How to Write a Return & Refund Policy (2026): The Complete Guide

Most small businesses treat return and refund policy as a bureaucratic afterthought — copy-paste a template, stick it at the footer, move on. That works right up until the first chargeback lands, the first state AG inquiry arrives, or the first EU customer invokes their statutory withdrawal right. At that point, your return policy is the single most-scrutinized document on your website.

A return and refund policy is not the same thing as your terms of service, even though the two overlap. Terms of service are contractual — they bind users to behavioral rules, liability caps, arbitration, intellectual property, and governing law. A return policy is a consumer-facing commitment that your payment processor, your state regulator, and your customer will each hold you to independently of your terms.

Here's why a specific, well-drafted return policy matters, in order of how often each reason actually costs money:

• Chargebacks. When a customer files a chargeback, Visa and Mastercard ask for your disclosed policy as part of the representment package. A vague policy loses. A specific one wins. The difference in outcome is cash.
• State disclosure requirements. California, Florida, New York, Massachusetts, Minnesota, and others require conspicuous disclosure of your return policy as a condition of enforcing any restriction on returns. Miss the disclosure and your customers get the default window (30 days in CA/NY, 7 in FL, indefinite in MA).
• EU and UK statutory rights. Customers have 14-day cooling-off rights that you cannot waive for distance sales, plus separate rights for faulty goods. Failing to inform them of these rights extends the withdrawal window by 12 months.
• Australian Consumer Law. The ACL's consumer guarantees are non-excludable. Blanket 'no refund' signs are explicitly unlawful and trigger ACCC action.
• Class action exposure. UDAP statutes in every US state give consumers a private right of action for misleading or missing return-policy disclosures. Massachusetts Chapter 93A is particularly aggressive.
• Customer lifetime value. Zappos built an empire on its 365-day return policy. Your return terms are a marketing input, not just a cost control.

The test for whether your current return policy is adequate is not whether it's polite or readable. It's whether a regulator, a card network, or a plaintiff's lawyer could point to it as disclosing the specific thing you're being challenged about. If your policy says 'returns accepted within 30 days' and a customer disputes a subscription cancellation, your policy has said nothing about the actual issue — and that's the compliance gap that costs money.

A return policy that works for a physical e-commerce store will sink a SaaS company on its first chargeback. A SaaS policy will confuse the daylights out of a mattress buyer. Before you write anything, pick the model that matches what you actually sell, and recognize that 'one-size-fits-all' policy language is the single most common failure mode.

Physical e-commerce. The canonical case. Default to a 30-day window (matching California and New York's statutory defaults). Offer full refund, exchange, and store credit. Require unused condition and original packaging for change-of-mind returns. Customer pays return shipping unless the item is defective. Include a defective-item process that doesn't deduct shipping or apply restocking fees. Disclose the return address clearly. Restocking fees (if any) should be 10-15% maximum and carve out statutory returns.

SaaS. 14-day window aligned with EU/UK statutory minimums. Refund methods limited to prorated refund or cancel-future-billing — you can't 'return' SaaS in any meaningful sense. Exclusions: digital content already accessed, custom implementations, services already rendered. The critical piece is the subscription-cancellation section — ROSCA and California's auto-renewal law require you to offer cancellation as easy as signup, via the same medium (online signup → online cancellation). Chargebacks typically come in as Visa 13.2 (cancelled recurring) or 13.3 (not as described); your policy needs to address both.

Digital products. 14-day window with an express-consent waiver for immediate access. The EU CRD's Article 16 only lets you exclude digital content from withdrawal rights if the consumer gives express prior consent and acknowledges losing the withdrawal right — a pre-ticked checkbox or a generic terms-of-service acceptance does not count. You need a specific checkbox at checkout, you need to document the consent in the order confirmation, and your policy needs to reference both.

Subscription services (non-SaaS). Physical subscription boxes, curated services, digital media subscriptions. 14-day window for the subscription-formation right. Cancel-future-billing is the primary remedy, with no refund for the current cycle if it's already been fulfilled. Subscription-specific exclusions: already-shipped items (current cycle non-refundable), perishable items (baseline non-returnable), already-rendered services. California auto-renewal compliance is the de facto US standard here; EU 14-day withdrawal still applies to the subscription contract.

