Chapter 9: Opening: The Pre-Opening Countdown, Hiring Up, Soft Open, and the First Ninety Days

"You do not open when you are ready. You open when the money says you open. The entire job is making those two dates the same date." — constructed; the sentence every second-time operator wishes somebody had said to them the first time

Overview

The room is a room. The hood runs, the hearth threw its first fire on Tuesday, and there is still an electrician somewhere behind the dish pit. Twenty-three people are on payroll and not one of them has served a guest. The bank account holds whatever survived the build-out, and from here every single week costs money in exactly one direction.

This is the part of an opening nobody photographs. The photographs are of the first Saturday — the full room, the bar three deep, the chef at the pass. What produced that Saturday was six weeks of paying people to practice, an inventory bought on credit you had to establish from nothing, and a sequence of inspections held by people who do not work for you and do not care about your lease.

Here is the operating question this chapter exists to answer: how much money do you need between the day the contractor hands you the keys and the day a guest first pays you — and what happens if the number in your plan is wrong? Bellwether's plan carries \$35,000 for pre-opening. That figure was set in Chapter 4, before anybody had costed a training hour or priced a bar's worth of opening liquor. We are going to build it from the bottom, line by line, and we are not going to flinch at the result.

The second question is about time. A restaurant opening is a chain of dependencies, and several of the links belong to a plans examiner, a health authority, a licensing board, and a factory that builds make-up air units. You cannot compress those links. What you can do is decide where in the chain your payroll starts, and that single decision is worth more than any negotiation in Chapter 6 except one — the sentence that tied rent commencement to the certificate of occupancy, which handed this chapter three months of free rent and made it the most valuable pre-opening asset on the balance sheet.

The third question is what to do once the doors are open, when everything is going better than you expected and that is the most dangerous condition a new restaurant can be in.

In this chapter, you will learn to:

• Build a pre-opening budget from the bottom up and say out loud, in dollars, where it lands against the figure the plan carries.
• Work a twelve-month countdown backward from opening day and find the four gates you do not control.
• Compute what an opening crew costs before revenue exists, and what one month of slippage costs after they are hired.
• Sequence hiring so people arrive when they can be trained, not when the schedule looks tidy.
• Run a menu rehearsal and a soft open that generate numbers, not compliments.
• Recognize the honeymoon period while you are inside it, and instrument the business for the week it ends.

Learning Paths

🏗️ Opening — all of it, twice. §9.1 and §9.2 are the two sections most likely to save you real money, and §9.2's thirty-day rule is the single most actionable paragraph in Part II. 📋 Managing — weight §9.3, §9.4, and §9.7. You will more often join an opening than own one, and the manager who understands the ramp is the one who does not panic in week nine. 🍸 Beverage — §9.1's opening-inventory build is yours; the bar is the most capital-hungry inventory in the building. §9.2's licensing gate decides whether you open wet or dry. 🚚 Small Format — the numbers shrink and the structure does not. A truck has a countdown, a commissary inspection, a soft open, and a honeymoon, and its reserve is proportionally thinner.

9.1 The pre-opening budget: what it covers and why it is always underestimated

A pre-opening budget is the money a restaurant spends between the point at which the space stops being a construction project and the point at which a paying guest first walks in. It covers payroll for people who cannot yet produce revenue, the food and beverage consumed in learning to cook and serve the menu, licenses and certifications, the first full inventory, and the ordinary operating costs — utilities, insurance, supplies, marketing — that begin before the register does.

It is the most reliably underestimated line in a restaurant project, and the reason is structural rather than moral.

Every other line in the budget is a thing. Construction is a building you can walk through. Equipment is a hearth and a walk-in and a dish machine you can stand next to and point at. Smallwares are plates in a box. When a number attached to a thing is wrong, somebody notices, because there is a physical object failing to match an invoice.

Pre-opening is not a thing. It is four hundred hours of a line cook learning where the low-boy is. It is a hundred and eighty pounds of protein cooked, tasted, adjusted, and thrown away. It is a food-handler card for a busser you will lose in ninety days. None of it is visible in the finished restaurant, which means nobody misses it when it is cut, and the operator who cuts it does not discover the cost for three months — at which point the cost shows up somewhere else entirely, under a different name, on a line called prime cost.

Chapter 1 named undercapitalization as the most common cause of first-year failure and described its mechanism precisely: an operator who budgets exactly enough to open has, by definition, budgeted nothing with which to operate. The pre-opening line is where that failure is manufactured. It sits at the seam between building a restaurant and running one, and it is the line an optimistic planner squeezes because it is the only line without a photograph attached.

⚠️ Where the Money Leaks

The four things a first-time pre-opening budget always forgets.

In rough order of how much they cost:

1. Opening inventory. People budget "food." A full-service restaurant with a bar needs a walk-in, a freezer, a dry store, a back bar, a wine inventory, kegs, and non-alcoholic product — all of it on day one, and most of it on cash terms because you have no trade credit history yet. This is routinely the largest single line and it is routinely the one that appears last.
2. The hours nobody scheduled. Training is not the only pre-opening labor. Receiving equipment, organizing a dry store, labeling shelves, building the point-of-sale menu, setting pars, walking a punch list, deep-cleaning after the contractor's "final clean" — all of it is compensable time and none of it appears on a training calendar.
3. Utilities from the moment the meter is in your name. A walk-in pulled down to temperature in week −5 runs twenty-four hours a day for five weeks before it holds a dollar of product.
4. The owners' own income. Two partners who leave jobs six months before opening have created a household cash-flow problem that appears nowhere in the project budget and does not care that it appears nowhere.

Building it from the bottom

Bellwether's plan carries \$35,000** for pre-opening, inside a **\$620,000 project, alongside a **\$45,000** working-capital reserve. That \$35,000 was an allocation, not a build — a reasonable placeholder set in Chapter 4 before there was a menu, a crew, a lease, or a schedule. There is now all four. So let us build the real number.

Every figure below is constructed and illustrative, chosen to be defensible for a 68-seat full-service restaurant with a full bar in a mid-size Midwestern market. Your market's wages, utilities, and fees will differ, sometimes by a lot. The structure is what transfers.

