What lenders actually underwrite

The lender package for a middle-market financing runs to eighty or ninety pages. Business overview, market position, adjusted EBITDA bridge, three-year projections, a covenant case, a downside case, customer concentration, management biographies. It takes weeks to build and several rounds to get right, and it is genuinely necessary — no credit committee approves anything without it.

It is also not what decides the deal.

The package gets a company into the conversation. What settles pricing, structure, and whether the lender leans in or quietly slows down is a set of questions that mostly get asked on calls, often in passing, and almost never in a form that appears in any document. I have been on the answering end of those calls as the operating CFO, and the pattern is consistent enough to name.

Four questions

Can you tell me what happened last month?

Not what the model says will happen. What happened, in the month that closed three weeks ago, and why it differed from what you told me the month before.

This sounds like a reporting question and it is really a control question. A management team that can answer it quickly and without hedging is telling the lender something no projection can: that the business is being watched at a resolution where problems surface early. A team that needs a week to reconcile the answer is telling them the opposite, regardless of what the numbers eventually show.

The tell is not accuracy. It is latency.

How wrong has your forecast been?

Every borrower presents a forecast. The useful question is what happened to the last four.

A company that has consistently landed within a few points of its own projection has established something valuable — not that it will hit the new number, but that its process for producing numbers is honest. A company that has beaten its forecast twice and missed it twice by wide margins in both directions has a forecasting problem, and a lender will price the resulting uncertainty even if the average is fine.

Most borrowers never volunteer this history. The ones who do, and who can explain the misses without defensiveness, materially change how the rest of the conversation goes.

Where does the covenant actually bind?

The covenant case in the deck shows headroom. The question underneath it is what has to go wrong, specifically, for that headroom to disappear — and whether anyone has modeled it against the operating plan rather than against a generic haircut.

This is the question where the gap between an advisor and an operator shows most clearly. An external advisor can tell you the leverage covenant is set at 4.5x with a 25% cushion. Someone who has run the business can tell you that the cushion evaporates if two specific renewals slip a quarter, that both are in the same vertical, and what the mitigation is. The first answer is arithmetic. The second is a credit view.

Lenders are not primarily pricing the asset. They are pricing their uncertainty about the people who will be running it when something goes wrong.

What is the working capital actually doing?

Revenue growth and cash generation come apart more often than most decks acknowledge. A company can grow twenty percent and consume cash the entire way, and whether that is a healthy investment in growth or a slow-motion collections problem depends on details that live below the caption level — DSO by customer cohort, whether payment terms have quietly extended to win deals, whether the receivable that makes the quarter is from a customer who always pays late.

A lender who has been doing this for fifteen years will find that. The only question is whether they find it from you in week two or from diligence in week eight, and those two paths lead to different deals.

Why fluency is the variable

What connects all four is that none of them can be answered from a document. They are answered live, by whoever is on the call, and the quality of the answer depends almost entirely on whether that person has been inside the numbers or has been briefed on them.

This is the structural problem with how most middle-market financings are staffed. The advisor who built the package is fluent in the package. The CFO is fluent in the business. The lender's questions sit precisely in the gap — they are about the business, but they arrive in the context of the package, and they get asked at moments when only one of those two people is in the room.

The version of this that works is unglamorous: the person answering lender questions should be the person who closes the books. Not briefed by them. Them. That is not always possible, which is why the role exists at all — but it is worth being clear about what is being traded away when it is not.

What this means before you start

The preparation that matters is not the deck. It is the four answers, ready before the first call, with the uncomfortable parts already surfaced.

Which means: a close fast enough that last month is genuinely available. A forecast history you are willing to show. Covenant headroom stress-tested against the real operating plan, including the scenario nobody wants to model. And a working capital picture where you already know what diligence is going to find.

None of that is about presenting better. It is about being the borrower who has already asked themselves the questions the lender is about to ask — which, in a market where nearly every structure is covenant-lite and lenders have correspondingly less recourse if they get it wrong, is worth more than it used to be.