A blank page is hard because it asks you to do two incompatible things at once: find out what you think, and put it in order. A mind map splits those apart. Get everything down as an unordered tree, then decide the order, and only then open the document.
The method below takes about forty minutes for an article and produces an outline you can write from without stopping to think about structure again.
Why the blank page beats you
Writing is generation plus sequencing. Generation wants to be loose, associative and non-judgemental. Sequencing wants to be strict. Doing both in the same pass means every new idea gets evaluated for where it goes before you have decided what the piece is about, so you write four openings and no middle.
A map fixes the order of operations. Nothing in a mind map is "before" anything else, so you can add a thought without answering the question of where it goes. The ordering happens later, in one deliberate pass, when you actually have the material to order.
Step 1: Write the promise in the centre
Not the topic. The promise: what the reader will be able to do or decide afterwards.
- Topic: "CSV files"
- Promise: "You'll know why your export broke and how to fix it"
The centre is the thing that makes every later cut easy. A branch that doesn't serve the promise is out, and you will not agonise over it, because the criterion is written in the middle of the page.
Step 2: Dump, don't organise
Ten minutes, no editing. Add whatever you have: arguments, examples, objections, a stray sentence you like, a link. One idea per topic, in whatever branch feels close enough. Wrong placement costs nothing here: dragging a topic to a different parent takes a second, and re-parenting is the cheap operation in a mind map.
Two things worth deliberately dumping, because they're the ones people leave out and then discover halfway through the draft:
- The objection. What a well-informed reader will say back. It usually becomes a section, and always improves the piece.
- The concrete example. If a branch has no example under it, that branch is currently an assertion.
Step 3: Find the real structure
Now look at the shape rather than the content.
Fat branches are sections. A branch with six children is a section that already exists. A branch with one child is either a sentence inside another section or the thing you haven't thought about yet. Decide which.
Repeated ideas mean a missing parent. If "cost" turns up under three branches, the piece probably has a cost section and you haven't drawn it.
Anything that survives no matter where you put it is the thesis. Promote it to the top.
This is the pass where you're allowed to be ruthless, because you're deleting topics rather than sentences. Cutting a branch you never wrote costs nothing; cutting a paragraph you spent twenty minutes on is a fight you will lose.
Step 4: Order the branches, once
Drag the top-level branches into reading order. Standard orders that work, depending on the piece:
- Problem → cause → fix → verification. For anything troubleshooting-shaped.
- Answer first → why → exceptions. For anything a reader arrived at from a search query. Give the answer in the first three lines, then earn it.
- Claim → evidence → objection → response. For an argument.
- Chronology. Almost always the wrong choice unless the sequence is the point; it's the order things happened to you, not the order they matter to the reader.
Then order the children inside each branch. That is your paragraph order, and it is the last structural decision you have to make.
Drag to re-parent and re-order, then flip to the outline view, the same tree as an indented list, and export it as Markdown straight into whatever you write in.
What it doesStep 5: Export the outline and write
A mind map is a tree. A document is a tree. So the outline is not something you retype. Export the map as Markdown or a plain text outline and the branches arrive as headings with your notes underneath.
Write into that skeleton, top to bottom. When you find yourself wondering what comes next, you are done thinking and the answer is on the next line.
Adapting it to different pieces
A blog post or article. As above. Aim for three to five top-level branches; more than seven and you have two articles.
An essay with a required argument. Put the thesis in the centre instead of a promise, and make the top-level branches your claims. Any claim without an evidence child is a claim you can't yet make.
A talk. Same map, different export: one branch per slide, and present the map itself branch by branch to rehearse. What survives a rehearsal is what belongs in the deck.
A book chapter or long report. Two maps: one for the whole document at chapter level, one per chapter. Trying to hold both levels in a single map produces something you can't read at either scale.
Something you write regularly. Save the finished map as a template. Weekly posts, release notes and status reports have the same skeleton every time, and re-deriving it every week is wasted effort.
What this method won't do
It won't write good sentences, and it won't rescue a piece with nothing to say. A well-structured article about nothing is still about nothing. The dump phase is where you find out; if ten minutes produces eleven thin topics and no example, the honest move is to go and learn more rather than to start drafting.
It also has a failure mode worth naming: map polishing. Restyling branches and picking themes feels like progress and is not. The map is scaffolding. Spend forty minutes on it, export it, and let it go.