Question Bank
This is a pooled bank of assessment items for Restaurant Management: Behind the Pass — roughly 240 items, six for each of the forty chapters, followed by an answer or a worked solution for every one.
It is not the chapters' exercise sets and it is not their quizzes. Those files are built to be worked in sequence, immediately after the reading, and they lean on the chapter's own figures. The items here are built to be recombined. Each one is self-contained: it carries its own data, names its own restaurant, and never assumes the class has seen a particular cost card or a particular P&L. You can pull item 11.3, item 19.4, item 28.6, and item 32.2 into a single ninety-minute exam and none of them will contradict each other or reference a figure the students have not been shown.
Every business in this bank is constructed. Every dollar figure is illustrative. Where an item draws on the book's running business-plan project, it is labeled [the Bellwether plan]; everywhere else the item builds its own café, bar, truck, taqueria, bistro, or catering operation, on purpose, so that a student who has memorized Bellwether's numbers gains nothing and a student who has learned the method gains everything. Fewer than one item in six uses Bellwether at all.
How to use the bank
The four item types.
- Recall — a definition, a distinction, a sequence, a threshold. These take a student thirty to ninety seconds and they are the only way to find out whether the vocabulary is actually there. Do not build an assessment out of them, and do not omit them.
- Computation — arithmetic with a determinate answer. This is roughly forty percent of the bank, which reflects the book: a manager who cannot do this arithmetic under time pressure cannot do the job. Every computation item here has been checked, and the answer key shows the steps rather than the result alone, so partial credit is assignable.
- Judgment — a short scenario, a decision, and no single right answer. The answer key does not give an answer; it gives what a strong response contains. Grade against that list. A student who reaches the opposite conclusion from the key's example, for defensible reasons, and who names the costs on both sides, has earned full marks.
- Artifact — a small document rendered in a
```textfence: three lines of a schedule, half a cost card, one row of a flash report, a settlement statement, an inspection excerpt. The student reads it and acts on it. These are the closest thing in the bank to the actual work, and they are the items students find hardest, because nothing in the artifact tells them which calculation to run.
The three difficulty levels.
- Level 1 — one step, one concept. A quiz item. A prepared student answers it in under two minutes.
- Level 2 — two or three steps, or one calculation plus a judgment about what the result means. An exam item. Budget four to six minutes.
- Level 3 — multi-step, and structured so that the obvious calculation is the wrong one. Something in the item is a trap: a rounded percentage that should not be reconstructed, a displaced sale that should be netted out, a cost that is not recoverable, a base that is not the right base. Budget ten to fifteen minutes, and expect a normal class to split roughly in half on them.
The distribution is roughly 40% Level 1, 40% Level 2, 20% Level 3, and it is deliberate.
Items are not interchangeable across difficulty. This is the standing warning and it is worth stating plainly: three Level 1 items are not a substitute for one Level 3 item. They test a different thing. A student can answer every Level 1 item in this bank correctly and still be unable to look at a distressed statement and say which line is recoverable — which is the entire skill the book exists to teach. If you assemble an exam from Level 1 items because they grade quickly, you will have measured vocabulary and called it competence. Keep at least one Level 3 item on every graded assessment longer than twenty minutes, and weight it accordingly.
Assembling a quiz (15–20 minutes, one or two chapters): four to six items, all Level 1 and Level 2, at least one Computation and one Recall. Skip Artifact items unless the students have the chapter's artifact conventions fresh.
Assembling a midterm or final (60–120 minutes, one or more parts): twelve to twenty items. Aim for roughly 30% Level 1, 45% Level 2, 25% Level 3 by point weight — which means Level 1 items should be worth two or three points and Level 3 items worth fifteen to twenty. Cover at least three of the four types. Always include at least one Artifact item; it is the only type that tests whether a student can find the calculation rather than perform it.
Assembling a problem set (take-home, a week): four to eight items, weighted toward Level 3, with at least two Judgment items so that the written argument carries real weight. Problem sets are where the Judgment items belong; they are hard to grade fairly under time pressure.
A note on grading computation. The book's own rule applies to the bank: dollars are canonical, percentages are rounded displays. Several items — 31.6 most explicitly — are built on the fact that reconstructing a dollar figure from a displayed percentage produces a wrong answer. Accept any answer that is correct to the cent from the dollars given, and treat a student who reconstructs from percentages as having made a substantive error, not a rounding one.
Part I — The Business Behind the Food (Chapters 1–5)
Chapter 1 — Why Restaurants Fail
1.1 [Recall · Level 1] A trade-magazine article, a banker, and a culinary instructor all tell a prospective operator the same thing: "Sixty percent of restaurants fail in their first year — some people say ninety." Which of the following is the defensible statement of what the published research actually shows? (a) 90% close in year one. (b) 60% close in year one. (c) Roughly 26–27% close in year one, with cumulative closure approaching 60% by the end of year three. (d) Failure rates cannot be estimated. State your choice and add one sentence on why the correction makes the picture more worrying rather than less.
1.2 [Recall · Level 1] Define a cover, and distinguish it from a check and from a table. A party of four orders on one check and a couple at the bar orders on two separate checks. How many covers, how many checks, how many tables?
1.3 [Computation · Level 1] A neighborhood café reports weekly sales of \$38,000, cost of goods sold of \$11,400, and total labor including payroll taxes and benefits of \$13,300. Compute prime cost in dollars and as a percentage of sales, and characterize it against the full-service rule of thumb.
1.4 [Computation · Level 2] A 50-seat bistro plans dinner five nights a week, 50 weeks a year, at 1.6 turns and a \$38 average check. (a) Estimate annual dinner revenue. (b) The owner believes a better beverage program can move the average check to \$40 with no change in covers. What is that worth annually? (c) Which of the two variables — turns or check — would you test first, and why?
1.5 [Artifact · Level 2] A taqueria's annual summary:
ANNUAL SUMMARY — 44-seat counter-service taqueria [constructed teaching example]
Revenue $780,000 100.0%
Cost of goods sold $249,600 32.0%
Labor, all-in $296,400 38.0%
Occupancy $54,600 7.0%
Compute prime cost in dollars and percent. The owner asks which half to attack first. What do you need to know before answering, and what is your provisional recommendation?
1.6 [Judgment · Level 3] An operator has been open fourteen months. Sales are steady, the reviews are good, and she tells you she is "past the dangerous part, statistically." Write a two-hundred-word response. Your response must correct the statistical claim, explain what the shape of the survival curve implies about causes, and name the specific things about months fifteen through thirty that make them more dangerous than months one through twelve, not less.
Chapter 2 — Concept and the Competitive Set
2.1 [Recall · Level 1] "Modern Italian" is not a concept. Explain what a concept actually is — the claim it makes and about whom — and state the three things a concept must specify that a cuisine does not.
2.2 [Computation · Level 1] A trade area within a ten-minute walk contains eleven restaurants at a comparable price point, with 1,140 seats among them. A new 60-seat restaurant opens. What share of the district's seat supply does it represent?
2.3 [Computation · Level 2] Analysis of a trade area suggests it supports about 2,400 dinner covers per week across eleven restaurants in a given price band. A twelfth restaurant is planned, needing 95 covers a night, five nights a week. (a) What weekly cover count does the plan require? (b) What share of the trade area's covers is that? (c) What would an equal share be? (d) State in one sentence what the gap between (b) and (c) obligates the plan to explain.
2.4 [Judgment · Level 2] A partner proposes "elevated comfort food, generous portions, approachable prices" for a district that already has three restaurants a guest would describe the same way. The food is genuinely better. Is that a concept? What would you change, and what evidence would tell you the change worked?
2.5 [Artifact · Level 2] A competitive-set worksheet:
COMPETITIVE SET — six blocks, dinner [constructed teaching example]
# Style Seats Avg check Turns Open nights Reservations
1 Trattoria 84 $34 1.6 6 yes
2 Brewpub 120 $27 1.9 7 no
3 Steakhouse 96 $71 1.2 6 yes
4 Ramen counter 38 $23 2.8 7 no
5 Wine bar / small pl. 46 $41 1.5 5 yes
Identify the gap in this set — a combination of price, format, and occasion that nobody serves — and state the single largest risk in filling it.
2.6 [Computation · Level 3] A plan for a 72-seat restaurant projects \$1,440,000 of first-year revenue. Benchmarking the four nearest comparable operators produces sales-per-seat figures of \$12,900, \$14,100, \$14,600, and \$17,600. (a) Compute the plan's sales per seat. (b) Compare it to the set. (c) The owner's response is that the concept is unique in the district, so the set does not apply. Evaluate that response, and state what the plan must now demonstrate.
Chapter 3 — Brand and Price Positioning
3.1 [Recall · Level 1] Distinguish a brand from a logo, a name, and a marketing budget. State the one-sentence test for whether a restaurant has a brand.
3.2 [Computation · Level 1] A restaurant's average check is \$38. The competitive set averages \$31. Compute the price-positioning index (your check as a percentage of the set's) and state in one sentence what the guest must perceive for that index to hold.
3.3 [Computation · Level 2] An operator considers repositioning upward, from a \$31 average check to \$38, and expects to lose 12% of covers. On a base of 100 covers per service: (a) compute revenue before and after; (b) state the percentage change; (c) name two cost lines that also move, and in which direction.
3.4 [Judgment · Level 2] A restaurant's brand promise is "your neighborhood living room." It takes no reservations, and the Friday wait is ninety minutes. A consultant says to add reservations. The owner says the wait is the brand. Who is right, and what would you measure to settle it?
3.5 [Artifact · Level 2] Three review excerpts and the positioning statement they were meant to reflect:
POSITIONING STATEMENT (internal) [constructed teaching example]
"A warm, unfussy neighborhood restaurant where a family can eat well
on a Tuesday without planning it in advance."
RECENT REVIEWS (excerpts)
★★★☆☆ "Beautiful food. We waited 50 minutes with a seven-year-old."
★★☆☆☆ "$34 for a pasta in a room with paper napkins."
★★★★★ "Best date night we've had in a year."
Name the gap between the statement and the reviews, and state which one of the two — the statement or the operation — you would change, and why.
3.6 [Computation · Level 3] An item sells at \$18 with a plate cost of \$5.40. Marketing proposes a 20% discount to drive traffic. (a) Compute the contribution margin before and after the discount. (b) How many additional units must sell, in percentage terms, merely to hold total contribution flat? (c) The proposal claims the discount "only costs us 20%." Explain precisely why that framing is wrong.
Chapter 4 — The Business Plan and the Assumptions Register
4.1 [Recall · Level 1] What is an assumptions register, and what four things must each row of one contain?
4.2 [Recall · Level 1] Name the sections of a business plan a lender reads first, in order, and state what each one is being read for.
4.3 [Computation · Level 1] A plan for a 60-seat restaurant assumes 1.6 turns; comparable local data suggests 1.25 is more realistic. At a \$40 average check, five nights a week, 50 weeks a year, compute the annual revenue at each assumption and the difference between them.
4.4 [Computation · Level 2] [the Bellwether plan] An early draft of the plan splits labor into \$252,000 of fixed cost plus 16.0% of sales. The plan's own labor model is \$191,895 of fixed cost plus 19.88% of sales. (a) Compute total labor under each model at \$1,550,000 of sales. (b) Recompute both at \$1,200,000. (c) The draft's fixed figure is \$60,105 too high and its slope is 3.88 points too low. Explain why the draft nonetheless looks correct at \$1,550,000, and state the general lesson about an assumption that has been fitted to a single point.
4.5 [Artifact · Level 2] Four rows of an assumptions register:
ASSUMPTIONS REGISTER (extract) [constructed teaching example]
ASSUMPTION VALUE SOURCE TEST
Dinner turns 1.6 "industry standard" none stated
Average check $40 menu priced at target FC% mystery-shop 3 comps
Food cost 30.0% cost cards, 22 of 24 items re-cost at open
Build-out cost $310,000 contractor budget estimate 3 bids by Mar 1
Which row is unsupported, and what specifically would you require before the plan goes to a lender?
4.6 [Judgment · Level 3] A plan describes its projections as "conservative." On inspection, turns, average check, food cost, labor, and the ramp are each set at the optimistic end of the defensible range. Each is individually arguable. Write the memo to the founders. Your memo must explain why individually reasonable assumptions compound into an unreasonable plan, propose a specific remedy, and say what a lender will conclude if the plan arrives as written.
Chapter 5 — The Capital Stack, Guarantees, and Failure
5.1 [Recall · Level 1] Define a personal guarantee. Name two things it survives that an operator often assumes it does not.
5.2 [Computation · Level 1] [the Bellwether plan] The capital stack is: owner injection \$150,000, tenant-improvement allowance \$75,000, equipment lease \$60,000, and an SBA 7(a) loan of \$335,000, against a project cost of \$620,000. (a) Confirm the stack funds the project. (b) What percentage of the stack is the owner's own cash?
5.3 [Computation · Level 1] A constructed operation projects net operating income of \$96,000 against annual debt service of \$64,000. Compute the debt service coverage ratio and state whether it clears a typical 1.25× threshold.
5.4 [Artifact · Level 2] A use-of-funds schedule:
USE OF FUNDS [constructed teaching example]
Construction and build-out $268,000
Equipment $142,000
Smallwares and FF&E $38,000
Pre-opening labor, training, licensing $29,000
Initial inventory $11,000
Construction contingency (10%) $26,800
─────────────────────────────────────────────────
TOTAL $514,800
Confirm the total. Then name what is missing, estimate what it should be, and explain what happens in month four without it.
5.5 [Judgment · Level 2] An operator is \$120,000 short. Option A: an investor takes 22% of the equity. Option B: a second loan at 13% over five years. The operator says the loan is cheaper because "equity is forever." Evaluate both, and name the circumstance that would flip your recommendation.
5.6 [Computation · Level 3] A restaurant fails in year three. At closing: the equipment lease has a \$60,000 balance and the equipment, originally \$185,000, liquidates at roughly 25% of cost. The SBA-guaranteed loan has \$298,000 outstanding, secured by business assets that realize about \$55,000 after the lessor is satisfied. The owner injected \$150,000. Both the lease and the loan are personally guaranteed. (a) Compute the deficiency on each obligation. (b) Compute the owner's total personal exposure after the equity is gone. (c) The owner told friends "the worst case is I lose my \$150,000." Explain exactly why that was wrong.
Part II — Getting Open (Chapters 6–9)
Chapter 6 — Occupancy Cost, Not Rent
6.1 [Recall · Level 1] An operator says "my rent is \$6,500 a month." List everything else that belongs in occupancy cost, and explain in one sentence why the distinction changes a lease decision.
6.2 [Computation · Level 1] A 2,400 sq ft space is offered at \$26 per square foot base rent plus \$7 per square foot NNN. Compute total annual occupancy cost and the monthly figure.
6.3 [Computation · Level 2] The operator of that space projects \$880,000 of first-year sales. (a) Compute occupancy as a percentage of sales. (b) What annual sales would be required to bring occupancy to 7.0%? (c) Which of those two numbers is the operator actually able to change in year one?
6.4 [Judgment · Level 2] Site A is \$33 per square foot all-in on a corner with heavy foot traffic. Site B is \$21 all-in, four blocks off the main street, in a building with parking. The concept is a 70-seat dinner house with a bar. Argue for one, and state the single piece of data that would change your mind.
6.5 [Artifact · Level 2] A lease abstract:
LEASE ABSTRACT (extract) [constructed teaching example]
Premises 2,600 sq ft, second-generation restaurant
Term 10 years, one 5-year option
Base rent $27.00/sq ft, years 1-3
Escalation 3.0% annually, compounding, from year 2
NNN estimate $6.50/sq ft, reconciled annually, no cap
TI allowance $60,000, paid on completion and lien waivers
Rent commencement the earlier of opening or 120 days from possession
Percentage rent 6% of gross sales above a $1,400,000 breakpoint
Identify the two clauses that carry the most uncapped risk, and state what you would ask for in negotiation on each.
6.6 [Computation · Level 3] Using that abstract's percentage-rent clause: base rent in the relevant year is \$90,000 and gross sales are \$1,750,000, against a \$1,500,000 breakpoint. (a) Compute the percentage rent owed. (b) Compute total rent and total rent as a percentage of sales. (c) The bookkeeper computes percentage rent as 6% of all sales and reports a much larger number. Explain the error and what a "natural breakpoint" means.
Chapter 7 — The Floor Plan and Two Kinds of Capacity
7.1 [Recall · Level 1] Explain the difference between capacity from the room and capacity from the kitchen, and state which one a business plan almost always computes and which one closes the gap between the plan and Saturday night.
7.2 [Computation · Level 1] A grill station produces 22 plates per hour under sustained load. Service runs five hours. About 60% of covers order a grill item. What cover count does the grill support across the whole service?
7.3 [Computation · Level 2] The same restaurant has 72 seats and expects 1.8 turns. (a) Compute capacity from the room. (b) Compare it with the grill's capacity from 7.2. (c) Which constraint binds, and what is the operational consequence of building a marketing plan against the other one?
7.4 [Judgment · Level 2] A designer can add four covers by narrowing the service aisle from 48 inches to 36 inches. Four covers a night at a \$42 check is roughly \$52,000 a year. Make the argument against, in cost terms, and name the two compliance questions you must answer before even considering it.
7.5 [Artifact · Level 2] A seat schedule:
SEAT SCHEDULE — 2,400 sq ft, 1,450 sq ft FOH [constructed teaching example]
Zone Tables Config Seats
Front window 6 2-tops 12
Main floor 9 4-tops 36
Banquette 4 4-tops 16
Bar — stools 14
Private nook 1 8-top 8
─────────────────────────────────────────────────
TOTAL 86
Confirm the seat count and compute square feet of FOH per seat. Then identify the flexibility problem this room has on a Tuesday and on a Saturday, and name one change to the mix.
7.6 [Computation · Level 3] The restaurant from 7.2 and 7.3 does 130 covers on a Friday across a five-hour service. Reservation data shows 40% of covers arrive within a single peak hour. (a) Compute average covers per hour and peak-hour covers. (b) Compute peak-hour grill items required. (c) Compare with the station's 22-per-hour capacity and state the deficit. (d) The chef's response is that the kitchen "did 130 covers, so it can do 130 covers." Explain why the whole-service average is the wrong calculation, and name two fixes that do not involve hiring.
Chapter 8 — Entity, Permits, Liquor, Insurance
8.1 [Recall · Level 1] An operator forms an LLC and concludes the business's debts can never reach personal assets. Explain the two most common ways that conclusion fails in a restaurant.
8.2 [Recall · Level 1] Name five distinct permits, licenses, or registrations a new restaurant typically needs before it may serve a paying guest, and state which authority issues each. Note the standing caveat that applies to your whole answer.
8.3 [Computation · Level 1] In a quota jurisdiction, a liquor license is purchased on the secondary market for \$140,000 and amortized over a ten-year lease term. Compute the annual amortization and express it as a percentage of \$1,200,000 of sales.
8.4 [Computation · Level 2] Build the annual insurance stack for a restaurant with \$460,000 of payroll: general liability \$6,400; property \$4,200; liquor liability \$5,800; workers' compensation at 2.8% of payroll; employment practices liability \$2,600. (a) Total the stack. (b) Express it as a percentage of \$1,200,000 of sales. (c) Which line moves if the concept drops table service for counter service, and which moves if it drops the bar?
8.5 [Artifact · Level 2] A permit tracker:
PERMIT TRACKER — target opening June 1 [constructed teaching example]
Item Filed Typical turnaround Status
Business license Feb 14 2 weeks issued
Seller's permit (sales tax) Feb 14 1 week issued
Building permit Mar 02 4-8 weeks issued Apr 22
Health department plan review Apr 26 3-6 weeks in review
Certificate of occupancy — after final inspect not filed
Liquor license (transfer) Mar 18 90-150 days in review
Food manager certification — 2 days not scheduled
Which item gates the opening date, and which item is most likely to move it? State the specific dependency chain, and one thing you would do this week.
8.6 [Judgment · Level 3] The landlord offers early possession and three extra free-rent months if the operator opens on schedule. The liquor license will not clear for another eight to ten weeks. A consultant suggests opening dry and adding beverage later. The plan assumes 28% of sales from beverage. Write the analysis. Address the revenue effect, the brand effect, the staffing effect, and the one compliance question that must be answered before anyone considers a workaround.
Chapter 9 — The Countdown, the Soft Open, and the Ramp
9.1 [Recall · Level 1] Define the ramp. Explain why the opening spike is not the baseline, and state what a plan that treats it as the baseline gets wrong.
9.2 [Recall · Level 1] What is a soft open for? Name three things a friends-and-family night genuinely tests and two things it cannot test at all.
9.3 [Computation · Level 1] Fourteen staff are trained for 22 hours each at a \$15.50 training wage before opening. Compute the pre-opening training payroll, before burden.
9.4 [Computation · Level 2] A plan assumes \$29,800 per week from week one. Actual results: weeks 1–4 average \$26,000; weeks 5–8 average \$24,000; weeks 9–13 average \$27,000. (a) Compute planned and actual revenue for the thirteen weeks and the shortfall. (b) Explain the weeks 5–8 dip in one sentence. (c) State what the shortfall means for the working-capital reserve, not for the P&L.
9.5 [Artifact · Level 2] A countdown extract:
90-DAY COUNTDOWN (extract) [constructed teaching example]
T-60 Post job listings; begin interviews
T-45 Order smallwares and china
T-30 Menu finalized; cost cards complete
T-21 POS installed and menu programmed
T-14 Staff training begins
T-10 Health department final inspection requested
T-7 Friends-and-family service
T-3 Initial food order placed
T-0 Open
Two tasks in this sequence are in the wrong order and one is dangerously late. Identify them and re-sequence, with a one-line reason for each change.
9.6 [Judgment · Level 3] Two weeks before the announced opening, the hood contractor reports that the exhaust system will not pass inspection without a change that takes fifteen working days. A consultant proposes opening with a limited menu that avoids the hearth. Payroll for a trained staff of 26 is running. Argue for a course of action. Your answer must quantify the cost of waiting, name what the limited-menu opening does to the brand and to the ramp, and identify the single non-negotiable constraint.
Part III — Menu, Product, and Cost (Chapters 10–16)
Chapter 10 — Menu Development and Cross-Utilization
10.1 [Recall · Level 1] Define cross-utilization. Name the two distinct costs it reduces, and the one risk it creates.
10.2 [Computation · Level 1] A 22-item menu currently requires 96 distinct purchased items. A redesign holds all 22 menu items but requires only 68. Compute the reduction in count and in percentage terms.
10.3 [Computation · Level 2] A braised pork preparation yields 60 portions per batch and holds three days. (a) If it appears on one dish selling 14 per day, how many portions are sold within the hold window and how many are discarded? Express the waste as a percentage of the batch. (b) If the same prep appears on three dishes selling 20 per day combined, recompute. (c) State the principle in one sentence.
10.4 [Judgment · Level 2] A chef wants a 34-item dinner menu because "we should be able to feed anybody." Make the operational and financial case against, in specific terms, and name the one circumstance in which the chef would be right.
10.5 [Artifact · Level 2] A cross-utilization map:
CROSS-UTILIZATION MAP (extract) [constructed teaching example]
Component Used on
Salsa verde Hearth chicken, pork shoulder, brunch eggs
Pickled shallot Burger, trout, charcuterie, house salad
Chicken jus Hearth chicken
Duck confit Duck entree
Whipped ricotta Bruschetta, gnocchi, brunch toast
Nduja butter Mussels
Identify the two components carrying the most risk of waste, and propose one specific change to each that does not remove a dish from the menu.
10.6 [Computation · Level 3] A chef proposes adding a dish that requires five new purchased items. Projected: 9 covers a week at \$26 with a \$9.10 plate cost. The five new items are expected to carry \$38 per week of spoilage, and the prep adds 20 minutes a day, five days a week, at \$17 an hour. Menu-mix analysis suggests 6 of the 9 weekly sales would come from an existing dish carrying an \$18.40 contribution margin. (a) Compute the new dish's weekly contribution. (b) Subtract the added costs. (c) Subtract the cannibalized contribution. (d) State the net and the decision. (e) Which of these four figures does an operator usually forget?
Chapter 11 — Cost Cards, Yield, and What COGS Actually Is
11.1 [Recall · Level 1] Distinguish as-purchased (AP) cost from edible-portion (EP) cost, and define yield percentage. Which of the two costs belongs on a cost card, and why?
11.2 [Recall · Level 1] State the inventory usage formula. Then explain, in two sentences, why dividing food invoices by food sales is not food cost, and name the two behaviors that make the shortcut most misleading.
11.3 [Computation · Level 2] A 12-pound case of whole beef tenderloin costs \$14.80 per pound. Fabrication yield is 62%. (a) Compute the case cost, the edible-portion weight, and the EP cost per pound. (b) Compute the cost of an 8-ounce portion. (c) The chef objects that the trim becomes tartare and stock, so the yield calculation overstates the cost. Is the chef right? State how you would handle it on the cost card.
