Chapter 10: Menu Development: Design, Pricing, Engineering, and the Menu as Your Most Important Marketing Tool

"Show me the menu and I will tell you your labor cost, your walk-in, your vendor list, and roughly how your Friday goes. I do not need to taste anything." — constructed; what a consultant says on the first morning of an engagement

Overview

Nine chapters in, and we finally get to the food.

Except we do not, and the fact that we do not is the whole point of this chapter. The document you are about to write is not a list of things you like to cook. It is the most consequential operating document in the building, and it is consequential in ways that have almost nothing to do with flavor. Every line on it commits you to a supplier, a delivery day, a shelf in the walk-in, a station on the line, a skill level you now have to hire, a piece of equipment you now have to buy, a prep task that has to happen before four o'clock, and a number of minutes that will be spent on a ticket at 7:40 on a Saturday when eleven other tickets are also live.

Write twenty-two items and a small line can execute them at ninety-five covers. Write thirty-eight and the same line cannot, and you will not discover that until the third Friday, when ticket times go to thirty-one minutes and a four-top walks out. The menu decided that. Not the cooks.

Here is the way to hold it. A restaurant does not buy ingredients for dishes. It buys deliveries, and then decides what those deliveries become. A whole air-chilled chicken arriving on a Tuesday is not one dish; it is a signature entrée, a starter, a side, a sauce base, a brunch plate, and a stock — or it is one dish and a bag of expensive garbage, depending entirely on what you wrote on this piece of paper months before the bird showed up. That discipline is cross-utilization, it is the central craft of this chapter, and it is most of the reason a thirty-percent food cost is achievable at all.

This chapter drafts Bellwether's actual menu — the one the next six chapters cost, engineer, purchase, cook, and analyze. We will name the signature dish, place it, price it, and prove that the price ladder produces the \$33.12 of food per guest that Chapter 2's check build already committed us to. We will not cost a single plate; Chapter 11 does that, and doing it here would let us cheat. And we will treat menu psychology as what it is: a body of confidently repeated claims, some supported, several not, and all of them worth less than a server who can describe the chicken.

In this chapter, you will learn to:

• Read a menu as five documents at once — sales, production, purchasing, labor, and brand — and say which one any given decision serves.
• Filter an item list through concept fit, skill fit, and equipment fit, and articulate why each rejected item was rejected.
• Draw a cross-utilization map that traces every purchased line to every menu item it reaches.
• Quantify what one more item costs before you have bought a single ingredient for it.
• Write a standardized recipe that survives a new cook, a busy Saturday, and your own absence.
• Separate the menu-design advice that has evidence behind it from the advice that merely has confidence behind it.
• Build a price ladder from a target check and demonstrate, with arithmetic, that it resolves.

Learning Paths

🏗️ Opening — this is one of your load-bearing chapters. Work §10.2 through §10.4 with your own concept, and draw the cross-utilization map before you write a single price. The menu you write here determines the equipment schedule you signed for in Chapter 7 and the purchasing system you will build in Chapter 13. 📋 Managing — weight §10.4, §10.5, and §10.8. You probably inherited a menu. Your leverage is item count, recipe discipline, and change cadence — and §10.5 is the single cheapest fix available to most restaurants in America. 🍸 Beverage — §10.7 is yours: a price ladder works the same way on a list as on a menu. Note throughout where the food menu creates beverage attachment points, and where it kills them. Chapters 15 and 16 build the program; this chapter builds the plate it sits next to. 🚚 Small Format — §10.3 and §10.4 matter more to you, not less. A truck with four square feet of cold storage lives or dies on cross-utilization, and a twelve-item menu in a kitchen the size of a closet is still probably four items too long.

10.1 What a menu has to do: five jobs at once

Ask ten operators what a menu is and nine will say some version of "it is what we sell." That is true, in the way that "a lease is where you put the restaurant" is true — accurate, useless, and the beginning of a very expensive misunderstanding.

A menu, as a document, is the restaurant's operating specification expressed as a list of things a guest can order. It is simultaneously a sales instrument, a production schedule, a purchasing specification, a labor model, and a brand statement — and every item on it is a commitment in all five registers at once. You cannot change one without changing the other four, and most menu disasters are the result of an operator optimizing one job while blind to the rest.

FIGURE 10.1 — One document, five jobs                         [constructed teaching example]

                          ┌───────────────────────────┐
                          │        THE MENU           │
                          └─────────────┬─────────────┘
        ┌──────────────┬────────────────┼────────────────┬──────────────┐
        │              │                │                │              │
   1. SALES       2. PRODUCTION    3. PURCHASING    4. LABOR       5. BRAND
   what a guest    what the        what arrives     how many        what the guest
   is asked to     kitchen must    on the dock,     hands, at       believes about
   choose from     make, and in    from whom,       what skill,     you before the
   and at what     what order      how often        at which        food arrives
   price                                            station

   MOVES:          MOVES:          MOVES:           MOVES:          MOVES:
   average check   ticket times    food cost        labor cost      who walks in
   attachment      86 rate         waste/spoilage   turnover        and how often
   turns           consistency     vendor leverage  training time   what they pay

   ─────────────────────────────────────────────────────────────────────────────
   Jobs 3 and 4 together are PRIME COST. The menu sets both halves of it before
   you have hired anyone or bought anything. This is why the menu is written
   before the staffing guide (Chapter 19) and before the order guide (Chapter 13).

Look at the bottom line of that diagram, because it is the argument of the chapter. Chapter 1 introduced prime cost — cost of goods sold plus total labor — as the number that predicts survival. The menu is the document that sets both halves of it, months before a single dollar moves. A menu built out of eleven proteins, four cooking methods you do not own, and six items that each require their own dedicated prep has already decided that your food cost will be high and your labor will be higher. No amount of downstream discipline recovers that. You do not manage your way out of a menu.

Menu categories, and what they are actually for

Menu categories are the named groupings a menu is organized into — starters, mains, sides, desserts, or a restaurant's own idiom for the same thing. They look like a convenience for the reader. They are not. They are three things at once:

1. A pacing instruction. Categories tell the guest what a meal here looks like — whether this is a two-course room or a graze-and-share room — and therefore how long they will occupy the table. That is a revenue-management decision made in typography (Chapter 24).
2. A production grouping. Everything in one category tends to be produced at one station. A category is, from the kitchen's side, a station's workload.
3. A measurement unit. Category is the level at which you will later analyze what sells. Chapter 12 does that work; it can only do it if your categories were drawn sensibly at the start.

Categories also cap things. If you have one dessert category with three items, you have decided that dessert is not a business here — a defensible choice with real consequences for check average and for the last ten minutes of a table's visit. Say it out loud when you make it, because the alternative is discovering in month seven that dessert attachment is 8% and nobody ever decided that.

👨‍🍳 On the Line

What a menu meeting actually looks like, and who should be in the room.

The romantic version is a chef alone with a notebook. The real version, in a restaurant that intends to survive, is four people and a whiteboard, and it takes several sessions.

Who is there: the chef, who proposes. The sous or lead cook, who has to execute it at 8:15 on a Saturday and who will tell you — if you have built the kind of kitchen where this is safe — that the item you love takes four minutes too long. The front-of-house partner, who knows what a server can actually describe. And whoever places the orders, because they are the only person who will notice that three new items each require one perishable thing nothing else uses.

What is on the whiteboard: not recipes. Three columns. Item · Station · What else uses it. The third column is the one that matters and the one nobody writes down.

How it goes wrong. The chef presents twenty-six items. Everyone is polite. The list survives intact because nobody wants to be the person who killed the dish, and because at no point did anyone say a number. Six weeks after opening the kitchen quietly stops doing the two hardest items well, and nobody decides that either — it just happens, at the pass, under pressure, which is how most restaurant decisions actually get made.

The fix is a rule, agreed before the first item is proposed: nothing goes on the list until somebody names the station it fires from and one other menu item that uses at least one of its components. If neither can be said, it is a candidate for a special (§10.8), not for the printed menu.

10.2 Building the item list: concept fit, skill fit, equipment fit

The item list is built by subtraction, not addition. You start with everything the chef wants to cook and everything the concept implies, and you run it through three filters in a fixed order. The order matters: concept fit first, because an item that fails it should never consume a minute of argument about equipment.

Filter one — concept fit. Does this item belong to the restaurant Chapter 2 defined? Bellwether positioned itself at the top of the neighborhood band and the bottom of the occasion band: the nice place you can go on a Tuesday without it being an event, serving four occasions in a specific ratio at a \$46 dinner check. An item fails if it would look strange to the weeknight guest who is 46% of covers, if it cannot be ordered by a two-top on a compressed timeline, or if it belongs to a price point the room does not occupy.

Filter two — skill fit. Can the cooks you can actually hire, at the wage the plan actually funds, produce this correctly on the four hundredth repetition? Not "can the chef do it." The chef will not be on the line in year two, and that is the whole point of building a business rather than a job. This filter kills more good dishes than the other two combined, and it is the one first-time operators skip — because the chef-owner is standing right there being able to do it.

