Chapter 32: Break-Even Analysis: How Many Covers Per Night to Keep the Lights On

"Ask an operator what their break-even is and you get one of three answers: a shrug, a number they heard from their accountant eighteen months ago, or — very occasionally — a figure in covers, said immediately, with no hesitation at all. The third kind of operator is almost always still open." — constructed; the pattern behind every business I have watched from close up

Overview

It is Tuesday, ten past seven, February. You are standing at the host stand looking at your room. You have four deuces seated, a four-top finishing, and two names on the book for 7:45. Twelve people are eating in a restaurant built for sixty-eight.

Here is the question that matters, and it is not "where is everybody." It is: at what point in this evening did this restaurant start making money instead of losing it? Was it the fourth table? The twelfth cover? The thirtieth? Because if you know the answer — if you can look at that room and say we are eleven covers short of paying for tonight — you are running a business. If you cannot, you are watching a room and hoping, and every slow Tuesday feels identical to every other slow Tuesday, which means you will not notice the difference between February and a structural problem until February is over and you are four months into the structural problem.

Break-even analysis answers that question. It is the arithmetic that converts a stack of cost lines into a single operating threshold, and then converts that threshold into the only unit a manager standing in a dining room can actually act on: guests through the door.

It is also, I have to say plainly, one of the most abused calculations in this industry. It gets run once, in a business plan, with a fixed/variable split that somebody guessed at in an afternoon, and then it gets quoted for years. Bellwether's own plan has already produced two break-even figures that disagree with each other — \$1,030,454 in Chapter 4 and \$986,700 in Chapter 5 — and this chapter is going to show you that both of them are wrong, for different reasons, and that the reason they are wrong is the fixed/variable split underneath them. Getting that split right is nine-tenths of the work. The division at the end takes four seconds.

By the end of this chapter you will be able to state Bellwether's break-even in dollars, in covers a year, and in covers per night by daypart — and you will be able to say what happens to all three when the labor line moves, when the first quarter runs hot on cost, and when you remember that the loan payment is not on the P&L.

In this chapter, you will learn to:

• Sort every line on a restaurant P&L into fixed, variable, and semi-variable, and split the semi-variable lines honestly instead of shoving them into whichever bucket is convenient.
• Compute the contribution margin ratio and read it as the fraction of each dollar that survives.
• Convert break-even sales into break-even covers and then into break-even covers per night, naming the revenue base at every step so you do not manufacture a phantom average check.
• Build a daypart contribution statement and answer, with numbers, whether brunch is carrying itself.
• Compute operating leverage and explain why a 10% revenue miss is a 33% profit miss.
• Compute margin of safety four ways and say what a cushion is actually worth.
• Use break-even to price a specific decision: a new service, a price change, an added position.
• Name what break-even hides — starting with the fact that it does not know cash exists.

Learning Paths

🏗️ Opening — this is the chapter your plan lives or dies on. Do §32.1 by hand on your own cost lines; the split is the whole exercise. §32.6 and §32.8 are the two sections a skeptical reader of your plan will go to first. 📋 Managing — weight §32.3, §32.4, and §32.5. If you run someone else's restaurant, break-even covers per night is the number that tells you whether tonight was a bad night or a bad business, and §32.4 is how you find the daypart nobody has ever priced. 🍸 Beverage — §32.2 is yours. Beverage carries a much better contribution margin than food, which means beverage attachment moves the CM ratio, which moves break-even for the entire building. §32.7's price-change example is a beverage argument in disguise. 🚚 Small Format — a truck has a tiny fixed base and enormous operating flexibility, which makes break-even low and operating leverage mild. §32.5 explains why that is your structural advantage — and §32.1's semi-variable discussion explains why "low fixed cost" is never as low as you think.

32.1 Fixed, variable, and the semi-variable costs that make restaurants tricky

Every cost in your restaurant behaves in one of three ways when volume changes, and the entire chapter rests on sorting them correctly.

A fixed cost does not move with sales volume within the relevant range. Rent is the pure case: you owe \$95,200 a year at Bellwether whether you serve twelve covers on a Tuesday or a hundred and thirty. Nobody at the landlord's office is watching your cover count.

A variable cost moves in direct proportion to sales. Food cost is the pure case: sell no chicken, buy no chicken. If you double covers you roughly double the protein bill, and if you close for a week the cost goes to nearly zero.

A semi-variable cost — also called a mixed cost — has both a fixed base and a variable slope. It does not go to zero when you close and it does not double when volume doubles. This is the category that makes restaurants awkward, and it is where most amateur break-even analysis quietly falls apart.

FIGURE 32.1 — The three cost behaviors                        [constructed teaching example]

  FIXED                      VARIABLE                   SEMI-VARIABLE
  cost                       cost                       cost
   │                          │                    /     │                   /
   │────────────────          │                 /        │              /
   │                          │              /           │         /
   │                          │           /              │────/
   │                          │        /                 │  ← fixed base
   └──────────── volume       └──/──────── volume        └──────────── volume

   rent, salaries,            food cost, card            utilities, repairs,
   insurance, G&A             processing, supplies       breakage, some labor

   LEGEND for every chart in this chapter: the horizontal axis is annual sales
   volume; the vertical axis is annual dollars. All figures are drawn schematically
   and are not to scale.

The semi-variable line is the honest shape of most of a restaurant's middle. Take utilities. Your walk-in compressors run twenty-four hours a day whether or not anybody eats. The hood runs when the hearth is lit. The water heater cycles. The dining-room lights and the HVAC are on for a fifty-hour week regardless. On top of that base sits a genuinely variable layer: more covers means more dish cycles, more gas through the hearth, more hot water, more make-up air.

Splitting a semi-variable line: the high-low method

There is a simple technique for this and you do not need software. Take your highest-volume period and your lowest-volume period, and let the slope between them tell you the variable rate.

🧮 Run the Numbers

Splitting Bellwether's utility bill. (constructed teaching example)

Two months from a full year of statements, chosen as the extremes:

	Sales	Utilities
August (highest)	\$148,000 | \$4,090
February (lowest)	\$92,000 | \$3,451
Difference	\$56,000** | **\$639

The only thing that changed between those months, to a first approximation, is volume. So the \$639 of extra utility spend is the variable component of \$56,000 of extra sales:

$$\text{Variable rate} = \frac{\$639}{\$56{,}000} = 0.0114 = \textbf{1.14\% of sales}$$

Now back out the fixed base from either month. Using August:

$$\text{Fixed} = \$4{,}090 - (0.0114 \times \$148{,}000) = \$4{,}090 - \$1{,}687 = \textbf{\$2{,}403/month}$$

Check it against February: $\$2{,}403 + (0.0114 \times \$92{,}000) = \$2{,}403 + \$1{,}049 = \$3{,}452$. That is the February bill to within a dollar of rounding. The split holds.

On the year: a fixed base of about \$28,830 and a variable layer of \$17,670 on \$1,550,000 of sales — \$46,500 of utilities, which is exactly Bellwether's budgeted line.

What this method cannot do. High-low uses two data points and throws away ten. If August was hot and busy, you have loaded a weather effect into the variable rate. If February included a compressor failure, the low point is contaminated. Run it, then sanity-check the fixed base against a month you were closed for a remodel or a holiday — that month is nearly pure fixed cost, and it is the best single check you will ever get.

Do this for every line that is not obviously one thing or the other. It takes an afternoon and you do it once a year.

Sorting Bellwether's actual P&L

Here is the plan as it stands after Chapter 31, and here is what each line actually does when volume moves. This is the table the rest of the chapter is built on.

(All Bellwether figures are constructed. Sums: \$930,280 + \$95,200 + \$217,000 + \$46,500 = \$1,288,980 of cost; \$1,550,000 − \$1,288,980 = \$261,020.)

Occupancy and G&A are the easy ones. Occupancy is base rent plus triple-net charges on a signed ten-year lease; it does not move. G&A is accounting, legal, licenses, and bank fees — I will treat it as fixed, while flagging that bank fees do creep with volume and that a restaurant which grows into a second bookkeeper has moved G&A into the semi-variable column.

COGS is the easy variable one: 27.8% of sales, held there by the cost cards from Chapter 11 and the pour-cost discipline from Chapter 15.

That leaves the two lines that carry all the difficulty: labor and other operating.

Other operating, line by line

Fourteen points of sales is too big a number to wave at. Broken out and sorted:

Three of those deserve a sentence.

Credit-card processing is the purest variable cost in the building after food. Chapter 26 established Bellwether's all-in effective rate at 2.81% of net sales — interchange plus assessments plus the processor's markup, blended across card types. At \$1,550,000 that is \$43,555 a year, and it scales with the last dollar you ring exactly as reliably as the first. Operators routinely file this mentally under "bank fees, fixed." It is not fixed. It is two hundred and eighty-one dollars for every ten thousand dollars you sell.

Repairs and maintenance splits about sixty-forty toward fixed because the contracted portion — quarterly hood cleaning, grease-trap pumping, refrigeration preventive maintenance, the annual fire-suppression inspection — happens on a calendar, not on a cover count. The break-fix half (a dish machine that runs four hundred more racks a week, a faucet, a hinge) tracks use.

Marketing is fixed because you decided it was. It is a budgeted number, not a consequence of volume. That is worth naming, because it means marketing behaves like rent for break-even purposes: every dollar you add raises the threshold you must clear. Chapter 27's cost-per-cover-acquired discipline exists precisely so you can tell whether the dollar came back.

⚠️ Where the Money Leaks

The three ways operators butcher this split — and what each one does to the answer.

Mistake one: shove everything ambiguous into "fixed." It feels conservative. It is not; it is just wrong in a specific direction. Overstating fixed cost overstates break-even, which makes you reject a marginal service that would actually have contributed. Chapter 4 made a version of this error, and we are about to quantify it.

Mistake two: treat all labor as variable. This is the most common single error in restaurant break-even, and it is seductive because labor feels like a dial you turn. It is not. Somebody unlocks the door, and somebody is on salary, and neither of them is a function of tonight's book. Treating all labor as variable understates fixed cost and therefore understates break-even. Chapter 5 did this deliberately, as a stress test, and said so.

Mistake three: run the split once and never again. The split is a property of your cost structure, and your cost structure changes when you add a salaried manager, renegotiate a processing rate, sign a maintenance contract, or move marketing from budgeted to performance-based. Re-run it annually and after any structural change. It takes an afternoon.

The reason to care is that the split does not affect your P&L at all — Bellwether still shows \$1,288,980 of cost and \$261,020 of profit no matter how you sort it. It affects only your ability to answer questions about other volumes. Which is to say: it affects every forecast, every decision, and every break-even you will ever compute.

