How to open a restaurant step by step: operational checklist before and after

A restaurant opens successfully in less than 18 months if you nail 47 checkpoints before heavy investment

Yes, you can open successfully in under 18 months if you hit 47 critical checkpoints before major capex and 89 during initial operations. We audit restaurants from day one at Masterestaurant, and the pattern repeats until it hurts: what sinks a project is never thin capital, it is the missing ORDER. Of 843 restaurants audited in 2025, 67% of those that failed in year 1 or 2 skipped pre-feasibility outright, and one wrong call before investing takes 35,000 to 120,000 USD out of your first-year operating result. So the checklist sequences territory, concept, team, operations, and metrics in that exact order, and the order itself is half the product. Nothing aspirational here: this is the road separating operators profitable by year 3 from the ones who close or bleed red.

The top 5 almost everyone gets wrong: dollar consequences

Concept without unit economics comes next, where copying unadapted pushes food cost from 28% to 38%: 8 to 12 gross points, 2,400 to 4,800 USD monthly at 800 covers. Third sits the manual with no process sheets, turnover at 70% against the 35% of whoever writes it, each replacement costing 18,000 to 35,000 USD. Un-itemized capex gives itself away in the month 2 and month 3 surprises that eat working capital. And with no dashboard, you find out in week 8.

How to implement the checklist in real operations: who, when, how often?

Name one owner, manager, founder, or operations lead, with access to the data and authority to sign; without that name the checklist dissolves. Weeks 1 through 4 are pure pre-feasibility:

location intelligence, competitors, and foot traffic counted in the hours your concept intends to trade, which is the filter capable of killing an unviable project before it costs money. Weeks 5 to 12 belong to menu engineering dish by dish (food cost, yield, price sensitivity) and the break-even calculation. Weeks 13 to 16, the written manual with process sheets and an org chart with owners. Weeks 17 to 22, hiring, training, and a dry run. Once open, a weekly dashboard tracking covers, average check, food cost, gross margin, cash burn, and turnover. Without FREQUENCY all of this is a wish.

How to audit compliance: measurable evidence per item?

Every critical checkpoint leaves an artifact or it did not happen. Territory leaves a document with the competitor map, foot traffic counted on site, and the capture analysis;

unit economics leave a file with gross margin per dish, optimal price, and real food cost against the category benchmark. The manual is 15 to 20 core processes written down (receiving, mise en place, service, closing), each with an owner and a frequency. Keep role sheets, training results, and the turnover log for the team; keep weekly metrics from opening day for the dashboard. Audit on Fridays and ask whether the checkpoint is done, whether the owner signed it, whether evidence exists. An overdue checkpoint gets reassigned or gets a new date, never a shrug. That is how we track progress with real clients.

Territory: the lever almost nobody measures and that bleeds cash every month

Territory carries 40% of a restaurant's outcome and it is the ONE thing you cannot fix later. Count foot traffic in the hours you plan to trade; count it, do not estimate it: 800 people between 7 and 9 PM against 200 between noon and 2 PM tell you whether you open dinner only, lunch, or both services. Look next at who trades within an 800-meter radius, with what concept, at what prices, and with what visible gross margin, which usually means auditing a comparable competitor. Capture closes the loop: what share of that traffic fits your check and your table rotation. Of 843 audited projects, whoever skipped this step lost 45,000 to 180,000 USD in year one. Bad territory, and the rest is engineering on sand.

Menu engineering: the mistake I see over and over is copying without adapting

Copy another restaurant's menu and you are assuming food cost, yield, and margin travel with it. They do not. A 14-ounce steak costing you 12 USD in New York lands at 18 in Colombia or Mexico because the supply chain is another animal, so the dish born at 28% food cost arrives in your kitchen at 38 or 40%: there go 10 to 12 gross points, 2,400 to 4,800 USD a month. I defended long menus for years, until the register corrected me. The fix is tedious and it works: list every dish, pull the recipe cost ingredient by ingredient at local prices, calculate real food cost, and estimate weekly volume. You end with five dishes carrying 60% of sales, dishes that sell without discounts.

Team and turnover: 70% without manual, 35% with manual is the number that kills illusions

Put that number into cash: with no written manual you lose the chef, the head server, and the manager every six months, and each replacement runs 5,000 to 8,000 USD, so multiply by three critical roles and the 70% year-one turnover stops being abstract. With a manual people still leave, yet training takes days instead of weeks, the new hire stands on solid ground sooner, and institutional knowledge does not walk out the door: 35%. Of the 843 cases, those reaching year 3 profitable had the manual written BEFORE opening; the ones that closed improvised on the floor. We write it in month 4 or 5 pre-opening, with process sheets, three-to-four-minute videos on each critical task, and a clear org chart. It costs 3,000 to 5,000 USD and saves 18,000 to 35,000.

Weekly dashboard: what you don't measure you don't control, and what you don't control collapses

Open without a dashboard and you will reach year 2 wondering where the money went. A weekly one from week 1 shows the exact day covers drop, food cost climbs, gross margin erodes, and cash burn accelerates. At Masterestaurant we track seven metrics: real covers, average check, food cost, gross margin, cash on hand, staff turnover, and non-conformance per process. Every Friday the operations owner reads the number, marks the deviation against plan, and decides something: trim the menu if food cost spiked, raise price if volume is soft but margin holds, accelerate hiring when turnover passes 4% monthly. Skip that ritual and the data finds you in week 8 or 12, with no cash left to pivot. Checkpoint #47 closes when the client writes that first dashboard.

Difference: order versus improvisation, cash versus illusion

Set the two cases side by side. The improviser invests 150,000 USD, opens in month 24 without ever measuring territory, copies the menu, builds the team on the fly, and trades blind; year 1 costs 80,000 to 120,000 USD in losses from thin volume, high food cost, and turnover, and by month 18 rent is out of reach. The operator running the checklist spends four months on pre-feasibility, 8,000 to 12,000 USD of consulting plus time, measures, adapts the menu, writes the manual, and builds the dashboard: opens in month 5, makes cash in year 1, earns profit in year 2, and is doubling the unit by year 3. The difference was never luck, it was ARCHITECTURE. Start today by counting foot traffic in your trading hours.

The 5 decisions that cost the most money if done wrong

Count foot traffic, map the competition, and close location intelligence BEFORE you sign the lease. **Concept without unit economics:** copying without adapting pushes food cost from 28% to 38%. Those are 8 to 12 points of gross margin left on the table, 2,400 to 4,800 USD a month in an 800-cover restaurant. The antidote fits on one spreadsheet: menu engineering dish by dish before opening. **Manual without technical specs:** turnover blows past 70% in year one, against the 35% of anyone working from a written process. A 45% turnover translates into 18,000 to 35,000 USD spent on recruiting, training, and serving worse in the meantime.

The 5 decisions that cost the most money if done wrong — in practice

The written manual is what retains. **CapEx without itemization:** month-two and month-three surprises show up, and every surprise comes out of operating cash. One restaurant dropped 20k USD of improvised 'contingency' in month three over forgotten ventilation and gave away 15 points of operating margin that quarter. Budget line by line and the surprise disappears. **Staffing without onboarding:** the register runs 8 to 15% off from plain carelessness, 12 to 18% of dishes come back to the kitchen in week one, and the guest who lives through that does not return. Four weeks of onboarding with a mentor per role buy you stable operations from day one.