Filter three — equipment fit. Does the equipment do this, at volume, without displacing something more valuable? Chapter 7 turned 2,800 square feet into 68 seats and a line and priced the hood and make-up air upgrade the hearth demanded. Every square inch of that fire is capital you are paying for, and an item occupying it four minutes for a \$16 sale is competing with one occupying it eight minutes for a \$29 sale.

⚠️ Where the Money Leaks

The item you added for one guest.

Every menu accumulates them. A regular asks for a gluten-free pasta. A server reports that "people keep asking" for a fish that is not on the menu. A vendor offers a good price on something and it goes on as a special that never comes off.

None of these is stupid. Each one solves a real, visible problem, and each one costs money that is invisible. §10.4 puts a number on it — roughly \$3,020 a year in identifiable non-ingredient cost for an item selling nine a week — but the number is not the whole leak. The leak is that nobody ever removes anything. Items arrive one at a time, in response to specific requests, and they leave only when a menu is deliberately rebuilt.

What the disciplined operator does: run a one-in, one-out rule on the printed menu, and give the requests somewhere else to go. A rotating special (§10.8) answers "people keep asking" without a permanent commitment. A single, well-executed accommodation — "we can do the chicken without the sauce and the roots without butter" — answers most dietary requests better than a dedicated item does, because it is a thing the kitchen already makes four hundred times a week and therefore makes well.

What Bellwether cut, and why

The cuts teach more than the list. Four items were on the whiteboard and did not survive.

Notice that only one of those four was rejected because it was a bad dish. Three of them were rejected for reasons that have nothing to do with how they taste — which is the chapter's version of the book's first theme. The food is the easy part. Everyone in that meeting could cook all four.

The menu

Here it is: the document the rest of Part III works on.

FIGURE 10.2 — BELLWETHER, dinner                                     [the Bellwether plan]
                     Tuesday through Saturday, 5:00 to 10:00

  ────────────────────────────────────────────────────────────────────────────────
  TO START

  Hearth bread, cultured butter                                                  7
  Wood-roasted olives, marcona almonds, orange                                   8
  Little gems, buttermilk, herbs, hearth breadcrumb                             12
  Ember beets, whipped ricotta, hazelnut, honey                                 13
  Chicken liver mousse, mostarda, grilled bread                                 14
  Charred broccoli, chile, anchovy, lemon                                       14
  Smoked trout rillettes, pickled onion, rye toast                              15
  Hearth flatbread, autumn squash, fontina, sage                                16

  ────────────────────────────────────────────────────────────────────────────────
  FROM THE HEARTH

  Hearth chicken — half a bird, wood-roasted, roasted roots, salsa verde        29
  Ember trout, brown butter, capers, charred lemon                              28
  Pork chop, apple mostarda, mustard greens                                     31
  Autumn squash and grains, ember tomato, ricotta salata                (v)     24
  Hanger steak, salsa verde, hearth potatoes                                    36
  Dry-aged ribeye, 14 oz bone-in, ember shallot                                 46

  FROM THE STOVE

  Hand-cut pasta, pork sugo, cured yolk                                         25
  Bellwether burger, aged cheddar, pickles, hearth potatoes                     21

  ────────────────────────────────────────────────────────────────────────────────
  ALONGSIDE

  Charred greens, lemon, chile                                                   9
  Hearth potatoes, garlic, rosemary                                             10
  Roasted roots, honey, yogurt                                                  10

  ────────────────────────────────────────────────────────────────────────────────
  TO FINISH

  Hearth crisp — the fruit that is good this week, crème fraîche                11
  Chocolate budino, olive oil, sea salt                                         11
  Buttermilk soft-serve, honey, hearth-toasted oat                              10

  ────────────────────────────────────────────────────────────────────────────────
  (v) vegetarian.  Please tell us about allergies before you order and we will tell
  you honestly what this kitchen can and cannot do.
  Parties of six or more are served on one check.
  Cocktails, beer, and the bottle list are on a separate card.
  Printed and dated every Tuesday.

Twenty-two food items. Eight to start, eight mains, three sides, three desserts. That is a short menu by American full-service standards and it is short on purpose; §10.4 shows what the twenty-third item would cost.

A few things to notice before we go further, because each of them is a decision that will be defended later in this chapter.

• The signature is named and placed. The Hearth Chicken — half a wood-fired air-chilled bird, roasted roots, salsa verde, \$29 — sits first in the hearth panel with the longest description on the menu. It is the dish the restaurant is for. It is also, as Chapter 11 will demonstrate with a cost card, the most profitable dollar-contributor in the building, which is a very pleasant thing for a signature dish to be and is not an accident.
• The mains are split into two panels by station, not by protein. "From the hearth" and "from the stove" is a brand statement — it tells the guest what the restaurant is — and it is simultaneously a load-balancing statement, made visible. §10.4 shows why that split is not decoration.
• The vegetarian item is an entrée in the main panel, not an afterthought at the bottom. Chapter 3 made the point that a menu with no vegetarian entrée has told a fraction of the market that the room was not built for them, at zero marginal cost to reverse. This is the reversal.
• There are no dollar signs, no decimals, and no leader dots. §10.7 explains why, and is honest about how thin the evidence for that convention is.
• The menu is dated and reprinted weekly. Chapter 3 costed the alternative: a menu that is expensive to reprint is a menu you will not reprint, and a menu you will not reprint is where food cost goes to drift.

10.3 Cross-utilization: how one delivery becomes six dishes

This is the section that pays for the chapter.

Cross-utilization is the deliberate design of a menu so that each purchased ingredient reaches several menu items — and, at its best, so that the parts of an ingredient that a careless kitchen throws away become menu items of their own. It is not thrift and it is not a sustainability posture, though it is both of those incidentally. It is the primary structural mechanism by which an independent restaurant achieves a food cost in the twenties or low thirties without buying cheap food.

The logic is worth stating plainly, because it is not obvious.

• It raises the volume of every line you buy. Buying fifty-two chickens a week rather than nineteen changes what a purveyor will do for you (Chapter 13).
• It cuts the number of lines you buy. Fewer lines means fewer invoices, fewer specs, fewer count-sheet rows, fewer things to receive incorrectly, and fewer things to run out of.
• It converts trim into sales. The livers, the fat, the bones, the smaller fish, the shoulder behind the chop — these are either revenue or they are garbage you paid for.
• It compresses the walk-in. Chapter 7 gave the plan 900 square feet of back of house and 200 of storage. A menu of forty items with forty ingredient families does not physically fit in it.
• It turns over inventory faster. An ingredient used by four dishes moves four times as fast, which means less of it sits, which means less of it spoils. Chapter 13 makes this a number.