🧾 Read the Numbers

```text FIGURE 9.1 — "The pre-opening budget, built from the bottom up" [the Bellwether plan] THE ARTIFACT A bottom-up pre-opening budget for a 68-seat chef-driven American restaurant with a full bar, dinner Tuesday–Saturday plus weekend brunch, built against the twelve-month countdown in Figure 9.2. THE CONTEXT Rivermill District. Lease signed, rent commencement tied to the certificate of occupancy, three months' abatement. Construction line $310,000, equipment $185,000, smallwares and FF&E $45,000. The plan's pre-opening allocation is $35,000 and its working-capital reserve is $45,000. Nothing below has been spent yet; this is what it will cost.

  LABOR BEFORE REVENUE
    Hourly wages, 1,504 hours (Figure 9.4)                       $ 27,640
    Employer burden at 14% — FICA, unemployment,
      workers' compensation (14% of $27,640 = $3,869.60)            3,870
    Partner draws (assumed at zero — see below)                         0
                                                                 ─────────
    LABOR SUBTOTAL                                               $ 31,510

  TRAINING, REHEARSAL, AND THE SOFT OPEN
    Menu rehearsal and R&D food                                  $  3,200
    Staff tasting; wine and cocktail education                      1,400
    Friends and family + soft open, 200 covers at $16               3,200
                                                                 ─────────
    TRAINING F&B SUBTOTAL                                        $  7,800

  OPENING INVENTORY
    Food — about 1.2 weeks of planned usage                      $  7,700
    Beverage — spirits, ~40-bottle list, beer, non-alcoholic        8,120
                                                                 ─────────
    OPENING INVENTORY SUBTOTAL                                   $ 15,820

  LICENSING, PERMITS, CERTIFICATION   (allocation — Chapter 8 prices it)
    Business license, food establishment permit, fire, sign,
      health plan review and pre-opening inspection fees         $  1,900
    Liquor license application and responsible-service training      1,800
    Food-handler cards (23 at $18 = $414) and two manager
      certifications (2 at $180 = $360)                                774
                                                                 ─────────
    LICENSING SUBTOTAL                                           $  4,474

  EVERYTHING ELSE THAT STARTS BEFORE REVENUE DOES
    Utilities, final three months ($450 / $850 / $1,100)         $  2,400
    Pre-opening marketing — site, photography, print,
      neighborhood announcement                                     3,000
    Uniforms and aprons, opening crew (23 at $52)                   1,196
    Smallwares and glassware broken in training                      1,200
    POS configuration, menu build, install labor                     1,800
    Opening chemicals, paper, packaging, cleaning                    1,400
    Recruiting — postings, screening, background checks                700
                                                                 ─────────
    OTHER SUBTOTAL                                               $ 11,696
                                                                 ═════════
    BOTTOM-UP TOTAL                                              $ 71,300
    THE PLAN'S FROZEN PRE-OPENING LINE                             35,000
                                                                 ─────────
    GAP                                                          $ 36,300

WHAT IT SHOWS The honest cost of opening this restaurant is $71,300, which is $36,300 — 104% — above the line the plan carries. Two blocks do most of the damage and neither is discretionary: labor before revenue at $31,510, and opening inventory at $15,820. Together they are $47,330, already $12,330 more than the entire budgeted line, before a single training meal is cooked or a single permit fee is paid. WHAT IT DOESN'T It carries nothing for the partners' own compensation; the build assumes they work the final six months unpaid. It excludes the security deposit and utility deposits, which are cash out the door but not expenses. It excludes the first premium instalment on the insurance schedule Chapter 8 builds — general liability, liquor liability, property, workers' compensation — which lands before opening and is not in this total. It assumes nobody crosses forty hours in a week, which in the final fortnight is fiction, and it therefore carries no overtime premium. And it assumes the schedule holds: every figure here is a function of a date. THE DECISION Do not revise the $620,000 and do not quietly move money between buckets. Publish the $71,300 in the plan next to the $35,000, name the $36,300, and state precisely which pot absorbs it — which, on these facts, means the rent abatement and the working-capital reserve. Then re-read Chapter 1 §1.4 and decide whether you are comfortable with what is left. THE LESSON The pre-opening line is the only line in a restaurant project with no physical object attached to it, which is exactly why it is the line that gets guessed. A budget nobody can walk through is a budget nobody audits. ```

Two clarifications about that build, because both are places where a careless reader would go wrong.

The 14% employer burden is illustrative and it is not benefits. It covers the employer share of FICA at 7.65%, federal and state unemployment insurance, and workers' compensation premium — and workers' compensation for restaurant classifications varies enormously by state and by experience rating. Some operators land nearer 11%, some nearer 18%. Verify yours before you build a budget on it, and note that this build carries no health insurance, no paid leave, and no retirement match.

The partner draws sit at zero, and that is a disclosure, not a saving. Bellwether's build assumes the chef partner and the front-of-house partner take nothing during pre-opening. This is what actually happens in most independent openings and it is why the plan can show \$35,000 at all. But the labor is real. An executive chef's salary in a market like this one runs, illustratively, somewhere in the seventies; leaving that job six months before opening forgoes roughly \$37,500 of household income, and the front-of-house partner's number is not much smaller. Call the pair of them \$68,500 of forgone earnings over six months — nearly twice the entire pre-opening line, appearing nowhere in the \$620,000, and still requiring two people to pay rent and eat.

Write it in the plan's risk section in plain words. A reader who has financed restaurants before will look for it, and its absence is more alarming than its presence.

🔍 Check Your Understanding

1. Why is opening inventory a larger pre-opening line for a full-service restaurant with a bar than for a fast-casual restaurant of the same seat count?
2. The build above puts partner draws at zero. Explain in one sentence why that makes the \$35,000 figure look more achievable and in one sentence why it makes the business more fragile.
3. Which two blocks of the \$71,300 are effectively non-discretionary, and what is their combined total against the frozen line?

(1: A bar carries depth, not turnover — a back bar, a forty-bottle list, kegs, and non-alcoholic product must all be present on night one even though a given bottle may sell over months, whereas a fast-casual kitchen buys a short list of high-velocity items. 2: It removes roughly \$68,500 of real labor from the visible budget; and it means the two people the business depends on enter the hardest ninety days of their working lives with no personal cash cushion. 3: Labor before revenue \\$31,510 and opening inventory \\\$15,820, totalling \\$47,330 against a \\\$35,000 line.)

9.2 The countdown: twelve months to opening day, working backward

Every first-time operator builds the schedule forward. You sign a lease, you start design, you hope, and opening day is whenever the chain of events happens to end. This is how the opening date becomes a rumor, and a rumor is not something you can hire against.

Experienced operators do the opposite. They pick a date and work backward, because a pre-opening timeline — the dated, dependency-ordered sequence of everything that must happen before a restaurant can serve its first guest — only becomes a management tool when it terminates in a commitment. Backward planning forces you to confront the two facts forward planning lets you avoid: which items have lead times you cannot compress, and which items you are not allowed to start until somebody else finishes.

Picking the date first

Bellwether's plan sets opening day as the first Tuesday in April, and the reasoning is commercial rather than sentimental.

The plan's \$1,550,000 first-year revenue includes a **\$139,240 bridge above the base dining room — patio dinner and brunch across a twenty-week season, fourteen private events, and a takeout line (Chapter 4 built it; Chapters 24, 28, and 29 own the components). Almost none of that bridge is available in a restaurant's first quarter. A patio is seasonal. Private events require a sales cycle and a reference. Takeout builds slowly from a standing start. So the calendar question is: do you want your first quarter to be the quarter with no bridge and no season, or the quarter with no bridge but a patio arriving in month two?

Open in April and the patio opens in month two, the summer lands in months two through six — exactly when the reserve is thinnest — and February, the cruelest month in most American markets, arrives in month eleven with eleven months of systems behind it. Open in November and you get the opposite of every one of those. Chapter 33 works the seasonality and the cash properly. For now, note the principle: the opening date is a revenue decision before it is a construction decision.

The twelve months

FIGURE 9.2 — The twelve-month countdown, worked backward from opening day  [the Bellwether plan]

  Legend:  ███ work in progress    ▼ a gate (an event you do not control the date of)
           ●━━● lead time on order  ▲ a fixed event you schedule    ? unknown duration

  MONTHS BEFORE OPENING DAY
                            12   11   10    9    8    7    6    5    4    3    2    1    0
                             │    │    │    │    │    │    │    │    │    │    │    │    │
  MONEY AND PAPER
   Plan, forecast, the ask   ██████████
   Entity, EIN, bank acct         █████
   Lender package, closing        ███████████
   Insurance bound                          ▼              ▼              ▼
  THE SPACE  (Chapter 6)
   Search, tour, shortlist   ███████████
   LETTER OF INTENT                    ▼
   Due diligence, 45 days               ███  ← the hood is found on day 9
   Lease negotiation                     █████
   LEASE SIGNED                              ▼
  DESIGN AND PERMIT  (Chapter 7)
   Drawings to 100%                          █████████
   Health plan review                             █████
   Building plan review                            ███████
   PERMIT ISSUED                                          ▼
  LICENSES  (Chapter 8 maps this)
   Liquor license filed                      ▲━━━━━━━━━━━━━━━━━━━━?  longest lead; binary
   Operating permits                                        ████████
  CONSTRUCTION  (Chapter 6)
   Long-lead equipment                            ●━━━━━━━━━━━━━●   hood, MUA, walk-in, hearth
   Demolition                                             ████
   MEP rough + inspections                                   █████
   Finishes, millwork                                          ██████
   Equipment set + connect                                        ████
   Final inspections + fire                                        ███
   CERTIFICATE OF OCCUPANCY                                          ▼   ← T−30 days
   Health pre-opening inspection                                     ▼
  PRODUCT  (Chapters 10, 11, 13, 26)
   Menu built and costed                                     ███████
   Vendor accounts, specs, pars                                 █████
   POS build, reservation platform                                ████
  PEOPLE  (Chapters 17, 18)
   Management on payroll                                            ███
   Hourly hiring and offers                                          ██
   Certification, paperwork                                          ██
   Menu rehearsal + plate-ups                                         ██
   Floor training, simulated service                                   █
   FRIENDS AND FAMILY ×2                                              ▲
   SOFT OPEN ×2                                                        ▲
   OPENING DAY                                                          ▼

  ── the thirty-day rule ─────────────────────────────────────────────────────────
  Rent commences on the earlier of opening or 30 days after the certificate of