11.4 [Artifact · Level 2] A partial cost card:
RECIPE COST CARD — pan-roasted trout [constructed teaching example]
Component Qty Unit cost Extended
Trout fillet, 6 oz 6 oz $0.62/oz $3.72
Fingerling potato 5 oz $0.14/oz $0.70
Brown butter, capers 1.2 oz $0.45/oz $0.54
Haricot vert 3 oz $0.24/oz $0.72
Lemon, herbs, oil — — $0.32
Garnish — — $0.15
─────────────────────────────────────────────────────────
COMPONENT TOTAL $6.15
Waste and spillage allowance (3%) [ ]
PLATE COST [ ]
Complete the two blank lines. Then price the dish to a 30% food-cost target and state the menu price you would actually print, with a one-sentence reason.
11.5 [Judgment · Level 3] A sous chef proposes removing the waste allowance from every cost card, arguing that the kitchen is disciplined, the allowance is guesswork, and it inflates prices against the competition. There is a real argument on that side. Write the response. Address what the allowance actually represents, what happens to the menu price and the reported variance if it is removed, and what you would do instead if you shared the sous chef's concern about accuracy.
11.6 [Computation · Level 3] A four-week period at a bistro: beginning food inventory \$21,400; food purchases \$47,900; ending food inventory \$19,800; food sales \$162,000. Also recorded: employee meals \$1,850 at cost, comped food \$1,240 at cost, and food transferred to the bar \$690 at cost. (a) Compute food usage and unadjusted food cost percentage. (b) Compute a defensible adjusted food cost percentage, and state exactly which of the three adjustments you applied. (c) One of the three should not be deducted from food cost. Name it, explain why, and state what reporting it as a deduction would do to the number.
Chapter 12 — Menu Engineering
12.1 [Recall · Level 1] Name the two axes of the menu-engineering matrix, the four quadrant names, and the action associated with each quadrant.
12.2 [Recall · Level 1] State both thresholds used to place an item in a quadrant — the popularity threshold and the margin threshold — including the formula for each.
12.3 [Computation · Level 2] A dinner menu carries 14 entrées. (a) Compute the popularity threshold. (b) An item sells 4.2% of entrée units. Which side of the threshold is it on? (c) Explain in one sentence why the threshold uses 70% rather than a simple equal share.
12.4 [Artifact · Level 2] A four-week menu-mix report:
MENU MIX — 4 weeks, 6 entrees [constructed teaching example]
Item Units Price Plate cost CM
Roast chicken 260 $27.00 $8.10 $18.90
Steak frites 180 $34.00 $13.60 $20.40
Mussels 150 $22.00 $6.60 $15.40
Burger 290 $19.00 $6.65 $12.35
Trout 60 $24.00 $10.80 $13.20
Cassoulet 60 $26.00 $9.10 $16.90
─────────────────────────────────────────────────────
TOTAL 1,000
Compute both thresholds and classify all six items. Then state the single action you would take first and why.
12.5 [Judgment · Level 3] A general manager reviews the menu-mix report and proposes cutting the item with the highest food-cost percentage, on the grounds that "it's dragging our food cost up." The item is the second-best seller and carries the highest contribution margin on the menu. Write the response. Explain the contribution trap precisely, state what would have to be true for the GM to be right, and describe the one circumstance — a real one — in which a high-food-cost, high-CM item genuinely should come off a menu.
12.6 [Computation · Level 3] Two items. Item A: price \$16.00, plate cost \$4.00, 40 units a week. Item B: price \$34.00, plate cost \$12.24, 55 units a week. (a) Compute food-cost percentage and contribution margin for each. (b) Compute weekly contribution dollars from each. (c) Compute the blended food-cost percentage of the two items together. (d) A manager proposes replacing B with a second item like A to "fix food cost." Compute what that would do to weekly contribution, assuming the replacement sells the same 55 units. (e) State the lesson in one sentence.
Chapter 13 — Purchasing, Receiving, and the Count
13.1 [Recall · Level 1] Define a par level. Explain how a par differs from an order quantity, and state the two variables that set a par.
13.2 [Computation · Level 1] The par for a produce item is 24 cases. The count sheet shows 9 on hand, and 6 cases are already on order from a previous delivery. What quantity do you order?
13.3 [Computation · Level 2] For a four-week period, the POS reports theoretical food usage of \$47,300 based on recipes and items sold. The physical count produces actual usage of \$49,900. Food sales were \$162,000. (a) Compute the variance in dollars, as a percentage of theoretical usage, and in points of food cost on sales. (b) Name the four most likely causes, and (c) state the first one you would test, and how.
13.4 [Judgment · Level 2] Variance has run between 4% and 6% of theoretical for three consecutive periods, concentrated in proteins, and is worst in the periods with the most new-hire hours. Rank three candidate causes by likelihood and describe the specific test for the one you rank first.
13.5 [Artifact · Level 2] A receiving log:
RECEIVING LOG — Tuesday [constructed teaching example]
Vendor Item Ordered Received Invoice unit Last unit
Produce Co Romaine, 24ct 4 cs 4 cs $28.40 $26.10
Produce Co Roma tomato, 25# 3 cs 2 cs $31.00 $30.75
Meat Co Chicken, whole 120 lb 118 lb $3.34 $3.20
Dairy Co Butter, 36# 2 cs 2 cs $132.00 $132.00
Two problems appear here. Identify both, state the dollar effect of each, and name the receiving-desk procedure that catches each one at the door rather than at month end.
13.6 [Computation · Level 3] Over a quarter, invoice prices in the protein and dairy categories rose an average of 4.1% while menu prices held. Those categories represent 55% of a \$49,500 monthly food usage. (a) Compute the monthly and annual cost increase. (b) A manager computes 4.1% of total usage and reports a much larger figure. Compute that figure and explain why it is wrong. (c) Compute the menu price increase, on a \$27 entrée, that would offset the correct figure if the restaurant sells 2,050 entrées a month.
Chapter 14 — Stations, Prep, and the Pass
14.1 [Recall · Level 1] What is the pass, and what does the expediter actually do there? Name three decisions that belong to the expediter and nobody else.
14.2 [Computation · Level 1] Over one service the kitchen fired 62 tickets with a total cook time of 1,984 minutes. Compute the average ticket time and state whether that figure alone tells you the kitchen was in trouble.
14.3 [Computation · Level 2] A restaurant does 130 covers over a five-hour service at 1.4 items per cover. The sauté station handles 45% of items and sustains 14 items an hour. (a) Compute total items, sauté items, and sauté items per hour. (b) Compare to capacity. (c) Name two changes that relieve sauté without adding a cook.
14.4 [Judgment · Level 2] It is 3:40 p.m. on a Friday with 142 covers on the books including a 40-top at 6:30. The grill cook has not shown up and is not answering. The sous is already on a double. Lay out three options, quantify the cost of each as far as the information allows, and state which you would take and what you would do at 4:00 p.m. tomorrow.
14.5 [Artifact · Level 2] A ticket-time report:
TICKET TIMES BY HOUR — Saturday [constructed teaching example]
Hour Tickets Avg time Longest Covers
5:00-6 11 17 min 24 min 19
6:00-7 19 21 min 33 min 36
7:00-8 34 38 min 61 min 67
8:00-9 27 34 min 52 min 51
9:00-10 12 19 min 27 min 21
Identify the pinch and describe what is most likely happening. Then name one change to the reservation book and one change to prep that would flatten it.
14.6 [Computation · Level 3] A kitchen makes stock in-house: four hours of labor per batch at \$18 an hour, one batch covering three service days, five service days a week, plus \$31 a week of ingredients. A commercial base of comparable quality costs \$4.20 a quart, and the kitchen uses 26 quarts a week. (a) Compute the weekly and annual cost of making it. (b) Compute the weekly and annual cost of buying it. (c) Compute the difference. (d) Explain what this decision does to food cost percentage, to labor cost percentage, and to prime cost, and state which of the three should govern the decision.
Chapter 15 — The Bar: Pour Cost and Control
15.1 [Recall · Level 1] Define pour cost and give the typical full-service range. Explain in one sentence why a beverage-weighted sales mix pulls blended COGS down.
15.2 [Computation · Level 1] A 750 ml bottle of spirit costs \$28.00 and is poured at 1.5 ounces. Using 25.36 ounces per 750 ml: (a) compute pours per bottle and cost per pour; (b) at a \$12.00 menu price, compute pour cost percentage.
15.3 [Computation · Level 2] The same bar's bartenders free-pour and average 1.75 ounces. (a) Recompute pours per bottle and cost per pour. (b) Across 40 bottles a week, compute the lost drinks and the retail value of the loss, weekly and annually. (c) State the cheapest control that addresses this and one reason bartenders resist it.
15.4 [Judgment · Level 2] A bar manager argues that jiggers slow service at the peak and cost more in lost throughput than they save in product. There is a real argument there. Evaluate it, and design a test that would settle the question in two weeks.
15.5 [Artifact · Level 2] A bar variance sheet:
BAR VARIANCE — 4 weeks [constructed teaching example]
Category Depletion at cost POS cost of sales Variance Variance %
Beer $7,140 $7,010 $130 1.8%
Wine $9,880 $9,430 $450 4.8%
Spirits $11,260 $9,940 $1,320 13.3%
N/A $890 $870 $20 2.3%
Which category is the problem, and by how much annually? Name the three most likely causes in that category specifically, and the order in which you would test them.
15.6 [Computation · Level 3] A bar's category pour costs are: beer 24%, wine 32%, spirits 16%. The current sales mix is beer 30%, wine 30%, spirits 40%. (a) Compute blended pour cost. (b) A wine promotion shifts the mix to beer 25%, wine 45%, spirits 30%. Recompute. (c) State the change in points. (d) The bar manager is asked to explain why pour cost rose, having changed no recipe, no price, and no supplier. Write the two-sentence explanation, and name the report that would have shown this coming.
Chapter 16 — The Wine List
16.1 [Recall · Level 1] Distinguish a flat multiplier from progressive markup and from fixed gross profit as wine-pricing methods. Explain in one sentence why a straight 3× multiplier fails at the top of a list.
16.2 [Computation · Level 1] A bottle costs \$14.00 wholesale and is listed at a 3× multiplier. Compute the list price and the pour cost percentage.
16.3 [Computation · Level 1] A bottle costs \$28.00 wholesale and is priced by adding a fixed \$32.00 gross profit. Compute the list price and the pour cost percentage.
16.4 [Artifact · Level 2] Four rows of a list:
WINE LIST PRICING (extract) [constructed teaching example]
Wine Wholesale List price
A $9.00 $36.00
B $14.00 $48.00
C $26.00 $74.00
D $48.00 $112.00
Compute pour cost percentage and gross profit dollars for each. Describe the pattern in one sentence, and state which of the four is most likely to sell and which is most likely to sit.
16.5 [Judgment · Level 2] A wine director proposes expanding the by-the-glass list from 8 to 16 selections to "give guests choice." The restaurant does 95 covers a night, five nights. Evaluate the proposal in terms of spoilage, training, and inventory dollars, and state the condition under which you would approve it.
16.6 [Computation · Level 3] A by-the-glass wine costs \$26.00 a bottle and yields 5 glasses at \$16.00. Of open bottles, 60% sell all five glasses within the freshness window; the other 40% sell three and are discarded. (a) Compute revenue and pour cost for each case. (b) Compute the blended pour cost across ten bottles. (c) The wine director's cost card shows 32.5%. Explain why, and name two changes that would close the gap between the card and reality.
Part IV — People (Chapters 17–21)
Chapter 17 — Hiring and the Pipeline
17.1 [Recall · Level 1] Name the five categories of cost that make up the true cost of one turnover event, and state which of them almost never appears in any report.
17.2 [Recall · Level 1] What is a hiring pipeline, and why does "post the job when someone quits" reliably produce a worse hire at a higher cost?
17.3 [Computation · Level 1] A restaurant has 22 budgeted positions and recorded 17 separations over twelve months. Compute the annual turnover rate and compare it to the industry's rough benchmark.
17.4 [Computation · Level 2] Build the turnover cost for one line cook: recruiting and advertising \$310; manager time 6 hours at \$32; trainer time 24 hours at \$21; the trainee's first 60 hours are 40% unproductive at \$19 an hour; overtime coverage during the gap \$380. (a) Total it. (b) The restaurant lost 9 cooks last year. Compute the annual cost. (c) Name the one line in this build that a skeptical owner will challenge, and defend it.
17.5 [Artifact · Level 2] A hiring funnel:
HIRING FUNNEL — line cook, one posting [constructed teaching example]
Job post views 120
Applications received 26
Screened by phone 14
Interviews scheduled 8
Interviews attended 5
Offers made 3
Offers accepted 2
Still employed at 90 days 1
Compute the yield from applications to a 90-day hire. Then identify the two worst leaks in this funnel and name a specific fix for each.
17.6 [Judgment · Level 3] Two finalists for a line cook opening. Candidate A has six years of relevant experience and one former chef who says, carefully, "I wouldn't work with him again." Candidate B has eight months of experience, is obviously capable, and would need four weeks of development. The station is open now and the sous is covering it. Argue for a choice. Address the cost of each path using the figures from 17.4, the risk each carries, and what you would do differently in the offer itself depending on which you choose.
Chapter 18 — Training and What a Trained Employee Costs
18.1 [Recall · Level 1] What is a training matrix, and what does a signed certification checklist protect against — operationally and legally?
18.2 [Computation · Level 1] A server trains for 32 hours at a \$12.50 training wage. Compute the direct trainee wage cost.
18.3 [Computation · Level 2] Build the fully loaded cost of producing that trained server: 32 trainee hours at \$12.50; a trainer premium of \$2.50 an hour for 24 of those hours; 4 hours of manager orientation at \$30; \$45 of materials; payroll burden of 12.4% applied to the hourly wage components only. (a) Total it. (b) State which component you would cut first if forced, and what it would cost you.
18.4 [Judgment · Level 2] The floor is short and a manager proposes cutting server training from five shifts to two. Name the specific costs that decision creates, say where each one shows up and how long it takes to appear, and state the condition under which the shortcut is nonetheless correct.
18.5 [Artifact · Level 2] A training checklist:
SERVER CERTIFICATION CHECKLIST [constructed teaching example]
Module Trained Verified Signed
POS and check handling ✓ ✓ ✓
Menu knowledge - food ✓ ✓ ✓
Menu knowledge - beverage ✓ — —
Allergen protocol ✓ — —
Steps of service ✓ ✓ ✓
Responsible alcohol service ✓ ✓ ✓
Table maintenance and pacing ✓ — ✓
Two rows here should stop this server from taking a section alone. Identify them, explain the exposure each creates, and describe what "verified" must mean for one of them to be signed honestly.
18.6 [Computation · Level 3] Post-training data suggests a fully trained server sells \$3.10 more per cover than a partially trained one, through beverage and appetizer attachment. A section runs 480 covers a month. Using a 40% contribution margin ratio and the fully loaded training cost from 18.3: (a) compute incremental monthly sales and incremental monthly contribution; (b) compute the payback period on the training investment; (c) name the assumption in this calculation that would most likely be challenged, and what evidence would support it.
Chapter 19 — The Labor Model
19.1 [Recall · Level 1] Define sales per labor hour (SPLH). Name two things it usefully reveals and two things it cannot tell you at all.
19.2 [Recall · Level 1] Distinguish fixed labor from variable labor, giving two examples of each, and state why the distinction matters more at low volume than at high volume.
19.3 [Computation · Level 2] One dinner service is scheduled as follows: 2 line cooks × 8.5 hours at \$20; 1 prep cook × 6 hours at \$17; 1 dishwasher × 6.5 hours at \$15; 4 servers × 6 hours at \$10; 1 host × 5 hours at \$15; 1 bartender × 7 hours at \$16; plus allocated salaried cost of \$185 (chef) and \$150 (manager). Payroll burden is 22.5% on everything. Sales for the service are \$4,600. (a) Compute unburdened and burdened labor cost. (b) Compute labor as a percentage of sales. (c) Compute SPLH on hourly hours only.
19.4 [Artifact · Level 2] A week at a glance:
SCHEDULED HOURS AND FORECAST [constructed teaching example]
Day Forecast covers FOH hours BOH hours Total Forecast sales
Tue 58 34.0 38.0 72.0 $2,320
Wed 71 36.0 38.0 74.0 $2,840
Thu 86 38.0 41.0 79.0 $3,440
Fri 132 49.0 50.0 99.0 $5,280
Sat 138 49.0 50.0 99.0 $5,520
Compute SPLH for each day. Identify the day with the worst labor productivity, state how many hours you would remove and from where, and name what you would check before removing them.
19.5 [Judgment · Level 3] A kitchen can be staffed with one salaried sous chef at \$58,000 or with two hourly lead cooks at \$24 an hour, roughly 38 hours each. Work the arithmetic far enough to compare, then argue the decision. Address coverage, overtime exposure, development, what happens when one person quits, and the classification question that must be answered before the salaried option is even legal.
19.6 [Computation · Level 3] [the Bellwether plan] The labor model is fixed labor of \$191,895 plus 19.88% of sales. (a) Compute total labor and labor percentage at \$1,550,000 of sales. (b) Recompute both at \$1,300,000. (c) Explain, in terms a lender would accept, why labor percentage rose when nothing about the schedule changed. (d) An owner who budgets labor as "32.3% of whatever we do" will be wrong in which direction, and when?
Chapter 20 — Wage and Hour
20.1 [Recall · Level 1] What is the tip credit under the Fair Labor Standards Act, in structure? State the two facts about it that a multi-state operator must check before writing a single schedule.
20.2 [Recall · Level 1] Who may not share in a tip pool, and why? Name the category the law is most concerned about and the practical test a manager should apply.
20.3 [Computation · Level 2] A line cook paid \$21.00 an hour works 47 hours in a workweek. (a) Compute straight-time pay, overtime premium pay, and gross pay for the week. (b) The cook also received a \$100 non-discretionary production bonus that week. Explain, without computing it, what that bonus does to the overtime calculation and why.
20.4 [Artifact · Level 2] A payroll register row:
PAYROLL REGISTER (one tipped employee, one week) [constructed teaching example]
Hours worked 30.0
Direct cash wage rate $5.00
Direct cash wages $150.00
Tips reported $120.00
Applicable minimum wage $10.50
Compute what this employee must be paid for the week and what the employer owes beyond the direct cash wages. Then state the standing caveat about jurisdiction that belongs on your answer.
20.5 [Judgment · Level 3] A general manager proposes classifying two assistant managers as exempt at a \$780 weekly salary to end the overtime problem. Both spend roughly half their shifts expediting, bussing, and running food. Write the analysis. Address both prongs of the exempt test, the specific risk the duties create regardless of salary, what an audit would cost, and what you would do instead.
20.6 [Computation · Level 3] Two assistant managers are paid \$52,000 a year each and work 52 hours a week. A review concludes they must be reclassified as non-exempt. (a) Compute each one's regular hourly rate from the salary. (b) Compute the weekly and annual overtime premium owed going forward for each, and for both together. (c) Add 22% payroll burden. (d) At a 40% contribution margin ratio, compute how much additional annual sales the restaurant must generate to stay at the same profit. (e) State what this reveals about compliance decisions and the break-even line.
Chapter 21 — Culture, Retention, and the Manager's Day
21.1 [Recall · Level 1] State the difference between what turnover costs and what retention costs to buy. Name three retention levers that cost nothing in wages.
21.2 [Recall · Level 1] Define pre-shift. What does a good one accomplish, and what is the most common way it is wasted?
21.3 [Computation · Level 1] A restaurant with 22 positions cuts annual turnover from 78% to 55%. Using a turnover cost of \$1,842 per separation, compute the annual saving.
21.4 [Computation · Level 2] The proposed mechanism for that retention improvement is a \$1.25 an hour raise across 14 hourly staff who average 28 hours a week, with 22% payroll burden. (a) Compute the weekly and annual cost of the raise. (b) Compare it to the turnover saving from 21.3. (c) The raise does not pay for itself on turnover alone. Name three other places the return could appear, and how you would measure one of them.
21.5 [Artifact · Level 2] A manager's Monday:
MONDAY TASK LIST — general manager [constructed teaching example]
08:30 Count walk-in and dry storage
09:15 Enter invoices; code to accounts
10:00 Build weekly flash report
10:45 Write next week's schedule against forecast
12:00 Vendor call - produce pricing
13:00 Interview: server candidate
14:00 Reconcile comps and voids report
15:00 Order for Wednesday delivery
16:00 Pre-shift
Three of these can be delegated, one must not be, and one is in the wrong place in the day. Identify each and justify the one that must not be delegated.
21.6 [Judgment · Level 3] A server is the highest-selling employee on the floor by a wide margin, is beloved by regulars, and is corrosive to the rest of the team: two cooks have complained, one host has quit, and the sous will not work Saturdays with them. The owner points at the sales numbers. Write the argument. Quantify what you can — using turnover cost — acknowledge what you cannot quantify, and state the specific sequence of steps you would take, in order, with a deadline.
Part V — Service, the Room, and the Guest (Chapters 22–26)
Chapter 22 — Service Flow and Table Management
22.1 [Recall · Level 1] Break table turn time into its four components, in order, and name the one a manager can compress with the least damage to the guest experience.
22.2 [Computation · Level 1] Average dwell time is 95 minutes and reset takes 8 minutes. Over a four-hour service window, how many turns can one table produce?
22.3 [Computation · Level 2] A restaurant reduces turn time from 103 minutes to 92 minutes across 18 four-tops over a 240-minute service. (a) Compute turns per table before and after. (b) Compute the additional covers. (c) At a \$38 average check, compute the additional nightly revenue. (d) Name the two ways this gain is commonly given back.
22.4 [Judgment · Level 2] A regular who spends heavily calls at 6:45 asking to hold a four-top for an 8:00 arrival. The book is full and the wait list is 40 minutes. Decide, and state what your decision implies as a policy if every server knew about it.
22.5 [Artifact · Level 2] A reservation book:
RESERVATION BOOK — Friday [constructed teaching example]
Time Covers booked Notes
5:30 6
6:00 10
6:30 18
7:00 46 three 6-tops, one 8-top
7:30 34
8:00 14
8:30 8
─────────────────────
TOTAL 136
Describe what will happen in this dining room, hour by hour, and name the three specific failures that will follow. State the pacing rule you would set.
22.6 [Computation · Level 3] The kitchen behind that book sustains 30 entrées an hour and service runs five hours. (a) Compute total entrée capacity for the night and compare it to 136 covers. (b) Compute entrées required in the 7:00–8:00 window as booked. (c) Compute the maximum covers per 15-minute slot that keeps the kitchen inside capacity. (d) The owner's objection is that the kitchen "can obviously do 136 because capacity is 150." Explain exactly why that is the wrong calculation.
Chapter 23 — Recovery, Reviews, and What a Guest Is Worth
23.1 [Recall · Level 1] Define guest lifetime value. Explain in one sentence why frequency, not reach, is where an independent restaurant's marketing money belongs.
23.2 [Computation · Level 1] A guest spends \$38 a visit. The restaurant's contribution margin ratio is 40%. The guest visits 8 times a year and stays a guest for 3 years. Compute contribution per visit and lifetime value in contribution dollars.
23.3 [Computation · Level 2] A manager comps a \$46 check to recover a service failure for a guest with the profile in 23.1–23.2 and 24 visits remaining. (a) Compute the ratio of preserved contribution to the cost of the comp. (b) State the two conditions that must hold for that ratio to be real. (c) Explain why this arithmetic does not justify comping every complaint.
23.4 [Judgment · Level 2] A two-star review describes a 45-minute entrée wait (true, it was a Saturday at 7:15), a rude server (the server's account differs), and a "\$19 hamburger" (the burger is \$17). Write the public response, then write the two internal actions. Explain what the public response is actually for.
23.5 [Artifact · Level 2] A review summary:
REVIEWS - 90 days, all platforms [constructed teaching example]
★★★★★ 118 Themes mentioned (count)
★★★★☆ 64 Food quality (positive) 141
★★★☆☆ 29 Service warmth (positive) 88
★★☆☆☆ 21 Wait time (negative) 47
★☆☆☆☆ 14 Noise level (negative) 31
──────────── Value/price (negative) 22
TOTAL 246 Reservation system (negative) 18
State what you would fix first and why, what you would deliberately not fix, and what this summary cannot tell you about the guests who did not write anything.
23.6 [Computation · Level 3] Two proposals, same quarter. Proposal A: \$4,800 on a reach campaign expected to bring 300 new guests, 22% of whom return once. Proposal B: \$1,200 on an email and guest-database program expected to lift 900 existing regulars from 8.0 to 8.6 visits a year. Contribution per visit is \$15.20. (a) Compute the contribution and net return of each. (b) State which wins and by how much. (c) The owner objects that Proposal B "doesn't grow the business." Answer the objection, then name the circumstance in which the owner would be right.
Chapter 24 — Turns, RevPASH, and Capacity Pricing
24.1 [Recall · Level 1] Define RevPASH. State what it measures that neither average check nor covers can measure alone.
24.2 [Recall · Level 1] Why can seat turns mislead? Give a concrete two-restaurant comparison in which the restaurant with more turns is the weaker business.
24.3 [Computation · Level 2] A 68-seat restaurant does \$4,600 of revenue over a five-hour service. (a) Compute RevPASH. (b) Compute average check if 100 covers were served. (c) Explain which of the two numbers you would put on a manager's daily report, and why.
