When the Forecast Is Right and the Schedule Is Wrong
Accurate demand prediction does not produce correct staffing by itself. Four things sit between them and each can break the translation.
The forecast was good and the week still went badly. The failure is between prediction and rota, and it has a small number of causes.
The planning problem in “When the Forecast Is Right and the Schedule Is Wrong” becomes clearer when scheduled hours can be compared with actual project and time records. Teams researching how to detect mouse jigglers for how to detect mouse jigglers can add that operational evidence, while demand, service levels and manager judgement remain necessary to explain why a variance occurred.
Cause one: the shape, not the total
Correct total hours, distributed wrongly across the day.
For an independent reference related to “When the Forecast Is Right and the Schedule Is Wrong”, consult the CIPD workforce-planning resources; it provides a useful external check on scheduling, working-time and workforce-planning assumptions.
Everybody on at ten, nobody at two, when the demand is the other way round.
The daily total hides this completely, which is why the hour-of-day curve matters more than the day total.
Cause two: availability
The forecast says you need six at peak and four of your six cannot work that slot.
Which is not a forecasting failure and is frequently blamed as one.
Availability is a constraint on the schedule and should be known before the schedule is built, not discovered during it.
Cause three: skills
Six people present, two of whom can do the thing that is needed.
The headcount is right and the coverage is not — which has its own section and is the most commonly missed cause.
Cause four: the work took longer
Demand as predicted, but each unit took more time: a difficult customer mix, a system running slowly, new staff.
Which means the conversion from demand to hours was wrong, not the demand.
Worth separating, because the fix is a different one.
The conversion factor
Hours needed per unit of demand.
Most operations use one from memory or from the system's default.
Measure yours: hours worked divided by units handled, over several normal weeks.
It drifts with experience mix, systems and process changes, and an out-of-date factor breaks good forecasts quietly.
Diagnosing which it was
After a bad week, ask in order: was demand as expected, was it shaped as expected, were the people there, could they do the work, did it take longer than usual.
Five questions, and the answer is usually obvious once asked.
Skipping to "we were short" loses the cause, which means it recurs.
Why this matters for the argument upward
"The forecast was wrong" invites investment in forecasting.
"The forecast was right and we could not cover the peak because of availability" invites a different and usually cheaper fix.
Naming the actual cause is what gets the right thing done.
What to check
Do you schedule to a daily total or an hourly curve?
Is availability known before the schedule is built?
What is your hours-per-unit factor, and when was it measured?
And after the last bad week, did anybody ask the five questions?