The highest-failure-rate category in small business lending earns the most skeptical studies — and the good projects survive them.
Restaurant lending’s reputation precedes it, which is precisely why the independent study matters: committees need the projects with real economics separated from the ones running on the founder’s palate. The separation happens in the trade-area math — daypart-level demand, traffic and visibility for the drive-thru formats, the employment base that funds weekday lunch — and in expense honesty: food cost, labor, and occupancy benchmarked line by line against industry composites, with any sub-benchmark assumption explained or removed.
Format determines the analysis. A QSR drive-thru is a throughput machine — stack depth, service times, and the traffic count decide revenue capacity before the menu matters — while full-service concepts live on covers, check average, and the turn count the floor plan permits. Franchise systems import brand pull, training, and documented unit economics; the study verifies the local set’s performance rather than accepting the system average.
Recent practice includes dual-tenant retail with a Baskin-Robbins component — multi-unit approval language and equipment budgets carried in the evaluated cost basis — and QSR analysis inside larger fuel and travel-center engagements, where the food operation’s contribution is modeled as its own profit center with its own benchmarks rather than blended into inside sales.
Restaurant demand uses spending-potential capture analysis, frequently disciplined by a gravity-model check for sites competing across nodes. Spending potential is defined as the trade area’s total food-away-from-home expenditure: category demand = households × average FAFH spend per household × segment share, built from consumer-expenditure data and segmented to the concept’s price tier and daypart structure. Daytime population — the employment base within the lunch radius — enters as its own demand pool for daypart-dependent formats.
The subject’s claim is then required capture = pro forma revenue ÷ category demand, the single statistic that converts the entire revenue projection into a market-share assertion the committee can judge: a concept needing 1.8 percent of its trade area’s segment spending is making a modest claim; one needing 9 percent is explaining itself. Capacity supplies the physical ceiling and the cross-check: maximum revenue = seats × turns per daypart × check average × operating days — and the pro forma must fit under it with room to spare. For drive-thru formats, throughput substitutes for seats: cars per hour at documented service times. The rationale: pairing a demand-share statistic with a capacity ceiling traps the forecast between two independent constraints, which is the strongest position a projection-based credit can occupy.
Restaurant files flag faster than any category because the tells are standardized: food cost three points under the segment composite with no sourcing story, labor scheduled below the format’s documented hours, occupancy cost above the viability threshold waved through on revenue hope, and check-average-times-turns arithmetic that exceeds the floor plan’s physics. Daypart blindness recurs — lunch-dependent locations underwritten on dinner concepts — as does franchise misreading: system-average unit volumes claimed for a trade area the franchisor’s own model would tier lower. And the first-time-operator file with no manager named, no ramp modeled, and the owner’s salary absent from the expense build is its own complete flag set.
Mitigation in food service is benchmark-driven and blunt. We rebuild the pro forma to the segment’s composite ratios and the floor plan’s real capacity, then identify with the sponsor what bridges the gap the honest model opens: a revised menu engineering toward margin, a smaller buildout, the drive-thru lane that changes the throughput math, or franchise affiliation where systems genuinely substitute for missing experience. Where the operator gap is the finding, the determination conditions on the named general manager or the executed franchise agreement rather than pretending biography is evidence. The dual-tenant Grain Valley engagement modeled the food component on its own economics inside the larger credit — the same discipline that keeps a marginal restaurant from sinking an otherwise sound real-estate deal.
Let the trade area choose the concept: daypart demand, traffic patterns, and the employment base are findings that should precede the menu, not follow it. Benchmark yourself before the underwriter does — if your food cost or labor line beats the composite, write down why and prove it. Respect the occupancy threshold; no concept outruns rent that consumes its margin. Name the manager, fund the ramp, and pay yourself in the pro forma. And for franchise candidates: the system’s value is its data and its training — use both, and claim the tier your trade area actually supports.
The category’s economics have been structurally reset: labor costs stepped up and stayed, delivery platforms take their commission from a margin that never had room, and the formats winning are the ones engineered for those facts — drive-thru-forward QSR, small-footprint concepts with counter service, menus built for off-premise. Underwriters have followed: throughput infrastructure now reads as collateral quality, and pro formas that ignore third-party-delivery economics on concepts that will depend on them are incomplete. The trade-area data revolution cuts both ways — mobile-traffic analytics make demand claims more provable and lazy claims more exposed.
Engagements are typically initiated by the borrower, with lender or CDC confirmation obtained before work begins — institutions apply differing rules, so sponsors should confirm the required path with their lending contact — and are delivered in 10 to 15 business days from complete project data, and built to the program framework that governs the credit — SBA SOP 50 10 8 coverage minimums of 1.15x operating and 1.00x global, the 37-factor structure of USDA RD Instruction 5001, or the 1.20x convention of conventional credit policy — with a ten-year pro forma, sensitivity at ±5/10/15 percent, rate stress to +3.0 percent, and Monte Carlo analysis as standard equipment.
President, Wert-Berater, Inc. — independent feasibility study consultants since 1998. More than 4,000 feasibility studies completed across all 50 states and internationally, evaluating $40.2 billion in project value for SBA, USDA, EB-5, conventional, and institutional financing decisions. Fiduciary duty runs to the lender and agency in every engagement.
+1 310-857-2443 ext. 800 · email · 1968 South Coast Hwy, Ste 2382, Laguna Beach, CA 92651 · 111 Town Square Pl Ste 1238 PMB 657834, Jersey City, NJ 07310 · 539 W. Commerce St #8486, Dallas, TX 75208
Schedule a ConversationIndependent feasibility studies since 1998 — 4,000+ engagements, $40.2 billion in evaluated project value. Standard delivery in 10 to 15 business days. Fiduciary duty to the lender and agency.