Don't Expand Your Restaurant Until You Can Answer This One Question - A Framework for Operators and Investors
- Apr 13
- 5 min read
The Most Expensive Sentence in Restaurants
"We should open another one."
It usually gets said on a packed Friday night. The energy is high, the bar is three deep, and the room feels electric. For the operator, it feels like validation. For the investor across the table, it sounds like opportunity. Both of them might be wrong.
The decision to expand a restaurant brand is one of the most capital-intensive moves in hospitality, and one of the most frequently made on the wrong evidence. Busyness is not a business case, it's a feeling. A full dining room is a demand signal. It is not a scalability signal. Before a second lease gets signed or a second round of capital gets deployed, the question isn't whether demand exists. It's whether the system behind that demand is actually transferable.
Busy Is a Feeling. Profit Is a System.
Volume is forgiving. Expansion is not.
A well-run single unit operating at capacity can sustain inefficiencies that would be fatal at scale, overstaffing absorbed by high revenue, inconsistent execution masked by guest loyalty, labor creep covered by strong weekend nights. These aren't just operational inconveniences. They're margin erosion events that multiply across every unit you add.
For operators, this means a full dining room can be hiding the very problems that will sink a second location. For investors evaluating a concept for capital deployment, it means top-line revenue at unit one is often a poor predictor of EBITDA performance at unit two, particularly if unit one's margins are owner-dependent rather than system-dependent.
The distinction matters enormously. A restaurant generating 18% EBITDA margins because the owner works the floor six nights a week is not the same business as one generating 18% margins through a documented operating system and a trained management layer. Only one of those scales.

The One Question That Determines Everything
Can this restaurant perform predictably, under pressure, without its founder?
This is the operator's readiness test and the investor's due diligence question. They're the same question. If the answer is genuinely yes, you have the foundation of a platform. If the answer is "mostly" or "with the right people in place," you have a high-performing single unit that carries significant key-person risk at scale.
That risk doesn't disappear when you open a second location. It compounds.
What Readiness Actually Looks Like In Numbers
Readiness isn't a feeling. It's a pattern you can measure.
At the unit level, a scalable operation shows consistent prime costs — labor and cost of goods combined, landing reliably below 60%, without daily intervention from ownership. Weekly EBITDA margins hold within a narrow band across peak and off-peak periods, not just on Saturday nights. Average Unit Volume is predictable enough to model against buildout costs with a clear payback window, which is typically 24 to 36 months for a well-capitalized expansion.
Operationally, ticket times hold when the room fills. The guest experience is consistent on a Tuesday as on a Friday. Managers run shifts and resolve problems without escalating everything upward. Standards don't drift when ownership isn't present.
For operators, this is the baseline before signing a lease. For investors, this is the minimum viable evidence before deploying capital into a multi-unit thesis.
The Management Independence Test
As an owner step away from operations for four consecutive weeks. No check-ins, no schedule fixes, no line coverage.
If performance holds margins stable, throughput consistent, guest experience intact, you have a management infrastructure that can support a second unit. If things drift, you have a business that is performing well precisely because of your involvement. That's not a platform. That's a job.
For investors, the equivalent test is simpler: ask to see four weeks of operating data from a period when ownership was traveling or otherwise removed from daily operations. The variance in that data tells you more about scalability than any pro forma.
Where Expansion Actually Breaks
The failure pattern is predictable and almost never dramatic.
Unit one starts slipping as ownership attention divides. Unit two never fully stabilizes because the operating system that existed at unit one wasn't documented, it was institutional knowledge held by a small group of people who are now split across two locations. Leadership gets stretched. Communication slows. Standards erode in increments too small to trigger intervention until the margin damage is already done.
For operators, this feels like losing control of something you built. For investors, it shows up as unit-level EBITDA compression at location one concurrent with underperformance at location two — a pattern that is difficult to reverse without significant operational restructuring and, often, additional unplanned capital.
Nothing collapses overnight. It degrades quietly, expensively, and on both P&Ls simultaneously.
The Math That Doesn't Get Modeled Correctly
A second location doesn't double profit. It doubles complexity.
If unit one requires constant ownership oversight to sustain its margins, unit two introduces an entirely new variable set, new team, new market, new physical plant, new supplier relationships on top of a system that was already fragile. The pro forma built on unit one's best periods will not survive contact with the operational reality of unit two.
Margins don't scale automatically. Systems do. The capital required to build those systems after expansion is significantly higher than the capital required to build them before it — because after expansion, you're paying for them under pressure, with two locations at risk instead of one.
For investors underwriting a multi-unit rollout, the most important number in the model isn't the projected AUV at unit two. It's the defensibility of unit one's margins during the distraction of opening it.
What Has to Exist Before Capital Goes In
These are prerequisites, not goals to work toward during expansion:
Documented operating systems that produce consistent results on the worst day, not just the best. A management layer, GMs and shift leads, that can hold standards, manage people, and run service without ownership involvement. Repeatable unit economics across multiple seasons, not just peak periods. Clear throughput capacity data showing how much revenue the operation can process per hour before guest experience degrades. A leadership bench deep enough that losing one key person doesn't destabilize either location.
If any of these are absent, expansion will force you to build them under pressure. That is the most expensive way to do it.
The Right Time to Expand (It's Not When You Think)
Counterintuitively, the right time to expand is when operations feel controlled. When nothing feels urgent. When your best night looks a lot like your average night, and your average night still works without you in the building.
For operators, that's readiness. For investors, that's a platform worth backing, a concept where the unit economics are proven, the management infrastructure exists, and the expansion thesis is supported by data rather than momentum.
That's not stagnation. That's the business actually working.
Final Thought
Expansion is not a reward for being busy. It's a test of whether the business works — operationally and financially, without the conditions that made the first unit successful.
Pass that test and you build a scalable brand. Fail it and you build a more expensive, more complex version of the same fragile system, now with two addresses and a much narrower margin for error.
I work with founders, operators, and investors to build restaurant platforms that are ready for growth before capital goes in — through documented systems, management infrastructure, and unit economics that hold under pressure. Whether you're preparing for a second location or evaluating a concept for investment, the question is the same: is this business actually ready? If you're not sure, that's exactly where the conversation starts.



Comments