Scaling a Restaurant: the numbers that decide, traditional method vs Masterestaurant method

What does a second location cost today, and how long until it pays back?

Opening a second location in Latin America starts at a median of 285,000 dollars per full-service unit, and that investment does not come back in under 18 months even in the fastest formats.

The hard reference comes from quick service, where BusinessDojo places break-even between 18 and 36 months; in full service, with heavier buildout and kitchen, the range stretches. Established franchises say it even more plainly: Restaurant Velocity calculates 3 to 5 years for a Domino's with an investment of 156,000 to 682,000 dollars, 5 to 7 years for a McDonald's at 525,000 to 2.7 million, and 4 to 6 years for a Chick-fil-A. The concrete decision that follows: if your financial plan promises to recover the investment in 12 months, the plan is wrong, not the market. Revenue measures demand; contribution margin per service hour measures the ability to generate cash with the structure already in place, and only the second one authorizes an expansion.

Contribution margin per service hour, not revenue

A location billing 95,000 dollars a month with a 68% prime cost is not ready to scale, it is ready to be fixed: against an industry reference of 60% to 65%, those three to eight extra points are 2,850 to 7,600 dollars a month evaporating before rent gets paid. Multiply that flaw by two locations and the second one is born eating the first one's cash. I got this wrong for years: I judged health by monthly net margin because it was the number the owner understood and the accountant already had ready, and that indicator mixes operations with financing. The decision: do not sign the second lease until the first one holds prime cost under 62% for three straight months. Chipotle, with a mature operating machine and its own capital, targets NET unit growth of 8% to 10% a year, according to CRE Daily, and that self-imposed ceiling is the lesson for any restaurant group with three to ten locations.

Opening pace: what Chipotle's 8%-10% teaches

If a company with that financial backing will not let itself double its footprint in a year, a four-location group planning three openings in twelve months is not scaling: it is gambling. Translated to your reality, 8% to 10% on four locations is less than half a unit a year, which in practice means one opening every two years with your current management team, or one opening a year if you first build a second line of command. Nation's Restaurant News confirms it from the other side: the 500 largest chains concentrate net unit openings in the United States, because they have the bench, not because they have the appetite. CAPEX takes the spotlight —buildout, kitchen, furniture— and working capital ends up as the leftover, which is exactly the reverse of how it should be planned.

The working capital nobody budgets, and it sinks the new location

A full-service location with 285,000 dollars invested and break-even at 24 months has to cover the ramp-up gap: during the first six to nine months the new site rarely clears 60% of the mother location's sales, while payroll and rent run at 100% from day one. With opening costs 18% higher than in 2023 and an average check that moved barely 7%, you pay for those scissors with reserves, not with optimism. A rule I apply without exception: set aside the equivalent of six months of the new location's fixed costs, outside of CAPEX and outside the mother location's operating cash. Without it, the expansion does not fail on concept, it fails on liquidity in month seven. The density figures of the big networks explain why replicating their model without their structure is the costliest trap in this business.

Network density: why the Domino's and KFC model does not copy over

Yum China reports 12,640 KFC stores in China as of September 2025; Domino's runs close to 7,000 locations in the United States and roughly 14,500 abroad, per Quartr; Starbucks closed fiscal 2025 with 8,011 stores in China, according to Statbase. Those networks work because their marginal cost per new unit falls with scale: centralized purchasing, a brand already installed, small formats and a manual a new manager executes in weeks. Your four-location group has none of those three levers, so your cost per new unit does not fall, it climbs. Operating decision: before opening, measure how much your input cost drops when you double purchase volume; if it does not drop at least 4%, you are not scaling a network, you are opening loose restaurants. Franchising means selling a system; opening your own means running a restaurant, and confusing those two trades costs you the estate.

Franchise or open your own: two businesses under one logo

Spain, a mature market, had 1,384 franchise networks in 2024 with 82.7% of domestic origin, according to the Spanish Franchise Association report, and those networks live off royalties and fees, not off plate margin. Mexico, with more than 428,000 establishments per CANIRAC, shows the opposite end: an enormous universe of independent operators where the brand is rarely monetized. The bridge between both ideas is uncomfortable but true: you cannot franchise what you have not documented, and you cannot document what still depends on you being in the kitchen on Fridays. What you can do today: time how many hours a week you spend on decisions a manual would settle. If it runs past ten, your system does not exist. These benchmarks read differently depending on size, and applying them flat is the most repeated mistake. In the SINGLE location of 40 to 80 seats, forget Chipotle's 8%-10% and work one figure: prime cost under 62% for six straight months, with contribution margin per service hour measured in both the strongest and the weakest shift.

How to read these numbers in YOUR operation: three scenarios?

In the MID-SIZED group of two to five locations, the governing number is the six-month fixed-cost reserve for the new site on top of that 285,000-dollar median, with break-even planned at 24 to 36 months, not 12.

In the GROUP of six or more, marginal cost per unit enters: if doubling purchase volume does not cut inputs by at least 4%, the scale is nominal. The diagnostic framework Diego F. Parra applies at Masterestaurant orders these three readings before anyone touches the expansion plan. The figures in this analysis are public and verifiable, and it is worth saying with equal clarity what they are NOT. The recovery ranges come from Restaurant Velocity and BusinessDojo on United States franchises and quick service; the net growth rate is Chipotle's corporate guidance as reported by CRE Daily; network density comes from Yum China, Quartr and Statbase reports; the market data, from the Spanish Franchise Association's 2024 report and from CANIRAC in Mexico.

Where these benchmarks come from and how far they reach?

None of it comes from a sample of ours or from an audit we ran. Two limits that matter when you use this table:

the big-franchise paybacks assume an installed brand and centralized purchasing an independent does not have, and figures from the United States and China do not transfer straight across to Latin American labor cost or rent. Use them as a reference ceiling and adjust with your own last twelve months of income statements. The first difference sits in what counts as sufficient evidence to open. The traditional method reads revenue as a health signal, when revenue only measures demand; the Masterestaurant method reads contribution margin per service hour, because that measures the ability to generate cash with the structure already in place. A venue billing 95,000 dollars a month at 68% prime cost is not ready to scale, it is ready to be fixed. That distinction, which sounds like semantics, explains much of the 60% of second locations in Latin America that close before year three.

The three differences that move the money

Next comes the treatment of working capital. In the traditional model CAPEX takes the spotlight — construction, kitchen, furniture — and working capital ends up being whatever sits in the account on opening day. The pattern repeats with uncomfortable regularity: the new venue takes seven to eleven months to mature, that negative flow gets financed with the flagship's cash, and five months in you have two sick locations instead of one healthy and one being born. Reserving 4.5 months of fixed costs is not conservatism; it is the entry price to scaling. Third stands the question of how central overhead grows. The traditional route adds people when it hurts: an accountant once volume overflows, a purchasing lead once three invoices cross, a coordinator once nobody answers the phone. Overhead then climbs linearly with sales and eats 11% to 14% of consolidated revenue by the third venue. Masterestaurant grows it in planned steps every four units, with the function defined before the hire, and holds it between 6% and 8%.

The three differences that move the money — in practice

In a four-unit group billing 4 million dollars a year, those five or six percentage points are worth 200,000 to 240,000 dollars of EBITDA that appear without selling one extra plate.