Voice Memo Triage: Turning Four Hundred Fragments Into a Short List

Name every memo with three fields

The naming convention is the whole method. Every memo gets renamed to `YYYY-MM-DD | stage | one-line description`. The stage is one of your working stages — for this library, idea, verse, chorus, bridge, or demo. The description must be a description, not a judgement: 'chorus, three chords, high ending', not 'good one'.

• Date first, so sorting by name sorts by time.
• Stage second, so you can filter to only choruses when you need one.
• Description last, and never a compliment — compliments are unsearchable.

Delete on a fixed rule, not a feeling

Feelings change between sessions, so deletion needs a rule. Use this one: any memo that does not contain a rhythm you could tap without listening twice gets deleted. That removes roughly forty percent of a typical pile in the first pass.

Promote, do not hoard

After triage, every surviving memo gets one of three labels: bed (a loop or groove), line (a lyric fragment), or song (something with two sections). Only the ones labelled song go onto the short list, and the short list should never exceed twelve items.

The twelve-item cap is deliberate. A list you can read in one screen is a list you will use; a list you have to scroll is one you will avoid. When a thirteenth song-quality memo appears, the oldest one either gets promoted to a working file or demoted to bed.

Run it on a schedule, not when you feel like it

Triage once every two weeks, for twenty minutes, at the same time. The interval matters because a pile of fewer than ten memos is not worth the ritual, and a pile of more than sixty takes an hour and you will skip it. Two weeks keeps you in the twenty-to-forty range, which is the sweet spot.

The measurable payoff appears around month three: you stop starting songs from nothing, because you now have a searchable list of beds, lines, and songs that have already survived one round of judgement.

Questions that come up here