The labor line: three numbers, one of which is right

Now the hard part, and the part this chapter must get right or everything downstream is decoration.

Bellwether's plan carries labor at \$500,000 = 32.3% of sales. Three chapters have now touched that line, and they disagree.

Chapter 4 split it as \$252,000 fixed plus 16.0% of sales variable. Notice that this reproduces the plan exactly at plan volume: $\$252{,}000 + (0.16 \times \$1{,}550{,}000) = \$252{,}000 + \$248{,}000 = \$500{,}000$. That is why nobody caught it. The total came out to the number the plan wanted, so the structure underneath went unexamined.

Chapter 19 built the same line from the schedule instead of from the target — three salaried positions and twenty-one scheduled hourly positions, 453.5 hourly hours a week at a \$14.70 blended rate — and found two things. First, that Chapter 4's fixed floor was \$60,105 too high and its variable rate was 8.4 points too low. Second, and more seriously, that the honest build of the schedule does not produce \$500,000 at all. It produces **\$570,461 = 36.8%.**

Here is the build.

🧾 Read the Numbers

```text FIGURE 32.2 — "The labor line, built from the schedule" [the Bellwether plan] THE ARTIFACT Annual labor build-up for Bellwether, Year 1, reconstructed from the Chapter 19 staffing guide and sorted by cost behavior. THE CONTEXT 68 seats; dinner Tuesday through Saturday plus two brunch services; 3 salaried positions and 21 scheduled hourly positions; 453.5 hourly hours a week at a $14.70 blended rate; burden (payroll taxes, workers' compensation, benefits) at 13.0% of gross wages.

               FIXED LABOR — does not move with covers
                 Chef-owner salary, burdened                          $62,150
                 FOH partner / GM salary, burdened                    $58,760
                 Sous chef salary, burdened                           $48,025
                 ────────────────────────────────────────────────────────────
                 Salaried subtotal (= $149,500 x 1.13)               $168,935
                 Open/close hourly floor, burdened (~26.5 hrs/wk)     $22,960
                 ────────────────────────────────────────────────────────────
                 TOTAL FIXED LABOR                                   $191,895

               VARIABLE LABOR — moves with the schedule
                 Scheduled hourly wages
                   453.5 hrs/wk x $14.70 x 52                        $346,655
                 Overtime allowance (2.5% of hourly wages)             $8,678
                 ────────────────────────────────────────────────────────────
                 Hourly wages                                       $355,333
                 Burden at 13.0%                                     $46,193
                 ────────────────────────────────────────────────────────────
                 Total hourly, burdened                             $401,526
                 Less the fixed open/close floor above              ($22,960)
                 ────────────────────────────────────────────────────────────
                 TOTAL VARIABLE LABOR                               $378,566   24.42% of sales

               TOTAL LABOR  $191,895 + $378,566 =                   $570,461   36.80% of sales

WHAT IT SHOWS The fixed floor is $191,895 — the three salaries fully burdened plus about twenty-six and a half hours a week of open-and-close hourly time that gets worked whether the room does sixty covers or a hundred and thirty. Everything above that floor is 24.42% of sales. Chapter 4's $252,000-and-16.0% split was $60,105 too high on the floor and 8.42 points too low on the slope.

               Two slopes, and do not mix them. Against THIS schedule -- the bottom-up
               $570,461 -- the variable rate is 24.42%, and Chapter 4 is 8.42 points
               under it. Against the rate that reproduces the PLAN's $500,000 on the same
               $191,895 floor -- 19.88% -- Chapter 4 is only 3.88 points under, and its
               two errors then cancel exactly at $1,550,000, which is Chapter 31's
               demonstration. Same floor, two questions: what labor WILL cost, and what
               the plan SAID it would. 24.42% answers the first; 19.88% answers the second.

WHAT IT DOESN'T It does not settle whether the plan's $500,000 is achievable — that is a staffing decision, not an arithmetic one. It does not include the sous chef reclassification from Chapter 20. It assumes the blended rate holds, which it will not: every raise, every promotion, and every minimum-wage step moves it. And it treats the open/close floor as a clean 26.5 hours, which is an average of a messier reality. THE DECISION Carry two labor numbers side by side in every break-even from here forward: the plan's $500,000, because that is what the plan promised, and Chapter 19's $570,461, because that is what the schedule produces. Do not average them. THE LESSON A total that lands on target tells you nothing about the structure underneath it. Chapter 4's two errors pointed in opposite directions and cancelled at exactly one volume — the one the plan was written for. Break-even is a question about every OTHER volume, which is precisely where a cancelled error stops cancelling. ```

Chapter 20 then added a third number. Correctly classifying the sous chef as non-exempt — a salaried kitchen position working roughly fifty-five hours a week, whose duties do not clear the executive exemption test — puts overtime back on the line. At a \$48,000 salary covering forty hours, the regular rate is \$23.08 an hour, the overtime rate is \$34.62, and fifteen hours a week for fifty-two weeks plus burden is \$27,000**. Labor goes to **\$597,461 = 38.5% of sales, and prime cost to 66.3%.

And note where that \$27,000 lands: in the fixed bucket. The sous chef works the same fifty-five hour week whether the room does sixty covers or a hundred and thirty. Reclassification does not make labor more variable. It makes the floor higher, which is the worse of the two outcomes for break-even.

⚖️ Code and Compliance

Why the exempt/non-exempt question is a break-even question.

Chapter 20 covers the substance. What belongs here is the structural consequence: a misclassification does not just create a wage liability, it hides a fixed cost. An operator who believes a salaried sous chef is a flat \$48,000 has understated their fixed floor by \$27,000, which understates break-even by more than sixty-six thousand dollars of sales — as we are about to compute. They will make expansion decisions, pricing decisions, and staffing decisions against a threshold that is materially too low.

The duties test under the Fair Labor Standards Act (FLSA) turns on what the person actually does, not on what the job is called or whether they are paid a salary. Exemption thresholds and duties tests vary by state, several states are stricter than federal law, and the salary thresholds change. Verify your specific positions with an employment attorney, and re-verify when someone's duties change — which in a restaurant is roughly every six months.

Sorted, then, here is the whole cost structure. Two schedules, and every number in this chapter comes out of them.

Fixed costs — the annual nut.

Variable costs — at Chapter 19's labor line.

Two checks before we go on, because a schedule that does not tie to the P&L is a schedule you invented. At the plan's \$500,000 labor line, variable labor is $\$500{,}000 - \$191{,}895 = \$308{,}105$, total variable cost is $\$430{,}280 + \$308{,}105 + \$112{,}960 = \$851{,}345$, and $\$437{,}635 + \$851{,}345 = \$1{,}288{,}980$ — the plan's total cost exactly. At Chapter 19's line, $\$437{,}635 + \$921{,}806 = \$1{,}359{,}441$, which is \$70,461 more, which is precisely the gap between \$570,461 and \$500,000. The schedules tie.

👨‍🍳 On the Line

What the fixed floor actually looks like at 4:15 on a dead Tuesday.

The floor is not an accounting abstraction. It is people, and you can point at them.

The sous is in at nine to break down the delivery, brine chicken, and get stocks moving. The dishwasher comes at ten to run the prep pit. The chef-owner is in the building from morning until close. The FOH partner is doing the schedule, the reservations, the vendor calls, and the deposit. Somebody polishes glassware, sets the room, cuts fruit, changes the kegs, checks the walk-in temps and writes them on the log.

That happens at forty covers and it happens at a hundred and thirty. It is roughly \$191,895 a year — about \$3,690 a week, or \$527 a day across a seven-service week — and it is spent before the first guest is seated.

Here is the operational meaning. At Bellwether's blended contribution of \$15.84 a cover (§32.2), that daily floor alone requires about 33 covers a service before you have paid for the people who opened the building — and that is before rent, insurance, marketing, or the loan. On a Tuesday forecast to 62, roughly the first half of the room pays for the fact that the door was unlocked.

This is why the cut order in §19.5 matters so much and why the cut you can never make is the opening one. You can send a server home at nine. You cannot un-open the restaurant at four.

32.2 Contribution margin ratio: the fraction of each dollar that survives

Chapter 12 taught you contribution margin on a plate: menu price minus plate cost, the dollars an item contributes. Scale that idea to the whole business and you get the single most useful ratio in this chapter.

Contribution margin at the business level is sales minus all variable costs — the dollars left over after you have paid for everything that arrived because the sale happened. The contribution margin ratio (CM ratio) is that figure as a fraction of sales:

$$\text{CM ratio} = \frac{\text{Sales} - \text{Variable costs}}{\text{Sales}} = 1 - \text{Variable cost ratio}$$

At Chapter 19's labor line:

$$\text{CM ratio} = \frac{\$1{,}550{,}000 - \$921{,}806}{\$1{,}550{,}000} = \frac{\$628{,}194}{\$1{,}550{,}000} = 40.53\%$$

Read it in words: out of every dollar Bellwether rings, forty and a half cents survives to pay rent, salaries, insurance, marketing, and whatever is left over is profit. The other fifty-nine and a half cents left with the dollar — food, hourly wages, the card fee, the napkin, the gas.

At the plan's \$500,000 labor line the CM ratio is $(\$1{,}550{,}000 - \$851{,}345) \div \$1{,}550{,}000 = 45.07\%$. Four and a half cents of every dollar is the entire difference between the plan's labor promise and the schedule's labor reality, and you are about to watch what four and a half cents does to a break-even.

Contribution margin per cover

The ratio is what you use for sales-dollar arithmetic. For covers arithmetic you want contribution margin per cover, and this is where you have to be careful about which average check you are using.

The daypart CM ratios come out of §32.4; take them on faith for a page. The blended check is Bellwether's base weekly revenue divided by base weekly covers: \$27,130 ÷ 695 = \$39.04.

Read the third column, because it is the most operationally useful set of numbers in the chapter. A dinner cover contributes \$19.37 toward the fixed nut. A brunch cover contributes \$8.22. A dinner cover is worth two and a third brunch covers, which is a fact with consequences for how you spend a marketing dollar, how you set a reservation policy, and whether you extend brunch to Friday.

🧮 Run the Numbers

The beverage lever, and why the bar changes break-even for the whole building.

Bellwether's plan is 72% food / 28% beverage, with food cost at 30% of food sales and pour cost at 22% of beverage sales — blending to 27.76% COGS. Suppose the beverage program in Chapters 15 and 16 does its job and the mix moves to 70/30 on the same total revenue.

• Food sales: \$1,550,000 × 0.70 = \$1,085,000. At 30%: \$325,500 of food cost.
• Beverage sales: \$1,550,000 × 0.30 = \$465,000. At 22%: \$102,300 of pour cost.
• **New COGS: \$427,800 = 27.60%** of sales, against \$430,280 = 27.76%.

Two points of mix moved COGS by 0.16 points and saved \$2,480. That sounds trivial. Watch what it does to break-even, using the CM-ratio machinery from §32.3: the CM ratio rises from 40.53% to 40.69%, and break-even sales fall from \$1,079,815 to \$1,075,570 — \$4,245 lower.

A sixteenth of a point on COGS moved break-even by four thousand dollars, because the CM ratio is a multiplier. Every dollar of fixed cost you must cover requires 1 ÷ CM ratio = \$2.47 of sales at Bellwether, and every basis point you take off the variable rate improves that multiplier for every dollar of fixed cost you own. This is the most important structural fact in the chapter, and §32.5 is about nothing else.

The limit: a mix shift toward beverage is not free. It usually means more bartender hours, more glassware breakage, more liquor-license exposure, and a guest who lingers — which costs you a turn. Run the whole calculation, not the COGS half of it.

🔍 Check Your Understanding

1. A restaurant does \$1,100,000 in sales with \$660,000 of variable costs. What is its CM ratio, and what does that number mean in words?
2. Bellwether's dinner CM per cover is \$19.37. If you comp a four-top's entire \$184 check, how much contribution did you give away — and why is it not \$184?
3. Why does the CM ratio change when the labor line changes, but not when the rent changes?

(1: (1,100,000 − 660,000) ÷ 1,100,000 = 40.0% — forty cents of every dollar survives variable cost and is available to pay fixed costs. 2: You gave away 4 × \$19.37 = **\$77.48 of contribution, not \$184, because the food, the hourly labor, and the card fee were spent either way — though on a comp there is no card fee, so the true figure is very slightly higher. The \$184 never existed as margin. 3: Because rent is a fixed cost and sits outside the CM ratio entirely; the ratio is built only from variable costs. Rent moves the break-even point, not the ratio.)

32.3 Computing break-even in sales, then in covers, then in covers per night

Now the arithmetic, which is genuinely simple once the split is done.

The break-even point is the sales volume at which total contribution exactly equals total fixed cost — where operating profit is zero. Below it you lose money; above it you make money.

$$\text{Break-even sales} = \frac{\text{Total fixed costs}}{\text{CM ratio}}$$

That is the whole formula. Here it is three times, once for each of Bellwether's live labor numbers.

Check each one. \$437,635 ÷ 0.4507 = \$970,915. \$437,635 ÷ 0.4053 = \$1,079,815. Fixed costs rise to \$464,635 in the third row because the sous chef's \$27,000 of overtime lands in the fixed bucket; the CM ratio is unchanged, so \$464,635 ÷ 0.4053 = \$1,146,435.

🧮 Run the Numbers

Do that division on a calculator and you will not get \$1,079,815. Good — that is the point.

\$437,635 ÷ 0.4053 is \$1,079,780, about \$35 low. The CM ratio is not 0.4053; it is \$628,194 ÷ \$1,550,000 = 40.528645%, and 0.4053 is that number displayed to four places. Divide by the full ratio and you get \$1,079,816.51, which the table carries as **\$1,079,815** — a residual of a dollar and a half that comes from where the variable-cost total was rounded on its way in.

Three things follow, and they are worth more than the \$1.51.

1. Round last. Every intermediate figure in this chapter carries its full precision and only the printed values are rounded. Round the ratio first and a \$1.08 million answer moves \$35 — which is nothing here, and is not nothing in a model with six chained divisions.
2. A residual is not an error, but it must be visible. A book that quietly printed \$1,079,780 in one place and \$1,079,815 in another would be asking you to trust it. Naming the gap is cheaper than hiding it, and it is the only way you can check the arithmetic yourself.
3. None of it changes the decision. \$1,079,780, \$1,079,815, and \$1,079,817 are all 66 dinner covers a night. Precision past the point where the answer changes is a hobby, not analysis — and knowing where that point sits is most of what separates a useful model from a tidy one.

Bellwether's break-even is \$1,079,815 of annual sales. That is the figure this chapter defends: the plan's labor line is a target, and Chapter 19 demonstrated that the schedule required to run the room does not fit inside it. Carrying the plan's \$500,000 into a break-even analysis is carrying a number the operation has already contradicted. The Chapter 20 row is not a footnote either — it is a live compliance question, and if it resolves against the plan, break-even is \$1,146,435.

How that compares to Chapters 4 and 5

This is worth doing carefully, because two earlier break-even figures are already in the reader's plan and both of them are lower than the honest one.

🧾 Read the Numbers

