How to open a restaurant step by step: the 2026 numbers and why the traditional order bankrupts sites

What is the real first step to opening a restaurant in 2026?

The first step is setting the cost structure you can carry, and only then looking for square footage that fits inside it. Written down it sounds obvious, and almost nobody respects it:

the lease gets signed during the week when the founder knows least about the business, with no plate costings, no finished menu and no projection of covers per shift. A ten-year lease at 9,400 dollars a month commits 1.13 million dollars in fixed spend, and at a prudent 8 % occupancy-to-sales target it forces the site to bill 117,500 dollars monthly from week one. No 62-seat dining room reaches that without two full services a day. The fix costs NOTHING: reversing the sequence takes no money, it takes patience. US restaurant industry sales passed 1.1 trillion dollars in 2025, up 4.1 %, reaching 1.5 trillion counting all foodservice, according to the National Restaurant Association.

The size of the industry will not protect you from its margin

That number comforts anyone about to sign, and that is exactly the trap. Typical operating margin for an independent runs between 3 % and 6 %, so on annual sales of 1.4 million dollars the operator keeps somewhere between 42,000 and 84,000. A giant sector on pharmacy margins means the mistake is not absorbed: it is paid. With six points of cushion, a badly calculated lease will not be fixed by selling more; it gets fixed by closing or renegotiating, and ten years into a term there is no leverage left to pull. The second breaking point is prime cost, food plus labor combined. A restaurant that starts with plates at 37 % food cost and payroll at 33 % lives on a 70 % prime cost, while the healthy trade range sits between 55 % and 60 %. Those ten points, on 1.4 million dollars of annual sales, are 140,000 dollars gone before rent, utilities or the buildout loan get paid.

Prime cost decides whether you open or merely inaugurate

Volume will not rescue them either: every extra dollar billed under that structure drags 70 cents of variable cost behind it. At Masterestaurant we cap food cost per plate at 32 %, never as a target but as the tolerable MAXIMUM, with payroll and rent kept out of the costing sheet and charged instead to the break-even calculation. The requirements — health, extraction, occupancy, food handling, liquor license — all have to be met, and meeting them produces not one dollar of contribution. The sequencing mistake gets expensive by another route: in many markets the full permitting process runs four to six months, and that stretch is paid with the shutter down if the lease clock already started. On the same 9,400 dollars a month from the example, five months of administrative waiting burns 47,000 dollars without serving a single cover. Negotiate a rent-free period covering buildout and licensing, or do not sign.

Licenses generate no margin, yet they burn rent

A founder who enters that process with an active lease loses every ounce of pressure, because the clock runs against him and the landlord knows it perfectly well. Chains publish what the independent guesses, which is reason enough to study them. Chick-fil-A sits near an average unit volume of 7.5 million dollars and Raising Cane's approaches 6.5 million, per the Restaurant Business AUV ranking; McDonald's averages 3.96 million per unit according to its 2024 FDD; Cava, the fast casual benchmark, lands around 2.93 million (Technomic), and Wingstop at 2.13 million per its 2025 FDD. Jack in the Box closes at 1,913,335 dollars over the twelve months to September 2025. That spread — 1.9 to 7.5 million — says something uncomfortable: formats with short menus and simple kitchens bill more per square foot. If your concept needs twenty fresh SKUs, your realistic AUV belongs at the bottom of that table, not the top.

A franchise is not a shortcut, it is a different calculation

For a group leader weighing expansion, the relevant fact is that the franchised network keeps growing while the total store count does not. Franchised QSR units passed 204,000 in 2025, advancing 2.2 %, according to the International Franchise Association, and Top 500 fast casual grew 6 % to nearly 77 billion dollars (Technomic). Meanwhile US chain locations fell to roughly 691,181 in 2024 from about 703,000 in 2019, per Technomic Ignite. Fewer sites billing more: consolidation, not expansion. For anyone opening a first unit the reading is blunt. The market rewards sales density per location rather than a count of flags, and whoever plans three openings before stabilizing the first is buying the risk of scale without the margin that pays for it. Staffing gets planned alongside the menu, and it almost always gets thought about last.

The team that will run it is also an opening constraint

There are now 6.2 million 16-to-19-year-olds in the US workforce, roughly 900,000 more than in 2019 according to National Restaurant Association figures drawn from BLS, which relieves the register and the pass, yet covers none of the posts that demand judgment: a head chef who controls waste, a manager who reconciles inventory weekly. If your model depends on that profile showing up the month you open, you have an operational risk hidden inside a financial plan. Yum China runs 17,514 stores as of September 2025 because it standardized processes until they stopped depending on individual talent, and that is the exam your menu must pass before the buildout starts. Three numbers, one action each. First, 8 % occupancy-to-sales: divide monthly rent by 0.08 and compare the result against what your dining room can bill full across two services a day; if it falls short, the site is not yours, however good the corner looks.

The 3 numbers you should tattoo on yourself

Second, 60 % prime cost: cost out the ten dishes you expect to sell most, add projected payroll, and if the total clears 60, cut the menu before buying a single oven. Third, the 3 % to 6 % operating margin typical of an independent: multiply it by your expected annual sales and you will see the real money the year produces. With those three calculations done on one sheet, before any signature, you already know whether you are opening a restaurant or buying a very expensive ten-year job. The gap is not about ambition, it is about WHICH decision comes first. A ten-year lease at 9,400 dollars a month commits 1.13 million dollars of fixed cost, and that signature usually lands in the very week a founder knows least about their own business. Flipping the order —lock the cost structure, then hunt for square footage that supports it— costs nothing but patience, and it is the only free correction available across the whole project.

The differences that land in the cash drawer

Food cost is the second breaking point. A restaurant opening with dishes at 37 % raw material cost and a 33 % payroll load lives at a 70 % prime cost, while the healthy range of the trade runs 55 % to 60 %; those ten lost points, on 1.2 million dollars of annual sales, equal 120,000 dollars that no longer exist to cover rent, debt, or investor return. Fixing it after opening means raising prices or rewriting a menu guests already got used to. The third difference is working capital, and I was wrong for years recommending a three-month cushion. A new restaurant's ramp is rarely linear: opening curiosity inflates the first six weeks, a trough follows and drains the team's morale, and recovery depends on repeat visits rather than launch noise. Six full months of fixed costs, set aside and untouchable, is what lets you cross that trough without selling the project cheap to an emergency partner.

The differences that land in the cash drawer — in practice

Then there is MTIE, which in the Masterestaurant methodology is literally the approval threshold: months to break-even. If the model, under conservative check and turnover assumptions, does not reach break-even within 9 to 12 months, the project is not approved —not adjusted, not optimized, not approved—. That hard rule saves more money than any supplier negotiation, because capital never committed never has to be recovered.