Restaurant Demand Forecasting: Sales, Reservations, Inventory and Purchase Quantity

Buying too much creates excess stock and waste, while buying too little causes stockouts and lost sales. Demand planning should combine historical data with future conditions.

Separate demand from purchase quantity

Demand is expected consumption or sales over a period. Purchase quantity also depends on usable inventory, confirmed inbound orders, lead time, pack size and safety stock. Combining the two concepts can cause overbuying or shortages.

Core data

• Sales on comparable days and hours
• Order count and product mix
• Confirmed reservations
• Events and large groups
• Current usable inventory and open purchase orders

Clean historical data

Closures, equipment failures, unusual promotions, limited menus and stockouts should be flagged. Sales recorded while a product was unavailable may understate real demand. Exceptional data should be separated with a reason, not silently deleted.

Convert product demand into ingredient need

If recipes, portion sizes and standard waste are reliable, forecast product volume can be converted into approximate ingredient requirements. Without reliable inputs, purchasing and kitchen teams should review the conversion.

Add reservation and event detail

Guest count is only a starting point. Menu selection, arrival time, event duration, drinks, dietary restrictions, possible changes and guaranteed purchases all affect demand. Unconfirmed reservations should not carry the same weight as fixed events.

Build three scenarios

• Base scenario from comparable periods
• High scenario for events, weather or campaigns
• Low scenario for cancellations, closures or weaker demand

Long-life goods can carry more buffer, while perishables require tighter control of waste and storage capacity.

Lead time and order cycle

The quantity must cover demand until the next delivery, not only tomorrow. Ordering days, actual supplier lead time, holidays, minimum order and partial delivery should be included.

Practical purchase-quantity logic

Expected need plus a justified safety margin, minus usable inventory and confirmed inbound quantities, gives a management estimate of what to order. Damaged, reserved or near-expiry stock should not automatically count as freely usable.

Packaging and rounding

The estimate may use consumption units while purchasing uses cases or packs. Rounding should consider pack factors, minimum order, storage and shelf life. Always rounding up can increase waste.

Review forecast error and bias

Compare forecast with actual consumption or sales. Both error size and repeated direction matter. Consistently high forecasts create overbuying; consistently low forecasts increase stockout risk.

Pre-order checklist

1. Select comparable periods.
2. Flag stockouts and unusual events.
3. Enter future reservations and confirmed orders.
4. Check usable stock and inbound orders.
5. Review lead time, packaging and shelf life.
6. Have the final quantity approved by the purchasing owner.

Conclusion

A demand forecast is not a certain number; it is a repeatable process for making better purchasing decisions.

Frequently asked questions

How is demand different from purchase quantity?

Demand is expected usage; purchase quantity also considers inventory, inbound deliveries, lead time and packaging.

How should stockout days be analysed?

Flag them because recorded sales may understate actual demand.

How is safety stock determined?

Using demand variation, lead time, shelf life, item importance and waste risk.

How does forecast quality improve?

By regularly comparing forecasts with actual results and reviewing error size and direction.