The financial close is the process that turns a month of transactions into a set of numbers the board, the bank and the market rely on, and it is audited less often than almost anything it produces. Auditors test journal entries, reconciliations, estimates and disclosures one at a time and rarely test the machine that manufactures them: the calendar, the task list, the sequence of subledger cut-offs, accruals, journals, reconciliations, intercompany eliminations, consolidation and review that has to run correctly, in order, every month, under a deadline. When the machine runs badly, everything downstream is late, rushed and patched, and the patching is where the misstatements and the overrides live. When it runs well, most of the controls the external auditor relies on operate as a by-product.
Speed and accuracy are not opposites in a close; they are the same measure read from two sides. APQC’s cross-industry benchmark, the most cited figure in the field, has for years put top-performing organizations at 4.8 calendar days or fewer to close the month, the median at 6.4 and bottom performers at ten or more, and the slow closes are slow because they are doing rework: chasing reconciliations that should have been done during the month, booking late entries that should have been accruals, and fixing consolidation packs that arrived wrong. This guide covers the close as a process: the terrain and its eight stages, a risk map, a twelve-control starter matrix, how to size and sequence the engagement, a 14-test program, the analytics that profile a close, Pennine Foods plc’s financial close audit with every test’s result, the findings that recur with wording that lands, and scoping variants for SOX and Provision 29 environments, multi-ERP groups, shared service centers and small companies. It sits beside the journal entry testing guide, which covers the entries the close produces, and the management review controls guide, which covers the reviews the close depends on.
In this guide
- Know the terrain: the eight stages and the two clocks
- The close risk map
- The starter RCM: twelve controls that carry the close
- The close calendar as a control: seven properties to test
- Sizing and sequencing the engagement
- The 14-test program
- The analytics that profile a close
- Worked example: Pennine Foods plc’s financial close audit
- The findings that recur, and wording that lands
- Scoping variants: SOX and Provision 29, multi-ERP groups, shared services, small companies
- Where to go next
Know the terrain: the eight stages and the two clocks
A close runs through eight stages, and the order matters because each stage consumes the previous one’s output. Subledger cut-off freezes sales, purchasing, payroll, inventory and fixed assets so that the ledger stops moving. Accruals and estimates book what the subledgers cannot: goods received not invoiced, revenue earned not billed, rebates, bonuses, provisions, impairments, the allowance for doubtful accounts. Manual journals adjust, reclassify and correct, and they are where the journal entry analytics look. Balance sheet reconciliations prove that every account’s balance is supported by something outside the ledger. Intercompany matching and elimination remove the group’s transactions with itself. Consolidation aggregates the entities, translates the currencies and applies the group’s adjustments. Review is the layer of management review controls that is supposed to catch what the earlier stages missed. Reporting and disclosure turn the numbers into the statements and the commentary.
Two clocks run over these stages. The speed clock is the calendar: day one is the first working day after period end, each task has a due day, and the close is finished when the last review is signed. The accuracy clock is the rework: late entries after the close, reconciliation differences carried forward, consolidation adjustments made by hand, restatements of prior periods. Most organizations measure the first clock and not the second, which produces a fast close that is wrong in ways nobody counts. The audit’s contribution is to measure both, and to show where they are in fact the same clock: a close that books 190 entries after its own deadline is slow because it is inaccurate, not in spite of it.
