Case Study 2: The Great Lease That Nearly Closed the Restaurant

This case is a clearly labeled composite. It is assembled from patterns that recur constantly in restaurant construction — concealed conditions, code-triggered upgrades, allowance reconciliation, plan-review delay, and the migration of money between budget buckets. No real business, landlord, or contractor is depicted, and every figure is illustrative. It is built as a composite deliberately: no single real project would make the argument as cleanly, and inventing details about a real one would be worse than useless.

The complementary case to the first one. There, a bad clause met a catastrophe. Here, an excellent lease meets an ordinary construction project — and the ordinary construction project wins.

Background

Two partners with strong operating backgrounds — one a chef, one a general manager — found 2,600 square feet of second-generation restaurant space in a neighborhood they knew well. They negotiated hard and they negotiated well.

Five-point-nine percent occupancy. Their broker congratulated them. Their accountant congratulated them. By every benchmark in Chapter 1 and every table in §6.4 of this chapter, that is a very good deal, and if the story ended here it would be a success case.

There was one term they traded. The landlord's draft set rent commencement at ninety days after delivery of possession. The partners asked for the certificate-of-occupancy trigger. The landlord countered: keep the delivery trigger and take a dollar a foot off the base rent.

A dollar a foot on 2,600 square feet is \$2,600 a year. Over a flat ten-year comparison that is \$26,000 of certain, bankable savings against a risk that felt theoretical, on a project they were confident would take five months.

They took the dollar.

The operating issue

Their construction budget:

  CONTRACT SUM (stipulated, drawings ~85% complete)        $240,000
    incl. electrical service upgrade ALLOWANCE   $12,000
  SOFT COSTS (architect, engineering, permits, utility)      34,000
  CONTINGENCY                                                11,000
  ─────────────────────────────────────────────────────────────────
  TOTAL CONSTRUCTION BUDGET                                 $285,000
      = $109.62 per square foot        contingency = 4.6% of hard cost

The contingency is the number to circle. Eleven thousand dollars against a \$240,000 contract sum is 4.6%, where the rule of thumb is 10–15% — which would have been \$24,000 to \$36,000. The partners knew this. They also knew what they had raised, and \$11,000 was what was left after the budget was built backward from the money available. That is how almost every thin contingency comes into existence: not from ignorance, from arithmetic performed in the wrong direction.

Then the project happened.

Plan review took eleven weeks instead of the five they had scheduled. Nothing went wrong; the jurisdiction was busy, one round of comments came back on the mechanical set, and the resubmittal went to the back of the queue. This is not a failure. It is the ordinary behavior of a system that does not work for you.

Four things then hit the budget:

\$60,200 against an \$11,000 contingency is a \$49,200 overrun. As a percentage of the contract sum it is 25%, which is well outside the 8–15% range §6.6 gives as a planning number — this is a bad project, not an average one. But look at the composition: only \$9,600 of it, 16%, was the owners' own doing. Three-quarters of the overrun was a collapsed pipe, a code requirement, and a placeholder number that turned out to be wrong.

And then the clause they traded came due. Construction ran eight months instead of five. Rent commenced ninety days after delivery — the beginning of month four. They opened at the start of month nine.

$$5 \text{ months} \times \$7{,}150 = \$35{,}750 \text{ of rent paid on a building that could not open}$$

Under a certificate-of-occupancy trigger, that number would have been zero. The clause they sold for \$2,600 a year cost \$35,750 in a single stretch — and, far more damaging, it cost it in the five months when there was no revenue at all.

What it shows

Total unplanned cash requirement: \$49,200 + \$35,750 = \$84,950.

There was no more money to raise, so the money came from inside the project:

Read that table as the case's central finding. The construction overrun did not come out of the construction budget. It came out of the restaurant's ability to operate. Chapter 1 warned that the contingency and the reserve are different money; this is the mechanism by which they get merged, and notice that nobody ever decided to merge them. It happened one invoice at a time, each of which had to be paid.

Three further observations.

