Your Q1 Pack Says the Quarter Was Fine. The Invoices That Would Change That Answer Arrived Last Week.
Your Q1 Pack Says the Quarter Was Fine. The Invoices That Would Change That Answer Arrived Last Week.
Completeness is the one assertion a fully reconciled ledger cannot prove — and the only one that turns a good quarter into a quarter you have to take back. A deliberate search, an evidence trail, and the part you cannot hand to a tool.
The September quarter closed three weeks ago. In most organisations under $30 million, the Q1 management reporting pack is being finalised about now — reconciliations done, variances explained, commentary written, board papers due. Everything agrees.
And the pack can still be wrong, in one specific direction, for a reason none of that work would detect: a cost that belongs in September was never recorded, because the paperwork for it turned up in October.
This is not an exotic risk. It is one of the most common reasons a quarter gets restated in a small finance function, and it is systematically under-searched because the controls that would catch it don't exist — not because they failed.
Why a reconciled ledger cannot tell you this
Reconciliation compares one record of what happened against another record of what happened. The bank reconciliation proves that every transaction the bank saw is in your ledger and vice versa. The creditor reconciliation proves your ledger agrees with your creditors' ledgers. Both are tests of things that exist.
Completeness is the opposite shape of problem. It asks whether something that should exist is absent — and no control that works by comparing two lists of present items will ever detect an absence from both of them. The bank does not know about the contractor who hasn't invoiced yet. Your creditors ledger does not know about an invoice sitting in a manager's inbox awaiting approval.
That is why completeness has to be searched for deliberately, on a schedule, rather than relied upon to surface. Purchase orders and approval workflows narrow the gap; they rarely close it, because the costs that arrive late are disproportionately the ones nobody raised a PO for. It is also, incidentally, exactly what an auditor spends a meaningful part of the year-end doing — and the organisations that find this painless are the ones that already did it themselves, quarterly.
Worth separating this from a related exercise I wrote about a few weeks ago. A supplier statement reconciliation asks one external party what it thinks you owe. It is excellent, and it only covers suppliers who send statements. The search below covers the costs where nobody sends you anything.
Seven places September is hiding in October
The method is a fixed set of sources, searched over a fixed window — I use three weeks past the period end, extended to the date of sign-off — with every hit tested for which period it belongs to.
| Where to look | What you are looking for |
| October bank lines and direct debits | Payments with no matching September invoice or accrual — the fastest, highest-yield test on the list |
| Invoices entered in October | Anything dated, or describing a service period, on or before 30 September |
| The AP inbox and approvers' inboxes | Invoices received but never entered, because approval stalled |
| Expense claims and card statements | September spend claimed late, and the statement period that straddles the date |
| Contracts and engagement letters | Work performed and not yet billed — legal, audit, recruitment, consultants, repairs in progress |
| Payroll and leave records | Overtime and allowances paid in October for September, the leave balance movement, and on-costs on both |
| Grant and funding files | Unspent funds that are repayable, and acquittal obligations crystallising at the period end |
The last one is where sector-specific exposure lives. In an NFP, NDIS or aged care setting, a funding condition can create a liability out of money already sitting in the bank — which is the most counter-intuitive line on the balance sheet to a non-finance board member, and the one most worth explaining in the pack rather than in a footnote.
Adjusting or not: the question that keeps the search honest
A search for missing costs can quietly become a tool for moving results around, which is why the classification step matters as much as the finding step.
AASB 110 Events after the Reporting Period draws the line that applies: an adjusting event provides evidence of conditions that already existed at the reporting date, and the numbers are changed. A non-adjusting event arose after it, and is disclosed rather than booked. In practice, an invoice dated 18 September arriving on 12 October is almost always adjusting — the service was consumed in the period, and you have merely learned the amount. A contract signed on 10 October is not, however much it affects the next quarter.
Management reporting is not a financial report and AASB 110 does not bind it. Borrowing the distinction anyway is what stops the exercise drifting: every item gets classified, and the ones you decide not to book are recorded as decisions with reasons, not quietly dropped. The auditing standard that covers the same ground, ASA 560 Subsequent Events, is a useful preview of the questions you will be asked later.
One page of evidence, every period
The output of the search is not the journal. It is a one-page schedule, filed with the period, carrying: each source searched, the date range covered, who searched it, every item found, its amount, its classification, and where it was posted or why it wasn't.
Three things that schedule does that the journals alone never will. It proves the search happened, to a board or an auditor, months later. It makes the omissions visible — a source you skipped is a blank row rather than an invisible gap. And it compounds: the recurring items show up in the same place each quarter, which turns a two-day exercise into a half-day one by about the third cycle.
If you do nothing else from this post, start the schedule this quarter and keep it for four. The fourth — a full year of them — is where the value shows.
Where AI earns its place, and the trap it sets
Three genuinely good uses. Matching October bank lines and newly entered invoices against the September accrual listing, and returning the unmatched items with a reason. Reading a folder of contracts and engagement letters to extract service periods and billing milestones, so you can see which obligations should have been accrued. And drafting the schedule itself from the evidence you supply.
Now the trap, because it is a specific one. Ask a tool "is anything missing from September?" and it will answer — fluently, and from nothing. It can only see what you handed it, so an absence from your extract is invisible to it in precisely the way it is invisible to your reconciliations. The judgement that cannot be delegated is the choice of which seven sources to search. The tool can tell you what is unmatched inside a dataset; only a person decides that the dataset should also have included the grant files.
We build this as a standing quarterly step inside clients' reporting packs, and the part worth paying for is never the matching. It is the source list, tuned to where that organisation's costs actually arrive late — which differs enough between a disability provider and a childcare operator that a generic template would mostly search the wrong places.
Why this is a leadership issue, not a technical one
Here is the reason I would put this in front of a board rather than leaving it in the close checklist.
The quarter you report is the quarter the organisation plans from. If Q1 is overstated by six weeks of late-arriving costs, the reforecast built on it is wrong in the same direction, the next budget conversation starts from a number that is about to move, and the correction arrives at Q2 — when it reads as a deteriorating business rather than a reporting artefact. I have watched that sequence cost a finance function more credibility than the original error ever warranted.
A variance explanation is only as good as the completeness of what it is explaining, which is the quiet dependency behind everything I wrote about letting AI draft the commentary. Three weeks after period end is the cheapest moment this will ever be fixable. By February it is a restatement, or a catch-up adjustment that reads like one.
Could you show a board the evidence that nothing is missing from Q1?
Reconciliations prove what's there. Almost nobody can produce the page that proves what isn't. PFL provides senior-level outsourced finance, management reporting, and AI automation for Australian NFP, NDIS, and SME organisations — including a completeness search built into the quarterly pack, with the schedule to prove it ran.
Talk to PFL →
Comments
Post a Comment