Case Study 2 — Profitable in November, Closed in March
This is a labeled composite. It is not a real restaurant. It is assembled from patterns that recur constantly in independent full-service operations: a rent escalation, a slow prime-cost drift, deferred revenue mistaken for income, payables stretched past terms, and a daily-remittance advance taken in the worst month of the year. Every figure is constructed and illustrative. No real business, person, or lender is depicted, and no participant is named — the roles are the chef-owner and the front-of-house partner.
Chapter 1 promised that most restaurants that close were, on paper, doing fine two months earlier. This is what that looks like from the inside, month by month, with the reports the operators were actually reading.
Background
A 78-seat neighborhood Italian restaurant in a mid-size American city. Year four. Well-reviewed, a genuine local institution, a Saturday you had to book. Two partners: a chef-owner and a front-of-house partner, the same structure Bellwether has.
The year-four profit-and-loss statement, which the outside bookkeeper delivered in the third week of January for the year just ended:
Revenue $1,640,000 100.0%
Cost of goods sold 495,280 30.2%
Labor, all-in 555,960 33.9%
──────────────────────────────────────────────────────────────────
PRIME COST 1,051,240 64.1%
Occupancy (year-three escalation, now 9.0%) 147,600 9.0%
Other operating 246,000 15.0%
General & administrative 49,200 3.0%
──────────────────────────────────────────────────────────────────
OPERATING PROFIT 145,960 8.9%
Interest on debt 41,000
──────────────────────────────────────────────────────────────────
NET PROFIT $104,960 6.4%
Read it the way Chapter 1 taught you and it is a good statement. Prime cost at 64.1% is four points above the full-service target — workable but tight, and worth fixing — but a 6.4% net on an independent is above average. Nothing on that page says "closing in nine weeks."
The operating issue
What the P&L could not show
Total debt service was $96,000** a year: roughly **$41,000 of interest and $55,000 of principal.** The partners took **$84,000 in distributions across the year, in four quarterly payments, which is not extravagant for two working owners.
Run the profit-to-cash bridge from §33.1:
Net profit as reported $104,960
− Debt PRINCIPAL repaid (not an expense) (55,000)
− Owner distributions (not an expense) (84,000)
──────────────────────────────────────────────────────────────────
= Change in cash before any timing effects ($34,040)
The restaurant reported $104,960 of profit and consumed $34,040 of cash. Neither number is wrong. The partners had never seen the second one, because nobody had ever built the bridge — the bookkeeper produced a P&L, and a P&L does not contain principal or draws.
What the bank balance could not show
On November 30 the operating account held $71,300, which felt like the best position the business had been in for two years. It was in fact the worst.
Bank balance, November 30 $71,300
Gift cards sold in the holiday push, not yet redeemed (37,000)
Deposits held on Q1 private events not yet catered (18,400)
Sales tax collected, remitted December 20 (9,600)
──────────────────────────────────────────────────────────────────
Genuinely free cash $6,300
Memo: days payable outstanding had drifted from 16 days
to 34 days over the year. At $1,357 a day of COGS, those
18 extra days represent $24,426 of purchases financed by
vendors beyond agreed terms — cash sitting in the balance
above that had already been spent by somebody else's
accounts-receivable department.
──────────────────────────────────────────────────────────────────
Adjusted position ($18,126)
Every dollar of the difference between $71,300 and negative $18,126 is a liability that looked like cash. This is §33.2's warning — getting paid instantly creates a dangerous illusion of liquidity — running at full strength, with three separate mechanisms stacked.
The trap in the trend line
Here is the part that makes this composite worth studying rather than merely worth pitying.
The partners were watching cash. They had started a simple weekly balance log in September, exactly as this chapter recommends. The log showed the balance rising: $38,000 in September, $47,000 in October, $71,300 in November.
A rising cash balance on a deteriorating business. Both of the mechanisms producing the rise — selling gift cards and stretching payables — generate cash today and an obligation later. A cash log that does not separate free cash from liability cash will report a recovery on a business that is accelerating toward a wall.
This is the most important limitation of everything in Chapter 33, and it is why §33.7's forecast carries the memo lines beneath the ending balance. A cash number without its composition is not information.
What happened
January. The gift cards came back. Roughly $21,000 of the $37,000 redeemed in the first six weeks of the year — revenue on the P&L, with no cash attached, because the cash had arrived in December and been spent. Two Q1 events were catered against deposits already consumed. The bookkeeper delivered the year-four statement showing $104,960 of profit, and the partners, reading it in a January that felt terrible, concluded that January was just January.
February. The month behaved exactly as §33.6 describes: the weakest revenue of the year colliding with the annual bills. The insurance renewal, the landlord's NNN reconciliation, and the license renewals landed inside three weeks. Two vendors — the produce house and one of the specialty purveyors — moved the account to COD after a payment ran nineteen days past terms. That single change removed roughly $9,000 of float in one week, which is the mechanism §33.4 warns about: once your DPO is 45 days, stretching is no longer a lever, you have spent it.
The decision. In the second week of February the partners disagreed, and the disagreement was legitimate.
The front-of-house partner argued for an orderly wind-down: close at the end of March, use the remaining unencumbered assets to pay staff through the notice period and settle with vendors, negotiate a lease termination while the landlord still had a paying tenant to work with, and stop the bleeding. Chapter 39's argument, made early.
