I counted, in January 2026, the number of unfinished blog drafts in my ~/writing/drafts/ folder. Seventeen. Some had been there since 2022. One of them was four sentences and a TODO that just said “finish this.” Helpful, past Marin. Thanks.
I have shipped six posts since then, which by my standards is suspicious productivity. The folder is currently down to nine drafts. I figured I would write down what changed, partly because somebody asked me at a coffee shop last week, and partly because writing it down makes me commit to keeping the habit.
The actual rule
One rule. No new draft until the current one ships.
That is it. That is the whole thing. If I have an open draft and I get an idea for a different post, I am not allowed to start a new file. I can jot the idea into a single ideas.txt file with one line. That is the cap. The new draft does not get to exist as a file with a slug and an outline and three paragraphs of an introduction until the current draft is published.
The reason this works, I think, is that my problem was never running out of ideas. My problem was that ideas are exciting and finishing a draft is not. Every time I hit the awkward middle of a piece, where the argument has stopped writing itself and I have to actually decide what I think, my brain would helpfully suggest a brand new topic. “Hey, you should write about CMS choices in 2026!” Yeah, I should, but I am supposed to be writing about freelance pricing right now, brain.
So I would open a new file. The new file would get to about 600 exciting words and then hit its own awkward middle, at which point my brain would suggest yet another topic. Repeat for four years. Folder fills up.
The rule kills this loop because the new draft cannot be opened. The exciting feeling has nowhere to go. I am stuck with the boring middle of the current piece, which means I either grind through it or I close the laptop. I cannot productively-procrastinate by writing a different post.
The supporting habits
The one rule does most of the work. But there are three smaller things I do alongside it, and I think they help, though it is hard to tell because I changed them all around the same time.
Drafting in plain text, no formatting
I write the first pass in iA Writer, which is basically just a plain text editor with a slightly nicer typeface. No bold, no italic, no headings, no nothing until the draft is structurally done.
This sounds dumb but the reason is real. When I draft in HTML or Markdown with full formatting, I spend an embarrassing amount of time adjusting the visual presentation of half-finished thoughts. I will write a heading, then re-style the heading, then realize the heading does not match the section yet because the section is not written, then noodle with the heading some more. Plain text removes the option entirely. There is just text. I have to keep writing text, because there is nothing else to fiddle with.
TK markers instead of inline research
This one I stole from journalists. If I am writing a sentence and I do not know a fact, I write TK in the spot. “Tailwind has had TK major versions in TK years.” Then I keep going. I do not open a tab. I do not check the docs. I do not fall down a thirty-minute hole reading release notes from 2022.
At the end of the draft I search the document for TK and fill them all in at once, in one focused research session that takes about twenty minutes. (TK does not appear naturally in English text, which is why journalists use it, in case you were wondering. Searching for “TBD” or “???” sometimes catches things you did not mean to flag.)
The math here is simple. Each inline research detour used to cost me twenty to forty minutes of context. Doing them all in a batch costs maybe twenty minutes total. The piece I wrote about SEO advice last month had eleven TK markers in it. Filling them in took eighteen minutes flat.
Time-boxing the first edit pass to 30 minutes
This one is the most fragile habit but the most useful when I actually do it. After I finish a draft, I close the file for at least a few hours (ideally overnight) and then come back and do exactly thirty minutes of editing. Timer set. When the timer goes off, I publish or I close the file for another day. No “just one more pass.”
Editing is where I used to lose entire weekends. There is always one more sentence that could be tighter, one more transition that could be smoother. The thirty-minute cap forces me to ship something at 92% rather than chase 99% forever. Nobody reading the post knows the difference. I know the difference, but I have made my peace with it.
The disclaimer
None of this might work for you. Writers who are paid by the word, writers with editorial deadlines, writers who outline meticulously before writing a sentence, writers who genuinely cannot start a piece without knowing the ending. Different machinery. The one-rule trick worked for me because my specific dysfunction was “too many open loops,” not “unable to start.” If your dysfunction is different, the fix is different.
Honestly, the only universal advice I would give is: actually count your unfinished drafts. The number is usually worse than you think. Mine was seventeen. I have a friend who admitted to over fifty when I asked. He looked physically pained.
Anyway. I have to go. The next draft is open. There is a TK in the second paragraph and I have not earned the right to start a new file yet.