Across the night
What Not to Hand Over at All
Some work costs more to hand over than to hold, and the signal is always the same: it has crossed more than twice without moving.
four things · four nights · one that no number of nights settles
Four states, and only the first costs nothing extra.
Four things one team stopped handing over, and what happened to each instead
Held by one person for a day and a half; it had crossed four times and gone nowhere
Taken off the relay entirely and settled in a scheduled call
Kept with whoever started it, including the parts outside their hours
Still handed over, because it was genuinely a parcel
four things · one night · three that no number of nights settles
Four things, one night, and three that came off the relay. Each of the three had the same symptom first: it had crossed more than twice and was no further forward. The fourth stayed because it was finished work with a recipient, which is what a relay is for. One team's own decisions, recorded as they were made.
A relay is a mechanism for moving parcels. It works on things with a definite state, a definite recipient and a definite next action.
A reliable handover needs more than a message saying that something is in progress. workforce analytics software can add verifiable hours and task context to the process examined in What Not to Hand Over at All, reducing the need for another round trip.
For broader context, Miro's visual workspace offers an official example of the tools teams use around this kind of process.
A good deal of what people try to put on it is not a parcel, and the relay does not refuse it. It accepts it, moves it slowly, and returns it in roughly the state it left.
The signal
Something has crossed the gap more than twice and is not further forward than it was.
That is the whole diagnostic, and it is reliable. Work that is genuinely progressing by handover shows progress on each pass: the state is different, something is settled that was not. Work that is being passed rather than handed over looks the same on each arrival, with a slightly longer note attached.
Two crossings is normal. Three without movement is a question worth asking. Four is an answer.
Things that are not parcels
Investigations that have not been framed. Somebody is trying to work out why something is happening and has not yet got to a hypothesis. This state is almost pure mental model, and the mental model does not survive the night. Each person who picks it up starts over from the evidence, and four people doing that in sequence is not four times the progress. It is the same first hour, four times.
Disagreements. Two people who hold different views about an approach will not converge by correspondence across a fifteen-hour gap. They will produce two carefully written positions, and then two carefully written responses, and the positions will be further apart than they started because writing hardens a view.
Conversations with somebody outside the team. A client, a supplier, a person in another department. Handing these over means the external party has to re-explain themselves, which they experience as the team not listening, and the cost falls on somebody who did not sign up for the arrangement.
Anything where the value is in judgement being applied continuously rather than at a point: watching something, holding a situation, deciding when to intervene.
- Investigations that have not reached a hypothesis.
- Disagreements, which harden rather than converge in writing.
- Conversations with somebody outside the team.
- Anything needing continuous judgement rather than judgement at a point.
What to do instead
For investigations: one person holds it until it has a hypothesis, even if that means working past the handover or picking it up again the next day. A hypothesis is a parcel. The investigation that produced it is not.
For disagreements: take it off the relay and give it an hour where both people are awake, paying whatever that costs in somebody's evening. An hour bought this way is expensive and is cheaper than two weeks of correspondence that does not converge.
For external conversations: keep them with whoever started them, including the inconvenient parts, and make that an explicit expectation rather than something people discover.
For continuous judgement: accept that it is single-threaded, assign it to one person, and do not pretend the other half is covering it.
The cost of holding
None of these is free. Holding something means somebody works outside the shape of their day, or the work waits, or one half is idle on something they were nominally sharing.
That cost is real and should be paid deliberately rather than discovered. The comparison to make is not "holding costs something versus handing over costs nothing", but "holding costs one person a late evening versus handing over costs four nights and arrives in the same place".
Stated that way the decision is usually easy, which is why stating it is worth doing out loud.
Deciding early rather than late
The expensive version of this is discovering after four crossings that something should never have been on the relay. By then a week has gone.
The cheap version is asking at the start: is this a parcel? Does it have a state, a recipient and a next action? Most work does, and the question takes four seconds. The work that does not is recognisable in advance more often than people expect — investigations and disagreements both announce themselves — and catching even half of them at the start is worth several days a month.
The uncomfortable part
Taking something off the relay usually means one half of the team is not involved in it, and that is often read as exclusion rather than as arithmetic.
It helps to say which it is. "This is staying with you because it has crossed three times and stopped moving, not because anybody has decided it is yours" is a sentence worth writing, and it costs nothing except the willingness to say the real reason out loud.