Why Genre Is a Bad Way to Recommend Music

Ask someone what music they like and they will name genres. Ask them to name three songs they have loved this year and the genres will rarely line up. That gap is not inconsistency on their part. It is genre doing what genre actually does, which is something other than describing sound.

Where the categories came from

Almost every genre label you use was invented for a commercial reason, by someone who was not a musician.

"Rhythm and blues" was coined by Billboard in 1949 as a replacement for its "race records" chart — a marketing category defined by the race of the performer, not by any musical property. "Alternative" meant alternative to commercial radio formats, a distribution fact. "Indie" originally meant released on an independent label — a contract, not a sound — which is why it now describes a guitar tone that has nothing to do with anyone's label.

Genres are, mostly, retail and radio infrastructure that outlived its purpose. They tell you which bin a record went in, and bins were a solution to a physical problem nobody has anymore.

Four ways genre fails as a recommendation

1. It groups by scene, not by sound

Two bands from the same city, playing the same venues, get filed together regardless of whether they sound anything alike. Meanwhile a producer in Reykjavík and a producer in Detroit making almost identical records land in different sections because their social worlds never touched.

2. It drifts

Genre labels move under your feet. "Emo" in 1994, 2004 and 2024 are three unrelated sounds. Someone who says they like emo is telling you which of those decades formed them, which is useful information, but it is biographical rather than acoustic.

3. It splinters into uselessness

The response to genre being too coarse has been to make more genres. There are now thousands of tracked microgenres. This does not fix the problem, it inverts it: a label so specific that only 300 people can define it is not a shared vocabulary, and cannot function as a recommendation to anyone outside those 300.

4. It says nothing about the situation

This is the one that matters most. Genre is a property of the record. What you need is a match between the record and your evening — and your evening is not in the metadata. "Jazz" contains both a frantic bebop session and a quiet solo piano record. They are useless substitutes for each other at midnight.

Where genre is genuinely useful

It would be overcorrecting to say genre is worthless. It carries real information about instrumentation, production conventions and lineage, and it is an efficient shorthand between people who share a reference frame. "Post-punk" tells another listener a great deal in one word.

It is also the right filter at the end of a search. Once you know you want something low-energy, warm and instrumental for the next hour, "jazz or ambient?" is a perfectly good next question. The mistake is making it the first question, because it closes off most of the catalog before you have established anything about what you actually need.

Ask about the situation first

The reordering is simple. Instead of "what genre am I in the mood for" — a question that most people cannot honestly answer — ask:

1. How much energy should the next hour have?
2. Do I want it bright, or do I want it to sit with me?
3. Words or no words? (This one decides more than genre ever does.)
4. Familiar, or something I've never heard?

Four answers, and the catalog collapses to something manageable — across every genre at once, which is the point. That's how the song quiz is ordered, and why genre is the last thing it asks about rather than the first.

For more on the listener's half of this — how music actually moves a mood, and why sad music helps a sad one — see what your mood says about your music.