Finish the Song: A Field Guide to Beating Perfectionism

The Graveyard in Your Projects Folder

Every artist I've ever worked with has one: a folder of thirty, fifty, a hundred unfinished songs. Verses without choruses. Hooks without verses. Full arrangements abandoned because the vocal "wasn't right yet."

Here's what A&R taught me: the difference between artists who build careers and artists who build folders is almost never talent. It's the willingness to call something done and let it leave the house.

Perfectionism Is Fear Wearing a Lab Coat

Perfectionism presents itself as high standards. Usually it's something simpler: as long as the song is unfinished, it can't be judged. An unfinished song is all potential. A finished song is a fact — and facts can be criticized.

So the endless tweaking isn't quality control. It's a stalling tactic your brain dresses up as craftsmanship. Naming that is step one, because you can't out-technique a fear problem. You have to change the rules of the game instead.

The Rules That Actually Ship Songs

Define done before you start. Before you record a note, write one sentence: "This song is done when it has two verses, two choruses, a bridge, and a mixed vocal." Done is a spec, not a feeling. Feelings will move the goalposts forever; a spec can be met.

Adopt the 90 percent rule. Your last ten percent of tweaking produces changes nobody but you can hear, and half of them make the song worse. When you catch yourself undoing yesterday's edits, the song is finished. That's not a failure state — that's the signal.

Cap your versions. Give every song a maximum of three full revision passes. Version one is the sketch, version two is the fix, version three is the polish. Whatever exists at the end of version three ships. Constraints don't limit creativity; they force decisions, and finishing is nothing but a stack of decisions.

Set a calendar deadline with a witness. "I'll release it when it's ready" means never. "It goes to my group chat Friday at 8" means Friday. The witness matters more than the date — perfectionism thrives in private.

Finish ugly first. Get a complete bad version of the song end to end before you improve anything. A rough one-take vocal over a simple beat — the kind you can knock out in Volmix in an evening — gives you a whole song to react to. Editing a complete draft is a fundamentally easier psychological task than perfecting fragments, because the question shifts from "is this good enough to exist?" to "how do I make this better?"

Ship small to build the muscle. If releasing a single feels enormous, post a chorus. Finishing is a muscle, and muscles grow with reps, not with one heroic lift. Every completed thing makes the next completion cheaper.

What Finished Actually Buys You

Released songs generate the only feedback that improves your next song: real listeners reacting to a real thing. Your folder of demos generates nothing. It can't find fans, it can't teach you what connects, and it can't get better sitting there — songs don't age into completion.

There's also a compounding effect nobody mentions: the artist who releases ten imperfect songs in a year learns roughly ten songs' worth of lessons. The perfectionist learns one, maybe. In three years that gap is unbridgeable.

Done Is a Skill

Your favorite artist has released songs they now find flawed. Every artist has. The flaw tolerance isn't a lack of standards — it's the understanding that a catalog is built out of finished songs, and finished is a skill you practice, not a verdict you await.

Pick the demo closest to complete. Give it a spec, a version cap, and a Friday. Then let it go be a song.