How to Audit Revenue Recognition Under IFRS 15

Why revenue recognition is a presumed fraud risk

ISA 240.26 presumes that revenue recognition carries a risk of material misstatement due to fraud. You can rebut the presumption, but ISA 240 .A30 sets a high bar. The engagement team has to document why the presumption doesn’t apply to the specific revenue streams of this client. In our experience, most teams accept the presumption and design their procedures around it. The question isn’t whether revenue is a significant risk. It’s which assertions within revenue carry the highest fraud risk for this particular client.

That distinction matters because it drives your response under ISA 330.18 . A significant risk requires a substantive response specifically designed for it. Pulling a random sample of sales invoices and agreeing them to goods dispatched notes doesn’t qualify as a specifically designed response. ISA 330 .A52 makes this explicit. The auditor needs procedures that address the particular way in which the significant risk could result in material misstatement. A blanket vouching test designed before the team even discussed fraud risks is the opposite of what ISA 330.18 requires. If your engagement uses a standardised revenue programme copied from the prior year without modification for the current year’s fraud risk assessment (what most of us call SALY with better narratives, where last year’s approach is dressed up for the current file), you don’t have a specifically designed response.

For most mid-tier engagements, the highest-risk assertions sit in two places. Cut-off (revenue recorded in the correct period) and occurrence (the transaction happened as recorded). Where contracts contain multiple deliverables, accuracy and allocation also become high-risk because IFRS 15.73 –86 requires the client to allocate the transaction price to each performance obligation based on relative standalone selling prices. A client that bundles installation with product delivery and recognises the full amount on shipment has an accuracy problem even if every invoice is genuine. And if the client has variable consideration (volume discounts, rebates, performance penalties), the estimate of the transaction price under IFRS 15.56 creates a measurement risk that standard invoice testing misses entirely.

The ciferi ISA 530 sampling calculator helps you size samples specifically for revenue populations where you’ve identified a fraud risk, adjusting for expected misstatement rates higher than zero.

The five-step model and what it means for your audit procedures

IFRS 15 replaced IAS 18 with a single framework that applies to every contract with a customer. The five steps aren’t just an accounting exercise for the client. They’re your audit roadmap. Each step creates a distinct assertion that needs a distinct procedure.

Step 1: Identify the contract ( IFRS 15.9 –16)

Your procedure here is confirming that a contract exists with commercial substance and that collectability of the consideration is probable. For most standard sales, this is straightforward. It gets complicated with framework agreements, call-off orders, contracts with variable terms, and multi-year arrangements. If the client has a master agreement with annual pricing reviews and separate purchase orders, you need to determine whether each purchase order is a contract or whether the master agreement is the contract. The answer changes when and how much revenue gets recognised.

Step 2: Identify performance obligations ( IFRS 15.22 –30)

Most audit deficiencies originate at this step. IFRS 15.22 requires the client to identify every distinct good or service promised in the contract. A good or service is distinct under IFRS 15.27 if the customer can benefit from it on its own (or with readily available resources) and the promise is separately identifiable from other promises in the contract. Installation services bundled with equipment delivery might be distinct or might not, depending on whether the installation is so intertwined with the equipment that the customer couldn’t use the equipment without it.

Step 3: Determine the transaction price ( IFRS 15.47 –72)

Variable consideration (volume discounts, rebates, penalties, performance bonuses) is the main audit risk here. IFRS 15.56 constrains the amount of variable consideration you can include in the transaction price, only to the extent that it’s highly probable that a significant reversal won’t occur. Test the client’s constraint assessment by examining historical reversal rates and management’s assumptions about the likelihood of reversal against current contract terms.

Step 4: Allocate the transaction price ( IFRS 15.73 –86)

If the contract has more than one performance obligation, the client allocates the transaction price based on relative standalone selling prices. Where the client doesn’t sell an item separately, IFRS 15.79 requires an estimation approach (adjusted market assessment, expected cost plus margin, or residual). Your procedure tests whether the client’s allocation method is appropriate and whether the inputs are reasonable.

Step 5: Recognise revenue when (or as) the entity satisfies a performance obligation ( IFRS 15.31 –45)

The distinction between over-time and point-in-time recognition is the final critical assertion. IFRS 15.35 sets out the conditions for over-time recognition. If none of the conditions is met, revenue is recognised at a point in time when control transfers. Your procedure verifies that the client applied the correct method and that the indicators of control transfer ( IFRS 15.38 ) align with the timing of recognition in the general ledger.

How to identify performance obligations the client missed

Reading the contract is the basic step. Reading it with IFRS 15.22 open next to you is what separates an adequate file from a good one. Most performance obligation testing in mid-tier files is ticking and bashing: agree invoice to delivery note, tick, move on. That catches existence but not bundled obligations. Start by listing every promise the client makes in the contract, including implied promises under IFRS 15.24 (customary business practices that create a valid expectation). Then test each promise against the two-part test in IFRS 15.27 .

A practical approach for mid-tier engagements: request the client’s five largest contracts by revenue value and two contracts from each material revenue stream. For each contract, create a two-column table. Left column: every promised good or service (explicit and implied). Right column: the client’s treatment (bundled or separated, with the performance obligation they’ve assigned it to). Your job is checking whether the right column matches what IFRS 15.27 requires.

The most common gap isn’t exotic. It’s installation or warranty services that the client bundles with the delivered product because they’ve always bundled them. Under IAS 18 , that often didn’t matter because the revenue recognition criteria were less prescriptive. Under IFRS 15 , each distinct obligation needs its own allocation and its own recognition point.