Marketplaces. The complex case. Each seller has their own policy, and the platform has its own supplementary policy, and the customer sees both. Platform-level policy should set minimum requirements (e.g., sellers must accept returns for defective items, minimum 14-day windows). Customer pays return shipping unless defective. Platform dispute-escalation process serves as a buffer before chargebacks.

Professional services. 14-day window from contract conclusion. Prorated refund for work not yet performed; credit notes for client-favor scenarios; cancel-no-refund where work is substantially complete. Exclusions: services already rendered, custom work product. For EU clients, you need express consent to immediate performance in order to bill for services performed during the 14-day cooling-off period, and you must still refund any unperformed portion.

Courses and info products. The chargeback magnet. Two approaches work: (1) full 14-day refund without strings (simple, customer-friendly, high refund rate), or (2) percent-accessed threshold paired with express consent to immediate access (more defensible against friendly fraud but requires a system that actually tracks access). Either way, your policy must be extremely specific about what counts as 'accessed' and what the threshold is, because this is the number one chargeback category for info products.

Mobile apps. You don't control refunds — Apple and Google do. Your policy explicitly states that app-store refunds are processed through the App Store or Google Play per their respective policies, and that you cannot process refunds for app-store transactions directly. This is both accurate and the correct representment document: if a customer chargebacks an IAP, your evidence is that the transaction was processed by the platform, not by you.

If you sell to EU consumers — even incidentally — the Consumer Rights Directive (2011/83/EU), as amended by the Omnibus Directive (2019/2161), applies. It doesn't matter where your business is based. It matters where your customer is. And the 14-day right of withdrawal is not a marketing concept or a competitive choice — it's a statutory right you cannot waive by contract.

The mechanics are straightforward but precise:

• For goods, the 14-day period runs from the day the consumer (or a designated non-carrier third party) acquires physical possession.
• For orders of multiple goods delivered separately, the period runs from the day the last item is received.
• For services and digital content not on a tangible medium, the period runs from the day the contract was concluded.
• Customers do not need to give a reason. The model withdrawal form in Annex I is sufficient but not required — any unequivocal statement is enough.
• You must refund within 14 days of being notified of the withdrawal, including the basic standard-delivery cost (but not supplements for expedited delivery).
• You may withhold the refund until the goods are returned or the customer has provided evidence of return.
• The consumer pays return shipping unless you have not informed them of this cost or you have agreed to bear it.
• The consumer is liable for diminished value from handling beyond what's necessary to examine the item (the 'shop inspection' standard).

The critical exceptions, from Article 16 of the Directive:

• Goods made to the consumer's specifications or clearly personalized.
• Perishables and goods with short expiry dates.
• Sealed goods unsuitable for return for hygiene or health reasons, once unsealed.
• Sealed audio, video, or computer software, once unsealed.
• Goods inseparably mixed with other items after delivery.
• Digital content supplied where performance has begun with the consumer's prior express consent and acknowledgment of losing the withdrawal right.
• Services fully performed where performance began with prior express consent and acknowledgment of loss of withdrawal right upon full performance.

The digital-content exception is where most merchants get this wrong. A generic 'digital products are non-refundable' clause does not satisfy Article 16. You need: (1) express consent, (2) acknowledgment that the consumer is losing the withdrawal right, (3) confirmation on a durable medium. The practical implementation is a checkbox at checkout with specific language, and the order confirmation email must memorialize both the consent and the acknowledgment.

Now for the change coming on 19 June 2026: Directive (EU) 2023/2673 amends the Consumer Rights Directive to require online sellers to provide a mandatory cancellation function on their e-commerce sites. The language requirements are specific:

• Clearly labeled with wording such as 'withdraw from contract here' or equivalent unambiguous language.
• Easy to find, legible, and continuously available during the withdrawal period.
• Must allow the consumer to identify the specific contract being withdrawn (partial withdrawal where relevant).
• Must enable submission of a clear withdrawal statement online without unnecessary re-identification if the consumer is logged in.
• Must generate an acknowledgment of receipt on a durable medium including date and time.

