Blog · August 24, 2026
How to finish a track: ship at 70%
Somewhere on your drive is a folder with thirty project files in it, and none of them is a song.
Perfectionism tells you they are unfinished because they are not good enough yet. The truth runs the other way: they are not good enough because they are unfinished. Arrangement, mixing, and the act of letting go are skills you only practice at the end of a track, and if you never reach the end, those are exactly the skills that never improve. Your tenth finished track will beat your first polished loop, and the only way to reach the tenth is to ship the first nine.
That is what “ship at 70 percent” means. The last 30 percent of polish has the worst exchange rate in music production: days of tweaks that only you can hear, learned on a track you are already sick of. The next track teaches you more than the next revision. So here is a way to get from eight bars to a finished two-minute track on a deadline, with the labor squeezed out of every step.
Step 1: define done before you touch anything
“Finished” is not a feeling. Write down a definition you can check:
- Two minutes, not six. At 128 BPM that is 64 bars, which is eight 8-bar sections. Countable on one screen.
- An intro, one peak section, one breakdown, an outro. No second drop, no bridge, no key change.
- An exported audio file with a real name, sent to one person, by Sunday.
Small scope is not a compromise. Plenty of great electronic tracks say one thing for two minutes and leave. A two-minute track you finish this week is worth more to you than the six-minute epic you finish never.
Step 2: set a deadline that hurts a little
“Could this be better?” is always true, which makes it useless as a stopping rule. “Is this good enough for Sunday?” has an answer. That is the entire function of a deadline: it converts an unanswerable question into an answerable one.
Pick a date close enough to feel. Today if the loop is strong, this weekend if not. Tell the person you are sending it to, so the deadline exists outside your head.
Step 3: freeze the palette
Option paralysis kills as many tracks as perfectionism does. The rule that fixes it: the sounds you have when arranging starts are the sounds in the track. You can swap a snare that turns out to be wrong. You cannot go looking for a better one that might exist. Browsing is scope creep with a search box.
If you genuinely need one more sound, make it one request and take what comes back:
Find a riser in my samples and put it on an audio track at bar 29. Pick the best one, don’t give me options.
Step 4: build the whole skeleton in one pass
This is where the deadline usually dies by hand: an hour of duplicating, muting, and dragging before you can hear whether the structure works. Dawbrain does that part in one round. Stretch the loop to length first:
Copy the loop’s clips into the Arrangement and tile them from bar 1 to bar 64.
Then carve it into sections in a single request:
Build it as: 8-bar intro with kick and hats only, 8 bars adding the chords, 16-bar full section, 8-bar breakdown with just chords, 16 bars full again, 8-bar outro thinning back to kick.
That is a complete 64-bar structure, generated and placed at once. The method behind those two prompts, subtracting instead of adding, thinking in sections, fills on the joins, is covered in turning a loop into a full song. For the deadline version, one amendment: you get one re-roll. Ask for another take of the section that bothers you most, keep the better one, move on.
What’s playing between bars 33 and 48? Which tracks are empty in the outro?
Asking beats squinting at the timeline, and the answer comes back in seconds across the whole set.
Step 5: one mix pass, timeboxed
Thirty minutes, and the scope is levels, pans, and at most one EQ move per track. That is not a real mixdown, and it does not need to be. It needs to not embarrass the arrangement.
Set the mix: pull the pads down 3 dB, pan the shaker left and the open hats right, and tell me the volume on every track.
Dawbrain reads the set and changes it, so “make the kick the loudest thing, then the bass 3 dB under it” is a request rather than a spreadsheet. Ask it to play the drop back and listen yourself, then ask for the next change. Judging it is your job; the typing is not.
If you have more than thirty minutes on another track, level-matching a reference and finding the mud in your low-mids are the two highest-value upgrades to this pass. Not this week.
Step 6: export it and let it be 70 percent
File, Export Audio/Video, Master, render. This step is yours: Dawbrain arranges and mixes with you, but exporting the file is a thing you do in Live. Give the file a name that is not “final_v2”, because there is no v3.
Then send it to the one person from step 1. This is the part that makes it real. A track only you have heard is still negotiable, and you will renegotiate it at midnight. A track someone else has heard is finished.
You will have a list of things you would fix given another week. Every producer ships with that list. It does not shrink with time, it just changes items, and the fastest way to get better at the items is to start the next track carrying what this one taught you.
What it will not decide for you
Dawbrain compresses the labor between your decisions, and on a deadline that compression is the difference between finishing and not. But it will not set the deadline, defend it against your own better taste at 1 a.m., or tell you the track is done. “Done” is not a measurement. It is a decision, and it is the one decision this whole method exists to make you practice.
Ship it at 70 percent. The missing 30 goes into the next one.
Dawbrain runs inside Ableton Live 12 and is free to start: 150 credits, no card, no expiry date. Install it, open the loop that has been eight bars for a month, and email info@dawbrain.com with what breaks.