The close risk map
| # | Risk | What it looks like in the file | Where it hurts |
|---|---|---|---|
| R1 | Calendar without teeth: tasks with no owner, no due day, or due days everyone ignores | Tasks completed after the close closes; the same tasks late every month; the calendar maintained in a spreadsheet nobody reconciles to the work | Every downstream stage starts late; review is compressed to a signature |
| R2 | Cut-off leakage: subledgers still moving after the freeze, or frozen before the last transactions arrive | Subledger postings dated in the period after the close; goods received in the period invoiced and expensed in the next | Misstated revenue, cost and working capital at every period end |
| R3 | Estimates by inertia: accruals and provisions rolled forward, re-based never, reversed late | Accrual balances flat for quarters; estimate reversals concentrated in one month; the allowance formula unchanged for years | Profit smoothed by default; the override mechanism the financial statement fraud guide describes, without anyone intending it |
| R4 | Late and post-close journals: entries booked after the deadline, top-side entries at consolidation with thin support | Journals dated after day N; top-side entries by the controller with one-line descriptions; entries reversed the next period | The population where override lives; the close’s accuracy clock, visible |
| R5 | Reconciliations in name only: accounts unreconciled, differences explained as timing for months, plugs | Reconciliations aged past policy; unexplained differences carried forward; reconciling items described identically each month | Balances nobody has proven; a suspense account that is a parking lot |
| R6 | Intercompany drift: mismatches settled by top-side entries rather than by the entities | Intercompany differences above tolerance cleared at consolidation; balances that never settle in cash | Eliminations that hide real differences; profit in the wrong entity, tax in the wrong country |
| R7 | Consolidation on spreadsheets: packs adjusted by hand, versions uncontrolled, formulas untested | Manual adjustments to entity packs; pack versions that differ from the ledger; translation rates typed in | A group number that cannot be traced to entity ledgers |
| R8 | Review as ritual: management review controls performed after the deadline, or signed without a question ever recorded | Review evidence with no queries; reviewer sign-offs seconds apart; reviews dated after the numbers were released | The last line of defense operating as acknowledgment |
The starter RCM: twelve controls that carry the close
Twelve controls carry the close in most organizations, and in a public company most of them are entity-level or financial reporting controls the SOX scoping already treats as key. Build them in the RCM template, mark which exist, and note that several of them are management review controls whose evidence has to meet the standard the management review controls guide sets: what was reviewed, at what precision, what was questioned, and what changed.
| Control | Stage | Risk | How to test it |
|---|---|---|---|
| A close calendar with every task owned, sequenced and dated, tracked in a tool that records completion time, and reviewed by the controller each month for lateness | All | R1 | Reconcile the calendar to the tasks actually performed; profile completion times against due days for twelve months |
| Subledger cut-off enforced by system period lock at a set day and hour, with re-openings approved and logged | Cut-off | R2 | Inspect the period lock configuration and the re-opening log; test postings dated after the lock |
| Accrual and estimate schedules prepared from source data under a documented method, reviewed against actuals when they reverse, and re-based at least annually | Estimates | R3 | Re-perform the largest accruals; compare reversals to actuals for twelve months; inspect the re-basing |
| Manual journals prepared and approved by different people, with support attached, thresholds routing large entries to the controller, and post-close entries requiring a second approval | Journals | R4 | Full-population analytics on the journal file; sample post-close and top-side entries for support and approval |
| Balance sheet reconciliations for every account above a threshold, prepared and reviewed by different people to a deadline, with differences aged, explained and cleared within policy | Reconciliations | R5 | Reconciliation status report for twelve months; a sample re-performed; aged differences traced to resolution |
| Intercompany transactions matched by the entities before consolidation, with differences above tolerance resolved by the entities and settled in cash on a schedule | Intercompany | R6 | Mismatch report at each period end; top-side intercompany entries traced to entity resolution |
| Consolidation in a controlled tool or a spreadsheet under change control, with entity packs loaded from ledgers, translation rates sourced automatically and manual adjustments logged and approved | Consolidation | R7 | Trace the group number to entity ledgers; inspect the adjustment log; test the spreadsheet’s controls where one is used |
| Flux and variance analysis at entity and group level, with explanations at a set precision and unexplained variances escalated | Review | R3, R8 | Inspect twelve months of flux files for precision, questions raised and answers obtained |
| Management review of the consolidated results, the estimates and the disclosure checklist by named reviewers, evidenced with queries and changes, before release | Review | R8 | Review evidence tested to the management review control standard; timing against the release date |
| Segregation between posting, approving, reconciling and reviewing, enforced in the ledger’s access model | All | R4, R5, R8 | Access extract joined to the journal and reconciliation logs |
| Close metrics reported monthly: days to close, late entries, unreconciled accounts, unexplained differences, adjustments after review | All | R1 to R8 | Twelve months of metrics inspected; a metric that moved traced to an action |
| Post-close review: adjustments made after release, external auditor adjustments and prior-period corrections analyzed for root cause each quarter | Reporting | R3, R4, R7 | Inspect the last four quarters’ post-close reviews and the actions they produced |
The close calendar as a control: seven properties to test
Every organization has a close calendar and most of them are documents rather than controls. The difference is testable. A calendar is a control when it has the seven properties below, and the audit’s first hour with the calendar is spent checking each one against the June or September close that actually happened. The template is written so that it can be handed to a controller as the standard the calendar will be tested against next year.
