Menu engineering: why busy restaurants still lose money

Start with dish-level contribution, not food cost percentage

A headline food cost percentage hides the problem. What matters is the cash contribution each dish makes after its own ingredient cost, and how many of each you sell. Plot dishes on two axes — popularity against contribution margin — and four groups appear:

• High popularity, high margin — the core. Protect these, feature them, never change them without reason.
• High popularity, low margin — re-engineer the recipe, renegotiate the supply, or adjust portion and garnish to lift contribution without losing the dish.
• Low popularity, high margin — a selling problem. Better description, better position on the menu, or a service suggestion from staff.
• Low popularity, low margin — remove it. Every one of these consumes kitchen labour, storage and menu attention.

Complexity is the hidden cost

Long menus cost money in ways that never appear on a recipe card: more suppliers, more deliveries, more storage, more prep labour, more training, more waste, and slower service at peak. Trimming a menu often improves both margin and guest experience at the same time.

Count the distinct ingredients behind the menu and ask whether each earns its place across several dishes. Ingredient overlap between dishes is one of the cheapest sources of margin available.

Fix purchasing before you fix prices

Raising prices costs you customers. Renegotiating supply, tightening specification and consolidating the supplier list usually costs nothing. For one client we seconded an Executive Chef for six months specifically to overhaul purchasing and production; the resulting EBITDA improvement across food and beverage was significant, and it came from the supply chain rather than the price list.

• Consolidate suppliers and re-tender by category
• Write specifications so you are comparing like with like
• Track yield and waste, not just invoice price
• Review portion control at the pass, not in the office

Then look at throughput

If the kitchen cannot produce fast enough at peak, guests wait, covers are lost and staff morale falls. The constraint is usually equipment, workflow or the number of times a dish crosses the pass — not the number of chefs. Speed of service and transaction processing were paramount at the Eden Project precisely because transaction volume was the business model.

A worked order of operations

1. Build dish-level contribution for the last 12 weeks of sales.
2. Cut the bottom group and consolidate ingredients across the rest.
3. Re-tender the top categories of spend.
4. Re-engineer the high-popularity, low-margin dishes.
5. Retrain the team on the new menu and the new specification.
6. Only then consider price changes, and only where demand is inelastic.

In our experience the first three steps usually deliver more than any price increase would, without risking a single customer.