The top line was the best on the block. The bottom line never showed up.

The subject was an independent full-service restaurant being acquired and lightly repositioned, on a strong retail corridor, by an operator whose pro forma projected revenue at the top of the local set — on the order of $1.8 million in year-two sales. The all-in financing ran to roughly $1.4 million across leasehold improvements, equipment, franchise-free goodwill, and working capital.

The deal was structured for SBA 7(a) financing — the dominant vehicle for an independent restaurant acquisition with a leasehold and a working-capital need. Because the projections carried the credit, an independent feasibility study was expected to give the lender a defensible basis for the forecast, and the going-concern appraisal had to allocate the leasehold improvements, equipment, and goodwill.

The sponsor's pro forma led with revenue: strong covers, a healthy check average, the best top line on the block. The analytical question was whether the top line described the deal, or whether the binding constraint sat several lines below it.

The intuitive way to underwrite a restaurant is sales: establish the covers and the check average, build a confident revenue line, and read strength off the top of the P&L. On that logic the concept looked like the best credit in its set — the highest projected sales on the corridor, a busy dining room, a top line that cleared the requested debt service with room to spare. Revenue was the headline, and the headline was strong.

It was also underwriting the line of the P&L that says the least about whether a restaurant survives — because in full-service, the top line is consumed by two costs that decide everything, and a strong revenue number reveals nothing about either.

A full-service restaurant lives or dies on prime cost and occupancy, not on revenue, and a high sales line can sit on top of an unviable cost structure. Prime cost — the cost of goods sold plus labor — benchmarks at roughly fifty-five to sixty-five percent of sales in full-service; above sixty-five percent, it is extremely difficult to make a profit regardless of how high the volume goes. Occupancy cost — rent and its associated charges — benchmarks at six to ten percent of sales, and ten percent is the point at which it seriously impairs profitability. What is left after those two is thin: full-service net margins run on the order of three to five percent. The industry's own operating data makes the mechanism concrete: profitable full-service operators run labor near a thirty-four percent median, while loss-making operators run it near forty-three percent — an eight-to-nine-point swing in a single cost line that separates a profitable restaurant from an unprofitable one at the same revenue.

Re-underwritten on prime cost, the subject's headline strength thinned. At the wage rates and food costs prevailing in that market, the concept's menu and labor model ran prime cost closer to sixty-eight to seventy percent than the high-fifties the pro forma implied, and the lease carried occupancy near ten percent. On an eighteen-hundred-thousand-dollar top line, the difference between a sixty-percent and a seventy-percent prime cost is roughly a hundred and eighty thousand dollars — more than the entire projected net profit. The strong revenue line was real; it simply sat on a four-wall margin that, as modeled, did not clear debt service. The bankable version of the deal was not the one with the highest sales — it was the one with a prime cost and a rent the revenue could actually carry. The relevant analysis was four-wall economics, not the top line.

Prime-cost build at market input costs. Cost of goods sold and labor were modeled at the food prices and wage rates prevailing in the subject's market and daypart mix — not at the pro forma's assumed percentages — so prime cost reflected what the concept would actually run rather than an industry ideal.

Occupancy-cost ratio. Rent and associated charges were measured as a percentage of projected sales and tested against the six-to-ten-percent threshold, isolating how much of the top line the lease consumed before any other cost.

Four-wall EBITDA, not revenue. Coverage was built on four-wall economics — revenue net of prime cost and occupancy — rather than on the sales line, so the analysis underwrote the margin that actually services debt.

Prime-cost sensitivity. The model stress-tested prime cost at fifty-five, sixty, sixty-five, and seventy percent and occupancy at six, eight, and ten percent, mapping the combinations under which the concept cleared coverage and the combinations under which it did not.

Delivery-channel margin. Third-party delivery was modeled at its effective cost — commissions of roughly fifteen to thirty percent headline and thirty to forty percent once packaging, promotions, and refunds are layered in — so delivery volume was treated as a margin question rather than a demand signal, and not allowed to flatter coverage.

Failure-rate context corrected. The analysis set the deal against the documented restaurant survival data — first-year survival near eighty-three percent, ahead of the small-business average, against the discredited folklore of near-certain early failure — so the lender's risk frame rested on evidence rather than myth, with location and four-wall economics identified as the real determinants.

The credit file replaced a revenue headline with a four-wall margin and explained why the strongest sales line in the set was not the strongest credit. The analysis showed prime cost and occupancy as the binding constraints, mapped the cost structure under which the concept cleared coverage, and identified the right-sizing — in menu and labor model, or in rent — that the deal required to be bankable. The SBA reviewer treated four-wall economics, not the top line, as the basis for the projections, and the going-concern appraisal's allocation across leasehold improvements, equipment, and goodwill clarified the collateral. The independent study answered the program's expectation by evaluating the margin that pays the loan, which is where restaurant credits are most often misjudged.

The 7(a) financing closed underwritten on four-wall economics rather than revenue, with the cost structure right-sized so that prime cost and occupancy left a margin the projected sales could actually carry. The inflection was not that the concept was weak. It was that the number the pro forma led with — the top line — described the restaurant's revenue, not its viability, and the bankable deal was the one underwritten on the small number several lines below it.

• 01Revenue is the most misleading line on a restaurant P&L. Prime cost and occupancy decide viability, and a high sales line can sit on top of a cost structure that does not clear debt service.
• 02Prime cost is the binding constraint. Above sixty-five percent of sales, profit is extremely difficult regardless of volume — and an eight-to-nine-point labor swing separates profitable operators from unprofitable ones at the same revenue.
• 03Occupancy cost compounds it. Rent above ten percent of sales seriously impairs profitability, so the lease is part of the underwriting, not a fixed given.
• 04Underwrite four-wall economics. Coverage depends on revenue net of prime cost and occupancy, not on the top line — and delivery volume is a margin question, not a demand signal.