Close calendar standard. 1. Ownership: every task has one named owner and one named reviewer, and a departure triggers reassignment before the next close. 2. Timing: every task has a due working day and hour, sequenced from the subledger locks forward, with no task due before the task it depends on. 3. Dependencies: the calendar records which tasks feed which, so that a late reconciliation shows which review it delays. 4. Evidence: completion is recorded in a tool with a timestamp and a link to the workpaper, never by editing a spreadsheet cell. 5. Escalation: a task late by more than a set number of hours is visible to the controller the same day, and a task late three months running is a calendar change or a staffing change, not a note. 6. Metrics: days to close, late tasks, post-close entries, unreconciled accounts and unexplained differences are reported monthly from the tool, not from the plan. 7. Review: the calendar itself is reviewed after every quarter-end close for tasks that no longer exist, tasks that exist but are not on it, and due days that reality has moved.
Sizing and sequencing the engagement
A close audit is sized by entities and by the state of the reconciliation population. A single-entity company with a reconciliation tool is a 200-hour engagement; a group with a dozen entities, a spreadsheet consolidation and a reconciliation backlog is 320 to 400, most of the difference in the reconciliation and intercompany work. The ranges below assume a group of about ten entities on one ERP with a separate consolidation tool, a few hundred reconciliations a month, and an auditor with the journal analytics already built. Time the engagement to observe at least one close live; a close audited only from its artifacts misses the calendar’s real behavior.
| Phase | What happens | Hours |
|---|---|---|
| Planning and criteria | Close calendar, policies for accruals, journals and reconciliations, the approval and threshold matrices, the consolidation manual and the external auditor’s recent points obtained; the RCM drafted | 30 |
| Observing a close | One month-end close followed live from day one to release, with the calendar tracked against reality | 30 to 40 |
| Data acquisition | Twelve months of journals, reconciliation status, intercompany reports, consolidation adjustments, task completion logs and close metrics, with completeness proofs | 30 |
| Close analytics | The profile below run on the year; hits scored | 40 to 50 |
| Control tests | The 14-test program, aimed by the analytics; reconciliations re-performed; estimates re-performed; review evidence tested | 110 to 140 |
| Consolidation trace | The group number for one quarter traced from entity ledgers through packs, eliminations and adjustments to the statements | 30 to 40 |
| Reporting | Findings with the two clocks quantified; actions agreed with the controller and, where material controls are affected, with the audit committee | 40 |
| Total | 310 to 370 |
Sequence the live observation early, because it tells you which tasks are theater before the analytics confirm it; run the analytics before the reconciliation work, because the aging report chooses which accounts to re-perform; and do the consolidation trace last, after the entity-level findings are known, so that the trace explains the group number rather than discovering it.
The 14-test program
| # | Test | Population and method | Risk |
|---|---|---|---|
| 1 | Observe one close from day one to release; reconcile the calendar to the tasks performed; redraw the RCM | One period | All |
| 2 | Calendar discipline: task completion times against due days; tasks routinely late; tasks with no owner | Twelve months of task logs | R1 |
| 3 | Cut-off: period lock configuration; re-opening log; subledger postings dated after the lock; the six-week search for unrecorded liabilities at two period ends | Configuration; full posting population; two period ends | R2 |
| 4 | Estimates: the ten largest accruals and provisions re-performed from source; reversals against actuals for twelve months; re-basing evidence | Ten estimates; twelve months of reversals | R3 |
| 5 | Journals: the analytics catalog run on the year; post-close and top-side entries sampled for support and approval; preparer-approver segregation from the log | Full journal file; sample 40 | R4 |
| 6 | Reconciliations: status report for twelve months; twenty-five re-performed including every account with a difference older than policy; suspense and clearing accounts aged | Full status population; sample 25 | R5 |
| 7 | Intercompany: mismatch reports at four period ends; top-side intercompany entries traced to entity resolution; cash settlement tested | Four period ends; full top-side population | R6 |
| 8 | Consolidation: the group number for one quarter traced to entity ledgers; adjustment log tested; translation rates traced to source; spreadsheet controls tested where used | One quarter, full trace | R7 |
| 9 | Flux analysis: twelve months of variance files tested for precision, questions raised and answers obtained | Twelve months | R3, R8 |
| 10 | Management review controls: review evidence for the consolidated results, estimates and disclosure checklist tested to the standard; timing against release | Four quarter ends | R8 |
| 11 | Segregation from access: post, approve, reconcile and review rights by user, joined to the logs | Access extract | R4, R5, R8 |
| 12 | Close metrics: twelve months of reporting; days to close reconciled to the task logs; a metric that moved traced to an action | Twelve months | R1 |
| 13 | Post-close adjustments: entries after release, external auditor adjustments and prior-period corrections for four quarters analyzed to root cause | Four quarters | R3, R4, R7 |
| 14 | Prior points: external auditor management letter points and prior internal audit findings on the close re-tested | All open and closed points | All |
The analytics that profile a close
The journal-level analytics live in the journal entry analytics catalog and are run in full here. The eight profiles below are the close-level ones: they read the process rather than the entries, and they choose which reconciliations, estimates and reviews the file work examines.
