Forecasting is the prediction of future revenue based on historical data, weekdays, seasonality, weather and local events. In hospitality it is the basis for revenue-based scheduling, purchasing of goods and liquidity planning.
Key factors
- Weekday: the biggest single factor — Saturday vs. Tuesday can mean a factor of 3
- Seasonality: monthly and yearly fluctuations by concept and location
- Weather: outside temperature, rain and sunshine influence guest numbers significantly
- Local events: trade fairs, concerts, sports events, school holidays
- Public holidays: Mother's Day, Valentine's Day, Christmas — with sometimes extremely high deviations
- Long-term trend: growth or decline over several months
In the Scheduling module of Kiosa the revenue forecast is the basis: the AI creates the schedule from it and translates expected revenue into staffing needs.
Methods
Simple methods use moving averages of the previous week or the previous year. Advanced approaches use machine learning models (e.g. gradient boosting, neural networks) that weigh all factors at the same time. Modern ML models achieve a Mean Absolute Percentage Error (MAPE) below 10% at daily level, which is accurate enough for precise scheduling.
MAPE as a quality measure
MAPE = ∅(|Forecast − Actual| / Actual) × 100
| MAPE | Quality |
|---|---|
| < 5% | Excellent |
| 5–10% | Good — suitable for staff planning |
| 10–15% | Acceptable |
| > 15% | Insufficient |
Reference values for orientation, not industry statistics. Concept, menu and location shift the values.
Application in practice
With a forecast for the coming week, schedules can be matched automatically to expected revenue. Instead of fixed weekly shifts, staffing levels are adjusted by day and shift. For a chain with 10 locations, cutting overstaffing by 10% quickly saves a six-figure sum per year.
Related terms
- Anomaly Detection — when actual values deviate strongly from the forecast
- Schedule Efficiency — forecast as the basis for staff planning
- Labor Cost Ratio — is optimised by forecast-based planning
- Break-Even Point — forecast vs. break-even as an early-warning indicator
Track this KPI automatically
Kiosa calculates all restaurant KPIs automatically every day — for every location, without spreadsheets.
Book a Kiosa demo →