Growing a restaurant: the numbers before and after the Masterestaurant method

Why does the second location sink the first one?

The second location sinks the first one when expansion CapEx exceeds twelve months of consolidated EBITDA and the mother store's cash ends up funding the new build.

Sector numbers explain why that mistake costs so much: net margin at a full-service restaurant runs between 3% and 5%, and fast casual sits at 6% to 9%, according to Peppr POS in its 2025 margin guide, while Statista places the sector range between 3% and 9%. At a 4% margin, every dollar of cost overrun demands twenty-five dollars of extra sales to recover. And the overrun always arrives, because a hospitality build rarely closes on budget. Before signing the lease on your second unit, calculate how many months of consolidated EBITDA that opening is eating; if it passes twelve, you are not growing, you are leveraging the business that already works against a promise. Talking about «sector growth» stopped making sense in 2025, because the market split into two branches heading opposite ways.

A K-shaped market: growth is no longer an average

Technomic measured, in its Top 500 published via Restaurant Business, that the 250 largest chains grew sales by 3% while the next 250 fell 6.2% that same year. Nine points of separation between two halves of one market is not statistical noise: it is a structural gap, and the independent operator planning expansion off the sector average is using a number that describes nobody. Datassential counted more than 860,000 restaurant locations in the United States as of November 2025, an all-time record, which means the problem is not weak demand but excess supply chasing it. The decision this figure triggers is uncomfortable and clear: if your concept lacks a defensible edge in cost or in brand, scale will expose you rather than protect you. Replicating a unit costs far more than operators budget, and franchises publish their figures under legal obligation, so they work as a hard benchmark.

What replicating a unit actually costs?

GrowthFactor analyzed 149 FDD documents in 2026 and found an average initial investment between 598,000 and 1.6 million dollars for a fast-food franchise, with an average fee of 35,000 dollars.

At the top end, McDonald's 2024 FDD declares between 1.47 and 2.73 million dollars per location according to Franchise Chatter, and Taco Bell reports a range of 1,584,750 to 3,980,200 dollars in its Item 7. Those brackets —nearly three million dollars wide in Taco Bell's case— are a confession that not even the most disciplined chains on the planet know in advance what an opening will cost. If McDonald's admits a 1.26 million dollar range, your budget without a contingency line is not a budget. No amount of operational effort compensates for a badly chosen format, and profitability by model proves it brutally. Peppr POS reports net margins of 10% to 30% for delivery-only concepts, against the 3%-5% of traditional full service.

Format outranks effort

A full-service operator running flawlessly, with food cost under control and excellent table turns, still lands below the floor of a mediocre delivery concept. A genuine concession belongs here: for years I recommended squeezing the existing location before touching the format, and in many cases that was right because the operation was dirty. But once the operation is clean and margin still sits at 4%, the ceiling is not the team, it is the model. Before opening a second unit identical to the first, ask yourself whether the correct replica is the same format or one with a different cost structure. Latin America offers today the best ratio between growth and saturation, and the projections back it up. Market Data Forecast values the Latin American fast-food market at 61.49 billion dollars in 2025, on a path toward 94.98 billion by 2034 —a 54% jump in nine years.

Where the money is growing: geography and category?

Mexico concentrates much of that momentum: Nation's Restaurant News reported in 2025, citing data from Wendy's expansion, a QSR burger market worth 2.4 billion dollars in 2024 growing at 14.3% annually over five years.

Compare that with the global fast-food market, which Market Research Intellect projects at 520 billion by 2033 with a CAGR of barely 4.7%. Three times the global pace in one specific category of one specific country. The Masterestaurant reading of these numbers is blunt: geography forgives execution errors that no mature market ever will. Starbucks closed 2024 with 38,587 locations worldwide, according to Restaurant Business, and that number usually reads as proof that replication works. It does work, yes, but look at the time denominator: the chain took more than five decades to get there, and every stretch of growth rested on an operating manual that allowed openings without the founder setting foot in the store.

Scale exists, but it takes decades

That is the point independent operators skip. Diego F. Parra insists, based on Masterestaurant's field work with operators across 43 countries, on a simple prior test: the mother store must hold its EBITDA corridor for three consecutive months with the founder off the floor. If it cannot, what you own is not a replicable business but a well-paid job. And a job does not multiply, it wears out. Expansion failure almost never announces itself with falling sales, and that is precisely the trap. A group can bill 2.4 times the prior year after two openings and watch consolidated EBITDA drop from the 14.8% the original location produced to a 4.1% that no longer covers the debt. It did not lose the business: it diluted it. With sector margins of 3% to 9% per Statista, dilution leaves no room to correct along the way.

Survival depends on cash, not on sales

Here is the counterfactual question every operator should ask before signing: if the new location takes eighteen months to reach break-even instead of the six your plan assumes, how many months of the mother store's payroll get consumed in that gap? If the answer passes three, the opening is already doomed and you have not laid a single brick. Three numbers decide whether your expansion adds or subtracts, and each carries its own action. First: twelve months of consolidated EBITDA as the ceiling for expansion CapEx —if the construction budget passes it, do not open, and if you already passed it, freeze the decoration line today and renegotiate the timeline with your contractor. Second: 23% contingency on the budget, because average construction overrun in hospitality hovers around that figure and always surfaces in electrical work and exhaust systems; park that money in a separate account before the first invoice, not once the problem lands.

The 3 numbers you should tattoo on yourself

Third: 4% net margin is the point where full service stops tolerating mistakes, per the ranges from Peppr POS and Statista; if your mother store sits below it, fix food cost and labor productivity before looking at a second lease. Start this week with the second one: open the account and move the money. The trigger for opening. Under the intuitive model the signal is external — a site appeared, a partner offered capital, a competitor closed — while the measured method waits for an internal signal: the mother unit hit its EBITDA corridor and released the founder. One logic opens when the market invites; the other opens when operations are ready, and that inversion of causality explains much of the 60% of restaurants that never reach year three, per National Restaurant Association data. How CapEx gets treated. An expansion budget without a contingency line is not a budget, it is a wish; average construction overruns in hospitality hover near 23% and always surface in the least visible line item, which is electrical and ventilation.

Five differences that change the outcome

When that overrun is funded from the mother unit's operating cash, you are not financing an opening — you are defunding the business that works to pay for the one that does not yet. Ownership of knowledge. A replicable operations manual converts judgment into procedure, and that conversion is what lets unit number four resemble unit number one. Without it, each opening reinvents the business with new staff, and the result is quality drift that guests notice before owners do, because a guest compares today's visit against one made three months ago at another address. How the site gets read. Serious due diligence blends hard location intelligence — pedestrian flow by daypart, competitive mix within 400 meters, corridor average check — with cannibalization arithmetic: how much sales volume the new unit steals from the old one. I have reviewed expansions where unit two captured 31% of its revenue from unit one, and the group celebrated growth that was really a partial relocation of guests.

Five differences that change the outcome — in practice

The stopping rule. Almost no growth plan carries a mechanism for NOT opening, and there lies the fatal asymmetry: every system pushes forward, nothing brakes. MTIE fills that role. It measures the incremental margin the expansive investment returns to the group, net of cannibalized sales, and its 18% threshold does not get renegotiated inside an optimistic committee.