.png)
Hardly any other sector of the food service industry produces as much surplus as institutional food service: company cafeterias, hospital kitchens, and college dining halls cook daily for a number of guests that isn’t determined until noon. What’s in the warming trays at 11 a.m. was decided hours earlier—based on experience and hope.
Every restaurant manager is familiar with the result: full containers at the end of the shift or an empty buffet at 12:30 p.m. Both come at a cost—one in terms of food costs and labor, the other in terms of revenue and guest satisfaction.
Why Employee Restaurants and Cafeterias Are So Difficult to Plan
The number of visitors varies from day to day. Work-from-home days, meetings, holidays, and the weather all determine how many guests come—on Fridays, there are often a third fewer than on Tuesdays, and even fewer on long weekends. Anyone who plans based on average figures will be off the mark almost every day.
The menu selection changes. Not only does the number of guests depend on the weather and what’s available, but so does what they order: a salad bar when it’s sunny, stew when it’s raining, and meat or vegetarian options depending on the day’s menu combination.
Production takes place in batches. Additional production during the service period is only possible to a limited extent—the initial quantity must be correct, and the decision to produce a second batch must be made under time pressure.
How Data-Driven Planning Is Changing Things
Fluctuations in demand in the institutional food service sector may seem chaotic, but they actually follow patterns: weekday rhythms, weather effects, school and holiday schedules, and location-specific characteristics. GoNina identifies precisely these patterns from sales data and provides daily forecasts for each component and location for the next seven days—offering concrete production recommendations rather than relying on gut feelings.
Practical experience shows that this works: In a pilot project with a leading Swiss provider of institutional food services, up to 48% less food waste was measured across ten locations—while maintaining the same level of availability at the buffet.
The Leverage in Everyday Cafeteria Life
Initial quantities per component. Instead of a total number for "meals," the system forecasts demand per menu item and component—which ensures that side dishes, proteins, and buffet items are produced in the correct proportions.
The weekly overview for purchasing and human resources. A seven-day lead time makes it possible to plan orders and shift schedules—a double benefit, especially when dealing with fresh goods and a tight staffing situation.
Transparency Regarding Locations. Those who manage multiple facilities can see in the dashboard where planning is on track and where there is systematic overproduction or underproduction—the foundation for targeted improvements rather than blanket directives. The article on plate return rates and production surpluses shows exactly where the surplus occurs in cafeterias.
Frequently Asked Questions
Our guest numbers depend heavily on working from home—can a forecast account for that?
Yes—working-from-home patterns show up as stable weekday rhythms in the sales data, and the model learns precisely those patterns. If behavior changes, the forecast adjusts continuously.
We have rotating menus—will this still work?
Yes. The forecast operates at the level of ingredients and categories, not just individual dishes—so the predictions remain reliable even with rotating menu cycles.
Is this worth it even for a single business?
The benefit per location is the same—less excess inventory, better availability, and less planning effort. Multiple locations multiply the benefits, but they are not a requirement.
Conclusion
Institutional food service operates daily to serve an unknown number of guests—and that is precisely why it is the sector where data-driven forecasts have the most immediate impact. A pilot project showing up to 48% less surplus demonstrates that these fluctuations are not inevitable, but rather a pattern that can be learned.
To learn how AI forecasting works in the restaurant industry as a whole, check out the complete guide for the restaurant industry.
