Guess What The Page You Want Would Actually Say

The strongest move in searching is to imagine the document before you go looking for it.

This is the single habit that most separates experienced finders from everybody else, and it can be described in one line: do not describe your question, describe the answer. Picture the page that would satisfy you. Who wrote it, and why did they bother? Is it a formal decision, a guidance note, a set of minutes, a datasheet, an annual report, a set of instructions for staff, a summary written for the public? Each of those genres has its own vocabulary and its own furniture, and once you have picked one you can guess several phrases that would almost certainly appear in it.

Those guessed phrases are what you type. Not your question, which will never appear in the document, but a fragment of the document's own language: the sort of stiff formal phrase that only appears in that genre, or the heading that such a page always carries, or the pair of technical terms that would sit next to each other only in the material you want. You are essentially trying to fingerprint the target. A good fingerprint is specific enough that almost nothing else could produce it, and generic enough that every document of that kind would contain it.

The habit has a second benefit that is easy to miss. Sometimes you sit down to imagine the page and realise that no such page exists, because nobody would have had a reason to write it, or because the thing you want is never recorded in one place and can only be assembled from several. That is enormously valuable to discover in ten seconds of thinking rather than an hour of fruitless typing, and it redirects you to the right activity, which is now assembly rather than retrieval.