Chapter 27 — Instructor Material

Teaching Notes

What this chapter is really for

Students arrive at a restaurant marketing chapter expecting channels, and the chapter is not about channels. It is about what a budget too small to do everything is therefore obliged to do, and the answer it reaches is the one nobody in the room predicts: at \$23,250 against a \$1,550,000 plan, acquisition is not the risk. A first cover costs \$2.57 and is worth \$18.40 immediately, and it still costs under ten dollars after you cut attribution to a third and price the owner's unpaid hours at replacement cost. The risk is the fourth visit, which Chapter 23 valued at three times the whole budget and which no line of it can buy.

Everything else hangs off that. The section order is the argument: what a cover may cost (§27.1), the free foundation almost nobody maintains (§27.2), the two things reviews and social actually do (§27.3–27.4), the only asset that appreciates (§27.5), the two things that cannot be scheduled (§27.6, §27.8), the mechanic with a computable bar (§27.7), and an honest reckoning with what can be measured at all (§27.9).

There is a second, quieter thing the chapter is doing, and it is worth naming for students on day one: it refuses two numbers on purpose. It will not tell you what a star is worth in revenue — Chapter 23 refused that claim and derived x = (A − 1) ÷ (5 − A) instead — and it will not give engagement benchmarks. Both refusals are stated in the text. Students who have taken a marketing course elsewhere will notice the absence and should be told why: a book that opened by dismantling the industry's most famous unsourced statistic cannot then quote an elasticity it did not verify. That is a lesson about evidence that will outlive everything else in the chapter.

Budget two sessions. One session forces you to drop either the arithmetic or the neighborhood, and both are load-bearing.

Common misconceptions, in the order they surface

"Marketing's job is to get people in the door." Only 25% of it, and the chapter's whole allocation follows from that. Three-quarters of the plan's covers are return visits, and marketing buys none of them. Land this in the first ten minutes or the rest of the chapter reads as a strange set of priorities.

"\$2.57 against \$18.40 means spend more." See the hardest point below. This is the single most common and most reasonable-sounding wrong answer in the chapter.

"A four-star review is a good review." It is a downgrade at any average above 4.0, which is arithmetically obvious and emotionally very hard to accept. The correction that matters is not "chase five stars" — it is stop counting four-star reviews as good news in your reporting, and understand that the defense of a high average is the volume of genuinely delighted guests, not the absence of unhappy ones.

"Undoing a bad review is about the review." It is about the operation. Nine five-stars at a 4.6 average, thirty-two to undo one bad Saturday — and the mechanism that produces them is staffing, pacing, and a manager who touches the table at minute twenty. The best review-management program in a restaurant does not have the word "review" in it.

"More reviews means more exposure to damage." Backwards. The identity is count-independent, but the drop is not: one one-star takes a 4.6 average to 4.32 at twelve reviews and to 4.59 at five hundred. The first hundred reviews are the fragile period, which is an argument for building volume early and deliberately.

"Followers are the objective." Social is a consideration channel, not an acquisition channel. Figure 27.6 makes this concrete: Instagram produced 97 covers of 3,240 in a month, against 559 from "a friend brought me." A restaurant with 8,000 followers and a wrong phone number on its Google profile is losing money it cannot see.

"A redemption is an acquisition." The denominator error, and the one that produces the most expensive mistakes downstream. If you hand a \$10 card to somebody already standing at your host stand, they were already standing at your host stand. Chapter 28 will make this worth six figures.

"The discount is the cheap option." It carries the only hard, computable break-even bar in the chapter — 54.3% of redemptions must be incremental at \$10 against \$18.40. Access and recognition mechanics set D = \$0 and therefore have no bar at all. A 68-seat room has scarcity to give away and should give that away instead of money.

"Press is good news." Press is a demand-side event that does nothing to your cost structure. §27.6 models a write-up that nets −\$4,839 because a 123-cover Saturday walked in at 160 and the first impression a hundred new guests formed was of a kitchen that could not keep up.

"Cash from gift cards is revenue." It is a liability until redemption, and it will make a December bank balance look far better than a December P&L. Students get this wrong nearly universally; Exercise 27.15 exists to catch it.

**"The \$0 plan is free."** It costs about five hours a week, forever — \$6,500 a year at replacement cost, 28% of the marketing budget, appearing nowhere in it. The chapter is explicit that this is the best trade available and equally explicit that it is not free.

The hardest point to teach

Why \$2.57 against \$18.40 does not mean "spend more."