24.4 [Artifact · Level 2] Hourly performance:
REVPASH BY HOUR — 68 seats, Thursday [constructed teaching example]
Hour Revenue Seat-hours RevPASH
5:00-6 $211 68 $3.10
6:00-7 $639 68 $9.40
7:00-8 $1,482 68 $21.80
8:00-9 $1,333 68 $19.60
9:00-10 $558 68 $8.20
Compute the service's average RevPASH. Identify the two hours worth attacking and state a different tactic for each — they are not the same problem.
24.5 [Judgment · Level 3] The proposal is a 30%-off "early seating" price for 5:00–5:45 reservations. Argue for or against. Your answer must address which guests actually take the offer, what it does to the brand's price positioning, what it does to the 7:00 book, and what a fixed-price early menu would do differently than a percentage discount.
24.6 [Computation · Level 3] [the Bellwether plan] Sixteen early seats are sold at 30% off a \$46 average check. The contribution margin ratio is 40.53%. (a) Compute discounted check, total discounted revenue, and contribution. (b) Analysis suggests 6 of the 16 guests would have booked at 7:00 and paid full price. Compute the displaced contribution. (c) Compute the net effect. (d) State the general rule this establishes about discounting into a constrained period.
Chapter 25 — Food Safety and the Inspection
25.1 [Recall · Level 1] State the temperature danger zone, the cold-holding maximum, the hot-holding minimum, and the minimum internal cooking temperature for poultry, as framed by the FDA Food Code. Add the sentence that must accompany all four.
25.2 [Recall · Level 1] What concentration range is typical for quaternary ammonium sanitizer in a three-compartment sink, and what two things must a manager verify about a sanitizer bucket beyond the number on the test strip?
25.3 [Computation · Level 1] A stock comes off the fire at 4:15 p.m. Under the two-stage cooling rule, by what time must it reach 70°F, and by what time must it reach 41°F?
25.4 [Computation · Level 2] A cooling failure requires discarding 40 pounds of chicken at \$3.20 a pound. The jurisdiction charges a \$185 re-inspection fee, and the correction consumes 6 hours of manager time valued at \$32 an hour. (a) Compute the direct cost of the incident. (b) Name three costs of the same incident that this total does not capture.
25.5 [Artifact · Level 2] An inspection report:
INSPECTION REPORT (extract) - Tuesday 10:40 a.m. [constructed teaching example]
1 Walk-in cooler ambient 46°F
2 Raw chicken stored above ready-to-eat greens
3 Sanitizer bucket at 50 ppm
4 No thermometer in reach-in
5 One employee without food-handler card
Result: re-inspection in 10 days
Rank these five findings by risk to a guest, not by the order listed. State which one you correct before the inspector leaves the building, and which one is a systems failure rather than an incident.
25.6 [Judgment · Level 3] The walk-in reads 46°F at 10:40 a.m. mid-prep. Nobody knows how long it has been there; the last recorded log entry is from 6:30 p.m. the previous day. Inside: raw proteins, dairy, prepped produce, and cooked braises from two days ago. Write the decision. Address what you keep and what you discard and on what basis, what you do about the log gap, what you tell the staff, and what you change so this is detectable within four hours next time.
Chapter 26 — The POS and What Technology Really Costs
26.1 [Recall · Level 1] Break a card processing rate into its three components and state which one is negotiable, which is not, and which the operator is usually quoted as if it were the whole thing.
26.2 [Recall · Level 1] Name three things a POS report can tell you with authority and three things operators routinely believe it tells them but it does not.
26.3 [Computation · Level 1] [the Bellwether plan] Processing costs 2.81% of sales. On \$1,550,000 of revenue, compute the annual processing cost.
26.4 [Computation · Level 2] Build the annual technology stack: POS \$189 a month per terminal on 3 terminals; kitchen display \$85; online ordering \$129; reservations \$249; scheduling \$79; inventory \$149; accounting \$85; payroll service \$220; internet, phone, and music \$140. (a) Compute the monthly and annual total. (b) Add processing from 26.3 and express the combined figure as a percentage of \$1,550,000. (c) Where does each of these two components live on the P&L?
26.5 [Artifact · Level 2] A merchant statement:
MERCHANT STATEMENT (extract) - one month [constructed teaching example]
Gross card volume $118,400
Interchange $2,462
Assessments $305
Processor markup $497
Monthly fees, PCI, statement $138
Chargebacks and adjustments $100
─────────────────────────────────────────────────
TOTAL FEES $3,502
Compute the effective rate. The operator was quoted "2.4% and ten cents." Explain the gap between the quote and the statement, and name the one line here that a better negotiation actually moves.
26.6 [Judgment · Level 3] A POS vendor offers free hardware and free installation, contingent on a three-year processing agreement at a rate the operator cannot verify against the current effective rate. Evaluate the offer. Quantify what "free hardware" is worth, identify what the operator is actually buying, name the contract terms you would insist on, and state the one question that settles it.
Part VI — Revenue Beyond the Dining Room (Chapters 27–30)
Chapter 27 — Marketing Built on Frequency
27.1 [Recall · Level 1] Distinguish reach from frequency as marketing objectives. State why frequency is cheaper for an independent restaurant, in one sentence, in cost terms.
27.2 [Computation · Level 1] A campaign costs \$3,600 and produces 240 identifiable new guests. Compute cost per acquired guest.
27.3 [Computation · Level 2] Those guests spend an average of \$38 at a 40% contribution margin ratio. (a) Compute first-visit contribution in total. (b) If 25% of them return twice more, compute the additional contribution. (c) Compute the campaign's net result. (d) State the single variable this whole calculation hangs on.
27.4 [Judgment · Level 2] \$5,000 is available. Option A: a billboard on the arterial for three months. Option B: building a guest email and text database with a birthday and lapsed-guest program. The owner likes the billboard because "everybody will see it." Argue the case, and name what Option B requires that Option A does not.
27.5 [Artifact · Level 2] A campaign report:
CAMPAIGN REPORT - 6 weeks [constructed teaching example]
Impressions 412,000
Clicks to site 3,180
Reservation page views 742
Reservations booked 119
Reservations seated 96
Spend $3,600
Compute the conversion at each stage and the cost per seated reservation. State which stage of this funnel is the actual problem, and what you would test first.
27.6 [Computation · Level 3] A "\$20 off \$60" offer is redeemed 310 times. Post-campaign analysis suggests only 40% of redemptions were incremental visits; the rest would have come anyway. The contribution margin ratio on a \$60 check is 40%. (a) Compute the total discount cost. (b) Compute the contribution from incremental visits only. (c) Compute the net result. (d) The owner points to 310 redemptions as proof the promotion worked. Explain precisely what the 310 measures and what it does not.
Chapter 28 — Delivery and the Third-Party Channel
28.1 [Recall · Level 1] Define channel contribution per order. State what a third-party commission does and does not cover, and name the cost it never covers.
28.2 [Computation · Level 1] A \$42.00 order comes through a third-party marketplace at a 28% commission. Compute the commission and the net remitted before any other cost.
28.3 [Computation · Level 1] Food cost on that order runs 30% of the \$42.00 menu value. Compute the contribution remaining after commission and food cost, before packaging and labor.
28.4 [Artifact · Level 2] A settlement statement:
DELIVERY SETTLEMENT - 4 weeks [constructed teaching example]
Orders 214
Gross order value $8,988.00
Commission (28%) -$2,516.64
Promotional fee participation -$430.00
Customer adjustments and refunds -$268.00
In-app marketing -$310.00
───────────────────────────────────────────────────
NET REMITTED $5,463.36
Confirm the arithmetic and compute net remitted per order. Then state which two deductions here were choices rather than terms, and what you would do about each.
28.5 [Judgment · Level 2] The kitchen is at capacity from 7:00 to 8:30 on Fridays and delivery orders are arriving through the same window. A manager proposes turning the channel off during those 90 minutes. Argue for or against, and name what the platform's algorithm may do in response.
28.6 [Computation · Level 3] [the Bellwether plan] Full channel arithmetic on a \$42.00 delivery order: 28% commission; 30% food cost on menu value; \$1.35 packaging; a dedicated packer works 4 hours at \$17 with 22% burden and handles 22 orders in that window. A dine-in cover averages \$46 at a 40.53% contribution margin ratio. (a) Compute unabsorbed labor per order. (b) Compute channel contribution per order. (c) Compute the contribution of one dine-in cover. (d) If each peak delivery order displaces 0.8 of a dine-in cover, compute the net effect per order. (e) State the rule this establishes, and say why the same order is a good order at 4:30 p.m.
Chapter 29 — Events, Catering, and the BEO
29.1 [Recall · Level 1] What is a BEO? Name six things it must specify, and state which one causes the most disputes.
29.2 [Computation · Level 1] A 40-guest private party carries a \$65 per person food-and-beverage minimum, a 20% service charge, and 7% sales tax on the total. Compute the F&B subtotal, the service charge, the tax, and the guest's total. Note the jurisdictional caveat on one of those lines.
29.3 [Computation · Level 2] On that \$2,600 of F&B: product cost runs 26%; five staff work five hours each at \$17 with 22% burden; rentals cost \$240. (a) Compute contribution in dollars and as a percentage of F&B revenue. (b) Compare that percentage with a normal dinner service and explain the difference. (c) Name the cost most often left out of event pricing.
29.4 [Judgment · Level 2] A buyout inquiry would close the dining room on a Saturday. The organizer's budget is firm. Describe exactly how you would price it — what floor you compute first, what you add, and what you refuse to discount — and name the two non-financial reasons to decline a buyout that clears the floor.
29.5 [Artifact · Level 2] A BEO:
BANQUET EVENT ORDER (extract) [constructed teaching example]
Event date Saturday, October 18
Guests guaranteed 55, expected 65
Space dining room, 6:00-10:00 p.m.
Menu three-course plated, $72/person
Beverage consumption bar
Deposit $1,000 received
Balance due day of event
Final count due —
Cancellation —
Two lines are blank and one is a material risk as written. Identify all three, state what each must say, and compute the exposure if 55 are guaranteed and 40 attend.
29.6 [Computation · Level 3] [the Bellwether plan] A normal Saturday does 132 covers at a \$46 average check with a 40.53% contribution margin ratio. A buyout offer is 90 guests at \$85 per person, with 26% F&B cost and \$900 of incremental labor and rentals. (a) Compute the displaced contribution the buyout must clear. (b) Compute the buyout's contribution. (c) Compute the margin over the floor. (d) The owner's instinct is that \$7,650 in one night obviously beats \$6,072. Explain what that comparison leaves out and why it happens to reach the right answer here anyway.
Chapter 30 — Trucks, Pop-Ups, and Ghost Kitchens
30.1 [Recall · Level 1] Define capital efficiency and distinguish it from scale. State which one a first-time operator should optimize for and why.
30.2 [Recall · Level 1] What is a commissary requirement, and why is "a truck has no rent" false? Name three fixed costs a truck carries that a first-time operator typically omits from the plan.
30.3 [Computation · Level 1] A truck requires \$185,000 of capital and produces \$420,000 of annual revenue. A restaurant requires \$620,000 and produces \$1,550,000. Compute revenue per dollar invested for each.
30.4 [Computation · Level 2] A truck's annual P&L: revenue \$420,000; COGS 30%; labor 26%; commissary \$14,400; fuel and maintenance \$18,000; permits and insurance \$11,500; event and location fees 8% of revenue; other operating \$21,000. (a) Compute each cost line, total costs, and operating profit in dollars and percent. (b) Explain why the profit percentage is so much higher than a restaurant's, and what it is compensating the operator for.
30.5 [Artifact · Level 2] A ghost-kitchen model:
GHOST KITCHEN - monthly [constructed teaching example]
Orders 960
Gross order value $33,600
Third-party commission (avg 26%) $8,736
Food cost (31% of gross) $10,416
Labor $8,900
Kitchen license fee $3,200
Packaging $1,296
Technology and other $940
Compute total costs and monthly result. Then state the structural problem this model has that a dining room does not, in one sentence.
30.6 [Judgment · Level 3] A truck operator clearing \$86,300 a year wants to open a 60-seat restaurant. Write the counsel. Address what the truck's profit does and does not predict, which skills transfer and which do not, what the capital requirement does to personal risk, and name the one intermediate step you would recommend first.
Part VII — The Numbers That Run the Business (Chapters 31–34)
Chapter 31 — The P&L, the Chart of Accounts, and the Flash Report
31.1 [Recall · Level 1] Distinguish a comp from a void. State where each belongs in the accounts and what a rising rate of each one tells you about a different problem.
31.2 [Recall · Level 1] What is a flash report, what does it contain at minimum, and on what cadence? State the specific consequence of running it monthly instead.
31.3 [Computation · Level 2] One week at a bistro: sales \$29,800; food purchases \$7,900; beginning food inventory \$18,400; ending food inventory \$17,650; beverage usage \$2,180; total labor \$9,640. (a) Compute food usage. (b) Compute total COGS. (c) Compute prime cost in dollars and percent. (d) State the one action this number demands on Monday.
31.4 [Artifact · Level 2] Four weeks of flash:
FLASH REPORT - four consecutive weeks [constructed teaching example]
Week Sales COGS Labor Prime Prime %
1 $29,400 $8,290 $9,410 $17,700 60.2%
2 $28,100 $8,180 $9,380 $17,560 62.5%
3 $27,600 $8,340 $9,520 $17,860 64.7%
4 $26,900 $8,410 $9,610 $18,020 67.0%
Describe what is happening in dollar terms, not percentage terms. Name which of the two halves of prime cost is actually moving and which is only appearing to move, and state what you would do first.
31.5 [Judgment · Level 3] The outside bookkeeper posts smallwares, cleaning chemicals, and to-go containers to cost of goods sold, arguing that they are consumables. Prime cost as reported has run 63% for two quarters. Write the response. Explain what this does to every benchmark comparison the operator makes, what it does to the flash report's usefulness, what the correct treatment is, and how you would restate the prior quarters.
31.6 [Computation · Level 3] [the Bellwether plan] The plan shows: revenue \$1,550,000 (100.0%); COGS \$430,280 (27.8%); labor \$500,000 (32.3%); prime cost \$930,280 (60.0%). A student adds 27.8 and 32.3, gets 60.1, and reports that the plan contains an arithmetic error. (a) Compute each percentage from the dollars to three decimal places. (b) Compute prime cost percentage from the dollars. (c) Explain what the student did wrong, and state the rule that prevents it. (d) Compute the dollar error a student would introduce by reconstructing COGS from the displayed 27.8%.
Chapter 32 — Break-Even and Sensitivity
32.1 [Recall · Level 1] Define contribution margin ratio and break-even sales. State the one classification error that makes a break-even calculation useless.
32.2 [Recall · Level 1] Define margin of safety, in both dollar and percentage form, and state what it is actually measuring about a business.
32.3 [Computation · Level 2] A restaurant has \$340,000 of annual fixed costs and a 40% contribution margin ratio. Average check is \$38, and it operates 260 services a year. (a) Compute break-even sales. (b) Compute break-even covers per year. (c) Compute break-even covers per service. (d) State what changes about (c) if the restaurant closes Mondays.
32.4 [Artifact · Level 2] A break-even worksheet:
BREAK-EVEN WORKSHEET [constructed teaching example]
FIXED COSTS
Occupancy $92,000
Salaried management and chef $146,000
Hourly kitchen and floor labor $214,000
Insurance, licenses, G&A $51,000
Fixed share of other operating $88,000
────────────────────────────────────────────────
TOTAL "FIXED" $591,000
Contribution margin ratio 40.0%
BREAK-EVEN SALES $1,477,500
Confirm the arithmetic, then find the classification error, correct it, and recompute break-even sales. State how large the error was in sales dollars.
32.5 [Judgment · Level 3] [the Bellwether plan] Two labor lines exist for the same restaurant at the same \$1,550,000 of revenue. The summary P&L carries labor at **\$500,000, which leaves variable costs of \$851,345 (54.93%) and a contribution margin ratio of **45.07%**. The chapter that builds the labor model from an actual schedule carries labor at **\$570,461, which leaves variable costs of \$921,806 (59.47%) and a contribution margin ratio of **40.53%**. Total fixed cost is **\$437,635 on both**. (a) Compute break-even sales under each ratio. (b) Explain what could make the same restaurant's labor line differ by \$70,461. (c) State which break-even you would put in front of a lender, which you would put on the wall of the office, and why those are not the same answer. (d) State the rule this establishes about quoting a contribution margin ratio.
32.6 [Computation · Level 3] [the Bellwether plan] Total fixed cost is \$437,635**: fixed labor \$191,895 (salaried \$168,935 plus a \$22,960 open-and-close hourly floor), occupancy \$95,200, genuinely fixed other operating \$60,950, the fixed base of the semi-variable other-operating lines \$43,090, and G&A \$46,500. The contribution margin ratio is 40.528645%. Annual debt service is \$69,500 and planned revenue is \$1,550,000 on about 95 covers a night. (a) Confirm the fixed-cost total, then compute accounting break-even sales. (b) Compute margin of safety in dollars and percent. (c) Compute cash break-even including debt service. (d) Convert both break-evens to covers a night — and name the base you converted on**, because a break-even in covers is meaningless without it. (e) Explain in one sentence why (c) is the number an operator should carry in their head.
Chapter 33 — Cash
33.1 [Recall · Level 1] Name four specific reasons EBITDA overstates the cash a restaurant actually generates. Which one surprises first-time operators most?
33.2 [Computation · Level 1] A restaurant collects 7% sales tax on \$29,800 of weekly sales, remitted monthly. Compute the weekly accrual and the approximate balance sitting in the account at the end of a month. State what that balance is not.
33.3 [Computation · Level 2] [the Bellwether plan] In February, revenue runs at 72% of the \$29,800 weekly average. Variable costs run 59.47% of sales. Annual fixed costs of \$437,635 and annual debt service of \$69,500 both land evenly across 52 weeks. (a) Compute the weekly fixed and debt outflows. (b) Compute weekly revenue, variable cost, and net cash flow. (c) Compute the month's cash burn over 4.33 weeks. (d) State what this figure obligates the operator to have arranged, and by when.
33.4 [Judgment · Level 2] October was strong and the account holds \$71,000. The owner wants a \$40,000 distribution. It is October 30. Argue the case, name the two documents you would want on the table before answering, and state the rule you would set so the question does not have to be re-argued next October.
33.5 [Artifact · Level 2] A six-week forecast:
CASH FORECAST - six weeks [constructed teaching example]
Week Opening Receipts Disbursements Closing
1 $38,000 $27,400 $31,200
2 $26,100 $29,800
3 $24,600 $38,900 (rent + annual insurance)
4 $23,900 $31,500
5 $25,800 $33,400 (quarterly sales tax)
6 $28,700 $29,600
Complete the closing balances. Identify the trough week and its balance, and state the size of the line of credit you would arrange and why that number is not the same as the trough.
33.6 [Computation · Level 3] A restaurant reports EBITDA of \$148,000. Also in the year: interest \$31,000; debt principal repayment \$38,000; capital expenditure \$26,000; an inventory build of \$9,000; and a favorable sales-tax timing swing of \$4,000. (a) Compute cash actually generated. (b) The owner planned \$120,000 of distributions off the EBITDA figure. State the shortfall. (c) Name which of these five items an operator can defer and which they cannot, and what deferring the deferrable one costs later.
Chapter 34 — Controls and the Variance Ladder
34.1 [Recall · Level 1] Name the three functions that must not sit with one person, define each in restaurant terms, and give the most common restaurant example of all three sitting together.
34.2 [Recall · Level 1] Describe the variance ladder from the P&L down to the individual item. State what each rung narrows and where most operators stop.
34.3 [Computation · Level 1] A period shows 84 voids against 2,180 tickets. Compute the void rate and state what threshold would make you look harder.
34.4 [Computation · Level 2] Comps for a month total \$4,180 against \$118,400 of sales; policy is 1.0%. (a) Compute the comp rate. (b) Compute dollars over policy for the month and annualized. (c) Name three legitimate causes and three illegitimate ones, and state which report distinguishes them.
34.5 [Artifact · Level 2] A comp report by employee:
COMPS BY SERVER - 4 weeks [constructed teaching example]
Server Sales Sales % Comps Comp % Comps as % of own sales
A $31,600 26.7% $410 9.8% 1.3%
B $28,900 24.4% $355 8.5% 1.2%
C $14,200 12.0% $1,714 41.0% 12.1%
D $24,300 20.5% $498 11.9% 2.0%
E $19,400 16.4% $1,203 28.8% 6.2%
Confirm the totals reconcile to \$4,180 of comps. Identify what stands out, state the three explanations you must rule out before drawing a conclusion, and describe the conversation you would have first.
34.6 [Judgment · Level 3] The chef-owner approves invoices, signs checks, does the bank deposit, has the only key to the office, and adjusts the inventory count when it "looks wrong." No theft is suspected. Write the case for changing this. Address what the controls protect the owner from rather than what they detect, what a lender or insurer will say about it, and design the minimum viable control structure for a business too small to hire anyone.
Part VIII — Growth, Endurance, and the Career (Chapters 35–40)
Chapter 35 — Growth Criteria and Owner Dependency
35.1 [Recall · Level 1] Define owner dependency. State the practical test for it and the two systems that most reduce it.
35.2 [Computation · Level 1] A restaurant reports \$96,000 of operating profit. The owner draws \$45,000, included in labor, and performs both the general manager and working chef roles. Hiring both would cost \$146,000 fully loaded. Compute the restaurant's profit on a fully staffed basis.
35.3 [Computation · Level 2] A second unit requires \$480,000 of capital. In year one it is expected to run at 60% of unit one's \$96,000 profit, and it adds \$34,000 of above-store overhead. (a) Compute year-one profit from unit two and the return on the new capital. (b) State the two things that must be true for that return to be acceptable.
35.4 [Judgment · Level 2] A landlord offers a second space two blocks away with 30 days to decide and \$100,000 of tenant improvement money. Unit one is profitable but the owner works six shifts a week in it. Argue the decision, and name the one condition that would make you say yes.
35.5 [Artifact · Level 2] A readiness checklist:
GROWTH READINESS [constructed teaching example]
Criterion Status
Unit 1 prime cost at or below target 6 months yes
A general manager who runs the unit alone no
Written systems: recipes, specs, training partial
12 weeks of operating cash after the raise no
Owner absent 2 consecutive weeks without loss untested
Second-site trade area analysis complete yes
State which unmet criterion is disqualifying, which are merely expensive, and what the "untested" row should become before a lease is signed.
35.6 [Computation · Level 3] Same operator. During the build and the first six months of unit two, unit one's profit is expected to fall 15% because the owner is absent. (a) Compute unit one's reduced profit. (b) Add unit two's year-one contribution from 35.3. (c) Compare the total to the \$96,000 standalone and compute the return on \$480,000 of new capital. (d) The owner's pitch to a lender is that a second unit "doubles the business." Explain what that framing gets wrong and what the honest version of the pitch is.
Chapter 36 — Franchising
36.1 [Recall · Level 1] Name the four components of a typical franchise fee stack and state which one continues regardless of profitability.
36.2 [Computation · Level 1] A franchise charges a 5% royalty and a 2% advertising fund contribution, both on gross sales. On \$1,300,000, compute the annual total.
36.3 [Computation · Level 1] The initial franchise fee is \$45,000, amortized over a ten-year term. Compute the annual amount and express it as a percentage of \$1,300,000 of sales.
36.4 [Artifact · Level 2] A disclosure summary:
FRANCHISE SUMMARY (extract) [constructed teaching example]
Initial franchise fee $45,000
Royalty 5.0% of gross sales
Advertising fund 2.0% of gross sales
Technology fee $650/month
Estimated initial investment $610,000 - $940,000
Term 10 years
Renewal fee $22,500
Territory "as designated by franchisor"
Required remodel every 7 years
Compute the total annual fee load on \$1,300,000 of sales. Then identify the two entries here that carry the most unquantified risk and state what you would demand in writing about each.
36.5 [Judgment · Level 2] A first-time operator with \$210,000 and no restaurant management experience is choosing between a franchise and an independent concept. Argue both sides honestly, then state what you would advise and what single fact about the candidate would change your advice.
36.6 [Computation · Level 3] Using 36.2 and 36.3, the franchisee's total annual fee load is \$95,500 on \$1,300,000. The contribution margin ratio is 40%. (a) Compute the incremental sales required to cover the fee load. (b) Express that as a percentage of the sales base. (c) State the question this reframes: what must the brand deliver for the arrangement to be worth it? (d) Name two costs a franchisee avoids that partly offset the fee load, and why they are hard to quantify.
Chapter 37 — Multi-Unit Management
37.1 [Recall · Level 1] Define comparable-store sales and the base-period rule. State the two things that most often make a comp number misleading.
37.2 [Recall · Level 1] What is span of control in a multi-unit restaurant business? Name three factors that determine how many units one above-store leader can carry.
37.3 [Computation · Level 1] A unit does \$1,412,000 this year against \$1,340,000 last year. Compute comparable-store sales growth.
37.4 [Computation · Level 2] Decompose that growth: traffic fell 2.1% and average check rose 7.6%. (a) Confirm the combination reproduces the comp figure. (b) State what this decomposition reveals that the headline number hides. (c) Name what you would look at next, and what you would expect to find.
