Theoretical vs Actual Food Cost: Finding Where Your Margin Leaks

Two numbers, one question

Most owners know their food cost as a single percentage: food purchases divided by food sales. That figure tells you whether you have a problem. It does not tell you where.

Splitting it in two does:

• Theoretical (ideal) food cost: for every item sold, multiply the quantity by its recipe cost, then add them up. This is what food should have cost if every plate matched the recipe and nothing was wasted.
• Actual food cost: opening stock plus purchases, minus closing stock. This is what you really used, measured by counting.

Variance is actual minus theoretical. If it is small and steady, your kitchen is under control. If it jumps, you know to look, and the ingredient-level breakdown tells you where.

If you have not yet set your target food cost, start with our restaurant food cost guide, then come back here to find where the money goes.

Getting the theoretical figure right

The theoretical number is only as good as your recipes. Before you trust it:

• Cost every menu item. Each dish needs a recipe with quantities for every ingredient, including garnish, sauce and oil. Our recipe costing template walks through this.
• Use yield-adjusted quantities. If a whole fish yields 55% usable fillet, the recipe should deduct the whole-fish weight, not the fillet weight. Otherwise theoretical usage will always look too low.
• Include modifiers. Extra egg, add cheese, upsize: each should deduct its own ingredients. Otherwise every add-on appears as variance.
• Update prices. Recipe costs should use current purchase prices, or the theoretical cost drifts away from reality as suppliers raise prices. When menu prices need to follow, see our guide to menu price increases.

With recipe-level stock deduction, each sale in the POS deducts the recipe’s ingredients from stock. At the end of the period, the system already knows how much of each ingredient your sales should have used. That is your theoretical usage, ingredient by ingredient.

Getting the actual figure right

Actual usage comes from three numbers per ingredient:

Actual usage = opening stock + purchases received − closing stock

The weak point is usually the count. Count the same way every time:

• same day and time each period (for example, Sunday night after close)
• same units as your recipes, or clear conversions (a 5kg bag counted as 5,000g)
• open containers estimated consistently, or weighed
• items in the chiller, freezer, dry store and on the line all included

Our stock take guide covers count sheets and routines. Start with a weekly count of your 15 to 20 most expensive ingredients rather than everything. That usually covers most of your spend and keeps the count to under an hour.

A worked example (hypothetical)

The figures below are invented to show the method. They are not benchmarks for any real outlet.

A hypothetical 60-seat Western café in Singapore runs a weekly count on key ingredients. Food sales for the week are $24,000. The POS, using recipe-level deduction, puts theoretical food cost at $6,240 (26.0% of sales). The stock take shows opening stock of $4,200, purchases of $6,800 and closing stock of $4,000, so actual food cost is $7,000 (29.2%).

The variance is $760, or 3.2 percentage points of sales. Over a year that would be roughly $39,500, which is enough to justify an hour of investigation each week.

The total does not say where the money went, so the café breaks it down by ingredient:

Ingredient	Theoretical usage	Actual usage	Variance	Cost of variance
Chicken thigh	42.0kg	48.5kg	+6.5kg	$58.50
Salmon fillet	18.0kg	21.2kg	+3.2kg	$108.80
Ribeye	14.0kg	16.9kg	+2.9kg	$145.00
Fresh cream	22.0L	25.0L	+3.0L	$27.00
Avocado	160 pcs	215 pcs	+55 pcs	$110.00
Other items				$310.70
Total				$760.00

Five ingredients account for about 60% of the variance. That gives the manager a short list to investigate: ribeye, avocado and salmon first, because they cost the most.

Investigating the gap: the five usual causes

1. Waste and spoilage

Avocados that ripen too fast, cream past its date, a tray of chicken dropped on the floor. Waste is real usage that no sale accounts for. For the national picture, see food waste figures for Singapore. Fix: keep a waste log at the pass or on a tablet, with item, quantity and reason, and review it at each count. Order smaller and more often for fast-spoiling items. Our food waste guide covers prep planning and first-in, first-out storage.

2. Over-portioning

If the recipe says 180g of ribeye and the grill cooks cut 200g “to be safe”, that is 11% over on every steak. Fix: portion high-cost proteins in advance, provide scales at the station, and use photos of the correct plate. Spot-check a few plates each shift.

3. Unrecorded meals, comps and remakes

Staff meals, a free dessert for a regular, a steak sent back and recooked. If these are not recorded in the POS, the food leaves stock with no matching sale. Fix: ring staff meals and comps through the POS as zero-priced items or with a reason code, so they deduct stock and show up in reports. Our staff meal policy guide helps you set the rules. Limit who can void or comp with staff permissions.

4. Theft

Uncomfortable but real. Signs include variance concentrated in portable, high-value items (spirits, premium meat, seafood), variance that appears on certain shifts, or deliveries that “go missing” between the loading bay and the store. Fix: lock high-value stores, restrict key access, count those items more often, and compare variance by shift where you can.

5. Receiving and pricing errors

You paid for 20kg of salmon but 18kg arrived. A supplier raised the price and the recipe still uses the old cost. Fix: weigh and count at delivery against the invoice, and reject or note shortfalls there and then. Update recipe costs when prices change. Our guide to reducing restaurant costs covers supplier checks.

Reading variance patterns

The pattern often points to the cause before you walk into the kitchen:

• Variance spread evenly across many items: likely a counting or recipe problem, such as wrong units or missing yields.
• Variance in one or two proteins: portioning, theft or receiving.
• Variance in produce and dairy: usually spoilage and ordering.
• Variance that swings wildly week to week: inconsistent counting, or deliveries recorded in the wrong week.
• Negative variance (used less than theoretical): usually a recipe that overstates quantities, or a delivery not recorded. It is not a saving to celebrate until you have checked.

Track the variance percentage each week on one chart. The trend matters more than any single week.

Food cost software vs a food cost tracker spreadsheet

You can do all of this in a spreadsheet: export sales by item, multiply by recipe quantities, and set that against your count sheet. Start with a clean item list, such as this inventory template for F&B. It works for a small menu, but it takes hours and breaks whenever the menu changes.

Food cost software built into the POS does the theoretical side automatically, because each sale deducts ingredients as it happens. The manager’s time goes into counting and investigating, which is where the money is recovered.

When comparing systems, ask each vendor to show:

• recipe-level deduction, including modifiers and sub-recipes such as sauces
• a report of theoretical against actual usage by ingredient
• how stock counts are entered (tablet, phone, import)
• how waste, staff meals and comps are recorded
• purchase orders and receiving against invoice

How ChaChaCha supports this

ChaChaCha, powered by AppsPOS, does recipe-level stock deduction, so each sale deducts ingredients by recipe, with recipe costing per menu item. That gives you the theoretical side of the comparison without spreadsheets. Staff permissions control who can void and comp, and reports show sales by item so you can match them with usage. For purchase orders, stock transfers between outlets, unit conversion and wastage logging, ask us to confirm how your set-up would work. See our inventory management page, or contact us for a walkthrough with your own menu.