You're Not a Vendor, You're an Unsecured Lender

You're Not a Vendor, You're an Unsecured Lender

You've seen the engagement letter. You've seen the SOW. The milestone billing schedule looks airtight on paper — the kind of document that makes you feel like the cash side is handled.

Then you pull the AR aging and the WIP report.

Something doesn't reconcile. Work is happening. Payroll went out. Materials moved. The senior tech's hours are logged. But the invoices haven't gone out. And it's not the usual story — the client isn't slow-paying. The trigger event hasn't happened yet. The project isn't "done." Or more precisely, the customer hasn't agreed it's done.

This is the spot where the financials stop being able to help you. The numbers describe what happened. They don't describe what was decided — or wasn't — before any of it started.

Here's the reframe that changes how you read the engagement: when delivery has no defined endpoint, the business isn't selling a service. It's extending credit. Unsecured. No collateral, no guarantee, no lien, no covenants, indefinite maturity, zero interest. The customer's continued good faith is the entire underwriting basis.

That's where margin disappears. That's where AR aging deceives. That's where the most valuable people in the business get consumed. And — the part nobody talks about — that's where the cash to keep operating gets pulled from work that was never properly billed.

The tell isn't in the financials. It's in what was, or wasn't, defined before the work began.

Three Cash Failures, One Common Cause

The owner sees one problem. There are actually three, and they look different on the balance sheet.

The visible one is margin compression. Labor performed at zero marginal price. Every "while you're in there" hour is an unpriced extension of credit. The owner's read on this is almost always "we underpriced it." That's not what happened. The work was scoped right at signing. The customer extended the principal afterward — added scope, added requests, added expectations — and the owner accepted the extension without renegotiating terms. That's a different problem than mispricing, and it doesn't fix with a rate sheet update.

The deceptive one is revenue recognition delay, and it's worth pausing on because it looks like a familiar problem and isn't. Slow-pay AR — clock running, customer dragging, you know how that conversation goes. This isn't that. Here the clock hasn't started. The contract ties billing to a definition of complete that the customer controls. So the receivable can't be booked. Can't be factored. Can't be pledged against a line of credit. A third party looking at the engagement sees nothing — no asset to recognize. The owner has performed the work. The instrument that would let anyone else acknowledge it doesn't exist.

The dangerous one is capital lockup, and this is where the recursion bites. While the unsecured exposure builds on the revenue side, the cost side keeps running. Payroll. Materials. Subcontractors. The business consumes cash to perform work it can't recognize as revenue. And the cash funding that consumption? It's frequently coming from prior un-invoiced work in the same relationship. Yesterday's unpaid loan financing today's bound labor. The customer is on both ends of the loop.

A bank wouldn't write this loan. Not at any rate. No defined obligation date, no covenants, no reporting requirement, and the only collateral is work that can't be redirected or sold. Your client signed it anyway.

One more thing worth naming up front: this failure mode lives almost entirely in whale relationships. Smaller clients trying to operate this way just get released — the engagement gets renegotiated or ended. The whale is structurally too important to release. That asymmetry is what lets the whole dynamic persist for months. Sometimes years.

Andre's Eight-Month Implementation

Andre runs a SaaS company for mid-market manufacturers. ARR looks healthy. Product has SKUs and tiers. By every quick read, this is a clean software business.

Then you look at where the engineering team actually spends its time.

Every enterprise deal closes with implementation support attached. Sometimes that's papered as a fixed-fee package. Sometimes it's a verbal "we'll get you set up." Neither version is bounded, and the floor of work expands to fit. The customer's ERP is twenty years old. The training got rescheduled twice. Someone wants a custom report. The integration broke when their IT pushed an update on a Tuesday. Andre's engineers are spending forty hours a week in a typical week on customer issues that aren't on any product roadmap.

The whale customer — a quarter of the book — has been "in implementation" for eight months. The annual subscription has been recognized. The implementation fee was bundled at signing. Everything past that, Andre's company is absorbing.

And here's the part the org chart doesn't show. The implementation isn't long because the work is unusually complex. The implementation is long because the customer is a quarter of the book and Andre didn't force a close. The cost of forcing one is potentially the relationship, and the relationship is too important to risk. A 5%-of-book customer in the same situation? Escalated, renegotiated, or written off months ago. Andre tells himself this is implementation. The accurate read is that the whale figured out — explicitly or otherwise — that staying in implementation costs them nothing and keeps Andre's engineers on tap.

So you walk the failures.

Engineering payroll keeps showing up against "implementation," dragging gross margin down quarter over quarter, and Andre's framing is "we just need to get them stable." That's the margin compression. The customer extended principal, Andre accepted, no renegotiation happened, the cost shows up in the engineering line.

The professional services revenue that was supposed to bill at go-live is sitting in deferred or hasn't been booked at all, because go-live keeps moving. That's the recognition delay. Andre's CFO is forecasting against a number that has no trigger event.

