Chapter 26 — Teaching Notes

What this chapter is doing

Chapter 26 closes Part V and has two jobs that pull in different directions.

The first is a cost job: put a defensible dollar figure and a percentage of sales on the technology stack, since almost no operator can. That work is arithmetic, it culminates in \$73,273 and 4.73%, and it is straightforward to assess.

The second is a systems job: establish that the POS is the system of record on which every metric in Parts III, IV, and VII depends. This is the harder teaching problem, because students arrive believing a POS is a cash register with a nicer screen, and nothing in their experience contradicts that. Until they see the dependency map in Figure 26.1, they do not understand that they have already spent twenty-five chapters learning to compute numbers that do not exist unless a server presses a button correctly at 7:40 on a Saturday.

Teach the second job first. If you open with the money, the room spends the session on processing rates and never internalizes why a mandatory cover count matters more than a nicer dashboard.

Timing

  • A 75-minute session: §26.1 with Figure 26.1 (15 min) → §26.7 payments, worked live, through Figure 26.8 (35 min) → §26.9 the budget and Figure 26.12 (20 min) → assign §26.2–26.6 and §26.8 as reading. This is the version to run if you only have one session; it hits both jobs.
  • Two sessions: split after §26.4. Session one is the system of record, the KDS, and integration — the data supply chain. Session two is money: payments, the budget, and the P&L placement. The break is natural and each half stands alone.
  • Three sessions (recommended for an operations program): add a session on §26.2 and §26.8 built entirely around the two contract documents (Figures 26.2 and 26.8) as a document-analysis workshop. This is the session students remember, and it is the one that transfers directly to a job.
  • Self-paced: 5–7 hours including exercises.

Common misconceptions

1. "Technology is a small line — it's just the POS subscription." This is the belief the chapter exists to break, and it is nearly universal. Before you show anything, ask the room to write down a guess: what does a \$1.55M restaurant spend on technology, all in, as a percentage of sales? Collect the guesses. In my experience they cluster around 1% to 2%. Then build the number. The gap between their guess and 4.73% is the lesson, and it lands far harder when they have committed to a number first.

2. "Payment processing costs about 2.5% of sales." Almost everyone, including experienced operators, uses net sales as the base. The gross-up — card volume includes sales tax and tips — is the single most surprising fact in the chapter. Work the build-up in §26.7 line by line on the board and stop after the tip line. Ask: whose money is that? The answer that one dollar in five of the processing bill is charged on money that was never the restaurant's is the moment the room sits up.

3. "You should shop processors aggressively; that's where the savings are." Students reliably overestimate the negotiable share. Figure 26.6 corrects it: 73% goes to issuing banks, 7% to networks, and only 20% to the company on the statement. Do not let this collapse into fatalism — that is misconception 4.

4. "So there's nothing you can do." The opposite failure. Push them to §26.9's four levers, which are worth roughly \$7,000 a year at Bellwether. The framing that works: you cannot negotiate the 73%, but you choose the pricing model, which determines how much of the 73% you actually see, and that is worth more than the markup. Figure 26.7 makes this concrete — same restaurant, same guests, same cards, \$3,475 apart.

5. "Free POS software with bundled processing is a good deal." Students take the bundle at face value because the software price is visible and the rate is not. The arithmetic in §26.2 — 28 basis points costs \$4,883, more than the \$4,800 subscription it was supposed to save — is short enough to do live and it ends the argument. Be careful to teach the honest version: bundles are often competitive and the integration is genuinely tighter. The error is not the bundle, it is failing to price it as a bundle.

6. "A KDS makes the kitchen faster." It might, marginally. That is not why you buy one. Students who leave believing the case for a KDS is speed will not defend the purchase when a chef pushes back. The case is the timestamp: without it, Chapter 14's 22-minute standard is a feeling.

7. "Integration means it works." The integration ladder (Figure 26.10) exists for this. Six rungs all get described by vendors with the same word. The rung that students consistently underrate is Rung 3 — the scheduled file that fails silently — because it looks automated. Emphasize that a silent failure is worse than a manual process, which fails loudly.