| Analytic | Logic | A hit usually means |
|---|---|---|
| Task lateness by task | Completion time minus due time, by task, over twelve months | The same tasks late every month: a calendar that is wrong, or an owner who is overloaded |
| Post-close journal share | Manual journals dated after the close deadline as a share of all manual journals, by month and preparer | The accuracy clock; rework concentrated in the people who fix what the close missed |
| Top-side concentration | Consolidation-level entries by preparer, description length and support attached | Where override would live if it lived anywhere; the population the quarterly program tests |
| Reconciliation aging and repetition | Accounts by months since last reconciled; reconciling items with identical descriptions across months | Reconciliations in name only; a difference being carried rather than resolved |
| Estimate flatness and reversal | Accrual and provision balances by month; reversal amounts against actuals | Estimates rolled forward by inertia, or trued up in one month to hit a number |
| Intercompany mismatch trend | Mismatch total and count at each period end, and the share cleared by top-side entries | Entities not resolving their own differences; eliminations hiding real ones |
| Period re-openings | Re-openings of locked periods by user, reason and approval | Cut-off enforced until it is inconvenient |
| Reviewer velocity and query rate | Time between reviewer sign-offs; number of queries recorded per review | Review operating as acknowledgment; the management review control without the review |
Worked example: Pennine Foods plc’s financial close audit
Pennine Foods plc is the 900-million-pound UK food manufacturer used in the UK Corporate Governance Code guide, with a December year-end and a first Provision 29 declaration built on 42 material controls in seven groups, twelve of them in the financial reporting group. It reports through nine entities in the UK, Ireland and the Netherlands on one ERP with a separate consolidation tool, runs about 340 close tasks a month, books roughly 2,400 manual journals a month across the group, and reconciles about 410 balance sheet accounts. Its close calendar says the month closes on working day eight, the quarter on day twelve, and the board pack goes on day twenty. The financial close audit followed the quarterly override program that tests the group’s top-side entries, and was timed to observe the June close live. It took 340 hours.
