Controls should reduce the chance that an error or fraud reaches the financial statements, cash account or board report undetected.
CFO and process owners
Continuous, quarterly testing
Map material finance processes.
Risk and control matrix.
Start with material processes: revenue, purchasing, payroll, payments, treasury, inventory, taxes, financial close and reporting. Identify where a wrong or unauthorised transaction could enter, change or escape review.
Design controls around clear objectives. Preventive controls stop an event; detective controls identify it; corrective controls resolve the cause and effect.
Smaller companies may use founder review or manual reconciliation, but the control must still be evidenced, timely and independent enough to challenge the preparer.
| Control | What it covers | Operating rule |
|---|---|---|
| Process risk | Material error or fraud scenario is defined. | Link to financial statement impact. |
| Control design | Owner, frequency and evidence are specified. | Avoid vague ‘management review’. |
| Operation | The control is performed on the complete population. | Retain dated evidence. |
| Testing | Exceptions and remediation are independently assessed. | Repeat after system change. |
Prioritise controls that protect cash, revenue, liabilities and statutory compliance. A short effective matrix is stronger than a hundred copied controls.
Reassess design after rapid hiring, ERP migration, new locations, acquisitions or a change in business model.
Record the decision, owner, due date and evidence expected. A verbal explanation should become an approved working, board note, contract amendment, statutory filing or reconciliation before the item is treated as closed.
Rules, forms, thresholds and procedures can change. Use the latest official source and the actual company facts rather than copying a prior-year control or another entity’s legal position.
Classify every exception as a timing difference, data error, missing document, legal non-compliance, control-design gap or control-operating failure. This prevents management from treating fundamentally different problems as one ageing list.
The exception file should show amount or exposure, root cause, immediate correction, preventive action, owner and board-escalation threshold. Repeated low-value issues can become material when they reveal weak systems or management override.
Close the item only after the evidence agrees across source documents, books, portal data and management reporting. A screenshot or email promise is not equivalent to a completed filing, lender waiver, signed contract or reconciled ledger.
The control should operate across the full transaction population, not only the samples management expects a reviewer to inspect. For this topic, the key stages are process risk, control design, operation, testing. Each stage should identify the source system, preparer, reviewer, deadline and evidence retained.
A useful management review asks whether the legal document, accounting entry, bank movement, tax treatment and public filing describe the same event. Differences may be valid, but they should be reconciled through a dated working rather than explained from memory during audit or diligence.
Materiality should determine escalation, not whether the company keeps a record. Repeated small exceptions can show weak master data, unclear authority, system bypass or management override. Root cause and preventive action should therefore be documented separately from the immediate correction.
Control evidence should show operation, not merely design. A policy document proves what management intended; a reconciliation, access review, approval log or exception report proves whether the control actually worked during the period.
Manual journals, spreadsheet uploads, administrator access and post-close changes deserve additional scrutiny because they can bypass automated workflows. The reviewer should assess both the entry and the reason normal processing was not used.