```text FIGURE 32.3 — "Three break-evens for one restaurant" [the Bellwether plan] THE ARTIFACT Reconciliation of every break-even figure produced for Bellwether across Chapters 4, 5, and 32, showing the fixed/variable split behind each. THE CONTEXT One restaurant, one $1,550,000 plan year, three different analyses done at three different points for three different purposes.

               MODEL                     FIXED       VAR %    CM %    BREAK-EVEN
               ─────────────────────────────────────────────────────────────────────
               Ch. 4 pro-forma preview   $517,700   49.76%  50.24%   $1,030,454
                 labor split: $252,000 fixed + 16.0% variable
                 other operating split:  $124,000 fixed + $93,000 (6.0%) variable

               Ch. 5 funding stress      $335,480   66.00%  34.00%     $986,706
                 labor treated as ENTIRELY variable, by design
                 fixed = $265,980 operating + $69,500 debt service

               Ch. 32, plan labor        $437,635   54.93%  45.07%     $970,915
               Ch. 32, Ch. 19 labor      $437,635   59.47%  40.53%   $1,079,815  <-- honest
               Ch. 32, Ch. 20 labor      $464,635   59.47%  40.53%   $1,146,435

WHAT IT SHOWS Bellwether's break-even is $1,079,815, which is $49,361 HIGHER than Chapter 4's $1,030,454 and $93,109 HIGHER than Chapter 5's $986,706. Chapter 4 was wrong in two directions at once: it overstated the fixed floor by $60,105 and understated the variable slope by 8.42 points, and it also parked $124,000 of other operating in "fixed" when only $104,040 belongs there. Chapter 5 was not wrong at all — it was a different question, honestly labeled: it treated all labor as variable to stress the funding case, said so, and stated that the true break point must therefore be HIGHER than its figure. It is higher, by $93,109. WHAT IT DOESN'T None of these five numbers is a cash break-even except Chapter 5's, which includes debt service by design. None of them knows about the ramp, the sales-tax remittance, inventory build, or principal amortization. And none of them is stated in covers, which is the only form a manager can act on. THE DECISION Replace both earlier figures in the plan with $1,079,815, show the $1,146,435 sensitivity beside it, and re-run the margin-of-safety section that depended on the old numbers. Then convert to covers per night, because nobody manages a dollar figure. THE LESSON Two break-evens that disagree are not a contradiction to be averaged. They are two different fixed/variable splits, and one of them is wrong. Find out which by rebuilding the split from the operation, not from the target. ```

There is a point in Figure 32.3 that deserves saying flatly, because it is counterintuitive and it matters for your own plan. Chapter 4's error made break-even look higher than it is on the fixed side and lower on the variable side, and the two effects did not cancel. Overstating the fixed floor by \$60,105 pushes break-even up; understating the variable rate by 8.42 points pushes the CM ratio up, which pushes break-even down. In Chapter 4's arithmetic the second effect won, and the answer came out \$1,030,454. Then the honest labor line added \$70,461 of real cost that Chapter 4 did not have at all, and the answer went past it in the other direction.

There is no rule of thumb here. There is only doing the split properly. Which is the point.

From sales to covers: name your revenue base

Now convert. And here is where most operators get it wrong, so we are going to be pedantic.

$$\text{Break-even covers} = \frac{\text{Break-even sales}}{\text{Average check}}$$

Simple enough — except that Bellwether has two different weekly revenue bases and they are not the same number.

The base is the dining-room cover pattern from Chapter 24: dinner runs Tuesday 62, Wednesday 78, Thursday 92, Friday 120, Saturday 123 — 475 covers a week — at a \$46 average check, plus two brunch services at about 110 covers each at \$24. That is 695 covers a week, 36,140 covers a year, and \$1,410,760 of revenue.

The plan says \$1,550,000. The \$139,240 difference is the revenue bridge, and it is real revenue that is not a base cover:

Note the patio split, because it is the kind of error that is easy to make and hard to spot: the patio's 1,600 covers are not all dinner covers. Twelve hundred are dinner at \$46 and four hundred are brunch at \$24. Pricing all 1,600 at the dinner check inflates the patio to \$73,600 and forces some other bridge line to absorb the difference — the total still reads \$139,240, and it is still wrong. A bridge that foots is not the same as a bridge that is right.

With the patio, annual covers are 37,740.

⚠️ Where the Money Leaks

The phantom \$51 average check.

Here is the error, and I have seen a version of it in a plan that went to a lender.

Take the plan's weekly revenue, \$29,808. Subtract brunch at the base level, \$5,280. You get \$24,528 of dinner revenue. Divide by 475 dinner covers:

$$\$24{,}528 \div 475 = \$51.64$$

A \$51.64 average check. It appears out of nowhere, it is five dollars and sixty-four cents above anything the menu can produce, and the plan never claimed it. It is an artifact of dividing one revenue base by the cover count of a different revenue base. The extra \$139,240 comes from patio seats, event contracts, and takeout orders — not from guests spending twelve percent more.

The same error runs the other way in covers. Divide break-even sales of \$1,079,815 by the blended base check of \$39.04 and you get 27,659 covers — which would imply about 73 dinner covers a night. The honest answer, computed below, is 25,177 covers and 66 dinner covers a night. The naive method overstates your break-even by roughly seven covers a night, every night of the year, which is exactly the size of the cushion you would spend defending yourself against a problem you do not have.

The rule: name your revenue base every single time, out loud, in the sentence. Not "break-even is twenty-seven thousand covers." Rather: "break-even is 25,177 covers on the base dining-room pattern, plus the patio, events, and off-premise revenue running at plan." It is clumsier. It is also correct, and it is the only version you can hand to someone else.

Here is the honest conversion, done two ways, because there are two legitimate questions hiding inside "break-even covers."

