Every Contractor Has a Profit Figure. Most of Them Are Reading a Timing Difference
· 7 min read · Faceela
Cost value reconciliation compares what a project has earned against what it has cost, measured to the same date, to produce the only honest margin figure a contractor has while the job is still running.
It fails, on most projects, for a single reason that has nothing to do with arithmetic: value and cost both arrive late, by different amounts, from different directions.
Value arrives late because work is executed weeks before it is certified. Cost arrives late because a subcontractor's invoice for work done in March may land in May, and materials delivered on site are consumed long before the supplier bills for them. Compare the two as they sit in the ledger and you are not measuring performance — you are measuring the difference between two lags.
The direction of the error is consistent and dangerous. Early in a job, cost lags more than value, so the project looks profitable. It stays looking profitable for as long as the invoices keep arriving behind the work, which on a long contract can be most of it. The correction arrives late, all at once, and is usually described afterwards as a problem that appeared suddenly. It did not. It was being reported wrongly the whole time.
That is the whole answer. The rest of this article is how each side is properly built, and what makes the difference between a CVR that informs a decision and one that confirms a hope.
Value is not what you invoiced
The value side has three components, and only the first one is a document.
Certified to date is what the consultant has approved. It is reliable and it is behind reality.
Applied but not yet certified is what you have submitted and are waiting on. It belongs in value at your assessment of what will be certified, not at what you asked for — the difference between those two figures is itself a useful number, and a contractor who cannot state it does not know how their applications are being received.
Executed but not yet applied for is work physically done since the last application cut-off. It is real earned value that appears in no document at all, and leaving it out is the most common way a CVR understates value and manufactures a fake loss at month end.
All three sit on the cumulative position the payment certificate revalues each period, which is why a CVR built on a system that stores certificates as monthly invoices rather than cumulative valuations is starting from a broken foundation.
Cost is not what you have been invoiced
The cost side has the same problem, more acutely, because more parties are involved.
| Component | Why it is missed |
|---|---|
| Supplier invoices received | It is not missed — this is the part everyone has |
| Goods received not invoiced | Delivered and consumed; the invoice has not arrived |
| Subcontract work executed not certified | Their application is pending your assessment |
| Labour and plant to date | Often posted on a payroll or hire cycle that does not align with the cut-off |
| Accruals for known commitments | Ordered, committed, not yet delivered or billed |
The rule that makes the whole exercise work is that cost must be recognised against the period the work happened in, not the period the paperwork arrived in. That is what accrual means here, and it is the discipline that separates a CVR from a printout of the purchase ledger.
It is also the single hardest thing to sustain, because it requires people who are not accountants — site and commercial staff — to record commitments and receipts as they happen. Software can make that cheap or expensive. It cannot make it optional.
Cut-off is the discipline nobody enjoys
Both sides must be measured to the same date. Stated that way it is obvious; in practice it is where most reconciliations quietly break.
The failure looks like this: value is taken to the certificate date, cost is taken to the accounting month end, and the two are a fortnight apart. A fortnight of cost with no matching value is a fortnight of pure invented loss — and it appears every single month, so it stops looking like an error and starts looking like the business.
The fix is not clever. It is a stated cut-off date, applied to both sides, on which everyone agrees before the month runs rather than during it. What software contributes is the ability to ask for the position as at a date and get a consistent answer — which requires transactions to carry the date the event happened, not just the date they were entered.
A system that cannot restate last month's position on demand cannot support a CVR at all, because the moment a late invoice is posted, the previous month's reported margin silently changes and no one can reproduce what was reported at the time.
The number that makes it useful: cost to complete
Everything above is history. The figure that turns a CVR from a scorecard into a decision tool is the forecast.
Cost to complete is what remains to be spent to finish the contract. Added to cost incurred, it gives forecast final cost; set against forecast final value, it gives the margin the job will actually deliver rather than the margin it has delivered so far.
This is the number that catches a problem while there is still time to act, and it is the number most often replaced by a guess. The honest version is built from the remaining measured work at current expected rates, plus known committed costs, plus a considered figure for risks that have already materialised — not the tender allowance, which is what it gets replaced with when nobody has time.
The connection to the rest of the chain is direct: remaining work comes from the bill, at rates whose build-ups tell you what they were expected to cost, which is why the estimator's bill dying in a spreadsheet is not a filing problem but the reason the forecast cannot be built. Variations that have been agreed but not yet certified belong in forecast value, and ones instructed but not yet valued belong in the risk position.
Why the general ERP cannot do this
A conventional ERP reports profit by comparing revenue recognised to cost posted, both driven by documents. That is correct for a business that invoices what it delivers when it delivers it.
Contracting is not that business. Revenue is certified by somebody else, on their timetable. Cost arrives from dozens of parties on their own timetables. The margin at any moment is a computed position across five or six accruals, not a subtraction of two ledger balances.
This is why contractors so often run two systems in effect — the accounts, which are correct for statutory purposes and useless for management, and a commercial spreadsheet, which is useful and owned by one person. The spreadsheet is not a failure of discipline. It exists because the ERP models a different business.
The recurring cost of maintaining it is a real, measurable annual figure, and it is the same category the cost of chaos calculator turns into an order of magnitude — worth establishing before anyone prices a replacement.
What to make a vendor show you
- Produce a CVR for one project as at a stated date, with value and cost both to that date.
- Show executed-not-applied value included, and where the figure came from.
- Show goods received not invoiced and subcontract work not certified inside cost.
- Post a late supplier invoice dated to a prior period, then reproduce the prior month's CVR unchanged.
- Enter a cost to complete and show forecast final margin move.
- Show the same reconciliation across all live projects on one screen.
- Decompose the margin on one project to the bill items that produced it.
Item 4 is the test that separates real systems from reporting layers. If posting a backdated invoice rewrites history with no ability to restate what was reported, the CVR cannot be trusted for any decision, because no version of it can be defended a month later.
Item 7 is the one that pays for itself over years, for the reason set out in the bill of quantities: a margin you cannot decompose teaches you nothing about your next tender.
The short version
A CVR compares earned value against incurred cost at one date. Both sides are mostly made of things that have happened and not yet been documented, which is why the naive version reports a timing difference and calls it profit.
Getting it right needs three things: accrual discipline on both sides, one cut-off date applied to both, and the ability to restate a prior period after late documents land. Getting it useful needs a fourth — a cost to complete built from remaining measured work rather than from the tender allowance.
That closes the chain this cluster describes: the bill prices the work, the certificate values it cumulatively, retention and advance recovery are deducted from it, variations change it, and the CVR is where you find out whether any of it made money.
What that looks like built, as one system rather than six, is on the contracting page.