🧾 Read the Numbers

```text FIGURE 10.3 — "What one week's deliveries actually become" [the Bellwether plan] THE ARTIFACT The cross-utilization map for the draft dinner and brunch menus: every purchased family on the left, every menu item it reaches on the right. Drawn before the order guide is written. THE CONTEXT Bellwether, planning. Twenty-two dinner food items and twelve brunch lines, five dinner services and two brunches, roughly 475 dinner covers and 220 brunch covers a week.

POULTRY — 52 whole air-chilled birds a week, two deliveries ├─ 104 halves ...................... HEARTH CHICKEN (dinner) ├─ backs, necks, wing tips ......... stock → pasta sugo, brunch gravy, braises ├─ livers .......................... chicken liver mousse (dinner) ├─ rendered fat .................... hearth potatoes · flatbread · brunch hash └─ Saturday's unsold halves ........ chicken hash (Sunday brunch)

THE HERB-AND-CAPER BOX — one order, one sauce, made twice a week └─ SALSA VERDE ..................... hearth chicken · hanger steak charred broccoli · chicken hash (brunch)

ROOT BOX — carrots, parsnips, celery root, turnips; two deliveries ├─ roasted ......................... the chicken's roots · roasted roots (side) ├─ ember-roasted ................... ember beets · chicken hash · grain bowl └─ trim and peel ................... stock mirepoix

FLOUR, LEVAIN, BUTTER — one dough ├─ baked as loaves ................. hearth bread · grilled bread for the mousse │ and the rillettes · brunch toast ├─ stretched ....................... hearth flatbread (dinner) · flatbread and │ egg (brunch) ├─ day-old ......................... hearth breadcrumb → little gems · pasta └─ laminated trim .................. brunch pastry

PORK — one rack and one shoulder a week ├─ rack, cut in house .............. pork chop (dinner) ├─ shoulder, ground ................ pasta sugo · brunch sausage gravy └─ fat, rendered ................... flatbread · potatoes

BEEF — hangers whole, ribeyes portion-cut, chuck and brisket for the grind ├─ hanger, trimmed in house ........ hanger steak ├─ hanger trim + chuck + brisket ... BELLWETHER BURGER (dinner and brunch) └─ ribeye, bought portioned ........ dry-aged ribeye

TROUT — whole fish, two deliveries ├─ fillets ......................... ember trout ├─ smaller fish, hot-smoked ........ trout rillettes · trout toast (brunch) └─ frames .......................... discarded. We do not make fish stock here.

DAIRY — buttermilk, cream, ricotta ├─ buttermilk ...................... little gems dressing · soft-serve · biscuits ├─ cream ........................... crème fraîche for the crisp · hearth eggs └─ ricotta ......................... ember beets · flatbread · grain bowl

WHAT IT SHOWS Thirty-four menu lines across two dayparts, built from about 54 purchased lines on the order guide. Seven of the eight programs above reach four or more menu items. The poultry program alone reaches five menu items in both dayparts, and the bread dough reaches six. Half of the menu's apparent variety is produced by re-cutting, re-cooking, and re-plating a small number of things. WHAT IT DOESN'T It does not say what any of this costs — every figure in this map is a quantity, not a dollar, and Chapter 11 supplies the dollars. It does not prove the volumes reconcile; par levels and order quantities are Chapter 13's work. It says nothing about whether the guest will find this much repetition monotonous. And it deliberately shows one thing being thrown away — the fish frames — because a map that claims to use everything is lying. THE DECISION Write the order guide from this map, not from the menu. Set par levels on the eight programs, not on twenty-two dishes. Then take the map to two purveyors and ask what fifty-two birds a week is worth as a committed volume (Chapter 13). THE LESSON You do not buy ingredients for dishes. You buy deliveries and decide what they become. A menu written the other way around is how a walk-in fills with forty things that each feed one dish. ```

The chicken, in detail

The poultry program is worth walking through slowly, because it is the clearest instance of the principle and because it produces the book's recurring example.

Bellwether sells roughly one Hearth Chicken for every four entrées ordered. On 95 covers a night at an entrée attachment of 0.90, that is about 86 entrées a service and about 21 Hearth Chickens — roughly 105 a week, which is 52 or 53 whole birds. Two deliveries a week, twenty-six birds each.

What arrives is fifty-two whole 3.5-pound air-chilled birds. What leaves the walk-in is: 104 halves, dry-brined, air-dried, par-roasted in the cooling hearth and finished to order over the fire — the \$29 entrée and roughly a quarter of everything the kitchen sells; fifty-two sets of backs, necks, and wing tips, roasted in the last heat of the previous night's fire and turned into the stock that bases the pasta sugo, the brunch gravy, and every braise on the specials board; livers, which become a \$14 starter built from a component that arrived attached to something else; rendered fat, which is what the hearth potatoes roast in and what makes the brunch hash taste like the restaurant rather than like a hash; and Saturday's unsold halves, which are Sunday's chicken hash — not thrift dressed as a menu item, but the reason the kitchen can carry the last seating on a Saturday without either running out or throwing money in a bin.

One delivery. Five menu items. Two dayparts.

👨‍🍳 On the Line

Tuesday, 9:40 in the morning, and the honest version of what this costs.

Cross-utilization is not free, and the place it is not free is a stainless table at nine-forty in the morning with twenty-six birds on it.

Somebody has to break them down. A cook who is good at it does a bird in about ninety seconds — halve it, remove the backbone and wing tips, trim, pull the liver if it came with one, save the fat. Twenty-six birds is forty minutes plus setup, cleanup, and the sanitation break between raw poultry and anything else on that table. Call it an hour and a quarter of a prep cook's day, twice a week. That hour is a labor cost, and Chapter 19 will put it in the staffing guide. This is the trade Chapter 1 described running in both directions: you are buying a lower food cost with labor, and the only way to know whether it was a good trade is to watch prime cost, not food cost.

Three things that go wrong, in order of frequency:

The birds come without giblets. Many air-chilled birds do. If your mousse assumes fifty-two free livers a week and the birds arrive empty, you have a menu item with a purchased ingredient nobody planned for. Ask before you write the item; put it in the product specification (Chapter 13).

The fat goes in the bin. Rendering happens only if it is on the prep list with a quantity next to it. "Save the fat" is not a system. "Render 2 quarts, Tuesday and Friday, label and date" is.

The butchery is the first thing cut on a short-staffed morning — and it should not be, because everything downstream of it fails silently three days later, at which point nobody connects it to the Tuesday somebody called in sick. Chapter 14 builds the prep system that prevents this.

The limits, which are real

Every method in this book states what it cannot do. Cross-utilization has four failure modes.

It concentrates supply risk. Where fifty-two chickens reach five items, one bad poultry delivery — wrong size, temperature abuse, short count — takes out five items at once. Chapter 13's receiving discipline is the countermeasure and it is not optional in a cross-utilized kitchen. Neither is a second approved supplier for anything reaching four or more menu items.

It can make a menu taste the same. If salsa verde appears on four items, a guest ordering two of them has eaten the same sauce twice. The discipline is to cross-utilize components aggressively and finished flavors carefully: three of the salsa verde's four placements are unlikely to appear on the same check, and the broccoli version carries more chile and less caper so it does not read as a repeat.

It creates false economy on labor. Breaking down whole animals is skilled work, and Chapter 11 shows you how to test whether the cost saved justifies the hours spent. Sometimes it does not — Bellwether buys ribeyes portion-cut for exactly this reason, and buying that one item "badly" is a decision, not an oversight.

It fails at very small volume. Cross-utilization needs throughput. Sell four chickens a week and the stock program is a fantasy and the livers spoil. A pop-up or a truck (Chapter 30) cross-utilizes differently — fewer programs, used harder — and should not copy a restaurant's map.

10.4 Menu size: the real cost of every additional item

Item count is the number of distinct items offered on a menu. It is the single most consequential number on the document, and virtually nobody manages it.

Item count is not a matter of taste. It is a throughput variable, an inventory variable, a training variable, and a spoilage variable. And the costs it drives are almost entirely invisible on a P&L, because they do not appear as a line called "too many items." They appear as slightly high food cost, slightly high labor, slightly slow ticket times, and a walk-in that is always full. Which is to say: they appear as Chapter 1's cost drift, arriving through a design decision.

The arithmetic of one more item

Here is what an item costs before you have bought a single ingredient for it.

🧮 Run the Numbers

The twenty-third item.

A dish is proposed for Bellwether's menu. It is good. It will sell about nine a week — modest, but real. What does it cost, ignoring ingredients entirely?

Cost	The build	Annual
Prep labor	25 minutes a day × 5 days = 2 hr 5 min a week, at \$19.00/hr fully loaded = \$39.58/wk	\$2,058
Spoilage	one unique perishable component; \$14 a week discarded on average | **\$728**
Ordering and counting	5 minutes a week added to the order guide and count sheet, at \$19.00/hr | **\$82**
Training and tasting	4 menu changes a year × 10 staff × 12 minutes, at \$19.00/hr | **\$152**
Total identifiable		\$3,020

The item sells $9 \times 52 = 468$ times a year. So:

$$\frac{\$3{,}020}{468} = \$6.45 \text{ per plate}$$

Six dollars and forty-five cents of cost attaches to every plate of that dish before anyone has put food on it. Chapter 11 will tell you what the ingredients cost. Chapter 12 will tell you whether the item earns its place on the menu. What Chapter 10 can tell you is that the number is not zero, that almost nobody computes it, and that it is the reason a menu of twenty-two items outperforms the same restaurant's menu of thirty-four.

