Case Study 35.2 — The Two-Unit Trap

This is a labeled composite. It is assembled from a failure pattern that is extremely common and well understood in this industry — expansion past the management capacity that supports it — and every figure in it is constructed for teaching. It is not any specific business, and no named restaurant's financials appear here. The pattern is real; the numbers illustrate it.

Background

A neighborhood restaurant opens in a walkable district in a mid-size city. Ninety seats, a chef-and-manager partnership, dinner six nights, a full bar. It is good. Not extraordinary — good, and consistent, and warm, and the neighborhood adopts it.

By year four it is doing \$1,720,000 with a 62% prime cost and an 11.5% operating margin — \$197,800 of operating profit — and after \$74,000 of debt service the partners are taking home about \$124,000 between them. They are exhausted and they are proud, and both are correct.

The restaurant's reputation reaches a developer who is finishing a mixed-use project eleven minutes away. He wants a signature restaurant on the ground floor and he is offering terms: \$85 a foot in tenant improvements, four months of free rent, and a below-market base for the first three years. The space is 3,100 square feet. He needs an answer in ninety days because the anchor tenants are being announced.

The partners run the numbers on a legal pad over two nights. The second restaurant will do about what the first one does. Say \$1,500,000, conservatively, at the same 11.5%: \$172,500. Debt service on the build-out, maybe \$95,000. That is \$77,500 a year of new profit, and the TI allowance means they only have to put in \$140,000 of their own.

Seventy-seven thousand dollars a year for a hundred and forty thousand dollars. They sign.

The operating issue

Here is what the legal pad did not have on it.

The management the second building forced. The chef-partner took unit two's kitchen; the manager-partner stayed at unit one's floor. That left unit one without a kitchen leader and unit two without a floor leader. They promoted the sous at unit one to chef de cuisine — a good, loyal cook, eleven months in the job, who had never written a schedule or run a food cost — and hired a general manager for unit two from outside, who lasted seven months.

Loaded, those two seats cost about \$168,000 a year. Neither had existed before.

The overhead that appeared. The bookkeeper went from twelve hours a week to full-time. The accountant's fee doubled and then some, because there were two entities. A second POS, a second reservations subscription, a second scheduling license, a second inventory seat. Two insurance policies. A vehicle. Recruiting for forty-one new hires. Call it \$81,000 in year one, of which not one dollar appeared on either restaurant's P&L.

The cannibalization. Eleven minutes. The partners had not pulled a postal-code report, because the neighborhoods "felt different." Roughly a third of unit one's covers lived closer to the new site. Unit one's Tuesday and Wednesday, which had been running just above cash break-even, went below it and stayed there.

And the thing that is not a line item. In year five, for the first time since opening, neither partner was in unit one's dining room on a Tuesday. Nobody wrote a review about it.

What the second year actually looked like

(Constructed. Unit one's decline: \$112,000 of transferred sales, the loss of the chef-partner from the kitchen, and a promoted chef de cuisine learning food cost in public. Unit two's \$121,400 is a respectable second year — it is not the villain here.)

Read the bottom two rows together, because they are the case.

The group's revenue rose 71%. Its operating margin fell from 11.5% to 5.0%. The partners went from taking home \$123,800 to funding a \$21,300 shortfall out of savings. And their personal exposure more than doubled.

Unit two did not fail. It cleared 9.1% in its second year, which is an ordinary, survivable margin for a restaurant that age. The failure was structural: a two-unit company built on a one-unit management team, paying company overhead on restaurant volume.

What happened next

Year seven: the promoted chef de cuisine at unit one resigned, having spent two years being blamed for a food cost nobody had trained him to manage. The chef-partner went back to unit one's kitchen, which meant unit two's kitchen was now run by its sous. Unit two's food cost moved three points inside a quarter. Nobody noticed for eleven weeks, because the weekly flash report had quietly become a monthly one when the bookkeeper's workload doubled.

Year eight: the partners closed unit two. The landlord was reasonable and took a negotiated surrender in exchange for a payment; the equipment sold for roughly a third of its cost; the partners carried the remainder personally, because they had guaranteed it.

Unit one recovered, slowly, over about two years, and is open today. It is a good restaurant. The partners are in their fifties and they are eleven years behind where they would have been.

What it shows

The decision was made against the wrong comparison. "Seventy-seven thousand a year for a hundred and forty thousand" compares unit two to zero. The right comparison — the group with it against the group without it — was negative from the first month, and the legal pad had no line for the four things that made it negative: the management the second building forced, the overhead that appeared, the debt service, and the damage to unit one.

The two-unit valley is real and it is not survivable on hope. \$78,000 of above-unit overhead is 2.6% of group sales. On an 11.5% margin business that is nearly a quarter of the profit, handed to the company that owns the restaurants. At five units it would have been a fraction of that. At two it was fatal.

The bench failed exactly where the chapter says it fails. Not on opening night — opening night went fine. It failed eleven months later, when a promoted sous was asked to hold a food cost he had never been taught to hold, and then again at year seven, when he left and the failure cascaded into the other building. §35.5's rule — the bench must be in seat for twelve months before you sign, and must have run a slow February, a December, an inspection, and a resignation — is written from exactly this.

And the timing pressure was the mechanism. Ninety days is not enough time to build a bench, pull a postal-code report, run a residency, or accumulate cash. The developer's deadline was real and it was also, from the partners' side of the table, the single most expensive feature of the deal. A growth opportunity with a deadline shorter than the time it takes to become ready is not an opportunity. It is a filter that selects for operators who are not ready.

The lesson

There is a version of this case where the partners say no, spend twenty-four months building a bench and filling their soft nights, and open unit two in year seven with five leaders on payroll and \$210,000 of accumulated cash. That version probably works. Nobody would write a case study about it.

The lesson is not "never open a second restaurant." It is the one Chapter 29 put best and this chapter borrows:

Growth that consumes the thing that was working is not growth.

Discussion questions

1. Reconstruct the partners' legal pad, then add every line it was missing. What was the honest year-one number, and at what point in the analysis would you have stopped?
2. The developer's ninety-day deadline was real. Write the response that a prepared operator sends — one that neither signs nor closes the door — and say what it would take to make that response credible.
3. Unit two cleared 9.1% in its second year and the partners still closed it. Argue that closing was the right call. Then argue it was the wrong call, and say what information would settle it.
4. The promoted chef de cuisine "was blamed for a food cost nobody had trained him to manage." Whose failure is that, and what specifically should have happened in the eleven months before the promotion? Reference §35.5 and Chapter 21.
5. The weekly flash report became monthly when the bookkeeper's workload doubled, and a three-point food-cost move went unnoticed for eleven weeks. Price that omission at unit two's volume. What does it say about where above-unit overhead should be spent first?
6. The partners guaranteed both leases personally, so the shortfall from the surrender landed on them. How would the outcome have differed if unit two had been a small format at a third of the capital, or a food truck? Use §35.7 and Chapter 30 to answer in numbers.
7. The uncomfortable one: the partners were exhausted and proud at year four, taking \$123,800 between them for two enormous jobs. Is "stay at one unit forever" an adequate answer to that? What does §35.8 offer them, honestly, and what does it fail to offer?