The hardest point to teach

That the fixed/variable structure of a cost matters more than its level.

This shows up three times and students miss it every time.

  • §26.6: a first-party platform fee is fixed (9.4% of channel sales at \$31,200, 5.1% at \$100,000 with no negotiation) while a commission is variable (the same percentage at any volume).
  • §26.9: \$29,700 of the stack is fixed and \$43,573 is variable, which means they do completely different things to break-even.
  • Exercise 26.37: at Year 3's \$1,850,000 the all-in percentage falls to about 4.53%, not because anything got cheaper but because the fixed half spread over more sales.

What works: put all three on the board together and ask what they have in common. Students who see it are ready for Chapter 32; students who don't will struggle with operating leverage. This is the single best predictor in the chapter of who is following the financial argument of the book.

A demonstration that works

The statement autopsy. Put Figure 26.8 on the screen with the effective-rate line blank. Give the class three minutes to compute it. They will divide fees by volume and get 2.51%.

Then reveal the composition of the \$158,400 — \$126,720 of sales, \$8,870 of tax, \$22,810 of tips — and ask them to compute it again against the restaurant's own net sales. When 2.51% becomes 2.82%, ask what that difference is worth annually. (\$4,823 at Bellwether.)

Then, and this is the part that transfers: ask what is missing from page 1. The good answers come slowly — which transactions downgraded, what the gateway fee buys, whether the deposit was gross or net. Let the silence run. The realization that a document can be entirely accurate and still tell you almost nothing is worth more than the arithmetic.

Extension if you have time: bring a redacted real merchant statement if you or a colleague can supply one. Nothing in the chapter lands like a real one, and every experienced operator in the room will discover they have never read theirs.

A second demonstration, for a shorter slot: the cover-count exercise from §26.1. Give them 123 covers and \$5,658 and have them compute PPA. Then tell them the floor entered 113. Ask which reports are now wrong and in which direction. The fact that average check goes up while covers per labor hour goes down — two managers, two wrong conclusions, no collision — is a five-minute demonstration that students remember for the rest of the course.

Assessment notes

  • Exercises 26.18–26.22 are the computational core. A student who can build card volume from net sales, compute an effective rate two ways, and compare interchange-plus against flat rate can do everything financial in this chapter.
  • Exercise 26.28 (the statement autopsy) is the best single diagnostic. Strong students find the non-compliance fee and notice that 26 batches against 30 open days means four batches were missed. That second catch is the one to reward — it is the only place in the exercise set where the arithmetic points to an operating failure rather than a billing one.
  • Exercise 26.16 (fixed/variable classification) predicts performance in Chapter 32 better than anything else here. Worth assigning even in a compressed course.
  • Exercise 26.35 (the surcharge argument) is the best writing prompt in the chapter. Look for students who resist deciding on the arithmetic alone and who name the measurement asymmetry — \$26,000 you can count against damage you cannot. Penalize confident answers in either direction that do not acknowledge it.
  • Exercise 26.13 (the guest note) is short, uncomfortable, and reveals a great deal about a student's instincts. Run it in class rather than assigning it.
  • Exercise 26.37 is the Business Plan extension and the natural exam question.

Connections forward and back

Flag these explicitly so students build the map: POS → Chapters 11, 12, 13, 19, 22, 24, 31, 34 (nearly the whole second half of the book). KDS → Chapter 14's 22-minute standard and 28-items-per-hour hearth. Reservation platform → Chapter 22's pacing caps and Chapter 23's recognition. Online ordering → Chapter 28, which owns the economics and which this chapter deliberately does not pre-empt. The fixed/variable split → Chapter 32. Settlement timing → Chapter 33. The audit trail → Chapter 34. Marketing placeholders in Figure 26.12 → Chapter 27.

Backward: this chapter is where Chapter 1's "counting problem" becomes concrete. The restaurant that believed its food cost was 30% for eleven months did not have a software problem — but it also had no system that would have told it otherwise. Make that connection out loud; it closes a loop the students opened in week one.