Reading 1 — everything falls together. This is the recession case, the bad-winter case, the new-competitor case: dinner, brunch, patio, events, and takeout all soften at once. Scale the whole business proportionally.

$$\text{Scale factor} = \frac{\$1{,}079{,}815}{\$1{,}550{,}000} = 69.67\%$$

Cross-check that against the direct division, done correctly: the base's share of plan revenue is \$1,410,760 ÷ \$1,550,000 = 91.02%, so the base share of break-even sales is \$1,079,815 × 0.9102 = \$982,807, and \$982,807 ÷ \$39.04 = 25,175 covers. That agrees with the 25,177 above to within rounding. The two methods agree when you name the base.

Reading 2 — dinner alone flexes. This is the different question: if brunch, the patio, events, and takeout all run at plan, how bad can dinner get? Brunch contributes \$94,037 a year and the bridge contributes \$55,696 (both from §32.4), so \$149,733 of the fixed nut is already handled. Dinner must cover the remaining \$287,902 at \$19.37 of contribution per cover:

$$\$287{,}902 \div \$19.37 = 14{,}863 \text{ dinner covers a year} = 286/\text{week} = \textbf{57 covers a night}$$

Both readings are correct and they answer different questions. Reading 1 is the number you plan with, because in a real downturn nothing falls in isolation. Reading 2 is the number you diagnose with, because when one daypart softens and the others hold, Reading 1 will scare you unnecessarily.

Notice something instructive: Reading 2's total revenue is \$683,698 of dinner plus \$274,560 of brunch plus \$139,240 of bridge = **\$1,097,498, which is higher than Reading 1's break-even of \$1,079,815 — and yet both break even. That is not an error. Break-even sales is mix-dependent.** A sales mix weighted toward brunch, whose CM ratio is nearly eight points worse than dinner's, requires more total revenue to clear the same fixed nut. Any operator who quotes a single break-even dollar figure without a mix assumption is quoting an incomplete number.

FIGURE 32.4 — Bellwether's break-even chart              [the Bellwether plan, Ch. 19 labor line]

  $ thousands
  1,600 ┤                                                        TR ●  1,550
        │                                                    ●
  1,400 ┤                                              ●         TC ○  1,359
        │                                         ●        ○
  1,200 ┤                                   ●        ○                  PROFIT
        │                             ●         ○                       region
  1,000 ┤                       ●  ✱ ○                       ✱ = BREAK-EVEN
        │                 ●    ○                                 $1,079,815
    800 ┤           ●   ○                                         66 dinner
        │      ●   ○                                              covers/night
    600 ┤ ●   ○                        LOSS region
        │    ○
    438 ┤○─────────────────────────────────────────────  fixed cost $437,635
        │
      0 └────┬─────┬─────┬─────┬─────┬─────┬─────┬─────┬──── annual sales
             600   800  1,000 1,080 1,200 1,400 1,550 1,700

        ● total revenue (TR)          ○ total cost (TC = $437,635 + 59.47% of sales)
        Vertical gap between the lines = operating profit or loss.
        Not to scale. All figures constructed.

Read the chart the way an operator should. The two lines are not steep and shallow — they are similar, diverging at only 40.53 cents per sales dollar. That narrow wedge is the entire profit of the business, and it is why the crossing point sits so far to the right: you have to travel almost seventy percent of the way to plan volume before the wedge is wide enough to cover the fixed nut. Above the crossing the wedge opens fast, which is the good news of §32.5. Below it, it closes just as fast, which is the bad news of §32.5.

🧮 Run the Numbers

The number to hold in your head at seven o'clock.

Forget the annual figures for a moment. Here is break-even expressed the way you actually need it.

Bellwether's fixed cost is \$437,635 a year. There are 364 services in the year — five dinners and two brunches a week, fifty-two weeks. So the fixed nut is:

$$\$437{,}635 \div 364 = \$1{,}202 \text{ per service}$$

A dinner cover contributes \$19.37. So a dinner-only view of the daily nut is:

$$\$1{,}202 \div \$19.37 = \textbf{62 covers}$$

Sixty-two covers pays for the day. Which is a startling number, because 62 is exactly Bellwether's forecast for a Tuesday. On a plan Tuesday, the restaurant breaks even and contributes nothing. Everything Bellwether earns in a year, it earns Wednesday through Saturday and at brunch.

Two caveats, because this is a rule of thumb and rules of thumb get quoted past their range. First, it allocates fixed cost evenly across services, which is a convention, not a fact — the fixed cost of a Saturday and the fixed cost of a Tuesday are the same \$1,202 only because we said so. Second, it ignores brunch's contribution, which is why the honest annual answer (66 dinner covers a night with brunch scaling down alongside) differs from this back-of-the-napkin 62.

Use it anyway. It is the version you can say out loud to a manager on the floor, and a manager who knows that sixty-two covers pays for the day will make a better decision at nine o'clock than one who knows nothing.

32.4 Break-even by daypart: which services actually carry the building

Chapter 1 promised that this chapter would tell you whether a slow daypart is carrying its weight or quietly subsidized. Here is the method, and here is Bellwether's answer.

Break-even by daypart means computing contribution separately for each service you run, then asking two different questions about each one:

1. The avoidable-cost question. If I closed this service tomorrow, what would I actually save, and what would I actually lose? This is the correct question for a close-it decision.
2. The fully-allocated question. If this service carried its proportional share of rent, salaries, insurance, and marketing, would it stand up? This is the correct question for a should-it-exist decision and for anything you put in a plan.

They give different answers. Both are legitimate. Confusing them is how restaurants end up keeping a Monday lunch for nine years.

Building the daypart contribution statement

You need three things per daypart: revenue, COGS at that daypart's own rate, and the hourly labor that service actually consumes. Everything else follows.

