The Restaurant Unit Economics Pro Forma to Run Before You Sign the LOI

A nine-unit group in Scottsdale sent us a one-page restaurant unit economics pro forma for a tenth location three days before their LOI expired. It showed $2.9 million in year-one sales, a 19% four-wall margin, and a build-out number copied from the last store. What it didn't show: ten weeks of management payroll before opening day, the deposits and opening inventory, or that unit six sat 2.4 miles away and would give up volume to it. The site was fine. The math was decoration.

By then they had already picked the tile. That's the problem with most new-unit models — they get built after the emotional decision, to justify it. A pro forma is only useful if it can still say no. Here's the version that runs before you sign: what the capital number really includes, how to build revenue from the bottom up, the screens that kill a deal outright, and the deals worth walking away from.

What a restaurant unit economics pro forma actually is

It's a cash model for one location: how much cash goes in, when it comes back, and what has to be true for that to happen. A P&L projection says the unit makes money in year two. It doesn't say you'll be $340,000 underwater in month five, and that's the number that ends groups. The failure mode is rarely "the restaurant lost money." It's "the restaurant was fine and the group ran out of cash funding it," which is why a group-level 13-week cash flow forecast belongs beside this model.

Two rules. Build the first 26 weeks in weeks, months after that. And source every line: a quote, a signed comp, a portfolio benchmark, or a flagged assumption. The flagged cells are the deal.

The total capital number nobody builds correctly

Total capital is build-out plus everything else, and "everything else" routinely runs 40% to 70% of the construction number. Groups budget the box and forget the business that lives in it. The line that disappears most often is working capital: the cash the unit burns between opening day and the week it funds itself, usually week 10 to 20.

Every range moves with format and market. It's a checklist, not a budget.

One tax note: how build-out spend is classified drives your depreciation schedule and first-year cash taxes. Check the IRS rules on depreciating property and run cost segregation before the invoices land.

Build revenue from the bottom up, not from the last store

Bottom-up means seats times turns times average check times daypart times operating days, with off-premise as its own channel and margin. "The last one did $3.2 million" is not a revenue build. It's a memory of a different trade area and labor market.

It looks like this: 96 seats, 1.3 lunch turns at a $19 check, 2.0 dinner turns at a $41 check, six lunches and seven dinners a week, plus delivery and pickup at a set percentage of dine-in with fee drag priced in.

The honeymoon, and what it hides

Weeks one through eight run hot. Opening volume commonly lands 15% to 40% above the stabilized run rate, then decays through month six and settles between months nine and fifteen. Build the annual plan off those eight weeks and you get a plan that fails in week twelve. Set the base case on the stabilized number and treat the honeymoon surplus as cash timing: it funds the working capital hole, it doesn't validate the site.

The cannibalization line nobody writes down

If the new unit sits in an existing unit's trade area, model the transfer. Overlapping sites within two to three miles commonly move 5% to 15% of the existing volume. Book it against the existing unit and judge the new one on incremental contribution to the group. A store doing $2.6 million by taking $400,000 from your best performer is a $2.2 million store carrying full capital cost.

Occupancy cost ratio: the single go/no-go screen

Occupancy cost ratio (rent plus CAM, taxes, insurance, and percentage rent, divided by net sales) is the fastest way to kill a bad deal. Full-service concepts want 6% to 10%. High-volume fast casual can survive into the low teens. Above that, everything else has to be perfect, and it won't be.

Run it against your downside case. That reframes the negotiation: maximum rent equals target ratio times downside-case sales, and that's the number you take to the broker.

Know what a rent win is worth. Saving $4 per foot on 3,200 feet is $12,800 a year. Missing revenue by 10% on a $3 million unit costs $300,000 of sales and roughly $90,000 of contribution at a 30% incremental margin. A great rent deal can't rescue a bad revenue assumption.

Why the tenant improvement allowance moves the answer more than rent

The TI allowance is the highest-value item in the negotiation, because it moves cash out of your stack into the landlord's. Shifting $250,000 of build-out to TI barely touches four-wall EBITDA. It can move cash-on-cash return ten points or more, because it shrinks the denominator.

Get three things in writing: how TI is paid (construction draws versus reimbursement at CO, which means you fund the build first), whether it's amortized into rent and at what rate, and what qualifies. Negotiate free rent separately. It's a cash line, not a courtesy.

The three cases, and the breakeven question underneath them

Model three cases every time: a base case on stabilized volume, a downside where year-one revenue lands 20% below plan, and a ramp case showing the honeymoon curve and the trough behind it. If the downside can't service debt and cover fixed costs without a capital call, you don't have a deal. You have a bet.

Underneath all three sits the question your operator needs: what weekly sales covers fixed costs? Divide weekly fixed costs (occupancy, salaried management, insurance, base R&M, technology, minimum labor) by contribution margin: one minus your variable percentage of food, labor, and operating expense. For a typical full-service box that lands between $24,000 and $34,000 a week before debt service. Add debt service and that's the number for the GM's door.