  occupancy (Chapter 6). Three months' abatement = $23,800. Open inside those thirty
  days and you burn none of it. Every week past it costs $1,831.

Read the diagram against Chapter 6's Figure 6.7 and you will see the same critical path with two months bolted onto the front — the plan, the financing, the entity, and the site search that had to happen before there was a letter of intent to draw. Chapter 6's figure ran from the LOI; this one runs from the decision to do it at all.

Four items on that chart are gates: events whose timing belongs to somebody else. Building plan review. The certificate of occupancy. The health department's pre-opening inspection. And the liquor license, whose timeline Chapter 8 maps and whose duration in some jurisdictions is measured in quarters rather than weeks. You can influence a gate by submitting a complete package the first time and by hiring people who have been through that particular department before. You cannot schedule one.

Everything else on the chart is yours. Which means the only real scheduling skill in an opening is knowing which of your own activities to start before which gate — and the answer is almost always "later than feels comfortable."

The thirty-day rule

Chapter 6 bought this chapter its single most valuable asset with one sentence: rent commences on the earlier of opening or thirty days after the certificate of occupancy, and the first three months are abated. That is \$23,800 of free rent, and where it lands is entirely up to you.

The abatement can be spent three ways.

Spent on the ramp — the intended use. Rent commences when you open, the three abated months cover months one through three of operation, and \$23,800 of occupancy cost simply does not happen during the quarter when revenue is least predictable. This is the version Chapter 6 negotiated for.

Burned before opening — the common failure. The certificate of occupancy is issued and then you take ten weeks to open. Rent commenced at day thirty; six of those weeks are unabated-in-effect, consuming \$10,986 of abatement while the dining room sits empty. Nothing about that is dramatic. It happens one week at a time, for good reasons, and nobody notices until the free rent runs out in month two of operation instead of month four.

Spent on the honeymoon — the trap. Because the abatement lands on months one through three, and because months one through three are also when novelty traffic is at its peak (§9.6), the first quarter's cash looks superb for two independent reasons, both temporary, both ending at almost the same time. We will come back to this, because it is where operators do the most self-inflicted damage.

So: the thirty days between the certificate of occupancy and rent commencement is your entire abatement-free training window, and it is exactly long enough. Design the countdown so that everything requiring a legally occupiable building — rehearsal in the actual kitchen, floor training, friends and family, the soft open — fits inside it, and so that everything that does not require the building is already finished when the certificate is issued.

Things that do not require a certificate of occupancy: menu development and costing, vendor accounts and specifications, par levels, the point-of-sale menu build, the reservation platform, recruiting, interviewing, offers, handbooks, classroom training, wine and spirits study, the service-standards document, cash-handling procedures, and the marketing calendar. That is most of the work. Do it in a borrowed office.

FIGURE 9.3 — The last ninety days, week by week                        [the Bellwether plan]

  WK    THE BUILDING                   THE TEAM                     THE MONEY
  ────────────────────────────────────────────────────────────────────────────────────────
  −13   MEP rough complete             Job descriptions written     Vendor credit
        Rough inspections passed       Postings live                applications in
  −12   Finishes start                 Management interviews        Insurance quotes
  −11   Millwork set                   Sous chef offer signed       Opening orders quoted
  −10   Equipment delivered            Reference checks             POS contract signed
   −9   Equipment set + connected      Hourly interviews begin      Deposits paid
   −8   Final trade inspections        Offers out to BOH            —
   −7   Fire inspection                —                            Utility accounts
        Punch list walk #1                                          transferred
   −6   Contractor corrections         SOUS CHEF ON PAYROLL ◄────── the payroll clock starts
                                       Menu costing, specs, pars    Opening orders sized
   −5   Punch list corrections         Offers out to FOH            Trade credit approved
   −4   CERTIFICATE OF OCCUPANCY ▼     Bar manager, cooks, prep     RENT CLOCK STARTS:
        Health pre-opening insp. ▼     on payroll                   30 days of abatement-
                                       I-9s, W-4s, handbooks        free runway
   −3   Deep clean                     MENU REHEARSAL begins        Opening food order
                                       Every item cooked to card,   placed
                                       weighed, plated, timed       Wine + spirits received
   −2   Signage, final finishes        PLATE-UPS: full menu tasted  Bar par set
                                       by every server              Banking, tip policy,
                                       Floor training, POS drills   comp accounts live
   −1   Nothing changes                FRIENDS AND FAMILY ×2        Soft-open comps coded
        Punch list walk #2, at night   SOFT OPEN ×2                 separately
    0   —                              OPENING DAY                  Day-one sales report
  ────────────────────────────────────────────────────────────────────────────────────────
  Weeks are numbered backward from opening day. Everything above week −4 can be done
  without a certificate of occupancy. Nothing below it can.

🧮 Run the Numbers

What a four-week slip costs after the crew is hired.

Suppose the certificate of occupancy lands on schedule at week −4 and then something ordinary happens — a failed final inspection, a health re-inspection, a licensing hold — and opening day moves four weeks to the right. The crew is already on payroll. What does that cost?

The abatement. Rent commenced thirty days after the certificate of occupancy, which is the original opening day. Four extra weeks of rent at \$95,200 ÷ 52 = **\$1,831 a week** is $4 \times \$1{,}831 = \$7{,}324$ of abatement consumed before a guest arrives.

The payroll. You will not send everyone home — you would lose them — so you carry management full-time and the hourly crew at reduced hours:

	Hours	Rate	Wages
Sous chef	40	\$24.00 | \$960
Bar manager	40	\$26.00 | \$1,040
Line cooks (3)	16 each	\$19.00 | \$912
Prep cook	16	\$17.00 | \$272
Bartenders (2)	12 each	\$16.00 | \$384
Servers (8)	10 each	\$15.00 | \$1,200
Hosts (2)	8 each	\$16.00 | \$256
Dishwashers (2)	8 each	\$15.00 | \$240
Runners / bussers (3)	8 each	\$15.00 | \$360
Weekly wages			\$5,624
Employer burden at 14%			\$787
Weekly cost of standing still			\$6,411

Four weeks: $4 \times \$6{,}411 = \$25{,}644$.

Total cost of the slip: $\$25{,}644 + \$7{,}324 = \$32{,}968$ — 94% of the entire frozen pre-opening budget, spent on nothing. And that ignores the two or three people who will take another job during the delay, whose replacement cost Chapter 17 prices properly.

The conclusion is not "hire later and train less." It is this: the payroll start date is the only variable in this equation you control, so do not commit it until the date is real. A date becomes real when the certificate of occupancy is issued and the licenses required to serve what you intend to serve are in hand. Before that, an opening date is a forecast, and you do not put twenty-three people on payroll against a forecast.

The item you must not put on the critical path

The liquor license is the sharpest version of this problem, because it is binary. A restaurant with a full bar, planning a 28% beverage mix, either has the license or does not, and no amount of management skill produces a partial one.

🧮 Run the Numbers

Open dry, or delay?

Bellwether's in-room plan is \$27,130 of weekly sales at a 72/28 mix, so beverage is $\$27{,}130 \times 0.28 = \$7{,}596$ a week. At a 22% pour cost the beverage line contributes $\$7{,}596 \times 0.78 = \$5{,}925$ a week.

• Open dry for four weeks: you forgo roughly $4 \times \$5{,}925 = \$23{,}700$ of contribution — assuming, generously, that you hold every cover, which a full-bar concept opening without a bar will not.
• Delay four weeks: \$32,968, computed above.

On arithmetic alone, opening dry is \$9,268 cheaper. On judgment, both answers are bad, and the lesson is that you have already lost by the time you are choosing between them.

The rule: never place opening day on the critical path of an item whose date you do not control. Set opening day after the license is in hand, and build the countdown backward from that. Chapter 8 maps the licensing path and explains why in some jurisdictions this single item can dominate the entire schedule.

⚖️ Code and Compliance

What has to be true before a guest — any guest, paying or not — is in your building.

All of this varies by state, county, and city, and it changes. Verify locally, in writing, with the actual authorities, before you put a date on anything. The structure below is the shape of the obligation, not the rule where you live.

• The certificate of occupancy is a gate, not a formality. Occupying a space for a use it is not certified for exposes you to enforcement and, in many policies, voids coverage. Chapter 8 owns this document.
• The health authority's pre-opening inspection is separate from the certificate of occupancy and frequently has its own lead time and its own queue. Schedule it early; a failed one is a re-inspection, not a conversation.
• A friends-and-family service is a food service. The food code does not contain an exemption for guests who did not pay. Your permits, your certified manager, your logs, and your handwashing discipline must all be in place for service one.
• Alcohol without a license is a licensing offense in essentially every American jurisdiction, and in many of them giving it away is treated the same as selling it. Do not pour at a soft open on the theory that nothing was charged. The downside is not a fine; it is your application.
• Whether you may cook and train in the space before the certificate issues varies. Some jurisdictions permit it, some issue a temporary certificate, some prohibit occupancy entirely. Ask the building official directly and get the answer in writing, because your whole rehearsal calendar depends on it.

9.3 Hiring and training up: sequencing, payroll before revenue, and the burn

This is where the money actually burns, and it burns for a reason that is not negotiable: a restaurant cannot be trained after it opens. Every hour of competence that does not exist on opening night is paid for in comps, in ticket times, in over-portioned plates, and in guests who do not return — and Chapter 23 will show you that the last of those is the expensive one.

Chapter 17 owns hiring — sourcing, the structured interview, the stage, references, Form I-9, onboarding. Chapter 18 owns training — the program, the manual, certification, menu knowledge. This section owns the two things neither of them can: the order the people arrive in, and what carrying them costs.

The sequencing rule

Hire backward from what must be trained, not forward from what you can afford.

The rule in one sentence: a position joins the payroll on the first day it can do work that could not have been done without it. Applied to Bellwether:

• The sous chef comes first, six weeks out, because the menu has to be costed, the specs written, the par levels set, the equipment received and organized, and the opening orders sized — and none of that can be done by someone who arrives two weeks before service.
• The bar manager comes next, four weeks out, because the bar has the deepest inventory in the building and somebody has to build the back bar, receive it, count it, write the cocktail specs, and train two bartenders on them.
• Line cooks and the prep cook come at four weeks, because they need to cook every item on the menu multiple times before anybody outside the building tastes it.
• Bartenders at three weeks; servers, hosts, and dishwashers at two; runners and bussers at one. Nothing is gained by having a busser on payroll for a month.

This is not stinginess. It is that a person hired six weeks early and given nothing to do learns that this restaurant wastes time, and that is a lesson you cannot un-teach.

🧮 Run the Numbers

What it costs to carry a team before a dollar arrives.

