Chapter 23 — Self-Check Quiz

Twenty-four questions. Answer key in the collapsed block at the bottom — write your answers down before you open it, and do the arithmetic on paper rather than in your head.

Bellwether reference figures: $46.00 dinner check · 40% contribution · 9,035 guests × 4 visits · 36,140 covers · $1,550,000** Year 1 revenue · Hearth Chicken **$8.52 cost / $29.00 menu.

Multiple choice (1–14)

1. A comped Hearth Chicken that the guest ate and would have paid for costs the restaurant:

  • A. $8.52, the plate cost
  • B. $20.48, the contribution margin
  • C. $29.00, the menu price
  • D. $37.52

2. Guest lifetime value for Bellwether's plan guest ($46 check, 4 visits/year, 3 years, 40% contribution) is:

  • A. $73.60
  • B. $184.00
  • C. $220.80
  • D. $552.00

3. Which quadrant of the service/hospitality matrix is the most financially dangerous?

  • A. Poor service, hospitality absent — the one-star review
  • B. Poor service, hospitality present — "they were lovely about it"
  • C. Good service, hospitality absent — "it was fine"
  • D. Good service, hospitality present

4. A 20% discount on a $46 check at a 60% prime cost changes contribution per cover from:

  • A. $18.40 to $14.72
  • B. $18.40 to $9.20
  • C. $18.40 to $11.04
  • D. $18.40 to $16.56

5. At a restaurant with a 40% contribution margin ratio, a discount promotion breaks even only if the incremental share of redemptions is at least:

  • A. the discount percentage
  • B. the discount percentage divided by the contribution ratio
  • C. the contribution ratio divided by the discount percentage
  • D. 100% minus the discount percentage

6. How many five-star reviews are needed to return a 4.6 average to 4.6 after a single one-star?

  • A. 4
  • B. 6
  • C. 9
  • D. 19

7. The offset formula $x = (A-1)/(5-A)$ does not depend on:

  • A. your current average
  • B. the number of reviews you already have
  • C. the star value of the new review
  • D. the platform

8. At a 4.6 average, a four-star review:

  • A. raises the average slightly
  • B. leaves the average unchanged
  • C. lowers the average, and takes 1.5 five-stars to offset
  • D. lowers the average, and takes 9 five-stars to offset

9. The correct sequence of service recovery is:

  • A. comp it, apologize, fix it, explain
  • B. notice before they do, own it without excuses, fix the thing, then decide about money
  • C. apologize, explain what went wrong, comp the table, follow up
  • D. find a manager, comp the check, apologize

10. Bellwether's annual comp budget at 0.8% of $1,550,000 must save approximately how many guests to break even at $220.80 each?

  • A. 12
  • B. 42
  • C. 57
  • D. 124

11. Each tenth of a visit per guest per year, across Bellwether's 9,035 guests, is worth roughly:

  • A. $1,411
  • B. $14,108
  • C. $70,538
  • D. $141,076

12. Which of these belongs in a guest profile note?

  • A. "Came in with a different woman this time"
  • B. "Tips 12%, seat in the back"
  • C. "Severe shellfish allergy; prefers table 9; liked the Barbera"
  • D. "Looks like the guy from the news"

13. A post-visit survey returning a 4.7 average on a 6% response rate is best used for:

  • A. the absolute level of guest satisfaction
  • B. trend over time and the free-text themes
  • C. server performance evaluation
  • D. comparison against a competitor's rating

14. The three prerequisites for hospitality that a three-week employee does not have are:

  • A. training, uniforms, and menu knowledge
  • B. capacity, standing, and information
  • C. empathy, patience, and product knowledge
  • D. authority, incentives, and supervision

Short answer (15–24)

15. In one sentence each, define service and hospitality, and then explain why only one of them is defensible against a competitor.

16. Chapter 14 costed the second Friday in October at $116.70. Using the chapter's modeled figures — eleven tables, about 29 guests, $220.80 each — compute the cost of that night at a 10% attrition rate, and state the ratio to the visible cost.

