How to Organise Your Notes So You Can Find Them Again

You know you saved it. You do not know where. The note exists, and so do four copies of the document, two of them called final.

Keeping things has never been the hard part. The hard part comes later, when you need one of them and have a search box and a vague memory. This article is about that later moment. It will not recommend an app. Everything here works in any tool, including a folder of plain text files. The quotations come from my book The Librarian.

Finding is the job, not storing

A hoard you cannot search is a landfill with good intentions. One question settles most organising decisions: will you, a colleague or an AI assistant actually find this again?

Organising always costs something, and the cost lands at one of two moments. You can pay when you save, by naming, filing and tagging with care. Or you can pay when you look, by searching, scrolling, opening things and asking around. There is no version where the cost disappears. You only choose when to pay, and how much.

That cuts both ways. Careful filing of something nobody will ever look for is wasted effort. Filing is a loan from the present to the future, and some of the future never collects.

The common mistake is sorting things by what they are instead of by how they will be asked for. People arrange files the way a museum arranges exhibits: by type or by origin. Then they need something, and the question they ask has nothing to do with either. Before you build a structure, write down the questions you most often ask of your notes, and arrange things so those are easy to answer.

Names, folders and one true home

Most finding never reaches the search box. You open a folder, glance down a list of names and recognise the one you want. Or you do not, open three wrong files, sigh and search. The filename is the first and often the only description anyone reads, so it is the cheapest useful metadata you will ever write.

A good test: a filename should still make sense after it has been emailed to a stranger.

Two habits do most of the work.

Put the date first. Prefix the name with year, month and day, as in 2026-10-07. Your file list then sorts itself into date order, on every operating system and in every tool, with nothing to install. The format is an international standard, ISO 8601.

Use short, plain slugs. Name files as if a script will read them, because one eventually will. Lower case, a few meaningful words, no "new document", no "final".

If you cannot think of a name, pay attention to that. It usually means you do not yet know what the thing is, and that is a thinking problem more than a filing one.

Names settle a good deal. Structure settles most of the rest. The argument between folders and tags has run for years, and each side is half right. Folders answer where a thing lives. Tags answer what it is about. Nearly every system that lasts uses shallow folders plus a small number of tags that earn their keep.

Keep the folders shallow because every level of nesting costs twice. When you save, it is a decision: which subfolder does this belong in? When you look, it is a guess: which one did I put it in? One level is easy and two is manageable. Beyond three, people stop filing correctly and drop things wherever is nearest.

Tags need a gardener, or the weeds win. One concept, one term.

Then deal with copies. Every piece of knowledge that matters needs exactly one authoritative location. Copies, summaries and links are fine. What does the damage is doubt about which copy is true. When the pricing sheet exists in a shared drive, a slide deck, two email attachments and a wiki page, and they disagree, nobody knows what the price is and everybody acts as if they do. Link to the original where you can. Keep old versions behind the document, in its history, and not beside it as a row of near-identical files.

Capture now, file later

Noticing that something is worth keeping and deciding where it belongs are different mental activities. Noticing is fast and reactive. Deciding is slow and comparative. Do both at once and both get worse: you capture less because filing feels like effort, and you file badly because you are in a hurry.

The fix is the inbox pattern. Everything lands unsorted in one place, and a regular ritual empties it. The two halves depend on each other. Without the inbox, the friction of filing kills capture. Without the ritual, the inbox is a bin with a nicer name.

Watch the rate, too.

If you save faster than you read, you are not collecting. You are accruing.

Saved articles, bookmarks and meeting recordings that outrun your ability to review them are a debt, and it compounds quietly. Every save is a small promise to your future self. Make fewer of them.

When you do keep something, add a sentence of your own. Copying is storage and rephrasing is understanding, and your own words are also the ones you will later type into a search box.

Get better at looking

Most people search the way they knock on a door: two words, a pause, and if nobody answers they go away.

Two words and hope is not a search strategy. It is a wish with a cursor.

Search is a skill, and the tools you already use almost certainly support more than you ask of them. A modest amount of practice makes a plain search box far more precise.

Decide what kind of search you are doing. Sometimes you want everything relevant and will put up with some noise. Sometimes you want one right answer and nothing else. The two goals pull against each other, so choose before you type.

Keep a note of the searches that fail. Each one tells you something precise: either the thing is missing, or it is filed under words you would never think to use. Rename it using the words you actually tried.

And leave room for browsing. Search needs you to know what you are looking for. Scanning a well-ordered folder or a short index note lets you find the thing you could not have named.

Review, weed and archive

No system stays organised on its own. Nobody decides to make a mess. It arrives one reasonable shortcut at a time. Design is a single event and maintenance is a habit.

The habit that matters most is a fixed weekly session to empty the inbox, prune and connect. It is unglamorous and regular, like brushing your teeth, and far cheaper than the alternative.

Weeding is part of the work.

A library that never removes anything is not a library. It is a storage unit with a reading room.

Librarians remove books on a schedule, because a collection that only grows crowds out its good material and loses its readers' trust. If deleting frightens you, archive first. Move things out of the working space and out of everyday searches. Once they have sat there long enough to prove nobody needs them, deleting is easy.

For anything time-bound, such as a price, a policy or a contact, add a review-by date when you save it. This matters more now that AI assistants read our files. An assistant will take a stale note at face value and repeat it with confidence. A short index note saying what exists and where helps a human who skims and a machine that believes every word.

What to take from this

• Judge the system by what comes back out when you ask, not by how much goes in.
• Organise around the questions you actually ask, not around what the files are.
• Name files so a stranger would understand them, with the date first: year, month, day.
• Keep folders shallow and tags few. Give every important thing one true home.
• Capture into a single inbox, and empty it on a schedule.
• Save less than you can review.
• Hold a short weekly review. Archive first, delete later, and date anything that will go stale.

If you want the full version, The Librarian: 100 principles for organising knowledge covers classification, names and metadata, structure, capture, search, memory, curation and making knowledge legible to AI agents, for one person or a team, in 100 short chapters with a labelled diagram in each. It is a 314-page PDF and costs £9 on Gumroad: https://matsiems.com/books/the-librarian.