The chef-owner argued that the business was profitable, that February is always terrible, that the spring had recovered the previous three years, and that closing a well-reviewed neighborhood institution over eight weeks of cash was an act of panic. That argument was not stupid. On the accrual evidence available, it was the better-supported one.
The chef-owner prevailed, and the business took a merchant cash advance: $60,000 advanced at a 1.38 factor rate, **$82,800 to be repaid, via a 13% holdback on daily card volume. At roughly $4,179 a day of card sales, that is about **$543 a day, retiring the advance in something on the order of five months — which is to say, through the entire spring and early summer, the exact period in which the business would otherwise have rebuilt.
$22,800 of cost on $60,000, consumed in about five months. The simple annualization is well into double digits; because the balance amortizes daily, the effective annual rate on a product like this is commonly quoted far higher. Nobody presented it that way. It was presented as "$60,000 tomorrow, and you barely notice the holdback."
March. The holdback took 13% off the top of every card batch, which meant it took the most on the best nights and left the least on the slow ones. Payroll cleared on the 14th by $1,100. On the 26th, with two vendors on COD, a payroll due Friday, and the April rent due in six days, the partners closed. The lease was surrendered rather than assigned, because a tenant with a March closure and vendor arrears is not an attractive assignment candidate — which is exactly what the front-of-house partner had been arguing in February, when a negotiated exit was still available.
What it shows
A restaurant does not need to be unprofitable to close. It needs to be insufficiently liquid when something ordinary goes wrong. This composite's business was profitable in every accrual sense for four consecutive years and never once computed the profit-to-cash bridge.
Deferred revenue is the most seductive form of borrowing in this industry, because it does not feel like borrowing at all. A gift card is a loan from a guest at zero interest with an unknown redemption date, and a holiday gift-card push is a very large loan taken out in the month before the two worst months of the year. Used deliberately by a business with a reserve, it is a fine tool. Used to fund February, it converts December's strength into February's weakness with a one-month delay.
Payables stretching has a terminal point, and it is not a warning — it is a stop. The move from terms to COD is made by a credit department, not by you, and it happens in one day. The float you have been relying on for six months disappears in the week you can least afford it.
A cash log without composition is worse than no cash log, because it manufactures confidence. Three of this restaurant's four "improving" months were improving because of liabilities.
The limits this case teaches
Chapter 33 is a chapter of instruments, and instruments have failure modes. This case shows four.
One: a thirteen-week forecast is only as good as its revenue line. The partners re-forecast four times between November and February. Each version showed the position recovering, because each version assumed a spring that arrived in April. The arithmetic was flawless. The assumption was the whole answer. The fix: forecast a base case and a case in which revenue runs 10% light, and act on the second one.
Two: the forecast tells you when you run out, not whether to continue. These are different questions and only one of them is financial. The front-of-house partner was not making a cash-flow argument in February; they were making a Chapter 39 argument about whether a business with a 9.0% occupancy line, a 64.1% prime cost, and no reserve had a route back. A thirteen-week forecast cannot answer that, and an operator who asks it to will get a number that sounds like permission.
Three: the "repaid inside the horizon" test is only a discipline if the horizon is fixed. §33.8's rule — never fund an operating loss with borrowed money; check that the drawn balance is repaid inside thirteen weeks — was technically satisfied every time, because the horizon kept moving forward with the forecast. A rolling window with a moving assumption will pass any test you give it.
Four: cash can rise while a business dies. Everything in §33.7 assumes you are reading the balance net of what is not yours. Strip out the memo lines and Bellwether's $42,598 becomes $24,855; strip them out here and $71,300 becomes negative $18,126. The memo lines are not an appendix to the forecast. They are the forecast.
Discussion questions
-
Reconstruct the November 30 position as the partners saw it and as it actually was. Which of the four adjustments — gift cards, event deposits, sales tax, or the DPO drift — would you consider the most dangerous, and why? Is your answer the same as the largest one?
-
The chef-owner's February argument was that the business was profitable and that spring had recovered the previous three years. Build the strongest possible version of that case, then say precisely which piece of evidence should have overturned it. Be specific: name the number and where it would have come from.
-
The composite restaurant took a $60,000 advance at a 1.38 factor with a 13% holdback. Compute the total repaid, the daily holdback, and the approximate repayment period. Then answer the harder question: at what price would the advance have been the right decision, and what would have had to be true about the business for any price to be right?
-
The front-of-house partner argued for an orderly wind-down in February. Chapter 39 covers that ground properly, but on the evidence in this case: was that the right call, and what would you have needed to see to be confident? Would your answer change if the lease had four years remaining instead of eight?
-
The partners' weekly cash log showed a rising balance for three consecutive months on a deteriorating business. Redesign the log so it could not have done that. Specify every line it should contain and the order they appear in — you are designing the one page from §33.5's Monday meeting.
-
Contrast with Case Study 1. In March 2020 an external shock stopped the revenue. In this case nothing external happened at all — the business simply ran a slow deficit while reporting a profit. Which failure is more common in this industry, which is more preventable, and which one does this book spend more of its pages on? Justify the allocation.
-
Business Plan connection. Bellwether opens with $8,700 against a $101,375 requirement. This composite closed with a structurally similar hole four years in. Write the two-sentence risk statement for Bellwether's plan that names the specific mechanism from this case that Bellwether is most exposed to, and the specific control from §33.9 that addresses it.