The Bad Version Goes First
Nothing you have written well was written well the first time. That is not encouragement. It is a scheduling constraint, and you have spent years refusing to build around it by asking attempt one to do a job it has never done for anybody.
I carried a piece about supplier risk for five months. I opened the file eleven times, wrote a first paragraph, decided it was the wrong first paragraph, and closed it. At some point I started describing this to people as thinking it through.
I was not blocked in any dramatic sense. I was running quality control on something that did not exist.
Call it the draft debt. Every hour spent making the first pass good is borrowed against the second pass, where quality actually gets made, and the interest comes due as work that never starts.
The diagnostic: count the things you have been carrying for more than a month, then count how many of them exist as a file with a bad version of themselves inside it. The gap is what you owe.
What I stole
The fix came from a friend who paints. She blocks the whole canvas in one ugly color before anything else happens, and she is completely uninterested in which color it is. The block is not a draft of the painting. Its only job is that the canvas is no longer white, because a white canvas is a decision about everything at once and a blocked canvas is a decision about one thing at a time.
I had been trying to make my first paragraph good, which is a white canvas problem. The first paragraph cannot be good. There is nothing behind it yet for it to be good relative to.
What I run now
Every first pass gets sixty minutes and is written to be thrown away. No research inside the hour, no rereading, no going back to repair a sentence I can hear is wrong. The file gets named draft so I cannot mistake it for the thing.
The output is genuinely bad, and that is the specification rather than a side effect. What sixty minutes of bad produces is an object with a shape, and a shape can be cut, reordered, and argued with. A shape in your head can only be admired.
Results so far
Fourteen first passes since March. Eight have become real pieces. Three are still sitting in draft and probably will not move, which is a smaller pile than the one I used to carry. Three died at minute sixty, and two of those I had held for over a year; seeing the actual argument on a page made it obvious the argument was thin, which is a result and a cheap one at the price.
The supplier risk piece took two hours once a bad version existed. It had cost me five months as an idea.
The good version is real, and you will write it. It has always been second in line. It is standing behind an ugly draft you will not produce, waiting for something to fix, and it will keep waiting for exactly as long as you keep guarding the first sentence.