There is a category of money that appears on no report in your business: work you delivered and never invoiced.
It is not a loss, so nothing flags it. It is not a receivable, so nobody ages it or chases it. It has no line in the P&L, no cell in the cash-flow forecast, no place in the month-end pack. It is simply revenue that never arrived, and in project businesses, businesses that deliver over weeks and bill in milestones, it is routinely the largest single leak in the building, precisely because it is the least visible.
This article is about why unbilled work accumulates structurally, how to find out in one question whether you have the problem, what a month-end should look like in a project business, and the checklist that recovers the money, on paper, before any software is involved.
Why it happens: the gap between "done" and "billed"
Start with an uncomfortable observation about roles. In almost every service firm, the person who delivers a milestone is not the person who raises the invoice. Delivery finishes the work; finance bills it. Between those two functions sits a gap, and the gap has a specific, repeatable failure mode.
Delivery finishes the milestone and moves immediately to the next urgent thing, because delivery always has a next urgent thing. In their world, the work is done, and done is the terminal state. Finance, meanwhile, bills from whatever information reaches them: a status update, a mention in a meeting, a spreadsheet that someone remembers to update. If the information does not reach them, nothing happens. Not a failure, not an error, nothing, which is the problem. A milestone can rest indefinitely in the state "done in one system, nonexistent in the other", and there is no alarm anywhere whose job is to notice.
Now add the textures of real project work. Partial deliveries that were "basically done" pending one small thing. Scope handled outside the original milestones. The change request that was agreed verbally and delivered helpfully. Projects that wound down gradually rather than ending crisply, so the final milestone never had its moment. Every one of these produces delivered value with no billing trigger attached, and each instance seems too small to matter, which is exactly how a large number is assembled from unremarkable parts.
The one-question test
You can diagnose this in your own firm in the time it takes to read this sentence: can you produce, right now, a list of every piece of delivered-but-uninvoiced work, with values?
If the answer is yes, one screen, current, trusted, you do not have the problem, and you are unusual.
If the answer involves opening more than one tool, asking more than zero people, or the phrase "let me check with", then you have the leak. You just have not measured it, and unmeasured leaks are always assumed smaller than they are. Firms that do the exercise for the first time, listing everything delivered in the last year against everything invoiced, are reliably unsettled by the result. Not because anyone was negligent, but because the structure guarantees accumulation: the gap between done and billed only ever collects work, it never empties itself.
Month-end as investigation, and what it should be instead
Here is the difference between a product business and a project business at month-end. A product business runs a report: units shipped, invoices issued, done. A project business, without connected systems, runs an investigation: someone from delivery, someone from finance, and someone with the spreadsheet, reconstructing the month from memory, email, and mutual interrogation. What did we actually finish? Was that invoiced? Did the client sign that PO? Why is this milestone marked both complete and pending?
The investigation costs three ways. It costs the hours themselves, senior hours, every month, forever. It costs timing: invoices raised on the 28th instead of the 5th push your own cash collection back most of a month, which in a business that pays salaries on the 1st is not an accounting subtlety. And it costs accuracy: memory-based reconciliation misses things, and the things it misses are precisely the unbilled items this article is about. The investigation is not the solution to the leak. It is the leak's most expensive symptom.
What month-end should be, in a project business, is the same thing it is in a product business: a report. That requires exactly one structural property, that delivery and invoicing live on the same record. When marking a milestone delivered creates a billable fact that finance can see without being told, the gap between done and billed closes by construction. Unbilled work becomes a list you clear weekly, not a mystery you excavate quarterly.
The checklist: six checks in the order that finds the most money
Whether or not you ever touch software, run this at month-end. The order matters; it is sorted by expected recovery.
Reconcile delivered against invoiced. Line by line, this month's completed milestones against this month's raised invoices. Every unmatched delivery is money waiting for a trigger. This single check usually pays for the whole exercise.
Chase approved-but-unsigned POs. Work that started on a verbal yes, with the PO "coming", is work you may struggle to bill at all if the relationship sours. Get the paper while everyone is still happy.
Confirm VAT treatment on new deals. Inclusive or exclusive, decided and written on the deal, before the first invoice makes the ambiguity expensive. (We wrote a whole piece on this one; the wrong word costs 5%.)
Age receivables past 30 days. Standard, but do it after the unbilled check, not instead of it. Most firms chase receivables diligently while sitting on unbilled work worth more, because receivables have a report and unbilled work does not.
Flag deals quiet for 14+ days. Silence in the pipeline is the same disease as unbilled work in delivery: a state nobody is watching. Two weeks of nothing is a flag, not a pause.
Match every invoice to its milestone. The reverse check: invoices that correspond to no recorded delivery are how billing errors and awkward client conversations are born.
Six checks, monthly, on paper if necessary. The first month is slow. By the third month the list is short, because the checks change behaviour upstream: once people know delivered work gets reconciled, delivered work starts getting reported.
The structural version
In Trazer, this checklist is mostly screens rather than meetings, because the structure does the work. Milestones live on the same record as the deal and the invoice, so delivered-but-unbilled is a standing list, visible any day of the month, not just at the excavation. Change requests are priced to exist, which means the helpful extras that used to vanish now carry values. The PO is a recorded stage, so work-started-without-paper is visible the moment it happens. The stale-deal flag watches the pipeline's silences automatically. And when the invoice is raised, its VAT rate freezes and it posts to Zoho Books in one click, so the FTA-facing books agree with operations by design.
The result, in our own operation, was that month-end changed in kind: from an investigation that produced a number nobody fully trusted, to a report that was finished before the coffee was. That is Case Study Zero, our own business, and it is the honest extent of our claims.
But the checklist comes first, and it is free. Run it once this month. The money it finds was always yours; it was just resting in the gap between two systems that never learned to talk.
The objection worth answering: "we'd rather not invoice aggressively"
When firms first confront their unbilled backlog, a hesitation often surfaces: chasing every small deliverable feels petty, and nobody wants to be the vendor who nickels clients. It sounds like commercial wisdom. It is worth taking apart, because it quietly defends the leak.
Billing what was agreed is not aggression; it is administration. The client signed a scope with milestones and prices, and invoicing a delivered milestone promptly is the most neutral act in commerce, far more neutral than the alternative, which is invoicing it months later, out of context, when the client's budget year may have turned and the person who approved the work may have moved on. Late invoices, not prompt ones, are what actually generate friction: they arrive as archaeology, they require both sides to reconstruct justification, and they are the invoices that get queried. The kindest thing you can do for a client relationship is bill accurately and immediately, while the value delivered is still fresh in everyone's memory.
As for the genuinely small extras, the half-hour favours, the fix is not to bill them grudgingly or absorb them invisibly, but to record them and decide deliberately. A recorded, priced change request that you then choose to waive is a gift the client can see, and generosity that is visible builds relationships. Absorbed work builds nothing; by definition, nobody knows it happened. If you are going to invest margin in goodwill, and sometimes you should, at least get the goodwill.
And if the list it produces is longer than you expected, bring it to a thirty-minute walkthrough. We will show you what it looks like when the gap does not exist.
Bring your messiest deal.
We will map it through Trazer live: where it would have been flagged, where the handoff would have been caught, where the invoice would have raised itself. Thirty minutes, no pitch deck.
Request a walkthrough
