One forecast runs the day

Six companies under one logo

One GM was excellent and another was not. One kitchen ran beautifully while another lost thousands of dollars of food a month, and nobody could say why. One location staffed well and another bled labour.

The owner thought the answer was stronger managers. That was partly true. The bigger problem was that each location had invented its own version of the business. One counted stock every Sunday night, another when things looked low. One chef over-ordered, another under-ordered.

The brand looked like one company. It ran like six.

One forecast per location, per daypart

Managers planned on feel. Fridays are usually busy.

Now each location has a forecast for every lunch and dinner, built from its own sales history, the day of the week, the season, holidays and promotions. It does not need to be right to the cent. It needs to say whether tonight is a $6,000 dinner or a $12,000 one, because those are run very differently.

Bookings and events feed it

Reservations sat in the book and the forecast never saw them.

Now the night's bookings, local events and the weather where it matters move the forecast as they change. A big booking or a game in town shows up before the doors open, not during the rush.

Prep from the forecast

Prep cooks used to hear: make a lot of fries.

Now prep targets come from expected covers, the usual item mix and what is already prepared. Too little prep hurts service. Too much becomes tomorrow's waste.

The schedule from the forecast

Schedules were built by habit. One person always works Tuesday, another wants Friday off, someone quit. Labour ran at 27% of revenue at one location and 34% at another.

Now projected covers set how many servers, cooks, bartenders and hosts a shift needs. Managers can still override, but the override shows. The question changes from why labour was bad to why 112 hours were scheduled against a forecast that needed 88.

Cut calls made during the shift

Labour problems used to surface at payroll, when nothing could be done.

Now live sales are read against the forecast during service. When a quiet night runs under, the manager sees it in time to send someone home. Staff clocked in but not scheduled, or close to overtime, show up the same week.

Orders from par and forecast

The chef texted one supplier, the GM emailed another, and someone ran to the cash and carry when onions ran low.

Now each location gets a suggested order from the forecast, its par levels, stock on hand and deliveries due. A person reviews it. Emergency buys are still allowed, and each one is logged, because it points at a bad forecast, a bad par or a supplier problem.

Waste logged by cause

Food went in the bin and food cost went up, and nobody knew which problem caused it.

Now actual usage is checked against what the recipes say was sold, and waste is logged as spoilage, overproduction, a guest return or a kitchen error. High spoilage points at ordering. High overproduction points back at the forecast.

One daily view per GM

Each GM kept their own notes, their own counts and their own idea of a good day.

Now every GM opens the same daily view: the forecast, the schedule against it, the cut calls, the orders and the exceptions that need a decision. The owner reads the same view for every location.

A standard a new location can open on

What happened each day flows back into the next forecast, so it gets better every week.

Every location now preps, staffs, cuts and orders from the same forecast. The next site opens on that standard, not on finding another great manager.