What Kills a Bank Deal Besides DSCR

Why the ratio is only the first gate

A bank funds most of a small acquisition. Depending on the structure and the lender, the senior debt commonly carries the majority of the purchase price, with the balance made up of buyer equity and some form of seller participation. That means the lender is the largest single capital provider in the transaction and is taking a position in a business it has never operated, run by a buyer who has usually never owned one.

Here is the part that is not obvious from outside the process. The bank has no acceptable exit. Calling a loan is expensive, slow, and bad for the lender's reputation in a market where reputation drives deal flow. Recovery is not a business line that anyone at the bank wants to be in, and a commercial banker who has to enforce is having a worse week than the borrower. A commercial lender I speak with regularly puts it plainly: the personal guarantee is about commitment, not about recovery. It exists so that the borrower has something at stake, not because the bank is planning a route to your house.

Follow that through and the logic of underwriting gets much clearer. If there is no good exit, the loan has to be right at the front end, because the back end is not a real remedy. Pricing cannot save a bad file. So the underwriter's job is to find, before funding, whatever could cause the cash flow to stop, and the single largest thing that could cause it to stop is the owner leaving.

That reframes the whole exercise. Bankability is transferability with a number attached to it. A bank is only willing to finance what it believes can survive the change of hands.

The objections that actually stop deals

1. Owner dependence

The most common and the most fatal. If the business is the owner, then the earnings history describes a person, not an asset, and the lender is being asked to finance a transfer of something that does not transfer.

What the file gets read for: whether there is a second in command who runs anything without the owner present, whether customer relationships sit with the owner personally or with the company, whether pricing and quoting are documented or held in one head, whether the owner is the licensed or certified person the business legally requires, and whether a transition period is contractually committed rather than promised verbally.

What fixes it: time. Delegation, a documented management layer, a real transition agreement, and where the licence sits on the owner personally, a qualified employee brought in before the sale. This is the objection you cannot patch during diligence.

2. Customer concentration

If one customer is a large share of revenue, that customer effectively holds a veto over the loan. Concentration matters more than most sellers expect because the risk it creates is correlated with the transfer itself. A large account is often a relationship account, and relationship accounts are exactly the ones that get reviewed when the owner they know retires.

What the file gets read for: revenue share of the top customer and the top few, contract term or absence of one, tenure, whether the relationship is institutional or personal, and whether there is any change-of-control provision.

What fixes it: partly time and partly disclosure. Diversification is a multi-year project. In the near term, a documented contract with real term, evidence that the relationship runs through the company rather than through the owner, and a candid explanation offered up front will do more than hoping it does not come up. It will come up.

There is also a second, less obvious cost to concentration, which is covered in the next section: the largest accounts are the ones most likely to sit on non-assignable contract terms, so concentration and collateral exclusion tend to strike the same customer.

And the risk does not rise in proportion to revenue. It rises faster. A contractual right is only as real as the counterparty's capacity to exercise it, and that capacity scales with size. A large institutional customer has a legal department, a procurement office, a vendor management process and a contract administration function. It will know that the business changed hands, because the change is disclosed, or required to be disclosed, or picked up in a vendor review. It has somebody whose job includes noticing. A small customer has nobody watching. Service continues, invoices arrive, payments are made, and a clause that technically permits termination is never read by anyone, because there is no one on the other side whose role is to read it.

So the same sentence in two contracts is a live risk in one and a dead letter in the other. Ten thousand small accounts with imperfect assignment language do not produce the exposure that five large ones do, and not only because they are diversified. Most of them were never on customer paper to begin with, and among those that were, the overwhelming majority of customers will neither notice the transfer nor care about it, and will simply carry on.

3. Contracts, licences and consents that do not transfer

This one kills deals late, which makes it expensive. Franchise agreements need franchisor consent. Many commercial leases have change-of-control or assignment clauses. Government and institutional contracts are frequently non-assignable or require a re-qualification. Certain licences attach to a named individual, not to the corporation. Key supplier agreements and dealer or distribution rights often require approval, and occasionally the approval is discretionary.

