Two deals stall in the same quarter. The pipeline review gives them the same sentence—“good business, hard to get done”—and moves on. But one stalled because its numbers never arrived in a form anyone could underwrite, and the other because nobody on the client’s side had the authority to sign what had already been agreed. Same epitaph; different causes of death. And because the debrief never asks where a deal died, both will be back next year under new names.
“Stuck” is not one condition; there are five, and they are fixed differently. Bankability itself—the short list of claims that survive being read by strangers—is defined in What Bankability Really Means. This piece is the working diagnostic: when a deal is stuck, where.
The five dimensions
1. Business quality
The substance everything else depends on. Is the model coherent? Is the revenue defensible? Does the margin survive the loss of one customer, one contract, one person? Business quality is not size or profitability. It is durability and legibility—whether the business makes sense to someone who does not already know it.
The tell: remove one name from the story—a customer, a supplier, the founder—and the model needs rebuilding. A fine business can be one relationship with operations attached; lending to it is a different proposition.
2. Information quality
A sound business can still be opaque on paper. Accounts organised for the tax office rather than the underwriter; management numbers that live in the accountant’s spreadsheet; forecasts that are ambitions with columns. Information quality is not perfection—it is sufficiency and coherence, enough reliable material to form a view. The gap between what an owner knows and what an institution can hold is nobody’s fault, and its cross-border shape is the third meeting problem.
The tell: every answer is correct—eventually. Each request routes through the owner, because the records cannot speak for themselves.
3. Structure quality
Whether the deal as drawn fits the business as run: the right facility type, tenor, security, covenants, the right answer to what happens if the plan is wrong—and the right fit between how the money moves and how the trade cycles. Structure rarely fails on purpose. It fails by default, when the conversation reaches for the most familiar structure rather than the most appropriate one—five-year money for a ninety-day need, a smooth-cycle facility lent to a lumpy business. The commonest version has its own field note.
The tell: the term sheet reads well until you ask what it does in the worst month of a normal year.
4. Execution quality
Structure is a plan; execution is what happens to it. Deals are moved by named people doing dated things—producing documents, clearing conditions, returning mark-ups—and an opportunity can be well understood, well structured, and still expire because nobody on either side can make it move. Decision-makers are unclear. Advisers are misaligned. The client wants the outcome but has not staffed the process. The bank is not exempt from this dimension either: a credit policy built entirely to select well, and never asked to move, fails here rather than at the credit decision—the argument of Policy as a Filter, Policy as Movement.
The tell: the deal only advances during meetings. Between meetings, nothing happens that was not chased.
5. Relationship quality
The most cited dimension and the least defined. Relationship quality is not warmth, frequency, or years of acquaintance. It is whether each side believes the other will act in good faith when the position becomes uncomfortable—and whether that belief has ever been tested. In cross-border and Asian-linked work it carries particular weight: trust is accumulated through behaviour, never assumed, and it is fragile in ways transactional relationships are not. The characteristic failure is timing—relationship energy arriving before structural readiness, a connection strong enough to open doors but not yet to bear the weight of an institutional process.
The tell: nobody can name the last time the relationship carried bad news. Warmth with no load history is not yet trust.
Where the cause hides
The five are not equally visible. Everyone can see the ends of the list: the business, because there are numbers for it, and the relationship, because there are feelings about it. When a deal strains, the bank’s file records a business problem—the numbers have weakened—while the client’s story is about the relationship—they stopped believing us. Each side reaches for its native language.
The causes concentrate in the middle. Information, structure, and execution belong to the process, and process is nobody’s first language—so a format gap gets read as bad faith, and a covenant calendar gets read as deterioration. The misdiagnosis is not carelessness but vocabulary: people place the cause in the dimension they have words for.
That is most of what the framework is for. Naming the middle gives a stalled deal an address that is neither the business is weaker than you say nor the bank has stopped believing you—those are fighting words—but the information runs six weeks behind the story, or the structure was drawn for a different cash cycle. Those are work.
Using it
The dimensions are rarely all strong at once, and the point is not to average them. A deal strong in four and missing one is not eighty per cent bankable; it is stuck at a specific place, and the fix depends on which. Weak information wants an accountant and six weeks. Weak structure wants a different drafting table. Weak execution wants a named owner and a dated list. The common error is treating them interchangeably—more meetings for an information problem, more documents for a relationship problem—which spends months and goodwill on the wrong repair. None of it is permanent: unbankable today is often bankable in six months, if the named dimension moves.
A caution from use: the framework misleads when it becomes a scorecard. The temptation is to rate each dimension out of five and add them up, which produces a number and destroys the insight—a nine that is missing information quality is a different animal from a nine that is missing execution, and neither is helped by being called a nine. It also misleads early. Relationship quality, in particular, is unknowable in the first meeting and easily faked in the second; scoring it there just records charm. The framework earns its keep from the third conversation onwards, once there is behaviour to read.
Two companion pieces sit alongside this one. Dimensions three through five—structure, execution, relationship—compress into the three pressure-point tests of Relationship, Structure, Execution, which is the version I reach for mid-process. And the most common failure shape—a strong business running weak on dimensions two through four—is examined in Why strong businesses can still be difficult to finance.