If you sell to EU consumers, you have a deadline. You need this function live by 19 June 2026. It is not optional and it is not satisfied by a contact form or an email address — the Directive specifically requires a dedicated function that allows structured withdrawal submission. Our generator flags this requirement automatically when EU is selected as a jurisdiction; if you're building out policy and workflow for 2026, treat this as a Q2 deadline.

Post-Brexit, the UK retained EU-derived consumer protection law but with distinct UK statutes. Two laws matter for returns: the Consumer Contracts (Information, Cancellation and Additional Charges) Regulations 2013 (the 'CCR'), and the Consumer Rights Act 2015 (the 'CRA'). They stack, not substitute.

The CCR covers distance sales (online, phone, mail order) and off-premises sales. Under the CCR, UK consumers have:

• A 14-calendar-day cooling-off period starting the day after delivery (or contract conclusion for services/digital).
• A further 14 days from the date of notification to return the goods.
• The right to a refund within 14 days of you receiving the goods back, or within 14 days of the consumer providing evidence of return, whichever is earlier.
• Refund of the basic delivery cost (not premium delivery supplements).
• Responsibility for return shipping (unless you haven't informed them of the cost or agreed to pay).

The CCR exceptions mirror the EU CRD: personalized goods, perishables, sealed hygiene/software goods unsealed after delivery, digital content where performance has begun with express consent, and fully-performed services with express consent.

The CRA runs alongside the CCR and gives UK consumers separate statutory rights for goods that turn out to be faulty. The headline right is the 30-day short-term right to reject: if goods are not of satisfactory quality, not fit for purpose, or not as described, the consumer can reject them within 30 days of delivery and receive a full refund. This right is independent of the 14-day CCR cooling-off period — even if the CCR period has expired, the CRA 30-day right still applies to faulty goods.

After 30 days and within six months of delivery, the consumer has the right to request a repair or replacement; if that's unsuccessful, they can reject for a refund. After six months, the burden of proof shifts to the consumer to show the defect existed at delivery.

The practical compliance checklist for UK sales:

• 14-day cooling-off window for distance sales, disclosed before purchase.
• Model cancellation form provided (CCR Schedule 2(B)).
• 30-day right-to-reject for faulty goods referenced in policy.
• Refund within 14 days of receiving goods back or proof of shipment.
• Refund includes basic delivery cost.
• Digital content express-consent model for CCR exception.

The failure mode to avoid: blanket 'no refunds' language in UK-facing copy. The CRA is explicit that attempts to exclude the consumer's statutory rights are void. A policy that tries to disclaim the 30-day right-to-reject is not just ineffective — it's evidence of unfair commercial practice under the Consumer Protection from Unfair Trading Regulations 2008 and exposes you to Competition and Markets Authority action.

Australia's consumer protection regime is fundamentally different from the US or EU. The Australian Consumer Law (ACL) — set out in Schedule 2 of the Competition and Consumer Act 2010 and enforced by the ACCC — doesn't focus on a cooling-off period for change-of-mind returns. Instead, it imposes non-excludable 'consumer guarantees' that apply to every sale, and allocates remedies based on whether a failure is 'major' or 'minor'.

The consumer guarantees include, for goods: acceptable quality, fitness for disclosed purpose, matching description, correspondence to sample/demonstration model, clear title, undisturbed possession, no hidden securities, durability, and availability of spare parts and repair facilities. For services: due care and skill, fitness for purpose, and delivery within a reasonable time.

These guarantees cannot be excluded, modified, or restricted by contract. A 'no refunds' sign, a store policy, or a terms clause that purports to disclaim these guarantees is not just ineffective — under ACL section 64, it's explicitly void, and under ACCC enforcement guidance, false representations of the non-existence of consumer guarantees are themselves prohibited representations.

