The Scheduling Software's Job

Music scheduling software does not choose songs. It chooses which rules matter most at any given moment, then finds a track that breaks the fewest of them. The distinction is important. Handed a library of several thousand songs and a set of instructions, the program works through each empty log slot methodically — matching, testing, rejecting, trying again — until it places something defensible. The result feels like curation. The mechanism is closer to constraint satisfaction.

01What goes in must follow rules out

The scheduler starts with a clock wheel: a template dividing the broadcast hour into labelled slots — a Gold here, a Current at the quarter, a Recurrent at the half, and so on. Each slot specifies not just a category but often a tempo range, a permitted mood, or whether the position must hold a vocal. That clock is the scheduler's outermost constraint. Everything else sits inside it.

Within those category slots, the rules stack up quickly. Artist separation is usually the first: the same artist cannot appear within a set number of minutes of themselves, whether on this hour's log or the previous hour's. Two hours is a common minimum; some formats push to three or four. Title separation goes further — a specific song might be locked out for a full rotation cycle, which on a tight current rotation could mean several days, and on a gold category might stretch to weeks. The scheduler tracks all of this in a matrix of what has played and when.

Then come the rules that shape texture rather than repetition. Tempo sequencing stops three slow songs from landing in a row. Era-spread rules ensure a classic-hits station doesn't front-load a quarter-hour with 1970s tracks and leave the 1990s until the end of the day. Gender balance rules — where used — prevent the log running consecutive male soloists across an entire drive show. Energy-arc rules attempt to build and release tension within a set. None of these is absolute; the software weights them and, when they conflict, resolves them by priority. Lower-priority rules break first.

02The scheduler's main problem: unschedulable songs

Even a well-maintained library develops what schedulers call "problem songs" — tracks that are so constrained by their own rules, their artist's separation timer, or their awkward position in a rotation cycle, that the software cannot legally place them anywhere in a given log without breaking something it is not permitted to break. Left unresolved, those songs accumulate. Run the log again without addressing them and the software will schedule around the gaps, warping everything else to compensate. In practice, music directors fix this partly by auditing rotation categories: if a category is too small for its scheduled frequency, separation rules become mathematically impossible to satisfy, and the library needs either more songs or a lower play-rate.

None of these is absolute; the software weights them and, when they conflict, resolves them by priority.

The two-hour artist rule is the one that bites hardest, because it operates across the boundary between logs. A song scheduled at 7:58 locks out that artist until at least 9:58. If the next log was built independently — as overnight automation logs often are — the software may not see the carry-over unless the systems share a real-time database. Stations running unified scheduling platforms avoid this; those pasting logs together from separate exports do not always catch it until a listener does.

03What the scheduler cannot do

Scheduling software is indifferent to narrative. It does not know that a lyric about loss sits badly after a news bulletin about a disaster, or that a song's tempo fits the moment only if the presenter's link before it runs long. Those judgements remain human. Most platforms allow a music director or presenter to lock, move, or override any placed track — and on a well-run station, some editing of the automated log happens every day before it goes to air.