How to open a restaurant step by step: traditional method vs Masterestaurant method

Where does each method start, and which one sequences the irreversible decisions correctly?

The traditional method starts with the lease and the Masterestaurant method starts with the arithmetic, and that single inversion of order decides your cash position at month 12.

The Colombian group that sat me down in front of its fifth opening on 31 July 2026 had budgeted 310,000 USD of CapEx and already signed a ten-year lease; the sales projection had been built on the average of the other four units, in a shopping center with 40% less foot traffic than the plaza hosting their best location. Nobody counted that traffic. It was assumed. Signing the lease before knowing your target average check turns the most expensive variable in the business into a fixed input, and you pay that number for 120 straight months with no way to renegotiate it downward when the location does not sell. The Masterestaurant method WINS on sequence: it accepts the rent last, once you know what you can pay.

Costing the menu before or after building the kitchen: six margin points at stake

Six points of contribution margin separate a menu costed before construction from one costed afterward, and that is the second axis where the two paths genuinely diverge. When costing arrives after the kitchen is installed, you sell what the equipment allows: a mix carrying a 35% weighted average food cost because the grill you bought cannot deliver the 29% dish. Costing first inverts the dependency and lets you buy the kitchen your profitable mix demands. On annual sales of 900,000 USD, those 6 points are 54,000 USD of margin that appear in no construction report and no equipment quote, because they were lost in a decision made six months earlier. The MR ceiling remains 32% food cost per dish as a maximum, never as a target. On this criterion the traditional method loses by design, not by execution.

Financing punishes each path differently: 23-28% charge-off on restaurant SBA loans

Banks have already put a number on the risk of opening badly: the charge-off rate on SBA restaurant loans runs from 23% to 28% according to PeerSense (SBA Default Rates by Industry 2026), while default across restaurants and food service under normal conditions sits between 12% and 15% according to Crestmont Capital. Set those figures against the 9,9% average SBA franchise loan default between 2010 and 2021 according to U.S. Small Business Administration data: nearly one in ten, yet less than half the pure restaurant charge-off. What makes money cheaper is not the concept or the brand, it is the verifiable financial model you bring to the table. The traditional method shows the bank a projection built on internal averages; the Masterestaurant method shows unit economics with counted foot traffic, target check and break-even point. The second one WINS, and the rate shows it.

Break-even: 18-36 months on one path, counted from day one on the other

A fast food restaurant takes between 18 and 36 months to recover its investment according to BusinessDojo (Fast Food Break Even 2025), and a Chick-fil-A franchise between 4 and 6 years according to Restaurant Velocity (Most Profitable Franchises 2025). Those ranges are not destiny, they are the consequence of the order in which you made your decisions. The traditional method discovers its break-even point once it is already trading, because rent, payroll and utilities arrived as accomplished facts and sales must be bent around them. The Masterestaurant method calculates break-even BEFORE signing and uses that figure to filter locations: if the plaza cannot support the sales your break-even requires, you do not sign. Payroll, rent and utilities never load onto the plate, they belong to the break-even calculation. The practical gap between both paths, across the groups I have worked alongside, is measured in quarters of cash, not decimals of margin.

The 310,000 USD case: what happens when foot traffic drops 40% and nobody measured it

Take that fifth Colombian opening and run the scenario to its end. If foot traffic in the new shopping center is 40% lower than in the plaza used as reference, and conversion and check hold steady, projected sales fall short by the same proportion: against 900,000 USD expected annually, roughly 540,000 real. The ten-year lease does not drop, the 310,000 USD of CapEx is already committed and the kitchen was sized for a volume that will never arrive. That location will not fail in year one — the first-year failure rate was just 0,9% in 2025 according to Datassential, the lowest since 2018 — but it drains cash from the other four units for the full ten years of the contract. It is the worst possible outcome: it neither dies nor wins. As Diego F. Parra, founding consultant of Masterestaurant, explains, a location that merely moves money costs more than one that closes fast.

What costs the same on both paths and what does not: construction, equipment, and the ten-year variable?