The remedy framework depends on the severity of the failure:

• Major failure: a failure serious enough that a reasonable consumer would not have bought the product had they known, or that makes the product significantly different from description, unfit for a disclosed purpose, or unsafe. For major failures, the consumer chooses between a refund, a replacement, or keeping the product and receiving compensation for the drop in value.
• Minor failure: a failure that can be fixed within a reasonable time. For minor failures, the business chooses between a repair, a replacement, or a refund. If a repair fails or the business refuses to remedy, or if two or more minor failures together amount to a major failure, the consumer then chooses.
• Services: major failures let the consumer cancel the contract for a refund (less a reasonable amount for services satisfactorily performed), or keep the contract and pay a reduced price.

Crucially, the consumer is not limited to the manufacturer — they can pursue the retailer directly, and the retailer cannot redirect them to the manufacturer. The retailer then has the right to claim reimbursement from the manufacturer.

The ACL also covers consequential loss: if the product or service failure causes the consumer reasonably foreseeable loss or damage beyond the cost of the product itself, the consumer can claim compensation for that too.

What this means for your return policy if you sell to Australian customers: you can still have a change-of-mind policy (the ACL doesn't require you to accept change-of-mind returns for any reason), but you cannot exclude or disclaim the consumer guarantees. Your policy needs to explicitly state that nothing limits the ACL consumer guarantees, that consumers may choose their remedy for major failures, and that the policy supplements but does not replace those guarantees. The generator produces this language automatically when AU is selected.

A common failure pattern: international merchants copy their US 'no refund on sale items' policy into their Australian storefront. That sign is per-se unlawful under ACL section 29 if the sale item has a major failure. The ACCC has taken enforcement action against large retailers for exactly this pattern.

There's no federal US law requiring merchants to accept returns, but roughly a dozen states impose specific disclosure requirements on merchants who want to restrict returns. These state-level rules are where the actual compliance exposure lives for US-only sellers.

California (Civil Code §1723). Any retailer with a policy of not giving full cash or credit refunds, or not allowing equal exchanges, for at least seven days following purchase must conspicuously display that policy — either on signs at each cash register and sales counter, at each public entrance, on tags attached to items, or on order forms. The display must state whether cash refund, store credit, or exchange is given; the time period; the types of merchandise covered; and any other conditions. If a retailer fails to comply, the consumer may return goods for a full refund within 30 days of purchase with proof of purchase. California's law also separately covers automatic-renewal subscriptions under Bus. & Prof. Code §§17600–17606, now widely treated as the de facto US subscription standard.

New York (GBL §218-aa). Mirrors California's model. Retailers must conspicuously post their refund policy; if they don't, consumers can return unused merchandise for a full refund within 30 days with proof of purchase. New York GBL §349 (deceptive acts and practices) supplements this with a private right of action for deceptive refund-policy disclosure.

Florida (§501.142). Specifically targets 'no refund' retailers. Any store not offering full cash or credit refunds, or not allowing equal exchanges, must conspicuously post the restriction at a point visible to buyers from the cash register. If the required disclosure is not posted, Florida consumers can return unused merchandise for a full refund within 7 days of purchase with proof of purchase. Florida's 7-day default is the shortest among disclosure states, but the disclosure failure makes the store's policy entirely unenforceable.

Massachusetts (940 CMR 3.13). The strictest framework in terms of consequences. Retailers must 'clearly and conspicuously disclose' their refund, return, or cancellation policy to the buyer before the transaction is completed. Failure to disclose is an unfair and deceptive act or practice under M.G.L. c. 93A. Massachusetts also requires that gift certificates and store credits remain valid for at least seven years, a requirement most retailers don't handle correctly in their policy language. Chapter 93A gives consumers a private right of action with minimum statutory damages and attorneys' fees, making Massachusetts the most attractive jurisdiction for plaintiffs' lawyers.

Minnesota (§325F). Requires the refund policy to be displayed in at least 14-point bold type (or the digital equivalent) at the point of sale. Failure to comply means the default rule applies: the consumer can return goods for a full cash refund. Minnesota's type-size requirement is the most specific in the country and is frequently overlooked in online policies, which may use smaller fonts to fit mobile screens.