The limit of this calculation, stated honestly. Every figure above is an estimate, and the spoilage figure is the softest of them — it could plausibly be \$5 a week or \$30 depending on whether the unique component is a dry good or a delicate herb. The \$19.00 fully-loaded hourly rate is a planning figure that Chapter 19 will rebuild properly from the staffing guide. And the largest cost of all is missing from the table entirely, because it cannot be honestly estimated in advance: what the item does to ticket times at peak. That one is not small, and §10.4's second half is about it.

What item count does at 95 covers

Chapter 7 designed a kitchen and a floor plan to carry 95 covers a night at 1.4 turns on 68 seats. The menu has to be executable inside that. This is where the abstraction ends and the fire starts.

Bellwether's line at peak is four people: one on the hearth, one on sauté and pasta, one on garde manger and cold pickups, and the chef expediting at the pass and stepping into whichever station is drowning. That is a small line for 95 covers, and it is small because the labor model has to come in at 32.3% of sales. The menu is the document that decides whether four people is enough.

FIGURE 10.4 — Where every item is fired, and what that does to the line  [the Bellwether plan]

                                       share of        on a 95-cover night
  STATION           ITEMS              entrées         (86 entrées, 43 starters
                                                        and sides, 14 desserts)
  ───────────────────────────────────────────────────────────────────────────────
  HEARTH            hearth chicken       25%            38 entrées fired to order
   fired to order   ember trout          13%          + 13 starters à la minute
                    dry-aged ribeye       6%            ──────────────────────────
                    ─────────────────────────             51 fire actions a night
                    subtotal fired       44%
  ───────────────────────────────────────────────────────────────────────────────
  PLANCHA           pork chop            13%            par-cooked over wood in
   par-cooked in    hanger steak         10%            prep, seared and finished
   prep, finished   bellwether burger    11%            to order off the fire
   off the fire     ─────────────────────────
                    subtotal             34%            29 a night
  ───────────────────────────────────────────────────────────────────────────────
  SAUTÉ / OVEN      hand-cut pasta       13%            19 a night, plus every
                    squash and grains     9%            sauce and every brunch egg
                    ─────────────────────────
                    subtotal             22%
  ───────────────────────────────────────────────────────────────────────────────
  GARDE MANGER      little gems · mousse · rillettes · beets · olives · bread
                    ~22 cold starter pickups + 14 desserts = 36 plates, none of
                    which touch the fire
  ───────────────────────────────────────────────────────────────────────────────
                    TOTAL ENTRÉES       100%            86 = 95 covers × 0.90

Read the second column and the design intent becomes visible. Only three of the eight entrées fire to order on the hearth. The other five — 56% of entrée orders — were deliberately built to reach the guest without occupying a fire position during service. That was not a culinary decision. It was a capacity decision, made on a whiteboard, and it is worth about thirty-five fire actions a night.

Note what did not move: the Hearth Chicken. The signature dish stays on the fire, because it is the reason people came. Everything else was negotiable.

Now the capacity question. Does the fire carry it?

🧮 Run the Numbers

Can one fire carry this menu?

Chapter 7 does not ask the menu what it would like. It specifies the hearth, measures it, and hands the kitchen two numbers: a sustained capacity of about 28 items an hour, and the design constraint that follows from it — no more than roughly half of entrée orders may route through the fire during service. That is the box this menu has to fit inside. The equipment is bought. The fire is the size it is.

The first draft of this menu failed that test, and it failed it the way most first drafts do. Written the way a chef wants to write it — every protein over wood, because that is the concept — it routed 67% of entrées through the fire. Against a 28-item ceiling that is a kitchen running roughly a quarter over capacity in the peak hour, and it is worth understanding what that actually looks like, because it does not look like a crisis. It looks like ticket times drifting from twelve minutes at 6:00 to twenty by 8:00, tables sitting eleven minutes longer than the model assumes, and a dining room turning 1.15 times instead of 1.40. Nobody has a bad night. The plan just quietly stops being true.

The fix was production, not menu. Not one item came off the list and not one price changed. Three dishes changed how they are made rather than what they are: the pork chop and the hanger steak are now par-cooked over wood during prep and seared to order on the plancha; the squash and grains moved to the oven entirely. Fire-time routing falls from 67% to 44%.

Now the demand. Suppose 55% of the night's covers arrive in the 90 minutes between 6:45 and 8:15 — a normal shape for a neighborhood dinner house (Chapter 24 measures this properly).

• Covers in the window: $95 \times 0.55 = 52$
• Entrées: $52 \times 0.90 = 47$
• Fired on the hearth: $47 \times 0.44 = 21$
• Starters fired à la minute in the window (charred broccoli, flatbread, greens): about 7
• Total: 28 fire actions in 90 minutes = 18.7 an hour.

$$\frac{18.7}{28} = 67\% \text{ utilization at peak}$$

The verdict: this menu is executable at 95 covers, with about a third of the fire in reserve. That is a real answer and it is not a comfortable one. Sixty-seven percent means the line works. It does not mean the line is relaxed, and it does not survive much.

And notice what just happened, because it is the chapter's whole argument in one move. The menu did not get worse. The guest cannot tell. What changed is that a document written in a chef's handwriting was made to satisfy a constraint written in a mechanical engineer's — and the only reason that was possible without cutting a dish is that somebody checked before the menu was printed rather than in week three of service.

That last sentence is the honest part, and it deserves specifics. Here are the four items on this menu that are throughput risks, in order of how much trouble they can cause.

1. The hearth flatbread (\$16) — the most dangerous item on the menu. It occupies prime fire real estate for three to four minutes, it is a starter competing with entrée fire in exactly the window when entrée fire peaks, and it is the item most likely to over-perform, because it is cheap, shareable, and the natural thing a bar-first guest orders. At 20% of starter orders instead of 13%, peak fire utilization moves from 67% toward 74% and ticket times move with it. The control: a hard nightly count — the dough is prepped to a par and when it is gone it is gone. Here, 86ing an item at nine o'clock is a design feature, not a failure (Chapter 14).

2. The dry-aged ribeye (\$46) — sixteen minutes on the fire plus a six-minute rest, at 6% of entrées. Individually rare, individually enormous, and it holds an entire table's ticket hostage because the other three plates must be timed to it. The control: it fires first, the ticket is coursed around it, and the point-of-sale flags it so the expediter sees it the moment the ticket lands (Chapters 14 and 26).

3. The Hearth Chicken's dependence on par-roasting. A raw half chicken is a thirty-five-minute proposition; the item only works at volume because it is par-roasted in prep, chilled properly, and finished in eight minutes. That makes the signature dish dependent on a prep task. If prep runs short on a Friday morning — Chapter 14's Friday night is exactly that scenario — the kitchen cannot simply cook more; it must cook from raw, at four times the dwell, into a fire already at 67%. The control: the par-roast is first on the prep list every day, with a count, and it is never the task that absorbs a short morning.

4. The hand-cut pasta, if sauté is single-manned. Eleven pastas a night is not a problem. Eleven pastas plus every pan sauce plus every vegetable finish on one body is. The pasta was moved to sauté to relieve the fire; that relief is only real if sauté can absorb it. Chapter 14 builds the station and Chapter 19 staffs it.

🔍 Check Your Understanding

1. Bellwether's menu has 22 food items. If the kitchen added six more and cover count did not change, what happens to the average number of times each item sells per week, and name two consequences that follow.
2. The burger and the pasta are 24% of entrée orders and neither fires on the hearth. If both had been designed as hearth items, roughly how many additional fire actions would the peak 90-minute window carry?
3. Why is the flatbread a bigger throughput risk than the ribeye, even though the ribeye occupies the fire four times as long?

(1: Total item-sales are fixed by covers, so spreading them over 28 items instead of 22 cuts the average item's weekly volume by roughly 21%. Consequences: slower inventory turnover and therefore more spoilage; each item's sales signal gets noisier, so Chapter 12's analysis needs a longer window to say anything; and more items sit below the volume at which a cook builds real muscle memory. 2: The window carries 46 entrées; 24% of those is 11 more items on the fire, taking peak fire actions from 46 to 57 in 90 minutes — 38 an hour against a 45-an-hour practical ceiling, or 84% utilization. The line would still function and would have almost no margin for a re-fire. 3: Volume and timing. The ribeye is 5–6 plates a night and its dwell is predictable and plannable; the flatbread is 6 a night at plan but is a cheap shareable starter with real upside variance, and it competes for the fire during the entrée peak rather than before it.)

10.5 Standardized recipes: why the same dish must be the same dish

A standardized recipe is a written specification for a menu item — components with exact quantities, method, plating, hold and reheat instructions, portion size, and a statement of what the finished dish should look and taste like — precise enough that any competent cook in your kitchen produces the same result on the four hundredth repetition as on the first.

It is the least glamorous document in this book and it is very close to the most important. Here is why, in one sentence: an unwritten recipe is a recipe that changes, and a recipe that changes is a food cost that changes. Chapter 1 listed the components of cost drift — a cook plating six ounces instead of five because it looks better, a new prep cook wasting a little more trim, a special that was never costed. Every one of those is a standardized-recipe failure, and together they were three to five points of prime cost.

