Asking at the Wrong End of Their Day

The two windows

For a counterparty behind you, their working day begins in your afternoon. Anything sent before then is waiting for them; anything sent after their day ends has missed it entirely.

For a counterparty ahead of you, their working day is largely over by your mid-afternoon. A question sent at four in the afternoon here reaches somebody who left an hour ago, and a question sent at eight in the morning reaches them mid-morning with most of their day available.

So the two populations want opposite habits: questions westward go out late in your day, questions eastward go out first thing.

The habit this collides with

Most people write the difficult messages when they have a clear hour, and the clear hour is usually either first thing or at the end of the day.

End-of-day is the worse of the two, because it is when the day's loose ends get tidied, and the loose ends are precisely the questions that have been waiting to be asked. Sent at six in the evening, they reach an office three hours ahead nine hours after everybody left.

The sorting that fixes it

Keep two lists rather than one: things going east, and things going west.

The eastward list is cleared first thing in the morning. The westward list is cleared at the end of the day. Neither requires anybody to change when they do the thinking, only when they press send, and most mail clients will hold a message until a chosen time.

Scheduled sending is not a trick

It is worth being direct about this, because people sometimes feel that holding a message until morning is a small deception.

It is not. Nobody is being misled about when the question was thought of, and the alternative — a message arriving at two in the morning local time — has a real cost to the recipient, who will see it on a phone, feel obliged to consider it, and be unable to act on it.

The only case where the timestamp carries information is where it evidences a deadline, and in those cases the message should say so explicitly rather than rely on being read.

The end of their day is the expensive end

One specific case deserves naming, because it recurs.

A request for something small — a confirmation, a number, a yes — arriving in the last half hour of somebody's day is very likely to be left. Not because the person is unwilling, but because it requires checking something, and the checking requires a colleague who has already gone.

The same request arriving at their mid-morning is answered in four minutes. The difference is a day, and it costs nothing to avoid.

When you cannot wait for the right hour

Then say what the constraint is, so the recipient can judge it against whatever else is in front of them.

I know this has reached you at the end of your day. If it is easier, the only part I need before tomorrow is whether the width is confirmed; the rest can wait until next week.

Splitting the urgent fragment out of a larger request is the single most effective thing available at that moment. It gives somebody who is on their way out a thirty-second action rather than a thirty-minute one, and thirty-second actions get done.

The week has the same shape as the day

The same logic applies at the scale of the week, and the stakes are larger because the unit is days rather than hours.

A question sent on the first two days of somebody's working week has their whole week in front of it. The same question sent on their last day has one day, and if it is not answered that day it has their weekend and then your weekend.

For a counterparty whose week runs Sunday to Thursday, that means anything substantial should leave here on a Monday or a Tuesday. By Wednesday it is landing on their fourth day; by Thursday it is landing on their last.

Nobody plans the week this way, because the natural rhythm here is to spend Monday organising and send things out from Tuesday onwards. For anything going to an office whose week ends on Thursday, that rhythm costs a day at exactly the point where it is most expensive.