🧾 Read the Numbers

```text FIGURE 32.5 — "Daypart contribution statement" [the Bellwether plan, Ch. 19 labor line] THE ARTIFACT Contribution statement for Bellwether Year 1, split into dinner, brunch, and the revenue bridge (patio, private events, off-premise). Variable costs only; fixed costs held below the line where they belong. THE CONTEXT Dinner: 475 covers/wk at $46. Brunch: two services, 110 covers each at $24. Hourly hours allocated from the Chapter 19 schedule: 297 to dinner, 100 to brunch, 41 to the bridge, per week (plus the 26.5-hour fixed open/close floor, which is FIXED and therefore excluded here).

                                     DINNER      BRUNCH      BRIDGE       TOTAL
               ───────────────────────────────────────────────────────────────────
               Revenue             $1,136,200   $274,560    $139,240   $1,550,000
               Cost of goods sold     318,136     74,131      38,013      430,280
                  as % of revenue        28.0%      27.0%       27.3%        27.8%
               Variable labor         256,776     86,377      35,413      378,566
                  as % of revenue        22.6%      31.5%       25.4%        24.4%
               Other variable          82,827     20,015      10,118      112,960
               ───────────────────────────────────────────────────────────────────
               Total variable         657,739    180,523      83,544      921,806
               ───────────────────────────────────────────────────────────────────
               CONTRIBUTION          $478,461    $94,037     $55,696     $628,194
               CM RATIO                 42.11%     34.25%      40.00%       40.53%
               CM PER COVER             $19.37      $8.22          —        $15.84
               ───────────────────────────────────────────────────────────────────
               Less total fixed cost                                     (437,635)
               OPERATING PROFIT                                          $190,559

               Fully-allocated view — fixed cost apportioned by share of revenue:
                                     DINNER      BRUNCH      BRIDGE
               Share of revenue         73.30%     17.71%       8.99%
               Allocated fixed cost   $320,787    $77,505     $39,343
               Allocated profit       $157,674    $16,532     $16,353
               Own break-even sales   $761,783   $226,292     $98,358
               Own break-even covers  64/night   91/service        —

WHAT IT SHOWS All three revenue streams carry themselves, fully allocated. Brunch contributes $94,037 against $77,505 of allocated fixed cost and clears its own break-even at 91 covers a service against 110 forecast — a 17% cushion. Dinner clears at 64 covers a night against 95 — a 33% cushion. Dinner's CM ratio is 42.11% against brunch's 34.25%: every dollar of sales moved from dinner to brunch costs about eight cents of contribution. WHAT IT DOESN'T The fully-allocated view is a CONVENTION, not a measurement. Apportioning fixed cost by revenue share is defensible and arbitrary in equal measure — apportion by hours of operation instead and brunch looks better, apportion by square-foot-hours and it looks worse. The statement also cannot see the fixed cost brunch CREATES (a sixth and seventh service on a crew scheduled five nights) or the demand brunch generates for dinner. And it says nothing about cash, which §32.8 handles. THE DECISION Keep brunch, and do not expand it at dinner's expense. Specifically: reject any proposal to add a Friday brunch that pulls hourly hours from a Friday dinner forecast to 120 covers. Work brunch's check average instead — §32.7 prices that. THE LESSON A daypart's contribution and a daypart's profitability are two different questions with two different right answers. Ask which one you are deciding. ```

Is brunch carrying itself?

Yes. Two ways, and the difference between them is the whole lesson.

On avoidable cost — decisively yes, and it is not close. If Bellwether closed brunch tomorrow, it would give up \$94,037 of annual contribution. What would it save? Not rent, not insurance, not technology, not the salaries — the chef-owner, the FOH partner, and the sous chef are salaried and their compensation is not brunch-specific. Realistically you save some utilities, a little wear, and perhaps some genuinely brunch-driven prep. Call it \$18,000 generously. Closing brunch costs Bellwether about \$76,000 a year. On that basis brunch clears its own avoidable break-even at \$18,000 ÷ 0.3425 = \$52,555 of revenue — 2,190 covers a year, 21 covers a service — against 110. It is not a marginal service. It is a very safe one.

On fully-allocated cost — yes, but thinly, and the thinness is the warning. Brunch's proportional share of the fixed nut is \$77,505, against \$94,037 of contribution: \$16,532 of allocated profit on \$274,560 of revenue, a 6.0% margin.** Its own break-even is \$226,292, which is 91 covers a service against 110 forecast.** Nineteen covers of cushion.

Now push on it. At Chapter 20's labor line the fixed nut rises to \$464,635, brunch's allocated share rises to \$82,287, and allocated profit falls to \$11,750 — a 4.3% margin, break-even at 96 covers a service. Fourteen covers of cushion. And brunch is the daypart most exposed to weather, holidays, and the school calendar. Brunch is carrying itself, and it would stop doing so on a labor reclassification plus one soft season.

That is a genuinely useful finding, and notice that it is not the finding either extreme would have given you. An operator who ran only the avoidable-cost test would conclude brunch is a slam dunk. An operator who ran only a crude "brunch labor is 31.5% of brunch revenue, that's terrible" test would close a service that contributes ninety-four thousand dollars a year.

👨‍🍳 On the Line

Why brunch labor runs 31.5% and dinner labor runs 22.6%.

The gap is not laziness and it is not bad scheduling. It is arithmetic about check average.

A brunch service at Bellwether needs a nearly complete building. You need a lead cook and two on the line, because eggs and griddle items are made to order and they do not hold. You need a dish/prep body. You need a host, three servers, a busser, and a bartender, because brunch cocktails are a meaningful part of the check. That is roughly fifty hours across the two services' worth of staffing per service — call it a hundred hours a week for both.

A hundred hours at \$14.70 blended, burdened at 13%, is \$86,377 a year. Divide by \$274,560 of brunch revenue and you get 31.5%.

Run the same crew against a \$46 check instead of a \$24 check and the labor percentage nearly halves. That is the entire mechanism. Brunch labor is not worse per hour, per cover, or per plate — it is worse per dollar, because the dollars are smaller. A brunch cover generates \$24 of revenue and takes almost as much service work as a dinner cover generating \$46.

Which tells you exactly where the lever is, and it is not the schedule. It is the check. §32.7 prices a two-dollar brunch check increase, and it is worth more than any staffing move available to you.

🤝 Hospitality

The thing the daypart statement cannot see, and why you should not close brunch on the numbers alone.

Figure 32.5 treats dinner and brunch as separate businesses that happen to share a kitchen. They are not. A neighborhood restaurant's brunch is the service where a guest tries you for the first time, because brunch is low-commitment: it is cheaper, it is daylight, it does not require a reservation three weeks out, and nobody has to decide whether the occasion justifies the check.

Chapter 23 works the arithmetic of the second visit properly. The relevant piece here is that a brunch cover at \$8.22 of contribution which converts into a dinner regular is worth many multiples of \$8.22, and none of that appears in a daypart statement. Brunch is, among other things, a \$274,560-a-year customer-acquisition channel that happens to be contribution-positive — which, compared to almost any marketing spend in Chapter 27, is a remarkable thing to own.

Do not turn that into an excuse. "It's a loss leader" is the sentence operators use to avoid costing a service. The discipline is: compute the contribution honestly, then argue the strategic case on top of a positive number, not instead of a negative one. At Bellwether the number is positive, so the strategic case is a bonus. If it were negative, the strategic case would have to be quantified — how many brunch guests convert, at what rate, worth what — before it justified a dollar.

The daypart nobody costs: the Tuesday

One more use of this machinery, and it is the one that changes behavior.

Bellwether's Tuesday is forecast to 62 covers at \$46 = \$2,852 of revenue. Contribution at 42.11% is \$1,201. The evenly-allocated fixed cost of a service is \$1,202. A plan Tuesday misses break-even by a dollar. Not a rounding artifact worth arguing about — the point is that a fully-forecast Tuesday contributes essentially nothing to the year.

That is not a reason to close Tuesday. Closing a night in a neighborhood restaurant costs you presence, costs your staff a shift, and does not reduce your rent. It is a reason to do three specific things, each of which is a later chapter: work the Tuesday check with a prix fixe or a wine feature (Chapter 24's shoulder-hour arithmetic), cut the Tuesday schedule to the forecast rather than to habit (Chapter 19), and stop treating a 55-cover Tuesday as normal variance. Below 62 covers, Tuesday is costing you money, and now you know the number.

32.5 Operating leverage: why restaurants swing so violently on small revenue changes

Here is a sentence I have heard in a hundred versions: "We were down about ten percent last quarter, but we'll be fine."

They were not fine. Let me show you why, because this is the most important structural idea in the chapter and almost nobody carries it correctly in their head.

Operating leverage is the sensitivity of operating profit to a change in revenue. It exists because fixed costs do not move. When revenue falls, variable costs fall with it — but the fixed nut sits there in full, so the entire shortfall in contribution comes straight off profit.

The measure is the degree of operating leverage (DOL):

$$\text{DOL} = \frac{\text{Contribution margin}}{\text{Operating profit}}$$

At Bellwether, on Chapter 19's labor line:

$$\text{DOL} = \frac{\$628{,}194}{\$190{,}559} = 3.30$$

Read that as: every 1% change in revenue produces a 3.30% change in operating profit, in whichever direction it moves. A 10% revenue decline is not a 10% profit decline. It is a 33% profit decline.

🧮 Run the Numbers

A 10% revenue miss, in dollars, in profit, and in covers.

Bellwether at plan, Chapter 19 labor line: revenue \$1,550,000, contribution \$628,194, fixed \$437,635, operating profit \$190,559.

Now take revenue down 10%, to \$1,395,000.

	Plan	−10% revenue	Change
Revenue	\$1,550,000 | \$1,395,000	−\$155,000 (−10.0%)
Variable cost at 59.47%	\$921,806 | \$829,625	−\$92,181
Contribution	\$628,194** | **\$565,375	−\$62,819
Fixed cost	\$437,635 | \$437,635	\$0
Operating profit	\$190,559** | **\$127,740	−\$62,819 (−33.0%)

The mechanism in one line: of the \$155,000 of lost revenue, \$92,181 of cost went away with it and **\$62,819 did not.** That \$62,819 is 40.53% of the lost sales — the CM ratio — and it lands entirely on profit, because rent, salaries, insurance, and marketing did not read the sales report.

Now in covers, which is how you will actually experience it. A 10% revenue decline is 9.5 dinner covers a night — 95 down to 85.5. Two and a half tables. On a Friday you would not notice; the room would still feel busy. Averaged across the year it costs \$62,819 of profit, which is a third of everything the restaurant earns.

And it runs upward with exactly the same violence. Revenue +10% to \$1,705,000 adds \$62,820 of profit, taking operating profit to \$253,379 — a 33% gain from nine and a half more covers a night. This is the arithmetic behind every operator who tells you the difference between a good year and a bad one was "a table or two a night." They are being literal.

FIGURE 32.6 — The operating-leverage curve         [the Bellwether plan, Ch. 19 labor line]

  operating
  profit
  ($000)
   +250 ┤                                                              ╱ 251
        │                                                        ╱
   +200 ┤                                                  ╱ 191  ← PLAN
        │                                            ╱          $1,550,000
   +150 ┤                                      ╱ 130
        │                                ╱
   +100 ┤                          ╱ 89
        │                    ╱
    +50 ┤              ╱ 49
        │        ╱
      0 ┼──────✱───────────────────────────────────────────────────────────
        │  ╱   $1,079,815 = BREAK-EVEN (66 dinner covers a night)
    -50 ┤╱ -32
        │
   -100 ┤ -73
        │
   -150 ┤-113
        └──┬─────┬─────┬─────┬─────┬─────┬─────┬─────┬─────┬──── annual sales
          800   900  1,000 1,080 1,200 1,300 1,400 1,500 1,700   ($000)

  Slope = the CM ratio, 40.53 cents of profit per additional sales dollar.
  The line is STRAIGHT — leverage is not curvature, it is the fact that the line
  does not pass through the origin. It crosses zero at $1,079,815, and the
  PERCENTAGE swing in profit is violent precisely because profit near the
  crossing is a small number. Not to scale. All figures constructed.

That figure repays a careful look, because operating leverage is routinely mis-taught as though profit accelerates. It does not. Every additional sales dollar is worth exactly 40.53 cents, at \$900,000 and at \$1,700,000 alike. **What changes is the base you are comparing against.** At plan, \$62,819 is 33% of profit. If Bellwether ever got to \$1,700,000 and \$251,000 of profit, the same \$62,819 would be 25%. Leverage falls as you move away from break-even, which is the entire financial argument for building a cushion.

Why the labor line changes the leverage, not just the level

Here is where §32.1's work pays off a second time.

Look at the third column. At Chapter 20's labor line, a 10% revenue miss takes 38.4% of operating profit. The same \$62,819 of lost contribution is a bigger fraction of a smaller profit. This is the compounding nobody expects: a labor problem does not just cost you \$97,461 a year, it makes every other problem hurt about forty percent more.

⚠️ Where the Money Leaks

The two directions operators get leverage backwards.

Backwards one: "we'll cut our way out of it." When revenue falls 10%, the instinct is to cut cost by 10%. You cannot. Only 59.47% of Bellwether's cost is variable, and much of the fixed 40.53% is contractually committed for years — the lease, the equipment lease, the insurance policies. To recover \$62,819 by cost reduction alone you would have to take it entirely out of the fixed base, which means a position, a marketing budget, or a maintenance contract. The math of leverage in a downturn points at revenue, not at cost, and the operators who forget that cut their way into a worse product and a worse guest experience and then lose the revenue too.

Backwards two: adding fixed cost in a good quarter. Leverage is symmetric and so is the trap. Two strong months feel like structural improvement, and the natural move is to hire, or to sign a bigger marketing commitment, or to take a distribution. Every one of those raises the fixed base, which raises break-even by 1 ÷ CM ratio = \$2.47 of sales for every \$1 of fixed cost added. Chapter 35 is a whole chapter about this mistake at a larger scale. §32.7 prices it at this one.

The disciplined version: in a good quarter, add variable capacity — hours, a shift, an extra prep day — and hold the fixed base flat until you have four quarters of evidence.

🔍 Check Your Understanding

1. A restaurant has contribution margin of \$540,000 and operating profit of \$90,000. What is its DOL, and what does a 5% revenue decline do to profit?
2. Two restaurants both earn \$150,000 of operating profit. One has a DOL of 2.0, the other 5.0. Which would you rather own going into an uncertain year, and what specifically makes the difference?
3. Bellwether's CM ratio is 40.53%. If you add \$25,000 of annual marketing spend, by how much does break-even sales rise — and how many additional dinner covers a night is that?

(1: 540,000 ÷ 90,000 = 6.0; a 5% revenue decline takes 30% of profit, \$27,000. 2: The DOL-2.0 restaurant, because the same revenue miss costs it less than half as much profit; the difference is the share of its cost base that is fixed — a lower fixed base means lower leverage and a lower break-even, generally bought by accepting a higher variable rate. 3: \$25,000 ÷ 0.4053 = **\$61,683 of additional break-even sales. At \$46 and 260 dinner services a year that is 1,341 covers, or about 5 covers a night — every night, forever, to pay for the marketing.)

32.6 Margin of safety and what a cushion is worth

Break-even tells you where the floor is. Margin of safety tells you how far above it you are standing, which is the number that actually predicts whether you survive a bad quarter.

$$\text{Margin of safety} = \text{Expected sales} - \text{Break-even sales}$$

$$\text{Margin of safety \%} = \frac{\text{Expected sales} - \text{Break-even sales}}{\text{Expected sales}}$$