There is a second reason, which operators discover later and value more. A standardized recipe is how a chef stops being the only person who can make the food. It is the difference between a restaurant and a job, and it is the first document you will need if you ever want a day off, a general manager, or a second location (Chapters 35 and 37).

What goes on the card

FIGURE 10.5 — Standardized recipe: HEARTH CHICKEN                    [the Bellwether plan]
  ────────────────────────────────────────────────────────────────────────────────
  MAKES              2 portions — one whole air-chilled bird, 3.5 lb, split
  STATION            Hearth · par-roast in prep, finish to order
  MENU PRICE         $29
  ────────────────────────────────────────────────────────────────────────────────
  BUILD                                        per portion
    Chicken          half a 3.5 lb air-chilled bird, dry-brined and air-dried
    Roasted roots    6 oz — carrot, parsnip, celery root, turnip, cut to 1 inch
    Salsa verde      1.5 fl oz
    Butter and       0.75 oz butter · 2 thyme sprigs · 1 crushed garlic clove
      aromatics
    Oil, salt,       to season
      pepper
    Garnish          3 g picked herb · a pinch of flake salt
  ────────────────────────────────────────────────────────────────────────────────
  PREP METHOD
    1. Dry-brine whole birds at 2% salt by weight, 12–24 hours, uncovered, on a
       rack in the walk-in.
    2. Split, remove backbone and wing tips (→ stock), pull liver (→ mousse),
       reserve fat (→ render).
    3. Par-roast halves in the cooling hearth to 145°F at the thigh.
    4. Cool through the two-stage cooling requirement, cover, label, date.
       Hold at or below 41°F. Use within 3 days.
    5. Roots: toss in rendered fat and salt, roast to just tender, cool, hold cold.
  ────────────────────────────────────────────────────────────────────────────────
  SERVICE METHOD
    1. Fire skin-down on the grate, 4 minutes.
    2. Move to the roasting side, 4–5 minutes, basting twice with butter and
       aromatics. Roots go on beside it at the second baste.
    3. Finish to 165°F internal at the thickest part of the thigh. CHECK IT.
    4. Rest 3 minutes on the warm shelf before it goes on the plate.
  ────────────────────────────────────────────────────────────────────────────────
  PLATE
    Roots down, slightly off-center. Chicken skin-up, leaning on the roots.
    Salsa verde over the LEG ONLY — never over the breast skin, it goes soft.
    Herb, flake salt, and a clean rim.
  ────────────────────────────────────────────────────────────────────────────────
  WHAT GOOD LOOKS LIKE
    The skin crackles under the back of a spoon. The leg pulls off the bone with
    no resistance. The sauce reads green, not brown. And there is a spoonful of
    salsa verde left on the plate when it comes back — if there is not, we sent
    it out short.
  ────────────────────────────────────────────────────────────────────────────────
  TICKET TIME TARGET   14 minutes from fire
  86 THRESHOLD         hold back 6 portions for the last seating
  ────────────────────────────────────────────────────────────────────────────────
  This card carries no dollar figures. The cost card that does is Chapter 11's.

Notice what that card contains that a home recipe does not: a station, a portion in weight rather than in adjectives, a hold instruction, a temperature you are legally obliged to hit, a ticket-time target, an 86 threshold, and — the field most kitchens omit and most need — a plain description of what correct looks like, written so a cook in week two can self-check without asking anyone.

Notice also what it does not contain: money. The recipe tells the kitchen what to make; the cost card tells the business what it costs. They share quantities, which is exactly why the recipe must be written in weights and fluid ounces rather than in "a handful" — Chapter 11 cannot cost a handful.

⚖️ Code and Compliance

Par-cooking is a legitimate technique and a controlled one.

The Hearth Chicken is only executable at 95 covers because it is par-roasted in prep and finished to order. That is standard practice in wood-fire kitchens and it is entirely acceptable — but it puts poultry through cooking, cooling, cold holding, and reheating, which is four opportunities to get it wrong.

Under the framework of the FDA Food Code, which states and localities adopt in varying versions and with local amendments, the relevant controls are: cooking poultry to 165°F; cooling cooked food through the two-stage requirement (from 135°F to 70°F within two hours, and to 41°F or below within a further four hours, six hours total); cold holding at or below 41°F; and reheating for hot holding to 165°F. The Bellwether card hits 165°F on the finish, which is the step that matters, and the par-roast is deliberately taken only to 145°F so the finish does real work rather than merely warming a cooked bird.

Two more items on this menu carry specific obligations worth naming here rather than discovering later. The burger is ground in house from hanger trim, chuck, and brisket — which is a genuine cross-utilization win and also means the restaurant is now grinding raw beef, with the record-keeping, equipment sanitation, and cooking-temperature requirements that attach to ground product rather than to whole-muscle steak. If the restaurant intends to serve it below the required temperature, that is a consumer-advisory question and in some jurisdictions a variance question. The dry-aged ribeye and the hanger steak are whole-muscle and are treated differently, which is a distinction a new cook will not know unless somebody teaches it.

All of this varies by state, county, and city, and all of it changes. Chapter 25 builds the food safety plan properly. Verify your local code before you write a card, and have your health department's plan reviewer look at your process — most of them would much rather advise you in advance than write you up afterward.

⚠️ Where the Money Leaks

The four-ounce that became six.

A standardized recipe that is written and then never enforced is a filing cabinet, not a control. The drift is always in one direction — portions grow, never shrink — because a cook who is unsure adds a little, and a cook who likes a guest adds a little, and nobody ever got yelled at for a generous plate.

Take the roasted roots on the Hearth Chicken: the card says six ounces. Suppose the actual average creeps to seven and a half — a spoonful, invisible on the plate, indistinguishable to the guest. That is a 25% overage on that component, on an item selling about 105 times a week and 5,460 times a year.

Chapter 11 will put the dollar figure on that spoonful and it is larger than you expect. What Chapter 10 can tell you is where the fix lives, and it is not in a memo. It lives in three things: a portioning tool that makes the correct portion the easy portion, a plated example at the start of service, and a weekly five-minute spot check where the manager weighs three plates before they leave the pass. The last is the one nobody does and the one that works. You can start it in week one for the price of a \$40 scale.

10.6 Menu psychology and layout: sequence, panels, anchors, and what the eye does

Now we come to the part of the subject that requires the most care, and I want to be direct about why.

Menu psychology is the study of how a menu's design, sequence, and language influence what a guest chooses and how much they spend. It is a real field, it contains real findings, and it is also — more than any other topic in this book — a body of confidently repeated claims whose evidence is thinner than the confidence. Chapter 1's first case study was about exactly this failure mode: a number everybody knows, cited by credentialed people, traceable to nothing. The menu-design literature has several of those, and you are going to be sold advice built on them.

So let us separate three categories: what is well supported, what is plausible but under-measured, and what is folklore.

What is well supported

People scan menus; they do not read them. Every host and manager can confirm this from a thousand hours of watching. The consequence is practical: set type a fifty-five-year-old can read in a room at 30% light, keep descriptions short, and do not bury the thing you want sold inside a paragraph.

Primacy and recency effects are robust. Across a very wide range of settings, people weight the first and last items in a list disproportionately — one of the oldest and most replicated findings in memory research, and not specific to restaurants. So the first and last positions in every category are your most valuable real estate, and you should decide what occupies them rather than sorting alphabetically and letting the decision happen to you.

Visual emphasis draws attention. Boxing an item, lengthening its description, or setting it in a different weight makes people look at it. That much is close to tautological. How much of the looking converts into ordering is another question.

Choice difficulty is real and its size is contested. A genuine experimental literature suggests very large choice sets can reduce satisfaction and even purchasing — the jam-tasting study is the famous example — and a substantial body of replication work finds the effect smaller, context-dependent, and sometimes absent. Honest summary: it probably exists somewhere, its magnitude is uncertain, and you should not build a business case on a specific number. You do not need to. §10.4 gave you four operational reasons to keep item count low that owe nothing to psychology.

What is plausible but under-measured

Price anchoring is the effect by which the first or most extreme price a person sees shapes their judgment of every subsequent price. The underlying psychology is well established — anchoring and adjustment is canonical work in judgment and decision-making research and it replicates broadly. What is not established is how much survives in a real dining room, with a real guest who has a budget, a companion, and a prior expectation of what dinner costs here.

Bellwether's menu has an obvious anchor: the dry-aged ribeye at \$46, ten dollars clear of the next item. The theory says its presence makes \$29 feel moderate, and the theory may be right. But the ribeye earns its place for reasons that do not require the theory at all — the celebration occasion is 30% of covers at a \$57 per-person average and the destination guest another 10% at \$50, and both want a large steak to exist. If anchoring is real, Bellwether captures it for free. If it is not, the item still sells five or six a night to people who came for it. Design so that you win either way and you will never need to know which explanation was correct.

