Case Study 2 — The Turnaround That Worked and Wasn't Enough

A clearly labeled composite, assembled from patterns common to independent full-service closures. It is not a real restaurant and no figure here is drawn from any actual business's records. Roles only — the chef-owner, the co-owner, the general manager, the landlord's asset manager. Every number is constructed and every number computes.

This case is the complement to Case Study 1. That one was about a legal instrument at scale. This one is about a single dining room, an operator who did almost everything this chapter recommends, and the one thing that could not be fixed by doing it well.

It is included to teach the limits of §39.3. A ninety-day turnaround is the cheapest and most under-used instrument in this chapter. It is also, routinely, the wrong one.

Background

A 74-seat neighborhood Italian restaurant in a coastal city of about 200,000. Dinner six nights, no lunch, a small private-event business, no brunch. Two owners: a chef-owner with eleven years on the line and a co-owner who handled the room and the books. Opened with \$205,000 of owner money and a \$290,000 bank note, personally guaranteed.

The space: 3,100 square feet at \$44 per square foot all-in — \$136,400 a year, on a ten-year lease with escalations, personally guaranteed, no good-guy clause. It was a beautiful room in a corridor that had just started to turn, and the rent reflected what the corridor was expected to become rather than what it was.

Write that number down, because it is the entire case. \$136,400 of occupancy is the fact that decides everything that follows, and it was decided before the restaurant served a plate.

Year two: a good restaurant, barely

The food was legitimately good. Reviews were strong. Saturday was full by seven. Year two delivered 96 dinner covers a night at a \$41 average check, plus \$11,968 of private-event revenue.

Two things stand out and they point in different directions.

Prime cost at 69.0% is an execution problem. Nine points over a 60% benchmark on \$1,240,000 is \$111,600 a year of margin that should have existed and did not. That is exactly the diagnosis §39.3 exists for, and it is genuinely fixable.

Occupancy at 11.0% is a math problem. Chapter 1's benchmark range for full-service occupancy is 6–10% of sales, and eleven points on a business with a four-to-six point margin is not a stretch, it is a structural fact. And it escalates.

The owners saw the first one. They did not see the second, because nobody had ever computed break-even covers against physical capacity — they had never run Step 5.

Month 14: the offer

In the spring of year two, the landlord's asset manager approached them. The corridor had improved faster than expected, a national tenant was interested in the space, and the landlord offered to terminate the lease early: a \$60,000 termination fee**, plus surrender of the **\$40,000 security deposit.

The owners had, at that moment, \$178,000 available — a home-equity draw taken the previous month plus the remainder of an undrawn facility. Their closure floor at that point was roughly \$71,000.

Here is the choice, laid out the way this chapter would lay it out.

They declined. Everyone would have. The restaurant was full on Saturday, the food was good, the corridor was improving, and \$100,000 to walk away from something you built is a hard sentence to say out loud. They took the \$178,000 and put it into a turnaround.

And the turnaround worked.

Months 15–30: a textbook operational fix

They did nearly everything §39.3 prescribes, and they did it well:

• Weekly inventory, counted by the chef-owner, every Sunday, without exception.
• Scales and portion tools on every station; six items re-spec'd and re-costed.
• A comp and void authorization policy with a daily review by the general manager.
• Bar controls: jiggers, a keg-yield audit, a spill log.
• A staffing guide built from sales-per-labor-hour data, and a cut order that was actually held.
• A menu re-price on eleven items.
• Linen, waste, and payment processing all re-bid.

Over nine months, prime cost fell from 69.0% to 62.8% — 6.2 points. By any professional standard that is an excellent turnaround. Most operators who attempt one do not get half of it.

Then the market moved. A larger, better-capitalized restaurant opened four blocks closer to the waterfront, and the corridor's foot traffic reorganized around it. Dinner covers fell from 96 to 84. Revenue fell to \$1,096,000. The lease escalated on schedule to \$139,700.

Now compute what the turnaround was actually worth, because this is the number the case exists to produce.

At the old 69.0% prime cost, year three's prime would have been \$756,240, operating profit would have been −\$30,640**, and net would have been **−\$74,640.

The turnaround was worth \$67,940 — 6.2 points on \$1,096,000 — and the restaurant still lost money.