Underneath both, the capital lockup. Andre's engineering team is the asset he sold. Their hours are committed against an engagement with no exit, which means they aren't building the next product release. The product roadmap — the long-term value of the company — is being delayed by the whale's indecision. And the cash paying those engineers? Some is recognized subscription revenue, sure. Some is operating cash that should be funding R&D. And some is the implementation fee, already spent, against work that hasn't ended.

Andre's contract called the work implementation. The customer reads it as standing access. Both readings are defensible because the document didn't decide.

Jeff's Build Bay

Jeff has three bays. Two for repair, one for custom builds. He runs a wrench himself, but Mark — the senior tech — anchors both sides. Mark is in the build bay most of the time and gets pulled into repair when the queue backs up.

Repair is a clean business. Estimate, approval, work, invoice, pickup. Cash conversion is fast. Margin is predictable. The shop survives on this.

The build bay is something else entirely.

A customer arrives with a vision. Jeff and Mark scope it as best they can. Deposit covers materials. Balance is "due on completion." The vehicle goes into the bay.

Nine months later, it's still there. Parts ordered, parts paid for. Mark's hours have been logged against the build week after week. The customer keeps revising — different wheels, different interior, "while it's apart can we also" — each revision small enough that pushing back feels petty. Each revision pushes "complete" out further. Jeff didn't push back because the customer is a referral source, and you don't get into shouting matches with the people who feed you new business.

Same financial mechanism as Andre. One structural twist that makes Jeff's case worse.

The build bay manufactures whales by design. One bay, one senior tech, one active build at a time. Whoever is in the bay is — by definition, by the geometry of the shop — the whale of that line of business. A hundred percent of build-bay revenue. A meaningful share of senior-tech hours. So Jeff's reluctance to push back on revisions isn't a personality thing. It's the bay's economics. Every build customer is structurally too important to lose, because while they're in the bay, there's only ever one of them.

The exposure compounds in two directions.

Every revision Jeff accepted without renegotiating was additional principal extended at zero price, against a receivable he can't invoice because the customer controls the trigger. That's the unsecured side.

The labor side is uglier. Mark isn't just one tech among many. He's the senior tech the whole shop depends on. Every hour on the build is an hour out of the repair bays. When repair backs up, Jeff pulls Mark over to clear it — which slows the build, which extends "complete," which deepens the cycle. The build customer didn't just bind Mark's hours on the custom job. They've effectively bound the throughput of the profitable repair side too.

And the cash to fund all of it has to come from somewhere. The deposit went to materials, and the materials cost more than the deposit covered. The cash funding nine months of Mark's logged hours is coming directly out of repair revenue. The profitable side of the business is financing the whale customer's indefinite hold on its own senior tech. The repair customers — who pay on time, who don't move goalposts — are funding the build customer's stay. They have no idea.

Same structure as Andre, more visible because there's a physical vehicle in a physical bay and a senior person being pulled in two directions.

Three Lenses, One Engagement

Pull back from the personas and look at what the position actually is.

As a credit instrument, it's unsecured exposure on terms a bank's credit committee wouldn't underwrite. No collateral, no guarantee, no reporting requirement, indefinite maturity, zero interest. The customer's good faith is the entire basis — the same kind of position institutional lenders charge premium rates and short maturities to compensate for. The owner extended it for free.

That's just the credit read. The labor side compounds it. The owner's senior people are bound to the engagement until the customer chooses to release them, and the cash funding their hours is coming from work the engagement hasn't legitimately produced as revenue. It's an indenture financed by un-invoiced receivables. Yesterday's unpaid work paying for today's bound labor. Both ends controlled by the same customer.

And underneath both of those is a risk position the owner never priced. Concentration risk, timing risk, scope risk — all absorbed, none compensated. An insurance company absorbs risk profitably because it prices the risk into the premium and holds reserves against the exposure. A vendor who priced concentration into their rate card, timing into their terms, and scope ambiguity into their change-order language has absorbed the same exposures but been paid to carry them. Andre and Jeff didn't price the risk. They didn't signal it. They're carrying the exposure for free.

Three readings, one engagement. The conclusion lands the same regardless of which lens you bring: the business accepted exposure that it did not price, did not bind, and didn't provide a means to close on its own authority.

What the Owner Decided, or Didn't

The whale isn't acting in bad faith here. They're operating rationally against the prices and signals the vendor presented, and those prices said zero. Extension was free. Ambiguity was free. Accommodation was free. The vendor's own behavior taught the customer that, and the customer responded the way any rational buyer would. This dynamic isn't being inflicted on the owner. The owner supplied the inputs that produced it.

This is the third place the same kind of owner decision shows up in this series. Concentration is who you sell to. Terms are when you get paid. Definition is what you agreed to deliver. All three, when undecided by the owner, become risks the owner's business absorbs without being compensated for them. The owner doesn't price what they didn't decide. The customer doesn't price what the owner didn't signal.

The numbers will tell you something is wrong. They won't tell you it started with a decision the owner didn't make.

Working on the operational structure underneath the numbers, not just the numbers themselves.