How to open a restaurant step by step: before and after applying a pre-feasibility framework

Chapter 1 — Why the opening sequence decides the outcome before the concept does

Validate the territory and the unit economics on paper before you sign anything: that single reversal of order explains most of the early closures I see across restaurant portfolios. A Wendy's franchisee must document one million dollars in liquid capital and five million in net worth before a single unit gets approved (Swoop / Wendy's FDD, 2025), and that filter does not exist because the business is expensive: it exists because the system learned, through bankrupt units, that working capital runs out long before enthusiasm does. Whoever opens alone skips that filter and pays for it in month eight. Diego F. Parra insists from Masterestaurant on a point that architects and designers find uncomfortable: the menu and its costing get written BEFORE the floor plan, because the plan costs whatever the margin can afford, never the other way around.

Chapter 2 — CapEx is not an amount, it is a calendar

Mistiming four hundred thousand dollars of construction turns an investment into short-term debt at commercial rates, and no menu rescues the project after that. Market reference helps size it: the SBA issued 57,362 7(a) loans worth over 31.1 billion dollars in fiscal year 2024, with an average ticket close to 542,000 dollars per operation (U.S. Small Business Administration, 2024). Read it backwards. A six-month gap between disbursement and actual opening burns, in interest and dead rent, something close to the full balance of one of those loans. My criterion, and here I am blunt: not one peso of construction leaves the account until three things are measured — foot traffic counted by hand, average check of the direct competition, and a menu costed with quoted supplier prices, not estimates.

Chapter 3 — Prime cost: the number that kills restaurants while the owner watches the rent

Nobody closes because of rent. They close because of prime cost — food plus labor — which in a healthy operation lives between 55% and 60% of sales and which in an improvised opening shows up at 68% by month three, when the lease is signed, the staff is hired and suppliers already extended credit. Translated to an operator's desk: a project costed with 2023 food cost enters the 2026 market with eight points less margin. The fix is not raising prices in month four; it is engineering the menu from day zero, knowing that 32% food cost per dish is the ceiling, never the target.

Chapter 4 — Revenue bands: the same decision costs different things

Each annual revenue band punishes a different mistake, and confusing them is the costliest error of all. Below 500,000 dollars, the owner IS the operation: one point of food cost equals their own salary, and any management software competes with a prep cook's wage. Between 500,000 and one million the first managerial hire appears, and with it the classic lag, because the salary lands whole while productivity takes a quarter. Above one million, royalties stop being a contractual footnote: between 4% and 8% of gross sales in restaurant franchises (Toast, 2025) means 40,000 to 80,000 dollars a year that now weigh on cash flow. Past five million the multi-unit logic rules — the average franchisee runs five locations (FRANdata) — and above ten million the problem moves to treasury and consolidated debt.

Chapter 5 — The high end: celebrity chefs, large formats and costs nobody budgets

Above five million dollars in revenue, a celebrity-chef or large-format themed restaurant stops behaving like a restaurant and starts behaving like an entertainment producer with a kitchen inside. What breaks the books there is not food — usually negotiated better than in any small venue — but structural payroll, the upkeep of the set, and the traffic drop when novelty fades, typically between month fourteen and month twenty. A large-surface format paying 12% royalty with heavy structure needs sustained volume, not an opening spike. If the project depends on the press continuing to talk, it is not a business: it is a campaign with a ten-year lease.

Chapter 6 — What would happen if the small venue copied the big one's playbook?

It usually ends with chain-sized structure and corner-shop sales. Picture the full exercise:

a venue billing 380,000 dollars a year hires an operations manager, builds a weekly reporting system and adopts the same centralized supplier scheme an eight-unit group uses. That structure adds roughly 14 points of expense over sales that scale dilutes for the big operator and does not for the small one. By month five the owner cuts — and cuts the kitchen first, the one thing sustaining repeat visits. The small band keeps its own recommendation, and it is not a watered-down version of the large one: measure by hand, price properly, work the peak shift yourself, and never fund structure before volume pays for it. That discipline is what makes a second location possible. Franchised QSR reached 321.8 billion dollars in 2025, growing 5.4% over the prior year (International Franchise Association, 2025), and that figure rewards documented replicability rather than size: written processes, per-dish costing, measured portion control. What scales is the manual, not the square footage.

Chapter 7 — Replicability: the only thing that turns an opening into a portfolio

A restaurant that cannot describe how it produces its margin is not replicable, and without replicability there is no second location and no reasonable credit. The hard evidence sits in multi-unit control: according to FRANdata, multi-unit operators control 54% of franchised units, which means the sector systematically rewards whoever documented the operation before growing. Portugal leads Spanish franchising abroad with 176 networks and 2,632 establishments (AEF, 2025), and those networks did not travel on charisma: they traveled with a manual, a standardized recipe and a technical spec sheet. This is where SATE Institute and Masterestaurant converge on the instrument: training operators to write down their process creates formal jobs that survive year one, while financing construction without process merely rotates venues and returns workers to informality.

Chapter 8 — The four documents you must have signed before the keys

Start with the one almost nobody demands: the menu costed dish by dish with current quotes from at least two suppliers. Next, the monthly break-even with payroll, rent and utilities charged to the operation — never to the dish — because that number tells you how many covers a day you need before signing a single square meter. Third, a traffic count at the site, taken across two time slots on two different days, contrasted with the average check of the three nearest competitors. And fourth, the working-capital cushion for six months of operating below break-even; with an average ticket of 542,000 dollars in SBA 7(a) loans during fiscal 2024 (U.S. Small Business Administration, 2024), reserving 15% to 20% of the credit for that cushion is what separates a bad quarter from a closure. When all four exist, sign the lease. Not before.

Chapter 9 — Where the project actually breaks

The mistake sits not in the construction budget but in its SEQUENCE. A 400,000-dollar CapEx mistimed becomes short-term commercial-rate debt; the same CapEx disbursed after territorial validation behaves like investment with an estimable return horizon. When the market reference in SBA 7(a) loans averages 542,000 dollars per operation (U.S. Small Business Administration, 2024), the gap between both sequences equals an entire loan balance. Nobody fails because of rent. They fail on prime cost, that sum of food and labor cost which in a healthy operation lives between 55% and 60% of sales and which in an improvised opening shows up at 68% in month three, when the lease is signed, payroll is hired and suppliers have extended credit.

Chapter 10 — Where the project actually breaks — in practice

Fixing it there costs three times what budgeting it would have. Royalties get signed without simulation. An operator projecting break-even at a 4% royalty and signing at 8% does not have a contract problem: the financial model was wrong from the first cell. Replicability cannot be improvised at unit two. According to FRANdata, multi-unit operators control 54% of franchised units. That profile gets built from unit one or not at all: whoever opens the first location without a manual is choosing, unknowingly, to stay a single unit. Territory is chosen by eye and paid for out of the register. Territorial pre-feasibility with location intelligence layers — competitive density, household income, daytime work flow, accessibility — shifts the project's supportable average check by ranges no later lease negotiation offsets.