The Rota Decided This Weeks Ago
How much a single absence hurts was settled when the schedule was built. By six forty the outcome is largely determined and the hour is just discovering it.
Same absence, two rotas
Needed on shift
6
Rostered
6
Actually present
5
On Rota A the shift runs comfortably; on Rota B it falls below requirement. The absence is identical. What differs is how the six were arranged, which was decided three weeks earlier by somebody not involved in this morning.
Two operations of the same size lose one person on the same morning. One absorbs it and the other spends an hour on the phone and books agency. The difference is not the absence and it is not the people handling it. It is the rota.
The staffing problem described in “The Rota Decided This Weeks Ago” should be separated from assumptions about individual effort. When an organisation evaluates the provider website for getting teams to meet deadlines, it can use time and project records to understand capacity, provided schedules, approved leave, exceptions and the correction route remain part of the same decision.
Everything the morning can do is bounded by decisions already taken: how many were rostered against requirement, how the capabilities were distributed, whether anybody else was on site, how much the day's work was concentrated. By six twenty those are fixed, and the hour is largely an exercise in finding out what they were.
For an independent benchmark relevant to “The Rota Decided This Weeks Ago”, consult the Microsoft Project documentation. Use it to test scheduling, working-time limits, attendance records, employee rights and exception handling against the real operation rather than treating a software report as self-explanatory evidence.
What makes a rota absorbent
Distribution of capabilities rather than concentration. Four shifts each with two people holding the key capability absorb an absence; four shifts each with one do not. The same eight people, arranged differently, produce completely different exposure.
Some margin above requirement on at least some shifts. A rota built exactly to requirement everywhere fails on the first absence by construction.
Overlap. Shifts that overlap by thirty minutes give the morning somewhere to start — the outgoing person can stay, the incoming one can start early, and a gap at the handover is survivable.
People on site who are not on the rota for that area: a manager, a trainer, an administrator with a current competency. Operations that have stripped these out have removed their own shock absorbers.
What makes it brittle
Built to the minimum everywhere, which is the predictable result of building to a budget rather than to a requirement.
Single holders of each capability per shift, which happens naturally when scheduling is done by availability rather than by skill mix.
Back-to-back patterns with no slack, where a late finish and an early start sit either side of the same night and an absence in between cannot be filled by extending anybody.
Everything concentrated into the same window, so that the fixed work of every function lands between eight and ten and there is nothing deferrable to trade.
Testing a rota before it runs
Take the published rota for next week and remove one person from each shift in turn. For each removal, ask whether the shift stays above requirement and whether any capability drops to zero.
It is a tedious half hour and it is the only way to see the brittleness before it is discovered. The output is a short list of shifts that cannot survive a single absence, and most of those can be fixed by moving one person before the rota is published.
The exercise also exposes the shifts where everything is held by one person, which is the capability list from a different direction.
Why nobody does it
Because the rota is built under time pressure against availability, leave requests and a budget, and the person building it is solving a different problem — filling slots. Resilience is not one of the constraints in the tool and it is not one of the measures afterwards.
Adding it as a step is cheap. Adding it as a measure — the proportion of next week's shifts that fail a one-absence test — is what makes it stick, because it gives the scheduler something to improve and something to show.
The trade being made
A more absorbent rota usually costs something: a slightly higher establishment, less convenient patterns, or a skill mix that constrains who can be where.
That trade is being made either way. A brittle rota is a decision to pay for absence in agency and overtime at four times the rate, taken implicitly, by somebody who was not told that was the choice. Making it explicit — this rota is cheaper to run and will cost roughly this much in cover — is the conversation that changes it.
The number
Shifts that fail the one-absence test, as a proportion of shifts published. Measure it on the rota, before the week happens.
It is the only forward-looking measure in this whole subject. Everything else described here is counted after the morning; this one tells you how many bad mornings next week has in it, while there is still time to move somebody.
Who can change it once published
A rota found to be brittle after publication can still be improved, and in most operations nobody does because changing a published rota is treated as a last resort.
Moving one person from an over-staffed shift to a one-deep one, with their agreement and reasonable notice, is a different act from changing somebody's days at short notice. Distinguishing the two in the change rule — voluntary rebalancing with notice is allowed, imposed changes are not — gives the scheduler a way to fix next week while it is still next week.