful neighbours.
- Sequencing decisions that used to be made by a station are now made for each listener separately.
- Records optimised for placement tend toward immediate legibility and away from slow builds.
Somebody has always chosen what comes next
Listeners have rarely assembled their own listening from scratch, since that requires knowing what exists before you have heard it. For most of the last century that choice was made by stations, whose programmers decided both what was played and what it sat beside.
Playlists occupy the same position but multiply it, so the choice is now made separately for each listener rather than once for a region. The function has not changed at all, and only the granularity and the identity of the chooser have. Understanding it as a continuation rather than a break makes the actual differences much easier to see.
Context is the real product
A playlist is not a container so much as a context, because the tracks around a record change how that record is heard. A song placed among things it resembles gets absorbed comfortably, and the same song among unlike neighbours can sound like an intrusion. This means placement is a genuine editorial act with consequences for how a piece of music is received.
Musicians have limited influence over it, which is a meaningful loss of control compared with sequencing your own release. It is also why two listeners can report entirely different impressions of the same track without either being wrong.
What gets rewarded
Placement favours records that establish themselves quickly, because a listener who is not paying full attention will skip anything that has not yet arrived. That pressure pushes toward shorter introductions, earlier hooks and a clearer identity in the first few seconds.
None of this is a new complaint, since broadcast radio applied similar pressure through its own constraints for decades. The difference is that the pressure is now measurable at the level of individual seconds rather than inferred from a station's output. Measurement makes the incentive sharper, and sharper incentives produce more uniform responses to them.
The mood category and what it displaced
A large share of listening now happens through categories organised by activity or mood rather than by genre or artist. This is enormously convenient and it detaches records from their origins, since a track's neighbours are chosen by function rather than by lineage. A listener can spend years inside such categories and accumulate very little sense of who made what or why.
On a second viewing, that is not a failure of the listener, because nothing in the interface encourages the connection to be made.
It does mean that the older habit of following an artist backwards through their catalogue now requires deliberate effort.
Discovery still happens, differently
Automated recommendation is genuinely effective at finding things adjacent to what you already like, which is a real and useful capability. It is much weaker at anything requiring a leap, because a leap by definition is not adjacent to your existing history.
Human curation remains better at that and is now a smaller share of the total, which narrows the average listener's range over time. The narrowing is gradual and invisible from inside, since each individual recommendation feels reasonable and usually is. The cumulative effect is what deserves attention rather than any single suggestion.
Production stories are told by people with an interest in how they sound.
What a listener can do about it
The straightforward correction is to deliberately introduce sources that do not know your history, such as another person's selection. Following an artist's own influences backwards works well, because it moves through relationships that recommendation systems represent poorly. Listening to a full album occasionally restores the sequencing decisions the musicians actually made, which placement discards entirely.
None of this requires abandoning the convenience, and the two modes coexist without much difficulty. The point is only to notice that convenience has a shape and that the shape is not neutral.
The takeaway
Someone has always chosen what comes next; the difference now is that they choose it only for you.
The premise gets you in. The structure decides whether you stay.
Questions readers ask
Are playlists worse than radio was?
They are different rather than worse. Radio narrowed everyone the same way; playlists narrow each listener individually, which is harder to notice and harder to argue about.
Does placement really change how a song sounds?
It changes what you compare it to, which is most of how impressions form. The same track among different neighbours genuinely produces different reactions.
Also by Reshma Qureshi
- Reshoots are ordinary film-making, not a distress signalFilm & Screen
- Documentary gets written in the edit because nobody could write it in advanceFilm & Screen
- What a music supervisor does, and why one song in one scene costs what it costsMusic
- Song introductions got shorter, and the incentive that shortened them is measurableMusic





