How auditors build expectations, investigate anomalies, and use analytics as audit evidence.
Analytical procedures evaluate plausible relationships among financial and nonfinancial data. Data analytics expands that work by letting auditors examine larger data sets, identify unusual patterns, and focus testing on items that carry higher risk. Neither technique is a shortcut around audit evidence. The auditor still validates data, develops an expectation, investigates differences, and documents the conclusion.
On the AUD exam, analytical procedures appear in planning, substantive testing, and final review. The key distinction is the objective. Planning analytics identify risk. Substantive analytical procedures can provide audit evidence when the expectation is precise enough. Final analytics help the auditor decide whether the financial statements still make sense before the report is issued.
flowchart LR
A["Define audit objective"] --> B["Develop expectation"]
B --> C["Validate data"]
C --> D["Compare recorded amount to expectation"]
D --> E{"Difference exceeds threshold?"}
E -- "No" --> F["Document conclusion"]
E -- "Yes" --> G["Investigate and corroborate explanation"]
G --> F
The same technique can have different evidential value depending on when and why it is used.
| Stage | Purpose | Common result |
|---|---|---|
| Planning | Identify unusual relationships and areas of possible risk | Audit plan is focused on higher-risk assertions |
| Substantive testing | Obtain direct evidence about recorded amounts | Difference from expectation is investigated and evaluated |
| Final review | Assess whether financial statements are consistent with audit understanding | Unresolved unusual relationships are reconsidered before reporting |
Planning analytics are required in many audits, but they usually do not provide enough evidence by themselves. Substantive analytical procedures require stronger precision, reliable data, and a clear threshold for investigation.
An analytical procedure is only as strong as the expectation behind it. A vague comparison such as “sales increased from last year” may identify a risk, but it is usually not precise enough to detect material misstatement. A stronger expectation is based on independent or reliable data and can predict a recorded amount within a narrow range.
| Analytical method | Example | Best use |
|---|---|---|
| Trend analysis | Compare monthly sales to prior periods and known seasonal patterns | Planning and final review |
| Ratio analysis | Compare gross margin, inventory turnover, or days sales outstanding | Risk identification and follow-up |
| Reasonableness test | Estimate interest expense using average debt and contractual rates | Substantive evidence when precise |
| Regression or model-based analytics | Predict utility expense using square footage, production volume, and rates | Routine high-volume accounts |
| Full-population scan | Identify weekend journal entries, duplicate payments, or unusual user activity | Targeting exceptions for further testing |
The auditor should define the expected relationship before looking only for a favorable result. A procedure designed after seeing the client number is less persuasive.
Precision is the closeness of the auditor’s expectation to the recorded amount. Higher precision makes the procedure more useful as substantive evidence.
Precision improves when:
Precision weakens when the account is volatile, judgmental, affected by one-time transactions, or based on unreliable client-prepared data. For example, a reasonableness test for fixed-rate interest expense can be precise. A broad comparison of legal expense to prior year is usually less precise because legal matters can change sharply.
Data analytics can improve coverage, but the auditor must first validate the data. A dashboard built from incomplete or incorrectly mapped data can produce persuasive-looking but unreliable results.
Before relying on analytics, the auditor considers:
| Data issue | Audit concern |
|---|---|
| Completeness | Are all relevant transactions, entities, locations, and periods included? |
| Accuracy | Do extracted fields agree to source systems or control totals? |
| Relevance | Does the data match the assertion being tested? |
| Transformation logic | Were joins, filters, exclusions, and calculations appropriate? |
| Access and change controls | Could the data or report logic have been altered without detection? |
| Repeatability | Can another auditor understand and rerun the analytic? |
Whole-population analytics are useful for identifying exceptions, but exception identification is not the same as misstatement detection. The auditor must investigate exceptions and determine whether they represent valid transactions, data errors, control failures, or possible misstatements.
When the recorded amount differs from the auditor’s expectation beyond the investigation threshold, the auditor obtains corroborating evidence. Management’s explanation is a starting point, not the conclusion.
| Explanation from management | Corroborating evidence to seek |
|---|---|
| Sales increased because of a new contract | Inspect contract, shipping records, invoices, and subsequent cash receipts |
| Gross margin fell because input costs rose | Inspect vendor invoices, inventory cost records, and production reports |
| Interest expense increased due to new debt | Inspect loan agreements, bank confirmations, and amortization schedules |
| Payroll rose because headcount increased | Inspect HR records, payroll registers, approvals, and tax filings |
| Journal entries are unusual because of year-end close | Inspect support, approval evidence, user access, and posting rationale |
If the explanation is unsupported or inconsistent with other evidence, the auditor revises the risk assessment and performs additional procedures.
Do not choose an answer that says analytics eliminates tests of details, controls testing, or professional judgment. Analytics supports risk assessment and evidence gathering, but the auditor still evaluates reliability and follows up.
Do not confuse a planning analytic with a substantive analytical procedure. Planning analytics may identify risk with broad comparisons. Substantive analytics require a precise expectation and reliable data.
Do not accept management’s explanation for a variance without corroboration. The auditor audits the explanation.
Do not assume a full-population analytic is automatically sufficient. A population scan may identify all items matching a rule, but the rule, data, and interpretation still need audit support.