Staffing a ski restaurant: planning a season the weather keeps rewriting

Access is the first difference, and it is underrated. Staff arrive by lift. Someone who misses the 07h50 gondola is not ten minutes late, they are twenty-five minutes late at best. Extending a shift into the evening means arranging the way down, which is settled before the person leaves the valley, not at 21h with the terrace still full. A town restaurant can call someone at 17h and have them behind the bar at 18h. On a mountain the useful call happens in the morning or not at all.

The second difference is that the replacement pool is closed. In a resort it is the people already on the mountain or in staff housing, and most managers can name all of them in ten seconds. No agency, no local student available for one service.

A bluebird Saturday and a whiteout Saturday can differ by a factor of three. Sun fills a terrace that holds half the covers. Wind empties it in twenty minutes. Fresh snow keeps skiers out until 15h and then delivers everybody at once.

The practical answer is a week built in two layers: a base sized on the quiet version of the day, and a flexible margin decided 48 to 72 hours out when the forecast is finally worth trusting.

The second layer only works if the people in it are treated properly. Standby is a real constraint on a day off, so it needs to be named as standby and to get an answer at a fixed time. "We will let you know Friday", said three times and delivered late twice, is a reliable way to lose seasonal staff before March.

Christmas to 2 January, the February school holidays as they roll across cantons and neighbouring countries, the sunny March weekends. Mid-January is dead. Staffing evenly overstaffs eight weeks and understaffs the four that matter.

The result is an hours balance that swings hard: 55 hours in a peak week, 25 in mid-January. That is normal in a seasonal operation and harmless as long as the balance is visible to both sides while it accumulates.

Usually it is not. It lives in the manager’s spreadsheet, staff keep their own count in their head, and the two are compared for the first time in April, four days before everyone leaves the country. Confirming hours weekly turns the end-of-season settlement into arithmetic rather than a negotiation over memory.

The question at 16h is not "who can come in". It is three at once: who is qualified for that post, who is not already working, and who is not already deep into overtime. The honest answer is four or five names, so the bottleneck is not the size of the pool. It is that every request travels through one person’s phone, one message at a time, while that person is also doing mise en place.

This is where offering a shift beats asking for one. If a cook who cannot make Saturday can put that shift in front of everyone qualified at once, and the first acceptance takes it, the problem resolves in fifteen minutes and the manager confirms rather than orchestrates. The same mechanism covers the sunny Thursday when an extra pair of hands would obviously pay for itself.