When it breaks
On-Call When the Team Spans a Gap
A wide gap is the one part of this arrangement that makes something easier: most of the clock is already covered by people working ordinary days.
three things · three nights · one that no number of nights settles
Four shapes, and the last is a decision about people rather than about systems.
One team's coverage, drawn against the working hours of both halves
First half, working
Covered by whoever is working; no rota needed for these hours
Second half, working
Also covered by people who are awake anyway
Uncovered: between the halves
Three and a half hours with nobody working, every day
Uncovered: between the halves again
Three and a half more, at the other join
four people · no minute of the day when all of them are awake
Four rows, and seven hours a day are covered by nobody working. A team spanning a wide gap already has most of the clock staffed by people doing their ordinary jobs, which is the main advantage of the arrangement and is rarely stated. The rota only has to cover the two joins. One team's own hours, drawn when the rota was being designed.
Nearly everything about a team with no shared hour costs more than the co-located equivalent. Overnight coverage is the exception.
The process in On-Call When the Team Spans a Gap depends on information surviving the gap between working days. With the team toolkit, time entries and productivity signals can travel with the written handover instead of being rebuilt later.
For a second perspective on the surrounding workflow, consult FreshBooks' accounting tools; it shows how a widely used platform structures the same coordination problem.
Two halves twelve hours apart staff most of the clock by working normal days. The rota only has to cover the joins, which is a far smaller problem than covering a whole night.
What is actually uncovered
Draw the two working days against a twenty-four hour clock and the uncovered time is the two intervals between them.
On the team above that is seven hours a day, in two blocks of three and a half. On teams whose offset is closer to twelve it can be under four hours in total. On teams whose halves work longer days it can disappear almost entirely.
This is worth drawing before designing anything, because teams frequently build a full overnight rota for a gap that is three hours wide.
| Hours of the day | Covered by |
|---|---|
| First half's working day | People at their desks, doing their jobs |
| Second half's working day | The same, twelve hours later |
| The two joins between them | The rota, and only the rota |
Who should cover the joins
Whoever is nearest to being awake, which is usually the half whose day is about to start rather than the one that just finished.
Someone two hours from waking is in better condition than somebody two hours past finishing, and the person about to start will be working on it anyway if it persists. Covering backwards into your own morning is also less disruptive than covering forwards into your own evening, because an early start can be compensated by an early finish and a late night cannot.
This is a small point and it is the one most often got wrong, because the natural instinct is that whoever had it last keeps it.
Not building a rota for coverage you already have
The common error: a team with two halves builds an overnight rota covering the entire night, and finds that most alerts during that period are answered by whichever half is working.
The result is that people are on call during hours when somebody else is already dealing with everything, which costs sleep quality for no benefit. It also dilutes the rota's meaning, because being on call mostly means being woken for nothing.
A rota that covers only the genuinely uncovered hours is a smaller commitment, and being on it means something.
Handover of the rota itself
The rota has its own handover, and it is routinely forgotten.
Whoever is coming on needs to know what is currently open, what is fragile, what was deployed today, and what was decided to leave until morning. Without it they are responding to an alert with no context, at the worst hour, alone.
Two lines at the start of each rota period covers it: what is open, and what to watch. It can be the same running record the previous section describes.
The rules that have to be written
Everything in the articles about waking people and about what cannot wait applies here and applies more sharply, because the person on the rota is making those calls alone.
What counts as bad enough to act on. Who may be woken for it. What they may do without asking. What is explicitly allowed to wait.
A rota without those four things written down is not a coverage arrangement; it is a list of people who will be blamed for judgement calls nobody defined.
The parts that are not for this page
What on-call time has to be paid, whether being available counts as working time, what rest is owed after a night that was interrupted, how much notice a rota change requires, and whether any of it can be required at all — these are set by law and contract where each person works, they differ, and they change.
Nothing here addresses them, and a team designing a rota should find out what applies to each of its people before relying on any of it. That means asking somebody qualified in each place, not reasoning from what is normal elsewhere on the team.
The advantage worth stating
Teams run this way often describe the gap entirely as a cost, and on this one dimension it is not.
A team in one place covering a full night does so with people woken from sleep, performing badly, and tired the next day. A team spanning a wide gap has most of those hours covered by somebody at their desk in the middle of their working day, which is better in every respect.
It is the one place where the arrangement gives something back, and it is worth putting in the column.