The arithmetic is clean, the ratio is 7:1, and every business student in the room reaches the obvious conclusion: if a cover costs \$2.57 and returns \$18.40 on the same night, buy more covers. It is a sound instinct and it is wrong here, and the reason is not an ROI argument — it is a constraint argument, which is a genuinely different shape of reasoning and the reason this is the hard one.

The chapter's answer has three parts, and students need all three:

1. There is no more room to sell. Figure 27.5 has Friday at 176% and Saturday at 181% of seats at a single turn. The covers you would buy have nowhere to sit.
2. The demand may not exist. Chapter 2's capture requirement is 7.7% of the ten-minute drive-time population, and this chapter says flatly that marketing does not create demand, it competes for it. Bellwether adds ~40% to the direct-occasion supply on the block; more spend buys a bigger share of a fixed pool.
3. The covers the plan actually needs are not for sale. 27,105 of 36,140 are return visits. No price clears that market. Chapters 22, 23, and 24 do.

How to land it. Put the ratio on the board and ask the question straight: "It costs \$2.57 and it's worth \$18.40. Why not spend \$100,000?" Let the room answer for a full three or four minutes — they will reach for diminishing returns, which is the plausible wrong answer, because it implies the right amount is merely larger. Then put Figure 27.5 next to it and ask where the covers would sit. The room reorders itself. Only after that, add the third part: even if the room were empty, the covers the plan needs are the fourth ones, and those are not on the market at any price.

Do not give them the constraint before they have committed to the ROI answer. The commitment is what makes the correction stick.

A demonstration that works

Derive the offset identity live, and let them watch the review count disappear.

Four minutes, a whiteboard, no materials. Give the class a 4.6 average and one new one-star review, and ask how many five-stars it takes to get back. Take guesses first — you will hear one, two, and four. Then derive it with them:

Let the current average be A over n reviews, so the total is nA. Add one 1-star and x 5-stars:

(nA + 1 + 5x) ÷ (n + 1 + x) = A

nA + 1 + 5x = nA + A + Ax

1 + 5x = A + Ax → x(5 − A) = A − 1 → x = (A − 1) ÷ (5 − A)

The moment is when n cancels. Stop there and say it out loud: this holds at 40 reviews and at 4,000. Students physically react to a count-independent result, and the reaction is the hook.

Then run it: 4.6 gives 9. 4.8 gives 19. 4.9 gives 39. Ask what that shape means for a restaurant that is genuinely excellent — the better you are, the more expensive one bad night becomes — and close with the epistemological point, which is the actual payload: this is a definition, not a statistic. It needs no source, cannot be out of date, and is the reason the chapter can refuse the number everyone else quotes.

Two alternates if you want variety.

• The live profile audit. Put a real local restaurant's Google Business Profile on the projector and run §27.2's field table against it in five minutes. Essentially every audit finds at least one broken thing — a PDF menu that will not open on a phone, no special hours, no exterior-at-night photo, an unanswered question in the Q&A, a pin in the wrong place. Pick a business nobody in the room works for, and do not name it in any handout or recording.
• The comp-code sort. Hand out the fourteen year-one budget lines from the Business Plan checkpoint and ask, for each: what evidence would this produce? Students discover on their own that most of the budget generates none, which is a far better setup for §27.9 than any lecture about attribution, and it makes the 46% unattributed block in Figure 27.6 feel inevitable rather than sloppy.

Timing (two 75-minute sessions)

Session one — what a cover may cost, and the free foundation. 0:00 Overview and the 9,035 × 4 decomposition (10) · 0:10 §27.1, cost per cover acquired and the four honest versions, worked on the board (20) · 0:30 the "why not spend \$100,000?" discussion (12) · 0:42 §27.2, the live profile audit and the previous-tenant problem (15) · 0:57 §27.3, derive the offset identity (10) · 1:07 the four-star problem and the fragile first hundred (6) · 1:13 assign 27.9, 27.13, 27.16, 27.23 (2).

Session two — frequency, measurement, and the neighborhood. 0:00 §27.4 what social can and cannot do; the cadence that survives a Friday (10) · 0:10 §27.5 the list, the break-even, and the two compliance regimes (15) · 0:25 §27.7 the D/C bar, built with the class rather than presented, then the visit-count distribution (20) · 0:45 §27.9 Figure 27.6 and the comp-code sort (15) · 1:00 §27.8 the neighborhood discussion (12) · 1:12 Business Plan assignment (3).

Build the D/C bar live. Do not hand it over finished. Ask: "We give \$10 off. Who pays for the people who were coming anyway?" Push until somebody says the discount is paid on every redemption and the contribution is earned only on the incremental ones. Write it as value = rN·C and cost = N·D, set them equal, watch N cancel, and land on r ≥ D ÷ C. Then change one variable — make the reward access instead of money — and let them work out that D = 0 makes the bar disappear. That last sixty seconds is the most useful thing in §27.7 and it does not survive being presented.

