Every project-based business hits the same moment at month-end. The project management system says progress is one thing. The finance system says something else has been billed. Neither number is necessarily what belongs on the books — earned revenue is its own figure, and it falls to someone to work out what it actually is before the month can be closed. That work often happens manually, subjectively, under time pressure — repeated project by project, every single month, with no consistent record of how revenue decisions were made.
"Progress" is usually someone's feeling, not a measurement. When a project manager reports a project as 70% done, that number is rarely backed by an auditable calculation. It's a gut sense of how the work feels relative to the plan — shaped as much by the client relationship as by the work itself: a project running behind for a difficult or at-risk client might get talked down to avoid a dispute later, while a strategic account nearing renewal might get the benefit of the doubt. Project managers also have a structural incentive to report progress, not to flag that it's behind. That feeling becomes finance's input, with nothing independently checking it.
Cost incurred gets treated as a stand-in for value delivered, and the two aren't the same thing. A project can burn through 60% of its budgeted hours and still be nowhere near 60% of the value delivered — inefficiency, rework, and scope creep all break that link. Using spend as a proxy for progress is common because it's easy to measure, not because it's accurate.
Finance's own read of the data is just as subjective as the PM's. Even when the project data is accurate, turning "75% delivered, three milestones hit" into an actual earned revenue figure is its own judgement call — how much weight a milestone gets when it's substantially but not fully complete, whether cost-to-date should be discounted for scope changes, what counts as earned versus merely underway. That interpretation happens in someone's head, differently each month, with no methodology written down anywhere.
Different people apply different standards, and nothing checks for it. Two account leads on two similar projects can land on completely different definitions of "earned" — one conservative, one aggressive — and there's no mechanism that flags the inconsistency. Finance ends up defending a portfolio number that was never actually one methodology applied twice.
A subjective number doesn't stay contained to the month it was posted in. It resurfaces — in next month's forecast or in an audit later in the year.
Next month pays for this month's guess. Finance has no clean way to know which number to report as earned revenue, so someone pulls the latest project report, cross-references it against what was recognised last period, and picks a number they're comfortable defending — hoping it holds, and that it doesn't need correcting later. That correction might mean recognising little or nothing the following month while the project catches up, or reversing what was already booked. Either way, it isn't limited to the month it lands in — it moves through every month after it: the trend line finance is reporting, the run rate feeding into next quarter's forecast, the number the board was told to expect for the quarter and the year. For a $50M firm, that's not a rounding error — a 2–5% swing is $1M–$2.5M sitting in work-in-progress, understated, overstated, or simply unknown until someone spends the time to find out.
None of it holds up later. An auditor asks why $10,000 was recognised on a project instead of the $15,000 recognised the month before — exactly the kind of judgement call that standards like ASC 606 and IFRS 15 focus on. The reasoning that justified it exists only in someone's memory, or an email thread nobody can find anymore. A number that can't be defended after the fact was never defensible in the first place.
Neither is a separate problem. Both are downstream of the same one — a process that depends on judgement, with no structured way to capture that judgement.
Right now, most project-based businesses are carrying this cost without realising it: a recognition process nobody can fully defend, reconstructed from scratch every month because there was never a structured way to capture the judgement behind it in the first place.
None of this is inevitable, and judgement itself isn't the problem — some level of judgement will always be part of estimating what's been earned. The problem is that it's currently undocumented, inconsistent, and rebuilt from scratch every month, by whoever's closest to the numbers. Structuring that judgement — a consistent methodology, the data to support it at your fingertips instead of reconstructed from memory, a recorded rationale that holds up when someone asks — doesn't remove the human call. It just makes it one a business can actually stand behind.