Compute it four ways, because each form answers a question somebody will actually ask you.

(Cushion in covers = 95 planned dinner covers a night minus break-even dinner covers a night. Weeks of trading = MOS dollars ÷ \$29,808 of plan weekly revenue. Cash break-even adds the \$69,500 of annual debt service to fixed cost; see §32.8.)

Read the second row, which is Bellwether's honest position. Break-even is \$1,079,815; the plan says \$1,550,000; the cushion is \$470,185, or 30.3% of revenue, or 29 dinner covers a night, or about sixteen weeks of trading.

Each of those four framings does different work:

• 30.3% is the form a financial reader wants. Revenue can miss the plan by nearly a third before operating profit goes negative. For a first-year independent restaurant that is a respectable, not luxurious, position.
• \$470,185 is the form you use to size a decision. Any commitment smaller than this cannot, by itself, sink the year.
• 29 covers a night is the form the manager on the floor uses. Bellwether can lose seven four-tops a night — every night — and still not lose money. That is a genuinely reassuring sentence, and it is the one worth saying to a nervous chef-owner in February.
• Sixteen weeks is the form that connects to the calendar. Bellwether could lose four full months of trading and still break even on the year, which sounds enormous until you remember that the COVID-19 shutdowns took longer than that from a great many American restaurants, and that they did it without reducing the rent.

The ramp, which nobody models and everybody lives through

Now the thing that makes every number above optimistic for the year that matters most.

Chapter 9 established that Bellwether's first quarter runs at roughly 66.6% prime cost on \$363,100 of ramp revenue — the systems are not built, the cooks are learning the portions, the schedule is written to a forecast nobody has data for, and the walk-in is a mess. For the year to land on 60.0% prime, weeks 14 through 52 must average 58.0%.

Check that, because it is the constraint the rest of the year lives under: $0.666 \times \$363{,}100 = \$241{,}878$ of Q1 prime cost; the plan's total prime cost is \$930,280; so weeks 14–52 must deliver $\$930{,}280 - \$241{,}878 = \$688{,}402$ of prime cost on $\$1{,}550{,}000 - \$363{,}100 = \$1{,}186{,}900$ of revenue, which is 58.0%.

A break-even computed on annual averages does not describe the quarter you are actually about to live. Here is Q1's own break-even.

🧮 Run the Numbers

Break-even during the ramp.

Q1 fixed cost is one quarter of the annual nut: \$437,635 ÷ 4 = **\$109,409.** Fixed labor is \$191,895 ÷ 4 = \$47,974.

So Q1's variable prime cost is \$241,878 − \$47,974 = \$193,904**, which on \$363,100 of Q1 revenue is 53.40% — against 47.64% for the plan year on the plan's labor line. Add the 7.29% of variable other operating and Q1's variable cost ratio is 60.69%, giving a Q1 CM ratio of 39.31%.**

$$\text{Q1 break-even} = \frac{\$109{,}409}{0.3931} = \$278{,}323 \quad \text{(an annualized rate of } \$1{,}113{,}292\text{)}$$

	Steady state, plan labor	First quarter
CM ratio	45.07%	39.31%
Break-even, annualized rate	\$970,915 | **\$1,113,292**
Break-even dinner covers a night	60	68

During the ramp, break-even runs eight covers a night higher than the annual figure implies. Q1 revenue of \$363,100 does clear it — contribution is \$142,735 against \$109,409 of fixed cost, so the quarter earns \$33,326 with a 23.3% margin of safety. That is the good news, and it comes mostly from the opening spike: Q1 revenue runs at about 94% of the plan's weekly rate, because a new restaurant is busy before it is good.

Now do it at Chapter 19's labor line, which adds \$17,615 of variable labor in the quarter — 4.85 points of Q1 sales. Q1's variable ratio goes to 65.54%, the CM ratio to 34.46%, and:

$$\text{Q1 break-even} = \frac{\$109{,}409}{0.3446} = \$317{,}496 \quad \text{(an annualized } \$1{,}269{,}984\text{)}$$

That is 78 dinner covers a night. Q1 still clears it — \$15,715 of profit — but the margin of safety collapses from 23.3% to 12.6%. One bad February inside the ramp, and the quarter is negative.

This is the single most important sensitivity in the chapter, and it is invisible in an annual break-even. The honeymoon revenue is what saves Bellwether's first quarter, not cost discipline — and the honeymoon ends, per Chapter 9, right around the time the cost discipline is supposed to arrive.

⚠️ Where the Money Leaks

"We're above break-even" is not the same as "we're fine."

Three specific ways a positive margin of safety misleads.

It is an annual average over a seasonal business. Bellwether's 30.3% cushion is a full-year figure. Chapter 33's February problem is a specific set of weeks in which revenue is at its lowest and the insurance premium, the liquor-license renewal, and the sales-tax remittance all land. A business with a 30% annual margin of safety can be below break-even for eleven consecutive weeks.

It is measured against a forecast, not a fact. Margin of safety uses expected sales. If the \$1,550,000 is optimistic — and Chapter 4's assumptions register flagged three variables it depends on — then the cushion is smaller than it reads by exactly the amount of the optimism. Compute margin of safety against your downside case as well as your plan case, always. On a 12% revenue miss — \$1,364,000 instead of \$1,550,000 — margin of safety drops from 30.3% to 20.8%, and the cushion from 29 dinner covers a night to about 18.

It does not include the loan payment. \$470,185 of margin of safety at the operating line becomes \$298,702 once the \$69,500 of debt service is in the fixed base — 30.3% becomes 19.3%, and sixteen weeks becomes ten. §32.8 pursues this.

32.7 Using break-even for decisions: a new service, a price change, an added position

Everything up to here has been description. This is the section where break-even earns its keep, because a break-even that only tells you where you are is a thermometer. Used properly it is a pricing tool for decisions, and the method is always the same three questions:

1. What does this decision do to fixed costs? Every dollar you add there costs you 1 ÷ CM ratio = \$2.47 of sales at Bellwether, permanently.
2. What does it do to the CM ratio? Anything that changes price, product cost, or the labor slope moves the multiplier itself.
3. What is the incremental test, stated in covers? Because a decision that requires nine more covers a night is a different decision from one that requires thirty, and "nine" and "thirty" are things an operator can judge. "\$144,984" is not.

Decision one: adding a Sunday dinner

The FOH partner wants Sunday dinner. The Rivermill District has almost nothing open Sunday evening, the argument goes, and the building is already paid for.

What it does to fixed cost. Not much, and that is the whole appeal. No new rent, no new insurance, no new salaries — the chef-owner and FOH partner are already salaried and would be there. What it does add: a shift-lead differential so somebody who is not an owner can close a sixth night, plus the utilities of running the building an extra evening. Call it \$11,000 a year, and be suspicious of any estimate that says zero.

What it does to the CM ratio. It improves it slightly, because a Sunday dinner cover carries dinner's 42.11% rather than the blended 40.53%.

The incremental test. A dinner cover contributes \$19.37. So:

$$\text{Break-even Sunday covers} = \frac{\$11{,}000}{\$19.37} = 568 \text{ covers a year} = \textbf{11 covers a Sunday}$$

Eleven covers. Which looks like the easiest decision in the book — until you ask the question Chapter 28 taught you to ask about delivery, which applies here identically: is it incremental? If Sunday pulls one cover in five from a Friday or Saturday that would have filled anyway, the honest test is:

$$\frac{\$11{,}000}{\$19.37 \times 0.80} = 710 \text{ covers a year} = \textbf{14 covers a Sunday}$$

🧮 Run the Numbers

Sunday dinner at 55 covers, priced properly.

Suppose Sunday settles at 55 covers — between a Tuesday and a Wednesday, which is a reasonable expectation for a neighborhood room on a night with little competition.

• Gross Sunday revenue: 55 × \$46 × 52 = **\$131,560**
• Assume 20% cannibalization from Friday and Saturday: incremental covers = 44 a Sunday = 2,288 a year, worth 2,288 × \$46 = **\$105,248** of incremental revenue
• Incremental contribution: 2,288 × \$19.37 = **\$44,320**
• Less new fixed cost: (\$11,000)
• Net gain to operating profit: \$33,320

And what it does to the whole business:

	Before	With Sunday
Revenue	\$1,550,000 | \$1,655,248
Contribution	\$628,194 | \$672,514
CM ratio	40.53%	40.63%
Fixed cost	\$437,635 | \$448,635
Break-even sales	\$1,079,815** | **\$1,104,197
Operating profit	\$190,559 | \$223,879
Margin of safety	30.3%	33.3%

Break-even went up \$24,382 and the margin of safety went up three points. That is the signature of a good marginal-service decision: you raised the bar and cleared it by more than you raised it.

What this does not price, and you must: a sixth service on a crew scheduled for five. Chapter 21's whole argument is that turnover is a line item, and a Sunday added without adding headcount is a Sunday paid for out of your staff's weekends. The \$33,320 is real. So is the cook who leaves in August. Price the second one before you sign up for the first.

Decision two: a price change

Chapter 11 taught you to price from a target and Chapter 12 taught you to think in contribution dollars. Break-even tells you what a price move is worth to the building, and it produces one specific number that no other analysis gives you: how much traffic you can afford to lose.

Take the lever §32.4 identified: brunch's check average, currently \$24.

🧮 Run the Numbers

Two dollars on the brunch check.

A \$2 increase on a \$24 check is an 8.3% price move, achieved across the brunch menu — a dollar on the egg dishes, two on the larger plates, a repriced brunch cocktail. Portions do not change, so plate cost does not change, and the entire \$2 is contribution.

	Before	After
Brunch check	\$24.00 | \$26.00
Brunch CM per cover	\$8.22 | **\$10.22**
Brunch CM ratio	34.25%	39.31%
Brunch revenue	\$274,560 | \$297,440
Brunch contribution	\$94,037 | \$116,917
Total revenue	\$1,550,000 | \$1,572,880
Total CM ratio	40.53%	41.39%
Break-even sales	\$1,079,815** | **\$1,057,344
Operating profit	\$190,559 | \$213,439
Margin of safety	30.3%	32.8%
Brunch's own break-even	91 covers/service	78 covers/service

Twenty-two thousand eight hundred and eighty dollars of pure contribution, break-even down \$22,471, and brunch's own break-even down thirteen covers a service.

Now the number that matters: how many guests can you afford to lose? At \$10.22 of contribution per cover, matching the old \$94,037 requires 9,201 covers a year against 11,440 today. You can lose 2,239 covers a year — about 21 a service, or 19.6% of brunch traffic — and be no worse off.

Compare the same \$2 on dinner: contribution per cover goes from \$19.37 to \$21.37, break-even falls to \$1,032,892, and the loss tolerance is 2,311 covers a year — about 9 covers a night, or 9.4%.

Read those two tolerances against each other, because the lesson is counterintuitive in both directions. The arithmetic cushion is twice as wide on brunch, because \$2 is a bigger share of a \$24 check. But the guest also feels 8.3% more than 4.3%, so the elasticity risk is higher too. Break-even gives you the first half of that trade with precision and the second half not at all. It tells you what you can afford to lose. It has no opinion whatsoever about what you will lose.

That is the correct division of labor. The arithmetic sets the tolerance; Chapters 10, 12, and 24 set the judgment. An operator who has both makes a decision. An operator with only the second one guesses.

Decision three: adding a salaried position

This is the decision break-even is best at, because a salary is pure fixed cost and the multiplier is brutal and easy to see.

The chef-owner has not had a full day off since the soft open. The proposal: a salaried assistant general manager at \$47,000**, which fully burdened at 13.0% is **\$53,110 a year.

A \$58,760 hire raises break-even by \$144,984 — the \$2.47 multiplier, exactly — and costs nine covers a night, every night, forever. It also takes five weeks off the cushion.

Stated as an incremental test instead: the position must generate \$58,760 of contribution, which at \$19.37 a dinner cover is 3,034 covers a year, or about 12 more dinner covers a night — or the equivalent in verifiable cost savings, or some combination.

⚠️ Where the Money Leaks

The most expensive mistake in this section is treating that as a verdict.

Twelve covers a night sounds like a lot. Break-even, used badly, stops there and the hire is rejected. Used properly, break-even frames the question, and the question is: can a competent AGM produce twelve covers a night or \$58,760 of savings?

Look at what is actually on the table at Bellwether:

• The sous chef's overtime. Chapter 20's reclassification costs \$27,000 a year, and a meaningful part of it is the sous closing four nights a week. An AGM who takes the close absorbs a real share of that — call it \$18,000 conservatively — without a single additional cover.
• Schedule discipline. Chapter 19's variance analysis is a weekly job nobody currently owns. Two points of labor on \$1,550,000 is \$31,000. One point is \$15,500. A manager who holds the schedule to the staffing guide pays for a fraction of themselves in hours alone.
• Revenue. Better pacing on Friday and Saturday — Chapter 22's host-stand discipline and Chapter 24's duration management — plausibly adds a turn's worth of covers on the two busiest nights. Six more covers on Friday and six on Saturday is 624 covers a year, \$12,087 of contribution.

\$18,000 + \$15,500 + \$12,087 = **\$45,587, against \$58,760. It does not quite close on those three alone, and that is the answer break-even is supposed to give you: not yes, not no, but "you are thirteen thousand dollars short, so either find the fourth source or make the hire knowing it costs you thirteen thousand dollars and buys the chef-owner a day off."**

Which may be an excellent trade. An owner who burns out is an existential risk to a business built on their presence, and Chapter 35's readiness test is explicitly about whether the business runs without them. Break-even prices the decision. It does not make it.

32.8 The limits: what break-even hides, especially about cash

I have now spent nine thousand words teaching you a calculation, and this section is where I tell you what it cannot do. That is the deal this book makes in every chapter, and here the gap between what break-even says and what actually closes restaurants is unusually wide.

It does not know that cash exists

This is the big one, and it is the reason Chapter 33 comes next.

Bellwether's break-even of \$1,079,815 is an accounting break-even: the point where operating profit is zero. But operating profit is not the last line of the plan. Below it sits \$69,500 of annual debt service — the payments on the equipment lease and the term financing that built the restaurant — and that money leaves the account whether or not the P&L is happy.

Cash break-even puts the obligations that must be paid in cash into the fixed base:

$$\text{Cash break-even} = \frac{\$437{,}635 + \$69{,}500}{0.4053} = \frac{\$507{,}135}{0.4053} = \$1{,}251{,}298$$

That is 77 dinner covers a night, not 66. Eleven more covers, every night, before the account stops going backwards. At Chapter 20's labor line it is \$1,317,918 and 81 covers a night, and the cushion against the plan's 95 falls to fourteen.

(One technical note, because precision matters here. Bellwether's plan — like most independent restaurant plans, and like the P&L in Figure 1.3 — shows debt service as a single figure below operating profit, so none of it has been expensed above. In a strict accrual statement the interest portion sits in the expense lines and only the principal repayment is the additional cash item. Know which format you are reading before you add anything.)

And even \$1,251,298 is not the real cash number, because at least four more things want money that no P&L line describes:

• Sales tax is not your money. It is collected daily, sits in your account for weeks making the balance look healthy, and is remitted monthly or quarterly. Chapter 31 said this and Chapter 33 will say it again: the cushion you think you see in the bank on the twentieth is partly a liability.
• Inventory build. Growing revenue requires more product on the shelf and in the walk-in. That is cash out with no expense attached until the food is sold. A restaurant scaling from \$1.4M to \$1.7M will fund several thousand dollars of additional inventory out of cash flow.
• Capital replacement. Bellwether's plan carries no depreciation line, so nothing is funding the fryer, the dish machine, the POS terminals, or the roof of the walk-in. A reasonable reserve is 1–2% of sales; at 1.5% that is \$23,250 a year**, which pushes cash break-even to **\$1,308,671 — 80 dinner covers a night.
• The owners. The chef-owner's \$55,000 salary is inside fixed labor, so it is covered. A distribution is not, and neither is the second partner taking market compensation.

🧾 Read the Numbers

