How to Cut Restaurant Food Cost — Without Cutting Corners
Guide for restaurant & café owners in Kuwait and the GCC · 8 min read
If your food cost is a number you only see once a month — after the stock count, after the invoices, after the guesswork — you are managing your kitchen with the rear-view mirror. This guide shows how to make food cost a live number you can act on the same day, using recipe costing, automatic stock depletion, and menu engineering.
The short version
- Good food cost is 28–35% of revenue (cafés/QSR closer to 25–32%). Consistently above 35% means a pricing, portion, waste, or tracking problem.
- The silent leak is “eyeballing.” Every un-measured gram of protein, cheese, or sauce is margin you never see leave.
- Fix it by making cost automatic: link each dish to a costed recipe so the POS shows ideal cost, food-cost %, and margin as you build it.
- Let sales deplete stock themselves (backflush) so cost of goods and inventory stay accurate with zero manual counting.
- Rank your menu by margin, not by cost % — you bank money, not percentages.
First, know your number: the food cost formula
Food cost percentage is what your ingredients cost divided by what you sold them for.
Food cost % = (Cost of ingredients used ÷ Revenue from dishes sold) × 100Per dish, it is the recipe's ideal food cost divided by its selling price. A plate that costs you 1.100 KWD in ingredients and sells for 3.500 KWD runs a 31% food cost — comfortably inside the healthy band. The trouble is that most kitchens never know the top of that fraction with any precision, because nobody weighs what actually goes on the plate.
Where the money actually leaks
Across the industry the same handful of causes show up again and again: inconsistent portions, over-ordering and spoilage, unnoticed supplier price creep, and a stock count that is either too infrequent or simply inaccurate. The common thread is measurement. You cannot control a cost you only estimate. When a line cook “eyeballs” 130g of chicken instead of 110g, that 20g doesn't show up as a loss anywhere — it just quietly widens your food cost until, months later, the P&L looks wrong and nobody can say why.
The mindset shift — Stop treating food cost as an accounting result you discover after the fact, and start treating it as an operational setting you configure up front — one recipe at a time.
Step 1 — Cost every dish with a recipe
The foundation is a costed recipe for each menu item. In Culina, you open Recipes, pick the dish, add each ingredient with its quantity and unit, and set the yield (how many servings the recipe makes). Because your raw materials are received into stock with a real purchase cost, the builder shows you three numbers live, as you type: the ideal cost per serving, the food-cost %, and the margin.
That single screen turns pricing from a hunch into a decision. Want a 30% food cost on a dish that costs 1.050 KWD to make? The math is in front of you. Add a premium ingredient and watch the margin move before you ever put the item on the menu.
Step 2 — Let each sale deplete stock automatically (backflush)
A recipe that only lives on paper still relies on someone counting. The second step closes that gap. With recipe depletion (also called backflush) switched on, every time a dish is sold the system automatically issues its ingredients from the kitchen / ingredient warehouse and books the real cost to the ingredient COGS account. Sell a burger, and the bun, patty, and sauce leave stock on their own.
You choose the timing: real-time posts one stock movement per order — best when you want up-to-the-minute inventory — or end-of-day posts a single aggregated entry per outlet, which is lighter for high-volume venues. Refunds and cancellations reverse the movement, so your numbers stay honest. The result: your cost of goods and on-hand inventory reflect what you actually sold, with no manual tally at midnight.
Why it matters — Once recipes and depletion are on, the COGS and margin lines in your branch report stop being estimates. They become the real ingredient cost of everything you sold — the number you have been missing.
Step 3 — Engineer the menu by margin, not by cost %
Now that every dish has a true cost, you can do the thing that actually moves profit: decide what to promote, reprice, or retire. Culina plots each recipe-backed dish on a menu-engineering matrix (the Kasavana-Smith model) using two axes — popularity (units sold) and profitability (contribution margin per unit = selling price − ideal food cost). Each axis splits at the median, so every dish lands in one of four quadrants.
| Quadrant | Popularity | Margin | What to do |
|---|---|---|---|
| ⭐ Star | High | High | Feature and protect — never cut quality or visibility |
| 🐎 Plowhorse | High | Low | Nudge price up, trim cost, or upsell paid add-ons |
| 🧩 Puzzle | Low | High | Promote, reposition, rename, or photograph it |
| 🐢 Dog | Low | Low | Rework, re-price, or retire |
Rank by margin (money), not by food-cost %. A dish with a “worse” cost ratio can be your most profitable item, because you bank KWD per plate — not percentages.
A worked example (real demo figures)
With the medians sitting at 101 units and 1.295 KWD contribution margin:
| Dish | Units | Price KWD | Cost KWD | Cost % | Margin/unit | Verdict |
|---|---|---|---|---|---|---|
| Chicken Caesar Salad | 128 | 3.250 | 1.138 | 35% | 2.112 | ⭐ Star |
| Iced Latte | 99 | 1.500 | 0.225 | 15% | 1.275 | 🐢 Dog |
The Caesar's 35% cost ratio looks worse than the latte's tidy 15% — yet the salad is a Star and the latte is a Dog. Why? The Caesar sells more (128 > 101) and keeps 2.112 KWD per plate, while the latte earns just 1.275 KWD and sells just below the median. Each Caesar contributes roughly 2.11 KWD versus the latte's 1.28 KWD. The latte's low cost percentage does not pay the rent. Push the salad.
Putting it together
Cutting food cost is not about buying cheaper chicken or shrinking portions until guests notice. It is about measurement replacing estimation: cost every dish once, let sales deplete stock automatically, and let the margin data tell you which items to promote and which to fix. Do that, and the 32% food-cost average stops being a number you hope for at month-end and becomes a dial you turn — this week, from one screen.
See your real food cost, live
Culina is a restaurant POS built on ERPNext — recipe costing, automatic depletion, and menu engineering included, bilingual EN/AR.
Book a demoFrequently asked questions
What is a good food cost percentage for a restaurant?
Most restaurants aim for 28–35% of revenue; cafés and quick-service often target 25–32%, and fine dining can run 30–40%. Consistently above 35% usually signals a pricing, portion, waste, or tracking issue. The best benchmark is your own: is your food cost stable and consistent with how you priced the menu?
How do I calculate food cost percentage?
Divide the cost of ingredients used by the revenue those dishes generated, then multiply by 100. Per dish, divide the recipe's ideal food cost by its selling price. A POS with recipe costing does this automatically for every item.
What is backflush / recipe depletion?
It means the POS removes a dish's ingredients from stock and books the real cost the moment it is sold — in real time or aggregated at end of day — so inventory and cost of goods stay accurate without manual counting. Returns reverse it.
Should I rank menu items by cost % or by margin?
By contribution margin — the money each sale keeps (selling price minus ideal food cost). A low cost percentage can still be a low-profit item. You bank money, not percentages.