Food cost in US restaurants: the complete 2026 guide

The short version. Food cost is the slice of your revenue that goes into raw ingredients. Target: 28-32% for most full-service concepts. The problem: most US operators find out their number in the year-end P&L — when the money is already gone. This guide shows you how to calculate it, where it leaks without you noticing, and how to manage it week to week.

What food cost actually is

Food cost is the ratio between what you spend on raw ingredients and what you bring in from food sales. It's one of the two numbers that determine whether your restaurant is viable — the other is labor cost. Together, food and labor typically eat 55 to 65% of revenue (that's your prime cost). What's left covers rent, utilities, equipment, and if you're lucky, paying yourself.

According to the National Restaurant Association, the average food cost across US independent restaurants runs 28-35%, with the gap between top performers and median operators sitting around 4-6 points — a difference that translates directly to the bottom line. Most operators know the term. Few actually manage it. Even fewer measure it in real time.

What is a good food cost percentage for a US restaurant?

Direct answer: it depends on your concept. But there are reference ranges the industry has validated for decades.

The problem is that many operators aim for "around 30%" without ever verifying they're actually there. Two or three points off on $800,000 of revenue is $24,000 of margin lost per year. That's not a rounding error — that's a dishwasher's annual salary.

Target food cost by concept type (2026)

These are anchors, not gospel. If you're at 36% in a casual dining spot, you have a problem. If you're at 24% in fine dining, either your suppliers are exceptional or your portions are too small. Either way, understand why — don't just hope the number is right.

What matters is knowing where you stand today — not when the accountant calls in January.

How to calculate food cost per dish

Base formula: (raw ingredient cost × waste coefficient) ÷ net selling price × 100.

Two distinct calculations exist. Theoretical food cost — what you run on paper from your recipe cards. And actual food cost — what your COGS calculation tells you after a real physical inventory. Both are useful. The gap between them is even more useful.

Per-dish worked example (US dollars)

Take a burger plate sold at $18 net (pre-tax).

• Ground beef patty (6 oz raw, 12% cook loss) → cost = $1.92 × 1.12 = $2.15
• Brioche bun → $0.45
• Lettuce, tomato, onion, pickles → $0.38
• Cheese slice → $0.30
• Fries (5 oz raw, 8% yield loss) → $0.55
• Condiments, packaging liner → $0.20
• Total ingredient cost: $4.03

Theoretical food cost = ($4.03 ÷ $18) × 100 = 22.4%

You're in good shape for that concept — on paper. But is your cook actually weighing 6 oz of beef every service? Did your ground beef price go up when Sysco sent the quarterly update? That's where theoretical and actual start to diverge.

Actual food cost — the COGS formula

Actual food cost (%) = (Opening inventory + Purchases − Closing inventory) ÷ Net food sales × 100

This is the only number that tells you what really happened in your BOH over that period. It captures waste, over-portioning, spoilage, employee meals, and everything that didn't leave on a sold plate.

One important US accounting note: track beverage cost separately from food cost. Beer, wine, spirits — those go in their own COGS line. Blending them obscures both. Standard US industry benchmarks treat food cost and beverage cost as distinct metrics. Don't mix them.

Where the gap comes from

Theoretical food cost is your perfect kitchen on paper. Actual is your kitchen the way it really runs. The average gap in independent US restaurants sits around 3 points (NRA data). In poorly managed operations, it can hit 6 or 7.

The most common sources

1. Involuntary over-portioning. Your cook plates 7 oz of protein instead of 6. Without systematic weighing, you'll never know.
2. Untracked waste. Produce damaged at delivery, poorly cut proteins, product that hits its use-by date before it moves.
3. Vendor price drift not reflected in recipe cards. Sysco and US Foods push quarterly price updates — sometimes mid-cycle on volatile commodities like beef and avocado. If your recipe cards still run last quarter's prices, your theoretical food cost is fiction.
4. Shrinkage. Unrecorded employee meals, walk-in snacking, the occasional free order for a regular.
5. Recipe drift. Staff turnover, new hires in the middle of a busy season: grammage shifts without anyone noticing.

That's why recipe cards kept current with live vendor prices are not a fine-dining luxury. They're the foundation of any real food cost management.

The 7 levers to cut your food cost

Cutting food cost doesn't mean cutting quality. It means plugging the leaks. Here are the 7 operational levers, in order of impact.

1. Update your vendor price list every week

The price of beef, salmon, and cooking oil moves — sometimes significantly between Sysco billing cycles. If your recipe cards are running on prices from last quarter, your theoretical food cost is wrong. You're making menu pricing decisions on wrong numbers.

A live vendor price list tells you instantly when an item costs you 15% more than two months ago. That's when you adjust — selling price, portion size, or substitution — before it costs you a full month.

