Case Study 1: What the Package Actually Requires

The SBA 7(a) program as a disclosure regime — and what a restaurant plan has to survive to get through it

Background

Most people describe the Small Business Administration's 7(a) program incorrectly, and the error matters for how you write your plan.

The SBA does not lend you money. In the 7(a) program the SBA provides a partial guarantee to a participating lender — a bank, a credit union, or a licensed non-bank lender — which reduces the lender's loss if you default. You apply to the lender. The lender underwrites you. The lender decides. The SBA's role is to set eligibility rules, publish the standard operating procedures the lender must follow, and stand behind a portion of the balance.

That structure has two consequences that shape everything in this chapter.

First, you are being underwritten twice. Your file has to satisfy the bank's own credit standards and the program's eligibility and documentation rules. A plan that a loan officer likes personally can still fail because the file is not documentable.

Second, the guarantee exists precisely because businesses like yours cannot be collateralized. A restaurant's assets are a leasehold improvement in somebody else's building, a used hearth, and a walk-in bolted to a wall. The program exists to make loans possible where the collateral does not support them — which means the underwriting weight shifts onto the two things that remain: the projections and the people. This is not a bureaucratic quirk. It is why the credibility of your forecast is, functionally, your collateral.

The SBA also runs a 504 program, which finances real estate and long-lived equipment through a Certified Development Company on different terms. Chapter 5 covers the choice between them. This case is about the document.

The operating issue: what a package contains

An SBA 7(a) application for a startup restaurant is not a business plan. It is a file, and the plan is one exhibit in it. What lenders generally require, in some form, includes:

  • The business plan itself, with a market analysis and a description of the concept and operations.
  • Financial projections, conventionally three years, with year one detailed monthly — and, this is the part operators underestimate, the assumptions behind them. A projection handed over without its assumptions is an unsupported assertion, and a lender who has to reconstruct your reasoning will reconstruct it uncharitably.
  • A use of proceeds schedule — what the money buys, totaling the request.
  • Personal financial statements for each owner above a threshold stake, listing assets, liabilities, and contingent liabilities.
  • Personal background and history information, and résumés establishing relevant management experience.
  • Business and personal tax returns, and for an existing business, historical financials.
  • Evidence of the equity injection — where your own money is, and that it is actually yours.
  • The lease or letter of intent, and often a landlord agreement regarding access to collateral.
  • Personal guarantees. Program rules generally require a guarantee from owners at or above a specified ownership stake — commonly cited as 20% — which means the failure scenario for a restaurant startup is not "the business closes." It is "the business closes and the owners personally owe the balance."

Two important honesty notes. Program rules change, sometimes substantially, and they differ in detail between lenders operating under delegated authority and those submitting for standard processing. And the equity injection expectation for a startup is conventionally described in the neighborhood of ten percent or more of total project cost, but the figure is lender-dependent and policy-dependent. Verify both with a participating lender before you build a capital stack around a number you read in a book.

What it shows

Three things, and each one has a direct consequence for how you write.

1. The file is a coherence test

Every document in that list restates the same business from a different angle. Your projections say the restaurant does $1,550,000. Your lease says you are paying $95,200 for 2,800 square feet. Your use of proceeds says construction costs $310,000. Your personal financial statement says you have $150,000 to inject.

An experienced underwriter reads those four documents against each other. Does the revenue number make sense for the square footage? Does the injection actually exist in an account, or is it a promise? Does the build-out figure look like something a contractor would sign? The most common way a package dies is not a bad number. It is two documents that disagree, because a contradiction tells the reader that nobody assembled the file with an understanding of what it says.

This is exactly what the assumptions register in §4.4 prevents. A register forces you to state each number once, in one place, with its basis — which makes internal contradiction visible to you before it becomes visible to a stranger.

2. Management experience is scored, and restaurant experience is not the same as ownership experience

Underwriting weighs relevant management experience heavily, and for good reason: in a business with no collateral and thin margins, the operator is the risk model.

