
The first half hour in a new note app goes almost the same way every time: a folder tree gets built, then that tree is found to be wrong, then it gets built again. The tree was never the problem. The problem is that nobody asked which question the structure was meant to answer.
The folder’s one rule, and its one flaw
A folder puts a note in exactly one place, and that is both its strength and its ceiling. The strength: every note has a home, the answer to “where does this go” is single, and that answer still holds six months later. The ceiling shows up with the first real note.
Say a new pricing model came up in a client meeting. Does the note go under Clients/Acme, under Pricing, or under 2026 Meetings? All three are defensible. Once you decide, whoever looks for it down the other two paths will not find it, and that person is most likely you.
The tag’s one rule, and its one flaw
A tag can put a note in several places at once, which answers the problem above directly. The same note can carry acme, pricing and meeting together.
The cost shows up a year later: you are holding acme, Acme, acme-corp and client-acme at the same time, none of them covers all the notes, and the tag list itself has grown past the point of being browsable. Tags are cheap. Being cheap makes them multiply, and multiplying makes them worthless.
The real question is how you start looking
Instead of choosing between the two, ask this: six months from now, what is the first thing that will come to mind when you go looking for that note?
If the answer is “which project was it”, folders are the right tool, because projects start and finish and a finished project’s folder moves to an archive and leaves your field of view. If the answer is “which topic was it”, tags win. Topics cut across projects and never end. If the answer is “it had this sentence in it”, neither is needed; full-text search will find it.
That third possibility is what changed most over the last decade. Once search works well, the depth of a folder tree stops being a way of finding things and becomes only a way of keeping order.
A small rule set that holds up
Five rules, all of them checkable.
Keep the folder tree to two levels. By the third level, filing a note takes longer than searching for it.
Folders show lifecycle. Active, Archive, Reference. Topic names do not belong in a folder.
Tags use one spelling. Lower case, hyphenated, singular. pricing, not Pricings.
Look at the list before opening a new tag. If you are landing near an existing tag then use that one, because opening a new tag costs five seconds while merging two tags afterwards and reviewing every note costs half an hour.
Read the tag list once a month. A tag used on a single note either gets deleted or gets folded into another. That maintenance takes ten minutes, and skipping it lets the system rot quietly.
What this looks like in Context
Because notes, tasks and files live in one workspace, a tag can cut across all three: acme brings back both the meeting note and the task attached to it. Turkish search normalises dotted and dotless letters against each other too, so typing fiyatlandirma still finds the note tagged fiyatlandırma.
The tool still will not make the decision for you. That is what the five rules are for.
