A restaurant forecast is not a promise. It is the best operating estimate available before the shift, built from relevant evidence and improved by recording why reality differed.
Build a comparable baseline
- Start with recent comparable weekdays, seasonal patterns, dayparts, and the restaurant’s own history. Avoid averaging unlike days when events, holidays, school schedules, or major promotions make them materially different.
Add known context
- Reservations, large parties, catering, local events, weather, road closures, school calendars, promotions, sports or community events can matter depending on the restaurant. Record only context that has a plausible relationship to demand.
Forecast in a way operations can use
- A daily total may be insufficient. Break the estimate into dayparts or intervals when staffing, prep, table turns, or kitchen capacity require more detail.
Track forecast error
- Compare forecast with actual results and record material reasons. Over time, this distinguishes a bad model from an unusual day and helps managers calibrate future judgment.
Connect forecast to decisions
- Translate demand into staffing, prep/production, purchasing, reservations, manager coverage, and contingency planning. A forecast that does not change an operating decision is only a number.
Gravity principle: Context before conclusions. Evidence before judgment. People before percentages.
Put the guide into a working system
Gravity Sales Intelligence & Forecasting provides the data structure for this process.