Cost control
Profitability
Food cost gone wrong: a worked example of where the money leaks
Restaurant food cost percentage is the share of your sales spent on the ingredients to make what you sold — and most operators run it three to eight percentage points higher than they should. The healthy benchmark is 28–32% for most concepts, but the average independent we audit lands closer to 35–38%. That gap between target and reality is rarely one big leak. It's usually four small ones running at the same time. We're going to walk through a real-numbers example to find them.
Key takeaways
Healthy food cost runs 28–32% for most concepts. Most independents run 35–38% - that gap is the leak.
Real food cost is (Beginning Inventory + Purchases − Ending Inventory) ÷ Sales. POS-reported food cost is theoretical and almost always wrong
The 30/30/30/10 rule is a useful starting framework - 30% food, 30% labor, 30% other, 10% profit - but real numbers vary by concept
The four most common leaks: portion drift, waste, theft, and vendor pricing creep - almost always running at once
A six-point gap on $200K monthly sales is $144,000 a year leaving the kitchen. The hard part isn't seeing the gap. It's splitting which leak is which
What is restaurant food cost percentage?
Food cost percentage is the dollar value of the food you actually used during a period, divided by your total sales for that period. It's expressed as a percentage. A restaurant doing $200,000 in monthly sales that used $66,000 worth of food is running a 33% food cost.
Food cost is half of prime cost — the other half being labor — and it's usually the half operators try to fix first because it feels more controllable. Labor has people attached to it, which makes it emotional. Food cost is just inventory and recipes, which makes it look like a math problem.
The benchmark for most concepts is 28–32%. Quick-service runs lower (often closer to 28%), full-service casual sits around 30–34%, and fine dining runs higher (32–38%) because higher menu price points absorb richer ingredient costs. Anything sustainably above 35% on a non-fine-dining concept is a structural problem.
How to actually calculate food cost (and why your POS gets it wrong)
Most operators read the food cost number off their POS dashboard and call it a day. That number is usually wrong, and not in a small way.
Your POS calculates theoretical food cost — what should have been used based on the recipes you entered for each menu item. If the dashboard says food cost is 30%, what it's really telling you is "if every cook portioned exactly to spec, with zero waste and zero theft, your food cost would be 30%." In a real kitchen, none of those conditions hold.
Real food cost is what your inventory actually shows:
Real Food Cost = (Beginning Inventory + Purchases − Ending Inventory) ÷ Sales
You count what's on the shelves at the start of the period, add what came in via purchases, subtract what's on the shelves at the end, and divide by sales. That's how much food you actually used.
The gap between theoretical and real food cost is the gap between what you assume and what's happening. Most operators we audit are 2–4 percentage points off the moment we count.
A worked example — where the leaks hide
A small full-service casual concept doing $200,000 in monthly sales. Here's the inventory math.
Line item | Amount |
|---|---|
Beginning inventory (start of month) | $14,000 |
Purchases during the month | $74,000 |
Ending inventory (end of month) | $12,000 |
Food used = Beginning + Purchases − Ending | $76,000 |
Food cost % = $76,000 ÷ $200,000 | 38% |
38% food cost. The benchmark for full-service casual is 30–34%. This restaurant is six to eight percentage points over target. On $200,000 monthly sales, that's $12,000–$16,000 leaving the kitchen every month — call it $168,000 a year at the midpoint.
Now compare to theoretical food cost. The POS says recipes should produce 30.5% — about $61,000 in usage. So actual ($76,000) minus theoretical ($61,000) is $15,000 in monthly variance. Where is the $15,000 going?
In our consulting experience, almost every food cost variance breaks into the same four buckets:
Portion drift: ~$3,000/month. Cooks under pressure portioning slightly heavy. A 7-ounce protein on a recipe that calls for 6 ounces, applied across 2,000 covers, is real money
Waste: ~$4,000/month. Trim, prep mistakes, items past expiration, food held too long during slow services. Roughly 2% of sales is typical for a kitchen without active waste tracking
Theft and shrinkage: ~$2,500/month. The hardest one to measure and the easiest to deny. Some is staff meals never logged, some is comps with no oversight, some is genuine theft. We rarely see a clean kitchen below $1,500/month here
Vendor pricing creep: ~$5,000/month. Suppliers raised prices in February. Menu prices haven't moved in eight months. The margin has been quietly compressing the whole time, and nobody noticed because the dollar amount of purchases looked normal
That's $14,500 of the $15,000 variance accounted for, give or take. None of it is one big leak. All of it is real money.
The four most common food cost leaks
The fix isn't dramatic. It's a little bit of attention paid to four things at once.
Portion drift is solved with portion training, the right scoops and ladles at every station, and a periodic line check during service. The cooks aren't trying to over-portion — they're trying not to under-portion in front of the customer. Make the right portion the easy default.
Waste is solved with a waste log. A clipboard at the prep station where every discarded item gets noted. Sounds tedious; takes 10 seconds per entry; pays for itself within the first month because it surfaces patterns (the same item gets thrown out repeatedly because the prep par is too high, or the rotation is wrong).
Theft and shrinkage is solved with controls, not surveillance. Tighter receiving (count it when the truck pulls up, not later), better comp authorization (only managers can comp; every comp gets a reason code), and inventory done by someone other than the person who orders it.
Vendor pricing creep is solved by reviewing your top 20 SKUs by spend every quarter and comparing prices to the previous quarter. Inflation isn't a single event; it's a thousand small price increases that compound. If you're not actively watching, your menu prices are always six months behind your costs.
The bottom line
Food cost gone wrong is almost never one disaster. It's four quiet leaks running at the same time, all of them small, all of them addressable, none of them obvious until you actually count the inventory and split the variance. The operators who run consistently low food cost don't have a secret. They have a calendar — count weekly, log waste, check vendor prices quarterly, train portions like it matters. Because it does.
Sources and further reading
Food cost percentage benchmarks (28–32% baseline, with concept-level variation): aggregated from US Foods, Square, ChowNow, and Black Box Intelligence
Real food cost formula (Beginning Inventory + Purchases − Ending Inventory ÷ Sales): standard restaurant accounting practice, documented at Square, Toast, and the National Restaurant Association
The 30/30/30/10 rule: industry framework with no single attributable origin; clearest published explanations at RestaurantOwner.com and Toast On the Line
The $200,000 monthly P&L example in this post is illustrative — a hypothetical full-service casual concept used to show how the math works. The four-leak framework and the variance breakdown reflect Baby Chef's consulting experience with independent and small multi-unit operators.
Baby Chef is a hospitality consulting team made up of operators who've worked the kitchen, the floor, and the back office. We help independent restaurant owners and small multi-unit operators audit, optimize, and future-proof their operations.
Food cost running hot and not sure which leak to chase first? Get in touch with us. We split variances for a living.
