Refund liabilities and rights of return

Money-back guarantees, cooling-off rights, and cancel-for-refund clauses mean the entity is not entitled to keep all consideration invoiced, so recognizing full revenue and separately expensing refunds when paid overstates revenue during the guarantee window and creates a misleading spike-then-correction pattern that investors can mistake for a genuine cancellation problem rather than a modeling one; a properly estimated refund liability instead smooths recognition to reflect only the consideration the entity actually expects to keep.

Do subscription plans include money-back guarantees, statutory cooling-off rights, or cancel-for-any-reason clauses within a defined window? What does historical cohort data show about the actual refund rate within the guarantee window, and is this used to estimate the refund liability? Is revenue for the guarantee-window cohort recognized net of the expected refund, rather than recognized in full and expensed only when refunds are actually paid?

1. Identify contract terms that give the customer a right to a full or partial refund (money-back guarantees, statutory cooling-off periods, cancel-for-any-reason clauses within a defined window).
2. Treat the refund right as variable consideration: do not recognize revenue for the portion of consideration expected to be refunded.
3. Estimate the expected refund amount using the expected value method (a portfolio of contracts with a range of outcomes, appropriate given typically large volumes of similar subscriptions) based on the entity's own historical refund/cancellation experience by cohort.
4. Apply the constraint on variable consideration, including in the transaction price only the amount highly probable not to result in a significant future revenue reversal, recognize revenue net of the constrained refund estimate from inception, not gross with refunds expensed later.
5. Recognize a refund liability (not a reduction of revenue recognized as a contra-asset against receivables only) for the expected refund amount, separate from any related asset for the right to recover product or service from customers if applicable.
6. Update the refund liability estimate at each reporting date using the latest cohort experience, recognizing the adjustment in current-period revenue.

Measure the refund liability at the amount of consideration received or receivable that the entity does not expect to be entitled to keep, using the expected value method applied to the relevant cohort of contracts (e.g., all subscriptions sold within a rolling 30-day guarantee window), refreshed with actual refund experience as each cohort's guarantee window closes. Measure revenue as the transaction price net of the refund liability, i.e., only the amount expected to be retained. Reassess the historical refund rate periodically (e.g., quarterly) for changes in product mix, customer segment, or marketing practices that could shift the refund rate, rather than relying on a single stale rate indefinitely.

Disclose: the refund liability balance at period end, presented separately from contract liabilities if material, and its movement during the period; the method and key assumptions used to estimate the expected refund rate, including the cohort/portfolio basis applied; the nature of refund rights and guarantee windows offered across the entity's main product lines; the effect of any material changes in the refund rate estimate on revenue during the period; the policy for discretionary (non-contractual) goodwill refunds and whether they are factored into the estimate.

• Recognizing the full invoiced amount as revenue at the outset and expensing actual refunds as a marketing or contra-revenue cost when paid, instead of constraining revenue and recognizing a refund liability from day one
• ignoring statutory cooling-off or consumer protection rights that create an effective refund obligation even without an explicit contractual guarantee
• using a single company-wide refund rate that blends materially different products or customer segments with very different actual refund behavior
• failing to true up the refund liability estimate as a cohort'
• s guarantee window closes and actual experience becomes known
• excluding recurring discretionary goodwill refunds from the estimate even though a predictable pattern indicates they should be treated as de facto variable consideration.