37.5 [Artifact · Level 2] A four-unit read:
COMP-STORE READ - trailing 12 months [constructed teaching example]
Unit Comp sales Traffic Check Prime % Turnover
1 +5.4% -2.1% +7.6% 59.8% 61%
2 +1.9% +2.4% -0.5% 61.2% 74%
3 -3.1% -4.0% +0.9% 64.9% 96%
4 +6.8% +5.1% +1.6% 58.4% 48%
Identify the healthiest unit and the one in the most trouble, and say why the answer to the second is not simply "unit 3." Name the unit whose growth is least durable.
37.6 [Judgment · Level 3] A district manager carries nine units. One is failing: prime cost 65%, turnover 96%, comps −3.1%. The DM is spending three days a week there and the other eight are drifting. Write the intervention plan. Address what the DM's actual job is, what you would take off their plate, what you would do about the failing unit specifically, and what number tells you the intervention worked.
Chapter 38 — Sustainability, Measured in Dollars
38.1 [Recall · Level 1] Name the three categories of food waste in a restaurant, define each, and state which is most recoverable and which is least.
38.2 [Computation · Level 1] A waste audit finds 38 pounds of pre-consumer prep waste a day at an average value of \$2.10 a pound, across 310 service days. Compute the annual value.
38.3 [Computation · Level 2] Spoilage runs \$14,200 a year. A \$6,400 investment in vacuum sealing, labeling, and rotation discipline is expected to reduce it by 35%. (a) Compute the annual saving and the simple payback period. (b) Name the recurring cost this calculation omits, and state whether it changes the decision.
38.4 [Judgment · Level 2] Compostable to-go packaging costs three times as much as the current containers, and the local waste hauler does not offer commercial composting. Decide, and state what you would tell a guest who asks. Name the one condition that would change the answer.
38.5 [Artifact · Level 2] A utility comparison:
UTILITY REVIEW - monthly averages [constructed teaching example]
Electricity $2,940
Gas $860
Water and sewer $610
Waste hauling $445
Proposed: LED retrofit + hood controls $9,800 capital
Expected electricity reduction 14%
Compute the monthly and annual saving and the payback period. Then name two non-electricity lines on this list that are usually negotiable and how you would attack each.
38.6 [Computation · Level 3] An operator proposes cutting a 12-ounce protein portion to 10 ounces to reduce plate waste, which averages 2.4 ounces per plate. Protein costs \$0.44 an ounce and the item sells 190 times a week. (a) Compute the weekly and annual product saving. (b) Compare it to the prep-waste figure from 38.2 and the spoilage figure from 38.3. (c) Explain why plate waste is the least recoverable of the three categories even though it is the most visible, and state where the operator's attention actually belongs.
Chapter 39 — Reading a Statement in Trouble
39.1 [Recall · Level 1] Name the four questions to ask, in order, of a profit-and-loss statement that is losing money. State why the order matters.
39.2 [Recall · Level 1] Distinguish a recoverable cost from an unrecoverable one in a restaurant. Give two examples of each, and state the test.
39.3 [Computation · Level 2] A distressed statement: revenue \$1,180,000; COGS \$401,200; labor \$460,200; occupancy \$118,000; other operating \$212,400; G&A \$47,200. (a) Compute each line as a percentage of revenue. (b) Compute prime cost in dollars and percent. (c) Compute total costs and the operating result. (d) State which single line you would look at first and why.
39.4 [Artifact · Level 2] The same statement against benchmarks:
P&L VS. BENCHMARK - trailing 12 months [constructed teaching example]
Line Actual Actual % Full-service benchmark
Revenue $1,180,000 100.0%
COGS $401,200 34.0% 28-33%
Labor $460,200 39.0% 30-36%
PRIME $861,400 73.0% at or below 60%
Occupancy $118,000 10.0% 6-10%
Other operating $212,400 18.0% 12-18%
G&A $47,200 4.0% 2-5%
OPERATING RESULT -$59,000 -5.0% 3-10%
Three lines are inside benchmark and three are outside. Identify each group. Then state which of the outside lines is the one this operator cannot fix this year, and why that changes the whole plan.
39.5 [Judgment · Level 3] Five proposals arrive for the restaurant in 39.3 and 39.4: raise menu prices 6%; cut two positions; renegotiate the lease; drop lunch service; refinance the equipment lease to lower payments. Rank them by what you would do first, second, and not at all. Justify the ranking in terms of speed, size, and reversibility, and name the proposal that treats a symptom.
39.6 [Computation · Level 3] Same statement. Suppose COGS can reach 30%, labor 34%, and other operating 15% within two quarters, while occupancy is contractually fixed in dollars. (a) Compute the dollar saving from each of the three recoverable lines. (b) Compute the resulting operating profit and margin. (c) Occupancy is the worst line against benchmark at 10.0%. Compute the revenue level at which it would reach 8.0% with no change in the lease. (d) State what (c) proves about the difference between a percentage problem and a dollar problem.
Chapter 40 — The Plan, the Lender, and the Career
40.1 [Recall · Level 1] What is a DSCR covenant? State what it measures, how it is typically tested, and what happens on a breach.
40.2 [Recall · Level 1] Name five conditions a lender may attach to an approval beyond the interest rate and term, and state what each one is protecting the lender against.
40.3 [Computation · Level 1] [the Bellwether plan] Operating profit on plan is \$261,020 and annual debt service is \$69,500. Compute the debt service coverage ratio.
40.4 [Computation · Level 2] [the Bellwether plan] The loan carries a 1.25× DSCR covenant tested annually against \$69,500 of debt service. (a) Compute the operating profit at which the covenant is exactly met. (b) Compute the cushion between plan and that trip point, in dollars and as a percentage of planned operating profit. (c) State what a covenant tested annually cannot detect.
40.5 [Artifact · Level 2] [the Bellwether plan] The lender's file:
CREDIT FILE SUMMARY - SBA 7(a), $335,000 [the Bellwether plan]
Covenant DSCR 1.25x, tested annually
Debt service $69,500
Trip point $86,875 of operating profit
MODELED SCENARIOS Operating profit DSCR Covenant
On plan $261,020 3.76x clears
Higher labor line $214,060 3.08x clears
Combined downside $154,985 2.23x clears
CASH FORECAST NOTE
Under the combined downside, the operating account reaches its low
point in the second week of February and goes negative before the
month closes.
Read the file and state the conclusion it supports. Then state, in one sentence, what is wrong with the covenant given what the cash forecast shows.
40.6 [Judgment · Level 3] The loan is approved at \$335,000 with conditions including the annual 1.25× DSCR covenant. Every modeled scenario clears the covenant, and the business still runs out of cash in February. Write the argument. Explain why the covenant is the wrong instrument although it was correctly applied; propose the covenant or condition you would write instead and say exactly what it would test and how often; and state what an operator should do about a covenant that will not warn them in time — regardless of what the lender writes.
Answers
Computation answers show the steps so partial credit is assignable. Judgment answers give what a strong response contains — a student reaching a different conclusion for stated, costed reasons earns full marks.
Part I
1.1 (c) — roughly 26–27% in year one, approaching 60% cumulatively by year three. The 90% and year-one-60% figures are folklore; the honest shape is that failure is front-loaded but not catastrophic in year one. Worse, not better, because most casualties are businesses that worked for a while and then bled — a pattern caused by cost drift, which is countable and therefore was preventable. Note also that "failure" in this literature generally means closed or changed ownership.
1.2 A cover is one guest served. A check is one bill; a table is one party's seating. The example: 6 covers, 3 checks, 2 tables (or 2 seatings — the bar counts as a table location). Covers is the denominator for average check, seat turns, and RevPASH; checks and tables are not.
1.3 \$11,400 + \$13,300 = \$24,700**; ÷ \$38,000 = 65.0%**. Above the ≤60% full-service rule of thumb: workable but tight, with essentially no cushion. Accept "distressed-adjacent"; do not accept "fine."
1.4 (a) 50 × 1.6 = 80 covers; × \$38 = \$3,040 per service; × 5 = \$15,200 a week; × 50 = **\$760,000. (b) 80 × \$2 = \$160 per service; × 5 × 50 = \$40,000**. (c) Turns — it is the least controllable and the most commonly overstated of the four variables, and an error there is unrecoverable; check average can be moved by menu and training after opening.
1.5 \$249,600 + \$296,400 = \$546,000 = 70.0% of sales. At 70% the business is consuming itself. Before answering: the service style (a counter taqueria should run lower labor than full service, so 38% is the anomaly), whether the owner's own labor is in that number, and whether COGS includes paper and packaging. Provisional recommendation: labor, because 38% on a counter model is far outside its own benchmark, and because labor is the half of prime cost that can move this week.
1.6 A strong response: (1) corrects the claim — first-year closure is roughly a quarter, not most, so surviving twelve months was never the hard filter she believes it was; (2) reads the curve — the largest share of casualties fall in years two and three, so she is entering the dangerous period, not leaving it; (3) explains the causal implication — a business that dies at month twenty-nine was working at some point, which means the cause is slow drift rather than a bad concept; (4) names the specific month-fifteen-to-thirty hazards: opening capital spent, novelty traffic normalized, the first rent escalation, the opening staff cohort turning over, and deferred equipment maintenance arriving; (5) prescribes something concrete — weekly prime cost, a thirteen-week cash forecast, a reserve she does not touch.
2.1 A concept is a testable claim about a market: who the guest is, what occasion they are buying, and why they choose you over the specific alternatives within their consideration set. Cuisine specifies none of those. A concept that cannot be wrong is not a concept.
2.2 1,140 + 60 = 1,200 total seats; 60 ÷ 1,200 = 5.0%.
2.3 (a) 95 × 5 = 475 covers a week. (b) 475 ÷ 2,400 = 19.8%. (c) As the twelfth operator, an equal share is 2,400 ÷ 12 = 200 covers, or 8.3%. (d) The plan must explain why this restaurant takes roughly 2.4 times an equal share — either by growing the market, by taking share from named competitors for a stated reason, or by serving an occasion the set does not.
2.4 A strong response: better execution of an existing position is not a concept, it is a bet that guests will notice — and guests reliably under-notice quality differences at the same price and format. Change one structural variable: the occasion (late-night, family early, counter lunch), the format (bar-forward, counter service, tasting-only), or the price band. Evidence: a shift in the source of trade (new guests rather than switchers), repeat rate at 90 days, and whether the average check holds without discounting.
2.5 Strong answers identify a mid-price, reservable, quiet, adult-occasion dinner around a \$45–55 check: the set has cheap-and-loud (2, 4), casual mid (1), one expensive room (3), and a small-plates bar (5) that is not a full dinner. Largest risk: at that check the room needs 1.5+ turns and a beverage attachment the district may not support on weeknights — the gap may exist because the demand does not.
2.6 (a) \$1,440,000 ÷ 72 = **\$20,000 per seat. (b) The set averages \$14,800 and its best performer does \$17,600; the plan requires 13.6% more than the best operator in the district (20,000 ÷ 17,600 = 1.136) and 35% more than the average. (c) Uniqueness explains why a restaurant might take share; it does not raise the ceiling on what a seat in this district can produce, which is set by traffic, price tolerance, and how many nights people go out. The plan must now show either a higher check with support, a turns assumption benchmarked to something real, or a lower revenue line — and if it holds \$1,440,000, it must state which competitor's covers it takes and why.
3.1 A brand is the promise a guest expects you to keep before they walk in; a logo is a mark, a name is a label, and a budget is spend. The test: if a regular described you to a friend in one sentence, would that sentence match what the restaurant actually does on a Tuesday? If not, there is no brand — there is decor.
3.2 \$38 ÷ \$31 = 122.6%. The guest must perceive roughly 23% more value than the set delivers — in product, room, service, or occasion — or the index reads as "expensive" rather than "premium."
3.3 (a) 100 × \$31 = \$3,100; 88 × \$38 = **\$3,344. (b) +7.9%**. (c) COGS falls with 12 fewer covers (fewer plates), and variable labor may fall slightly; but occupancy and fixed labor do not move at all, and marketing usually rises during a repositioning. Note that contribution per cover rises even faster than revenue, which is why repositioning can work even with lost covers.
3.4 A strong response: both are describing real things, and the question is empirical. The wait contradicts "without planning it in advance," which is the operational half of the promise; but reservations move covers from the wait list to the book without creating any new seats, and they change who gets served, not how many. Measure: walk-away rate at the door, the mix of parties with children before 7:00, repeat rate of walk-in versus reserved guests, and RevPASH by hour. A partial book — half the room reservable — usually settles it in practice.
3.5 The statement says unfussy, family, spontaneous; the reviews say date night, long waits, and price-value friction at \$34 pasta with paper napkins. The operation has drifted upmarket in price without moving the room or the service to match. Strong answers change the statement — the operation is already succeeding at a different, coherent position — or change the operation on price, but not both halves at once. Either is defensible; changing nothing is not.
3.6 (a) Before: \$18.00 − \$5.40 = \$12.60**. After: price \$14.40 − \$5.40 = **\$9.00. (b) \$12.60 ÷ \$9.00 = 1.40 → 40% more units just to stand still. (c) The discount is 20% of price but 28.6% of contribution, because the plate cost does not discount. Discounts come entirely out of the margin, never out of the cost — which is why a 20% discount needs a 40% volume lift and a 30% discount needs one the business cannot produce.
4.1 A register lists every assumption the financial model rests on, with, for each: the value, the source, the test that would confirm or refute it, and the consequence if it is wrong (ideally sized in dollars). Its purpose is to convert a projection from an assertion into a set of falsifiable claims.
4.2 Typically: executive summary (is this coherent, and is this operator credible?); the financial projections and their assumptions (does it service the debt?); use of funds and the capital stack (how much of their own money is in it?); the operator's experience (have they done this before, and who covers what they cannot?); then market and concept. The lender is reading for repayment risk, not for the menu.
4.3 At 1.6: 60 × 1.6 = 96 covers; × \$40 = \$3,840; × 5 × 50 = \$960,000**. At 1.25: 60 × 1.25 = 75 covers; × \$40 = \$3,000; × 5 × 50 = **\$750,000. Difference: \$210,000.
4.4 (a) Draft: \$252,000 + (16.0% × \$1,550,000 = \$248,000) = **\$500,000. Plan model: \$191,895 + (19.88% × \$1,550,000 = \$308,140) = **\$500,035. The two agree to within \$35. (b) At \$1,200,000 — draft: \$252,000 + \$192,000 = \$444,000**; plan model: \$191,895 + \$238,560 = **\$430,455; the draft overstates labor by \$13,545**. (c) The two errors run in opposite directions and cancel at exactly one volume: \$60,105 of excess fixed cost against 3.88 points of missing slope, and 3.88% of \$1,550,000 is \$60,140. The draft was fitted to the plan's own revenue number, so it reproduces the plan's own labor number — and nothing else. The general lesson: a cost model that agrees with the plan at the planned volume has been tested against the one point at which no split is needed. A split exists to answer what happens when volume is different; it must be validated at a volume the plan does not assume. Note the direction, too: this draft overstates labor below \$1,550,000 and understates** it above, so it flatters the downside case and the upside case both, in the ways that matter least and most respectively.
4.5 The turns row. "Industry standard" is not a source and "none stated" is not a test. Require: turns observed at three comparable operators on comparable nights (counted, not estimated), the reservation-system or POS data behind any figure cited, and a stated sensitivity showing the plan at 1.25 and 1.4 as well as 1.6.
4.6 A strong memo: (1) each assumption at the optimistic end may be individually defensible, but the model multiplies them, so five 85th-percentile assumptions produce a result far above the 85th percentile — probabilities compound downward; (2) the remedy is a base case built from central estimates with the optimistic case shown separately, plus a downside case that the working-capital reserve is sized against; (3) a lender will read "conservative" against the numbers, find it untrue, and discount everything else in the plan including the parts that were honest — the credibility cost exceeds the modeling cost.
5.1 A personal guarantee makes the individual personally liable for the business's obligation if the entity does not pay. It commonly survives (1) the closure or dissolution of the entity — dissolving the LLC does not extinguish it — and (2) the sale of the business, unless the guarantee is expressly released in writing by the lender or landlord. Many operators also do not realize a lease guarantee can run for the remaining term, not just to the date they hand back the keys.
5.2 (a) \$150,000 + \$75,000 + \$60,000 + \$335,000 = \$620,000** — it funds the project exactly, with no cushion beyond the \$45,000 reserve already inside the project cost. (b) \$150,000 ÷ \$620,000 = 24.2%**.
5.3 \$96,000 ÷ \$64,000 = 1.50×. Clears a 1.25× threshold with room.
5.4 The listed items total **\$514,800** (268,000 + 142,000 + 38,000 + 29,000 + 11,000 + 26,800). Missing: a **working-capital reserve**, separate money for payroll, inventory, and rent during the ramp. On this project size, \$40,000–\$60,000 is a defensible range; less than a month of operating costs is not. Without it, month four — after the opening spike normalizes and before the systems mature — is funded by stretching vendors and personal credit cards, which is how an undercapitalized restaurant becomes a distressed one.
5.5 Strong answers note: the loan is cheaper only if it is repaid, and a 13% five-year note lands a fixed payment on a business with no operating history — it converts a business risk into a cash-flow obligation in exactly the year the business is least able to carry one. Equity is expensive and permanent, but it absorbs a bad year. What flips it: a business already generating stable cash with a demonstrated cushion, a materially lower rate, or an investor whose terms include control provisions the operator cannot live with.
5.6 (a) Equipment liquidates at \$185,000 × 25% = \$46,250 against a \$60,000 lease balance → **\$13,750 deficiency. The loan has \$298,000 outstanding against \$55,000 of realizable collateral → \$243,000 deficiency**. (b) Total personal exposure under the guarantees: 13,750 + 243,000 = **\$256,750, after the \$150,000 of equity is already gone. (c) Equity is the floor of what you lose, not the ceiling. A personal guarantee converts every unsatisfied business obligation into a personal one; the owner's real worst case here is roughly \$407,000 of value destroyed, plus the lease deficiency if the landlord pursues the remaining term.
Part II
6.1 Occupancy cost = base rent plus NNN or CAM charges, property tax, property insurance, and any percentage rent — plus, in some leases, a share of capital repairs. The distinction changes the decision because two spaces quoted at the same base rent can differ by \$3–8 per square foot all-in, which on 2,500 sq ft is \$7,500–\$20,000 a year, and because NNN is usually reconciled annually and often uncapped.
6.2 2,400 × (\$26 + \$7) = 2,400 × \$33 = **\$79,200 a year, or \$6,600 a month**.
6.3 (a) \$79,200 ÷ \$880,000 = 9.0%. (b) \$79,200 ÷ 0.07 = **\$1,131,429, call it \$1,131,000. (c) Neither is easy, but sales** is the only one that moves in year one — rent is fixed the day the lease is signed. That asymmetry is the whole reason occupancy is evaluated before signing, not after.
6.4 Strong answers price the difference: 12 × the square footage in annual rent, then ask what traffic buys. A 70-seat dinner house with a bar draws by destination and reservation more than by walk-by, which argues for Site B and its parking; a walk-in-dependent lunch or counter concept argues for Site A. The data that would change the answer: observed walk-by counts at 6:30 p.m. on a weekday at Site A, and the parking reality at Site B on a Saturday.
6.5 The two uncapped risks are the NNN estimate with no cap (an "estimate" that reconciles annually can rise without limit, and a roof or parking-lot repair can land in it) and the 3% compounding escalation (by year ten, base rent is roughly 30% higher than year one, against a business whose prices may not move that much). Ask for: a cap on controllable CAM increases (3–5% annually), exclusion of capital items from CAM, and either a fixed-dollar escalation or escalation tied to a published index with a ceiling. Also negotiate rent commencement to opening plus a defined fixturing period, not a flat 120 days.
6.6 (a) (\$1,750,000 − \$1,500,000) × 6% = \$15,000**. (b) Total rent \$90,000 + \$15,000 = **\$105,000, which is 6.0% of sales. (c) The bookkeeper computed 6% × \$1,750,000 = \$105,000 of percentage rent on top of base rent, reporting \$195,000. Percentage rent applies only to sales **above the breakpoint**. A *natural* breakpoint is base rent ÷ the percentage rate (\$90,000 ÷ 0.06 = \$1,500,000), so the landlord earns the same rent up to that point and shares only in growth beyond it. An artificial breakpoint set lower than that is a rent increase in disguise.
7.1 Capacity from the room is seats × turns — how many people can physically be seated and served in the service window. Capacity from the kitchen is the sustained output of the binding station — how many plates can actually leave the pass per hour. Business plans compute the room, because seats and turns are the numbers everyone has. The kitchen is what fails on Saturday.
7.2 22 × 5 = 110 grill plates; 110 ÷ 0.60 = 183 covers across the service.
7.3 (a) 72 × 1.8 = 130 covers. (b) Grill supports 183. (c) The room binds at 130. The consequence of marketing against the kitchen number is spending money to create demand the dining room cannot seat — producing waits, walk-aways, and a compressed peak that then breaks the kitchen anyway.
7.4 Cost argument: a 36-inch aisle slows every plate and every bus tub in the room, which lengthens turn time — and if turn time rises even four minutes, the four extra covers are given back with interest across all 86 seats. Add server injury risk, breakage, and guests bumped by service traffic, which shows up in reviews rather than on any report. The two compliance questions: ADA accessible-route and clearance requirements, and local fire and building egress width. Both are jurisdictional; verify with the local authority and the architect before pricing the four seats at all.
7.5 12 + 36 + 16 + 14 + 8 = 86 seats; 1,450 ÷ 86 = 16.9 sq ft of FOH per seat, which is tight-but-normal for casual full service. Flexibility problem: the room is built for parties of four (13 four-tops, one 8-top) but most parties are two, so Tuesday runs deuces at four-tops and burns capacity, while Saturday cannot flex up without combining. Change: convert two or three four-tops to movable 2-tops that combine, and put the banquette on a rail so it seats 2s and 4s.
7.6 (a) 130 ÷ 5 = 26 covers per hour on average; peak hour = 130 × 40% = 52 covers. (b) 52 × 0.60 = 31.2 grill items in the peak hour. (c) Capacity 22 → a deficit of about 9 items an hour, or roughly a nine-minute ticket-time penalty compounding through the window. (d) The average is the wrong calculation because demand is not uniform and the kitchen cannot bank capacity from 5:00 to spend at 7:30 — an unused hour is gone. Fixes without hiring: pace the reservation book (cap covers per 15-minute slot), and move work off the grill — par-cook or finish one grill item in the oven, or re-engineer the menu so no more than 45% of covers route to one station.
8.1 (1) Personal guarantees — the lender and the landlord almost always require one, and it reaches personal assets directly, entity or no entity. (2) Personal conduct and trust-fund obligations — unremitted sales tax and payroll withholding, and individual liability for one's own negligent or wrongful acts, generally follow the person; so can liability where the entity's formalities were never observed. Add that an entity does not create liability insurance, and that dram-shop exposure is insured, not incorporated, away.
8.2 Typical stack (jurisdiction-dependent, and this is the caveat that must accompany the whole answer): business license (city or county); seller's permit / sales tax registration (state revenue department); health department permit and plan review (county or city health authority); building permit and certificate of occupancy (building department); food manager certification and food-handler cards (state or county, via an accredited program); liquor license (state alcohol authority, often with local approval); plus sign, fire, and grease-interceptor permits. Specifics vary by state, county, and city — verify locally.
8.3 \$140,000 ÷ 10 = **\$14,000 a year; ÷ \$1,200,000 = 1.17%** of sales. Note this is an amortization convention, not a cash outflow — the cash left the business at purchase, which is exactly why it belongs in the capital plan and not only on the P&L.
8.4 (a) Workers' comp = \$460,000 × 2.8% = \$12,880. Total = 6,400 + 4,200 + 5,800 + 12,880 + 2,600 = \$31,880**. (b) ÷ \$1,200,000 = 2.66% of sales. (c) Dropping table service for counter service moves the workers' compensation line (different class codes and payroll) and generally reduces general liability exposure modestly; dropping the bar removes liquor liability** (\$5,800) and typically reduces GL as well. Rates and class codes vary by state and carrier.
8.5 The certificate of occupancy gates opening — nothing else lets you admit the public — and it depends on the building final, which depends on the health plan review and inspection. But the item most likely to move the date is the liquor license transfer: a 90–150 day window filed March 18 lands anywhere from mid-June to mid-August, past a June 1 opening. This week: schedule the food manager certification (two days, not filed, and a health-inspection blocker), and get the license status in writing from the state authority so the opening decision in 8.6 can be made on facts rather than hope.
8.6 A strong analysis: revenue — 28% of sales is beverage, and at the plan's volume that is roughly \$400,000 a year, so a ten-week dry opening forfeits a large share of the contribution the ramp was supposed to build (beverage carries the best margin in the building); brand — a full-bar concept that opens dry trains its first several thousand guests to think of it as something else, and first impressions are the most expensive to change; staffing — you cannot hire and hold bartenders for a bar that does not exist, and the FOH labor model was built around beverage attachment; compliance — any "workaround" (BYOB with a corkage, a third-party pouring, a "donation" bar) is regulated, is often illegal, and can jeopardize the pending license itself. Verify with counsel and the state authority before considering it. Most strong answers take the free rent, open dry only if the license is confirmed within weeks, and otherwise delay.