```text FIGURE 9.4 — Pre-opening payroll build [the Bellwether plan]

POSITION # ON PAYROLL HOURS TOTAL RATE WAGES FROM EACH HOURS ───────────────────────────────────────────────────────────────────────────── Sous chef 1 week −6 240 240 $24.00 $ 5,760 Bar manager 1 week −4 120 120 26.00 3,120 Line cooks 3 week −4 100 300 19.00 5,700 Prep cook 1 week −4 100 100 17.00 1,700 Bartenders 2 week −3 60 120 16.00 1,920 Servers 8 week −2 46 368 15.00 5,520 Hosts 2 week −2 40 80 16.00 1,280 Dishwashers 2 week −2 40 80 15.00 1,200 Runners / bussers 3 week −1 32 96 15.00 1,440 ───────────────────────────────────────────────────────────────────────────── TOTALS 23 1,504 $ 27,640 Employer burden at 14% 3,870 ───────────────────────────────────────────────────────────────────────────── PRE-OPENING PAYROLL, ALL IN $ 31,510 ```

Three readings of that number.

As a share of the budget: \$31,510 is 90% of the entire \$35,000 pre-opening line, spent on labor alone, leaving \$3,490 for inventory, licensing, utilities, marketing, and everything else.

As time: the plan's all-in labor is \$500,000 a year, or \$9,615 a week. So the pre-opening payroll is $\$31{,}510 \div \$9{,}615 = 3.28$ weeks — just under three and a third weeks of steady-state payroll, spent before the first dollar of revenue exists. That is the honest measure of what "hiring up" means.

As a rate: 1,504 hours across 23 people is about 65 hours each. For a chef-driven menu, a cocktail list, and a forty-bottle wine list, 46 hours of server training is not generous; it is about six shifts, and most of the operators I know who opened well wish they had bought more.

One sensitivity worth carrying. The industry runs roughly 75% annual turnover, and opening attrition is front-loaded — some of the opening crew were never going to work out and both of you will find that out in week three. Hiring two extra servers costs $2 \times 46 \times \$15 = \$1{,}380$ in wages plus \$193 of burden = **\$1,573**. Against the cost of running a Saturday section short in week three, that is the cheapest insurance in this chapter. Chapter 17 prices turnover properly; here it is simply a reason to over-hire the floor slightly and not the kitchen.

⚖️ Code and Compliance

Training hours are hours worked. All of them.

This is the single most common wage-and-hour exposure in a restaurant opening, and it is almost always created by enthusiasm rather than malice. Verify the specifics in your jurisdiction; several states are stricter than the federal floor and several have no tip credit at all.

• Under the Fair Labor Standards Act, time an employer requires or permits an employee to work is compensable. Training sessions, mandatory meetings, orientation, uniform fittings, menu study conducted on premises, and cleaning before opening are hours worked. The sentence "come help us set up and we'll take care of you when we open" is a wage claim with a friendly tone.
• Servers in training are usually not performing tipped work, and in tip-credit jurisdictions the credit generally cannot be applied to hours in which no tips can be earned. Budget the full applicable minimum for pre-opening server hours. Chapter 20 covers the tip credit and its conditions in detail; do not improvise here.
• Overtime applies during pre-opening exactly as it does after. Figure 9.4 assumes nobody crosses forty hours in a week. In the final fortnight, several people will. Every hour over forty for a non-exempt employee is time-and-a-half, and a chef who is salaried but does not meet the applicable exemption tests is non-exempt regardless of the title on the offer letter.
• Form I-9 must be completed within the statutory window from the first day of work, which during an opening means you need a system, not a folder. Chapter 17 covers the paperwork of day one.
• Certifications have lead times. Food-handler cards and manager certification take scheduling, sitting, and sometimes a wait for results. Book them in week −5, not week −1.

👨‍🍳 On the Line

What week minus two actually looks like.

Fourteen days out. The room smells like new paint and degreaser. The point-of-sale terminals came up yesterday and half the modifiers are wrong, so somebody is sitting at the bar with a laptop fixing "no onion" while eight servers do a walkthrough of the dining room learning table numbers from a printed floor plan that does not match the banquette the millworker moved.

The walk-in is at temperature and half full. Nobody knows where anything is, so everything takes four times as long, and the sous is relabeling shelves for the third time because the first two layouts did not survive contact with an actual delivery.

In the kitchen, the line is cooking the menu for the fourth time. The third time it was good. The fourth time it is good and the same, which is the actual goal. Somebody is standing at the pass with a timer.

Here is what a first-time owner does at this point and should not: adds things. A new dessert. A different plate. A cocktail somebody thought of. Every addition costs a rehearsal cycle you no longer have, and the menu you open with should have been frozen in week −4.

Here is what an experienced one does: stops changing things and starts repeating them. The last two weeks are not for improving the restaurant. They are for making the restaurant boring enough to execute on a Saturday. You have your whole first year to make it better. You have fourteen days to make it consistent.

And one small thing that matters more than it sounds: feed them. Family meal starts on day one of training, not on opening day. Twenty-three people who have eaten together for two weeks are a different organism from twenty-three people who met at a meeting. Chapter 21 makes the argument properly. It costs about eighty dollars a service and it is the highest-return money in this chapter.

9.4 Menu rehearsal, plate-ups, and timing the kitchen before guests exist

A menu rehearsal is the systematic production of every item on the opening menu, to the standardized recipe, under service conditions, before any guest is served — for the purpose of verifying the recipe, the yield, the plating standard, the station assignment, and the fire time.

That definition is deliberately unromantic, because the failure mode here is romance. What most first-time restaurants do in the last month is cook for each other — the chef makes the menu, the partners taste it, everyone is delighted, and nothing is written down. That is a tasting. A tasting is a pleasant evening. A rehearsal is a measurement exercise that happens to produce dinner.

What a rehearsal actually produces

Six artifacts, and if you do not have all six at the end you did a tasting.

1. A confirmed cost card for every item. Chapter 11 builds cost cards; the rehearsal is where you find out whether the yields on them are true in your kitchen with your equipment and your cooks. A three-percent yield difference on a protein is a real number and it is discovered here or it is discovered in month five as an unexplained variance.
2. A plating photograph for every item, taken from the guest's angle, printed, and hung where the line can see it. This is the single cheapest consistency control in a restaurant and almost nobody does it before opening.
3. A weighed portion standard for every component that is portioned by eye. Chapter 11 owns portion control; the rehearsal is where the number gets set and the scoop gets chosen.
4. A station map — which station touches which item. If four of your nine entrées come off the same station, you do not have a cook problem, you have a menu problem, and the time to find it is now, when the fix is moving a dish rather than hiring a body.
5. A fire time for every item, measured with a timer, not estimated. These become the coursing logic the expediter uses. Chapter 14 owns ticket times and the pass; you are giving that chapter its inputs.
6. An 86 plan. What the kitchen does, and what the floor says, when an item runs out on a Saturday at 7:40. Decide it in rehearsal and it is a procedure. Decide it in service and it is a scene.

Plate-ups

A plate-up is a service-staff tasting at which every item on the menu is plated exactly as it will go to a table, presented by the person who cooks it, tasted by every member of the service staff, and documented.

