Unit economics for a restaurant: what one cover really earns

We've taken apart a D2C order, a SaaS subscription, and a billable hour. A restaurant is all three at once, which is exactly why it's the hardest of the four to run. It carries physical inventory that spoils, like the D2C brand. It sells perishable capacity — an empty seat tonight is gone forever — like the services firm. And it does both on a margin so thin that the difference between a great year and a closed sign is a couple of guests a night. Part four of the series is about that knife-edge, and the two numbers that tell you which side of it you're on.

The reason so many good restaurants with full dining rooms still fail isn't bad food or a bad location. It's that the model has almost no slack in it: a huge fixed cost sits under a wafer-thin contribution per guest, so profit lives entirely in the last few tables of the night. Miss them and a "busy" restaurant loses money. Three rungs show you where the money is, and two reads show you how little room there is.

What is the "unit" in a restaurant?

For D2C it was one order; for SaaS, one customer; for a services firm, one billable hour. For a restaurant, the smallest repeating thing you sell is one cover— one guest served, one check. Not one dish (too granular to compare) and not one night (that's the whole P&L). Every question about whether the restaurant works is a story you can tell about a single cover: what they spent, what it cost to feed and serve them, and what was left after the rent.

And the number you're chasing isn't food-cost percentage, the figure most operators fixate on. Food cost is only the first rung. The honest measure is what a cover contributes after both variable costs — food andthe labor to cook and serve it — and then what survives the fixed nut behind the whole room. That's the atom of the business.

The cover that keeps less than the food cost suggests

Let's run one cover all the way down. Our example is a full-service neighborhood restaurant with an average check of $45. Food and beverage run 30% of the check — a textbook food cost — so $13.50 of that $45 is on the plate and in the glass. That leaves a 70% gross margin, which sounds wonderful right up until the rest of the costs arrive.

Seventy percent on the food-cost line; six percent in the bank. The famous food-cost number told you almost nothing about whether the night made money — the two lines that decided it were the labor, nearly as large as the food itself, and the $16.35 of fixed costeach cover had to carry. That last one isn't really a per-cover cost at all — it's a fixed monthly nut divided by however many guests walked in, which is the whole game, and it deserves its own look.

Prime cost: the number you actually run on

Add the two variable costs back together — food and labor — and you get the single most important number in a restaurant: prime cost. $13.50 + $12.60 = $26.10, or 58%of the check. Operators live by this figure because it's the part of the P&L they can actually move day to day, and the rule of thumb is stark: keep prime cost under about 60%of sales and there's room for the rent and a profit; let it drift past 65% and there simply isn't, no matter how full the room gets.

Our restaurant is at 58% — healthy, with a little daylight. That daylight is the 42% contribution — the $18.90each cover throws off toward the fixed costs. Hold that number; it's the one the entire break-even calculation runs on, and it's far thinner than the 70% food margin made the business look.

C1, C2, C3: the restaurant ladder

D2C operators count in CM1/CM2/CM3, SaaS in L1/L2/L3, services in R1/R2/R3. A restaurant climbs the same three steps — gross, then contribution, then net. Call the rungs C1, C2, C3, C for cover.

• C1 — gross margin.Check minus food & beverage cost. $45 − $13.50 = $31.50. The food-cost line everyone watches — necessary, flattering, and nowhere near the whole story.
• C2 — contribution. C1 minus the variable labor to cook and serve. $31.50 − $12.60 = $18.90. What a cover really throws off after the two costs that scale with every guest — food and labor together being prime cost. This is the cash that has to pay for the room.
• C3 — net per cover. C2 minus the fixed nut spread across every cover. $18.90 − $16.35 ≈ $2.55. The real profit on a guest — about 6% of the check, which is a good full-service restaurant. This is why the industry is famous for thin margins: even done right, there's barely anything here.