The lease was never the problem. First-year revenue came in at \$1,310,000 — below the \$1,450,000 forecast, which is entirely ordinary. Occupancy was 6.5%, still comfortably inside the healthy band. Every congratulation the partners received about their rent was correct. A great occupancy percentage does not protect you from a construction failure or a cash-timing failure, because they are different failures.

Used equipment is a loan from year two. The used dish machine failed in month seven. With \$3,050 in the reserve, the repair was funded on a personal credit card, and the partners spent the following eighteen months paying down opening debt instead of building a cushion. This is the pattern Chapter 1 described from the outside: a business that worked, bled slowly, and had no margin for an ordinary event.

The single highest-value thing they did not do was finish the drawings. At 85% complete, the electrical service was carried as an allowance. Completing the drawings and re-bidding would have cost perhaps three weeks and a few thousand dollars in design fees, and would have converted a \$17,500 surprise into a known number available at the moment they were still deciding whether to sign anything.

Outcome

The restaurant survived. It is open, it is reasonably successful, and by year three it was performing close to its original plan.

But the partners spent two years operating with no reserve, carrying personal debt, running equipment they had not chosen, and making purchasing decisions from a position of weakness — paying for produce at delivery instead of on terms, buying in small quantities at higher unit cost, and unable to take a prime-vendor position (Chapter 13) because they could not commit volume. Every one of those is a cost that shows up somewhere else on the P&L, and none of them appears on any line labeled "construction overrun."

Asked afterward what they would do differently, neither partner said anything about the rent. They said two things: take the certificate-of-occupancy trigger and let the landlord keep the dollar, and do not start a build with a contingency under ten percent, even if it means opening three months later.

Lesson

Case Study 1 taught that the clauses you never negotiate decide catastrophes. This one teaches the harder half: the clause you do negotiate can still be the wrong one to trade.

Three transferable principles.

One: trade duration for money, never money for timing. The partners exchanged a certain, small, annual saving for an uncertain, large, concentrated exposure that landed in the months of maximum cash weakness. The saving was real and the arithmetic on a spreadsheet even favored it over ten years. It was still the wrong trade, because \$2,600 a year is a rounding error in a healthy business and \$35,750 in one quarter with no revenue is an existential event. When you are comparing a stream to a lump, ask when the lump lands.

Two: the contingency is not a budget line, it is the boundary of your project. A project you cannot fund with a 10% contingency is a project you cannot fund. Building it anyway does not remove the risk; it relocates the risk onto the operating business, where it does more damage.

Three: an unfinished drawing set is a decision to be surprised later. Every allowance carried into a contract is scope that will be priced at a moment when you have no leverage, no alternative bidders, and a crew standing in the space.

And one limit on this chapter itself, worth stating plainly. Chapter 6 spends most of its length teaching you to negotiate a lease, and that teaching is correct. But negotiating the lease well is necessary and not sufficient. These partners did the lease part better than most operators ever will, and it did not save them, because the money and the schedule were decided in a different document — one that most first-time operators sign with far less scrutiny than the lease and that Chapter 7 is about to make considerably more expensive.

Discussion questions

1. Reconstruct the rent-commencement trade as a decision under uncertainty. What probability of a three-month overrun would have made the dollar-a-foot saving the correct choice? What was the actual probability, given industry experience, and how should they have estimated it?

2. The partners built their contingency backward from the money available rather than forward from the risk. What should they have done instead when the arithmetic said \$11,000 — and what does each option cost?

3. Only \$9,600 of the \$60,200 in change orders was owner-requested. Does that exonerate the partners? Which of the other three lines could a more disciplined process have caught, and at what stage?

4. Trace the used dish machine backward: from the month-seven failure, through the equipment-budget raid, through the change orders, to the drawings. At how many points could the chain have been broken, and which break would have been cheapest?

5. The restaurant survived and is now doing fine. Does that make this a success story? Argue both sides, and consider what "success" means for two people who spent two years servicing opening debt in a business with a four-to-six-point margin.

6. Compare this case with Case Study 1. In one, external catastrophe met bad drafting; in the other, ordinary events met a thin budget. Which risk should a first-time operator spend more of their limited attention and leverage on — and does your answer change if the space is a raw shell rather than second generation?