Structure is a claim, not decoration
Where you put a branch says something. Placing it under "money" claims it is fundamentally a money problem. Putting it at the first level claims it stands alongside the other major areas. Burying it three levels deep claims it is a detail of a detail.
Those are all assertions about the subject, and they are frequently wrong on the first attempt. Which is fine — a mind map is useful partly because it makes wrong assertions visible enough to fix. But it means structure deserves thought rather than being whatever the drag-and-drop produced.
The three parts every useful map has
A centre that is a question. One subject, phrased so that some things obviously belong and others obviously do not. "Career" admits everything. "How do I move into work that does not need night shifts?" excludes most of your life, which is the point.
A first level of independent areas. Three to seven branches, each able to fail for its own reasons rather than as a symptom of a sibling. This is the load-bearing layer. Get it wrong and everything below inherits the error.
Somewhere the work attaches. End branches that can hold a next action, a note, a date, evidence. Without this the map is a diagram — accurate, admired once, never reopened.
A map missing the third part is the most common failure, and it is invisible while you are making it.
Independence is the test for the first level
The most useful question when deciding whether something is a branch: could this fail on its own terms?
"Build a portfolio" can fail independently — wrong thing built, or nothing built. "Open the editor" cannot; it fails only because the portfolio branch failed. So the portfolio is a branch and opening the editor is a Todo inside it.
Apply this and first levels shrink dramatically. Most things people put at the top are tasks wearing a category's clothes.
Three to seven, and why the ceiling is real
Below three, one branch is usually concealing two. Above seven, you have stopped structuring and started listing.
The ceiling is not arbitrary. Once you cannot see all first-level branches without scanning, the map stops delivering its main advantage — noticing at a glance that two branches say the same thing in different words. That noticing is most of what a map is for.
If you genuinely have nine areas, some of them group. If they truly do not, you may have two subjects sharing a canvas, and they should be two maps.
Depth: uneven on purpose
Even depth looks tidy and is almost always waste. Real subjects are lumpy: one area is well understood and needs no subdivision, another is where all the uncertainty lives.
Go deep only where more detail would change what you do next. If subdividing a branch would not alter a decision, leave it alone — you can always expand later, and detail added early tends to get defended rather than revised.
Two levels handles most subjects. Three is common in one branch. Four almost always means something should have become its own map.
When a branch should stop being a branch
The signal is that it starts generating its own structure — sub-branches, its own todos, its own dates. At that point it is a project living inside a map about something else, and it distorts both.
In Minty Path, Copy, Move, and Link each promote a branch into its own Growth Path — Move takes it out entirely, Link leaves a shortcut behind. Splitting is the normal, healthy outcome of a branch going well, not an admission the original map was wrong.
Hierarchy is not dependency
Two different relationships get drawn the same way, and conflating them makes maps lie.
Belongs under is hierarchy: deployment is part of the launch. Depends on is sequence: deployment cannot happen until the code exists.
These are unrelated claims, and a branch frequently depends on something in a completely different part of the map. Minty Path keeps them visually separate — solid lines for parent-child, dashed relationship lines for cross-links — because a map that renders dependencies as hierarchy will push you to reorganise the tree to express an ordering, and the tree is then wrong about both things.
What does not belong on the map
- Everything you know about the subject. A map is a filter. Reference material belongs in a note attached to a branch, not as branches.
- Items already done. Unless you are keeping evidence deliberately, finished work is clutter.
- Aspirations you are not acting on. They convert the map into a guilt inventory, which is how maps stop being opened.
- Structure copied from a template you have not edited. Someone else's categories describe someone else's situation.
A structure that survives contact
The realistic test is not whether the map looks good when finished. It is whether you can open it in six weeks, understand it in ten seconds, and know what to do.
That requires: a centre still phrased as a live question, a first level you still believe in, exactly one to three branches marked as current, and end branches holding actual work rather than labels. In Minty Path, Focus enforces the third of those and the Proven-requires-Proof rule enforces the fourth — a branch cannot be closed on your say-so, so the map cannot quietly fill up with things you decided were finished.
You can build one without an account and check whether your structure survives that six-week test on something real.