9.1 The ramp is the period between opening and stabilized volume, during which sales spike on novelty, fall back, and then climb to a true baseline over roughly three to nine months. The opening spike is curiosity traffic with no repeat behavior attached; a plan that treats it as the baseline over-hires, over-orders, sets par levels too high, and — most damaging — spends the working-capital reserve believing revenue has arrived.
9.2 A soft open genuinely tests the flow of a ticket from POS to pass to table, station timing and prep quantities under real fire, and whether the room's physical layout works with bodies in it. It cannot test whether guests will pay (the guests are friends and the food is free or discounted, so no price signal exists) or sustained volume across a full week (one night of forgiving guests tells you nothing about a Saturday at 7:15 with strangers).
9.3 14 × 22 = 308 hours; × \$15.50 = **\$4,774** before payroll burden. At a 20% burden the real number is about \$5,729 — flag students who omit burden.
9.4 (a) Plan: 13 × \$29,800 = **\$387,400. Actual: (4 × 26,000) + (4 × 24,000) + (5 × 27,000) = 104,000 + 96,000 + 135,000 = \$335,000**. Shortfall: **\$52,400. (b) Weeks 5–8 are the post-novelty trough — curiosity traffic has been consumed and repeat traffic has not yet compounded. (c) The shortfall is a cash** event before it is a profit event: \$52,400 of revenue that never arrived, against fixed costs that did, comes straight out of the working-capital reserve. On a \$45,000 reserve, this quarter alone exhausts it.
9.5 T-3 initial food order is dangerously late — it must precede training (staff cannot train on food that is not in the building) and the friends-and-family service; move it to roughly T-16. T-30 menu finalized / cost cards complete is out of order relative to T-45 smallwares — you cannot order plateware and pans intelligently before the menu exists; move the menu to T-50 or later move smallwares to T-35. T-10 health final inspection request is out of order — request it early enough to allow a failed first inspection and a re-inspection before T-7; move to T-21.
9.6 A strong answer: quantify waiting — fifteen working days is roughly three weeks of payroll for a trained staff of 26 plus rent, easily \$60,000–\$90,000 on a business of this size, plus the reservations and press timing already committed. The limited-menu opening protects cash but opens the concept without the thing the concept is named for; it also trains the ramp on the wrong menu and risks the reviews that will define the first year. The non-negotiable constraint is compliance: a hood that will not pass inspection cannot operate the hearth, full stop, and there is no version of this decision in which the hearth runs uninspected. Best answers propose a third path — open on the limited menu only with the hearth items explicitly framed as "coming," hold the press, and cut pre-opening payroll to a core crew for the interval.
Part III
10.1 Cross-utilization is the practice of using the same purchased item or prepared component on multiple menu items. It reduces inventory carrying cost and spoilage (fewer SKUs, each turning faster) and prep labor (one batch serves several dishes). The risk it creates is correlated failure: when one component is 86'd, mis-made, or contaminated, several dishes go down at once, and a menu can start to taste the same.
10.2 96 − 68 = 28 fewer items; 28 ÷ 96 = 29.2% reduction.
10.3 (a) 14 × 3 = 42 sold within the window; 60 − 42 = 18 discarded; 18 ÷ 60 = 30% waste. (b) 20 × 3 = 60 sold; 0 discarded, 0% waste. (c) Waste is a function of turnover against shelf life, not of how carefully anyone works — cross-utilization increases turnover without increasing production.
10.4 Against: more SKUs, more spoilage, longer prep, more stations in the weeds, longer ticket times, harder training, more menu items that are never costed, and a guest decision problem that lengthens turn time. A 34-item menu also guarantees a long tail of items selling under the popularity threshold, which is dead inventory occupying real estate. The chef is right when the menu is built on deep cross-utilization — 34 items off 40 components, as in a trattoria or a diner — where the item count is a recombination, not an expansion.
10.5 Highest risk: duck confit (single-use, expensive, one dish) and nduja butter (single-use, one dish, and mussels are weather- and season-sensitive). Changes that keep both dishes: put duck confit on a second application — a brunch hash, a salad, or a small plate — so the batch turns; move nduja from a dedicated compound butter to a component the kitchen already carries (use the same chile-and-fat base on a pasta or a crostini). Chicken jus is also single-use but is a by-product of a high-volume item, so it costs almost nothing to carry.
10.6 (a) CM = \$26.00 − \$9.10 = \$16.90; × 9 = **\$152.10 a week. (b) Added cost: \$38 spoilage + prep (20 min × 5 days = 100 min = 1.67 hours × \$17 = \$28.33**) = **\$66.33; net so far \$85.77. (c) Cannibalized: 6 × \$18.40 = \$110.40**. (d) Net = 152.10 − 66.33 − 110.40 = **−\$24.63 a week, about −\$1,280 a year. Do not add the dish as configured. (e) The cannibalization** figure — operators reliably count the new item's sales as incremental. Strong answers note the dish could still be justified if it draws a guest who would not otherwise come, but that is a claim requiring evidence, not an assumption.
11.1 AP cost is what you pay per unit as delivered; EP cost is the cost per unit of the usable product after trim, peel, bone, and cooking loss. Yield percentage is EP weight ÷ AP weight. The EP cost belongs on the cost card, because a card built on AP cost understates every plate by the yield loss — on a 62%-yield protein, by more than a third.
11.2 Usage = beginning inventory + purchases − ending inventory. Invoices ÷ sales is not food cost because it measures what you bought, not what you used, and inventory levels move between the two. It is most misleading when (1) purchasing is lumpy — buying heavy in the last week of a period inflates the number, running the walk-in down flatters it — and (2) inventory is being built or drawn down deliberately, as during an opening or a menu change.
11.3 (a) Case cost = 12 × \$14.80 = **\$177.60; EP weight = 12 × 0.62 = 7.44 lb; EP cost = 177.60 ÷ 7.44 = \$23.87/lb**. (b) An 8-oz portion is 0.5 lb → **\$11.94. (c) The chef is partly right: if the trim is genuinely sold, its value should be credited. But it must be costed and sold, not merely used — and the honest treatment is a butcher's yield test** that assigns value to each fabricated stream (steak, tartare trim, bones for stock) so the entrée card carries its own share. Crediting trim you do not actually sell is how a cost card becomes fiction.
11.4 Waste allowance = \$6.15 × 3% = **\$0.18; plate cost = \$6.33**. Price at target = \$6.33 ÷ 0.30 = \$21.10**. Print **\$21 — the target is a guide, not a rule, and the round number reads better on the page than \$21.10; accept **\$22** with a stated reason (the item's CM, its position on the page, competitive set), which is the better answer if the student argues from contribution margin rather than from the percentage.
11.5 A strong response: the allowance is not an estimate of carelessness, it is the difference between recipe yield and real-world yield — over-portioning within tolerance, product that fails QC, the last serving in a pan, breakage, and the plate that goes back. It exists in every kitchen, including disciplined ones. Removing it does two things: it lowers every menu price by 2–3% of plate cost (small), and it moves that same variance into the ideal-versus-actual gap (large), where it looks like theft or waste and triggers investigations of a problem that is really a costing convention. The right response to the sous chef's accuracy concern is not to delete the allowance but to measure it — run yield tests on the ten highest-volume items and set item-specific allowances instead of a blanket percentage.
11.6 (a) Usage = 21,400 + 47,900 − 19,800 = \$49,500**; unadjusted food cost = 49,500 ÷ 162,000 = **30.56%**. (b) Deduct **employee meals (\$1,850) and transfers to the bar (\$690)** = \$2,540 → adjusted usage \$46,960 → 28.99%. Employee meals belong in labor or a benefits account; transfers belong in beverage cost, where the corresponding sales are. (c) Comps should not be deducted. A comped plate consumed real food; the loss is on the sales** side, not the cost side, and it must stay visible so the comp rate is managed. Deducting it would report 28.20% — a number that is 0.79 points better than the truth and improves precisely when comps get worse, which is exactly backwards.
12.1 Axes: menu mix / popularity (units sold as a share of category units) and contribution margin (price − plate cost, in dollars). Quadrants: Star (high popularity, high CM) — protect, never discount, feature it; Plowhorse (high popularity, low CM) — re-cost, re-portion, nudge the price, or reduce the plate's expensive component; Puzzle (low popularity, high CM) — reposition, rename, describe better, train the sell, or move it on the page; Dog (low popularity, low CM) — cut it, unless it exists for a reason outside its own economics.
12.2 Popularity threshold = (1 ÷ number of items in the category) × 70%. Margin threshold = the weighted average contribution margin across the category — total CM dollars ÷ total units sold, not the simple average of the item CMs.
12.3 (a) (1 ÷ 14) × 0.70 = 7.143% × 0.70 = 5.0%. (b) 4.2% is below the threshold — low popularity. (c) The 70% factor accepts that a menu cannot have every item at an equal share; it defines "unpopular" as meaningfully below equal share rather than merely below it, which prevents a fourteen-item menu from classifying half its items as failures by construction.
12.4 Popularity threshold = (1 ÷ 6) × 0.70 = 11.67%. Total CM dollars = (260 × 18.90) + (180 × 20.40) + (150 × 15.40) + (290 × 12.35) + (60 × 13.20) + (60 × 16.90) = 4,914 + 3,672 + 2,310 + 3,581.50 + 792 + 1,014 = \$16,283.50**; ÷ 1,000 units = **\$16.28 weighted average CM. Mix: chicken 26.0%, steak 18.0%, mussels 15.0%, burger 29.0%, trout 6.0%, cassoulet 6.0%.
| Item | Mix | CM | Classification |
|---|---|---|---|
| Roast chicken | 26.0% | \$18.90 | Star |
| Steak frites | 18.0% | \$20.40 | Star |
| Mussels | 15.0% | \$15.40 | Plowhorse |
| Burger | 29.0% | \$12.35 | Plowhorse |
| Trout | 6.0% | \$13.20 | Dog |
| Cassoulet | 6.0% | \$16.90 | Puzzle |
First action: the burger. It is 29% of units at the lowest CM on the menu, so it sets the restaurant's blended margin more than any other item; a \$1.50 price move or a \$0.60 cost reduction is worth roughly \$430–\$625 per four weeks, more than cutting the trout could ever return.
12.5 A strong response: food cost percentage is a ratio; you bank dollars. Cutting a high-volume, high-CM item removes the largest single block of contribution on the menu and improves the percentage precisely because it removes the dollars. The GM would be right only if the item's contribution could be replaced — that is, if guests who order it would buy a comparably priced item with a higher CM rather than leaving or trading down, which requires evidence, not assumption. The genuine case for removing such an item: when it is operationally destructive — it bottlenecks the binding station at the peak, it requires a SKU nothing else uses, or its execution is inconsistent enough to damage the room. That is an operations decision, not a food-cost decision, and it should be argued as one.
12.6 (a) A: food cost 4.00 ÷ 16.00 = 25.0%, CM \$12.00**. B: 12.24 ÷ 34.00 = **36.0%**, CM **\$21.76. (b) A: 12.00 × 40 = \$480**. B: 21.76 × 55 = **\$1,196.80. (c) Blended: sales = (40 × 16) + (55 × 34) = 640 + 1,870 = \$2,510; cost = 160 + 673.20 = \$833.20; 33.2%. (d) Replacing B with an A-like item at 55 units: 12.00 × 55 = \$660, against \$1,196.80 — a loss of **\$536.80 a week**, roughly \$27,900 a year, in exchange for a blended food cost of 25%. (e) You can starve to death with an excellent food cost percentage.
13.1 A par level is the quantity of an item you keep on hand to cover expected usage until the next delivery, plus a safety margin. An order quantity is par minus what you already have (and minus what is already on the way). The two variables that set a par: usage rate between deliveries and delivery frequency and reliability; shelf life caps it.
13.2 24 − 9 − 6 = 9 cases. Students who answer 15 have forgotten the open order — the most common real-world over-ordering error.
13.3 (a) Variance = 49,900 − 47,300 = \$2,600; ÷ 47,300 = 5.5% of theoretical; ÷ 162,000 = 1.6 points of food cost on sales. (b) Likely causes: over-portioning, waste and spoilage, unrecorded transfers or employee meals, receiving errors (short deliveries paid in full, price increases unnoticed), uncosted specials and off-menu items, POS recipe errors, and theft. (c) Test portioning first — it is the most common, the cheapest to check, and the fastest to confirm: weigh 20 plates of the three highest-volume items across two services against the spec.
13.4 Ranking: (1) portioning by new hires — the pattern points directly at it, since variance tracks new-hire hours and concentrates in proteins, which are the items where a half-ounce is worth money; (2) receiving errors on protein — price creep and short weights, which would also concentrate in proteins but would not correlate with new-hire hours; (3) theft — possible, but it is the explanation that fits the pattern worst and costs the most to investigate wrongly. Test for (1): a portion audit — scale checks on the four highest-volume protein items, 20 plates each, sorted by who plated them, over two services, compared against the spec and against the veteran cooks' plates.
13.5 Two problems: (1) Romaine at \$28.40 against a \$26.10 last price — an 8.8% increase accepted without challenge; on four cases that is \$9.20 today, but at that usage it is roughly \$480 a year, and nobody decided to spend it. (2) Tomatoes short-shipped — 2 cases received against 3 ordered; if the invoice bills 3, the restaurant pays \$31.00 for nothing, and the kitchen is a case short for Wednesday. The chicken is also up \$0.14/lb (4.4%), worth about \$16.50 on this delivery. Procedures: check every delivery against the purchase order and sign only for what is physically counted or weighed, with the driver present; and run a price-variance flag that compares each invoice unit price to the last one and requires a manager's initial on any increase above a set threshold.
13.6 (a) Affected usage = \$49,500 × 55% = \$27,225; × 4.1% = \$1,116 a month**, **\$13,395 a year. (b) The manager's figure: \$49,500 × 4.1% = \$2,030 a month, \$24,360 a year** — wrong because it applies a category-specific increase to the entire usage base, including categories whose prices did not move. Using the wrong base nearly doubles the number and will produce a menu price increase the market did not require. (c) \$1,116 a month ÷ 2,050 entrées = \$0.55 per entrée** — a \$27 item goes to \$27.55, so a **\$0.50 to \$1.00 increase** covers it. On the manager's incorrect figure it would have looked like \$0.99, or roughly a \$1.00–\$1.50 move.
14.1 The pass is the physical and organizational boundary where the kitchen hands a plate to the dining room — where tickets are called, plates are checked, and the table is assembled. The expediter sequences and times the board. Three decisions that belong to nobody else: when a ticket fires (and which tables fire together), whether a plate leaves the pass or gets re-fired, and when to hold, stagger, or re-sequence the board when a station falls behind.
14.2 1,984 ÷ 62 = 32.0 minutes average. No — an average conceals the distribution. A kitchen with a 32-minute average made up of consistent 30s is fine; the same average made up of 18s at 5:30 and 61s at 7:45 is a kitchen that failed for ninety minutes. Ask for ticket times by hour and the longest ticket, not the mean.
14.3 (a) 130 × 1.4 = 182 items; × 45% = 81.9 sauté items; ÷ 5 hours = 16.4 an hour. (b) Capacity is 14 — the station is 17% over on the flat average, and far worse at the peak. (c) Two changes without hiring: re-engineer the menu so a sauté item moves to the oven, the grill, or the cold station (the fastest lever is usually a single high-volume dish); and shift prep upstream so sauté fires components rather than building them — pre-portioned, pre-blanched, sauces held hot. A third acceptable answer: pace the book so the peak-hour load flattens.
14.4 Options and costs, as far as the data supports: (1) Call someone in at overtime — a cook at time-and-a-half for a six-hour shift, roughly \$180–\$220 fully burdened, and it works only if someone answers. (2) Run the station short — costs nothing in labor and costs everything in ticket times; with a 40-top at 6:30 the board will back up by 7:15, and slow tickets on 142 covers damage a Friday that the restaurant only gets once. (3) Cut the menu — pull the two most grill-dependent items, protecting timing at the cost of the check average and some guest disappointment; cheapest in dollars, real in brand terms. Most strong answers combine 1 and 3: call for coverage, and pre-emptively simplify the 40-top's menu to fixed courses fired together. Tomorrow at 4:00 p.m.: a written call-in list with a stated premium, and a standing rule for who calls whom by when.
14.5 The pinch is 7:00–8:00: 34 tickets at a 38-minute average with a 61-minute outlier, on 67 covers — nearly half the night's covers in one hour. What is most likely happening: the board is over-committed, one station (probably the one carrying the most items) has gone into the weeds, and the expediter is holding plates for tables that cannot be completed, which lengthens every ticket behind them. Reservation fix: cap covers per 15-minute slot so the hour lands nearer 45 covers than 67. Prep fix: move the most-ordered item's assembly upstream — par-cook, pre-portion, or hold a component hot — so the peak hour is assembly rather than cooking.
14.6 (a) Making it: 5 service days ÷ 3-day coverage = 2 batches a week; 2 × 4 hours × \$18 = \$144 labor + \$31 ingredients = **\$175 a week, \$9,100 a year**. (b) Buying it: 26 × \$4.20 = \$109.20 a week**, **\$5,678 a year. (c) Buying saves \$65.80 a week**, **\$3,422 a year. (d) Buying raises food cost percentage (the base is COGS) and lowers labor cost percentage (the four hours leave the schedule), and lowers prime cost by the \$3,422 — which is the only one of the three that should govern the decision. Add the two limits: the saving is real only if those labor hours actually leave the schedule rather than being absorbed, and quality must genuinely be comparable, since stock is a base flavor in many dishes.
15.1 Pour cost = beverage COGS ÷ beverage sales; the typical full-service range is 18–24%. A beverage-weighted mix pulls blended COGS down because beverage product cost is roughly a third lower than food's — every point of sales shifted from a 30%-cost plate to a 22%-cost drink adds eight cents of contribution per dollar.
15.2 (a) 25.36 ÷ 1.5 = 16.9 pours per bottle; \$28.00 ÷ 16.9 = **\$1.66 per pour. (b) 1.66 ÷ 12.00 = 13.8%. Note this is the theoretical pour cost; the actual will be worse for exactly the reasons in 15.3.
15.3 (a) 25.36 ÷ 1.75 = 14.5 pours; \$28.00 ÷ 14.5 = **\$1.93 per pour — 16% more cost per drink. (b) Lost drinks = 40 × (16.9 − 14.5) = about 96 drinks a week; at \$12 that is **about \$1,159 a week and roughly \$60,000 a year of retail value that was poured and not sold. (c) Cheapest control: jiggers, or measured pour spouts, plus a posted spec. Bartenders resist because free-pouring is faster at the rail and reads as craft and generosity — and because a heavy pour buys regulars.
15.4 A strong evaluation: the throughput claim is testable and is sometimes true at a crushed bar, but the arithmetic rarely favors it — 96 drinks a week of product loss is a large number to make up in speed, and speed is limited by the guest, the POS, and the well layout more than by the measure. Design a test: two weeks, one bar, alternating conditions — measure drinks per bartender-hour and average ticket time in each condition, and measure spirit variance from depletion versus POS over the same period. Hold the bartender roster constant across conditions and blind nothing (you cannot). Decide on the pair of numbers together, not on either alone.
15.5 Spirits, at 13.3% variance — \$1,320 in four weeks, roughly **\$17,160 a year. Three likely causes in order of testing: (1) over-pour (test with a jigger-and-measure audit and a re-count after two weeks); (2) unrecorded comps and staff drinks (test by reconciling the comp report and the shift-drink policy against the POS); (3) theft — unrung drinks or product leaving the building** (test with a spot inventory on the top five moving bottles, mid-shift, and a review of voids at the bar terminal). Test in that order because the first two are far more common and cost nothing to rule out.
15.6 (a) (0.30 × 24) + (0.30 × 32) + (0.40 × 16) = 7.2 + 9.6 + 6.4 = 23.2%. (b) (0.25 × 24) + (0.45 × 32) + (0.30 × 16) = 6.0 + 14.4 + 4.8 = 25.2%. (c) +2.0 points. (d) "Pour cost rose because the sales mix shifted toward wine, which carries the highest product cost of our three categories; no recipe, price, or supplier changed. The promotion moved 15 points of mix from spirits and beer into wine, and the blended number moved with it." The report that would have shown it coming: a beverage mix report by category, tracked alongside pour cost — the two must always be read together.
16.1 A flat multiplier applies one factor to every wholesale cost. Progressive markup applies a declining multiplier as cost rises. Fixed gross profit adds a set dollar margin regardless of cost. A straight 3× fails at the top because it produces prices guests will not pay — a \$90 wholesale bottle at 3× is \$270, so the bottle never sells and the restaurant earns nothing on it, whereas the same bottle at \$150 earns \$60 that would otherwise not exist.
16.2 \$14 × 3 = **\$42.00; 14 ÷ 42 = 33.3%** pour cost.
16.3 \$28 + \$32 = \$60.00; 28 ÷ 60 = 46.7% pour cost.
16.4 Pour cost is wholesale ÷ list; gross profit is list − wholesale.
| Wine | Wholesale | List | Pour cost | Gross profit |
|---|---|---|---|---|
| A | \$9.00 | \$36.00 | 25.0% | \$27.00 | |
| B | \$14.00 | \$48.00 | 29.2% | \$34.00 | |
| C | \$26.00 | \$74.00 | 35.1% | \$48.00 | |
| D | \$48.00 | \$112.00 | 42.9% | \$64.00 |
The pattern is progressive markup: pour cost percentage rises and gross profit dollars rise as you move up the list. A is most likely to sell (lowest price point, and the cheapest bottle on a list is the most-ordered position after the second-cheapest); D is most likely to sit — and that is the point of pricing it this way, since \$64 of margin on an occasional sale beats \$0 on a bottle priced at 3×.
16.5 Against: doubling by-the-glass selections roughly doubles open-bottle exposure, and spoilage is a function of how fast each open bottle turns; at 95 covers a night across five nights, sixteen selections means many bottles will not clear five glasses inside the freshness window. It also doubles the training load and the inventory dollars sitting on the back bar. Approve it if: the restaurant adopts preservation (argon, Coravin, or half-bottle formats) and the wine director commits to a BTG sales report by selection, cutting anything that does not turn a bottle within the window. A defensible middle path is 10–12 selections with two rotating.
16.6 (a) Sells all five: revenue 5 × \$16 = **\$80, cost \$26 → **32.5%**. Sells three then discarded: revenue 3 × \$16 = \$48**, cost \$26 → 54.2%. (b) Across ten bottles: revenue = (6 × 80) + (4 × 48) = 480 + 192 = \$672; cost = 10 × \$26 = \$260 → **38.7%**. (c) The card shows 32.5% because it assumes a full yield of five glasses from every bottle — it prices the recipe, not the operation. Two changes: **preservation systems** to extend the window (which moves the 40% toward zero) and **re-pricing or re-sizing** the slow selections — a 5-ounce pour at \$16 on a bottle that will not turn should either be a 6-ounce pour at \$18 or should not be on the list.
Part IV
17.1 Separation costs (final pay, unused leave where applicable, the exit itself); recruiting (advertising, screening, manager and interviewer time); training (trainee wages, trainer time, materials, certification); lost productivity (the ramp during which the new hire is slower and wastes more, plus the overtime and coverage that fills the gap); and guest impact (slower service, mis-fired tickets, the guest who does not return). The last one almost never appears in any report — which is precisely why it is the one that gets ignored when someone decides to save money on training.
17.2 A pipeline is the practice of recruiting continuously against expected turnover rather than reactively against a vacancy: a standing posting, a stage list of people you have already met, relationships with schools and other operators, and a referral program. "Post when someone quits" produces a worse hire because it collapses the candidate pool to whoever is available in the two weeks you are desperate, and it costs more because the gap is filled with overtime while you search.
17.3 17 ÷ 22 = 77% annual turnover — roughly the industry's rough benchmark of about 75%, which means it is normal and normal is expensive.
17.4 (a) \$310 + (6 × \$32 = \$192) + (24 × \$21 = \$504) + (60 × 40% = 24 unproductive hours × \$19 = \$456) + \$380 = \$1,842**. (b) 9 × \$1,842 = \$16,578**. (c) The **lost-productivity line** (\$456) draws the challenge, because it is an estimate and because the trainee was paid anyway. Defense: the restaurant is paying full wage for partial output — that gap is real cost whether or not anyone books it, and if you reject it you must also reject the idea that a trained employee is worth more than an untrained one, which no operator actually believes.
17.5 1 ÷ 26 = 3.8% — roughly 26 applications per retained 90-day hire. Two worst leaks: 8 scheduled → 5 attended (38% no-show, an enormous waste of manager time — fix with confirmation texts the day before, same-week scheduling, and an interview slot that respects the candidate's shift schedule) and 2 accepted → 1 at 90 days (a 50% early-attrition rate, which is an onboarding failure, not a hiring failure — fix with a structured first-thirty-days plan, a named trainer, and a check-in at day 7 and day 30). Accept 120 views → 26 applications as a third answer if the student argues the posting itself is the problem.
