Side-by-side comparison
| Dimension | Tests of details | Substantive analytical procedures |
|---|---|---|
| What it tests | Individual transactions, balances, disclosures, or estimates | Relationships and patterns across an entire population |
| Evidence strength | Direct: confirms whether a specific item is correctly stated | Indirect: identifies unexpected fluctuations that may indicate misstatement |
| Precision required | Determined by sample size and selection method | Determined by the precision of the independently developed expectation (ISA 520.5(c)) |
| Best suited for | Assertions where individual large items drive the balance (existence, valuation) | Assertions where the population is predictable and homogeneous (completeness of revenue, payroll) |
| What triggers follow-up | Misstatements found in tested items | Fluctuations that exceed the threshold set at planning |
Key Points
- Tests of details give you direct evidence about individual items; analytical procedures give you indirect evidence about an entire population.
- Analytical procedures work best when the underlying data is predictable and the expectation can be set precisely.
- For significant risks, ISA 330.21 requires substantive procedures that are specifically responsive, which often means tests of details.
- Using the wrong type for the assertion wastes time and leaves gaps a reviewer will find.
When the distinction matters on an engagement
ISA 330.18 requires substantive procedures for every material balance and class of transactions, as well as for disclosures. The engagement team must decide which type (or combination) is appropriate for each assertion. That decision depends on the nature of the assertion and the predictability of the underlying data.
Where a population is stable and the relationship between financial data and an external driver is well understood, analytical procedures can be more effective than testing individual items. ISA 520.5(a) requires the auditor to determine whether the analytical procedure is suitable for the assertion. Payroll expense at a client with 200 employees and no restructuring is a good candidate. A provision for litigation with one large uncertain claim is not. If the team applies an analytical procedure where the data is too volatile to set a precise expectation, the procedure produces weak evidence. ISA 520.A5 makes clear that the auditor needs to evaluate whether the expectation is precise enough to identify a misstatement at the relevant level of materiality.
Worked example: Lefèvre Holdings S.A.
Client: Belgian holding company, FY2024, consolidated revenue €120M across four subsidiaries, IFRS reporter.
Assertion: Completeness of payroll expense (€14.2M group-wide)
The engagement team applied a substantive analytical procedure. The expectation used headcount per subsidiary (obtained from HR records), multiplied by the average salary per grade (confirmed to employment contracts on a sample basis), plus employer social contributions at the Belgian statutory rate of 27%.
Documentation note: "Expectation developed independently of the general ledger. Data sources: HR headcount report (confirmed to December payroll run), average salary by grade (agreed to 15 contracts across all four subsidiaries), statutory social contribution rate per Belgian social security legislation. Ref: ISA 520.5."
Expected payroll: €13.9M. Recorded payroll: €14.2M. Difference: €300K (2.1%). The team set the threshold at performance materiality of €350K. The difference fell within the threshold. No further investigation required.
Assertion: Existence of a €2.8M intercompany receivable from subsidiary Lefèvre Industrie N.V.
Analytical procedures would not work here. The balance is a single item with no predictable pattern and no external data to benchmark against. The team performed a test of details: confirmed the balance directly with Lefèvre Industrie N.V., agreed the underlying invoices to delivery records, verified elimination entries in the consolidation, and inspected the cash settlement in January 2025.
Documentation note: "Substantive test of details over intercompany receivable existence. Direct confirmation received and agreed to subsidiary records. Underlying invoices agreed to signed delivery notes. Elimination entries traced through consolidation workpaper. Ref: ISA 330.18, ISA 505.7."
If the team had applied an analytical procedure to the intercompany balance, the file would contain no direct evidence that the receivable exists. A reviewer would flag the assertion as unsupported.
What reviewers get wrong
The most common error is applying a substantive analytical procedure without documenting the independently developed expectation and the threshold for investigation. ISA 520.5(b) and (c) require both. A file that shows "compared current year to prior year, no significant fluctuation noted" does not meet the standard. The expectation must be independent of the recorded amount, and the team must define what constitutes a significant difference before performing the comparison.
Teams sometimes apply substantive analytical procedures to assertions involving significant risks. ISA 330.21 requires substantive procedures that are specifically responsive to the assessed risk. For most significant risks (revenue fraud, management override), this means tests of details, not analytics. Analytical procedures alone rarely provide sufficient precision for a significant risk assertion.
Key standard references
- ISA 330.18: Requires substantive procedures for every material class of transactions, balance, and disclosure.
- ISA 330.21: Requires substantive procedures that are specifically responsive to significant risks.
- ISA 520.5: Sets requirements for substantive analytical procedures, including suitability, independent expectation, and precision.
- ISA 520.A5: Requires the auditor to evaluate whether the expectation is precise enough to identify material misstatement.
Related terms
Related tools
Related reading
Frequently asked questions
Can an auditor use only analytical procedures for a significant risk?
Rarely. ISA 330.21 requires substantive procedures that are specifically responsive to the assessed risk. For most significant risks, such as revenue fraud or management override, tests of details are necessary because analytical procedures alone rarely provide sufficient precision. The auditor must evaluate whether the analytical procedure can identify misstatement at the required level of materiality for that specific risk.
What makes a substantive analytical procedure fail under inspection?
The most common failure is not documenting the independently developed expectation and the threshold for investigation. ISA 520.5(b) and (c) require both. A file that shows only a comparison of current year to prior year without an independent expectation does not meet the standard. The expectation must be built from data that is independent of the recorded amount.