‘Call about the thing before Thursday.’ When you wrote it, this was a complete thought—you knew the person, the thing, and which Thursday. The note has not changed since. What changed is you: the surrounding situation that made it obvious has been replaced by dozens of newer ones. Notes do not rot. Context evaporates.
Meaning decays in a predictable order
The first casualties are pointers: he, she, it, this, the other one. Within weeks, ‘ask him about the estimate’ has three plausible hims. Next go the time anchors—‘tomorrow’, ‘after the call’, ‘next sprint’—which become unreadable the moment the reference week is forgotten. The reasons and stakes fade last, but they fade too: the note survives while the problem it was solving becomes invisible.
The words that age worst
Certain phrasings are almost designed to expire. They feel efficient at capture time precisely because they lean on context you currently have.
- Pronouns without names: ‘she said it’s fine.’
- Relative dates: ‘tomorrow’, ‘in two weeks’, ‘after the holidays’.
- Unnamed referents: ‘the document’, ‘that option’, ‘the second version’.
- Conclusions without their question: ‘42 is fine’—fine for what?
Media notes decay fastest of all
A photo of a whiteboard, a screenshot, a forwarded message—each depends entirely on remembered context, because the item contains no statement of purpose at all. One sentence of ‘why I saved this’ multiplies a media note’s lifespan.
Insurance is cheap at capture time, expensive later
The fix is not writing formal notes—it is spending ten extra seconds on the parts that decay first: a name instead of a pronoun, a date instead of ‘tomorrow’, a half-sentence of why. And when you skip that, a timely clarification question—asked while you still remember—repairs the note at a fraction of the future cost.
Audit your own archive in twenty minutes
This is worth doing once, because the result is specific to how you write rather than to how anyone else does. Open ten notes from at least a year ago, chosen without looking at their content—oldest first, or every fiftieth item.
Sort each into one of three piles. Still fully readable: you know what it means and could act on it today. Partially readable: you get the gist but one crucial detail—who, which, when—is gone. Dead: the words are there and the meaning is not.
The proportions tell you where your capture habit actually leaks, and the answer is rarely what people expect. Most archives turn out to be roughly half partially readable, and almost all of those failures trace to two or three habits: pronouns, relative dates, or a photo saved with no sentence. Fixing the specific habit that produced your dead pile is worth more than any general resolution to write better notes.
What this implies for old notes you cannot repair
The dead pile cannot be recovered, and the correct response is to stop treating that as a failure. Those notes were cheap to make and their loss is the ordinary cost of fast capture—not evidence that you needed a stricter system.
It does change what you should do with them. Leave them in place rather than deleting: a dead note still carries a date and a fragment, and it occasionally becomes readable again when a neighboring note supplies the missing name. Deleting forecloses that. What you should not do is run a cleanup project over them, because the information required to fix them stopped existing years ago and no amount of afternoon effort brings it back.
The partially readable pile is different and worth an hour. Those notes have one missing anchor, and often you can still supply it from a nearby note or a calendar entry. Repair them when you encounter them naturally—in a search result or a review—rather than in a sweep, and append the clarification with today's date instead of editing what the note originally said.
Write for the person you will be in six months: name the people, date the times, and state the why in half a sentence. That person remembers nothing you currently know.