Descriptive language. Advice to write evocative, provenance-heavy descriptions is usually supported by pointing at food-psychology research on descriptive menu labels. Readers should know that a substantial portion of the most-cited work in that specific area has been subject to correction and retraction in recent years, and that anyone quoting a precise percentage lift from a menu-description study is quoting something you should check before repeating. The practical guidance survives, but in a different shape: write descriptions the way your staff would actually say them out loud, because what reliably sells food is a server who can describe it, not an adjective on paper. Chapter 18 makes that a training subject.

What is folklore

The "golden triangle." You will be told, confidently, that the eye moves first to the center of an open menu, then top right, then top left, and that you should place your most profitable items in that triangle. It appears in trade publications, consulting decks, and design courses. Try to trace it to a study and you find citations pointing at other citations pointing at a consultancy's unpublished work from decades ago; later eye-tracking research has not consistently reproduced the geometry. I am not telling you it is wrong. I am telling you it is not established — and the difference between "not established" and "known to be true" is what this entire book is about.

Removing dollar signs increases spend, usually attributed to a hospitality-school study. Such work exists; the reported effects were modest, the samples small, and it has been repeated far more often than replicated. Prices ending in .95 signal value and whole numbers signal quality is widely believed, weakly evidenced in restaurants specifically, and probably swamped by everything else the room is telling the guest.

🤝 Hospitality

The menu is the first conversation, and most of it is happening before anyone speaks.

Strip out the psychology entirely and a menu is still doing something enormous: it is telling a guest whether this room was built with them in mind.

A guest with a shellfish allergy reads a menu looking for whether it is safe to ask. A vegetarian reads it looking for whether they are a guest or an accommodation. A guest who is watching what they spend reads it looking for a price they can order without embarrassment in front of the person across the table. A guest who does not know what mostarda is reads it hoping there is a way to find out that does not make them feel foolish. And a guest with tired eyes reads it in a dim room hoping the type is big enough.

All five of those readings happen in the first twenty seconds, before a server says a word, and Bellwether's menu answers all five deliberately. The vegetarian item is an entrée in the main panel. The allergen line is written in the restaurant's own voice — "tell us and we will tell you honestly what this kitchen can and cannot do" — rather than as a legal disclaimer, because the honest version invites the conversation and the legal version ends it. There is a \$21 entrée and a \$7 snack, so the range of ways to be a guest here is wide. And the type is set to be read.

None of that is on any list of menu-engineering techniques, and all of it is worth more than the golden triangle. You sell hospitality, not plates — and the menu is the first place a guest finds out whether you meant it. Chapter 23 works out what that is worth in return visits.

The panel and the sweet spot, used honestly

Two terms you will hear constantly. A menu panel is a bounded visual region of the menu — a column, a boxed section, or one side of a folded card — that the eye takes in as a unit. The sweet spot is the position within a panel that is claimed to receive the most attention, conventionally the top of the right-hand panel or the first line of a category.

Use both terms, and use them with the appropriate confidence. The panel is real and useful: it is how you control what is compared with what. Putting the hearth items and the stove items in separate panels means a guest compares chicken against pork rather than chicken against pasta, which is a genuine design lever regardless of any psychology.

The sweet spot is a hypothesis. Bellwether puts the Hearth Chicken first in the hearth panel with the longest description on the menu, and puts the ribeye last. It does not box it, star it, or write "chef's favorite" beside it — partly on brand grounds (Chapter 3), partly because a menu with one starred item invites the guest to notice they are being sold to.

Here is the honest argument for the placement, and it is different from the usual one: it costs nothing, it does no harm, the item deserves the position on its merits, and if the effect is real you have captured it. That reasoning is available to you whether or not the research holds up. The reasoning that says "studies show the top-right position increases sales by X percent" is not, and you should not use it, and you should be a little suspicious of anyone who does.

10.7 Pricing signals: whole numbers, decoys, and honesty

Chapter 11 prices from cost. This section prices from the guest's side — the shape of the list, the signals the numbers send, and the arithmetic that has to reconcile.

A price ladder is the ordered set of prices within a category, considered as a structure rather than as individual numbers. It is the thing a guest actually perceives. Nobody reads eight prices; they perceive a range, a middle, and an outlier, and they locate themselves inside it.

FIGURE 10.6 — Bellwether's entrée price ladder                       [the Bellwether plan]

   21  ██████            bellwether burger            11% of entrées
   24  ███████           squash and grains (v)         9%
   25  ███████▌          hand-cut pasta               13%
   28  ████████▌         ember trout                  13%   ┐
   29  █████████         HEARTH CHICKEN               25%   │  the working middle:
   31  █████████▌        pork chop                    13%   ┘  73% of entrées, $24–$31
   36  ███████████       hanger steak                 10%
   46  ██████████████    dry-aged ribeye               6%   ← the anchor

   Lowest $21 · highest $46 · ratio 2.19 · weighted average $29.00 · median $28.50
   Largest gap: $10, between the hanger and the ribeye — deliberate.
   Smallest gaps: $1, between 24/25 and 28/29 — also deliberate.

Four things about the shape of that ladder, each of which is a decision.

The middle is tight and crowded. Five of the eight entrées sit between \$24 and \$31, and they carry 73% of entrée orders. That is on purpose. When adjacent items are a dollar apart, the guest stops choosing on price and starts choosing on desire, which is where you want the decision to happen. It also means a single price change of a dollar or two does not re-rank the menu.

The bottom is a real item, not a loss leader. The burger at \$21 exists to serve the bar-first occasion, which Chapter 2 established at 14% of dinner covers and a \$36 per-person average. A menu whose cheapest entrée is \$24 has quietly told that guest to go to the taproom.

The top is ten dollars clear. The ribeye is the anchor discussed in §10.6, and it is also the item that carries the celebration table and pulls a bottle of wine with it (Chapter 16).

The weighted average and the median nearly agree. The median entrée price is \$28.50; the weighted average is \$29.00. When those two numbers diverge sharply, one item is carrying your menu and you should know which one — here they agree because the Hearth Chicken sits almost exactly at the center of the ladder and takes a quarter of the mix. That is a healthy shape.

The conventions Bellwether uses, and the honest reason for each

Every one of those justifications is operational or aesthetic. Not one of them requires a study to be true. That is the test to apply to any menu-design convention somebody sells you: if the only argument for it is a research finding, ask to see the finding.

Does the ladder produce the check?

Chapter 2 did not assert Bellwether's \$46 dinner check; it built it, from four occasions in a specific ratio, and then decomposed it into \$33.12 of food and \$12.88 of beverage — the frozen 72/28 mix. That decomposition is a set of claims about what guests order. This menu has to be capable of producing it, and "capable" means the arithmetic resolves.

Here is the reconciliation, computed from the actual prices on Figure 10.2.

Figure 10.7 — Does the menu produce \$33.12 of food per guest? (the Bellwether plan)

The attachment rates are Chapter 2's. The weighted average prices are outputs of this menu — and they only land on \$29.00, \$12.00, and \$10.80 if the assumed share of orders within each category is what the plan says. Here is the entrée column in full, since it is the one carrying \$26.10 of the \$33.12:

And the starters-and-sides column, whose eleven items have to average \$12.00:

Desserts: the crisp at \$11 (45%), the budino at \$11 (35%), and the soft-serve at \$10 (20%) give $0.45(11) + 0.35(11) + 0.20(10) = 4.95 + 3.85 + 2.00 = \$10.80$.

Every column resolves. But a table of assumed shares is a set of beliefs, and beliefs should be tested against something concrete. So here is the same claim from the other end — one Thursday's first seating, table by table, with actual orders off the printed menu.

🧮 Run the Numbers

Twenty-five covers, food only.

	Table	Covers	What they ordered (food)	Food
1	two-top, weeknight	2	bread 7 · little gems 12 · burger 21 · chicken 29	\$69
2	four-top, celebration	4	flatbread 16 · little gems 12 · ribeye 46 · chicken 29 · pasta 25 · trout 28 · greens 9 · crisp 11	\$176
3	two at the bar, drop-in	2	bread 7 · broccoli 14 · olives 8 · little gems 12	\$41
4	two-top, destination	2	rillettes 15 · trout 28 · chicken 29 · roots 10 · crisp 11	\$93
5	six-top, neighborhood group	6	olives 8 · flatbread 16 · squash 24 · pasta 25 · chicken 29 · chicken 29 · burger 21 · trout 28 · potatoes 10	\$190
6	two-top at the bar	2	little gems 12 · broccoli 14 · burger 21	\$47
7	four-top, weeknight	4	bread 7 · little gems 12 · chicken 29 · burger 21 · pasta 25 · trout 28	\$122
8	three-top, weeknight	3	olives 8 · chicken 29 · trout 28 · pasta 25	\$90
	Total	25		\$828

$$\frac{\$828}{25} = \mathbf{\$33.12} \text{ of food per cover}$$

Look at how wildly the individual tables vary. Table 2 spent \$54.00 a head on food; table 3 spent \$20.50 and did not order an entrée at all — which is precisely the bar-first guest Chapter 2 described, and precisely why the entrée attachment rate in the plan is 0.90 rather than 1.00. Across these twenty-five covers there were 22 entrées, an attachment of 0.88. Table 6 spent \$23.50 a head. Not one table hit \$33.12, and the average is \$33.12 exactly.

Now the honest caveat, because this is the kind of figure that flatters an author. One seating is not a mix. Seven of those twenty-two entrées were Hearth Chickens — 32%, against the 25% the plan assumes. On one night that is noise. On a quarter of point-of-sale data it would be a finding, and it would change the weighted average entrée price and therefore the check. That is exactly the distinction Chapter 12 is built on, and it is why Chapter 12 refuses to draw a matrix until it has enough weeks of data to mean something.