```text FIGURE 32.7 — "The ladder of break-evens, in dinner covers a night" [the Bellwether plan] THE ARTIFACT Every break-even threshold computed in this chapter, converted to the same unit — dinner covers per night — and stacked against the plan and the ceiling. THE CONTEXT 68 seats. Dinner Tuesday-Saturday plus two brunches. Plan is 95 dinner covers a night at 1.4 turns. All conversions scale the whole cover pattern proportionally against $1,550,000 of plan revenue (Reading 1, §32.3).

covers/night
     0 ├── closed
    43 │ ▓                          a break-even TUESDAY (62 x 69.67%)
    60 │ ▓▓▓▓▓▓                     accounting break-even, PLAN labor $500,000
    62 │ ▓▓▓▓▓▓▓                    forecast Tuesday = the daily-nut rule of thumb
    66 │ ▓▓▓▓▓▓▓▓  <-- THE ANSWER   accounting break-even, Ch.19 labor $570,461
    68 │ ▓▓▓▓▓▓▓▓▓                  Q1 RAMP break-even, plan labor
    70 │ ▓▓▓▓▓▓▓▓▓▓                 accounting break-even, Ch.20 labor $597,461
    77 │ ▓▓▓▓▓▓▓▓▓▓▓▓▓              CASH break-even (+ $69,500 debt service)
    78 │ ▓▓▓▓▓▓▓▓▓▓▓▓▓▓             Q1 RAMP break-even, Ch.19 labor
    80 │ ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓            cash + 1.5% capital replacement reserve
    81 │ ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓           cash break-even, Ch.20 labor
    95 │ ████████████████████████   THE PLAN
   132 │ ░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░  OPERATING CEILING
       └────────────────────────────────────────  (hearth sustains 28 items/hr;
                                                   151 -> 144 -> 132 after demand
                                                   shape and table fit)

WHAT IT SHOWS Bellwether's viable operating band is roughly 66 to 132 dinner covers a night, and the plan sits at 95 — comfortably inside it, with 29 covers of cushion below and 37 of headroom above. But the useful reading is the CLUSTER between 77 and 81: cash break-even, the first-quarter ramp, and the Chapter 20 labor exposure all land within four covers of each other, and all of them are 11 to 15 covers above the headline figure of 66. If two of those three conditions coincide, Bellwether needs about 81 covers a night, and the cushion is 14, not 29. WHAT IT DOESN'T It cannot show timing. Every figure on this ladder is an annual rate expressed per night; none of them says WHICH nights, and the year is not flat. A business averaging 95 covers a night runs 62 on Tuesday and 123 on Saturday, and it runs a February that looks nothing like an August. It also cannot show the step costs: sustaining even 110 covers a night every night would require a station Bellwether has not built. THE DECISION Publish the number 66 to the management team as the operating threshold and 81 as the planning threshold, and put both on the weekly flash report from Chapter 31. Then build the 13-week cash forecast, because the ladder cannot tell you which week is the dangerous one. THE LESSON There is no such thing as THE break-even. There is a ladder of them, and the one you should manage against is the one that includes every dollar that actually has to leave the building. ```

It assumes a mix, a check, and a linear world

Four more limits, each of which has bitten somebody.

Break-even is mix-dependent, and the number moves without anything real changing. §32.3 showed two break-even revenue figures for the same restaurant — \$1,079,815 and \$1,097,498 — differing only in whether the softness landed on dinner or spread across everything. If your brunch share grows and your dinner share shrinks, your break-even in dollars rises even if every cost line is identical, because brunch's CM ratio is nearly eight points worse. Quote a break-even with its mix assumption attached or do not quote it.

It assumes the average check holds. The whole conversion from dollars to covers runs through \$46 and \$24. Chapter 31's comps, voids, and discounts sit underneath those figures, and a comp rate that drifts from 0.8% to 2.0% moves the effective check by more than fifty cents. So does a shift toward guests who skip the second glass of wine — which is exactly what happens in a soft economy, at the same time your cover count is falling. Check average and cover count are correlated, and break-even analysis treats them as independent.

It is linear, and restaurants have steps. The straight line in Figure 32.6 is honest only within a relevant range — roughly speaking, the volume band you can serve with the building, the equipment, and the org chart you have. Bellwether cannot serve 160 covers a night on this line at any price; the hearth sustains 28 items an hour and Chapter 7 walked the ceiling down from 151 to 144 to 132 on demand shape and table fit. Push toward the ceiling and you do not slide along the line, you hit a step: a new station, a second hood run, an added salaried position. Every step resets the fixed base and therefore resets the break-even.

It says nothing about whether the revenue is good revenue. Break-even is indifferent to whether your \$1,079,815 arrives from full-price dine-in guests or from a discount platform at a 28% commission. It would happily be cleared by revenue that damages the brand, exhausts the staff, and does not repeat. Chapters 23, 27, and 28 exist because the quality of a dollar is not visible in this arithmetic.

