Case Study 2: Profitable on Paper, Three Weeks Late, and Gone

A labeled composite on the difference between having numbers and reading them

This case is a constructed composite, assembled from operating patterns common across independent restaurants that use an outside bookkeeper and no internal weekly reporting. It is not a real business, and no figure in it should be read as reported data. It exists because this failure is extremely common, entirely preventable, and almost never written up — the restaurants it happens to have accurate books and no idea what is in them.

Background

The restaurant in this composite did everything a responsible small business is told to do about its accounting.

It engaged a competent outside bookkeeping firm before it opened. It used a foodservice chart of accounts. It reconciled its bank accounts monthly. It filed on time, paid its sales tax on time, and produced a clean monthly profit-and-loss statement that its accountant described — accurately — as well-kept. In four years it never once failed an audit of anything.

The statements arrived, on average, nineteen days after the close of the period.

In its fourth year the restaurant closed. Its final twelve monthly statements show, in aggregate, a small operating profit. Not a comfortable one, but a profit. On paper the business was viable in every month it operated.

The operating issue

Four things went wrong, and every one of them was visible in data the restaurant already possessed.

The lag was the whole problem. A statement for March arriving on the nineteenth of April describes decisions made between one and seven weeks earlier. By the time the owner read that food cost had run 34.6% in March, it was late April, April was nearly over, and April had run at 34.9% for reasons nobody had yet examined. The reporting cycle was slower than the decision cycle it was meant to inform, so every corrective action was aimed at a month that had already ended.

The monthly period hid the pattern. Chapter 31 §31.7 explains why: a month is not a comparable unit. The restaurant's March contained five Fridays and five Saturdays and February contained four of each, so March looked strong and February looked alarming, and both readings were calendar artifacts. The owner spent a genuinely worried month investigating a February that had been fine.

Nobody computed prime cost. The statements reported cost of sales and labor as separate lines, and correctly. They never subtotalled. So the owner watched food cost, which drifted from 30% to 34% across two years — bad, and visible — and never noticed that labor had gone from 32% to 36% at the same time. Prime cost went from 62% to 70%. The number that mattered was never on any page.

And comps were in the wrong place. At the bookkeeper's reasonable suggestion, comps were recorded as a marketing expense rather than as contra-revenue. Operating profit was identical either way — which is why it seemed harmless. But it meant net sales were reported higher than they were, so every cost percentage above the marketing line was reported lower than it was. The statement understated food cost and labor percentage in exactly the period the owner was trying to diagnose them.

What it shows

The composite's failure was not a failure of accounting. It was a failure of latency and aggregation.

Run the arithmetic. In the final year the restaurant did roughly \$1,180,000, or \$22,700 a week. Prime cost ran about 68% against a 60% target — eight points.

$$0.08 \times \$1{,}180{,}000 = \$94{,}400 \text{ a year}$$

That is the whole business, several times over. And the eight points did not arrive at once; they accumulated at roughly a third of a point a month over two years. A weekly flash report would have shown the drift inside six weeks. A monthly statement nineteen days late, with no prime-cost subtotal and net sales inflated by misplaced comps, showed it in year three — as a mystery.

The most useful way to state the lesson is as a comparison of two costs:

The restaurant paid for accuracy and needed speed, and those are different products.

Outcome

In the composite the restaurant closed at the end of its fourth year, in an orderly way, with its obligations mostly met — which is the characteristic ending. There was no fraud, no catastrophe, and no single decision anyone would identify as the fatal one. The accountant's file was in perfect order.

The owner's account of it, which is the reason this case is in the book: "I had numbers. I didn't have information. Nobody ever told me those were different things."

Lesson

A statement that arrives after the decision it should have informed is a historical document.

The countermeasure is the one Chapter 1 promised in §1.3 and this chapter builds in §31.6: a weekly flash report, produced within one business day of the operating week, computing prime cost against a ramped target, and read by the person who can act on it. It is deliberately fast and approximate. It does not wait for the bookkeeper. It does not accrue. It is not a substitute for the monthly statement — it answers a different question, and the two are complements.

Three specific commitments follow, and each one is a direct fix for a failure above:

1. Subtotal prime cost, weekly. The number that predicts survival must appear on a page somebody reads.
2. Compare like periods. Use a 4-4-5 or 13-period calendar, or compare the same week year over year. Never March against February.
3. Keep comps in contra-revenue. Exercise 31.21 asks you to defend this against a bookkeeper who suggests otherwise, and the reason is exactly this case: moving them inflates net sales and quietly improves every percentage above the line.

Chapter 33 takes the argument one step further. This restaurant was profitable on paper and closed. The next chapter is about the businesses that are profitable on paper and run out of cash — which is a different failure with the same root: a number nobody looked at in time.

Discussion questions

1. The bookkeeping was accurate and the restaurant failed. Separate precisely what accurate books can tell you from what they cannot, and state which of the two this owner needed.

2. Compute the composite's eight-point prime-cost overrun in dollars, then work out how many weeks of a weekly flash report it would have taken to detect a third-of-a-point monthly drift. Show your reasoning about detection thresholds and noise.

3. The comps reclassification left operating profit identical and made every cost percentage look better. Is that a bookkeeping error, a policy error, or neither? Argue it, then say what you would have done as the owner when it was proposed.

4. Building the weekly report costs about two hours of a manager's week. Chapter 19 showed Bellwether's salaried week is already fully committed. What comes off the list to make room, and how do you defend that trade?

5. The owner says: "I had numbers. I didn't have information." Write the one-page brief you would have given this owner in year two. What is on it, and what deliberately is not?

6. This is a labeled composite. Its central claim is that latency and aggregation, not inaccuracy, destroyed the business. What real evidence would establish that in an actual restaurant — and why do the restaurants this happens to almost never produce it?