17.6 A strong response: prices both paths using 17.4 — Candidate B costs more in training and in coverage during four weeks of development (a real \$1,500–\$2,500 of the \$1,842 build), while Candidate A costs nothing to develop but carries the risk that the entire \$1,842 is spent twice if the reference is right. Notes that the reference's phrasing describes a behavior problem, and behavior problems in a kitchen cost more than skill gaps because they radiate — one corrosive cook can produce a second separation. Names the sous's current double coverage as a real, dated cost that argues for speed. Then adapts the offer to the choice: with A, a documented 90-day probationary period, explicit written expectations, and a second reference call before the offer; with B, a training plan with dated milestones and a scheduled wage step on completion so the development is a contract, not a hope.
18.1 A training matrix maps every position against every competency, with a status for each — trained, verified, signed. Operationally it tells a manager who can actually be scheduled on which station without supervision. Legally and in risk terms, a signed certification checklist is the documentation that the employee was trained on the things that matter — allergen handling, alcohol service, food safety, harassment policy — which is the record you need when something goes wrong and the record a regulator, an insurer, or an attorney will ask for.
18.2 32 × \$12.50 = **\$400.00**.
18.3 (a) Trainee wages \$400; trainer premium \$2.50 × 24 = \$60; manager orientation 4 × \$30 = \$120; materials \$45; burden 12.4% on (\$400 + \$60) = \$57.04. Total = **\$682.04. (b) The trainer premium** is the cheapest to cut in dollars — and cutting it costs you the willingness of your best people to train, which is worth far more than \$60. The honest answer is that the only cuttable line without consequence is materials, and it is \$45.
18.4 Costs created: more errors and re-fires (COGS, immediately); slower tickets and lower beverage attachment (revenue, within days); guest complaints and comps (weeks); reviews (two to six weeks, and they persist); and a materially higher chance the server quits inside 90 days, resetting the whole \$682 (one to three months). The shortcut is nonetheless correct when the alternative is not opening the section at all and turning guests away — a poorly trained server on the floor beats an empty section for one or two shifts. What makes it a mistake is doing it as policy rather than as a dated exception with make-up training scheduled.
18.5 Allergen protocol (trained but not verified or signed) and menu knowledge — beverage (same) are the stoppers; strong answers rank allergen first. The allergen exposure is a guest-safety and liability event, not a service quality issue: an untested server who says "I think that's fine" can send someone to a hospital, and the restaurant will be asked for the signed record. "Verified" for allergen protocol must mean an observed practical test — the server correctly identifies the allergens in five specific menu items, states the kitchen notification procedure, and describes what happens to a modified ticket at the pass — not that they read the sheet. The "table maintenance" row is signed without verification, which is a paperwork failure and should be corrected on principle.
18.6 (a) \$3.10 × 480 = **\$1,488 of incremental monthly sales; × 40% = \$595.20** of incremental monthly contribution. (b) \$682.04 ÷ \$595.20 = **about 1.15 months**. (c) The challenged assumption is **the \$3.10 attribution** — that the difference is caused by training rather than by the section, the shift, or the individual. Supporting evidence: compare the same server's per-cover sales before and after the training module (a within-person comparison), across comparable shifts and sections, and check that the lift persists past 60 days rather than fading.
19.1 SPLH = sales ÷ labor hours worked. It usefully reveals whether a shift was staffed proportionally to the volume it produced and how productivity compares across days, shifts, and units. It cannot tell you whether the guest was well served (a fast SPLH can be a room full of neglected tables) and it cannot tell you whether the mix of hours was right — the same SPLH results from too many servers and too few cooks as from a correct schedule, because it collapses everyone into one denominator.
19.2 Fixed labor does not vary with volume within the range of a normal week: salaried managers and chef, the opening prep cook, the closing dishwasher, the host at the door. Variable labor scales with covers: additional servers, additional line cooks, a second bartender, a busser. The distinction matters more at low volume because the fixed floor is a constant number of dollars — at 40 covers it is a catastrophic percentage of sales, and at 140 it is trivial. This is exactly why a slow daypart can lose money at a labor percentage no schedule change can fix.
19.3 (a) Hourly wages: cooks 2 × 8.5 × \$20 = \$340; prep \$102; dish \$97.50; servers 4 × 6 × \$10 = \$240; host \$75; bartender \$112 = \$966.50**. Plus salaried \$335 → unburdened \$1,301.50**; × 1.225 = burdened **\$1,594.34. (b) 1,594.34 ÷ 4,600 = 34.7%. (c) Hourly hours = 17 + 6 + 6.5 + 24 + 5 + 7 = 65.5; \$4,600 ÷ 65.5 = **\$70.23 SPLH.
19.4 SPLH: Tue \$2,320 ÷ 72.0 = **\$32.22; Wed \$38.38**; Thu **\$43.54; Fri \$53.33**; Sat **\$55.76. Tuesday is worst by a wide margin. Remove roughly 12 hours — one FOH shift and one BOH shift — which lifts Tuesday to about \$38.67 and still leaves it the weakest day. Note that BOH hours are 38.0 on Tuesday and also 38.0 on Wednesday for 13 more covers: that is the fixed floor, and it is where the hours actually are. Before removing them: check the last four Tuesdays' actual** covers against forecast; confirm none of those hours are weekly prep or deep-clean that will simply reappear on Wednesday; and confirm the cut does not push someone to a sixth day later in the week and create overtime.
19.5 Salaried sous: \$58,000, no overtime exposure, one point of accountability, and continuity. Two hourly leads: 2 × 38 × \$24 × 52 = \$94,848 before burden — far more expensive at those hours, but it buys coverage on both ends of the week, redundancy when one quits, and a development path that grows two people instead of one. A strong answer notes the salaried option is really "one person covering six days," which is how sous chefs burn out and leave, taking the whole investment. The classification question comes first: a salaried sous is exempt only if both the salary basis and the duties tests are met — a "sous chef" who spends most of their shifts on a station is likely non-exempt regardless of how they are paid, and the exempt option evaporates. This varies by jurisdiction; verify with counsel.
19.6 (a) Variable = \$1,550,000 × 19.88% = \$308,140; + \$191,895 = **\$500,035; ÷ 1,550,000 = 32.3%. (b) At \$1,300,000: variable \$258,440 + \$191,895 = **\$450,335; ÷ 1,300,000 = 34.6%. (c) Labor percentage rose 2.3 points because \$191,895 of the labor line does not move with sales — the same fixed dollars divided by a smaller denominator is a larger percentage. Nothing was mismanaged; the volume assumption was wrong. (d) An owner budgeting a flat 32.3% will be under-budgeted whenever sales fall short, which is exactly when cash is tightest — the error is largest in the ramp, in February, and in any week a forecast misses.
20.1 The tip credit lets an employer count a portion of an employee's tips toward the federal minimum wage obligation, paying a lower direct cash wage, provided the employee's cash wage plus tips reaches at least the applicable minimum for every workweek, the employee is notified in advance, and the employee retains their tips (subject to a valid pool). Two facts to check before scheduling: whether the state allows a tip credit at all — several do not, and require full minimum wage before tips — and what the state's own minimum wage and maximum tip credit are, since state law governs where it is more protective. Verify locally; this area also has active regulatory history around non-tipped side work.
20.2 Managers and supervisors may not participate in a tip pool, and neither may the employer — including through the entity, an owner working a shift, or a "house" share. The law is most concerned about the employer or its agents taking a share of employees' tips, in any form. The practical test a manager should apply: do I have authority to hire, fire, direct, or discipline the people in this pool? If yes, stay out of it — even when working a station, even when the staff offer. Whether back-of-house employees may be included depends on whether a tip credit is taken and on the jurisdiction; verify locally.
20.3 (a) Straight time 40 × \$21.00 = **\$840.00; overtime 7 hours at \$31.50 = **\$220.50; gross = **\$1,060.50**. (Equivalently: 47 × \$21 = \$987 plus a half-time premium of 7 × \$10.50 = \$73.50.) (b) A **non-discretionary bonus must be included in the regular rate** for the workweek it is attributable to, which raises the regular rate above \$21.00 and therefore raises the overtime premium owed. The employer cannot simply add \$100 to the check — it must be allocated and the overtime recomputed on the higher regular rate. This is one of the most common wage-and-hour errors in restaurants.
20.4 Required: 30 hours × \$10.50 = **\$315.00. Received: \$150 cash wages + \$120 tips = \$270.00**. The employer must pay an **additional \$45.00 in cash wages, because the tip credit claimed can never exceed the tips actually received — the employee must reach the applicable minimum wage in every workweek. Caveat: the applicable minimum, the maximum tip credit, and whether a tip credit exists at all are set by state and sometimes local law**; verify locally and use the more protective standard.
20.5 A strong analysis: exemption requires both a salary-basis test (a minimum weekly salary set by regulation, which \$780 may or may not satisfy depending on the current threshold and the state's own, often higher, threshold) and a duties test (primary duty is management, customarily directing two or more employees, with genuine authority over hiring and firing or meaningful input into it). Spending roughly half of each shift expediting, bussing, and running food is strong evidence the primary duty is not management — and the duties test is where restaurants lose. The risk is not a fine alone: it is back overtime for the lookback period, liquidated damages, attorney's fees, and the same exposure for every similarly titled employee. What to do instead: classify them non-exempt, control the hours with scheduling and a second key-holder, and if the manager genuinely runs a shift, restructure the job so the duties match the title.
20.6 (a) \$52,000 ÷ 52 weeks = \$1,000 a week; ÷ 52 hours = \$19.23 regular rate**. (b) Overtime premium is half-time on 12 hours: \$9.615 × 12 = \$115.38 a week**, **\$6,000 a year each, \$12,000 for both**. (c) With 22% burden: **\$14,640. (d) \$14,640 ÷ 0.40 = **\$36,600 of additional annual sales — roughly 100 additional covers a month at a \$30 check — simply to stand still. (e) Compliance decisions are break-even decisions. They are not administrative; they move the line the business must clear, and an operator who discovers this during an audit discovers it with back pay attached rather than with a plan.
21.1 Turnover costs are incurred per separation and are invisible — they appear scattered across recruiting, training, overtime, and waste. Retention costs are paid in advance and are highly visible — wages, schedules, benefits — which is why they lose every budget argument to costs nobody has counted. Three levers that cost nothing in wages: a schedule published far enough ahead to plan a life around (two weeks or more, and honored); a real path — a written progression from station to station with what each step pays; and a manager who runs a competent shift, since people leave chaos more reliably than they leave money.
21.2 Pre-shift is the short all-staff meeting before service. A good one accomplishes four things: the 86 list and the specials with tasting notes, the night's shape (covers on the book, large parties, VIPs, timing pressure), one teaching point, and one specific goal with a number attached. It is most commonly wasted by becoming an announcements meeting — a list of things people did wrong, delivered standing up, that nobody can act on during service.
21.3 22 × 78% = 17.2 separations; 22 × 55% = 12.1; a reduction of about 5 separations; × \$1,842 = **\$9,210 a year**.
21.4 (a) 14 × 28 = 392 hours a week; × \$1.25 = \$490; × 1.22 burden = \$597.80 a week**, **\$31,086 a year. (b) Against \$9,210 of turnover saving, the raise costs **about \$21,900 more than it returns on that measure alone. (c) Three other places the return appears: food cost (trained, tenured cooks portion correctly and waste less — one point of food cost here is worth roughly \$12,000–\$15,000); labor productivity (a tenured crew runs the same covers on fewer hours, and SPLH is the measure); and revenue** (tenured servers sell more per cover and produce more repeat visits — measure per-cover sales by tenure band and repeat rate). Measure the first one by tracking ideal-versus-actual variance monthly against average crew tenure, which is a number the restaurant already has.
21.5 Delegate: invoice entry and coding, the Wednesday order, and the vendor call — all three are rule-based, verifiable, and are how you develop an assistant manager. Do not delegate: writing the schedule against the forecast. It is the single largest controllable cost decision of the week, it requires judgment about who can carry which section, and delegating it is how a labor model quietly becomes a habit. (Strong answers may argue instead for the walk-in count on the same logic; accept it with reasoning, but note the count can be delegated once it can be audited, while the schedule cannot.) Out of place: the interview at 13:00 sits in the middle of ordering and reconciliation; move candidate interviews to a block that does not compete with a hard vendor cutoff — and note the order should be placed before the vendor call, not after.
21.6 A strong argument: quantify what is known — the host who quit is one full turnover event (\$1,842 on the 17.4 build, more for a trained position), two complaining cooks represent a measurable flight risk, and a sous who will not work Saturdays is a scheduling constraint with an overtime cost attached. Acknowledge what cannot be quantified: the covers the team does not run well, the candidates who hear about it, and the standard everyone else concludes is optional. Then state the sequence with dates: (1) a documented conversation this week naming specific behaviors, the standard, and the consequence; (2) a written expectation with a 30-day review, signed; (3) schedule changes now to protect the people who are staying; (4) at 30 days, either the behavior has changed or the employment ends. Note the general rule: the cost of keeping a corrosive high performer is paid by everyone else, in a currency the sales report does not show.
Part V
22.1 Seat-to-order (greet, drinks, decision), order-to-entrée (kitchen time and course pacing), entrée-to-check (dessert, coffee, the ask), and check-to-reset (payment, clear, wipe, re-set). The one a manager can compress with the least damage is check-to-reset — the guest has already decided to leave, and every minute of a dirty table is pure loss. Compressing order-to-entrée damages the meal; compressing entrée-to-check is the one guests actually notice and resent.
22.2 95 + 8 = 103 minutes; 240 ÷ 103 = 2.33 turns per table.
22.3 (a) 240 ÷ 103 = 2.33 before; 240 ÷ 92 = 2.61 after. (b) Difference 0.28 turns × 18 tables × 4 seats = about 20 covers. (c) × \$38 = **about \$762 a night**. (d) The gain is commonly given back by (1) the kitchen, which cannot absorb 20 more covers in the same window and lengthens ticket times until dwell time returns to 103 minutes, and (2) the floor, where the same server count now runs more tables, service degrades, and the turn slows for a worse reason.
22.4 Either answer defends, but the reasoning must be explicit. Holding: the guest's lifetime value is real and identifiable, and the cost is one table for 75 minutes. Declining: the table produces roughly a full turn in that window — at a \$38 check on a four-top, about \$150 of revenue and \$60 of contribution given away, plus a wait-list guest who now waits longer. As policy, the honest framing is: if every server could grant this, the book would be unmanageable and the wait list would become a fiction. So the correct structure is a named, limited authority — the manager may hold one table per service for a defined window, logged — not an ad hoc favor. Strong answers say what happens at 8:15 if the guest has not arrived.
22.5 Hour by hour: 5:30–6:30 the room is nearly empty and the labor is already on the clock; at 6:30 eighteen covers arrive; at 7:00 forty-six covers arrive at once, including three 6-tops and an 8-top, which is a single fire of enormous size; at 7:30 another 34 land on top of a kitchen already behind. Three failures that follow: ticket times blow out (large parties fire together and monopolize stations), the floor cannot greet, order, and run for 46 simultaneous covers, so first-drink times stretch and tables sit un-greeted, and the 8:00 and 8:30 guests inherit a broken room — the failure propagates forward and the last covers of the night get the worst service. Pacing rule: cap covers per 15-minute slot — roughly 7 to 8 here — and place large parties in slots of their own, at the shoulders rather than at the peak.
22.6 (a) 30 × 5 = 150 entrées of capacity against 136 covers — the night fits, with 9% of headroom. (b) The 7:00–8:00 window as booked demands 46 + 34 = 80 entrées in one hour, against 30 of capacity — 167% over. (c) 30 an hour ÷ 4 slots = 7.5 covers per 15-minute slot, so a rule of 7 to 8, with large parties counted against the slot they occupy. (d) The owner's calculation treats capacity as a stock that accumulates — as if the unused 5:30 hour banks entrées for 7:00. It does not. Kitchen capacity is a rate, and unused capacity in an empty hour is gone forever, exactly like an empty seat-hour. The night's total is irrelevant if no 60-minute window is inside the rate.
23.1 Guest lifetime value is the total contribution a guest produces over the period they remain a guest — visits per year × contribution per visit × years retained. Frequency belongs ahead of reach because a repeat visit costs essentially nothing to acquire and contributes the same margin as an expensive first visit, so a program that adds one visit a year to existing guests almost always beats one that adds first visits from strangers.
23.2 \$38 × 40% = **\$15.20 per visit; 8 × 3 = 24 visits; 24 × \$15.20 = **\$364.80 of lifetime contribution.
23.3 (a) \$364.80 ÷ \$46 = 7.9×. (b) Two conditions: the guest must actually return (the recovery has to work, which depends far more on the response than on the comp), and the guest must have been a repeat guest — the same arithmetic applied to a one-time visitor from out of town returns nothing. (c) Because the arithmetic prices a saved relationship, not a free meal. Comping every complaint teaches the floor that the comp is the recovery, trains a subset of guests to complain, and — worst — removes the pressure to fix the cause. The comp is the last step in a recovery, not the recovery.
23.4 Public response: short, named, non-defensive, no litigation of the facts — thank them, acknowledge the wait specifically because it is true, state what is being done about it, invite them back directly, and give a way to reach a manager offline. Do not correct the burger price in public and do not contradict them about the server; you cannot win either exchange and the burger correction makes you look like the kind of place that counts. Internal actions: (1) pull the ticket and the timing data for that Saturday at 7:15 and see whether the 45 minutes was systemic or singular; (2) talk to the server, get their account, and treat it as coaching rather than discipline unless a pattern exists. What the public response is for: it is not for the reviewer, who has already decided. It is for the hundreds of prospective guests who will read the exchange, and it is evidence about how this restaurant behaves when something goes wrong.
23.5 Fix first: wait time (47 mentions, and it is the theme that converts a 5-star experience into a 3-star review). Attack it through pacing and the reservation book, per 22.5–22.6, not by rushing tables. Deliberately do not fix: noise (31 mentions) — a loud room is a design and energy characteristic that a substantial share of your five-star reviewers are enjoying, and softening it costs real money to please the guests least likely to return. State it as a positioning choice, not neglect. What it cannot tell you: reviewers are a self-selected extreme — 246 reviews against tens of thousands of covers. The silent majority's experience is unmeasured here, and, critically, the guests who quietly stopped coming leave no review at all. Repeat rate and guest-database frequency measure them; this table does not.
23.6 (a) A: 300 × \$15.20 = \$4,560 first visits; returns 300 × 22% = 66 × \$15.20 = \$1,003.20; total \$5,563.20 − \$4,800 = net \$763.20**. **B:** 900 × 0.6 additional visits = 540 visits × \$15.20 = \$8,208 − \$1,200 = net \$7,008**. (b) **B wins by \$6,244.80, on a quarter of the spend. (c) The objection confuses new guests with growth. Growth is contribution, and B produces nine times more of it; moreover B compounds, because a database is an asset that can be used again next quarter at almost no cost, while a reach campaign must be re-bought. The owner is right when there is no database to work with** — a new restaurant, or one that has never captured guest contact information — in which case A's real product is not the 300 visits but the 300 records, and it should be evaluated that way.
24.1 RevPASH is revenue ÷ (available seats × hours open) — revenue per available seat-hour. It measures what neither of the others can: how much of your perishable inventory you actually sold. Average check ignores empty seats; covers ignore price and ignore how long each guest occupied the seat. RevPASH prices the seat-hour, which is the thing you cannot store.
24.2 Turns count bodies, not dollars, and they ignore how long the room was open. A counter concept doing 2.4 turns at a \$22 check produces \$52.80 per seat; a dinner house doing 1.3 turns at a \$52 check produces \$67.60 per seat — fewer turns, more money. Add a second failure: a restaurant open eleven hours and one open four hours can post identical turns while one of them has sold a fraction of its seat-hours.
24.3 (a) \$4,600 ÷ (68 × 5) = \$4,600 ÷ 340 = \$13.53 RevPASH**. (b) \$4,600 ÷ 100 = \$46.00 average check. (c) Put RevPASH by hour** on the manager's daily report. Average check is a menu-and-training number that moves slowly; RevPASH by hour is the one that tells the manager which hours were sold and which were given away, which is the decision they can act on tomorrow.
24.4 Total revenue \$4,223 ÷ 340 seat-hours = **\$12.42 average RevPASH. The two hours worth attacking are 5:00–6:00 (\$3.10)** and **9:00–10:00 (\$8.20) — and they are different problems. The 5:00 hour is a demand problem: the seats exist and nobody wants them, so the tools are early-seating offers, a bar menu, a reason to come at five, or shortening the service. The 9:00 hour is usually a conversion and pacing** problem: guests are still in the building, so the tools are dessert and after-dinner beverage attachment, a late bar program, and not turning the lights up at 9:15. Cutting labor helps in both, but only the second one can be fixed by selling more to people already there.
24.5 A strong argument, either way, must address four things. Who takes it: early-seating discounts are disproportionately taken by guests who would have come anyway at a nearby time, and by the most price-sensitive segment — which is not the segment a \$46-check restaurant is building around. **Positioning:** a standing percentage discount teaches the market your real price, and it is very hard to withdraw. **The 7:00 book:** every guest who shifts from 7:00 to 5:00 empties a seat you could have sold at full price *only if* there is unmet demand at 7:00 to fill it — if there is not, the shift is pure discount. **A fixed-price early menu** differs because it changes the **product**, not the price: three courses at \$38 with a shorter menu costs the kitchen less, cannot be compared line-by-line to the full menu, and can be withdrawn without a visible price increase.
24.6 (a) \$46 × 0.70 = **\$32.20; × 16 = \$515.20**; × 40.53% = **\$208.81 of contribution. (b) Displaced: 6 × \$46 × 40.53% = **\$111.86. (c) Net = \$96.95 — barely half of what the gross figure suggested. (d) The rule: discounting into a period only pays to the extent the discounted guests are incremental. Every guest who would have paid full price and instead pays the discount is a pure transfer out of margin. Note the limit in the other direction: if the 7:00 seats those six guests vacated are re-sold at full price, the displacement reverses — so the rule really is that the answer depends on whether the peak is capacity-constrained, and you must know that before you print the offer.
25.1 The temperature danger zone is 41°F to 135°F. Cold holding at or below 41°F; hot holding at or above 135°F; poultry cooked to 165°F (as an internal temperature, held for the required time). The sentence that must accompany all four: these figures follow the FDA Food Code, which states and localities adopt in different versions and with local amendments — verify the requirements of your own jurisdiction.
25.2 Quaternary ammonium sanitizer is typically used at 200–400 ppm, per the manufacturer's label and the local code. Beyond the test strip a manager must verify water temperature (quats are temperature-sensitive; too cold and the concentration is irrelevant) and contact time (the surface must stay wet for the label's stated dwell), plus that the strips themselves are for the right chemical and are not expired, and that the bucket has been changed — soiled solution loses efficacy long before the strip stops reading.
25.3 Two-stage cooling: 135°F to 70°F within 2 hours → by 6:15 p.m.; then to 41°F within 6 hours total → by 10:15 p.m. If it misses the first stage it may be reheated to 165°F and the clock restarted once, or discarded. Verify locally; some jurisdictions vary.
25.4 (a) 40 × \$3.20 = \$128 product + \$185 re-inspection + (6 × \$32 = \$192) manager time = **\$505. (b) Not captured: the menu items 86'd that night and the sales they would have produced; the re-inspection risk — a second failure can close the doors, and the cost of a closure is not a fee, it is a week; and the reputational and record cost**, since inspection results are public in most jurisdictions and follow the restaurant. Strong answers add the staff time lost to the correction and the possibility that the cooling failure indicates a walk-in problem with a capital cost attached.
25.5 By risk to a guest: (2) raw chicken over ready-to-eat greens first — direct cross-contamination of food that will never be cooked; (1) walk-in at 46°F second — temperature abuse across everything in the box, though it takes time to become dangerous; (3) sanitizer at 50 ppm third — an ineffective control across every surface; (4) no thermometer in the reach-in fourth — it removes your ability to detect the first two; (5) the missing food-handler card last in guest risk, though it is a compliance violation and a citation. Correct before the inspector leaves: re-shelve the chicken. It takes ninety seconds, and it is the highest-risk finding in the room. The systems failure is the missing thermometer — the others are incidents; a box with no thermometer means nobody could have known about the 46°F, which means it will happen again.
25.6 A strong decision: discard all time/temperature-control-for-safety foods whose safety cannot be established — the cooked braises from two days ago, all prepped produce and dairy, and any raw protein that cannot be verified — because with the last log at 6:30 p.m. the previous day, the box may have been out of temperature for sixteen hours and there is no basis to claim otherwise. Product that can be verified by internal temperature at 41°F or below and is still within its date may be moved to a working cooler and used; product that will be cooked to 165°F is not automatically safe, because toxins from bacterial growth are not destroyed by cooking. The log gap is the real finding: reconstruct nothing, document the gap honestly, and record the discard. Tell the staff plainly what happened, what was discarded and why, and that nobody is in trouble for reporting a temperature — the moment a cook believes a bad reading is a problem for them, the logs become fiction. Change: twice-daily logging with a manager's initial, a maximum/minimum recording thermometer or a cheap networked temperature alarm with a phone alert, and a written response protocol taped to the box.