Most sellers who know about the clause think of it as a consent errand. Get the counterparty to sign off, move on. It is worse than that, and the reason is worth understanding because it changes what you do about it.

A change-of-control provision does not just put the revenue at risk. It can remove the contract from the collateral entirely.

In several industries, recurring service agreements are a recognized collateral class in their own right. Monitoring contracts, scheduled inspection agreements, maintenance and route-based service books: lenders in those sectors will advance against the contract book, because the revenue is contracted, dated and predictable in a way that ordinary receivables are not. This is routine practice rather than an exotic structure, and in some businesses it is the main reason a deal is financeable at all.

Which is exactly what makes the exclusion so sharp when it happens. A lender that secures against a contract book as a matter of course will still refuse an individual contract inside it if that contract can be terminated on a change of control. The logic is self-referential and there is no arguing with it. A security interest attaches at closing. Closing is the event that triggers the counterparty's right to walk. So the bank would be taking security in an asset the customer is entitled to cancel at the precise moment the security comes into existence. There is no version of that position that is worth anything, so the lender does not take it.

And the exclusion is not random. It works its way down from the top of the customer list.

Small and mid-size customers sign the vendor's standard form, and vendor standard forms are assignable, because the vendor's lawyer wrote them. Large customers do not sign your paper. They impose their own, and their terms very commonly restrict assignment or allow termination on a change of control, because those terms were drafted to protect the buyer from precisely this event.

Which customers have their own paper is predictable. Government agencies and national laboratories, universities and school boards, hospitals and health networks, municipalities, large property managers and institutional landlords, national chains and public companies. Any counterparty with a procurement function has standard terms, and imposing them is what the procurement function is for. These organizations do not negotiate assignability with a mid-sized service vendor. They hand over the master agreement and the vendor signs it, gratefully, because the contract is worth having.

The result is that the exclusion does not strike one unlucky agreement. It can take out the entire top of the book at once, because the customers large enough to occupy the top five are the same customers large enough to have imposed their own terms. A seller can lose the majority of contracted revenue from the collateral calculation in a single pass and be left borrowing against the small accounts at the bottom of the schedule.

Now notice which way the quality signals point. Institutional counterparties are what a seller is proudest of. They are creditworthy, they pay on time, they renew, they are multi-year, they carry names that make the customer list look formidable, and they took years to win. Every one of those attributes is a reason the business is genuinely better. And every one of them travels with the procurement terms that remove the contract from the security position. The sentence a broker writes as a selling point, a blue chip institutional customer base, is the same fact a lender reads as a collateral problem.

This is the second cost of concentration flagged earlier, now visible in full. The concentrated account is disproportionately the non-assignable one, so the revenue risk and the collateral exclusion do not spread across the book. They land together, on the same contract, the one the business can least afford to lose.

The practical consequence is that a contract book has two different values and they are not the same number. There is contracted revenue, which is what goes in the memorandum, and there is assignable contracted revenue, which is what a lender will actually underwrite. Most sellers have never calculated the second one and are surprised by how far apart they are.

How this surfaces, and why it surfaces late.

Contract eligibility is generally not assessed by the banker across the table. It is tested by an outside firm the lender engages to examine the collateral, agreement by agreement, against the eligibility criteria written into the credit. That review lands well into the process, after the memorandum has been written and read, after the buyer is committed, and frequently after a price has been agreed. Each contract either qualifies or it does not, and the seller learns the answer one contract at a time.

It is worth being precise about what the lender is saying at that point, because sellers routinely hear something worse than what is meant. A contract ruled ineligible is not a contract the lender thinks is weak. Institutional accounts are usually the strongest thing about the business and the underwriter knows it perfectly well. The lender is answering a narrower question than the buyer is. Not whether this revenue is valuable, but whether it can be written into the credit. A contract can be the entire reason someone wants to buy the company and still be ineligible as security, and those two judgments do not contradict each other.