⚠️ Where the Money Leaks

The four sentences that mean somebody has misunderstood break-even.

"We're above break-even, so we're fine." Above the accounting break-even, on an annual average, at an assumed mix, before the loan payment. Four qualifiers, any one of which can be false in the week you are standing in.

"Our break-even is X." Whose? Computed when? With which labor line, and does it include debt service? At Bellwether, five defensible answers to that question span \$970,915 to \$1,308,671 — a range of \$337,756, or twenty covers a night.

"Break-even is low, so the risk is low." Break-even tells you where the floor is. It says nothing about whether you can reach the floor. A restaurant with a \$600,000 break-even in a trade area that will only ever produce \$500,000 has a low break-even and no business.

"We'll make it up in volume." Sometimes true, and §32.5 is why it is so tempting. But volume above the relevant range requires a step cost, and volume bought with discounting moves the CM ratio down while it moves revenue up. Compute the new break-even before the promotion, not after.

🔍 Check Your Understanding

1. Bellwether's accounting break-even is \$1,079,815 and its cash break-even is \$1,251,298. Explain the difference in one sentence, and name two things missing from both figures.
2. Why does adding \$1 of fixed cost raise break-even by \$2.47 rather than by \$1?
3. An operator tells you break-even is \$1,030,454 and margin of safety is 33.5%. Both figures come from the business plan. What is the first thing you would check?

(1: The cash figure adds the \$69,500 of annual debt service, which leaves the bank account but sits below the operating-profit line; both omit the sales-tax liability sitting in the account and any funding for capital replacement — and both omit inventory build and owner distributions. 2: Because only 40.53 cents of each additional sales dollar survives variable cost, so it takes \$1 ÷ 0.4053 = \$2.47 of sales to produce \$1 of contribution. 3: The fixed/variable split — specifically the labor line, because that is where Chapter 4's \$1,030,454 went wrong, and because a plan-era break-even almost always uses the plan's labor target rather than the schedule's labor reality.)

🍽️ The Business Plan

Checkpoint 32 of 40 — the Break-Even Analysis.

This is the section a skeptical reader turns to second, right after the revenue forecast. Its job is not to be reassuring. Its job is to demonstrate that you know exactly where your floor is and how much room you have above it.

Bellwether Restaurant — Break-Even Analysis (constructed teaching example)

A. Fixed costs — \$437,635 a year

B. The contribution margin ratio — 40.53%

Forty and a half cents of every dollar through the register survives to pay the fixed costs above.

C. Break-even sales — \$1,079,815

$$\frac{\$437{,}635}{0.4053} = \$1{,}079{,}815 \quad = \quad 69.7\% \text{ of the \$1,550,000 plan}$$

Sensitivities, stated rather than buried:

D. Break-even covers per night — 66 at dinner, 77 at brunch

Converted against the base cover pattern of 475 dinner covers a week at \$46 and two brunch services at 110 covers at \$24, scaled proportionally with the patio, event, and off-premise revenue:

What this section settles. Bellwether's floor, in the only two units that matter: \$1,079,815 of annual sales, or 66 dinner covers a night. The plan clears it by 30.3%, or twenty-nine covers a night, or about sixteen weeks of trading. All three revenue streams — dinner, brunch, and the bridge — carry their fully-allocated share of fixed cost. Operating leverage is 3.30, so a 10% revenue miss costs \$62,819, which is a third of the year's profit and nine and a half covers a night.

What this section does not settle, honestly stated.

1. Which labor line is real. The plan says \$500,000; the schedule says \$570,461; the compliance read says \$597,461. Break-even moves \$175,520 across that range — ten covers a night. This is now the single largest unresolved item in the plan and Chapter 40 has to answer it.
2. The first quarter. At the Chapter 9 ramp and Chapter 19's labor line, Q1's break-even runs at an annualized \$1,269,984 — 78 covers a night — against 95 forecast. The quarter clears by \$15,715 and a 12.6% margin. That is thin, and it is thin in the quarter with the least data and the most chaos.
3. Cash. Every figure above is an accounting threshold. The \$69,500 of debt service moves the floor to 77 covers a night, and nothing here funds capital replacement or the sales-tax remittance. This analysis cannot tell you which week is dangerous. Chapter 33 builds that.
4. The mix. Break-even in dollars assumes the plan's daypart mix. Brunch's CM ratio is 34.25% against dinner's 42.11%, so any drift toward brunch raises the dollar break-even with no cost line changing.

Open questions carried forward:

1. Can the labor line be brought to 34–35% without damaging the product, and if not, does the plan carry 66.3% prime cost? (Chapters 19, 20 — resolved in Chapter 40)
2. Which thirteen-week window in the year sits below break-even, and is there enough cash to cross it? (Chapter 33)
3. Does the \$2 brunch price increase modeled in §32.7 survive contact with guests? (Chapters 10, 24)
4. Is the working-capital reserve sufficient to cover the first-quarter gap between an accounting profit of \$15,715 and the cash the quarter actually consumes? (Chapter 33)
5. If revenue lands at the downside case rather than the plan, at what point does the operator act, and on what? (Chapter 39)

Conclusion

Break-even analysis is one division. Everything hard about it happens before the division, in the question of which costs move when volume moves — and Bellwether is a case study in how badly that can go in a document everyone has read three times. Chapter 4 produced \$1,030,454 from a labor split that was \$60,105 too high on the floor and 8.4 points too low on the slope, and nobody caught it because the total landed on the number the plan wanted. Chapter 5 produced \$986,700 by treating all labor as variable, said so out loud, and correctly warned that the real figure was higher. It is higher, by \$93,109.

Bellwether's break-even is \$1,079,815 of annual sales — 66 dinner covers a night, 77 per brunch service, 25,177 covers on the base pattern. The plan's 95 covers a night clears it with twenty-nine covers of cushion, about sixteen weeks of trading, and a margin of safety of 30.3%. That is a defensible position for a first-year independent, and it is not a comfortable one, because three separate conditions each push the threshold into the high seventies: the loan payment, the first-quarter ramp, and the sous chef's classification. If two of them coincide, Bellwether needs 81 covers a night and the cushion is fourteen.

Operating leverage is why those numbers deserve respect. At a degree of 3.30, a 10% revenue miss — nine and a half covers a night, two and a half tables, a number you could not see from the host stand — takes a third of the year's profit. The same arithmetic is why every dollar of new fixed cost costs \$2.47 of sales, and why the \$47,000 assistant general manager the chef-owner badly needs raises the floor by nine covers a night. Break-even does not decide that. It prices it, which is more useful.

And then there is what break-even cannot see, which is the whole reason it is not the last chapter of this part. It has no idea when money arrives or leaves. It cannot tell you that the sales tax in your account is not yours, that inventory has to be bought before it is sold, that payroll clears on a Thursday, or that the insurance premium and the liquor-license renewal both land in the month your revenue is lowest. Bellwether's break-even says the year is fine. It says nothing at all about February.

Chapter 33 is about February. Profit is an opinion formed over a period; cash is a fact on a Thursday morning, and the restaurants that close were, on paper, doing fine two months earlier. We are going to build the thirteen-week forecast that finds the week this chapter cannot.

Key Terms

Fixed cost — a cost that does not change with sales volume within the relevant operating range: rent, salaried compensation, insurance, budgeted marketing, technology subscriptions. Bellwether's fixed base is \$437,635 a year. (Ch. 32)

Variable cost — a cost that moves in direct proportion to sales: food and beverage cost, hourly wages, credit-card processing, guest supplies. Expressed as a percentage of sales, Bellwether's is 59.47%. (Ch. 32)

Semi-variable cost (also mixed cost) — a cost with both a fixed base and a variable slope, such as utilities, repairs, and breakage; split it with the high-low method before using it in any break-even calculation. (Ch. 32)

Contribution margin ratio (CM ratio) — sales minus variable costs, divided by sales; the fraction of each sales dollar that survives to pay fixed costs. Bellwether's is 40.53%. (Ch. 32)

Break-even point — the sales volume at which total contribution exactly equals total fixed cost and operating profit is zero; computed as fixed costs ÷ CM ratio. (Ch. 32)

Break-even covers — the break-even point expressed as guests rather than dollars; break-even sales ÷ average check, using a named revenue base. Bellwether's is 25,177 covers a year on the base dining-room pattern. (Ch. 32)

Margin of safety — expected sales minus break-even sales, in dollars, as a percentage of expected sales, in covers per night, or in weeks of trading; how far above the floor you are standing. (Ch. 32)

Operating leverage — the sensitivity of operating profit to a change in revenue, caused by the presence of fixed costs; the reason a 10% revenue decline produces a much larger percentage decline in profit. (Ch. 32)

Degree of operating leverage (DOL) — contribution margin ÷ operating profit; the multiplier that converts a percentage change in revenue into a percentage change in operating profit. Bellwether's is 3.30. (Ch. 32)

Break-even by daypart — computing contribution and break-even separately for each service, then testing it on both an avoidable-cost basis (for a close-it decision) and a fully-allocated basis (for a should-it-exist decision). (Ch. 32)

Cash break-even — the sales volume at which cash inflow covers all cash obligations, including debt service and capital replacement, not merely accounting expenses; always higher than the accounting break-even. Bellwether's is \$1,251,298 before a replacement reserve. (Ch. 32)

Spaced Review

1. Without looking back: state the break-even formula, then explain why getting the fixed/variable split wrong changes nothing on this year's P&L and everything about next year's decisions.
2. From Chapter 1. Prime cost is COGS plus total labor. Bellwether's plan says 60.0%; Chapter 19's schedule says 64.6%; Chapter 20's classification says 66.3%. Using the prime-cost diagnostic ranges from §1.3, characterize each of those three, and say which one you would put in a document a lender will read.
3. From Chapters 12 and 19. A restaurant's owner proposes cutting a \$14/hour prep shift, four hours a day, six days a week, to improve labor cost. The shift produces the house pickles, the stocks, and the salsa verde. Using contribution-margin reasoning and the fixed/variable distinction from §32.1, name three ways this could raise total cost rather than lower it.
4. From Chapters 24 and 26. Bellwether's break-even is 66 dinner covers a night against an operating ceiling of about 132. A delivery platform offers to fill shoulder hours at a 25% commission on a \$38 average order. Card processing on a dine-in cover is 2.81%. Sketch, without computing precisely, why that offer might clear break-even and still be a bad decision.
5. The recurring question. An operator computes break-even once, in the business plan, and quotes it for four years. Name four specific things that will have changed the answer by year three, and say which of the four you would expect them to notice on their own.