The Shift Pattern Built for Demand That Has Gone
How to tell whether a pattern still matches the work, the simplest comparison that shows it, and why demand shifts are almost never followed by pattern changes.
What one hourly comparison changed at a single site
Two posts moved five hours later in the day and the peak stopped needing overtime, at the cost of five people losing a start time they liked. That cost was real and was raised at consultation. This is one employer's own result, not a benchmark.
Shift patterns are built against a picture of demand, and demand moves. The picture does not, because nothing in the normal running of a service compares the two.
The rule discussed in “The Shift Pattern Built for Demand That Has Gone” is easier to test when the published schedule can be compared with a consistent record of actual hours. A team reviewing open the official workflow overview for online timesheets for shift teams can add current attendance and project context while keeping contractual rules and manager decisions explicit.
The result is familiar everywhere: a pattern with people on at the quietest hours and too few at the busiest, defended by everybody, chosen by nobody currently present.
The assumptions behind “The Shift Pattern Built for Demand That Has Gone” can also be checked against Skills for Care workforce-planning guidance. Record what applies locally and what requires qualified advice.
The comparison that shows it
Two lines on one chart: demand by hour of day, and staffed hours by hour of day, for a typical month. Any spreadsheet does it in an hour.
That chart is the single most useful artefact in rota planning and almost nobody produces it, because demand lives in an operational system and staffing lives in a rota system and they have never been put on the same axis.
What counts as demand
Whatever the service exists to handle: calls, admissions, orders, presentations, jobs, transactions. Something countable, by hour, over at least a month.
Perfect data is not needed. A rough hourly profile is enough to show a mismatch of the size that matters, and arguments about the exact definition usually come from people who do not want the chart.
What the chart usually shows
Three things. A peak that is understaffed by one or two people. A trough staffed at the same level as the peak because the pattern has no middle shift. And an edge — the first or last hour of the day — where the staffing and the work disagree completely.
- A peak understaffed by one or two people.
- A trough staffed at peak level because the pattern has no middle shift.
- An edge — the first or last hour — where staffing and work disagree outright.
The edge is usually the cheapest to fix and the one nobody has looked at.
Why patterns do not follow demand
Because changing a pattern is a consultation and demand changes continuously. The threshold for opening the conversation is high, so it is opened rarely, and between openings the mismatch accumulates.
That is a structural reason rather than a failure, and it argues for scheduled reviews rather than for responsiveness.
The review cycle that works
Once a year, the chart is produced and looked at, whether or not anything seems wrong. Most years nothing happens. Every few years the mismatch is large enough to act on, and it is caught while it is still small enough to fix without a crisis.
A pattern that has never been reviewed against demand is not stable; it is unexamined, and the difference shows up during the next surge.
Seasonal demand and the pattern
Where demand is seasonal, a single pattern cannot fit both halves of the year, and the usual arrangement — one pattern, overtime in the busy season — is the most expensive available.
Two patterns, or a pattern with a seasonal variant, is normal in some industries and unthinkable in others for no reason anybody can give. Whether it is possible depends on contracts and agreement, and is a question for somebody qualified where you are.
The shift that exists for one task
A frequent finding: a whole shift exists because of a task that used to take four hours and now takes twenty minutes, or that moved to another team entirely.
Nobody notices, because the shift is full — people fill the time with useful work. The chart finds it, because the demand that justified the shift is no longer in the data.
What to do with the finding
Not an immediate pattern change. Start with the cheapest adjustments that move staffing towards demand: start times, break timing, the allocation of a single post, the hours of a part-time role.
Those changes are small enough to make without consultation in most places, and they frequently close half the gap. The remaining half is the case for a bigger conversation, and it is now evidenced.
Three lines a year
The date the comparison was last made, the three largest mismatches it showed, and what was done about each.
Three lines a year. After five years it is a history of how the service has moved, which is a document no workforce team currently has and every one of them would use.
The data everybody already has
Most services are already counting the thing they would need: a call system, a patient administration system, a job dispatch record, a till. The hourly profile is one export away and nobody has ever asked for it by hour.
Asking for it by hour rather than by day is the whole trick. Daily totals hide every mismatch that matters, because the pattern is a decision about hours and the report is a decision about days.
Why this is the most important page in this section
Because every other inherited constraint is small by comparison. A pattern that is an hour out of step with its demand costs more than all the keys, rooms and assumptions in the register put together, and it is invisible until somebody draws two lines on a chart.