That reconciliation is the thing this chapter owed the plan. The \$46 check is not an aspiration sitting on top of a menu that cannot produce it; it is an arithmetic consequence of twenty-two prices and a set of ordering assumptions, all of which are now written down and all of which are falsifiable from day one of point-of-sale data (Chapter 26).

Where the beverage attaches

Chapters 15 and 16 own the other \$12.88. But the food menu determines where the beverage program can reach, and three attachment points were designed in rather than hoped for.

• The bar snack ladder. Hearth bread at \$7 and olives at \$8 exist so the bar-first guest has something to order alongside a cocktail without committing to a meal. That guest is 14% of dinner covers, the highest-frequency guest in the building, and their check is beverage-weighted.
• The flatbread as the bar's anchor food item. It shares, it eats with a drink, and it is the single most likely thing a two-top at the bar orders. It is also the throughput risk from §10.4, which is the trade.
• The large format pulls bottles. The ribeye and the pork chop are the items a table orders when it is going to drink a bottle rather than glasses, and Chapter 16 will build the by-the-glass and bottle ladders to sit alongside them.

10.8 Seasonal menus, specials, and the change cadence

A short seasonal menu was part of Bellwether's concept from Chapter 2, which means it is a promise, which means somebody has to actually change the menu on a schedule. The gap between "seasonal" as a brand adjective and "seasonal" as an operating cadence is where a lot of concepts quietly die.

The change cadence

FIGURE 10.8 — Bellwether's menu change calendar                      [the Bellwether plan]

   FOUR SEASONAL REBUILDS A YEAR — the printed menu changes substantially
     early March · late May · early September · late November
     4–6 items replaced · every new item costed BEFORE it prints (Chapter 11)
     every new item tasted by the whole staff before it sells (Chapter 18)
     ─────────────────────────────────────────────────────────────────────────

   WEEKLY REPRINT — same items, new date, and the ability to move a price
     every Tuesday · in-house, on the locked template · about 120 menus
     the crisp fruit changes with no reprint required ("the fruit that is
     good this week") — one line written to absorb weekly change
     ─────────────────────────────────────────────────────────────────────────

   NIGHTLY SPECIALS — 1 or 2, never more
     off the fire, from what came in, costed on the card before service
     announced at pre-shift with a price and a description (Chapter 21)
     ─────────────────────────────────────────────────────────────────────────

   THE ANNUAL RE-PRICE — one deliberate pass, every twelve months
     every item re-costed against current invoices, prices moved together
     (Chapter 11 does the costing; Chapter 12 says which items move)

Four rebuilds a year is a deliberate middle setting. A menu that changes monthly never lets the kitchen get good at anything, never lets a server learn to describe a dish, and never accumulates enough sales data for Chapter 12's analysis to say anything. A menu that changes once a year is not seasonal, whatever it says on the front. Four is enough to honor the claim and slow enough that the kitchen achieves mastery — and mastery is a cost control, because a cook who has made a dish six hundred times wastes less of it.

The weekly reprint is the piece most restaurants skip and should not. About fifty dollars a week on a locked in-house template buys the ability to move a price on a Tuesday afternoon when a protein contract renews, and it puts a date on every menu so that six months from now somebody can answer "when did we last change this?" Chapter 3 costed the alternative and found that a beautiful \$900 print run pays for itself in delay after fewer than six weeks.

⚠️ Where the Money Leaks

The special nobody costed.

This is the single most common uncontrolled cost in an otherwise well-run independent restaurant, and it hides perfectly, because specials do not appear on the printed menu that anybody audits.

The mechanism: a chef writes a special at four o'clock from what looks good and what needs to move. It is priced by feel — "call it thirty-two." Nobody costs it. Multiply by three specials a week, fifty-two weeks a year: 156 items a year priced by intuition, several of them the most expensive proteins in the building, because that is what specials tend to be.

Notice the selection effect. Specials are frequently the highest-cost items on the menu and they are the only items nobody costs. That is why "our food cost is running a little high and we cannot see why" is such a common sentence.

What the disciplined operator does: a special gets a cost card before it gets a price, exactly like a printed item — eight minutes on a template once the cards exist (Chapter 11). It gets written on the pre-shift board with its price and its description, so servers sell it accurately. And it gets counted, so it can become a candidate for the next seasonal rebuild instead of vanishing.

The best restaurants I know treat the specials board as a research and development pipeline, where next spring's menu items audition in front of real guests at a real price. That only works if somebody writes down what happened.

Brunch: the same kitchen, the same walk-in, a different menu

Bellwether serves brunch on Saturday and Sunday at a \$24 average check — about 110 covers a service, 220 a week, 11,440 a year. It is a second daypart with its own guest, its own labor model, and its own economics (Chapter 2 defined daypart; Chapter 24 manages it).

The design rule for a second daypart is simple and almost always violated: brunch must be built from the dinner menu's programs, or it will double your purchasing. Here is the menu, and then the map of what it borrows.

FIGURE 10.9 — BELLWETHER, brunch                                     [the Bellwether plan]
                      Saturday and Sunday, 9:30 to 2:00

  Hearth eggs, greens, cream, grilled bread                                     16
  Chicken hash, ember roots, soft egg, salsa verde                              18
  Buttermilk biscuits, pork sausage gravy                                       16
  Hearth flatbread, farm egg, greens, fontina                                   17
  Smoked trout toast, soft egg, pickled onion                                   18
  Grain bowl, ember vegetables, yogurt, herbs                            (v)    15
  Bellwether burger, aged cheddar, pickles, hearth potatoes                     21

  ALONGSIDE
  Two eggs, any style                                                            5
  Pastry, one — whatever came out of the hearth this morning                     6
  Seasonal fruit                                                                 6
  Hearth potatoes, garlic, rosemary                                              7
  Thick-cut bacon                                                                8

  Coffee 4 · Pot of tea 6 · Juice 3
  The brunch cocktail list is on the back.

Twelve lines. Count the new ingredient families it introduces: eggs, bacon, and yogurt. That is all. Everything else on that menu comes off the map in Figure 10.3 — the chicken program, the bread dough, the pork, the trout, the roots, the dairy, the beef grind, the grains, and the salsa verde.

And it reconciles the same way the dinner menu did:

Figure 10.10 — The \$24 brunch check (the Bellwether plan)

Food is \$17.28 of \$24.00 — 72.0% — holding the plan's mix in the second daypart as well as the first. The seven main plates weight to \$17.00 exactly: hearth eggs \$16 at 22%, chicken hash \$18 at 18%, biscuits \$16 at 16%, flatbread and egg \$17 at 12%, trout toast \$18 at 8%, grain bowl \$15 at 14%, and the burger \$21 at 10%.

One deliberate inconsistency worth naming, because a reader will spot it: hearth potatoes are \$10 at dinner and \$7 at brunch. Same prep, smaller portion, different price. That is a real operator decision with a real risk attached — a guest who eats both dayparts will notice — and the defense is that the portion genuinely differs and the menus never appear on a table at the same time. The alternative, holding one price across dayparts, is also defensible. What is not defensible is having it happen by accident, which is how most daypart price inconsistencies arise.

🔍 Check Your Understanding

1. Bellwether's brunch introduces only three new ingredient families. Name the four dinner programs the brunch menu leans on hardest, and one brunch item from each.
2. Why does the chapter argue for four menu rebuilds a year rather than twelve, given that the concept promises "seasonal"?
3. A chef proposes running three specials a night instead of one. Give two reasons to refuse that have nothing to do with the food.

(1: The poultry program → chicken hash; the bread dough → flatbread and egg, pastry, toast; the pork program → sausage gravy; the trout program → trout toast. The dairy and root programs are equally defensible answers. 2: Twelve changes a year never lets the kitchen achieve mastery, never lets servers learn to describe a dish, never accumulates enough sales data for menu engineering to say anything, and multiplies the costing and training work by three. Four is frequent enough to honor the claim and slow enough to get good. 3: Three specials a night is 468 uncosted items a year instead of 156 — three times the uncontrolled-price exposure; it triples the pre-shift briefing load and therefore the chance a server describes an item wrong; and it adds fire actions at peak that were not in the throughput math in §10.4.)

🍽️ The Business Plan

Checkpoint 10 of 40 — The Menu (draft).

Nine chapters produced a concept, a brand, a forecast, a capital stack, a lease, and a kitchen. This one produces the thing the restaurant actually sells. Everything before this was a claim about a market; this is the first section of the plan a reader can hold in their hand and say yes, I would eat there, or no, I would not.

What goes in the plan now:

Section 10 — The Menu (draft) (the Bellwether plan; all figures constructed)

10a. The dinner menu. Twenty-two food items across four categories — eight to start, eight mains split into hearth and stove panels, three sides, three desserts. Reproduced as Figure 10.2. Dinner Tuesday through Saturday, 5:00 to 10:00.

10b. The signature. The Hearth Chicken — half a wood-fired air-chilled bird, roasted roots, salsa verde — at \$29, first in the hearth panel, carrying an assumed 25% of entrée orders and reaching five menu items across two dayparts through the poultry program. It is the dish the restaurant is for.

