Chapter 28 — Instructor Material

Teaching Notes

What this chapter is really for

Students arrive wanting a verdict on delivery, and they expect it to be yes or no. The chapter refuses both and answers "it is a schedule." If the session works, that reframe is what they take away, and it transfers: most operating questions that present as binary are actually questions about when, or how much, or for whom.

Two organizing ideas carry the whole thing.

One: every channel a restaurant adds is a separate business line with its own margin structure that happens to share your kitchen. Delivery, catering, retail, a second daypart. Price each one on its own numbers or one of them will quietly fund somebody else's growth.

Two: contribution per order is the wrong denominator whenever something is scarce. This is the chapter's real intellectual content and it generalizes past restaurants entirely.

It is also the book's cleanest demonstration that a percentage can be arithmetically correct and completely useless. A 29% food cost on a delivery order is a true statement that will bankrupt you.

Budget two sessions. One forces you to cut either §28.2 or §28.6, and the chapter does not survive either cut: §28.2 without §28.6 produces students who think delivery is fine at 25%, and §28.6 without §28.2 produces students who feel the constraint but cannot price the trade. If you genuinely have one session, see the timing note below.

Common misconceptions, in the order they surface

"It's incremental, so any margin is good margin." The most expensive sentence in the chapter, and it is half true, which is exactly why it survives. The fixed-cost-absorption argument is legitimate — and only when the resource was going to sit idle. Almost nobody states the condition. Give them the test question early and make them repeat it all session: what would that resource have done in the next ten minutes if this order had not existed?

"The commission is the problem." The single most common wrong answer, and the chapter is engineered to break it. On the \$65 order, commission at 25% takes \$16.25; the beverage nobody ordered was carrying \$20.09. Students who genuinely internalize this have got the chapter.

"Our delivery food cost is only 29%, so it's our efficient channel." A percentage measured against a denominator you never receive. Drill it numerically: 29% ÷ 0.75 = 38.7% against net remittance, and 29% ÷ 0.65 = 44.6% at a realistic effective take rate. Then Chapter 12's rule: bank dollars, not percentages.

"Sales are flat, so the channel shift is a wash." Total sales cannot detect cannibalization — by construction. This one is hardest for the students who like data, because their instinct is that more data would settle it, and more of the same data never will.

"Commission hurts your expensive items because they're expensive." No. It hurts in proportion to food cost: the share of contribution margin a commission consumes is commission ÷ (1 − food cost %), and price cancels out entirely. A \$16 salad and a \$38 short rib differ only because their food costs differ. This surprises nearly everyone, including instructors who have taught the chapter before, and exercise 28.15 exists to make them prove it to themselves.

"An uplift fixes it." Show them the gross-up. A 30% commission requires a 49% uplift, not a 30% one, because the fee is charged on the raised price and packaging has to come through the same gross-up. Then the \$40.27 Hearth Chicken, which nobody is paying.

"A commission cap makes the channel fine." Fifteen percent on a business that keeps five is still enormous, and every other cost in §28.2 is unchanged.

"First-party is obviously better, so why does anyone list?" If you let the class dismiss the reach argument, they will sign a marketplace agreement at 28% within three years, because they will never have been taught what it actually buys. §28.4 makes the case properly. Make them make it too.

"Delivery was a technology change." It was a habit change. The apps existed before 2020 and after. What persisted was a saved card, a default address, and forty completed orders.

"Hire a packer." At Bellwether the dedicated position costs \$129.90 a week more than the entire channel produces. Not "reduces the margin" — consumes it, and then some.

The hardest point to teach

Contribution per unit of the constraint. Students can compute contribution per order all day. Changing the denominator is the leap, and it is a genuine conceptual jump rather than a harder sum.

It is hard because it requires holding two true things at once. The delivery order really does earn \$25.89, and that really is good money. The hearth slot it consumed really was producing \$52.71. Neither number is wrong, and students badly want one of them to be the answer.

The way to land it: withhold the subtraction. Put both numbers on the board separately. Ask "is a \$25.89 order a good order?" They will say yes, and they are right. Then ask "what was that hearth slot doing at 7:40 on a Saturday?" Then write the minus sign. Do not present −\$26.82 as a fact to be accepted — the reordering of their own reasoning is the entire lesson, and it does not survive being handed to them finished.

Then run the payoff immediately: the same order at 6:15 on a Tuesday. Same food, same commission, same packaging — and the fire is at 45% of rate, so displacement is zero and the \$25.89 is real, incremental contribution. A swing of \$52.71 decided entirely by the clock. That is the moment the room stops arguing about delivery and starts arguing about hours, which is where you want them.

A demonstration that works

Build the three-column board with the class, and let them get it wrong first. Twenty minutes, needs nothing but a board.

Write the \$65 order across the top — Hearth Chicken \$29, hand-cut pasta \$24, salad \$12. Have the class compute the plate cost (\$19.32), then the packaging (build the card live, component by component — they will guess \$0.75 and it is \$2.24), then the packing labor. Now put up three columns side by side and have them fill each one: dine-in two covers \$61.76 · first-party pickup \$38.65 · marketplace at 25% \$25.89.

Then ask: "Point at the single biggest difference between column one and column three."

Every hand points at the commission. Let that sit. Then circle the \$25.76 of beverage in column one and write \$20.09 under it.

It is the best twenty minutes in the chapter, because they built the wrong answer themselves and you did not have to argue them out of it.