Then the decision metrics. Cash-on-cash return is annual unit cash flow after maintenance capex, divided by total cash invested net of TI. Payback period is months until cumulative cash flow crosses zero, measured from the first dollar of capital, not opening day. For independent groups, a box below 20% cash-on-cash or past five years to payback rarely justifies the risk. EBITDA is what a buyer eventually pays a multiple on (how EV/EBITDA multiples work), but you can't make payroll with a multiple. The go/no-go is cash.

The 7-step order of operations, from site identification to signed lease

Sequence matters as much as math. Plenty of groups run steps 1, 6, and 7, then build the model.

Screen the trade area before you tour. Daytime and residential population, drive-time, daypart traffic, competitive density, parking, co-tenancy. Kill most candidates here, on paper, at zero cost.

Build the bottom-up revenue model in all three cases before you discuss price. Anchoring is real and it works on you.

Set the occupancy ceiling. Maximum rent equals target ratio times downside-case sales. That's what you negotiate toward.

Cost the box with a contractor walk and a test fit. Bring your GC and an MEP engineer in pre-LOI. Confirm what the space can't do: power, gas, grease, hood exhaust, ADA, restrooms, patio rights.

Convert the capital stack into cash-on-cash and payback across all three cases. Include working capital and contingency or it's theater.

Negotiate the LOI on terms that move cash, not base rent alone: TI and draw mechanics, free rent, delivery condition and date, rent commencement tied to CO, exclusive use, kick-out, assignment, personal guarantee.

Finish lease review and financing commitment before signature. Counsel reads the lease against the LOI; the lender's term sheet is executed. The signed lease is last, not first.

When the answer is no, and what a personal guarantee actually costs

Walk away when the downside case doesn't cover debt service. That's the whole test, and it's the one groups negotiate with themselves about. If revenue at 20% below plan requires a capital call, the site is asking your other units to underwrite it.

Three more that earn a hard no:

The site needs your best-ever unit's volume. Your best unit is an outlier. Outliers aren't a plan.

You don't have the GM. No bench, no build. A promoted-too-early manager costs more than a delayed opening.

Financing that only works in the ramp case. If the amortization assumes the honeymoon holds, you're absorbing a risk the lender priced.

Then the personal guarantee, which almost never gets priced. A full-term guarantee on a ten-year lease at $18,000 a month is roughly $2.16 million of exposure, and it typically survives the closing of the restaurant. Ask for a cap of twelve to eighteen months of rent, or a burn-off after 36 months of on-time payments. On the debt side it usually isn't optional: SBA's 7(a) loan program generally calls for guarantees from owners at or above a 20% stake, and your counsel should confirm how the lease and loan guarantees interact.

The National Restaurant Association's 2026 State of the Restaurant Industry report projects $1.55 trillion in sales on 1.3% real growth, with 42% of operators reporting they were not profitable. Expansion works there for groups with a disciplined model and a funded balance sheet, and against everyone else.

Where we land

The pro forma is the one artifact in expansion built to tell you the truth before you're committed, and it earns that only if it runs pre-LOI and carries the lines everyone skips: pre-opening payroll, working capital through day 90, cannibalization, contingency. Screen on occupancy cost ratio against the downside case. Decide on cash-on-cash return and payback, not EBITDA. Negotiate TI harder than base rent. And be willing to walk, because the discipline to pass is what makes the good sites affordable.

Common questions

What should a new restaurant location cost to open?

Total capital commonly runs $600,000 to $2.5 million for a second-generation space, more for ground-up builds. Think in ratios: budget 40% to 70% on top of hard costs for FF&E, technology, permits, pre-opening payroll, inventory, deposits, marketing, working capital, and contingency.

What is a good cash-on-cash return for a new restaurant unit?

For independent multi-unit groups, a base case producing 25% to 35% cash-on-cash with payback inside three to four years is a strong deal. Below 20%, or past five years, the return rarely compensates for the balance sheet risk and personal guarantee.

What occupancy cost ratio is too high for a restaurant?

Full-service concepts want total occupancy (rent, CAM, taxes, insurance, percentage rent) at 6% to 10% of net sales; high-volume fast casual can work into the low teens. Test it against your downside case, since occupancy is fixed and sales are not.

How much working capital do I need for a restaurant opening?

Plan for eight to fourteen weeks of operating burn beyond expected receipts, modeled weekly. The honeymoon masks the need for about two months, then volume decays toward the stabilized run rate while payroll and inventory stay at opening levels. That gap is behind a great many unplanned capital calls.

Should I model cannibalization when opening near an existing location?

Yes, and book it against the existing unit rather than the new one. Overlapping trade areas within two to three miles commonly transfer 5% to 15% of the existing location's volume. Judge the new site on incremental contribution to the group.