10c. The brunch menu. Twelve lines at a \$24 check, Saturday and Sunday, introducing only three new ingredient families (eggs, bacon, yogurt). Reproduced as Figure 10.9.

10d. Cross-utilization. Eight purchasing programs feed thirty-four menu lines from roughly 54 lines on the order guide (Figure 10.3). Seven of the eight reach four or more menu items. The map, not the menu, is the input to Chapter 13's purchasing system.

10e. The check reconciliation. The menu produces Chapter 2's frozen check build: entrées at 0.90 attachment and a \$29.00 weighted average, starters and sides at 0.45 and \$12.00, desserts at 0.15 and \$10.80 → **\$33.12 of food per dinner cover, against \$12.88 of beverage, for the \$46.00** dinner check and the 72/28 mix. Brunch reconciles the same way at \$17.28 food and \$6.72 beverage.

10f. Throughput. At 95 covers, the menu produces about 86 entrées, 43 starters and sides, and 14 desserts. Sixty-seven percent of entrées fire on the hearth; the burger and the pasta — 24% — were deliberately built off it. Peak fire utilization is about 67% against Chapter 7's 28-item ceiling. Named throughput risks: the flatbread, the ribeye, the chicken's par-roast dependency, and a single-manned sauté station.

10g. Change cadence. Four seasonal rebuilds a year, a weekly in-house reprint on a locked template, one or two nightly specials — each costed before it is priced — and one deliberate annual re-price.

What this checkpoint settles. The plan now has a product. More usefully, it has a product whose arithmetic connects to the rest of the document: the price ladder produces the check average, the check average produces the revenue line, the item count and the station map are executable inside the kitchen Chapter 7 designed, and the purchasing programs are few enough to fit the storage the lease provides. Twenty-two items is a defensible answer to "can four people on a line do this at 95 covers?" and the chapter has shown the arithmetic rather than asserting it.

What it does not settle. This is a menu, not a business.

• Nothing here costs anything yet. Not one plate cost, not one food cost percentage, not one contribution margin. The 30% food-cost line remains an assertion until Chapter 11 builds the cost cards, and some of these twenty-two prices may well turn out to be wrong.
• The mix is assumed, not measured. Every share in Figures 10.6 and 10.7 is a belief. If the Hearth Chicken runs at 32% rather than 25% — as it did on the sample Thursday — the weighted average entrée price and the check both move. Chapter 12 replaces these assumptions with point-of-sale data and should be expected to change some of them.
• The volumes have not been reconciled to purchasing. Fifty-two birds a week is an output of the mix assumption. Chapter 13 sets par levels and finds out whether the map survives contact with real delivery schedules and minimum order quantities.
• Nobody has cooked this menu yet. Ticket-time targets on a recipe card are hypotheses. Chapter 9's menu rehearsal and Chapter 14's ticket-time measurement are where they become facts.
• The vegetarian entrée is a single point of failure, carrying an entire occasion at an assumed 9% of entrée orders. At 4% the menu has under-served a segment; at 15% the squash-and-grains prep is undersized. We will not know until we are open.
• The \$139,240 revenue bridge from Chapter 4 is untouched here. Takeout was not considered in the menu design at all and several of these items travel badly. Chapter 28 owns that question and will have to answer it against this menu.

Open questions carried forward:

1. What does each of these twenty-two plates actually cost, and are any of the prices wrong? (Chapter 11)
2. Does the assumed mix hold, and which items are carrying the menu financially rather than popularly? (Chapter 12)
3. Can eight purchasing programs actually be bought this way, at these volumes, from this district? (Chapter 13)
4. Can four people on a line execute 86 entrées with 67% peak fire utilization, at the labor the plan funds? (Chapters 14 and 19)
5. Does the beverage program deliver \$12.88 a cover against a menu built to attach it? (Chapters 15 and 16)
6. Which of these items travel, and what does the menu look like if takeout becomes a real channel? (Chapter 28)

Conclusion

A menu is not a list of food. It is a production schedule, a purchasing specification, a labor model, a brand statement, and a sales instrument, and it commits you in all five registers simultaneously. Both halves of prime cost — the number Chapter 1 said predicts survival — are set by this document before a single cook is hired or a single case is delivered.

Bellwether's menu is twenty-two dinner items and twelve brunch lines. It is short because short is executable, and §10.4 put a number on the alternative: about \$3,020 a year of identifiable non-ingredient cost, or \$6.45 a plate, for one more item selling nine a week — before you have bought an ingredient for it. It is cross-utilized because eight purchasing programs feeding thirty-four menu lines is the structural mechanism that makes a thirty-percent food cost achievable without buying cheap food. And it is load-balanced: two of the eight entrées were deliberately moved off the fire, which is the difference between 67% peak utilization and something the line cannot survive.

The signature is named and placed. The Hearth Chicken at \$29 — half a wood-fired air-chilled bird, roasted roots, salsa verde — sits first in the hearth panel, carries a quarter of entrée orders, and reaches five menu items across two dayparts. We have said what it costs the guest. We have not said a word about what it costs us, and that restraint is deliberate: a chapter that priced from cost while pretending to price from the menu would have taught you the wrong sequence.

Two of the book's themes did the work. The food is the easy part — three of the four items Bellwether cut were rejected for reasons with nothing to do with flavor, and the throughput analysis in §10.4 is a manufacturing problem wearing an apron. And prime cost is the number that keeps you open — because the menu, uniquely among the documents in this plan, sets food cost and labor cost at the same time, with the same pen.

We were careful about one more thing. The menu-psychology material in §10.6 is exactly the kind of confidently repeated advice Chapter 1 warned you about, and the honest position is uncomfortable: the golden triangle is not established, the dollar-sign finding is weakly supported, and some of the most-quoted research on descriptive menu language has been corrected or retracted. Bellwether still puts its signature first in the panel and its anchor last — but for reasons that hold whether or not the psychology does. Design so that you win either way.

Chapter 11 turns all of this into money. Every quantity on the recipe card in Figure 10.5 becomes a line on a cost card, the Hearth Chicken becomes eight dollars and fifty-two cents of plate, and the plan's thirty-percent food-cost line stops being an assertion and becomes something you can check. It may be the most important chapter in this book, and it begins with the observation that most operators guess at their prices — and that the guess is exactly where the margin goes.

Key Terms

Menu (as a document) — the restaurant's operating specification expressed as a list of things a guest can order; simultaneously a sales instrument, a production schedule, a purchasing specification, a labor model, and a brand statement. (Ch. 10)

Menu categories — the named groupings a menu is organized into; at once a pacing instruction, a production grouping, and the unit at which sales are later analyzed. (Ch. 10)

Cross-utilization — the deliberate design of a menu so each purchased ingredient reaches several items, and so the parts a careless kitchen discards become items of their own. The primary structural mechanism for a low food cost without cheap food. (Ch. 10)

Item count — the number of distinct items on a menu; a throughput, inventory, training, and spoilage variable whose costs are largely invisible on a P&L. (Ch. 10)

Standardized recipe — a written specification giving components with exact quantities, method, plating, hold and reheat instructions, portion size, and a description of what correct looks like, precise enough that any competent cook produces the same result every time. Carries no dollar figures; the cost card does. (Ch. 10)

Menu psychology — the study of how a menu's design, sequence, and language influence what a guest chooses and how much they spend; a real field containing both well-supported findings and a substantial amount of confidently repeated folklore. (Ch. 10)

Price anchoring — the effect by which the first or most extreme price a person sees shapes their judgment of every subsequent price. The psychology is well established; its magnitude in a real dining room is not well measured. (Ch. 10)

Menu panel — a bounded visual region of a menu that the eye takes in as a unit, and therefore the device that controls what is compared with what. (Ch. 10)

Sweet spot — the position within a panel claimed to receive the most attention, conventionally the top of the right-hand panel or the first line of a category. A useful convention supported by weaker evidence than its confidence implies. (Ch. 10)

Price ladder — the ordered set of prices within a category considered as a structure rather than as individual numbers; what a guest actually perceives, and the thing that must reconcile to a target average check. (Ch. 10)

Spaced Review

1. Chapter 7 turned 2,800 square feet into 68 seats and a line with a fixed number of stations. Name two ways this chapter's menu decisions were constrained by that design, and one way the menu pushed back on it.
2. Chapter 9's menu rehearsal and plate-ups happen before a single paying guest arrives. Given what §10.4 and §10.5 established, what are the two most valuable things to measure during that rehearsal, and why?
3. Chapter 2 built Bellwether's \$46 dinner check from four occasion segments in a specific ratio. If the celebration segment came in at 20% of covers rather than 30%, which items on Figure 10.2 would you expect to lose the most volume, and what would that do to the weighted average entrée price?
4. Chapter 1 defined cost drift as the slow upward movement of prime cost when nothing is measured. Identify three distinct mechanisms in this chapter through which a menu produces cost drift, and name the countermeasure for each.
5. The recurring question: the kitchen decides to stop buying ribeyes portion-cut and to break down whole primals in house instead. Does this decision move prime cost, and in which direction? What would you have to measure to find out — and when would the decision show up in the bank account, as opposed to on the P&L?