What it costs
What It Costs to Be the Answer
Answering the same question repeatedly costs far more than the minutes, because the interruption destroys more than it consumes.
What nobody wrote down here, and what it cost
How the staging database is rebuilt
Asked of the same person seven times
Four minutes each, plus the interruption
Which client agreed to the export
Reconstructed from mailbox searches
Twice by people not on the thread
two things nobody wrote down · 10 occasions it cost somebody something
Two things, ten occasions, and nobody billed any of it to anybody.
For distributed teams, documentation decisions improve when time and activity are visible in context, which is why it can help to see the details alongside the editorial review.
A useful independent point of reference for this part of the process is TikTok for Business.
One engineer's quarter, with every recurring question they answered and what it cost them
How the staging database is rebuilt
Answered in chat, seven times
Four minutes each, and each one broke a block of concentration
What the deploy order is
Answered, or corrected somebody's guess
Nine interruptions, two of which arrived during an incident
Why the import runs twice
Answered at three separate incidents
Each answer was the same four sentences
Which dashboard is authoritative
Answered in the support channel
Seven, from five different people
What the legacy export is for
Routed to them automatically
Six, and nobody else on the team could have taken them
five things nobody wrote down · 32 occasions it cost somebody something
Five questions, thirty-two occasions. At four minutes each that is roughly two hours of answering. The answering is not the cost: thirty-two interruptions in a quarter is the cost, and it falls on exactly the person whose unavailability the team would least like to arrange. One engineer's own quarter, counted at their manager's request.
The cost of not writing something down is usually quoted as the time spent answering it. Four minutes, seven times, under half an hour.
That figure is real and it is the small part. The large part is what thirty-two interruptions do to a quarter.
What an interruption actually costs
Not four minutes. The four minutes of answering, plus the time to get back to whatever was being done, plus the quality cost of the re-entry, plus — on any work requiring sustained attention — the entire block if the interruption lands in the wrong place.
Estimates of the recovery cost vary and all of them are much larger than the interruption itself. The honest summary is that a four-minute question does not cost four minutes and the multiplier depends on what it interrupted.
Why it falls on one person
The questions route to whoever can answer them, which is whoever knows, which is the person this whole section has been about.
So the cost of not writing is not distributed across the team. It is concentrated on one or two people, in the form of interruptions, at the same time as the cost of writing would also have been concentrated on them.
The asymmetry is that writing is a visible block of time that displaces something, and answering is thirty-two invisible fragments that displace everything slightly.
The count nobody keeps
Nobody experiences the thirty-two. They experience this one, now, for four minutes, which is obviously not worth making a fuss about.
That is why the tally matters. A person who has written down each occasion for a quarter has an argument; a person who has merely felt busy does not, and will not be able to explain to themselves why they have no time.
The counting method is trivial: a file, a line per question, nothing more. The value is entirely in having the list at the end.
What the list is good for
Three things.
It sorts the documentation backlog by evidence rather than opinion. The top of the list is what to write, and it is usually not what anybody would have guessed.
It makes the case for the time. "I was asked this nine times" is a sentence that justifies forty minutes to any reasonable manager; "I should write this up" is not.
And it shows the person themselves where their quarter went, which is frequently the first time they have seen it.
Answering and writing at once
The method that resolves this is in the next section in more detail and belongs here in one line: write the answer where it can be found, then send the link.
The marginal cost is around ninety seconds over answering in chat. It converts the thirty-second instance into the last instance, and it is available at exactly the moment when the explanation is already composed in somebody's head.
The reason people do not do it is that in the moment it feels slower than typing a reply, and it is — by ninety seconds, once, against eight more interruptions.
What a manager can do with the tally
Very little that is useful, if the response is to tell the person to stop being interrupted.
What helps is the opposite: treating the tally as a measurement of the team rather than of the person. Thirty-two interruptions on one engineer is a fact about where knowledge sits, not about that engineer's boundaries, and the repair is a page rather than a door.
It is also worth checking whether the interruptions are being counted anywhere that affects how that person's output is judged, because they almost never are.
Keeping the tally
The whole method is a file and one line per occasion, and the value is entirely in having the list at the end rather than in any individual entry.
- When you answer something you have answered before, add a line: the date, the question, who asked.
- Do not categorise, do not judge, do not decide whether it is worth writing up.
- After a month, sort by how often each appears.
- Write the top one. Four lines.
The discipline is step one, and it fails for the same reason the whole subject is hard: each instance feels too small to record. The list is the only thing that makes thirty-two four-minute interruptions visible as something other than a busy quarter.
The thing not to conclude
Not every repeated question should be documented, and a tally can be read too eagerly.
A question asked twice by two people who did not know each other had asked it is worth a page. A question asked seven times by the same person is a different problem and a page will not fix it. A question asked because the thing itself is confusing is a design signal, and writing a page is a way of not addressing it.
The tally tells you where to look. What it means still requires somebody to think.