That distinction is small comfort in the moment. Watching the best agreements in the book come back declined, one after another, is a particular kind of surprise, and it is the wrong way round from every instinct a seller has about which parts of their business are the strong parts. The instinct is not wrong. The test is just measuring something else.

The principle underneath all of it: a lender can advance against a distribution, but not against a coin flip.

This is the single idea that explains why a book of small contracts is treated so differently from a handful of large ones, and it is worth holding onto because it generalizes well beyond contracts.

A large book of small agreements has a churn rate. Some fraction does not renew every year, the fraction is reasonably stable, and it can be measured from history. That is a distribution, and a distribution can be underwritten. The lender applies a haircut that reflects the observed attrition, advances against the remainder, and is right on average even though it is wrong about any individual account. Losing customers is not a surprise in that structure. It is a line in the model.

A single contract that represents a large share of revenue and can be terminated on a change of control is not a distribution. It is one event with two outcomes, and there is no haircut that makes a binary safe. Advance sixty percent against it and the lender is not sixty percent protected, it is fully exposed forty percent of the time. So the lender does the only rational thing available and assigns it zero.

Read that back into the earlier sections and the pattern is consistent. Owner dependence is a binary. A single non-assignable anchor contract is a binary. A concentrated customer is close to one. Diversification is usually described as reducing risk, which undersells it. What diversification actually does is convert risk from a form that cannot be underwritten into a form that can. That is why it moves a financing outcome so much further than the arithmetic alone suggests.

Two separate hits fall out of one clause. The revenue cannot be relied on in the cash flow test, because it is not contractually certain to survive the transfer. And the contract cannot support the security position, because a right the counterparty can extinguish is not an asset a lender can hold. On a business whose value largely is its contract book, that combination can take the underwritten value close to nothing while the business itself is unchanged and performing well.

What the file gets read for: whether the revenue the loan is underwritten on legally survives the closing, and whether anything the lender is being asked to secure against can be cancelled by a third party.

What fixes it: pull your top ten agreements before the business is listed and check two things on each. Whose form is it on, and what does it say about assignment and change of control. That is an afternoon of work and it produces the number a lender is going to calculate anyway.

Then start the consent conversations early, because with institutional counterparties they are slow. A consent, a waiver, or a written acknowledgment obtained in advance converts the objection from a structural one into a formality, and many agreements that restrict assignment also provide that consent will not be unreasonably withheld, which is a much better starting position than it sounds. But a university, a government body or a national account moves at its own speed, and a request that takes two months to work through a procurement office is fatal in month five of a transaction and merely administrative in month one of preparation.

4. Earnings nobody can substantiate

The purchase price comes off adjusted EBITDA, and the adjustment is where small deals go wrong. The gap between what the statements report and what the memorandum claims is made up of add-backs, and each add-back is a claim that a cost will not recur for the buyer.

Some are unarguable: a genuine one-time legal matter, an owner's above-market salary being reset to a market wage. Some are arguable: a vehicle, travel, a family member on payroll who does actual work. And some are not add-backs at all, they are the business.

What the file gets read for: whether each adjustment is supported by a document, whether the total adjustment is a modest reconciliation or a rebuild of the earnings, and what the tier of the underlying financial information is. A notice to reader, a review engagement and an audit are three very different levels of assurance, and they are not interchangeable in an underwriting file. Stating one and delivering another is a repricing event waiting in diligence, and rightly so.

What fixes it: build the normalization as a document trail rather than a spreadsheet total. Every add-back should resolve to the page it came from. If it cannot, it should not be in the number.

5. Collateral shortfall on a goodwill-heavy purchase

Most small business purchases are mostly goodwill. The tangible assets are a fraction of the price, and the lender advances against tangible assets at a discount that varies by asset class, with receivables treated more generously than inventory and inventory more generously than equipment. Real property is its own conversation.

The consequence is that on a service business with almost no hard assets, the security position is thin regardless of how good the cash flow looks. And as the contract point above shows, an agreement that looks like the single best asset on the seller's side of the table can be worth zero on the lender's side, which is how a business ends up simultaneously valuable and unfinanceable. That does not necessarily stop the deal, but it changes the structure: more buyer equity, more seller participation, a shorter amortization, tighter covenants, or a government-backed program where one is available.