Assessment guidance

27.9, 27.13, 27.16, and 27.23 form a coherent problem set: what a cover costs, what an offer must achieve, what the cheapest channel must achieve, and what a real campaign actually costs. Assign them together; they take about ninety minutes and they cover the chapter's entire computational spine.

27.23 is the discriminating computation. A weak student computes \$38.98 at 60% incrementality and stops. A strong student asks what incrementality would clear \$18.40, finds that it requires 127% — impossible — and identifies the cause: the \$960 offer is 42.8% of the campaign. The offer, not the media, is what breaks it. That inversion is the mark of a student who has understood the numerator rule rather than memorized it. Weight it accordingly.

27.18 is the best analytical item. The tell is whether the student challenges the code design rather than the business. The strongest available observation is that GC-FIRST is classified as frequency while its own name says these are first visits — reclassify it and the 72/28 split inverts to 43/57. A student who finds that has learned the most transferable lesson in §27.9: a measurement system reports the shape of its own instrumentation.

27.31 is the best judgment item and the one to grade hardest. It has a right answer with a counter-intuitive core: at a \$29 check, contribution falls to \$11.60, and the \$10 offer's break-even incrementality rises to 86.2% — so the price-sensitive neighborhood is exactly where you can least afford to discount. Also watch for whether the student notices that cost per cover acquired does not move (neither the budget nor the first-visit count changed) while every ceiling above it falls 37%. Mark down any answer that "solves" the \$29 concept by enlarging the marketing budget; that is §27.1's percentage-of-sales fallacy with the sign flipped.

27.29 is the best ethics item. The mature answer is not "no" — it is no, plus a counter-offer that costs nothing, plus a statement of what would change my mind. Answers that only decide are incomplete. The sharpest available test, and few students reach it: would the partner be willing to write this as a cash purchase of a positive post, disclosed as such? If that sentence is uncomfortable, it is the same transaction.

Mark down anywhere in the chapter's assessments: an answer that quotes a review-elasticity figure or a social engagement benchmark. The chapter refuses both by name, and reproducing an unsourced industry statistic is exactly the professional habit this book is trying to break. It is worth saying in advance that you will be doing this.

Connections

Backward, this chapter is almost entirely built on Chapter 23 — every ceiling, the \$70,538 half-visit prize, the \$98,223 reach cost, and the offset identity all originate there — and on Chapter 2, which produced the 9,035 × 4 decomposition that the entire allocation follows from. Chapter 1 supplies both the press-does-nothing-to-cost-structure argument and the epistemic warrant for refusing unsourced numbers.

Forward, Chapter 28 takes the deliberate opposite case: where these channels were nearly free and hard to measure, off-premise is expensive and exactly measurable, and it formalizes the incrementality problem introduced here. Chapter 29 answers whether private events can fill the Tuesdays six events cannot. Chapters 31, 33, and 34 pick up the three financial loose ends — where the offer's discount cost lands on the P&L, whether the cash forecast can carry \$5,250 of pre-opening spend and what December's gift-card cash really is, and who reconciles the comp report to the marketing line. Chapter 35 carries the warning not to expand on a press spike, and Chapter 39 carries the open question of what happens if the review average settles at 4.2.

A note on tone

Two passages need care.

§27.8, on the neighborhood. The chapter says plainly that Bellwether's 68 seats add ~40% to the direct-occasion supply, that some of its covers will be taken from operators already there including a fifteen-year-old family restaurant, and that the \$46 check is levered to a gentrification process raising rents on the people who were there first. It also says the chef-owner is a cook with a personal guarantee, not a developer, and that restaurants are among the earliest and most exposed businesses in a changing neighborhood. Hold both directions. Classes drift to one pole or the other — either the restaurant is a villain or the discomfort is somebody else's problem — and the chapter's position is that both facts are true and neither is resolvable by a marketing plan. The sentence to keep in front of the room is the chapter's own test: if you would be embarrassed to have that other owner read the caption, do not post it.

§27.3 and §27.5, on compliance. These are structural, not advisory, and every one closes with verify-locally. Students will ask what the penalty is, whether anyone actually enforces it, and whether their cousin's restaurant got away with it. Answer the structure — statutory damages assessed per message, on a list of 600 people, against a marketing budget of \$23,250 — and then decline the prediction. Never let a compliance discussion end at "in practice nobody checks." That sentence is the beginning of every expensive story in this book.