It is not a perk and it is not optional. A server who has never eaten a dish cannot describe it, and a server who cannot describe it defaults to "that one's really popular," which is the sound of your average check going down. Every plate-up should capture, in writing that the staff keeps: the ingredients in order, the cooking method, the allergens, what can and cannot be modified, the fire time, and one honest sentence about who the dish is for.

Run it at least twice — once at week −3 when the kitchen is still adjusting, and once at week −2 with the final version. And run the beverage equivalent: every cocktail made and tasted, every by-the-glass wine poured and discussed. Chapter 16 will show you that a wine list nobody can describe is inventory that ages on your balance sheet.

Timing the kitchen against a number

Chapter 7 designed this kitchen to produce 95 covers on a plan night. The rehearsal is where you find out whether the design is true, and the way you find out is to give it a number to hit.

🧮 Run the Numbers

What the kitchen actually has to do in the peak hour.

A plan dinner is 95 covers across roughly two and a half hours of service. Demand is not flat: in a neighborhood dinner house, something on the order of 35% of covers arrive in the peak hour — call it 7:00 to 8:00 on a Friday.

$95 \times 0.35 = 33.25$ covers in sixty minutes.

But the kitchen does not send covers, it sends plates. At a chef-driven American restaurant where a typical table takes an appetizer to share and most guests take a dessert or the cheese, assume about 1.8 plates per cover:

$33.25 \times 1.8 = 59.85$ plates in the peak hour — one plate a minute, for sixty consecutive minutes, across four stations, correctly, hot, and together.

That is the number the rehearsal exists to test. Run a simulated rush: someone writes sixty tickets reflecting a realistic menu mix and fires them at the kitchen on a real Friday-night arrival curve. Time every ticket from fire to pass. You are not looking for the average; you are looking for the ticket at minute forty-one that took nineteen minutes, and for which station it was waiting on.

The limit of this exercise: a simulated rush has no modifications, no allergy tables, no two-top that orders forty minutes after being seated, and no server who forgets to fire a table. Real service is harder than the simulation by a margin you cannot model. Which is the argument for running the simulation at 110% of the number, not at 100%.

👨‍🍳 On the Line

The simulated rush, and what it always finds.

Every simulated rush I have ever run has found the same three things, in the same order.

One station is the bottleneck, and it is not the one you expected. Usually it is the station with the most à-la-minute work rather than the most items — the one with three sauté pans going and a component that cannot be held. On a hearth-driven menu it is frequently the station next to the hearth, because the hearth cook is now also managing the fire.

The pass is understaffed. Everyone plans the cooking and nobody plans the expediting. At sixty plates an hour somebody has to be reading tickets, calling, wiping rims, checking temperatures, and holding the ticket until the table is complete, and that somebody cannot also be cooking. If your plan has the chef expediting and working a station, the simulation will tell you within twenty minutes.

Two dishes take twice as long as anybody thought. They are always the ones that were designed last, that were never cooked in sequence with anything else, and that the chef is most attached to. You have three choices: re-engineer the dish, pre-portion a component, or accept the time and let the expediter course around it. Pick one now. In week four of service, with a full room, you will pick "86 it," which is the most expensive of the three.

What to do with the findings. Write them down the same night, while the crew is still standing there, and change one thing at a time. A rehearsal that produces a list of eleven changes and implements all of them at once has taught you nothing, because the next run will not tell you which change worked.

⚠️ Where the Money Leaks

The rehearsal budget is the first thing cut and the most expensive thing to cut.

Bellwether's build carries \$3,200 of rehearsal and R&D food plus \$1,400 of staff tasting and beverage education — \$4,600 that produces no revenue and appears, to a partner staring at a shrinking balance, like the obvious place to save.

Here is what cutting it buys you. An item whose yield was never verified runs two points of food cost high all year. On \$1,116,000 of food sales, one point is \$11,160. A menu of fourteen items in which three were never cooked to a timer produces ticket times that stretch on Friday, which produces a table that turns in a hundred and ten minutes instead of ninety, which Chapter 24 will show you is measurable revenue. And a service staff that has never tasted the menu sells the cheapest thing on it, because that is the thing they are confident describing.

The disciplined move: protect the rehearsal budget by cutting somewhere the guest cannot feel. Not the training food. Not the training hours. Cut the opening marketing before you cut the rehearsal — a restaurant that executes will be found; a restaurant that is found and does not execute has spent money to accelerate its own bad word of mouth.

9.5 Friends and family, the soft open, and how to actually use them

Two terms that get used interchangeably and should not be.

A friends-and-family service is an invitation-only, non-revenue service for people connected to the restaurant and its staff, run to test whether the building, the kitchen, and the floor can produce a meal at all. Expectations are low, forgiveness is high, and the covers are deliberately few.

A soft open is a limited-capacity, limited-publicity period of real service — sometimes invitation-only, sometimes reservations-only, sometimes with a reduced menu — run before the announced opening, for the purpose of finding and fixing operational failures at a volume the restaurant can survive.

The difference that matters: friends and family tests whether you can do it; the soft open tests whether you can do it repeatedly, at pace, in front of people who will talk.

Both cost money and neither produces revenue. Bellwether's plan carries \$3,200 across four services and 200 covers — about \$16 a cover of food and beverage cost, which is above the plan's steady-state COGS per cover because a rehearsing kitchen over-produces and wastes. That is the correct budget for what these services are: an experiment, not a party.

Design the four services as four different experiments

FIGURE 9.5 — The four-service soft-open plan                          [the Bellwether plan]

  SVC  WHEN   COVERS  WHO                       MENU        WHAT IS BEING TESTED
  ──────────────────────────────────────────────────────────────────────────────────────
   1   wk −1   32     Staff families and the    6 of 14     Can the line fire a ticket at
       Mon            partners' immediate       items       all? Station-to-station timing.
                      circle                                Does the POS ring correctly?
   2   wk −1   48     Friends and family;       10 of 14    Coursing. The pass. Sequence of
       Tue            invited neighbors                     service. First real bar volume.
   3   wk −1   60     Invited guests, industry, full        Full menu under load. Ticket
       Wed            neighborhood businesses               times measured on every ticket.
                                                            Dish pit under real volume.
   4   wk −1   60     Reservations only, taken  full        The whole system at roughly
       Thu            on the actual platform                two-thirds of a plan night.
                                                            Nothing is simulated except
                                                            the payment.
  ──────────────────────────────────────────────────────────────────────────────────────
  TOTAL          200 covers · food and beverage cost $3,200 · zero revenue recognized

  Each service is followed the same night by a thirty-minute debrief with the whole crew,
  standing, in the dining room, before anybody goes home.

Notice the shape. Covers rise, menu widens, forgiveness falls, and the last service is deliberately almost real — a reservation taken on the actual platform, seated by a host using the actual table management system, rung on the actual point of sale, with a check presented and then comped. If your last soft-open service does not use every system in the building, you have left a system untested and it will fail on a Friday.

Sixty covers against a plan night of 95 is deliberate. You are not trying to prove you can do 95. You are trying to find failures at a volume where failure is survivable.

The rule that makes it worth the money

A soft open that everybody enjoys and nobody learns from is a party you paid for.

The mechanism of that failure is social. You invited people who love you. They will tell you it was wonderful. They mean it. And you will walk out of four services with a warm feeling and no data, having spent \$3,200 and about 400 labor hours to be congratulated.

The fix is instrumentation. Decide before each service what you are measuring, assign somebody to measure it, and write the numbers down. At minimum:

And the qualitative instrument that actually works: ask a closed question. "How was everything?" produces "wonderful." "What was the slowest part of your night?" produces an answer. Give every soft-open guest one specific question and collect the answers. Ten specific answers beat a hundred compliments.

