How to Run a Restaurant: Costs, Benchmarks & Daily Systems

Running a restaurant comes down to controlling two costs —food and labor—while giving enough people a reason to come back. In the National Restaurant Association's 2026 State of the Industry survey, only 42% of operators said they were profitable, and more than six in ten saw guest traffic fall.

What separates the two groups is rarely talent. It's whether the owner knows their prime cost to the week, staffs against forecast covers instead of last month's, and owns the channel their regulars order through.

Below: the benchmarks a healthy restaurant hits, the rules that move them, and the weekly routine that keeps them honest.

What running a restaurant actually costs

A healthy restaurant spends 55–65% of every sales dollar on food and labor combined. That figure is called prime cost, and it's the single number most profitable operators watch weekly.

Everything else — rent, utilities, insurance, marketing, repairs, and your own pay — comes out of what's left.

If your prime cost sits above 65%, no amount of menu tweaking fixes it on its own — the problem is structural, and it's usually the schedule.

The gap between a 3% return and an 8% one is roughly five points of prime cost: one badly staffed Tuesday a week, repeated. Quick-service and fast-casual sites shift the whole table down, to a prime cost nearer 55–60%.

The money: four numbers you actually control

Sales are the number you can't move directly. These four you can.

1. Know your prime cost before anything else

Add food and beverage cost to total labor cost — wages, salaried management, payroll taxes, benefits — and divide by sales for the same period.

Do it weekly: a monthly number tells you what went wrong four weeks ago; a weekly one gives you time to fix the schedule before the next payroll run.

2. Price the menu from the plate cost up

Every dish needs a costed recipe before it needs a price. Cost the ingredients at current supplier prices, add a realistic waste allowance, then price so food cost lands inside the 28–35% band. A dish costing $4 in ingredients should sell somewhere between $11.50 and $14.30.

Then look at the mix, not just the dish. Menu engineering—sorting items by popularity and contribution margin, then repositioning or cutting accordingly—moves profit without changing a single price.

Keep the menu short while you're at it: fewer items means less prep, less spoilage, simpler training, and a kitchen that runs at speed. Once a quarter, run the food-cost math on every dish.

3. Staff against forecast covers, not last month's

Labor is the fastest line on your P&L to move and the one most often built out of habit. Start the schedule from forecast covers — same weekday last year, adjusted for bookings, weather, and anything happening locally — and staff to that, not to what you ran last week.

Paying people properly isn't the opposite of controlling labor — underpaying drives turnover, and replacing someone costs more than keeping them. The target is the right people scheduled at the right hours, which is what your labor line should actually look like when it's working.

4. Own the channel your regulars order through

Every order that arrives through marketplace costs you a commission and the customer's contact details. Orders through channels you own — your own site, your own app — cost neither.

The difference compounds. In our analysis of 72,468 customers across 1,337 restaurants, the same person ordered 87% more often a year after installing the restaurant's own app, and spent 129% more. Same people, same restaurants; the only variable was where they ordered from.

The practical version is short. Install a till system that gives you clean sales data, add a commission-free online ordering system to your own website, and give regulars an app to reorder from in two taps. Marketplaces still have a role in discovery — they just shouldn't be where your regulars live.

The team: your only real product differentiator

Two menus can be identical. Two dining rooms rarely are. Alongside the wider picture on how to manage a restaurant, here are the two parts that decide whether the room runs without you.

Hire slowly, train on a written standard

Chefs decide whether the food is good; servers decide whether people come back. Hire for attitude, train for skill, and take the extra week rather than filling a gap with whoever is free this Friday.

Then train on something written down. A restaurant staff training program that lives in a document—not in the head of whoever is on shift—is what makes service consistent from Tuesday lunch to Saturday night.

Write the system down so it survives your day off

Ask an owner why the restaurant runs worse on their day off and the answer is almost never "bad staff." The standard lives in the owner's head.

Chains don't outperform independents on talent. They outperform because every task has one written way to do it, and it doesn't change when a different person is on shift. Opening the line, plating the six best sellers, cashing up, handling a complaint — each can be a single page with photos and a checklist.

Start with the five tasks that go wrong most often in your absence. Write them as standard operating procedures, train two people on each, and have someone other than you sign off that it was done.

The guests: turning first visits into regulars

A full dining room of strangers is a worse business than a half-full one of regulars.