| Test | What it found | Disposition |
|---|---|---|
| 1. Observed close | The June close released on working day eleven against a calendar that said eight; the last three days were spent on reconciliation differences and two entity packs that arrived wrong. | The two clocks quantified for the report |
| 2. Calendar discipline | Of 340 tasks, 31 were late in every one of the twelve months, most of them reconciliations owned by two accountants; 22 tasks had no owner since a departure in March and were being done by whoever noticed. | Finding, Medium: the calendar |
| 3. Cut-off | Period lock at 18:00 on day three; 41 re-openings in the year, 29 requested by one accountant and approved by email after the fact; 1,900 subledger lines posted after the lock. Unrecorded liabilities at December: nine items, 1.1 million pounds, below materiality and inside the pattern the external auditor had raised. | Inside finding 2; the re-opening approval moved into the system |
| 4. Estimates | The rebate accrual, a Provision 29 material control, re-performed and supported; the bonus accrual rolled forward unchanged for three quarters; onerous contract provisions unchanged since 2024 with two of the contracts since renegotiated. | Finding, Medium: estimates by inertia |
| 5. Journals | 28,900 manual journals in the year; 2,150 dated after the day-eight deadline, a median of 190 a month; 22 top-side consolidation entries with one-line descriptions and no attached support beyond a spreadsheet reference; no second approval on post-close entries. | Finding, High: post-close and top-side journals |
| 6. Reconciliations | 410 accounts; 38 unreconciled for more than two months at year-end; fourteen with unexplained differences totaling 2.1 million pounds gross and 0.6 million net; nine accounts carrying identical reconciling items for six months or more, described as timing. | Finding, High: balance sheet reconciliations |
| 7. Intercompany | Mismatches of 1.3 million pounds at year-end cleared by top-side entries rather than by the entities; cash settlement quarterly for five of nine entities and never for the other four. | Finding, Medium |
| 8. Consolidation | Packs loaded from the ledger for seven entities; the two Dutch entities’ packs prepared in spreadsheets with manual adjustments; 2.9 million pounds of consolidation adjustments logged in the year, eight of them unapproved. | Finding, Medium: consolidation controls |
| 9. Flux analysis | Precision of 250,000 pounds or 5 percent; questions recorded in four of twelve months; the other eight months’ files showed explanations written by the preparer with no reviewer query. | Inside finding 5 |
| 10. Management review | The chief financial officer’s review of the consolidated results evidenced by sign-off alone in three of four quarters, with queries recorded only for the December close. | Inside finding 5, reported as affecting the group consolidation review material control |
| 11. Segregation | Seventeen users held post-and-approve rights for manual journals; twelve had used both in the year. | Finding, Medium |
| 12. Close metrics | Days to close reported to the board as eight every month, the calendar’s figure; the task logs showed a median of eleven. | Finding, Low: the metric measured the plan |
| 13. Post-close adjustments | Six external auditor adjustments in the year, 1.7 million pounds net, four of them arising from two of the fourteen unexplained-difference accounts; no root-cause review performed. | Finding, Low |
| 14. Prior points | Three of five management letter points open for more than twelve months, all on reconciliations. | Inside finding 6, escalated |
The report carried nine findings, two High, five Medium and two Low, and an overall rating of Needs Improvement. The two High findings were reported to the audit committee as affecting two of the twelve financial reporting material controls in the Provision 29 declaration, the group consolidation review and the manual journal control, which turned a close audit into a governance conversation. Three things generalize. The two clocks were the same clock: the close was three days late because 190 post-close entries a month were fixing what the reconciliations had not caught, and the reconciliations were late because the two accountants who owned them were also booking the post-close entries. The board had been told the close took eight days for two years because the metric measured the calendar rather than the task logs, which is the most common close finding on this site and the cheapest to fix. And the external auditor’s six adjustments had come from the same two accounts the internal reconciliation status report had shown as unexplained for most of the year; the information existed, and nobody had joined the two lists.
The findings that recur, and wording that lands
Close findings land when they quantify both clocks and name the stage, because a controller who hears “the close is slow” hears an opinion and a controller who hears “190 entries a month after your own deadline” hears a number. The two below recur in most engagements, in the five-Cs form.
Balance sheet reconciliations. Condition: at year-end, 38 of 410 balance sheet accounts had not been reconciled for more than two months, fourteen accounts carried unexplained differences totaling 2.1 million pounds gross and 0.6 million pounds net, and nine accounts had carried identical reconciling items described as timing for six months or longer; three external auditor management letter points on the same accounts have been open for more than twelve months. Criteria: the group accounting manual requires every account above the threshold to be reconciled by working day six, reviewed by day eight, and cleared of differences within sixty days; the reconciliation review is a material control in the Provision 29 declaration. Cause: two accountants own 140 of the reconciliations alongside the post-close journal work, the status report is produced but not escalated, and the reconciliation tool permits roll-forward of a prior month’s reconciling items without re-investigation. Consequence: balances of 2.1 million pounds gross whose support cannot be demonstrated, four external auditor adjustments in the year arising from two of these accounts, and a material control that cannot currently be shown to operate. Corrective action: the controller will redistribute ownership by the next close, disable roll-forward of aged items in the tool, escalate the status report to the chief financial officer monthly, and clear the fourteen differences by the half-year; internal audit will re-perform the fourteen at that date.