Read left to right and the restaurant works — but only just. At 160 covers a night(80 seats turned twice) across 26 nights, that's about 4,160 covers and $187,000 of revenue a month, dropping roughly $10,600 to the bottom line. Real money — but notice how much had to go right to get it, and how little would have to go wrong to erase it. That fragility is the whole point, and the next number names it exactly.

Build your own cover

Your restaurant isn't this one. A quick-service spot lives on a $12 check turned six times; fine dining bills $120 but eats a brutal prime cost on few slow turns; a café survives on volume at a tiny ticket. Put in your real check, food and labor percentages, the size of your room, and your monthly nut, and watch the ladder — and the two reads that decide everything — resolve:

Two knobs move everything. Drag turns from 2.0 down toward 1.5 and watch net per cover fall off a cliff while every cost stays put — that's the fixed nut with fewer guests to carry it. Then push food or labor costup a few points and watch prime cost cross 60% and the profit vanish. The room's capacity and its prime cost are the only two levers, and both are unforgiving.

The two reads a restaurant lives by

The first you've met — prime cost %, food plus labor as a share of sales, kept under ~60%. The second is where the fragility becomes a number: break-even covers. Divide the fixed monthly nut by the contribution per cover — $68,000 ÷ $18.90 ≈ 3,600 covers a month, about 138 a night. You seat 160. So 138 of every 160 covers exist only to pay the rent; the profit is the last 22.

Put in turns instead of covers and it's even starker: 1.73 of your 2.0 turns just pays the fixed costs. That's the restaurant's version of being profitable but broke: the room can be humming, the reviews glowing, and a two-week slow patch that trims turns from 2.0 to 1.7 flips the whole month into a loss. A 15% dip in guests doesn't cost you 15% of profit — it costs you all of it, and then some.

Operating leverage: why turns swing everything

That knife-edge has a name — operating leverage— and it's the defining physics of the model. Because so much of the cost is fixed for the night, every cover past break-even drops almost entirely to the bottom line, and every cover short of it bleeds almost entirely from it. The swing is violent in both directions.

• A good quarter — turns 2.0 → 2.3. Covers rise to about 4,780 a month, revenue to $215,000, and net profitmore than doubles to roughly $22,000. A 15% bump in guests, a 110% bump in profit.
• A slow quarter — turns 2.0 → 1.7. Covers fall to about 3,540, and the same restaurant swings to a ~$1,200 monthly loss. Nothing changed but the door count.

This is why restaurateurs are obsessed with reservations, bar seating, turning tables faster, and filling the dead Monday: at this margin, throughput isn't one lever among many — it's the business. And it's where the restaurant meets the services firm: both sell perishable capacity, both live and die on how much of it they fill. Except a restaurant carries a second perishability the services firm never faces — the inventory itself spoils. Every unsold seat expires at closing time, and so does every unsold oyster.

Four levers that move the cover

When C3 is thin or break-even sits too close to a full house, there are only a handful of places to push — and the one owners reach for first (cut food cost) is rarely the biggest.

• Fill more seats before anything else. Throughput is the highest-leverage number you have because your costs are mostly fixed for the night. A dead Tuesday brought to life, a bar that seats walk-ins, a faster turn at dinner — each drops almost entirely to net. You're not paying more rent to seat the 161st cover.
• Engineer the menu, not just the price. Raising the average check lands mostly in contribution — but the sharper move is mix: steer guests toward the high-margin dishes and drinks (the classic "stars"), reprice or retire the low-margin ones. A well-designed menu quietly lifts the check without a single price feeling higher.
• Then attack prime cost — both halves. Portion control, waste tracking, and vendor terms defend the food line; scheduling labor to the actual covers you expect (not a flat roster) defends the labor line. A point off prime cost is a point straight onto a 6% net — proportionally enormous.
• Respect the fixed nut — you sign it once. Rent is the cost you can't fix later; it's set the day you sign the lease and it sets your break-even for years. The deadliest number in the whole model is the one most owners negotiate least — occupancy as a % of sales, ideally under ~8-10%.