A buffet always needs more food per head than the same menu plated — and chefs who size a buffet by plated portions either run a station empty in front of guests or cook a third of it into the bin. The difference is not a fudge factor you guess. It is a model with three numbers: how much each guest is meant to take, how much extra has to be in the pans so the line never looks bare, and how many guests actually reach for each dish. Get those three right and a buffet sizes itself.
Why self-serve eats more than a plate
On a plate, you control the portion — the guest gets the 150 g you decided. On a buffet, the guest controls it, and three behaviours push consumption up. People over-serve themselves at a buffet (a heaped spoon, not a measured one). They come back for seconds on the dishes they liked. And the buffet has to look abundant the whole time — a pan scraped to the metal stops guests serving themselves even when there is plenty elsewhere. Plated portions ignore all three. That is why they undersize a buffet every time.
The three numbers that size a buffet
1. Per-guest amount
How much of a dish one guest is meant to take — the target serving, in grams or portions. This is the appetite number, set by the same things that move any catering quantity: event length, time of day, how many other dishes compete for the plate. A buffet with eight dishes has a smaller per-guest amount of each than a buffet with three, because the guest is spreading one appetite across more options.
2. Refill multiplier
How much you cook relative to the per-guest amount so the line stays full and the dish never runs out mid-service. A refill multiplier of 1.2 means you prep 20% above expected consumption — the buffer that keeps the pan presentable and covers the heavier-handed guests. Higher multiplier means a safer line and more waste; lower means leaner but riskier. This is the dial you set by how unrecoverable a bare station is for that event.
3. Pickup ratio
What share of guests actually take a given dish. Not every guest takes every item — maybe 70% reach for the salmon, 90% for the roast potatoes, 40% for the slaw. Sizing every dish as if 100% of guests take a full serving is the single biggest source of buffet over-production. The pickup ratio is how you stop cooking salmon for forty people when twenty-eight will take it.
Four numbers, one line per dish. Plated portioning has only the first and third of these — which is exactly why it is wrong for a buffet.
A worked dish
120-cover buffet, salmon as one of two proteins. Per-guest amount 90 g (it is one of several mains, so the serving is modest). Pickup ratio 0.65 — about two-thirds of guests take salmon when beef is the alternative. Refill multiplier 1.15 to keep the pan looking full.
Size that same salmon by plated logic — 120 guests × 140 g — and you would prep 16.8 kg, more than double. Half of it would leave as waste, on a job you already priced. The model is not academic; it is the difference between a 28% food cost and a 50% one.
Benchmark the waste, don't just feel it
Buffet waste is normal — the industry runs at roughly 15–25% on a well-managed self-serve spread. Below 15% and you are probably running stations close to empty, which guests notice. Above 25% and the model is too loose or the pickup ratios too generous. Knowing the band turns "that felt like a lot of leftovers" into a number you can act on, event over event. The same forecasting that cuts over-production also keeps less food sitting in hot-holding past its safe window — quantity discipline is quietly a food-safety discipline too.
For the operational research behind buffet waste and forecasting, hospitality studies on food-service production planning — including work from the Cornell Nolan School of Hotel Administration — cover the same consumption-versus-production gap this model manages.
In ProChefDesk
The Buffet tool runs exactly this model: per-guest amount, refill multiplier and pickup ratio per item, scaled to the cover count, with a waste-percentage benchmark that flags when you are heading above the 15–25% band. It outputs the prep quantities, the cost per cover, and a shopping list net of stock — so the line stays full and the food cost stays where you priced it. Open the app to build a buffet.
What to do this week
- Stop sizing buffets by plated portions. They have no pickup ratio and no refill buffer — the two numbers a buffet lives on.
- Estimate pickup ratios for your dishes from what actually comes back. The crowd-pleasers run high; the side salads run low.
- Set a refill multiplier per event by how visible a bare station would be — tighter for a casual lunch, safer for a showpiece.
- Weigh your leftovers once. Put a real waste percentage on a real event and the model gets sharper every time.
The point
A buffet is not a plated meal served from pans. It is a different appetite with a different model — per-guest amount, refill multiplier, pickup ratio — and the chefs who run profitable buffets are the ones who size each dish from those three numbers instead of guessing high to be safe. Guessing high keeps the line full and the food cost ruinous. The model keeps both right at once.