26.1 Interchange — set by the card networks and paid to the issuing bank; the largest component and not negotiable. Assessments — the network's own fee; not negotiable. Processor markup — the processor's margin, quoted as basis points plus a per-transaction fee; this is the negotiable part, and it is often the smallest. Operators are usually quoted the markup, or a blended teaser rate, as if it were the whole cost, which is why the statement's effective rate always exceeds the quote.
26.2 With authority: what was rung (items, prices, modifiers), when it was rung (timing by hour, by server, by station), and how it was paid and by whom it was voided or comped. Not with authority: what was actually used (the POS computes theoretical usage from recipes, which is a model, not a count); what the guest experienced (a fast ticket time and a bad night look identical); and why an item sells — the POS records the order, never the reason, and menu decisions made from POS data alone routinely misattribute a position on the page to a preference.
26.3 \$1,550,000 × 2.81% = **\$43,555**.
26.4 (a) Monthly: (3 × \$189 = \$567) + 85 + 129 + 249 + 79 + 149 + 85 + 220 + 140 = \$1,703**; × 12 = **\$20,436. (b) + \$43,555 = **\$63,991, or 4.13% of \$1,550,000. (c) The subscription stack sits in other operating (technology); processing** is also an operating cost, though many operators mistakenly net it against sales — it must appear as a cost so it can be managed. Neither belongs in COGS, and neither belongs in G&A; putting them there breaks every benchmark comparison.
26.5 \$3,502 ÷ \$118,400 = 2.958% effective. The gap between "2.4% and ten cents" and 2.96% is: the quote described the processor's blended rate on qualified transactions, while the statement includes interchange downgrades (keyed and card-not-present transactions cost more), assessments, monthly and PCI fees, and chargebacks. The only line a better negotiation actually moves is the **processor markup (\$497)** — interchange and assessments are pass-through. Strong answers add that reducing keyed transactions and chargebacks is worth more than the negotiation, and that on this volume even a full 20% cut in the markup saves under \$100 a month.
26.6 Free hardware is worth what the equivalent terminals cost outright — typically \$3,000–\$8,000 for three terminals and a kitchen display, so call it \$5,000 amortized over three years, about \$139 a month. What the operator is actually buying is a three-year processing rate they cannot verify, on card volume of over a million dollars a year — where a 25-basis-point difference is \$2,500 to \$3,000 a year, dwarfing the hardware. Terms to insist on: interchange-plus pricing with the markup stated in basis points and cents, a cap or no early-termination fee, the right to a monthly statement in a readable format, portability or a buyout number for the hardware, and no automatic renewal. The one question that settles it: "Will you put the markup in basis points, in writing, on interchange-plus, for the full term?" A vendor who will not is selling the ambiguity, not the hardware.
Part VI
27.1 Reach is how many different people encounter you; frequency is how often the people who already know you come back. Frequency is cheaper because a repeat guest costs nothing to acquire — you already paid — and delivers the same contribution as a first-time guest who cost \$15 to find.
27.2 \$3,600 ÷ 240 = **\$15.00 per acquired guest**.
27.3 (a) 240 × (\$38 × 40% = \$15.20) = \$3,648**. (b) 240 × 25% = 60 guests × 2 visits = 120 visits × \$15.20 = \$1,824**. (c) \$3,648 + \$1,824 = \$5,472 − \$3,600 = **net \$1,872. (d) The return rate.** At 0% return the campaign loses \$0 net on the first visit alone (\$3,648 against \$3,600 — a rounding error, not a business), so every dollar of actual return comes from guests who come back. The campaign is not a customer-acquisition program; it is a bet on the restaurant's ability to convert a first visit.
27.4 Argue B. The billboard buys reach with no measurement, no addressability, and nothing retained when the three months end; "everybody will see it" is precisely the property that makes it unaccountable. The database is an asset — it can be used again next quarter at near-zero cost, it targets people who have already chosen you once, and every campaign against it is measurable to the cover. What B requires that A does not: a way to capture contact information at the table or the reservation, a consent and privacy practice that is honest and lawful in your jurisdiction, and someone who will actually run the program every month. B fails when nobody owns it.
27.5 Conversion: 3,180 ÷ 412,000 = 0.77% impressions to clicks; 742 ÷ 3,180 = 23.3% clicks to reservation page; 119 ÷ 742 = 16.0% page views to bookings; 96 ÷ 119 = 80.7% booked to seated. Cost per seated reservation = \$3,600 ÷ 96 = **\$37.50. The actual problem is the reservation page: 742 people wanted a table and 623 did not book.** Test that first — availability at the times people want, the number of clicks to complete, whether a deposit or card hold is required, and whether the page works on a phone. Note also the 19% no-show/cancel rate between booked and seated, which is a confirmation problem. The impression-to-click rate is the least interesting number here and is the one the vendor will want to discuss.
27.6 (a) 310 × \$20 = **\$6,200. (b) Incremental visits 310 × 40% = 124; contribution per visit \$60 × 40% = \$24; 124 × \$24 = **\$2,976. (c) Net = 2,976 − 6,200 = **−\$3,224**. (d) The 310 measures **redemptions**, which is the count of people who accepted free money. It does not measure incrementality — it cannot distinguish a guest the offer created from a regular who would have come Friday anyway and simply paid \$20 less. Redemption is the one number a discount promotion is guaranteed to produce, which is why it is the number promotions are always reported on.
28.1 Channel contribution per order is what an order through a given channel leaves behind after every cost that order causes — commission, product, packaging, and the labor the channel requires. The commission covers the platform's demand generation, its app, its payment processing, and its courier network. It never covers your labor — the person who packs the bag, the line that fires the food, the manager who deals with the missing-item complaint — and it never covers your packaging.
28.2 \$42.00 × 28% = **\$11.76 commission; net remitted before other costs = \$30.24**.
28.3 Food cost \$42.00 × 30% = \$12.60; \$30.24 − \$12.60 = \$17.64.
28.4 Arithmetic confirms: 8,988 − 2,516.64 − 430 − 268 − 310 = \$5,463.36**; ÷ 214 = **\$25.53 per order on a \$42.00 order — the restaurant keeps 60.8 cents on the dollar before touching food, packaging, or labor. The two deductions that were **choices, not terms**: **promotional fee participation (\$430) and in-app marketing (\$310)** — \$740 in four weeks, roughly \$9,600 a year, spent to be more visible inside a channel that already takes 28%. What to do: switch both off for one full period and measure order volume and, more importantly, **contribution**, not orders. Strong answers note the customer adjustments line (\$268) is neither a term nor a choice but a control problem — find out whether it is a packaging failure, a missing-item failure, or fraud, because each has a different fix.
28.5 Either answer defends. For turning it off: the peak is capacity-constrained, and every delivery order fired between 7:00 and 8:30 consumes station capacity that a full-margin dine-in cover would have bought (see 28.6, where the net is negative). Against: the channel's algorithm typically rewards consistent availability and acceptance rates and penalizes intermittent operators with lower placement — turning off the busiest 90 minutes of the week can cost visibility across all the hours you do want orders. A strong answer proposes the middle path most operators land on: stay on but throttle — lengthen quoted prep times during the window, trim the delivery menu to items that do not route to the binding station, and cap concurrent orders — then measure order volume in the off-peak hours to see whether the algorithm actually punished you.
28.6 (a) Packer cost: 4 × \$17 = \$68 × 1.22 = \$82.96 ÷ 22 orders = **\$3.77 per order. (b) \$42.00 − \$11.76 − \$12.60 − \$1.35 − \$3.77 = **\$12.52 of channel contribution. (c) Dine-in cover: \$46 × 40.53% = **\$18.64. (d) Displaced: 0.8 × \$18.64 = \$14.92; net = 12.52 − 14.92 = −\$2.40 per order**. (e) The rule: **a channel's contribution must be judged against what the capacity it consumes would otherwise have earned, not against zero.** At 4:30 p.m. the same order is excellent — the kitchen is idle, the packer's hours are already sunk, nothing is displaced, and \$12.52 lands against capacity that would otherwise have produced nothing at all. Delivery is not a good or bad channel; it is a good channel in the hours you cannot fill and a bad one in the hours you can.
29.1 A banquet event order is the single controlling document for an event — what the restaurant will deliver and what the client will pay. It must specify at minimum: date, times (setup, service, and end), and space; guaranteed guest count and the final-count deadline; the menu and beverage arrangement with prices; all charges — food and beverage minimum, service charge, rentals, labor, and tax; deposit, payment schedule, and method; and cancellation and force-majeure terms. The final count and its deadline cause the most disputes, because the client's expectation of "about sixty" and the kitchen's obligation to produce are settled by that one line.
29.2 F&B = 40 × \$65 = **\$2,600; service charge 20% = \$520**; tax 7% on \$3,120 = \$218.40**; total **\$3,338.40. Caveat: whether a service charge is itself taxable varies by state, as does whether a mandatory service charge is a taxable sale, wages, or a gratuity for wage-and-hour purposes — a mandatory service charge is generally not a tip, and how it is distributed has payroll consequences. Verify locally.
29.3 (a) Product \$2,600 × 26% = \$676; labor 5 × 5 = 25 hours × \$17 = \$425 × 1.22 = \$518.50; rentals \$240. Contribution = 2,600 − 676 − 518.50 − 240 = \$1,165.50, or 44.8% of F&B revenue. (b) That is materially better than a normal dinner service, because an event has a known count, a fixed menu, and a fixed end time — you buy exactly what you need, waste almost nothing, and staff to a number rather than to a forecast. (c) The cost most often omitted is the displaced dining-room business — the covers the space would have produced anyway — followed by the manager and chef time spent selling and planning the event, which is real and is never on the BEO.
29.4 A strong pricing method: compute the displacement floor first — normal covers × average check × contribution margin ratio for that night — then add the incremental costs the event itself causes (extra labor, rentals, security, cleaning), then add the margin you want. What you refuse to discount is the floor; everything above it is negotiable, and nothing below it is. Two non-financial reasons to decline a clearing offer: the event would displace your regulars on a night they rely on (a Saturday buyout is a message to everyone who was turned away), and the event carries operational or reputational risk you would not choose — a beverage-heavy party you cannot control, a client who wants to bring outside food, or a group whose behavior you would have to police.
29.5 Blank: final count due and cancellation. The final-count line must state a deadline (typically 72 hours before, in writing) and that the count may increase but not decrease after it. The cancellation line must state a sliding schedule — deposit non-refundable after a date, a percentage of the minimum owed inside a window — and a force-majeure carve-out. The material risk as written is "guaranteed 55, expected 65" with no final-count deadline: the kitchen must decide what to produce, and if it produces for 65 and 40 attend, you bill the guarantee — 55 × \$72 = **\$3,960 — but you have produced ten covers of food you cannot sell, roughly \$187 of product at 26%, plus over-staffing for the night. **Exposure if 55 are guaranteed and 40 attend: you bill \$3,960 and are protected on revenue; the loss is the over-production and the over-staffing, which is exactly what the missing deadline was supposed to prevent.
29.6 (a) Displaced contribution: 132 × \$46 = \$6,072 × 40.53% = \$2,461**. That is the floor. (b) Buyout: \$7,650 − (\$7,650 × 26% = \$1,989) − \$900 = **\$4,761. (c) Margin over the floor = \$2,300**. **Take it.** (d) The instinct compares **revenue to revenue** — \$7,650 against \$6,072 — which happens to reach the right conclusion here and would reach the wrong one easily. It ignores that the two revenue lines carry different cost structures: the buyout's product cost is 26% and its labor is incremental and known, while the dining room's contribution is already net of variable labor. Compare contribution to contribution**, always; the fact that the shortcut works on this particular set of numbers is luck, not method.
30.1 Capital efficiency is revenue and, more importantly, contribution produced per dollar invested; scale is absolute size. A first-time operator should optimize for capital efficiency, because the binding constraint is not how large the business could become but how much money they can lose before they are out of the game — and because a smaller invested base makes the personal guarantee survivable.
30.2 A commissary requirement obliges a mobile operator to base out of a licensed commercial kitchen for prep, water, waste disposal, and overnight parking in most jurisdictions — so the truck pays rent, it just pays it somewhere else. Three fixed costs commonly omitted from a truck plan: commissary fees, permits and insurance across every jurisdiction the truck operates in (each city or county may require its own), and maintenance, generator fuel, and depreciation on a vehicle that is also a kitchen — plus event and location fees, which are effectively rent charged as a percentage.
30.3 Truck: \$420,000 ÷ \$185,000 = \$2.27** of revenue per dollar invested. Restaurant: \$1,550,000 ÷ \$620,000 = **\$2.50. Note for discussion that the restaurant is slightly more capital-efficient on this measure, which surprises students — the truck's advantage is the absolute size of the loss if it fails, not the ratio.
30.4 (a) COGS \$126,000; labor \$109,200; commissary \$14,400; fuel and maintenance \$18,000; permits and insurance \$11,500; event fees \$33,600; other \$21,000. Total = **\$333,700. Operating profit = 420,000 − 333,700 = \$86,300, or 20.5%. (b) The percentage is high because the truck carries almost no occupancy in the conventional sense and a far smaller fixed overhead — no dining room, no host, no bussers, minimal utilities. It is compensating the operator for the labor they personally supply (most trucks are owner-operated, and the owner's own hours are not in that labor line), for revenue volatility (weather, events, seasons), and for the depreciation of a vehicle** that will need replacing on a much shorter cycle than a dining room. Adjust for a market wage for the owner and the margin looks ordinary.
30.5 Costs: 8,736 + 10,416 + 8,900 + 3,200 + 1,296 + 940 = \$33,488** against \$33,600 of gross order value → a monthly result of \$112, which is break-even and not a business. The structural problem: every order arrives through an intermediary, so the operator owns no guest relationship** — no email, no reservation, no recognition at the door, no way to lift frequency directly, and no ability to raise price without losing placement. A dining room's second visit is free; here every visit is re-purchased at 26%.
30.6 A strong counsel: (1) What the truck's profit predicts — demand for the food, the operator's cooking and cost discipline, and their tolerance for the work. What it does not predict — running a dining room, a floor staff, a lease, a liquor license, a P&L with an 8% occupancy line, and a fixed cost base that arrives whether or not it rains. (2) Transferable: menu, purchasing, speed, food cost instinct, brand. Not transferable: service management, scheduling to a forecast, cash forecasting against fixed obligations, and hiring at scale. (3) Capital and risk: \$620,000 with a personal guarantee converts a business whose worst case is a repossessed truck into one whose worst case is the arithmetic in 5.6 — an entirely different bet on the same person's life. (4) The intermediate step: a fixed location with a small fixed cost — a stall, a kiosk, a residency inside another business, or a second truck — which tests fixed-cost discipline and management-through-others at a fraction of the exposure. Add: hire and keep a general manager on the truck operation for six months first, because that is the skill the restaurant will actually require.
Part VII
31.1 A comp is a sale that was rung, delivered, and then given away — the guest received the product and did not pay. A void is a sale that was rung and then removed before delivery — a keying error, a changed order. A comp belongs in a contra-revenue or marketing account (the product cost stays in COGS, where it belongs, because the food was consumed); a void should never reach revenue at all. A rising comp rate points at service failures, over-generous discretion, or theft dressed as generosity; a rising void rate points at order-entry problems, training, or a terminal being used to remove sales after payment — different problems, which is exactly why they must never be reported as one number.
31.2 A flash report is a one-page weekly estimate of the numbers that move: sales, food and beverage usage from counts and purchases, labor from the payroll system, and prime cost in dollars and percent, against budget and against last week. Run monthly and delivered three weeks after close, it tells you about a problem seven weeks after it started — by which time the cause has become a habit and the cost is seven weeks of it. Weekly, you find it in eight days.
31.3 (a) 18,400 + 7,900 − 17,650 = \$8,650**. (b) 8,650 + 2,180 = **\$10,830. (c) 10,830 + 9,640 = \$20,470; ÷ 29,800 = 68.7%. (d) At 68.7%, the action is not analysis — it is the schedule for the week that starts tomorrow, because labor is the half of prime cost that can be changed inside seven days. Then count again next week: one week's number is a reading, not a trend.
31.4 In dollars: sales fell from \$29,400 to \$26,900, down \$2,500 or 8.5%**, while prime cost dollars rose only \$320 — from \$17,700 to \$18,020. So most of the 6.8-point move in the percentage is the denominator falling, not costs exploding. Within that, labor is the line actually failing: it rose \$200 (from \$9,410 to \$9,610) while volume fell 8.5%, which means the schedule did not flex at all — labor went from 32.0% to 35.7% of sales. COGS rose \$120 on lower sales, a real but much smaller deterioration (28.2% to 31.3%), and part of that is the fixed component of waste and spoilage spreading over fewer covers. First action: flex the schedule to the forecast for the coming week**, and simultaneously find out why sales fell 8.5% in four weeks — because if the trend continues, no schedule saves it.
31.5 A strong response: smallwares, cleaning chemicals, and to-go containers are operating supplies, not cost of goods sold. Cost of goods sold is the product you sold to the guest — food and beverage. Putting supplies there inflates COGS, inflates prime cost, and destroys every comparison the operator makes: the 63% prime is being read against a ≤60% benchmark that was computed on a different definition, so the operator either panics about a problem that is smaller than reported or, worse, cuts food quality to fix an accounting artifact. It also makes the flash report useless, because weekly usage from counts will never reconcile to a COGS line that contains chemicals. Correct treatment: supplies to other operating; to-go packaging to other operating (or to a delivery/packaging account so channel contribution can be computed per 28.6). Restate by reclassifying the prior quarters' amounts out of COGS and into other operating and re-presenting both quarters on the corrected basis — do not silently change the definition mid-year, because a chart of accounts that changes without a restatement destroys every trend line it touches.
31.6 (a) \$430,280 ÷ \$1,550,000 = 27.760%; \$500,000 ÷ \$1,550,000 = 32.258%. (b) \$930,280 ÷ \$1,550,000 = 60.018% — call it 60.0%. (c) The student added two rounded displays. 27.760 and 32.258 sum to 60.018; rounded independently to 27.8 and 32.3 they sum to 60.1, because each was rounded up. There is no error in the plan. The rule: dollars are canonical, percentages are rounded displays — compute from the dollars and round last. (d) Reconstructing COGS from 27.8%: \$1,550,000 × 0.278 = **\$430,900, against the true \$430,280 — an invented **\$620. Small here; on a percentage displayed to one decimal on a larger base, the same error runs into thousands, and it always appears as a variance nobody can explain.
32.1 Contribution margin ratio is the share of each sales dollar left after variable costs — (sales − variable costs) ÷ sales — and it is what pays fixed costs and then profit. Break-even sales = total fixed costs ÷ contribution margin ratio. The one classification error that ruins the calculation: putting variable costs in the fixed bucket (most often hourly labor), which double-counts them — once in the fixed total and again inside the contribution margin ratio — and grossly overstates break-even.
32.2 Margin of safety = actual or planned sales − break-even sales, in dollars; as a percentage, that difference ÷ actual or planned sales. It measures how far revenue can fall before the business stops covering its costs — that is, how much bad news the cost structure can absorb.
32.3 (a) \$340,000 ÷ 0.40 = **\$850,000. (b) \$850,000 ÷ \$38 = 22,368 covers a year. (c) ÷ 260 services = 86 covers a service. (d) Closing Mondays takes the year to 208 services, so break-even becomes 22,368 ÷ 208 = 108 covers a service — unless fixed costs fall. They will fall a little (utilities, some fixed labor) and variable costs disappear with the covers, but rent, insurance, salaried management, and G&A do not care. Closing a day only helps if the day's contribution was less than the fixed cost it avoided, which is a calculation, not an instinct.
32.4 The listed items sum to \$591,000** and \$591,000 ÷ 0.40 = \$1,477,500, so the arithmetic is right and the answer is wrong. **The error: "Hourly kitchen and floor labor \$214,000" is variable labor. A 40% contribution margin ratio means variable costs are already 60% of sales — which necessarily includes variable labor — so listing it again as fixed double-counts it. Corrected fixed costs: \$591,000 − \$214,000 = \$377,000**; break-even = 377,000 ÷ 0.40 = **\$942,500. The error overstated break-even by \$535,000 — more than a third of a year's sales, and enough to make a viable business look unfinanceable.
32.5 (a) \$437,635 ÷ 0.4507452 = **about \$970,915 on the 45.07% ratio; \$437,635 ÷ 0.40528645 = **\$1,079,815 on the 40.53% ratio — a difference of \$108,900**, or about 7% of planned revenue. (b) The same restaurant's labor line differs by \$70,461 because the two numbers are built differently: a summary P&L labor line is a planned target — often the number the plan needs — while a labor line built from an actual schedule is a bottom-up sum of positions, hours, rates, and burden, and it includes hours the summary assumed away. Neither is dishonest; they answer different questions. (c) Put the conservative one (\$1,079,815, the schedule-built basis) in front of a lender**, because a plan that survives its own least flattering assumption is the only kind a lender should believe, and because the schedule-built number is the one you can defend line by line. Put the **same number on the office wall** — and this is the point: they should be the same answer, and the reason they feel like different answers is that operators are tempted to manage against the flattering basis while presenting the conservative one. Managing to \$970,915 while telling the bank \$1,079,815 means every week between those two numbers feels like success and is not. (d) The rule: a contribution margin ratio is meaningless without the cost base it was computed on. Never quote one without naming which variable costs are inside it, and never pair a ratio from one base with a fixed-cost or profit figure from another.
32.6 (a) 191,895 + 95,200 + 60,950 + 43,090 + 46,500 = \$437,635** ✓. Break-even = 437,635 ÷ 0.40528645 = **\$1,079,815. (Using the displayed 0.4053 gives \$1,079,780 — a \$35 gap. Round last.) (b) Margin of safety = 1,550,000 − 1,079,815 = \$470,185**, or **30.3%** of planned revenue. (c) Cash break-even = (437,635 + 69,500 = \$507,135) ÷ 0.40528645 = \$1,251,298**; margin of safety against it is \$298,702, or 19.3%. (d) Converting on the plan's own cover base — 95 covers a night at the plan's blended check, holding the check constant: accounting break-even is 1,079,815 ÷ 1,550,000 = 69.7% of plan → 66 covers a night; cash break-even is 80.7% of plan → 77 covers a night. On the annual base instead — 37,740 covers a year at a \$41.07 blended check — accounting break-even is 1,079,815 ÷ 41.07 = **26,292 covers a year, about 506 a week**, which is the same answer expressed differently. A student who converts without naming the base, or who divides by a dinner-only check on an annual cover count that includes brunch, will manufacture a phantom average check and overstate break-even by roughly seven covers a night. (e) Because the P&L never shows debt principal: a restaurant sitting between \$1,079,815 and \$1,251,298 is profitable and losing cash every month, and cash is what closes restaurants.
33.1 (1) Debt principal repayment — cash out, not an expense. (2) Capital expenditure — equipment, repairs that get capitalized, the eventual replacement of everything in the kitchen. (3) Working capital movements — inventory builds, deposits, prepayments, and the timing of payables. (4) Sales tax and payroll withholding timing — money in the account that was never yours. Add income taxes and owner distributions. Debt principal surprises people most: an operator who covers debt service comfortably on the P&L discovers that only the interest was ever an expense.
33.2 \$29,800 × 7% = **\$2,086 a week; across roughly 4.33 weeks, about \$9,032 sitting in the account at month end. That balance is not the restaurant's money** — it was collected on behalf of the taxing authority and is owed. An operator who reads the bank balance as available cash is borrowing from a creditor who charges penalties and, in many jurisdictions, holds individuals personally responsible for the shortfall.
33.3 (a) \$437,635 ÷ 52 = **\$8,416 of fixed outflow a week; \$69,500 ÷ 52 = **\$1,337 of debt service. (b) Revenue = \$29,800 × 72% = **\$21,456; variable = \$21,456 × 59.47% = **\$12,760; net = 21,456 − 12,760 − 8,416 − 1,337 = −\$1,057 a week**. (c) × 4.33 weeks = **about −\$4,577 for the month. (d) It obligates the operator to have that cash already in place before February begins — in an untouched reserve or an undrawn, already-approved line of credit. A line arranged in February is a line arranged from a position of weakness, and a bank asked for money during the month it is needed will want to know why the forecast did not exist. Note also that this is the average burn: the week the rent, the insurance, and the sales-tax remittance coincide is far worse, which is why the forecast must be weekly, not monthly.
33.4 A strong argument: not yet, and probably not \$40,000. The two documents you want on the table are the thirteen-week cash forecast — which will show the January and February trough this account is about to fund — and the debt schedule with any covenant test. The rule to set so it is never re-argued: distributions are taken from a defined surplus above a stated minimum operating cash floor, tested at a fixed point in the year (most operators use after the winter trough, not before it), with the floor set at some multiple of a payroll cycle plus the largest single obligation of the quarter. Write it down once; it converts an annual argument into an arithmetic check.
33.5 Closing balances: week 1 \$34,200**; week 2 **\$30,500; week 3 \$16,200**; week 4 **\$8,600; week 5 \$1,000**; week 6 **\$100. The trough is week 6 at \$100**. The line of credit is **not \$100 and it is not the \$37,900 peak-to-trough swing either: it must cover **the trough plus a minimum operating floor plus a surprise**. With payroll alone running near \$10,000 a week here, a defensible line is \$50,000–\$75,000 — enough to hold a floor of one full payroll cycle and absorb a compressor failure or a bad week without a phone call. The trough figure tells you the shape of the year; the line is sized against what must never be missed.
