Why you forget why the song idea was good

I have a folder of voice memos. Most of them are useless.

Not because the playing is bad, though it often is. They’re useless because I can hear the thing perfectly well and have no idea what I meant by it. Forty seconds of a descending line, recorded at some point last spring, apparently urgent enough to reach for a phone. Now it’s just a chord sequence, and chord sequences are cheap.

The memo did its job. It saved the fragment. What it didn’t save was the reason.

The artifact and the intent

When an idea arrives it comes as a package. There’s the fragment — the actual notes, the sentence, the shape. And there’s everything around it, which is what made you reach for the phone:

• what it felt like, and how specifically
• what it reminded you of
• what you imagined it becoming
• the thing that was slightly wrong with it, that you were going to fix
• the one experiment you’d have tried next if you’d had another twenty minutes

Recording captures the first item. The rest evaporates, and it turns out the rest is most of the value. The fragment is reproducible — I can find another descending line in about a minute. What I can’t reconstruct is why that one felt worth interrupting my evening for.

This is the creative version of mental offloading: the artifact survives, but the working state doesn’t. Creative work just makes it more obvious, because the gap between capture and return is measured in months instead of hours, and because a half-finished song has no ticket number to give you a running start.

Thirty seconds after the memo

The fix is small and slightly annoying, which is why it took me years to actually do it: after you stop recording, before you put the phone down, write five lines.

The fragment. What it literally is, in enough detail to find it again. “Descending line under the second half of the verse, ~96bpm, capo 3.” Include the tempo and the tuning. You will not remember the tuning.

The feeling. Be specific and resist the obvious word. “Sad” is useless — it describes half of everything. “Resigned rather than sad, someone who’s already decided to leave” tells six-months-from-now-you what they’re aiming at. This line is the one that does the most work and the one people skip.

Where it came from. The thing it’s borrowing from, arguing with, or accidentally quoting. “The way that old folk melody drops at the bridge instead of lifting.” This is often faintly embarrassing to write down, which is a good sign. It’s also the fastest route back into the mood.

What’s wrong with it. There’s almost always something. “The chorus lifts too early and undoes the mood the verse just built.” Naming the problem is what makes the idea workable later rather than merely pleasant — you return with a job to do instead of a nice thing to admire.

What to try next. One concrete experiment. “Hold the verse chord under the first chorus line and see if the lift survives.” Not a plan. One move.

Thirty seconds. It’s the difference between a folder of artifacts and a folder of open work.

Why you don’t do it

Two real reasons, and I don’t think either is laziness.

The first is that it feels like it breaks the spell. You’re in the middle of something and stopping to annotate feels like stepping outside it. My experience is the opposite — writing “what’s wrong with it” usually generates the next idea directly, because naming the problem is already halfway to solving it. But the fear is genuine and it’s why the note has to stay at five lines. Anything more and it really does become admin.

The second is optimism. In the moment it’s inconceivable that you’ll forget. The idea is right there, vivid, obviously memorable. Research on confidence-guided offloading shows that confidence influences whether people create external reminders, independently of their actual unaided performance. In the moment, confidence is abundant. Six months later, the missing context says it was misplaced. I keep proving I’m not the exception.

Six months later

The test isn’t whether you can find the idea. Search solves that. The test is whether, on opening it, you know what to do in the next ten minutes.

That’s a much higher bar, and it’s the one that decides whether a folder of fragments turns into finished work or just gets larger. Ideas don’t usually die because they were bad. They die because by the time you came back, the only thing left was the artifact, and the artifact alone isn’t enough to act on.

I’m building something around this — holding onto the whole idea, not just the recording — and it’s early enough that the shape can still change. If you have a folder like mine, I’d like to know what you’d want to find in it.