Two alternates, both cheap:

• The container demo — physical, and worth the trouble. Twenty minutes before class, seal something crisp in a closed takeout container; put an identical portion in a vented container with a paper liner; leave a third on a plate. Open all three in front of the room and pass them around. The condensation on the inside of that sealed lid is the entire physics of §28.5 in one object, and nobody who has held it forgets Figure 28.2. Cost: one takeout order.
• The blind rate reveal. Give each small group a different commission rate — 15, 20, 25, 30 — without telling them the others have different numbers. Same order, same menu, same question: keep or kill? Then reveal. The groups reach genuinely opposed recommendations from identical food, which makes the point that the decision hinges on a number they do not control and cannot negotiate (§28.9) better than any lecture about leverage.

Timing (two 75-minute sessions)

Session one — the money. 0:00 the framing question, "should we do this at all," and §28.1's conditional (10) · 0:10 §28.1, the four changes — ask for them before you give them (12) · 0:22 the three-column board, built with the class (20) · 0:42 the packaging cost card, built live from components (8) · 0:50 Figure 28.4, the payout statement: what is the take rate, and which lines are choices? (15) · 1:05 §28.5's uplift arithmetic and the \$40.27 line (8) · 1:13 assign 28.12, 28.15, 28.16, 28.20 (2).

Session two — the constraint. 0:00 §28.3, incremental vs. cannibalized, and why a sales figure can never tell you (15) · 0:15 the break-even rate and the six tests (15) · 0:30 §28.6, the hearth at the binding hour — the two-number reveal (20) · 0:50 Figure 28.6, build the window with the class rather than showing it (10) · 1:00 the fifth-person arithmetic (8) · 1:08 §28.9, caps, leverage, and the contract-clause list (5) · 1:13 the Business Plan assignment (2).

If you truly have one session: keep the three columns and the two numbers. Cut §28.7 and §28.9 entirely and assign 28.16 and 28.35 to carry them. Do not cut §28.3 — a student who can compute channel contribution but has never been asked whether the sale was incremental has been made more dangerous, not less.

Assessment guidance

28.20 → 28.21 → 28.22 is the spine. Assign the three together. 28.20 computes the channels, 28.21 converts the same numbers into a tolerance for being wrong, and 28.22 shows the case against peak-hour delivery strengthening as the restaurant improves.

28.22 is the discriminating item. A strong student notices that the physical pressure eases — hearth load falls from 29 to 26 items, 104% down to 93% of rate — while the economic case hardens, and says both out loud. A weak student computes \$66.42 and stops. Weight it accordingly. The underlying insight is genuinely counterintuitive: a busier, higher-check, better-performing restaurant is less able to absorb a delivery channel, not more.

28.23 is the best item for business students: prime cost can improve while profit is destroyed, which is why prime cost is a management ratio and not a channel-selection tool.

28.29 is the best sorting item. Grade the justifications, not the placements. The trout, the pasta, and the chocolate tart are all legitimately arguable; fried items, crudo, dressed leaves, and anything that melts are not.

28.35 is the best diagnostic. The tell for a mature answer is whether the student notices packaging at 10.3% of menu sales and flags the ticket-size caveat before quoting a saving. A student who confidently banks \$32,800 a year without asking what the average order is has learned the arithmetic and not the discipline.

28.41 is the best ethics item, and the tell is subtle: does the student engage the food argument, or only the honesty argument? "Lying is wrong" is a C. Noticing that the proposal also degrades the food and probably fails to move the metric it targets is an A.

28.43 is the best Business Plan item. The tell: does the student see that the break-even cannibalization rate does not change with volume — it is a ratio of per-order contributions — while the dollar consequence of being wrong doubles?

Mark down, anywhere in the chapter's assessment: quoting a specific platform's commission rate as though it were a published fact; asserting a market share; computing contribution per order in the presence of a stated constraint without changing the denominator; and treating "sales are up" as evidence about a channel.

Connections worth making out loud

• Chapter 12 returns with one axis swapped. Make them notice that the quadrants moved — an item that arrives wrong has negative popularity, because it costs you the next order too.
• Chapter 15 is the punchline. If you taught beverage margin well, the \$25.76 lands in a sentence.
• Chapters 22 and 24 are the constraint argument. Without the shape of the week and the 132-cover ceiling, §28.6 is assertion rather than evidence.
• Chapter 27 supplies the acquisition framing. Commission as a cost per cover acquired is the only reading under which a marketplace is rational for a restaurant this size.
• Chapter 29 is next week, and the same idle Tuesday gets a second bidder. Say so at the end of the session. Catering will probably win that auction, and students should see it coming rather than discovering that this chapter's conclusion gets partly overturned.
• Chapter 30 is where the students who are genuinely excited about this channel get a proper hearing. Tell them, so their enthusiasm does not read as having been dismissed.
• Chapter 34 — platform promotions and sponsored listings are a discount nobody had to authorize. Same failure, different dashboard.

A note on tone

Some of your students have worked this channel — as drivers, as the person packing bags, as the person who was supposed to be watching the tablet. Their stories are the most valuable material in the room and they will not offer them unless you make it safe to.

Two things keep the session honest. First, the reach argument is real, and §28.4 makes it well; a class that leaves believing platforms are simply predatory has learned less than one that leaves able to read a payout statement. Second, 6.9 of the 10 points of take-rate gap in Figure 28.4 are the operator's own choices — promotions and ads nobody turned off. Point at that every time the conversation drifts toward grievance.

And hold one line about drivers. The chapter says the tip goes to someone you do not employ, cannot train, and cannot discipline, and that is true — but it must not slide into contempt. The driver is working a hard job under a dispatch algorithm they did not design, and the general manager's proposal in exercise 28.41 makes that driver wait unpaid at your door to improve your score. If a student proposes it approvingly, that is the moment to say so.