If you find a performance obligation that the client hasn’t separated, you need to quantify the misstatement. The ciferi IFRS 15 revenue recognition calculator walks you through the allocation calculation and produces a working paper you can file directly.

Designing substantive tests that go beyond vouching invoices

Vouching a sample of invoices to delivery notes tests occurrence and, to some extent, cut-off. It doesn’t test allocation, accuracy of variable consideration estimates, or whether over-time recognition calculations are correct. For a presumed fraud risk, you need procedures that target the specific way revenue could be materially misstated. Nobody enjoys challenging management on revenue cut-off, but skipping that conversation is how the file gets flagged.

Here’s what a substantive programme for revenue should cover at minimum on an IFRS 15 engagement.

For cut-off: select a sample of transactions around the period end (the last two weeks before and first two weeks after year-end). For each, trace the revenue entry to the point at which control transferred under IFRS 15.38 . Compare that date to the period in which the client recorded the revenue. If you’re using the ciferi ISA 320 materiality calculator, set your tolerable misstatement for the cut-off population based on performance materiality (PM), not overall materiality.

The most effective cut-off procedure for product sales is matching the invoice date to the shipping or delivery documentation. But IFRS 15 changed the benchmark. The question isn’t “when was it shipped?” It’s “when did control transfer?” IFRS 15.38 lists five indicators of control transfer: the entity has a present right to payment, the customer has legal title, the entity has transferred physical possession, the customer has the significant risks and rewards of ownership, and the customer has accepted the asset. For most standard product sales, physical delivery is the transfer point. But for bill-and-hold arrangements, consignment sales, delivered goods subject to customer acceptance, or goods shipped with right-of-return clauses, the transfer point may differ from the delivery date by weeks or even months. Test the control transfer point, not the logistics date.

For occurrence: confirm a sample of large or unusual transactions directly with the counterparty under ISA 505 . Select transactions that fall outside normal trading patterns (large one-off sales, sales to new customers, sales near the period end, or sales with unusual terms). Positive confirmation works better than negative for fraud-risk populations because the absence of a response from a negative confirmation tells you nothing. If the counterparty doesn’t respond to a positive confirmation, ISA 505.12 requires you to perform alternative procedures. For revenue, that means examining subsequent cash receipts and proof of delivery signed by the customer, together with contract documentation that corroborates the sale independently of the client’s own records.

For accuracy and allocation: for any contract with multiple performance obligations, recalculate the client’s allocation. Obtain standalone selling prices (or the estimation approach the client used) and verify the mathematical allocation. This isn’t a sampling exercise. You need to test every contract where the allocation is material.

For variable consideration: examine the client’s constraint assessment under IFRS 15.56 –58. Pull historical data on actual versus estimated variable consideration for prior periods. If the client consistently overestimates or underestimates, you have a bias indicator that needs to go into the ISA 240 fraud risk discussion.

What to do when the client has over-time recognition

Over-time revenue under IFRS 15.35 introduces a measurement risk that point-in-time sales don’t have. The client recognises revenue progressively based on a measure of progress (input method or output method, per IFRS 15.41 –45), and the audit risk shifts from existence and cut-off to accuracy. Your procedure tests whether the measure of progress is appropriate and whether the inputs to the calculation produce a reasonable estimate at year-end.

For input-method contracts (typically costs incurred to date as a proportion of total estimated costs), the two critical inputs are costs incurred and total estimated costs. Costs incurred can be tested against invoices and timesheets. Total estimated costs are an estimate, and that estimate sits squarely within ISA 540 . Test the estimate by comparing prior-period estimated costs to actual costs at completion. If the client consistently underestimates total costs (resulting in early revenue acceleration), you’ve identified a pattern that requires disclosure in the ISA 540 retrospective review and may indicate management bias.

For output-method contracts (milestones, units delivered, surveys of work performed), verify the milestone or output measurement independently. If the client claims 75% completion based on an internal project manager’s assessment, obtain supporting evidence: inspection reports, customer sign-offs on milestones, independent quantity surveys, or third-party assessments. The client’s own project tracking system is not independent evidence of the output measure.

Worked example: Brouwer Machining B.V.

Client scenario

Brouwer Machining B.V. is a Dutch precision engineering company. Revenue for 2024: €48M. Primary revenue stream: custom industrial components with installation and a two-year maintenance contract. Brouwer sells the components, installs them on-site, and provides preventive maintenance for 24 months. The client records the full contract value as revenue on delivery of the component.

Conclusion. The file now shows a quantified misstatement of €1.1M arising from failure to separate performance obligations and allocate the transaction price. The reviewer sees the contract analysis and the allocation calculation with verifiable standalone selling prices, together with a misstatement schedule cross-referenced to ISA 450 . This is defensible because every number traces to a source document.

Your engagement checklist

Common mistakes regulators flag

• The AFM has flagged insufficient documentation of the fraud risk discussion for revenue in multiple inspection cycles. ISA 240 .A30 requires a specific rationale if the team rebuts the presumption. A one-line note saying “management is trustworthy” doesn’t meet the standard.

Related content

• Revenue recognition. The ciferi glossary entry covers the IFRS 15 five-step model with a quick-reference summary of each step and the key paragraph references.
• IFRS 15 revenue recognition calculator. Allocates transaction prices across performance obligations and produces a working paper schedule you can file directly.
• How to audit estimates and fair value measurements. Revenue estimates (variable consideration, percentage of completion) follow the same ISA 540 framework covered in this post.

Word count target: 3,000–4,500 words.

Related ciferi content

Related guides:

Put audit concepts into practice with these free tools:

Frequently asked questions