Construction and equipment cost virtually the same under both methods, and that is precisely where nearly every expansion committee concentrates its debate. The figure that changes the outcome is not in that column.

It sits in the rent, the only CapEx line you pay across 120 months with no chance of renegotiating downward when sales fail to appear; suppliers get swapped, menus get redesigned, staff rotates, the lease does not. The traditional method treats rent as an input and everything else as adjustment variables. The Masterestaurant method reverses it: fix the target average check and the weighted food cost first, derive the sales you need, and from there comes the rent ceiling your model tolerates. Any location above that ceiling gets dropped without debate. The MR method WINS on the one variable that admits no later correction. Chains that sustain net openings year after year are not doing it on real estate instinct.

Scale and survival: why chains that grow well do not improvise the sequence

Jersey Mike's closed fiscal 2025 near 3,300 stores, with more than 250 net openings and system sales above 4,000 million USD according to Restaurant Dive; Chipotle holds a net unit growth target of 8% to 10% annually according to CRE Daily; Subway ended 2024 with 19,502 US locations according to QSR Magazine. None of the three signs a lease before running the model. And the most misread figure in the industry is survival: 51% of restaurants are still operating after five years according to the UC Berkeley 2014 study, which means the half that closes does not close over bad food but over arithmetic solved too late. Repeating the method at scale demands that the sequence be explicit, not intuitive. If you are opening your first location with your own capital and no partners demanding returns on a date, the traditional method will work for you: the penalty of the wrong order lands in your own pocket and you will learn from it.

What to choose by profile: single-unit operator, four-unit group, or franchise expansion?

If you run a group of three to five units and the sixth is financed by the cash of the others, the Masterestaurant method is not preferable, it is mandatory, because one location draining cash for ten years takes consolidated profitability down with it.

And if you plan to franchise, the sequence stops being your decision and becomes a transferable asset: the franchisee buys the order, not the recipe. My recommendation for the hospitality group leader in 2026 is concrete: before the next letter of intent, count real foot traffic for two weeks across the dayparts that matter to you and calculate the rent ceiling your break-even tolerates. That exercise costs under 2,000 USD and decides 120 months. The first difference is sequence, not effort. Under the traditional method the site arrives before the arithmetic, so the financial model is born with a constraint already imposed —the signed rent— and everything else bends around it.

Where the two roads genuinely diverge?

Under the Masterestaurant method rent is the last variable accepted, because it is the one CapEx line you pay for ten years without any chance to renegotiate downward when the unit underperforms.

The second sits in menu costing. Costing after you build the kitchen forces you to sell whatever the kitchen allows; costing first lets you build the kitchen a profitable mix demands. A menu with a 29% weighted food cost in casual format carries roughly 6 more points of contribution margin than a 35% one, and those 6 points on 900,000 USD of annual sales are 54,000 USD that decide whether the unit repays CapEx in two years or four. The third difference is information speed. A group reading its P&L 35 days late corrects a month behind reality; a group reading weekly prime cost fixes on Tuesday what broke on Saturday.

Where the two roads genuinely diverge — in practice?

According to Hudson Riehle, senior vice president of research at the National Restaurant Association, labor and input cost pressure is now the sector's dominant operating constraint, and in that environment reporting lag is paid in cash, not in theory.

The fourth —the one I argue about most in board rooms— is how you deal with restaurant investors. The traditional method asks for money against a dream and a projection; the Masterestaurant method asks against a model with auditable assumptions and a declared MTIE. The second raises cheaper capital, and in expansion the cost of capital belongs inside unit economics, not in some separate finance conversation. There is a real tension between the two positions and I will not hide it: the traditional method opens faster. Five months can pass between signing a decent site and opening the doors, while filtering ten candidates and killing six adds six to eight weeks.

Where the two roads genuinely diverge — key points?

That delay costs money if you are paying warehouse rent or already carry payroll. The bridge is simple: filter sites IN PARALLEL with menu costing and design, never in series.

You then pay the extra weeks in management effort rather than in calendar.