Tests of details and analytical procedures are the two categories of substantive audit work, and they differ in what they examine and what kind of evidence they produce. A test of details verifies individual transactions or balances against source documents, producing direct proof about specific items. An analytical procedure compares a recorded amount to an independently developed expectation, producing indirect evidence that the balance as a whole is plausible. Auditors use both, and the assessed risk of material misstatement for each account determines the mix.
What a Test of Details Actually Does
A test of details works by pulling specific items out of an account and checking them against the underlying documentation. The auditor picks a recorded revenue entry and traces it back to the signed contract, the shipping record, and the customer invoice. Or picks an inventory line and physically counts it. Either the document supports the recorded amount or it doesn’t. That directness is what makes tests of details the strongest form of substantive evidence, and also the most expensive to perform.
Vouching and Tracing
Vouching starts with a number already recorded in the ledger and works backward to the source documents that should support it. If the auditor selects a $50,000 expense entry, vouching means pulling the vendor invoice, purchase order, and receiving report to confirm the transaction happened and was recorded at the right amount. Vouching primarily addresses whether recorded transactions are real.
Tracing runs in the opposite direction. The auditor starts with a source document and follows it forward into the accounting records to confirm it was captured. Selecting a batch of receiving reports and checking whether each one appears in inventory and accounts payable tests whether everything that should have been recorded actually was. Tracing targets completeness rather than existence.
The direction of the test determines the kind of error it catches. Vouching catches fictitious or overstated entries. Tracing catches omissions. An audit that only vouches from the ledger could miss transactions that were never recorded in the first place.
External Confirmations
Confirmations involve contacting a third party directly to verify information the client has recorded. The most common example is confirming accounts receivable by sending a letter to the client’s customer asking them to verify the balance owed. Because this evidence comes from an independent external source, it ranks among the most reliable forms of audit evidence available.1PCAOB. AS 1105: Audit Evidence
Auditing standards effectively require confirmations for accounts receivable and cash held by third parties. The auditor must either confirm those balances or obtain equivalent evidence by directly accessing information maintained by an external source. Positive confirmations ask the customer to respond regardless of whether they agree with the stated balance. Negative confirmations ask for a response only if the customer disagrees, and they produce significantly weaker evidence because silence doesn’t prove agreement. On their own, negative confirmations aren’t enough to address the risk of material misstatement.2PCAOB. AS 2310: The Auditor’s Use of Confirmation
Sampling and Its Limits
Auditors almost never test every transaction in an account. They select a sample designed to let them draw conclusions about the whole population. The auditor first identifies items that individually could contain a misstatement large enough to matter, and those items get examined outside the sample. Everything else goes into the sampling population, sometimes stratified by size or type to improve efficiency.3PCAOB. AS 2315: Audit Sampling
This produces an important limitation. A test of details only directly covers the items examined. Finding zero errors in a sample of 60 invoices out of 10,000 supports a conclusion about the population, but it doesn’t guarantee every invoice is correct. Sampling reduces the risk of a wrong conclusion to an acceptable level; it doesn’t eliminate it.
What an Analytical Procedure Actually Does
An analytical procedure takes a completely different approach. Instead of examining individual transactions, the auditor builds an independent estimate of what an account balance should be and compares that estimate to what the client recorded. If the two numbers are close, the balance is probably reasonable. If they diverge significantly, something may be wrong.
The power of an analytical procedure depends almost entirely on how precise the expectation is. A vague sense that “revenue looks about right” provides almost no assurance. A calculation that derives expected revenue from units shipped, average selling prices by product line, and contractual price changes can be remarkably precise.4PCAOB. AS 2305: Substantive Analytical Procedures
Common Approaches
The simplest form is trend analysis: comparing this year’s balance to last year’s and investigating anything that moved more than expected. If office supply expense doubled with no change in headcount or office space, that warrants a closer look. Trend analysis is quick but imprecise, which limits how much assurance it provides on its own.
Ratio analysis adds context by comparing financial metrics to industry benchmarks or to the company’s own historical patterns. A gross margin that suddenly jumps five percentage points above the industry average, with no corresponding change in business strategy, could signal that costs are understated or revenue is overstated.
Reasonableness testing is the most powerful form because it builds an expectation from operational data rather than financial history. The auditor might estimate depreciation expense by applying known rates to the recorded asset balances, or estimate payroll by multiplying headcount by average compensation. These calculations create a mathematically independent benchmark that doesn’t rely on the client’s accounting system.
What Makes an Expectation Precise Enough
Standards require the expectation to be precise enough that a material misstatement would actually show up as a noticeable difference. Several factors drive that precision. Relationships involving income statement accounts tend to be more predictable than balance sheet relationships, because income statement accounts reflect transactions over a period rather than a snapshot at a single moment. Expectations built from disaggregated data catch more problems; monthly figures by business segment will flag issues that annual company-wide totals would obscure. A stable operating environment produces more predictable relationships than a volatile one.4PCAOB. AS 2305: Substantive Analytical Procedures
The data feeding the expectation also needs to be reliable. Data from sources outside the company is generally more trustworthy than data from inside it, and data produced under effective internal controls is more reliable than data from weak systems. Before relying on a substantive analytical procedure, the auditor must either test the controls over the data used in the calculation or perform other procedures to confirm the data is accurate and complete.4PCAOB. AS 2305: Substantive Analytical Procedures
Investigating Differences
When the recorded balance deviates from the expectation by more than a predetermined threshold, the auditor investigates. That threshold is anchored to materiality and should be set before performing the procedure, not adjusted after seeing the results.
