Chapter 4 — Key Takeaways

The core claims

1. A plan is an argument, not a prediction. Nobody believes a first-year restaurant forecast is accurate. The reader is testing whether you know which assumptions carry the weight, what each one rests on, and what you would do the Monday you found out it was wrong.

2. Write for the lender. Four of the five readers want to know what they get. Only the lender is professionally paid to imagine your failure — satisfy that reader and the others are handled.

3. The credibility of your forecast is your collateral. A leasehold improvement and a used hearth do not secure a loan. A revenue line a reader can trace to a physical constraint does.

4. Build revenue bottom-up; cross-check it top-down; never build top-down. Seats × turns × average check × services. A market-share forecast does not eliminate the turns assumption — it hides it.

5. The assumptions register is the chapter's instrument. One table, every belief, with its value, confidence, plain-language basis, dollar exposure per unit of movement, and the chapter or the week in which it gets tested. A plan without a register hardens. A plan with one stays alive.

6. Express exposure in dollars, never in adjectives. A point of food cost is $11,160, a point of labor is $15,500, a point of pour cost is $4,340. You can sort dollars by consequence. You cannot sort the word "significant."

7. Percentages that divide a fixed dollar cost by revenue are restatements of the revenue assumption. Occupancy at 6.1% is really "$95,200, and we believe the sales number." The rent does not care about your forecast.

8. Every table must foot, and the use of funds must foot to the ask exactly. A reader who finds one column that doesn't add up stops trusting all of them.

9. Break your own forecast on purpose. Sensitivity analysis produces a ranking, not a range — it tells you which assumption to watch first, defend hardest, and stop worrying about.

10. Name the gap. Bellwether's base case is $1,410,760 and its headline is $1,550,000. The plan prints both, itemizes the $139,240 into four named claims, discloses $1,240 of rounding, and flags the whole thing as its lowest-confidence row. A weaker plan would have nudged turns until the arithmetic landed on a round number.

11. Write the executive summary last, and put the three unresolved risks on it. A risk the reader finds unnamed discounts every other number in the document. A risk you named, with a dollar consequence and an action attached, tells them the rest is probably sound.

The formulas

$$\text{Revenue} = \text{seats} \times \text{turns} \times \text{average check} \times \text{services}$$

$$\text{Covers per service} = \text{seats} \times \text{turns} \qquad \text{Turns} = \frac{\text{covers}}{\text{seats}}$$

$$\text{Operating profit} = \big(\text{Revenue} \times \text{contribution per dollar}\big) - \text{fixed costs}$$

$$\text{Exposure per point} = \text{the base the percentage applies to} \times 0.01$$

The Bellwether numbers to carry forward

Exposure per unit of movement — the register's most useful column

Key terms

business plan · executive summary · pro forma · bottom-up sales forecast · assumptions register · use of funds · sensitivity analysis

What you should be able to do Monday morning

Take any restaurant forecast handed to you — yours or someone else's — and in under ten minutes: convert the revenue back into seats, turns, check, and days; ask what turn rate the cover count implies and whether the room can physically do it; find the number of operating weeks assumed; and name the one assumption that, if wrong by a plausible amount, costs the most dollars. Then ask the only question that matters about each of them: how do you know?