How to open a restaurant step by step: the mistakes that burn capital vs the MASTERESTAURANT method

Step 1: calculate the margin on ONE dish before you look at a single site

The first deliverable of an opening is not a floor plan, it is a spreadsheet showing one dish's contribution margin after waste, and that spreadsheet gets finished before you visit your first site. You verify it like this: pick the dish you expect to sell most, weigh every input in grams on a scale, load in the real trim and cleaning loss —on protein that runs somewhere between 12 % and 25 % depending on handling— and divide the resulting cost by the menu price. If that food cost clears 32 %, the dish does not make the menu; 32 % is a ceiling, not a target. In the MASTERESTAURANT method, payroll and rent never get charged to the dish, because they distort the decision you need to make today: that fixed cost gets settled at the break-even point in step 2. It is done when you hold the margin in currency, not in percentage, for your ten anchor dishes.

Step 2: set break-even in covers per day, not in monthly sales

A useful break-even point gets expressed in covers per day and per service window, because a monthly money target tells nobody whether the seating actually supports it. Add up rent, base payroll, utilities, insurance and construction amortization; divide that by the weighted average contribution margin you produced in step 1; the result is how many dishes your kitchen must send out every month just to avoid losing money. Divide by operating days, then by real selling windows. This is where projects die: that owner with two healthy locations arrived holding a signed lease that demanded he sell 1.9 times what his seating allowed during the lunch window. The deliverable is a single figure —«148 covers a day»— that you can hold up against chairs, turns and service hours.

Step 3: rule out sites with the calculator before ruling them out on foot

A site gets ruled out in the spreadsheet, not on the visit, and with the method properly applied 70 % of the options fall away on the first pass. The working rule is simple: rent plus common charges should not exceed 8 % to 10 % of the projected sales from step 2, and you do not invent that projection, you count it. I have people count foot traffic by hand, two weeks, during real selling windows, with a clicker. There is one legitimate exception to the ceiling, and here sits the paradox that trips up first-timers: the expensive site is usually the cheap one. A space at 12 % rent that fills every day returns more than one at 6 % on a dead street. The difference is that the exception gets taken with counted data, never with a hunch on a Sunday afternoon.

Step 4: negotiate the lease the way you negotiate a risk contract

The lease carries the highest reversal cost of the whole opening, which is why it gets signed last, with three clauses fought for in advance. First, the grace period during construction: if you build for three months, do not pay for those three months. Second, an annual increase that is indexed AND numerically capped, because an open-ended raise turns a healthy business into a sick one by year three. Third, the exit: who keeps the improvements, and how fast you can walk without surrendering the full deposit. Add the working capital reserve, the line item I see underestimated more than any other: six months of payroll and rent in cash, separate from the construction budget. Funding the build with money meant to operate is the most common way to close in month fourteen with a kitchen that was running perfectly well.

Step 5: paperwork belongs here, not earlier, because its cost is predictable

Health permit, zoning approval, fire and safety clearance, trademark registration and tax registration sit in fifth place for a measurable reason: they cost PREDICTABLE time and money, and the predictable should never lead a sequence of decisions. Zoning is the one ordering exception, since it is worth confirming before you sign; the rest run in parallel with construction. Budget calendar, not just cash: across most Latin American markets the full package eats eight to sixteen weeks, and every week of delay with the lease already running is rent paid without any till to carry it. The deliverable is a schedule with an owner and a date per filing, taped up on the job site. Let me admit a mistake I made for years: I put trademark registration last, until a client lost his restaurant's name three months after opening.

Step 6: size the payroll against the industry, not against your enthusiasm

Headcount gets built from the covers-per-window figure in step 2 and then checked against what the sector actually looks like, which is labor intensive: the United States restaurant industry employs 15.9 million people and added 200,000 jobs over the year, according to the National Restaurant Association (2025); in Mexico it is 2.1 million direct jobs and 3.5 million indirect ones, according to CANIRAC (2024). Those figures carry an operating message: labor is your second largest line and it gets decided on the menu, not in the interview room. A 47-dish menu demands stations, waste and hours that a 22-dish menu never asks for. Trim the menu and you trim payroll without letting anyone go. It is done when you hold headcount by shift, with contracted hours and cost per productive hour, squared against the sales of the window that funds it.

The mistakes that sink an opening, and how you avoid them

Four mistakes account for most early closures, and every one of them is a sequencing failure. One: signing the lease before step 1, which is exactly what happened to the owner with the architect's plan and the 47-dish menu. Two: funding construction out of working capital, an error that stays invisible on the P&L until month four. Three: copying a model without copying its revenue structure; some concepts live off one specific channel —drive-thru accounts for 60 % of Chick-fil-A's sales and 90 % of Dutch Bros' revenue, according to QSR Magazine— and without that channel the whole model collapses. Four: reading aggregate sector growth as though it were insurance. An expanding industry never rescues a badly costed location; it only makes the failure harder to notice among so many openings.

Closing checklist: how to know everything is right before you open the door

You know the opening is ready when you can answer six things with a number rather than an impression. One: the margin in currency on your ten anchor dishes, with real waste measured on a scale. Two: daily break-even covers by window, held up against chairs and turns. Three: rent plus common charges inside 8 % to 10 % of projected sales, or the exception documented with a two-week foot traffic count. Four: six months of payroll and rent in cash, untouched, outside the construction budget. Five: every filing with an owner and a date, and the trademark registered. Six: two full closed-door trial services, with tickets timed and the real cost of those services compared against the theoretical cost from step 1. If even one of the six lives in your head instead of on a page, do not open yet. Fix that one and count again.

Where an opening actually breaks?

The gap between the two routes is not the cook's talent, it is SEQUENCE. In the usual route the lease fixes the variables — seats, hours, extraction, neighbours — and the concept adapts to whatever remains;

in the method, the concept and its numbers decide which sites qualify, and 70 % of the options fall out on the first spreadsheet. Discarding fast and cheap is the most profitable financial skill in an opening. One paradox trips up first-time owners: the expensive site is often the cheap one. A space at 12 % rent with footfall that fills it every day can beat a 6 % site on a dead street, and that is the legitimate exception to the cap. The difference is that the exception gets taken on foot traffic counted by hand over two weeks, not on the broker's instinct. With no count, the 8 % cap rules. The costliest miss I keep finding in site due diligence is not rent, it is electrical load and extraction.

Where an opening actually breaks — in practice?

A space without the amperage the planned kitchen demands can add 15,000 to 60,000 dollars of upgrade work — sometimes months of utility paperwork on top — a line that rarely appears in the first budget and gets paid out of working capital.

Restaurant unit economics come down to three figures nobody can dress up: food cost per dish with real waste, labour cost per service hour, and rent over sales. The first two together are prime cost, and above 60 % the operation cannot absorb a slow month. Diego F. Parra makes the same point in every Masterestaurant audit: no sales volume rescues a negative unit margin, it only makes it bigger. Scaling changes the paperwork rules. Opening one is a permit; opening five is a permit system, and the group that documents recipes, suppliers and timings from the first opening measurably shortens the second cycle. Documentation is not bureaucracy: it is the only thing that turns a restaurant into an asset you can transfer to an investor or a franchisee.