🧾 Read the Numbers

```text FIGURE 9.6 — "The go / no-go sheet, ten days out" [constructed teaching example] THE ARTIFACT An opening-readiness punch list from a 70-seat full-service opening, reviewed jointly by the two owners at T−10 days. Borrowing Chapter 6's construction term for a different list: this one is operational. THE CONTEXT Certificate of occupancy issued on schedule. Friends and family in four days. Announced opening in ten. Everything on the list is either GO, FIX (has an owner and a date), or STOP (opening moves).

  ITEM                                       STATUS   OWNER    BY
  ─────────────────────────────────────────────────────────────────────
  Certificate of occupancy                   GO       —        issued
  Health pre-opening inspection              FIX      GM       T−7
  Liquor license in hand                     GO       —        issued
  Certified food manager on every shift      FIX      chef     T−6
  Food-handler cards, all staff              FIX      GM       T−5
  I-9s complete, all staff                   GO       —        —
  Insurance certificates on file             GO       —        —
  POS: menu, modifiers, tax, tip lines       FIX      GM       T−8
  Reservation platform live + tested         GO       —        —
  Merchant account settling to the bank      STOP     owner    T−9
  Cost cards for all 14 items                FIX      chef     T−6
  Plating photos hung on the line            FIX      sous     T−6
  Par levels set; first two orders placed    GO       —        —
  Walk-in and reach-in at temperature,
    thermometers in place, logs started      GO       —        —
  Hood cleaned + fire suppression tagged     GO       —        —
  Grease interceptor service scheduled       FIX      GM       T−4
  Emergency numbers, plumber, refrigeration  FIX      GM       T−3
  Cash handling, drops, safe, banking        FIX      owner    T−5
  Comp / void authorization policy written   FIX      owner    T−5
  Allergen protocol trained and posted       FIX      chef     T−6
  Music licensing in place                   FIX      owner    T−4
  ─────────────────────────────────────────────────────────────────────
  GO 8 · FIX 12 · STOP 1

WHAT IT SHOWS A normal opening ten days out. Twelve open items is not alarming; all twelve have an owner and a date inside the window. The single STOP is the merchant account: until card settlements are actually reaching the bank account, this restaurant cannot take money, and that is not a soft failure — it is the difference between opening and not. WHAT IT DOESN'T It says nothing about whether the food is good, whether the staff are ready, or whether the room works. It is a compliance-and-systems list, and a restaurant can pass every line on it and still be unable to run a Saturday. The soft open tests the other half. THE DECISION Hold the announced opening date. Move the merchant-account item to a named owner with a daily check-in, and run friends and family on schedule — a comped service does not need a merchant account, which is precisely why it is scheduled first. THE LESSON Separate the list that can move the date from the list that cannot. Most first-time operators treat every open item as equally urgent, which means they treat the one genuinely fatal item as ordinary. Three columns — GO, FIX, STOP — is the whole discipline. ```

🤝 Hospitality

The two hundred people you are about to make into advocates, or not.

Everything above treats the soft open as a test, which it is. It is also, quietly, the highest-value marketing you will ever do, and treating it purely as a rehearsal wastes half of it.

Two hundred people are about to be in your dining room before anyone else in the city. Handled well, they leave as insiders. They tell people they went before it opened. They bring the friend they told. That is the second-visit engine Chapter 23 prices, starting from a standing start, and it costs you nothing beyond the food you were already buying.

Three moves that convert them:

• Tell them the truth about what they are. "You are our rehearsal. Some of this will be slow. Tell us what was." People rise to being useful. They resent being flattered.
• Close the loop. Email the soft-open list the week after opening and tell them, specifically, what you changed because of them. We moved the appetizer course. We fixed the draft line. Thank you. Almost nobody does this and it is remembered for years.
• Treat the staff as guests too. Family meal on the last night of training, sat at real tables, served by the managers. The service staff's understanding of hospitality is formed by how they were treated in the week they learned it. Chapter 21 makes this argument at length; it starts here.

And a fourth, for opening week itself: the walk-in you have to turn away. A new restaurant at capacity will refuse people, and refusing them badly is the most expensive thirty seconds in the business. Take their name, give them a specific alternative and a specific date, and write it down. "We're full tonight — can I put you in on Thursday at seven?" costs nothing and converts a rejection into a reservation.

9.6 Opening day, opening week, and the honeymoon that will end

Opening day

The temptation on opening day is to open at full volume with the full menu and a press list. Resist all three.

Cap the covers for the first services. Bellwether opens reservations-only with a soft cap for the first three services, then releases. This costs you goodwill with people who cannot get in — which is a far cheaper problem than a full room served badly by a crew that has run four services in its life.

Do not invite press to opening week. A review written in week one is a review of your worst version, and it lives on the internet forever. If you want coverage, invite it at week six, when the kitchen is repeating itself. Chapter 27 owns the marketing calendar; this is the one date in it that is an operations decision.

Decide the day's job. Opening day is not for maximizing revenue. Its job is to get every system through a real service intact and to send the crew home believing they can do it again tomorrow. Cover count is not the score.

Somebody must be free. Both partners on the floor with jobs is a mistake. One of them holds a station or the pass; the other holds nothing — walking, watching, touching tables, catching the things nobody assigned. In week one the most valuable person in the building is the one with no assignment.

The honeymoon

The honeymoon period is the opening weeks during which a new restaurant experiences demand and guest tolerance that are both above its sustainable level — driven by novelty, concentrated attention, and a guest population disproportionately composed of friends, neighbors, and industry.

It is real. It is not a myth operators tell themselves, and this book is not going to pretend that opening weeks are hard to fill. They are usually easy to fill. That is precisely the problem.

The honeymoon does two things at once, and both flatter you:

It inflates revenue. People come once because it is new. A meaningful share of your opening-month covers are one-time novelty visits that will not repeat on any schedule you can forecast.

It biases the sample. The guests in the room in week two are systematically more forgiving than the guests who will be in the room in month six. They know somebody. They want you to succeed. They do not send the fish back. So your comp rate looks good, your reviews look good, and your service failures go unreported — which means the instrument you would normally use to find problems is temporarily broken.

FIGURE 9.7 — Weekly dinner covers, the plan's opening ramp        [the Bellwether plan]

  covers
  per wk
   520 ┤   ████ 511
   500 ┤   ████        ████ 490
   480 ┤ ··████········████··················· plan: 475 dinner covers a week (95 × 5)
   460 ┤   ████        ████        ████ 454
   440 ┤   ████        ████        ████
   420 ┤   ████        ████        ████
       └────────────────────────────────────
         weeks 1–4    weeks 5–8   weeks 9–13
         +7.6% v plan  +3.2%       −4.4%

  The plan projects the room's demand ABOVE forecast for eight weeks and BELOW it by
  week thirteen. The quarter still beats plan. The exit rate does not.

The shape of that curve is not pessimism; it is the pattern Chapter 1 §1.1 described from the other end. Revenue normalizes downward from the opening spike while costs, which were supposed to normalize downward too, do not — because the systems that would have brought them down were never built. Your entire defense against that is to know the curve is coming and to have built the systems during the weeks when the curve was flattering you.

Where the honeymoon lives, and how to watch it die. It lives on the weekend. Friday and Saturday hold up longest because they are the nights people plan around, and they are therefore the least informative. Tuesday tells the truth.

(The plan's dinner week is 95 × 5 = 475 covers. Weeks 1–4 project 108% of that; weeks 9–13 project 96%. All figures are the plan's projection, not results.)

Look at the two columns and notice that a manager watching only Friday and Saturday would conclude that nothing had changed. Nothing did change — on Friday and Saturday. The business changed by twenty-nine percent on Tuesday, and Tuesday is the night that decides whether your fixed labor floor is affordable.

⚠️ Where the Money Leaks

The distribution out of the honeymoon quarter.

This is the single most damaging financial decision made by new restaurant owners, and it is made by people who did everything else right.

Here is the mechanism. The first quarter produces cash — genuinely, in the bank. It produces cash for two reasons that are both temporary: the honeymoon is inflating revenue, and the rent abatement is suppressing the largest fixed cost in the building. Two partners who have not been paid in six months look at a healthy balance in month three and take a distribution, or stop watching, or both.

Month four arrives and three things happen at once. Rent restarts at \$7,933 a month. The novelty traffic is gone. And the first wave of opening-crew turnover lands, because the people who took a job to be part of an opening have now been part of an opening.

The discipline: treat the abated quarter's cash as what it is — a loan from months four through twelve, made by your landlord, that must be repaid in the form of a reserve you did not spend. Compute the quarter as if rent had been charged and manage to that number. If the quarter does not look good with \$23,800 of rent added back, it did not look good.

Chapter 33 builds the cash discipline properly. Chapter 34 builds the controls. This chapter's job is only to tell you that the danger arrives dressed as good news.

9.7 The first ninety days: what to measure, what to fix, what to leave alone

The first ninety days is the period from opening through roughly week thirteen, during which a restaurant's operating habits, cost structure, and guest expectations are established — and after which changing any of them costs several times what changing them now would.

It is a genuinely distinct period, for three reasons. The numbers are not yet meaningful, because there is no baseline to compare them to. The habits are not yet set, which makes them cheap to change. And the honeymoon is running, which makes the numbers look better than the business is.

Everything in this section is about resolving that contradiction: how do you manage tightly using numbers you cannot yet trust?

What to measure

The answer is: measure everything, but decide on very little. Instrument the business fully from day one, because you are building the baseline you will need in month six, and refuse to act on anything until you have a repeated reading.

The weekly prime-cost count is non-negotiable and it is the one every opening skips, because everybody is exhausted and the numbers are ugly. Count anyway. Ugly and known beats unknown, and the operator who starts counting in month four has already established that counting is optional.

🧮 Run the Numbers

The plan's own first ninety days, priced honestly.

A business plan that projects 60% prime cost in week one is not credible, and a reader who has financed restaurants will stop reading. Here is what Bellwether's plan should actually show for its first quarter — a projection, not a promise, built from the ramp in Figure 9.7.