Fix the friction points before the food

The experience starts before the first plate and ends after the check. Parking, how long the phone rings, how long a table waits to be greeted, whether a pickup order is ready when you promised — any one of them can undo a good meal. Walk the path yourself once a quarter as a customer would, including your own online ordering flow. Fixing friction is usually cheaper than fixing food.

Ask for feedback on a schedule

Ask once the check has landed, not mid-service, using a short survey, your order confirmation message, and the review sites you already appear on. Read everything once a week rather than whenever something stings — the pattern across twenty comments is worth more than any single one.

Give regulars a reason to come back

Most restaurants know their busiest night. Far fewer know how many people in the room have been before.

We looked at 13 million orders from 1,709 restaurants in 18 countries to find out what that gap is worth. A customer who orders ten or more times a year is worth around €420; a one-time customer, around €31—14 times less —from a group that makes up only about 7% of a typical base.

Two things build that base: collecting contact details with consent at the point of ordering, and giving people a structured reason to return — a loyalty program, a reorder prompt, a birthday offer. Across 69 restaurants that launched one, orders rose 16% and revenue 22% within a year.

Your operating rhythm: what to check daily, weekly and monthly

Profitable restaurants aren't run on instinct; they're run on a calendar. The owners who can take a Tuesday off without the numbers moving are the ones who turned "keeping an eye on things" into fixed checks at fixed intervals.

• Every day — 15 minutes, before service. Yesterday's sales against forecast. Today's labor schedule against today's booked covers. The prep list against tomorrow's expected volume. Three comparisons, each a difference between two numbers you already have.
• Every week — about an hour. Count inventory and calculate prime cost for the past week. Read every review and reply to the negative ones. Build next week's schedule from forecast covers rather than copying last week's, and check that stock counts match what the system says they should be.
• Every month — half a day. Re-cost your ten best sellers against current supplier prices. Compare the P&L against the same month last year, not last month — restaurants are seasonal and month-on-month comparisons mislead. Review which restaurant KPIs are worth tracking weekly and drop any you haven't acted on.
• Every quarter. Menu review: cut what doesn't sell, reprice what has drifted, check the mix still supports your margin.

Problems compound quietly — a two-point food cost drift, a schedule built on habit. A fixed cadence catches them while they're small, and everything else in day-to-day restaurant operations sits on top of it.

What actually closes restaurants

The claim that 90% of restaurants close in their first year has no traceable source. It appears to have entered circulation through a television ad and repeated itself ever since.

The measured figures are different. A longitudinal study of restaurants in the western United States found a first-year failure rate of about 17% — lower than the 19% for other service-providing businesses — and a median restaurant lifetime of 4.5 years. Restaurants aren't unusually fragile. They're unusually thin-margined, which is a different problem.

That thinness is what makes failures look dramatic. A restaurant returning 4% net has almost no buffer: a two-point drift in food cost, a lease renewal, or a soft quarter hits the owner's pay.

Restaurants rarely close because of one catastrophe. They close because several small, measurable things drifted at once and nobody was measuring — which is the whole argument for the operating rhythm above. More detail on why restaurants fail.

Staying legal: permits, licenses and food safety

A business license, a food service permit, a certificate of occupancy, and — if you pour — a liquor license. Renewal dates don't announce themselves, so put every one in the same calendar you use for the operating rhythm.

Food safety sits alongside them and closes doors fastest: keep temperature logs daily, separate raw and ready-to-eat, stock handwashing stations, and name one person responsible per shift.

Rules vary by country and often by city, so confirm yours with the local authority rather than any general guide — including this one.

Standing out: your concept and the reason to choose you

Your concept is the promise: cuisine, price point, service style, room. It should be specific enough that a guest can describe it to a friend in one sentence, and consistent enough to survive a chef change.

The restaurant location strategy behind it matters just as much — whether the people walking past are the people your menu is priced for.

Then decide what that sentence is. Not a list of pleasant things, but one thing you'd defend even when it costs money: the best Neapolitan dough in three counties, open at 6 a.m. when nobody else is, a wine list nobody in town would attempt.

Filling tables: marketing you can sustain

Marketing that survives a busy month is marketing you can run in twenty minutes a week: a reason for someone to give you their email or phone number, one message a week to the people who already did, and an up-to-date Google profile.

Everything else — paid social, influencers, events — is optional, and judged on bookings you can trace. Pick two channels, run them for a quarter, and drop whichever you can't measure. Learn more about restaurant marketing strategies.