2. Weigh systematically at the pass

Put a scale on the line and weigh protein portions before they go out. Not to police your team — to establish a real baseline. In 3 weeks, you'll know whether your actual portions match your recipe cards. They often drift 10 to 15%.

3. Optimize your vendor orders

Too much stock = spoilage. Too little = emergency orders at walk-in prices from a restaurant supply store. The right level is ordering from actual sales history and real inventory. Simple in theory. Hard at 6am when you're also prepping service.

4. Work your menu engineering

Some dishes drive volume but kill your margin. Others are profitable but rarely ordered. Crossing food cost percentage with popularity tells you what to keep, what to reprice, and what to 86 from the menu. This is often where the fastest margin recovery sits.

5. Inspect deliveries and log losses

A case of romaine arriving 20% damaged = your actual food cost climbing with no traceable cause. Check deliveries, log shortages, and enforce return clauses with your vendors. That's cash recovered directly.

6. Standardize and train continuously

One dish, one recipe, one portion weight. Not "roughly." If you run a seasonal brigade or see high BOH turnover (US restaurant industry turnover runs 75%+ annually, per NRA), standardization is the only way to maintain consistency without a full-time expediter.

7. Track food cost at least monthly — aim for weekly

A food cost number read once a year is endured. A food cost number read every week is managed. Quarterly is too slow — vendor price drift from Sysco's Q1 pricing update can blow your Q2 food cost before you've even noticed.

Case study — La Verrerie, 2016

In 2016, La Verrerie was running well. Revenue was climbing. Service was solid, reviews were improving. Then the year-end accounts came in. My accountant told me actual food cost for the year was 36%. My target was 28%. Eight points of margin gone across the whole year.

On roughly $500,000 of revenue at the time, that was $40,000 of margin evaporated — without me seeing a thing.

How? I had recipe cards. I was ordering regularly. But my vendor prices hadn't been updated since we opened. My cooks were portioning without weighing. And for six months I'd been doing emergency orders myself after my head chef walked out in peak season — at unnegotiated prices, in small quantities.

My theoretical food cost was at 29%. Actual was 36%. Seven-point gap. The lesson: the year-end accounts are a post-mortem. By the time you read them, the money is gone. You need at minimum a monthly number — and ideally a system that recalculates automatically every time a vendor invoice lands.

That experience is exactly why I built Onrush. Not a dashboard with nice graphs. A tool that stops you finding this kind of surprise in the P&L.

Managing food cost: year-end P&L, spreadsheet, or real time?

Three practices coexist in 2026. They don't produce the same decisions.

The question to ask any food cost tool: does it recalculate all my recipe cards automatically the moment a vendor price changes?

If the answer is "no" or "you have to update it manually," that's not management — that's disguised bookkeeping.

Onrush works like this: you photograph your Sysco or US Foods invoice. OCR reads the items and prices. Every recipe card that uses those ingredients recalculates automatically. Your food cost is current. The whole thing takes 30 seconds.

Common food cost mistakes

• Reading food cost once a year in the P&L. By then, the money has been gone for months. Monthly COGS calculation is the minimum.
• Not building waste into theoretical food cost. A 0% waste coefficient on proteins doesn't exist. Ground beef loses 15-20% to cook shrinkage. Whole fish loses 40-50% to butchery.
• Using stale vendor prices in recipe cards. If your Sysco price list is from Q4 and it's now Q2, your theoretical food cost is fiction.
• Ignoring the theoretical/actual gap. Most operators calculate theoretical and stop. The COGS-based actual figure — the one that really matters — only around 20% of US indie operators track systematically (NRA).
• Not training the BOH on portion weights. A recipe in the head cook's memory doesn't survive a brigade of four in a Saturday dinner rush.
• Mixing food cost and beverage cost. Keep them separate. Different benchmarks, different vendors, different levers.

Conclusion

Three things to take from this guide.

First: food cost is not managed in the year-end P&L. It's a living number that moves every time a vendor changes a price, a cook over-portions, or an ingredient gets wasted at delivery. Discovering it once a year is driving while looking in the rearview mirror.

Second: the gap between theoretical and actual is your real gauge. Theoretical at 29% with actual at 36% is 7 points of margin disappearing silently. On $800,000 of food revenue, that's $56,000 a year. Not an accounting anomaly — an operational leak.

Third: structure makes the difference. Current recipe cards, live vendor price list, monthly inventory minimum, portion training for the BOH. Not bureaucracy. The line between a restaurant that endures its margins and one that steers them.

If you're ready to move from post-mortem to real-time, see what Onrush does — or book a 20-minute demo and run the numbers on your own operation.

Written by Cyril Quesnel, founder of Onrush. 20 years in restaurant kitchens in France, two turnarounds. Building the tool I wish I had for US indie restaurants.