You Can Grind to Five. You Have to Build to Twenty.

Every operator who wants to scale a restaurant past a few locations gets the same advice: run a tight operation and keep going. It’s wrong in a way that costs people their groups.

Scaling isn’t one problem repeated. It’s a stack of different problems behind each other, and the difficulty doesn’t climb in a straight line. One to five units is hard. Five to ten is harder. Ten to twenty is a different category of hard, the kind that breaks operators who got everywhere they got on grit. The grind that built the first five is the exact thing that can’t carry the next fifteen.

The difficulty comes in cliffs

Picture the climb as a smooth ramp (more units, proportionally more work) and you’ll plan for the wrong thing. The real shape is stairs with cliffs at the edges: long flat stretches where you think you’ve got it, then sudden drops at the inflection points where the model that worked stops working all at once. The cliffs sit at predictable places (roughly five, ten, twenty) and each one breaks something different.

Five tests whether the concept’s economics survive being copied. Ten tests whether the business can run without you in the building. Twenty tests whether you’ve built a company, or just a very busy founder with a lot of restaurants. Operators who plan for “more of the same” walk off each cliff at full speed.

One to five runs on you

At one to five units, the founder is the infrastructure. You, maybe an assistant and a few trusted people, hold the standard by being close to it. You taste the food, walk every dining room, catch the problem before the guest does, know every manager by name and every number by feel. It works, and it can be profitable; plenty of excellent operators run three or four restaurants this way for years and do very well.

But understand what’s holding it up: it’s you. That’s fine at five. It’s the trap at ten.

Five to ten is where grit stops working

The first hard cliff is the one between five and ten, and it’s brutal precisely because grit got you this far. The reflex when things slip is to work harder: be in more restaurants, drive more miles, personally fix more problems. At five that still mostly works; your attention can still reach.

Somewhere past it, it can’t. One person’s presence does not stretch to ten locations. And the harder they push, the clearer it gets that effort isn’t the missing ingredient. The thing that built the business, the founder being everywhere, becomes the ceiling on it. You don’t grind through this cliff. You build through it, or you stall at the size your own attention can cover.

You can’t scale past ten units without infrastructure

This is the line nobody believes until they’re living it: you cannot run that kind of operation without the infrastructure to support it.

Infrastructure isn’t software or a bigger office. It’s the unglamorous machinery that replaces the founder’s presence. None of it is exciting. All of it is what the next ten units run on.

And the cruel part of the timing: it has to exist before the cliff, not after. You build the infrastructure for ten while you’re still at five, because the unit you open into a broken operation doesn’t get fixed by becoming part of a bigger one. It just copies the break.

Ten to twenty is a different company

If five-to-ten is about replacing your presence with systems, ten-to-twenty is about changing what your job even is. Past ten units you’re no longer running restaurants. You’re running an organization that runs restaurants. The work stops being operating and starts being building operators.

The operators who stall here are often the most talented restaurant people, so good in the building they can’t stop being in it. A twenty-unit group can’t be run by the best restaurant operator in it. It has to be run by the person who built the thing that runs the restaurants.

Build the infrastructure before you need it

The discipline almost nobody has is building for the next stage while still comfortable in the current one. It feels premature: why build a management layer for ten units when you have six and they’re running fine? Because the moment you need it, it’s too late to build it calmly; you build it in a panic, mid-stumble, while the new units bleed.

~5 units

What Has to Exist by Then: Proof the flagship’s numbers hold on a schedule a hired manager runs, not only on yours.

~10 units

What Has to Exist by Then: A management layer between you and the GMs, standing on a labor model and a weekly scorecard that already run without you.

~20 units

What Has to Exist by Then: People who develop the people who develop people, a standard held across leaders the owner doesn’t see daily, and decisions made about capital and bench depth, not any single dining room.

Growth is the reward for the build, not a substitute for it.

What to build while you’re still at five

The objection deserves a straight answer: a corporate hire at five units is margin you’re spending on units you don’t have yet. A director of operations at an illustrative $120K loaded lands on five P&Ls at $24K a unit, straight off the bottom line, for capacity you won’t fully use until unit eight. That instinct isn’t wrong; it’s just aimed at the wrong target. The mistake isn’t hesitating on the salary. It’s treating the salary as the whole build.

Most of the infrastructure that carries a group past ten costs discipline, not payroll, and the payroll piece has a number attached: one good regional runs four to six units well. Which means the moment you can see unit seven on the horizon, the first above-store operator isn’t premature overhead. It’s the load-bearing wall for the units already signed. What you build before that hire is the part that’s nearly free:

• —Write the standard down. Triggered the day you have a manager you didn’t personally train. The recipes, the service steps, the way a Friday gets set up: on paper, so a manager you didn’t train can teach it at pre-shift the way you would.
• —Build the labor model. Triggered the first period a unit misses its labor line without you noticing for two weeks. Hours tied to forecast volume, by daypart, executable by a manager in their second week.
• —Install the weekly cadence and scorecard. Triggered the first week you don’t set foot in every unit. The handful of numbers, read the same day every week, that tell you the truth about a restaurant you didn’t visit.
• —Bench one level deep. Triggered the moment you can name the next opening but not its GM. Every unit develops its own successor, so growth pulls from a bench instead of gutting a store.

Stamina does not scale

Grit is the price of entry to this business, not the strategy for scaling it. The operators who get past the cliffs aren’t tougher than the ones who stall. Plenty of people who stalled worked themselves nearly to death trying. They built something underneath themselves before they needed it, so each unit stood on infrastructure instead of on the founder’s stamina.

The concept and the instinct get you onto the climb; the infrastructure is what gets you up it.