Asking
Asking Somebody Who Can Actually Answer
Across a gap, sending a question to somebody who has to forward it costs a full extra night, so the obvious recipient and the right one are not the same.
three things · six nights
Three questions, six nights, and two of them avoidable.
Six questions from one week, sorted by whether the recipient could answer them alone
Sent to the person who owns the service
Sent to a shared channel with nobody named
Two people each assumed the other had it; nobody replied for a day
Sent to a lead who had to ask the person who owns it
The lead was the obvious recipient and not the informed one
Sent to the person who built it two years ago
They had forgotten, and the current owner was somebody else
Sent to two people, with the question split explicitly
Sent to a manager for an approval they did not have
Three crossings to find out who did, and one to get the answer
six questions · eight round trips beyond the first
Six questions, nine round trips beyond the first, and five of those nine were addressing errors. Nothing here is about anybody being unhelpful: each recipient did the sensible thing with a question they could not finish. One team's own week, logged by the person asking.
In a team sharing a room, asking the wrong person costs ten seconds. They point at somebody else and you walk over.
Teams should not confuse presence with output, especially in the situation described by Asking Somebody Who Can Actually Answer. remote workforce management helps shift the discussion toward recorded work patterns, project time and outcomes.
For a second perspective on the surrounding workflow, consult GitHub's development platform; it shows how a widely used platform structures the same coordination problem.
Across a gap it costs a night, and the person who forwards it does so having spent most of a day with your question sitting unanswered in front of them.
The two kinds of wrong recipient
Somebody who does not have the information. Usually a person who is senior to whoever does, or who used to own the thing, or who is simply the most visible member of a team.
Somebody who does not have the authority. They know the answer perfectly well and cannot give it, because the decision is not theirs. This is the more frustrating case because the reply often looks like an answer and is actually an opinion, which gets acted on and then overturned.
Both cost the same night and both are avoidable by asking a question before asking the question.
Both cost the same night, and both are avoided by asking a different question first: not who knows about this, not who is responsible for the area, but who owns the specific thing.
The question before the question
Who owns this?
Not who knows about it, not who is responsible for the area, not who is in charge. Who owns the specific thing being asked about.
Most teams can answer this in a few seconds for most things, and the cases where they cannot are the cases where the question was always going to take three crossings — which is worth knowing in advance rather than discovering.
Where the answer is genuinely unknown, asking who owns it is a better first crossing than asking the substantive question of somebody plausible, because it returns a usable answer in one pass.
Naming somebody in a shared channel
Posting to a channel without naming anybody is the most common addressing failure and the most expensive, because the failure mode is silence rather than a wrong answer.
Everybody who could answer assumes somebody else is dealing with it. Nobody is being unhelpful; each person is making a reasonable guess and the guesses are correlated. The question sits for a full day, sometimes two, and when it is finally answered the asker has usually worked around it badly.
A name in the message fixes this entirely. If several people could answer, say that explicitly — "anybody who has touched the export job this month" — which is a different and also workable instruction.
Asking two people at once
Legitimate, and it needs to be done deliberately rather than by copying people in.
Splitting the question works: "Two things — A, for you; B, for her." Each person knows which part is theirs and both answer in one pass.
Sending the whole question to two people does not work. It produces either two answers that have to be reconciled, which costs a crossing, or nothing, for the same reason as the unaddressed channel message.
The person who used to own it
A specific trap on teams that have been running a while.
Somebody built the thing, left the area, and remains in everybody's head as the person to ask. They answer from memory, their memory is two years stale, and the answer is confidently wrong in a way that takes a day to discover.
The fix is institutional rather than personal: ownership written down somewhere people check, and kept current when it changes. Where that exists, this failure disappears. Where it does not, no amount of care by individuals prevents it.
Making yourself askable
The reverse side, and it is cheaper than it looks.
Anybody who owns something can reduce the number of wrong-recipient questions aimed at them and at others by saying, once, in a place people will find, what they own and what they do not. Two lines.
The same applies to authority. Somebody who can approve things up to a limit and not beyond it should say what the limit is, because the alternative is that people either ask them about everything or ask them about nothing.
When the right person is asleep in the wrong direction
Occasionally the correct recipient is in your own half and awake, and the question still gets sent across the gap because the asker is following a habit about who gets asked things.
This is worth noticing because it is the one case where the whole cost is self-inflicted. Before sending anything across the gap, it is worth four seconds to check whether anybody awake can answer it, which on most teams is true more often than people assume.