The investigation typically begins with asking management to explain the difference, and this is where auditors routinely get into trouble. Management’s explanation alone isn’t enough. The auditor must corroborate whatever management says with independent evidence.4PCAOB. AS 2305: Substantive Analytical Procedures If management attributes a spike in repair expense to a major equipment overhaul, the auditor needs to see the invoices and work orders. When the difference remains unexplained, the auditor expands tests of details for that account.
How the Evidence Compares in Quality
Auditing standards establish a hierarchy for audit evidence: evidence from independent external sources outranks internal evidence, evidence the auditor obtains directly outranks evidence obtained indirectly, and original documents outrank copies.1PCAOB. AS 1105: Audit Evidence
Tests of details score well on this hierarchy. When an auditor physically inspects a signed contract or receives a bank confirmation, the evidence is both external and direct. The connection between the evidence and the assertion is immediate: the document either supports the recorded amount or it doesn’t.
Analytical procedures produce indirect evidence. The auditor isn’t proving a specific transaction is correctly recorded; the auditor is demonstrating that the overall balance falls within a range consistent with independent data. That’s still useful evidence, and a well-designed analytical procedure can be highly persuasive. But it works through inference rather than direct verification, and it can miss misstatements that offset each other within the account.
Neither procedure can rest on inquiry alone. Inquiry never constitutes sufficient audit evidence for a relevant assertion, whether it arises during a test of details or during the investigation of an analytical variance.5PCAOB. AS 2301: The Auditor’s Responses to the Risks of Material Misstatement
Choosing Between Them
Every significant account requires some substantive testing regardless of how strong the controls are. The question is what form that testing takes.5PCAOB. AS 2301: The Auditor’s Responses to the Risks of Material Misstatement
As the assessed risk goes up, the auditor needs more persuasive evidence, which pushes the plan toward tests of details. More evidence can come from a larger sample, from more direct procedures like confirmations, or both. The nature of the potential misstatement matters as well. If the risk involves management override of controls or complex judgmental estimates, the procedures need to be designed to address those specific concerns.
When risk is low and the balance is predictable, analytical procedures alone may provide sufficient evidence. Interest expense on fixed-rate debt is a classic example: the auditor can independently calculate what the expense should be from the loan balance and the interest rate. The result is precise, and the data is easy to verify. Running tests of details on every interest payment would burn hours without adding meaningful assurance.
The Significant Risk Floor
Standards draw a hard line in one place. For significant risks of material misstatement, analytical procedures alone are unlikely to provide sufficient evidence.4PCAOB. AS 2305: Substantive Analytical Procedures Significant risks involve complex transactions, areas of high estimation uncertainty, or elevated fraud risk. Revenue recognition is a common example. In these areas, the audit plan needs to include tests of details, even if the analytical procedures look clean.
The reasoning is straightforward. Analytical procedures work by assuming that predictable relationships will hold. Significant risks exist precisely because the relationships are unpredictable, or because someone with the means and motive to manipulate the numbers could make a misstatement look normal on the surface.
Using Both Together
In practice, most audit areas use both procedures together. Analytical procedures performed early in fieldwork help the auditor identify which accounts need heavy testing and which look routine. A margin analysis across product lines might reveal one segment with an unexplained profitability spike. That finding redirects the team’s tests of details toward the accounts driving the anomaly rather than spreading effort evenly across the ledger.
The reverse also happens. Results from tests of details can inform whether an analytical procedure’s conclusion holds up. If a reasonableness test on depreciation expense looks clean but tests of details on asset additions reveal several items capitalized at incorrect amounts, the analytical procedure’s conclusion needs to be revisited. The two approaches act as cross-checks.
Where Auditors Get This Wrong
The most common deficiency PCAOB inspectors find with analytical procedures is insufficient precision. The auditor runs a year-over-year comparison, sees that revenue grew 8%, decides it “looks reasonable,” and moves on. That isn’t a substantive analytical procedure. Without a specific, independently developed expectation and a predetermined threshold for investigation, the comparison provides almost no assurance.
The second recurring problem is failing to corroborate management explanations. When the analytical procedure flags a variance and management offers a plausible story, some auditors accept the explanation without obtaining supporting evidence. That collapses the entire procedure into a management inquiry, which standards explicitly say is not sufficient on its own.
On the tests of details side, auditors sometimes make the opposite mistake: testing extensively but not linking the work to assessed risk. Running a large sample of vouching procedures on a low-risk account while under-testing a high-risk one is a misallocation that no amount of extra documentation will fix. As assessed risk increases, the evidence obtained from substantive procedures must increase to match.5PCAOB. AS 2301: The Auditor’s Responses to the Risks of Material Misstatement
The best audit plans treat tests of details and analytical procedures as tools calibrated to different situations, not interchangeable alternatives. When the relationship is predictable and the data is reliable, a well-built analytical procedure can provide strong evidence efficiently. When the account involves judgment, complexity, or elevated risk, there is no substitute for examining the documents.