Install a Ceiling Cap

Set a hard maximum on the resource being overdrawn and decide in advance where spare capacity goes.

Draft seven of the proposal is open on your screen. Rounds one through four made it better. Rounds five and six mostly moved commas. You keep polishing because stopping feels like a lack of ambition, and every hour you spend here is an hour not spent on the client work that actually pays you. Effort is expanding with no natural brake.

Install a Ceiling Cap is the move for exactly this: when effort, spending, dose, risk, or polish keeps growing because quitting feels lazy. Run it in three steps. First, name the resource being overdrawn, here it's revision hours. Second, set a hard maximum for one cycle: two revision rounds, and a third only if you can name a specific reader failure it will fix. Third, decide in advance where the freed capacity goes, say, to the next client deliverable, so the cap creates something instead of just guilt. What changes: the endless-improvement loop gets a wall, and your best hours land where they matter. The warning from the spec: a cap should be revisable after evidence, not a magic number you picked to dodge the real work of judging quality.