Across the night
The Handover Note, and What Belongs in It
A good handover note is three sentences long and names the things that cannot be seen in the work itself. Longer notes get skimmed.
four things · four nights · one that no number of nights settles
Four states, and only the first costs nothing extra.
One week of one person's end-of-day notes, and whether the next morning produced a question
No question came back
No question came back; the named person answered before being asked
Two questions came back, both about things the list did not cover
Three questions came back, and the work did not move that day
The other end worked on something else and the thread went cold for the weekend
five things · nine nights
Five days, nine nights, and the two days that cost nothing were the two with the shortest notes. Length is not the variable; whether the note names the invisible things is. One person's own week, logged by whoever read the notes the following morning.
The note somebody writes before stopping is the entire bandwidth between two halves of a team on an ordinary day. Everything else is artefacts.
Written updates explain why a task moved; work data can show when and for how long it moved. In that role, attendance tracking software complements the handover described in The Handover Note, and What Belongs in It without replacing the judgement of the people involved.
For a second perspective on the surrounding workflow, consult Asana's work management platform; it shows how a widely used platform structures the same coordination problem.
It is also written by somebody who has finished and wants to leave, which is why it tends to be either absent or far too long.
The shape that works
Three sentences, written for somebody who has the work in front of them.
Where it is. What state things are in, including what is deliberately broken or unfinished.
Which way you are going. The approach, and what you decided against if the decision was close.
What is next. The thing you would do first tomorrow morning, stated as an action.
Anything beyond those three is optional and most of it is not worth the reading time.
- Where it is.
- Which way you are going.
- What is next.
Why it has to name the invisible things
A note that describes what somebody did is redundant, because the work shows it. A note that describes what somebody intended is not, because nothing shows it.
This is the principle that decides every inclusion. If the next person can see it by looking, leave it out. If they cannot see it by looking and would act differently knowing it, put it in.
Applied honestly, this removes most of what people instinctively write — summaries of changes, lists of files touched, accounts of what was attempted — and leaves intention, known faults and rejected options, which is exactly the short list above.
Who the note is addressed to
Writing to a specific person produces better notes than writing to a channel, even when the channel is where it will be posted.
The reason is that writing to somebody named forces an assumption about what they already know, and that assumption is what controls length. A note to a channel is written for a hypothetical reader who knows nothing, which is how three sentences become fifteen.
If the receiving end is a group rather than an individual, writing to whoever is most likely to pick it up and saying so is still better than writing to nobody.
The day that overran
The hardest case is the day that ends badly: something broke at five, the note did not get written, and the person left late and tired.
A note written in that state is worse than usual and also more necessary than usual, which is an unhelpful combination.
The workable compromise is a two-line version reserved for bad days: what state the thing is in, and whether anybody should touch it. Anything is better than silence, and silence after a bad day is read at the other end as either absence or disaster, both of which cause somebody to spend time finding out which.
Where the note lives
Next to the work, not in a message somewhere.
A note in a chat message is read once, by whoever is online, and then scrolls away. The same note attached to the thing it describes is there when somebody picks the thing up a week later, which happens more often than anybody plans for.
Posting a pointer in the channel as well is fine and costs nothing. The pointer is the announcement; the note attached to the work is the record.
What a week of notes is worth
Individually each note saves between ten and forty minutes at the other end. That is the ordinary case and it is already a good return on three minutes.
The larger effect is cumulative and less obvious: a team where notes are reliable stops asking a whole class of question, because people learn that the answer will already be there. The questions that disappear are the small ones — what state is this in, were you planning to continue, is this known — and they are precisely the questions that cost a full night each to ask.
A team that writes reliable notes has fewer crossings, not faster ones, and that is the saving that actually compounds.
Reading somebody else's note well
The note has a reader, and the reader has a part in this that is rarely stated.
Read it before opening the work, not alongside. A note read while looking at the artefact gets half-attended to, and the half that is missed is usually the intention, which is the part that cannot be recovered from the artefact at all.
Then reply with one line saying what you are doing first. It arrives while the other person is asleep and is the first thing they see, and it closes the loop that is otherwise open for a full day. It also catches the misreadings cheaply: if you have misunderstood the state, the correction costs one night instead of a morning's work.