```text FIGURE 9.8 — The first ninety days, the plan's ramp projection [the Bellwether plan]

PERIOD WEEKLY WEEKS REVENUE PRIME % PRIME $ WHAT IS HAPPENING ──────────────────────────────────────────────────────────────────────────────────── Weeks 1–4 $29,890 4 $119,560 71% $ 84,888 Full room, green kitchen, heavy waste, over-staffed floor Weeks 5–8 $28,300 4 $113,200 66% 74,712 Stations settling, comps falling, cuts being made on time Weeks 9–13 $26,068 5 $130,340 63% 82,114 Novelty gone; the real baseline appears ──────────────────────────────────────────────────────────────────────────────────── QUARTER 13 $363,100 66.6% $241,714 At the plan's 60% target 217,860 ──────────────────────────────────────────────────────────────────────────────────── EXCESS PRIME COST, FIRST NINETY DAYS $ 23,854 ```

Read the last line and then read Chapter 6 again. The rent abatement is \$23,800. The projected cost of learning to run this restaurant is \$23,854. The landlord's three free months pay for your education, to within fifty-four dollars, and if you spend the abatement on a slipped schedule instead, the education gets billed to the working-capital reserve.

Now the harder arithmetic, which almost no plan shows.

The frozen Year 1 model is \$1,550,000 of revenue at 60.0% prime cost, or \$930,280. Subtract the quarter above:

• Revenue remaining: $\$1{,}550{,}000 - \$363{,}100 = \$1{,}186{,}900$ across 39 weeks.
• Prime cost remaining: $\$930{,}280 - \$241{,}714 = \$688{,}566$.
• Required prime cost for weeks 14–52: $\$688{,}566 \div \$1{,}186{,}900 = \mathbf{58.0\%}$.

The plan's 60% annual prime cost requires the last thirty-nine weeks to run at 58.0%, because the first thirteen will not. That is a demanding but achievable number for a well-run full-service restaurant, and it is a completely different management assignment from "hold 60%."

And the revenue side. Weeks 14–52 must produce \$1,186,900. Of that, the **\$139,240 bridge — patio, events, takeout — was always going to arrive after the first quarter, which leaves $\$1{,}186{,}900 - \$139{,}240 = \$1{,}047{,}660$ from the dining room across 39 weeks, or \$26,863 a week**. The week-13 exit rate in the projection is \$26,068. So the dining room needs to climb 3.0% from its post-honeymoon baseline and hold it — which is a real assignment, and a considerably more modest one than the raw annual number implies.

This is the most useful thing this chapter does for the plan. It decomposes an annual target into a ramp, a bridge, and a residual, and it says which of the three is doing the work.

What to fix, in order

Four tiers, and the order is not negotiable.

Tier one — fix immediately, no discussion. Anything that is a food-safety, alcohol-service, or wage-and-hour failure. A cooler that will not hold temperature. A sanitizer bucket nobody tests. An employee working off the clock. A bartender serving without checking identification. These are not performance issues to be coached; they are existential and they get fixed the day they are seen.

Tier two — fix this week. Anything that breaks the guest experience repeatably: a specific dish that consistently takes nineteen minutes, a bar that cannot keep up at 7:30, a section that is always the last one greeted, a table that everybody hates. Repeatable is the operative word. A bad Tuesday is noise. The same failure three times is a system.

Tier three — fix by week eight. Anything costing more than a point of prime cost that you can identify and measure. On Bellwether's projected quarter, one point of prime cost is \$3,631. Portion drift, an uncosted special, a par level that is generating waste, a shift that is consistently over-staffed by one person.

Tier four — everything else. Write it down and wait. You will have a list of forty things by week three. Thirty of them are not worth the cost of changing them while the crew is still learning the first version.

What to leave alone

This is harder than the fixing, and it is where the discipline actually shows.

Leave the menu alone for six weeks. You cannot read menu mix from a honeymoon sample. The guests in your room in week two are ordering out of curiosity, ordering what their friend on the staff told them to order, and over-ordering because it is an occasion. Chapter 12 builds the menu-engineering matrix; it needs a settled sample, and week two is not one.

Leave prices alone for ninety days. A price change in month one tells every early guest that you did not know what you were doing, and it destroys the only comparable you have.

Leave the staffing structure alone until you have four weeks of real covers. Chapter 19 builds a staffing guide from sales-per-labor-hour data. You have no data. Schedule generously for three weeks, measure, and then cut to the number rather than to the feeling.

Leave the concept alone. Entirely. Chapters 2 and 3 built it against a trade area and a competitive set. Ninety days of operation is not evidence that a concept is wrong; it is evidence about execution. The operators who panic and pivot in month three almost always pivot into something they have not tested at all. Chapter 39 covers the genuine turnaround diagnostic, and its first question is whether the problem is concept, execution, or math — a question you cannot answer with a honeymoon sample.

One change a week. If you change five things and the following Tuesday improves, you have learned nothing about which change worked, and you have taught the staff that the standards are provisional. Change one thing, tell everyone you are changing it, measure it, and keep or reverse it.

🔍 Check Your Understanding

1. Bellwether's projected first quarter runs 66.6% prime cost against a 60% annual target. What prime cost must weeks 14–52 achieve for the year to land on plan, and why is that a different management assignment from "hold 60%"?
2. In the projection, Tuesday dinner covers fall 29% from weeks 1–4 to weeks 9–13 while Saturday falls 2%. Why is the Tuesday number the more useful one, and what decision does it inform?
3. An owner in week three wants to remove two dishes, reprice three, and cut a server from the Thursday schedule. Which of those, if any, should happen, and what is the rule?

(1: 58.0%, because $\$688{,}566 \div \$1{,}186{,}900 = 0.580$. It is different because it requires a specific, sustained improvement over a known ramp rather than a flat average, and it makes the weekly prime-cost count the mechanism by which the year is actually won. 2: Weekend demand is the least elastic and therefore the least informative; Tuesday reveals the underlying repeat demand and determines whether the fixed labor floor is affordable on a slow night — it informs whether Tuesday should be staffed differently, priced differently, or eventually reconsidered as a service. 3: None of them yet, except possibly the Thursday cut if there are three weeks of consistent labor data supporting it. The menu needs a settled sample (six weeks), prices need ninety days, and the rule is one change at a time, measured.)

🍽️ The Business Plan

Checkpoint 9 of 40 — the Pre-Opening Budget & Timeline, and the soft-open plan. Opening day has a date.

Part II ends here. Chapters 6, 7, and 8 gave the plan a space, a capacity, and a legal path. This chapter gives it a date — and, less comfortably, an honest number for what happens between the last inspection and the first check.

What the plan gains

Attach Figure 9.1 (the bottom-up budget), Figure 9.2 (the countdown), Figure 9.3 (the last ninety days), Figure 9.5 (the soft-open plan), and Figure 9.8 (the ramp projection). Five pages, and a reader can see the whole opening.

What this section settles

The date is real and it is reasoned. April is not a preference; it is the month that puts the patio, the summer, and the events pipeline in front of the reserve rather than behind it, and puts February at the far end of year one.

The thirty-day rule turns Chapter 6's best clause into an operating instruction. Rent commences on the earlier of opening or thirty days after the certificate of occupancy. Therefore the certificate is the starting gun for a thirty-day program, and everything that does not require a legally occupiable building — menu costing, vendor accounts, pars, the POS build, recruiting, classroom training, the service-standards document — is finished before it is issued.

The ramp is decomposed. The plan no longer claims 60% prime cost from week one. It claims 66.6% for a quarter, 58.0% for the remaining thirty-nine weeks, and a dining room that must climb 3.0% from its post-honeymoon baseline while the \$139,240 bridge arrives on schedule. Those are three separate, testable claims, which is what an assumptions register is for (Chapter 4).

