A system you keep

The test is simple: what is lost forever if you skip this move for two weeks?

Capture and title are load-bearing because they are the only two that happen at a moment which does not come back. The supplier said the thing once. If it was never captured, no amount of filing next month recovers it, and if it was captured as "Notes", the bench showed exactly what happens: a note with no searchable word and nothing pointing at it is lost in all five filings, permanently and quietly.

Filing, linking and reviewing are deferrable because the information still exists and the work merely accumulates. An inbox of forty is an afternoon. It is unpleasant, it is not a loss. The distinction matters because guilt attaches to the visible backlog, which is the cheap part, and never to the untitled capture, which is the expensive one.

So the floor is two moves and about six seconds per item. Write it down as a sentence you can read on a bad Tuesday: everything goes in one place, and everything gets a real title.

The author of the best-known method says this himself, and it is his second listed myth about his own work: the system is "adaptive", and "any incremental improvement immediately makes a difference, whether or not the other parts of the system are already in place." That is not encouragement, it is a structural claim, and the bench supports it. Going from one folder to PARA moved the lost count from four to two without a single link being added. Adding links moved reachability from 38 percent to 73 percent without changing a single folder. Each move paid on its own.

The practical consequence is that "I am behind on my system" is usually a category error. You are behind on the deferrable half. The half that cannot be caught up on is the half you did at the time, and if you captured and titled, you are not behind at all, you have a queue.

A month-old inbox item that you have not needed in a month is, empirically, an item you did not need. Archiving it costs nothing that was not already gone: it keeps its title, it stays searchable, and it sits exactly where an unread inbox item sat. The only thing you lose is the feeling that you were going to get to it.

What the digital hoarding research measures is relevant here and worth stating carefully. Neave and colleagues found "difficulty deleting" to be a real, measurable trait across two samples, correlating with how much people accumulate at work. It is not a character failing and the qualitative work refuses the pathology framing, finding a spectrum from hoarding to deliberate minimalism with most people in between. But a collection with no discard step only ever travels one way, and a person who finds discarding hard needs the discard to be a scheduled, low-stakes, reversible move rather than a judgement call made ninety times in an afternoon. That is precisely what an archive is.

Sometimes it is not a month, it is three years and eleven thousand items across five tools. The instinct is a migration project. Do not do the migration project; they fail, and they fail late, after a lot of weekends.

Do this instead. Declare the old collection an archive, entire and untouched, and leave it exactly where it is. It stays searchable, which is the only thing it was ever reliably giving you. Start a new place with the two floor moves, and let the archive be the thing you search when the new place does not have something. Within a couple of months the new place holds everything current, the archive holds the past, and you have spent no weekends. The cost is that the old collection never becomes navigable, and the bench is clear-eyed about what that means: it will stay at search-only retrieval forever. That is a real loss, and it is smaller than the migration you were considering.

Same method, so the comparison means something. Scroll to a random point, take ten notes in a row, and ask the three questions of each: could somebody search for this by its title alone, does anything link to it or sit beside it in a place small enough to scan, and was it ever going to be reread. Two yeses or more is findable.

Compare with the number you wrote down in session 1. Be careful about what the comparison proves: the notes you have captured since are the ones made under the new rules, and the old ones have mostly not been touched, so an improvement is partly a measure of how much you have added recently. That is a real limitation of measuring yourself, and it is still the most honest signal available. If the number has not moved at all, the likeliest cause is that the sample is dominated by old notes, and the second likeliest is that you have been filing rather than linking.