33.6 (a) 148,000 − 31,000 − 38,000 − 26,000 − 9,000 + 4,000 = \$48,000** of cash actually generated. (b) Against \$120,000 of planned distributions, a \$72,000 shortfall** — the owner planned to distribute 2.5 times the cash the business produced. (c) **Deferrable: capital expenditure (\$26,000) — and to a degree the inventory build. Not deferrable: interest, principal, and the tax timing.** Deferring capex is the standard move and it is a loan from your own equipment: the compressor, the fryer, the roof, and the chairs all fail eventually, and deferred maintenance arrives later as a larger, unplanned number, usually at the worst moment. Defer it once knowingly; defer it three years running and you have a restaurant that looks it.
34.1 Custody (holding the asset: cash in the drawer, product in the walk-in, the deposit bag), recording (entering it: the POS, the invoice, the count sheet, the books), and authorization (approving it: the comp, the void, the purchase, the payment). The classic restaurant example of all three in one person: the manager who counts the drawer, makes the deposit, enters the sales, and approves their own comps and voids — a common arrangement in an independent, and the one that makes a loss undetectable rather than merely possible.
34.2 The variance ladder descends from the aggregate to the actionable: the P&L line (food cost is high) → the category (which of meat, produce, dairy, dry goods) → the period and the shift (when did it start; which services) → the item (which recipes, compared theoretical to actual) → the transaction (which invoices, which plates, which employee). Each rung narrows the population of possible causes. Most operators stop at the category, which is exactly one rung above where a decision can be made, and then explain the number instead of fixing it.
34.3 84 ÷ 2,180 = 3.85%. Sustained voids above roughly 2–3% of tickets, or any concentration in one employee, one terminal, or one time of day, warrants a look — and voids after payment warrant one regardless of the rate.
34.4 (a) \$4,180 ÷ \$118,400 = 3.53%. (b) Policy at 1.0% = \$1,184; over policy = **\$2,996 a month, \$35,952 annualized. (c) Legitimate: genuine service recovery, a kitchen error, a marketing or press comp with an authorized purpose. Illegitimate: comping friends and family, comping to cover an unrung or mis-rung sale, and comping after payment to remove cash. The report that distinguishes them: a comp report requiring a reason code, the approving manager, and the time relative to the check's closing** — a comp entered after the check closed is a different event from one entered during service, and no legitimate recovery needs to happen after payment.
34.5 Comps reconcile: 410 + 355 + 1,714 + 498 + 1,203 = \$4,180** ✓, and sales sum to \$118,400 ✓. Server C produces 12.0% of sales and 41.0% of the comps — 12.1% of their own sales comped, roughly ten times servers A and B. Server E is a secondary flag at 6.2%. Three explanations to rule out before concluding anything: (1) section and shift — C may work the section or the daypart where recovery is genuinely most needed, or may be the one on the floor when a manager is not; (2) a training or approval problem — C may be comping with authorization, correctly, because nobody told them the policy or because a manager approves everything they ask; (3) a data artifact — C may be the server whose name is on the terminal used for house comps, press comps, or manager entries. The conversation to have first is with the approving manager, not with Server C** — every one of those comps was authorized by somebody, and if it was not, that is a much bigger finding.
34.6 A strong case: reframe the controls as protecting the owner, not catching the staff. With custody, recording, and authorization all in one person, there is no evidence the owner is honest — and that matters when an insurer investigates a claim, when a lender's covenant requires reviewed statements, when a partner or an investor asks, when the tax authority examines a return, and when the business is eventually sold and a buyer's accountant looks at the books. A sole-signer owner who adjusts inventory counts has also destroyed the variance ladder: the one number that would reveal a problem is being edited by the person the number is about. What a lender or insurer says is simpler — weak controls raise the price of credit and can void a fidelity claim. Minimum viable structure for a business too small to hire: (1) the outside bookkeeper reconciles the bank statement and receives statements directly from the bank, never through the owner; (2) two-person deposit verification, or a smart-safe / armored pickup that produces an independent record; (3) inventory counted by one person and extended by another, with adjustments requiring a written reason; (4) the owner's own comps and voids appear on the same report as everyone else's, reviewed monthly by the bookkeeper or the accountant; (5) invoice approval separated from payment — the owner approves, the bookkeeper pays, or the reverse, but not both. None of this costs a salary; all of it costs discipline.
Part VIII
35.1 Owner dependency is the degree to which the business's performance requires the owner's personal presence, relationships, and judgment. The practical test: what happens if the owner is gone for thirty consecutive days? If sales fall, costs drift, or the staff cannot answer a question, the owner is not an owner, they are an employee with liability. The two systems that most reduce it: written operating standards (recipes, specs, checklists, training) and a management layer with real authority and a number they are accountable for — a general manager who cannot approve anything has not reduced dependency at all.
35.2 \$96,000 + \$45,000 − \$146,000 = **−\$5,000**. On a fully staffed basis the restaurant loses money; the reported \$96,000 is the market value of the owner's unpaid second job. This is the most common concealed fact in independent restaurant financials, and it is the first adjustment a buyer, a lender, or a second-unit plan must make.
35.3 (a) \$96,000 × 60% = \$57,600 − \$34,000 = **\$23,600; ÷ \$480,000 = 4.9% return in year one. (b) Two things must be true: unit two must improve materially in years two and three toward or beyond unit one's maturity, and unit one must not deteriorate** while the owner's attention is elsewhere. A 4.9% first-year return is acceptable only as the first year of a curve; it is unacceptable as a steady state, since it is below the cost of the capital that funded it.
35.4 A strong argument: almost certainly no, as offered. Thirty days is not enough time to do a trade-area analysis, three contractor bids, a lease review, and a financing package — and a deadline set by the counterparty is a negotiating instrument, not a fact. The \$100,000 of TI money is real, but TI allowances are the cheapest thing a landlord gives and the most expensive thing to accept, because they buy a ten-year obligation. The disqualifying condition is in the question: the owner works six shifts a week in unit one, so there is no management capacity to staff unit two without breaking unit one. The one condition that would make it yes: a general manager already in place and proven in unit one — someone who has run it, alone, for a sustained period with the numbers holding. Absent that, the correct answer to a short fuse is to ask for a longer one and to be willing to lose the space.
35.5 Disqualifying: "a general manager who runs the unit alone — no." Everything else can be bought or built; this one is the difference between opening a second restaurant and opening a second job, and 35.6 prices it. Merely expensive: the partial systems documentation and the missing 12 weeks of operating cash — both are money and time, and both are solvable before a lease is signed. The "untested" row must become a completed test: the owner takes two consecutive weeks away, unreachable except for emergencies, before the lease is signed, and the unit's sales, prime cost, and guest feedback for those two weeks are compared to the four weeks either side. It is the cheapest due diligence available on the most expensive assumption in the plan.
35.6 (a) \$96,000 × 85% = **\$81,600. (b) + \$23,600 = **\$105,200. (c) Against \$96,000 standalone, that is **+\$9,200, a 1.9% return on \$480,000 of new capital in year one. (d) "Doubles the business" gets three things wrong: it assumes unit two performs like a mature unit one immediately, it ignores the new above-store overhead that did not exist before, and — the one operators never model — it ignores the cost of the owner's absence from the unit that currently works. The honest version of the pitch is a multi-year one: year one is roughly break-even on the new capital and is paid for out of unit one's degradation; the return arrives in years two and three, if unit two matures and if** unit one is protected by a manager who does not yet exist. A lender who hears that version will believe the rest of the plan.
36.1 Initial franchise fee (one time, at signing); royalty (a percentage of gross sales, continuing); advertising or brand fund contribution (a percentage of gross sales, continuing); and technology, systems, or support fees (usually a fixed monthly amount). Add renewal and transfer fees at the boundaries. Royalty and the ad fund continue regardless of profitability — they are charged on sales, so a franchisee losing money pays them in full, every month, which is the single most important structural fact in the arrangement.
36.2 \$1,300,000 × 7.0% = **\$91,000**.
36.3 \$45,000 ÷ 10 = **\$4,500 a year; ÷ \$1,300,000 = 0.35%** of sales.
36.4 Annual fee load on \$1,300,000: royalty and ad fund \$91,000 + amortized initial fee \$4,500 + technology (\$650 × 12 = \$7,800) = **\$103,300, or 7.95% of sales. (Note that 36.6 works from \$95,500 because it excludes the technology fee; be explicit with students about which scope you are grading.) The two entries carrying the most unquantified risk: "Territory: as designated by franchisor" — which as written protects nothing and permits a second unit across the street — and "required remodel every 7 years" — an unbounded capital obligation on someone else's schedule, easily six figures, landing in a year you do not choose. Demand in writing: a defined, mapped, exclusive territory with a stated radius or population, and a remodel scope with a cost cap and a notice period**, ideally with a franchisor contribution.
36.5 For the franchise: proven systems, a costed menu, a supply chain, training, brand demand on day one, and — genuinely important for a first-timer — a lender who has seen the concept's unit economics before, which makes financing easier and cheaper. Against: \$103,300 a year of fees, no menu control, no ability to reposition when the market moves, an obligation to remodel on demand, and a business that is far harder to sell as an independent asset. For the independent: full control, no fees, and the entire upside. Against: the operator must build every system in 36.1's list while also cooking, and a first-time operator's most common failure is precisely the systems. **Advice: the franchise**, for an operator with no restaurant management experience — the fee is the price of not learning cost control with their own \$210,000. The fact that would change it: relevant management experience — if the candidate has run someone else's restaurant profitably for two or more years, the systems argument collapses and the fees become a pure tax on their own competence.
36.6 (a) \$95,500 ÷ 0.40 = **\$238,750** of incremental sales. (b) ÷ \$1,300,000 = **18.4%**. (c) The question is reframed from "are the fees reasonable?" to **"can this brand deliver at least 18.4% more revenue than this operator would generate on their own, at the same cost structure — every year?"** For a strong national brand in a good site, plausibly yes. For a weak or regional brand, almost certainly not, and the arithmetic says so without any debate about the brand's merits. (d) Two offsetting avoided costs: **marketing the operator would otherwise have to buy** (the ad fund is not purely incremental — an independent spends 2–4% of sales on marketing anyway) and **purchasing power** (franchise supply contracts can be worth one to three points of food cost, which on \$1,300,000 is \$13,000–\$39,000 a year). Both are hard to quantify because the counterfactual is unobservable: nobody knows what this operator's food cost or marketing spend would have been independently, which is exactly why franchisors and franchisees argue about it.
37.1 Comparable-store sales measures sales growth for units open throughout both the current and the prior period, excluding new and closed units, so growth from opening restaurants is stripped out and only same-unit performance remains. The base-period rule is that a unit enters the comp base only after a defined maturity — commonly 12 to 18 months — so the opening spike and the ramp do not contaminate the comparison. Two things that most often mislead: price increases masquerading as growth (see 37.4), and base-period distortions — a prior year with a closure, a remodel, a road project, or a pandemic makes any comparison meaningless.
37.2 Span of control is how many units one above-store leader can effectively supervise. Three determining factors: geographic density and drive time (six units in one city is a different job from six across three states); unit complexity and volume (a full-service unit with a bar and events demands far more than a counter unit); and the strength of the unit-level management — a district with five strong general managers supports a much wider span than one with two vacancies and three new promotions. Typical spans run roughly five to ten units, but the number is an output of those three factors, not an input.
37.3 (\$1,412,000 − \$1,340,000) ÷ \$1,340,000 = \$72,000 ÷ \$1,340,000 = +5.4%.
37.4 (a) 0.979 × 1.076 = 1.0534, or +5.3% — which reproduces the +5.4% headline within rounding. (b) The decomposition reveals that all of the growth is price and the restaurant is losing guests: traffic is down 2.1% and check is up 7.6%. The headline says a healthy unit; the components say a shrinking one that has raised prices faster than it lost people. (c) Look next at traffic by daypart and by channel, and at the menu mix — specifically whether the check increase came from price or from mix (guests trading up, or guests forced onto more expensive items because cheaper ones were removed). Expect to find the traffic loss concentrated in the price-sensitive dayparts — lunch, early dinner, weeknights — which is where a price increase bites first and where it is hardest to win people back.
37.5 Healthiest: unit 4 — growth driven by traffic (+5.1%) rather than price, the best prime cost at 58.4%, and the lowest turnover at 48%. Those three facts are the same fact. In the most trouble: unit 3 on its face — comps −3.1%, prime 64.9%, turnover 96% — but the honest answer is that unit 3 is the most visibly broken while unit 1 may be the most dangerous, because unit 1 looks like the second-best performer and is actually losing guests behind a price increase (traffic −2.1%). Unit 3's problem is diagnosed and can be attacked; unit 1's is disguised by a good headline and will be discovered a year later. Least durable growth: unit 1 — price-only growth has a ceiling, and each increase costs traffic.
37.6 A strong intervention plan: (1) The DM's job is the district, not the unit. Three days a week in one restaurant means eight units are running without a supervisor, and drift across eight units costs more than the failure of one — this is a span-of-control failure, not an effort failure. (2) Take off their plate: the routine unit visits at the four healthiest units (move to a scheduled cadence with a reporting checklist), and any above-store project work. (3) The failing unit specifically: the numbers say the sequence is people first — 96% turnover produces the 64.9% prime cost, not the other way around, because a crew that is always new over-portions, wastes, runs slow, and requires overtime. So: assess the general manager honestly and decide within two weeks whether they are the fix or the problem; stabilize the schedule and the pay structure; then attack prime cost with a weekly flash and a portion audit. Put a temporary manager or a strong GM from another unit into the failing one rather than the DM — a DM covering a unit is the most expensive labor in the company. (4) The number that tells you it worked: not comps, which lag by months — turnover and prime cost, weekly, in that order. Turnover stabilizing at or below the district average, followed by prime cost falling toward 60%, is the leading indicator; comps confirm it two quarters later.
38.1 Pre-consumer or prep waste — trim, peel, over-production, mistakes, product that never reaches a guest. Spoilage — product that expires, is stored badly, or is ordered beyond usage. Plate waste — food the guest was served and did not eat. Most recoverable: prep waste, because it is entirely within the kitchen's control and is addressed by yield tests, specs, batch sizing, and cross-utilization. Least recoverable: plate waste, because the guest has already paid for it — the money is in the register, and the only lever is portion size, which trades directly against the guest's perception of value.
38.2 38 × \$2.10 = \$79.80 a day; × 310 = \$24,738 a year.
38.3 (a) \$14,200 × 35% = **\$4,970 a year; \$6,400 ÷ \$4,970 = 1.29 years simple payback. (b) Omitted: the recurring cost of the bags, labels, and the labor to seal, label, and rotate — plausibly \$1,500–\$3,000 a year, which lengthens payback toward two years or more. It does not** change the decision — a two-year payback on a control that also improves food safety and inventory accuracy is a good investment — but it changes the honesty of the pitch, and an operator who presents a 1.29-year payback and delivers 2.1 loses credibility on the next proposal.
38.4 A strong decision: do not switch, and say so honestly. Paying three times as much for compostable packaging in a market with no commercial composting means the containers go to the same landfill as the current ones, where compostable material does not reliably break down — the restaurant would be buying the appearance of sustainability at real cost, with no environmental benefit. What to tell a guest who asks: exactly that, plus what you are doing instead — waste reduction where it is measurable, energy retrofits, sourcing, food donation. Guests respect a specific answer far more than a virtuous one. The condition that changes it: commercial composting becomes available, or the packaging's cost premium falls enough that the decision no longer trades against something with a real return.
38.5 \$2,940 × 14% = **\$411.60 a month, \$4,939 a year**; \$9,800 ÷ \$4,939 = **1.98 years** payback. Two negotiable lines: **waste hauling (\$445) — right-size the container and the pickup frequency, which almost every restaurant over-buys, and bid it, since haulers compete on route density; and water and sewer (\$610) — much of a restaurant's sewer charge is computed from water consumption, and many jurisdictions offer an evaporation or irrigation credit**, while low-flow pre-rinse spray valves and fixing a single running toilet or leaking ice machine can cut the bill materially. Gas and electricity are also often negotiable in deregulated markets; the retrofit and the rate are separate savings and can be taken together.
38.6 (a) 2 oz × \$0.44 = \$0.88 a plate; × 190 = \$167.20 a week**, **\$8,694 a year. (b) Against prep waste of \$24,738** and spoilage of **\$14,200 — a combined \$38,938** of genuinely recoverable dollars, more than four times what the portion cut returns. (c) Plate waste is least recoverable because **the guest already paid for it**: that food generated revenue, so "recovering" it means reducing what you sell, and the only mechanism is a smaller portion — which the guest may notice, which affects perceived value, and which can cost more in lost visits than the \$8,694 it saves. Prep waste and spoilage, by contrast, are product you paid for and never sold; recovering those dollars costs the guest nothing at all. The operator's attention belongs on the walk-in, the prep list, and the order sheet — and if the 2.4 ounces of plate waste is genuinely consistent, the right response is a yield-tested portion spec supported by a plate-waste audit**, taken as a deliberate menu decision with the price examined at the same time, not as a quiet shave.
39.1 In order: (1) Is the revenue line real and is it moving? — everything below is a percentage of it, and a cost problem and a revenue problem look identical in percentage terms. (2) What is prime cost, in dollars and percent? — it is the majority of controllable cost and it is where the answer usually is. (3) Which lines are outside benchmark, and of those, which can actually be changed? — separating recoverable from unrecoverable. (4) What does cash look like, and how long is the runway? — because it sets how much time the plan has. The order matters because fixing a cost line on a falling revenue base is treating a symptom, and because a plan that is correct but slower than the cash runway is not a plan.
39.2 Recoverable costs can be changed by decisions the operator can make inside the current period, without renegotiating a contract or moving. Unrecoverable costs are contractually or structurally fixed for the horizon in question. Recoverable: food and beverage cost (specs, portions, purchasing, price), variable labor (schedule, staffing guide). Unrecoverable: base rent under a signed lease, debt service on an existing note. The test: can I change this with a decision I have the authority to make this month, and will the change show up in this quarter? Note that "unrecoverable" is horizon-dependent — rent is unrecoverable this year and negotiable at renewal.
39.3 (a) COGS 34.0%, labor 39.0%, occupancy 10.0%, other operating 18.0%, G&A 4.0%. (b) Prime = 401,200 + 460,200 = \$861,400**, or **73.0%**. (c) Total costs **\$1,239,000; operating result = 1,180,000 − 1,239,000 = **−\$59,000**, or **−5.0%**. (d) **Prime cost**, and within it labor — at 73% prime the business has no chance of profitability at any occupancy, and labor at 39% is both the largest overrun in dollars (\$59,000 above a 34% target, which happens to be exactly the loss) and the one that can move fastest.
39.4 Inside benchmark: occupancy (10.0%, at the top of the 6–10% range), other operating (18.0%, at the top of 12–18%), and G&A (4.0%, comfortably inside 2–5%). Outside: COGS (34.0% against 28–33%), labor (39.0% against 30–36%), and prime (73.0% against ≤60%). The line this operator cannot fix this year is occupancy — \$118,000 is a contractual dollar amount, and it is at 10.0% only because revenue is low. That changes the whole plan because it means the recovery cannot be built on cost cuts alone: two of the three fixable lines are in prime cost, and the fourth lever — revenue — is the only thing that moves occupancy at all. Note also that "at the top of the range" on two lines means there is no slack anywhere; this statement has no cushion in any direction.
39.5 A defensible ranking: first, cut two positions — fastest, largest, and within the operator's authority; at 39% labor on a business losing \$59,000, this is the only lever that can change the number inside one payroll cycle. **Second, raise menu prices 6%** — roughly \$70,800 of revenue at nearly full contribution if traffic holds, and it also improves the occupancy percentage; it is slower and riskier than the labor cut because traffic response is unknown, so it should follow a menu-engineering pass rather than be applied across the board. Third, renegotiate the lease — potentially large and permanent, but slow, uncertain, and dependent on a landlord's willingness; start the conversation now precisely because it takes months. Not at all: refinance the equipment lease to lower payments — it treats a symptom. It changes nothing about a business losing \$59,000 at the operating line; it converts a cash problem into a longer, more expensive cash problem and consumes management attention that belongs on prime cost. Dropping lunch ranks below the first three and above the last: it may be correct, but only after computing whether lunch's contribution exceeds the fixed cost it absorbs (see 32.3(d)) — closing a daypart on instinct is how operators discover that dinner was being subsidized.
39.6 (a) COGS 34% → 30% saves 4 points = \$47,200**; labor 39% → 34% saves 5 points = **\$59,000; other operating 18% → 15% saves 3 points = \$35,400**. Total **\$141,600. (b) −59,000 + 141,600 = \$82,600**, or **7.0%** operating margin — a healthy full-service result. (c) \$118,000 ÷ 0.08 = \$1,475,000** of revenue, a 25% increase. (d) It proves that **occupancy was never a cost problem; it was a revenue problem wearing a cost problem's clothes.** The dollar figure never changed. An operator who reads 10.0% and concludes "my rent is too high" will spend a year trying to renegotiate a lease when the actual finding is that this restaurant is under-performing its space by 25%. **Percentages diagnose; dollars decide.** Note also the sequencing consequence: the \$141,600 of recoverable savings is worth more than any plausible rent concession, and it is available now.
40.1 A DSCR covenant requires the borrower to maintain a debt service coverage ratio — net operating income ÷ total debt service — at or above a stated level, commonly 1.25×. It is tested at a stated frequency (here, annually) against financial statements the borrower must deliver. A breach is an event of default: in practice the lender's options range from a waiver, often for a fee, through additional reporting, a cash sweep, or additional collateral, to acceleration of the loan. Most breaches on a performing loan are waived or amended; the covenant's real function is to give the lender a seat at the table before the situation becomes unrecoverable.
40.2 Five common conditions, each with its purpose: a personal guarantee (aligns the borrower's own assets with repayment); a lien on business assets and a UCC filing (secures the collateral and prevents it being pledged elsewhere); life insurance on the key operator assigned to the lender (protects against the person the loan is actually underwriting); required financial reporting on a schedule — monthly or quarterly statements, annual tax returns (gives the lender visibility rather than surprise); and limits on distributions, additional debt, and capital expenditure without consent (prevents cash leaving the business ahead of the lender). Others include a required minimum owner injection, hazard and business-interruption insurance with the lender named, and a landlord's waiver or subordination agreement.
40.3 \$261,020 ÷ \$69,500 = 3.76×.
40.4 (a) 1.25 × \$69,500 = **\$86,875 of operating profit. (b) Cushion = \$261,020 − \$86,875 = \$174,145, which is 66.7% of planned operating profit — operating profit could fall by two thirds and the covenant would still be met. (c) An annually tested covenant cannot detect anything that happens and resolves inside a year** — a seasonal cash trough, a two-month labor overrun, a missed payroll funded by a personal loan, or a vendor stretched to sixty days. It measures a twelve-month average of profitability, and a restaurant does not fail on a twelve-month average.
40.5 The file supports one conclusion: the loan is approvable on covenant grounds under every scenario modeled. On plan, at a higher labor line, and in the combined downside, operating profit clears the \$86,875 trip point with coverage of 3.76×, 3.08×, and 2.23× — the covenant is never close to being breached. What is wrong with it, in one sentence: the covenant tests annual profitability, and this business's actual risk is a February cash trough — so the instrument is measuring the thing that is not going to fail while the thing that is going to fail happens between tests, invisibly.
40.6 A strong argument contains four things. (1) Why the covenant is the wrong instrument although correctly applied. A 1.25× DSCR test measures whether a full year of operating profit covers a full year of debt service. It is the right test for the risk that a business is structurally unprofitable. It is the wrong test for a business that is structurally profitable and seasonally illiquid — which is the ordinary condition of a restaurant in a cold-weather market. Both facts in the file are true at once: every scenario clears at 2.23× or better, and the operating account goes negative in February. Nothing about the covenant is misapplied; it is simply pointed at the wrong failure mode, and its annual frequency guarantees that the failure occurs and resolves between two tests that both read "compliant." (2) The instrument to write instead. A minimum liquidity covenant — a required minimum of unrestricted cash plus availability on a committed line, tested monthly, sized to something like one payroll cycle plus one month of fixed obligations — with a 13-week rolling cash forecast delivered monthly. That tests the thing that actually kills the borrower, and it tests it often enough to matter. Reasonable variants: a seasonal working-capital line sized to the modeled trough; a distribution block below a stated cash floor; a springing DSCR test triggered by a liquidity breach rather than a calendar. (3) What the operator should do regardless of what the lender writes. Assume the covenant will not warn you. Build the 13-week cash forecast and run it weekly; know the trough week by name before the year begins; arrange the line of credit in October, not February (see 33.3–33.5); set a distribution rule tied to a cash floor; and treat covenant compliance as evidence of nothing. A covenant is the lender's instrument for the lender's risk. (4) The general lesson — worth stating explicitly, because it is the book's closing argument: a business can be profitable, compliant, well-financed, and correctly measured by every instrument pointed at it, and still run out of money in February. Cash is not profit, and no covenant anyone writes will make it so.