Stop searching, start finding: faster file search habits

Most people search their computer the way they searched it fifteen years ago — open a file manager, click through folders, scan for something familiar. Meanwhile the search tools on both major platforms have quietly become very good, and almost nobody uses them properly.

Search from the keyboard, always

The single habit that changes everything: never navigate to a file. Hit the launcher, type three or four characters of the name, press enter. This is faster than folder navigation in essentially every case, and it removes the requirement to remember where you put things — which is the part of file management people are genuinely bad at.

Operators worth knowing

• Filter by kind — restricting to images, PDFs or documents cuts most noise instantly. On both platforms this is a keyword you type, not a menu you open.
• Filter by date — "modified this week" is the highest-value filter there is, because the file you want is usually one you touched recently.
• Exact phrase — quotation marks. Turns a vague match into a precise one.
• Exclude a term — a minus sign before a word. Essential when one project name appears in two hundred files.
• Search contents, not names — both platforms index inside documents. Searching for a phrase you remember writing beats searching for a filename you do not.

Naming so that search works

Search quality depends on what you gave it to find. Three conventions do most of the work.

Date first, in a sortable format. Year-month-day at the start of a filename means chronological order is also alphabetical order, in every application, forever. This costs nothing and pays out permanently.

Include the identifying word you will actually recall. Not the internal project code — the client's name, or the thing the document is about. You will search for what you remember, not for what was correct.

Avoid "final". There will be a final-final and a final-v3. Use a date or a version number instead, and let sorting do the work.

Where the built-in tools stop

Two limits are worth knowing. Indexing usually excludes external drives and network shares unless you add them explicitly, which is why searching a backup drive often returns nothing. And scanned documents are images until something runs text recognition over them, so a folder of scanned receipts is invisible to content search until you fix that.

Both are one-time configuration jobs. Add your external volumes to the index, and turn on text recognition for anything you scan. After that, the honest answer to "how do I organise my files?" is that you mostly do not need to — you need to name them predictably and let the index do the rest.

One folder structure that survives

If you want any structure at all, keep it shallow: a small number of top-level folders by area of life or work, one level of project folders inside, and nothing deeper. Deep trees fail because filing something requires a decision about which branch it belongs to, and ambiguous items end up wherever you were when you saved them. Shallow and searchable beats deep and correct, every time.