Here is the trap for a chef. Fourteen years of kitchen experience is real and it counts — but it is evidence of operational competence, not of financial management. A file that presents a chef-owner and a front-of-house partner, neither of whom has carried profit-and-loss responsibility, has a visible gap. Bellwether's executive summary in this chapter states that gap in its own third paragraph, and then states the countermeasures: an outside accountant engaged pre-opening, a weekly prime-cost review from week one, and a reporting calendar in the plan.

That construction — name the gap, then close it with a system — is worth more than any attempt to make the résumés look like something they aren't. A reader who finds an experience gap you disclosed weighs it. A reader who finds one you obscured re-reads everything.

3. The projections are the collateral

Since the physical assets do not secure the loan, the forecast has to carry the weight, and a forecast carries weight only if a reader can trace it.

This is why the bottom-up build in §4.3 is not a stylistic preference. A revenue line that reads "$1,550,000 based on comparable restaurants in the market" cannot be examined. A revenue line that reads "68 seats at 1.4 turns and a $46 average check across five dinner services, plus 110 brunch covers at $24 across two, plus a named bridge of $139,240 from the patio, private events, and takeout" can be examined line by line — and every line an underwriter can examine and accept is a line they no longer have to discount.

Outcome

The observable pattern across the industry is that restaurant applications are declined for a recognizable and mostly avoidable set of reasons. Assembled from what lenders and SBA-focused advisors describe consistently — treat this as attributed industry practice rather than as a published statistic — the recurring causes are:

  • Insufficient or unverifiable equity injection. The money is not there, or it is borrowed, or it is a family member's promise rather than a balance.
  • Projections without assumptions. Numbers that cannot be traced to anything physical.
  • Optimistic revenue with no comparable support, especially revenue that grows to fit the debt service rather than the other way around.
  • No relevant management experience, or experience that is culinary only, with no financial countermeasure described.
  • A use-of-proceeds schedule that doesn't tie to the request, the lease, or the contractor estimates.
  • Personal credit problems disclosed late, or not at all.
  • A lease that has not been negotiated, so the largest fixed cost in the model is a placeholder.

Notice how many of those are document failures rather than business failures. A viable restaurant with a badly assembled file gets declined. That is a genuinely unjust feature of the system, and the only useful response is to assemble the file properly.

Lesson

A business plan submitted for financing is a disclosure document, and it should be written like one.

The instinct is to treat the plan as marketing — to lead with the concept, the food, the room, and the founder's story, and to keep the difficulties in a section near the back that nobody reads. The structure of an SBA package punishes that instinct at every turn. The file demands your personal financial statement, your background, your lease, your injection, and your assumptions, and it is read by someone whose professional function is to imagine your failure.

So write toward the reader you actually have. State the ask early. Show the arithmetic. Attach the assumptions to the projections rather than making the reader excavate them. Name the experience gap and the system that answers it. Disclose the risks in the summary rather than the appendix.

None of that makes a weak business fundable. What it does is stop a fundable business from being declined for reasons that have nothing to do with whether it would have worked.

Discussion questions

  1. The case argues that "the projections are the collateral." Reconstruct that argument from the structure of the 7(a) program. What would have to be true about restaurant assets for it not to hold?

  2. A restaurant applicant has a strong concept, a great location, and a personal financial statement showing very little liquid net worth. Which parts of the package can compensate, and which cannot? What does your answer imply about who gets to open a restaurant?

  3. The file is described as a coherence test — several documents restating one business from different angles. Pick any two documents from the list and describe a specific, realistic way they could contradict each other in a restaurant application.

  4. Bellwether's executive summary discloses that neither partner has carried P&L responsibility. Argue both sides: does disclosing that make the file stronger or weaker? What would you need to believe about how underwriters read to hold each position?

  5. The personal guarantee means a failed restaurant produces personal debt. Should that change how you write the risk section of a plan — and should it change the plan itself? Where, specifically?

  6. The case lists "projections that grow to fit the debt service" as a recurring cause of decline. Describe the mechanism: how does a forecast get quietly reverse-engineered from a loan payment, and what feature of the process in §4.3 and §4.4 would catch it?