Read those two sentences together. The operational fix delivered essentially the maximum a ninety-day-style program can deliver. It moved the business from losing seventy-five thousand dollars a year to losing seven. It could not reach a rent that had become 12.7% of sales, and it could not manufacture twelve covers a night that had walked four blocks toward the water.

Month 30: the end, and what it cost

They closed at month 30. By then:

• Accumulated deficit and stretched payables totaled \$118,000.
• The closure floor had grown from \$71,000 to roughly **\$96,000**, almost entirely in the vendor line — the arithmetic of §39.7, exactly as described.
• They had \$31,000 in the bank.

The final payroll was made, but only because the chef-owner took a second draw against their home to fund it. Three vendors were paid in part. The space was surrendered; the landlord re-let it after seven months to a different tenant at a lower rent, and the realized claim on the personal guaranty came to approximately \$164,000.

Set the two options from month 14 side by side.

Sixteen months cost \$167,000 and ended in the same room with the same locked door.

The contested decision

Here is why this is a case study and not a morality tale: at month 14, the owners could not have known.

The turnaround was a genuinely reasonable bet. Nine points of prime cost is a real, identified, recoverable number, and they recovered two-thirds of it. Had covers held at 96, year three at 62.8% prime cost would have been a comfortably profitable restaurant. They were not being sentimental; they were being analytical, with the wrong analysis.

What they were missing was one afternoon of arithmetic. Step 5 of the diagnostic asks whether break-even covers exceed what the room can produce. At \$136,400 of occupancy, a \$41 check, and 74 seats, the answer was that the restaurant needed to run near the top of its realistic range every single night, in every season, forever, with no cushion for a competitor, a bad winter, or an escalation — all three of which are ordinary events.

That is not a business with an execution problem that also has high rent. It is a business whose arithmetic required perfection, and the correct name for that is a math problem.

The turnaround was the right thing to do and the wrong thing to bet the exit on. They could have done both — run the turnaround and taken the termination, closing at month 14 with the lease extinguished and their operational lessons intact — except that the termination was a one-time offer, and one-time offers arrive when the counterparty wants something, not when you are ready.

Which is the final lesson, and it is the same one as §39.2: the offer comes while you still have options, and it does not come back.

What it shows

1. An execution fix cannot reach a math problem, no matter how well it is executed. 6.2 points of prime cost is a professional achievement. It was worth \$67,940 and it was not enough.
2. Occupancy above 10% of sales is a structural fact, not a line item. It compounds through escalations and it does not respond to portion control.
3. A turnaround and an exit are not mutually exclusive — but the exit has a window, and the window is set by the counterparty.
4. Step 5 takes an afternoon and was never run. The most expensive omission in this case was not a decision. It was a calculation nobody made.
5. The closure floor grows in the vendor line. \$71,000 became \$96,000 without anyone deciding anything, because stretching payables is the last available lever and it borrows from the people who get hurt worst.
6. The guaranty outlived the restaurant by years. \$164,000, realized after mitigation, against a face obligation many times larger — and against a good-guy clause that was available at signing and never asked for.

Discussion questions

1. At month 14, with \$178,000 in hand and a \$171,000 termination-plus-closure cost on the table, was declining the offer the wrong decision — or the right decision made with incomplete information? Defend a position, and be precise about what information was missing and what it would have cost to obtain.

2. The turnaround delivered 6.2 points of prime cost. Compute what revenue would have had to hold at for year three to net positive at 62.8% prime cost and \$139,700 of occupancy. Is that a number the room could produce? What does your answer say about the diagnosis?

3. The competitor opening four blocks away is presented here as bad luck. Is it? What would a Chapter 2 trade-area analysis and a Chapter 32 margin-of-safety calculation have said about a business with no cushion in an improving corridor?

4. The chef-owner took a second draw against their home to fund the final payroll. Discuss this decision. What does it say about the closure floor as a concept, and what would you have done differently at month 24 to avoid the choice entirely?

5. Three vendors were paid in part. Write the phone call the chef-owner should make to the produce supplier — what is said, in what order, and what is promised. Then write the version that would be a mistake.

6. Suppose the lease had contained a good-guy clause with six months' notice. Recompute the total in Option B, and state in one sentence what that clause was worth. Then say what you would now do differently in a letter of intent.

7. This case and Case Study 1 both end with leases as the binding constraint — one at national scale, one in a single dining room. What does that convergence suggest about where an operator's attention should be concentrated in the first month of a project rather than the last?