What this section does not settle

The \$36,300. This is the chapter's central finding and it must be stated in the plan, not buried. The honest cost of opening this restaurant is \$71,300 against a \$35,000 line. There are four ways to close it and none of them is free:

The first three are false economies with arithmetic attached. The fourth is what actually happens, so price it honestly. Bellwether's fixed monthly obligations — the plan's fixed labor of \$21,000, rent of \$7,933, fixed other operating of \$10,333, G&A of \$3,875, and debt service of \$5,792 — total \$48,933 a month**. A \$45,000 reserve is 27.6 days of that. A \$8,700 reserve is 5.3 days**.

Five days is not a reserve. It is a rounding error with a name, and Chapter 1 §1.4 has already told you what happens next. Chapter 33 must size this properly, and the plan should carry the question openly rather than let a reader discover it.

Chapter 6's two open questions, answered as far as this chapter can. Question 4 asked where the missing construction contingency comes from without touching the reserve. The answer from here is blunt: not from pre-opening. The pre-opening line is itself \$36,300 short; it has nothing to give, and any transfer between the two is the exact failure Chapter 6 described one invoice at a time. Question 5 asked what the deposits total and which line absorbs them. The first month's rent is \$0, because Chapter 6 tied commencement to the certificate of occupancy and abated three months — a direct, countable return on one negotiated sentence. The security deposit and utility deposits are not: a two-month security deposit, a common ask, would be \$15,867 at Bellwether's \$7,933 monthly rent, plus roughly \$1,500 of utility deposits. Neither is an expense and neither fits inside the \$35,000. Chapter 33 must fund them out of working capital and say so.

The insurance premium. The schedule Chapter 8 builds — general liability, liquor liability, property, workers' compensation, business interruption — produces a first payment before opening that is not in Figure 9.1. Add it when Chapter 8 prices it.

Overtime. Figure 9.4 assumes nobody crosses forty hours in a week. In the final fortnight several people will, and roughly \$2,300 of overtime premium is unbudgeted.

Open questions carried forward

1. Where does the \$36,300 actually come from, and what is a working-capital reserve of \$8,700 worth in weeks? (Chapters 32, 33)
2. Does the \$139,240 revenue bridge land in weeks 14–52 as this ramp requires? (Chapters 24, 28, 29)
3. Can weeks 14–52 run 58.0% prime cost with a crew that is four months old? (Chapters 19, 31)
4. What does the liquor license actually cost and when does it actually arrive — and what is the plan if it is late? (Chapter 8)
5. What is the opening crew's ninety-day attrition, and what does replacing four people in month four cost? (Chapters 17, 21)
6. What do the security deposit and utility deposits total, and which source funds them? (Chapter 33)
7. Two partners will have gone six months without income by opening day. What is the household plan? (risk section; Chapter 33)

Conclusion

Most restaurants open before they are ready, and the reason is arithmetic rather than impatience. From the moment the first manager joins the payroll, a restaurant burns money at a rate that only one event can stop, and that event is opening. Every week you wait to be ready costs more than the readiness is worth — which is why the discipline in this chapter is almost entirely about sequencing rather than quality. You do not get to be more ready. You get to arrange the work so that less of it happens while the meter is running.

Bellwether's plan carries \$35,000 for pre-opening. Built from the bottom — 1,504 training hours, a walk-in and a back bar filled from empty, four rehearsal services, the permits, the utilities, the uniforms, and the glassware that gets broken learning where the shelf is — the honest number is \$71,300**. The gap is **\$36,300, and there is no clever way to close it. Cutting training costs more than it saves inside eight weeks. Cutting inventory defers the cost into a worse week. Cutting the soft open puts an untested kitchen in front of paying guests. What actually happens is that the working-capital reserve absorbs it, falling from \$45,000 to \$8,700 — about five days of this restaurant's fixed obligations — and that is precisely the mechanism Chapter 1 called undercapitalization, arriving on schedule, dressed as a budget line nobody could photograph.

Two things save it, and both are already in the plan. The first is the sentence Chapter 6 negotiated: rent commences on the earlier of opening or thirty days after the certificate of occupancy, with three months abated. That is \$23,800 — almost exactly the \$23,854 of excess prime cost the first ninety days will cost — and it is only worth that if you open inside the thirty days. Every week past it burns \$1,831 of the best asset you have. The second is the decision about when to put people on payroll. A four-week slip after the crew is hired costs \$32,968 and produces nothing, which means the single highest-leverage control in an opening is refusing to commit the payroll date until the certificate is issued and the licenses are in hand.

Then the doors open and the numbers get better than they should be. The room is full, the rent is abated, the guests know somebody, and the first quarter throws off cash for two reasons that both end at about the same time. Measure everything from week one, decide almost nothing until week six, fix safety immediately and repeatable failures within a week, and leave the menu, the prices, the structure, and the concept alone. Watch Tuesday. Count prime cost every Monday even when the number is ugly, because the habit is the point and the habit is not available retroactively.

Part III begins the work that makes the 58.0% possible. Chapter 10 builds the menu — the document that is simultaneously a production schedule, a labor model, a purchasing spec, a brand statement, and the most important marketing tool the restaurant owns. Everything the kitchen rehearsed in this chapter, it rehearsed against a menu somebody had to design first, and that design decides more about prime cost than any decision made after opening day.

Key Terms

Pre-opening budget — the money a restaurant spends between the end of construction and the first paying guest: payroll before revenue, training and rehearsal food, licensing and certification, the opening inventory, and the operating costs that begin before the register does. The most reliably underestimated line in a restaurant project, because it is the only one with no physical object attached to it. (Ch. 9)

Pre-opening timeline — the dated, dependency-ordered sequence of everything that must happen before a restaurant can serve its first guest, built backward from a committed opening day rather than forward from a lease signature. Its purpose is to expose which items have uncompressible lead times and which are gated by an authority. (Ch. 9)

Opening inventory — the full stock of food, beverage, and supplies that must be present on the first day of service: walk-in, freezer, dry store, back bar, wine, beer, and non-alcoholic product. Frequently the largest single pre-opening line, and usually purchased on cash terms because the business has no trade-credit history. (Ch. 9)

Menu rehearsal — the systematic production of every item on the opening menu, to the standardized recipe, under service conditions, before any guest is served, in order to verify the recipe, the yield, the plating standard, the station assignment, and the fire time. Distinct from a tasting, which produces a pleasant evening and no artifacts. (Ch. 9)

Plate-up — a service-staff tasting at which every menu item is plated exactly as it will go to a table, presented by the person who cooks it, tasted by every member of the service staff, and documented with ingredients, method, allergens, permissible modifications, and fire time. (Ch. 9)

Friends-and-family service — an invitation-only, non-revenue service for people connected to the restaurant and its staff, run at low volume to test whether the building, the kitchen, and the floor can produce a meal at all. Subject to the same permits, certifications, and food-safety obligations as a paid service. (Ch. 9)

Soft open — a limited-capacity, limited-publicity period of real service before the announced opening, run to find and fix operational failures at a volume the restaurant can survive. A soft open that produces compliments rather than measurements is a party the operator paid for. (Ch. 9)

Honeymoon period — the opening weeks during which a new restaurant experiences both demand and guest tolerance above its sustainable level, driven by novelty, concentrated attention, and a guest population disproportionately made of friends, neighbors, and industry. It inflates revenue and biases the feedback sample at the same time, and it ends. (Ch. 9)

The first ninety days — the period from opening through roughly week thirteen, during which a restaurant's operating habits, cost structure, and guest expectations are established, and after which changing any of them costs several times what changing them now would. The period in which you measure everything and decide very little. (Ch. 9)

Spaced Review

1. Chapter 6 tied rent commencement to "the earlier of opening or thirty days after the certificate of occupancy," with three months abated. State in one sentence what that clause is worth in dollars, and in a second sentence what operating rule this chapter derived from it.
2. Chapter 8 maps the licensing path, including the liquor license. Explain why this chapter insists that opening day must be set after that license is in hand, and compute the two costs an operator is choosing between if it is not.
3. From Chapter 1: undercapitalization was named as the most common cause of first-year failure, and defined as opening with less money than the business needs to reach sustainable operation. Using Bellwether's figures, show the specific arithmetic by which a \$35,000 pre-opening allocation manufactures that condition, and state what the working-capital reserve is worth afterward in days of fixed cost.
4. From Chapter 4: the plan's assumptions register makes beliefs visible and testable. This chapter decomposed the frozen 60% annual prime cost into a 66.6% first quarter and a 58.0% remainder. Write the two new register entries that decomposition creates, and say how each one would be tested.
5. The recurring question: an operator decides to hire the full opening crew two weeks earlier than planned so that everyone is thoroughly trained. Does this decision move prime cost, and in which direction? When does it hit the bank account as opposed to the P&L — and what would you need to measure, ninety days later, to know whether it was worth it?