Post-close and top-side journals. Condition: 2,150 of the year’s 28,900 manual journals were posted after the close deadline, a median of 190 a month, without the second approval the policy requires for post-close entries; 22 top-side consolidation entries carried one-line descriptions and no attached support; the chief financial officer’s review of consolidated results was evidenced by a sign-off alone in three of four quarters. Criteria: the journal policy requires post-close entries to be approved by the controller with support attached, and the group consolidation review to be evidenced with the matters questioned and resolved. Cause: post-close entries are the mechanism by which reconciliation differences and pack errors are corrected after the deadline, the workflow does not distinguish them from in-period entries, and the review is scheduled after the numbers have been released to the board. Consequence: the population where management override would live is the least controlled population in the close, and two material controls in the declaration lack the evidence to support it. Corrective action: management will configure the workflow to route post-close entries to the controller with mandatory support, require the consolidation review to precede release with queries recorded, and reduce post-close volume by resolving reconciliation differences in-period; the quarterly override program will report the post-close share to the audit committee.
Scoping variants: SOX and Provision 29, multi-ERP groups, shared services, small companies
In a SOX 404 or Provision 29 environment, most of the twelve controls are already designated as key or material, and the close audit’s findings translate directly into deficiency evaluations and declaration consequences; coordinate scope with the SOX or Provision 29 program so that the internal audit engagement tests the process and the compliance program tests the controls’ evidence, rather than both testing the same reconciliations twice, and use the control deficiency evaluation guide to rate what the close audit finds. Multi-ERP groups, after acquisitions or by history, move the risk to the mapping and the packs: each ledger’s chart maps to the group chart, each mapping is a control, and the consolidation tool’s load is where entities’ errors become the group’s; the audit adds a mapping test per ledger and a pack-to-ledger trace per entity. Shared service centers concentrate the reconciliations and journals and leave the estimates and reviews with the entities, which puts the risk at the boundary: whether the center’s reconciliation status reaches the entity controllers who own the balances, and whether the center can refuse a bad accrual. Small companies with one accountant cannot segregate posting, approving, reconciling and reviewing; the compensating control is an owner or finance director who reviews the post-close entries and the reconciliation list with the evidence in front of them, and the audit tests whether that review is real, which is usually the whole audit.
Where to go next
Audit the close as the machine it is. Observe one live, measure both clocks, run the journal analytics in full, re-perform the reconciliations the aging report points at, trace one quarter’s group number to the entity ledgers, and write the findings so that the stage and the number are in the first sentence. The entries the close produces are covered in the journal entry testing guide and scored in the journal entry analytics catalog; the reviews it depends on are covered in the management review controls guide; and the bank side of the reconciliations is covered in cash management and bank reconciliations. When the post-close population starts to look like a mechanism rather than rework, the financial statement fraud guide explains which of the five it is.
Related guides
- How to audit journal entries — the management override lens, with Lakeshore’s year-end run
- The journal entry analytics catalog — risk-scoring the general ledger
- Financial statement fraud — the mechanisms, with Pennine’s quarterly override program
- Management review controls — documenting and testing the reviews the close depends on
- SOX scoping and risk assessment — why the close’s controls are key
- UK SOX in 2026: Provision 29 — Pennine’s declaration and the material controls behind it
- Evaluating control deficiencies — from exception to material weakness
- How to audit cash management and bank reconciliations — the bank side of the reconciliation population
- How to audit order-to-cash end to end — where revenue cut-off and the allowance come from
- How to audit procure-to-pay end to end — where accruals and unrecorded liabilities come from
- How to audit fixed assets — depreciation, impairment and construction in progress at the close
- The risk and control matrix template — where the twelve-control matrix lives
- Writing the audit work program — the five-element procedure standard
- The sampling memo template — recording the reconciliation and journal samples
- Segregation of duties beyond the ERP — post, approve, reconcile and review kept apart
- The 5 C’s of audit findings — the form the two findings above are written in
- The finding and issue log template — tracking nine findings and two material-control consequences
- Fieldwork and testing guides and internal controls guides — the full collections
- Pre-IPO SOX readiness — building a close that meets public-company deadlines
Leave a Reply