Theoretical vs actual food cost

Last updated: October 21, 2026 · By the Varastiq team

Key takeaway: Theoretical food cost is what your food should have cost: each menu item's recipe cost × the number sold. Actual food cost is what it did cost: beginning inventory + purchases − ending inventory. The gap between them is variance, the dollars that left your kitchen without a sale to explain them. Actual tells you where you are. The gap tells you what to fix.

What is theoretical food cost?

Theoretical food cost (also called ideal food cost) assumes every plate was made exactly to recipe, nothing was wasted, comped or stolen, and inventory was counted perfectly.

Theoretical food cost $ = Σ (Plate cost of each item × Number sold)
Theoretical food cost % = Theoretical food cost $ ÷ Food sales × 100

You need two things: a costed recipe for every item, at current prices, and a product mix report from your POS showing how many of each item sold.

What is actual food cost?

Actual food cost is what the shelves and invoices say you used:

Actual food cost $ = Beginning inventory + Purchases − Ending inventory
Actual food cost % = Actual food cost $ ÷ Food sales × 100

Full method: how to calculate food cost percentage.

How do you calculate the gap?

Variance $      = Actual food cost $ − Theoretical food cost $
Variance points = Actual food cost % − Theoretical food cost %

Both numbers use the same period and the same food sales, so the difference is directly comparable. The food cost calculator does this if you enter a theoretical figure.

What does a worked example look like?

A full-service restaurant, one week, $38,400 in food sales. Theoretical food cost from recipes × product mix:

• Theoretical food cost % = $11,142 ÷ $38,400 = 29.0%
• Actual food cost from counts and invoices = $12,200, or 31.8%
• Variance = $12,200 − $11,142 = $1,058 this week, or 2.8 points
• Over 52 weeks, about $55,000 of food the recipes can't account for.

Illustrative example, not customer data.

The restaurant-level number tells you there's a problem. It doesn't tell you where. For that, go item by item.

How do you find which items cause the variance?

For each high-value item, compare usage, not dollars:

Actual usage      = Beginning count + Received − Ending count
Theoretical usage = Dishes sold × Recipe quantity per dish
Item variance     = Actual usage − Theoretical usage

If the wings recipe says 0.6 lb per order and you sold 520 orders, you should have used 312 lb. If counts and invoices say you used 335 lb, that's 23 lb unaccounted for, at whatever wings cost you per pound. The free food cost variance calculator does this for a list of items at once.

Start with proteins, seafood, cheese and liquor. They carry the most dollars per unit, and they're where small portioning drift adds up fastest.

What causes the gap between theoretical and actual?

In rough order of how often they turn out to be the cause:

1. Counting errors. Wrong unit, missed shelf, guessed partials.
2. Invoice problems. Deliveries not entered, credits not applied, invoices in the wrong week.
3. Recipe drift. The recipe says 6 oz, the line plates 7.
4. Unrecorded waste. Spoilage, dropped food, re-fires, staff meals and comps not logged.
5. POS mapping. Modifiers or specials rung under the wrong item, so theoretical usage is wrong.
6. Theft. Product leaving without a ticket. It's real, but rule out 1–5 first.

How big should the gap be?

We haven't found a published industry standard for acceptable variance that we can cite, so we won't give you one. Practical guidance: some gap is normal, because no kitchen runs at zero waste. What matters is the size in dollars, which items drive it, and whether it's growing. Track it weekly and set your own thresholds per item. Varastiq's calculator defaults to flagging ±2% for a look and ±5% to investigate, but those are starting points, not a benchmark.

Why do so few restaurants track theoretical food cost?

Because it takes three systems lined up: costed recipes at current prices, item-level sales from the POS, and reliable counts. In a spreadsheet, that's hours each week, so it gets done at month-end, if at all. By then, the evidence is gone.

How does Varastiq track theoretical vs actual automatically?

Varastiq connects the three pieces:

• Recipes at current prices. Live plate cost updates from captured invoices.
• Sales by item. From Toast (read-only), or a sales report uploaded from a supported sales export.
• Counts. From the Stiq app, in walk order, with item search, manual quantities and offline drafts.

Then it does something a monthly report can't: it flags item-level variance during the count, so the person in the walk-in can recount or log the reason on the spot. See variance and theft flags.

Frequently asked questions

What is the difference between theoretical and actual food cost?

Theoretical food cost is what food should have cost based on recipes and items sold. Actual food cost is what it did cost based on inventory counts and purchases. The difference is variance.

How do you calculate theoretical food cost?

Multiply each menu item's plate cost by the number sold in the period, add them up, and divide by food sales for a percentage.

What causes a high variance between actual and theoretical food cost?

The most common causes are counting errors, missing or misdated invoices, recipe drift, unrecorded waste and comps, POS items mapped to the wrong recipes, and theft.

Is theoretical food cost the same as ideal food cost?

Yes. Both terms describe the food cost you would have if every item were made exactly to recipe with no waste.

Do I need a POS to track theoretical food cost?

You need item-level sales, which usually come from a POS product mix report. Varastiq reads them from Toast or from a sales report uploaded from a supported sales export.

What is an acceptable variance?

There is no published standard we can cite. Track the gap in dollars every week, focus on the items that drive it, and watch whether it is growing.

Sources

All figures on this page are illustrative examples. No external benchmarks are cited.

See the gap every week, item by item

Varastiq Pro: $149/month for your first location, plus $99/month for each additional location, or $1,639/year (1 month free). Optional $299 guided setup. Money-back guarantee on the software: 14 days on monthly plans, 30 days on annual.

Request a demo → · Try the variance calculator → · See variance flags →