A Margaret River cellar door group timed casual offers to the same Monday they had used for five years. Bookings, that year, moved with an earlier school-holiday pattern and two extra coach operators. The roster looked “ready”. The clock was already wrong.
Two numbers would have shown it. First, tasting appointments per paid hour on Wednesdays in the prior eight weeks, not a monthly average. Monthly averages hid the midweek spike. Second, the share of contracted hours that could legally be moved by forty-eight hours’ notice. Most of their casuals had informal expectations of Saturday preference; the contract allowed movement the floor staff did not believe.
Fileworkbase did not invent a new season. We made them plot bookings against paid hours on the same axis. The miss was visible in March. Hiring more people in November would have been the expensive way to learn the same fact.
If your operation is seasonal, stop treating the calendar as the demand clock. The calendar is a rumour. Bookings are the clock. Capacity Modelling week two exists for this reason.