What the file gets read for: what the lender is actually secured on if the cash flow argument turns out to be wrong.

What fixes it: nothing on the seller's side, but knowing it in advance changes how the deal is structured and priced, and it explains why an all-cash-at-close expectation on a service business is frequently unrealistic.

6. Working capital left short at close

A deal that funds and then starves. If the purchase agreement does not specify a working capital target, or the buyer's model assumes the receivables come with the business when they do not, the business opens on day one unable to make payroll without drawing the operating line it does not have yet.

What the file gets read for: whether the buyer has an operating facility sized to the actual cash conversion cycle, and whether the transaction structure leaves any of it behind.

What fixes it: a defined working capital target in the letter of intent, not in the final agreement, and an operating line arranged alongside the term loan rather than after it.

7. The buyer's own file

The business can be perfectly financeable and the loan still declines because of who is buying it. Relevant management experience in or adjacent to the industry, the source of the equity injection (borrowed equity is not equity), personal credit history, and the buyer's own financial position all get underwritten.

The failure mode worth naming is a good business matched to a buyer who cannot carry the specific risks that business holds. A concentrated industrial supplier is a different proposition in the hands of someone who has run a similar operation than in the hands of a first-time buyer from an unrelated field, and the lender will price or decline on that difference.

8. Deterioration in the stub period

Underwriting runs on the trailing twelve, but the interim statements get read. A business whose current-year numbers are sliding while it is on the market invites the question of whether the trailing twelve is a history or a peak. A seller who cannot explain the trend cleanly is negotiating against their own financials.

Why a decline is the most useful free information a seller ever gets

Sellers treat a financing decline as an obstacle. It is better understood as an assessment, and it is the only rigorous one available at no cost.

The bank has looked at the business with real money on the line, with no interest in the outcome other than being repaid, and reached a conclusion about whether the earnings survive the transfer. Nobody else in the process is doing that. The broker is being paid on the transaction. The buyer is arguing their own case. The accountant is working from what they are given. The lender is the only participant whose incentive is aligned purely with whether this thing keeps working after everyone else has been paid.

So the useful response to a decline is not to shop the file to three more lenders. It is to ask what the objection was and treat it as the finding it is.

What this means if you are getting ready to sell

The order matters. Every one of the objections above is cheaper to fix before listing than during diligence, and several of them cannot be fixed during diligence at all.

A business that goes to market unfinanceable does not get a lower price. It gets a smaller buyer pool, because the only buyers left are the ones who do not need a bank, and those buyers know exactly why they are the only ones there. The price follows the pool. This is the part sellers consistently get backwards: financing is not the buyer's problem to solve after agreeing on a number. It is the constraint that determines which numbers were ever real.

The practical version is short. Run the bank's test on yourself first. Ask what happens to the earnings if you leave, and answer it honestly with reference to who does what. Read your material contracts for change-of-control language. Know your concentration and be ready to speak to it. Make sure every add-back resolves to a document. Know what your financial statements actually are and never describe them as something else.

The honest limits

Every lender is different, and the same file can clear one credit desk and fail another. Program-backed lending, asset-based facilities and conventional cash flow term loans have materially different tests, and the security regimes in Canada and the United States are not the same. Nothing here is a credit decision or a substitute for a conversation with a commercial lender who will actually look at your file.

What generalizes is the reasoning, not the parameters. Lenders finance what they believe transfers. Everything above is a specific instance of that one question.

Where we sit

JCoBee builds the small-deal version of this test. The system reads the financial statements, rebuilds the earnings, checks each add-back against the document it came from, and runs the bank math including the coverage and leverage tests, so that a seller or a buyer can see the objections before a lender does. Every figure stays connected to the page it came from, and where a number has nothing behind it, it prints as incomplete rather than as a confident guess.

You can walk a sample deal without signing up for anything at jcobee.com/samples.