17. A $77 recovery is spent on that Friday night. Compute the break-even: how much of one guest relationship must it preserve, and what percentage of the 29 guests is that?

18. Give two reasons the $220.80 lifetime value figure should be treated as a modeled estimate rather than a measured fact, and state the honest band you would put around it.

19. Explain why a comped entrée is judged against a probability rather than against a revenue line, and compute the break-even probability for a $29 comp against a $220.80 guest.

20. Name three things that must never appear in a public review response, and explain the specific damage each one does.

21. Why does an uncontrolled comp line distort your food cost percentage, and in which direction?

22. §23.7 lists behavioral signals to watch instead of waiting for complaints. Name four, say where each one lives, and identify the one the chapter calls the highest-value signal that nobody runs.

23. Chapter 21 priced a retention bundle at $19,716 that came out roughly break-even on turnover alone. Explain the mechanism by which this chapter's arguments produce a return on top of that, and then state honestly why you cannot prove it.

24. A guest offers to remove a one-star review in exchange for a gift card. Give the answer, and then give the reason that has nothing to do with ethics.

Answer key **1. C — $29.00.** The $8.52 of food was spent whether or not you charged for it, so the only variable is whether $29.00 of revenue arrives. Charging produces +$20.48 of contribution; comping produces −$8.52. The difference is $29.00. ($37.52 is the swing for a *re-fire* that is also comped, because the food cost is incurred twice.) **2. C — $220.80.** $46.00 × 4 × 3 × 0.40. **3. C — "it was fine."** A complaint is information: you know who, you know the table, you have a chance. A guest who had a fine time and forgot you generates no signal at all, and never returns. In a plan requiring four visits a year per guest, indifference is a larger financial event than anger. **4. B — $18.40 to $9.20.** Prime cost attaches to what the guest consumes ($27.60), not to what you charge. Revenue falls from $46.00 to $36.80; $36.80 − $27.60 = $9.20. A 20% discount cuts contribution exactly in half at a 40% margin. **5. B — the discount percentage divided by the contribution ratio.** 0.20 ÷ 0.40 = 50%. And the corollary: a discount deeper than the contribution margin ratio can never pay, because nothing is left to earn. **6. C — 9.** $x = (4.6 - 1)/(5 - 4.6) = 3.6/0.4 = 9$. **7. B — the number of reviews you already have.** $n$ cancels out of the algebra entirely. The damage depends on how good your average is, not on how many reviews you have. (Volume still matters enormously — it governs how fast a *run* of bad reviews can move you, which is why soliciting broadly is the real defense.) **8. C — lowers it, and takes 1.5 five-stars to offset.** $y = (A-4)/(5-A) = 0.6/0.4 = 1.5$. Above a 4.0 average, every four-star review is a downgrade. **9. B.** The most common error is skipping to money. A manager who says "I've taken care of your entrées" and leaves has purchased the complaint without addressing it — the guest's problem was that they waited thirty-one minutes and nobody told them anything. **10. C — 57.** $12,400 ÷ $220.80 = 56.2, round up to 57 — which is 0.63% of the 9,035-guest base. (42 is the number of guests the *overrun* in Figure 23.4 would have to save: $9,145 ÷ $220.80 = 41.4.) **11. B — $14,108.** $1,410,760 × (0.1 ÷ 4) × 0.40 = $14,107.60. One guest in ten coming one extra time in a year. **12. C.** An allergy is service-critical, a seating preference is useful, and a wine they liked is a gift you can give back. A, B, and D are gossip, judgment, and speculation respectively. The governing rule: never write a note you would not be comfortable reading aloud to the guest it describes. **13. B — trend and free-text themes.** With 6% responding, 94% of the room said nothing, and the respondents are the two tails — the delighted and the furious. The level is close to meaningless; the direction over six months and the words in the box are worth reading weekly. Answer C is actively dangerous: tying a server's outcomes to a survey score creates an incentive to solicit selectively, which is review gating with extra steps. **14. B — capacity, standing, and information.** Capacity: a new server's attention is fully consumed by mechanics, and reading a table is a residual skill. Standing: authority you have not earned is authority you will not use. Information: recognition requires having been present on the prior visit. **15.** *Service* is the technical delivery of the product — the steps, the timing, the sequence. *Hospitality* is how that delivery makes the guest feel. Only hospitality is defensible because every other asset can be taken: the menu reverse-engineered, the room copied, the price matched, the chef hired, the sequence of service duplicated in an afternoon. What cannot be copied is a neighborhood's accumulated preference for being in your room, which is rebuilt or eroded one table at a time. **16.** 29 guests × 10% = 2.9 guests × $220.80 = **$640.32**, which is **5.5×** the visible $116.70. **17.** $77.00 ÷ $220.80 = **0.35 of one guest** — about a third of one relationship out of 29 guests, a **1.2%** success rate. Note what the break-even does for you: it removes the need to estimate the attrition rate at all, which is the number you cannot measure. **18.** Any two of: (a) the 40% treats all labor as variable, but there is a fixed labor floor, so the marginal cover is worth more and the lost cover less; (b) three years is an assumption — no independent restaurant has clean guest-level retention data; (c) money later is worth less than money now, and discounting three annual $73.60 payments at 10% gives $201.34; (d) a guest whose visits include brunch is worth $187.39 at the blended $39.04 check. Honest band: roughly **$185–$225**, and every conclusion in the chapter survives anywhere inside it. **19.** Because you can never observe the return — you will not know whether the guest came back because of the comp, or would have anyway, or was lost regardless. What you *can* compute with no assumptions is the bar: $29.00 ÷ $220.80 = **13.1%**. If the failure happened on visit one of twelve, the eleven remaining visits are worth $202.40 and the bar is **14.3%**. A one-in-seven shot at a $220 asset for $29 is a bet you take every time. **20.** Any three, with the damage: **arguing the facts** teaches every reader that you will do that to them; **revealing anything about the guest** is a privacy problem that reads as a threat; **offering a comp publicly** teaches every reader the price of a one-star review; **responding the same night** produces writing nobody would approve in the morning; **a visible template** is worse than silence; **admitting fault in writing on an allergen or illness claim** creates exposure your insurer and attorney should have handled first. **21.** A comped $29.00 item removes the revenue but leaves the $8.52 of food cost in COGS. Food cost percentage is COGS ÷ food sales, so the numerator holds while the denominator falls — **food cost percentage rises**. You then spend a month hunting a portioning or purchasing problem that is actually a comp problem. Chapter 31 §31.8 handles the accounting treatment. **22.** Any four of: plates returning with food on them (the dish pit); the second drink not ordered (POS beverage attachment); dessert declined at a table that lingered past 100 minutes (POS + turn time); the check requested while food is still on the table (server observation); a booked party that shrinks (reservation platform); a second no-show from a reliable guest (reservation platform); "fine" without eye contact at a table touch (the server). The highest-value signal nobody runs: **a guest record with a 90-day gap.** **23.** Mechanism: hospitality requires capacity, standing, and information, all three of which are functions of tenure. Longer tenure raises the room's capacity to notice, recover, and recognize, which raises frequency. At $14,108 per tenth of a visit, a 0.10 improvement returns 72% on the $19,716 bundle and a 0.25 improvement returns 179% — on top of a program that was already break-even. Why you cannot prove it: you would need guest-level visit histories, server-level tenure, and a control group, and a 68-seat restaurant has none of the three. What you can honestly do is track average server tenure and repeat-visit rate side by side, quarterly, and see whether they move together. **24.** Decline, politely and briefly, and do not send anything. The non-ethical reason: paying for review removal is a transaction with a **price and a precedent**. It establishes what a one-star review is worth from you, and that information does not stay with one person. It is also squarely inside the territory the Federal Trade Commission's rules on deceptive and suppressed consumer reviews address, and platform terms prohibit it. Respond publicly if the review warrants it, invite them privately, and be genuinely generous in that private conversation — which costs the same money and buys a guest instead of a rate card.