Texas (Business & Commerce Code §17.41). Texas doesn't have a specific return-disclosure statute but the Deceptive Trade Practices Act (DTPA) covers misrepresentations about refund policies as actionable deceptive practices. The DTPA gives consumers a private right of action with treble damages for knowing violations. For Texas-specific enforcement, the focus is usually on misleading disclosures rather than missing ones.

Other states with notable rules: Virginia requires 20-day refunds absent disclosure. Hawaii defaults to 60-day refunds. New Jersey requires clear display, with 20-day default otherwise. Ohio requires posted policies with implied full-refund default. The state patchwork means that if you sell nationwide, the practical minimum is a 30-day window with conspicuous online disclosure.

The key insight: the disclosure is the compliance. A state-compliant policy disclosure at checkout doesn't require you to offer generous terms — it just requires you to clearly state what you actually offer. Failure to state, in most of these states, defaults to the most consumer-favorable interpretation.

On October 16, 2024, the Federal Trade Commission finalized the Rule Concerning Recurring Subscriptions and Other Negative Option Programs — better known as the 'Click-to-Cancel Rule.' It would have been the most consequential US subscription-commerce regulation in decades. On July 8, 2025, the Eighth Circuit vacated it. Understanding exactly what happened, and what's still in force, is essential for anyone selling subscriptions in 2026.

The Rule, as finalized, imposed four core requirements on any seller using a 'negative option' (where the consumer's silence is treated as continued acceptance — auto-renewing subscriptions, free-trial-to-paid conversions, continuity plans):

• Misrepresentation prohibition: no misrepresentation of any material fact in marketing goods or services with a negative-option feature.
• Disclosure: clear and conspicuous disclosure of all material terms before obtaining billing information.
• Consent: express informed consent to the negative-option feature separately from the rest of the transaction (no pre-ticked boxes).
• Simple cancellation: a cancellation mechanism as easy as the signup mechanism, in the same medium.

The Rule was scheduled to take effect in stages, with the misrepresentation provisions effective January 14, 2025 and the rest on May 14, 2025 (later deferred to July 14, 2025). Industry groups challenged it immediately. The Eighth Circuit consolidated the challenges and ruled on July 8, 2025 that the FTC's rulemaking process was procedurally insufficient. Specifically, the court held that the FTC failed to issue a required preliminary regulatory analysis for a rule estimated to have over $100 million in annual economic impact — a procedural requirement under Section 22 of the FTC Act. The court did not address the substantive challenges; it vacated on procedural grounds only.

The current status (as of 2026): the Rule is not in force. The FTC under its current leadership has signaled it will not attempt to re-promulgate the Rule without congressional direction, citing a policy of 'regulatory humility.' Whether that position holds through 2026 and beyond is uncertain.

But the subscription-cancellation landscape has not returned to 2022. What's still enforceable, right now:

• ROSCA (Restore Online Shoppers' Confidence Act, 15 U.S.C. §8403). Federal law requiring clear disclosure of material terms, express consent for recurring charges, and a simple mechanism to stop recurring charges for online transactions. The FTC has enforced ROSCA actively, including against Amazon in 2023 and Uber in 2025, with settlements and consent decrees in the hundreds of millions of dollars.
• FTC Act §5. The general prohibition on unfair and deceptive practices. Subscription cancellation failures are prosecuted as §5 violations whenever they're not covered by a specific rule.
• California auto-renewal law (Bus. & Prof. Code §§17600–17606). The strictest widely-applicable state law. Requires clear and conspicuous disclosure of automatic-renewal terms, affirmative consent, an acknowledgment in retainable form, and a cost-effective, timely, and easy-to-use online cancellation mechanism for any subscription originated online. California's AG and private plaintiffs under CA's UCL can pursue violations.
• New York Auto-Renewal Law (GBL §527-a). Expanded in 2024 to require notice before renewal for most subscriptions, with specific disclosure mandates.
• Other state laws: Connecticut, Colorado, Illinois, Oregon, Vermont, and roughly two dozen other states have enacted auto-renewal statutes with varying requirements.

The practical consequence: if your subscription policy was designed to comply with the FTC Click-to-Cancel Rule, it's already broadly compliant with what's still in force. California's law is the strictest widely-applicable standard, and drafting to California makes your policy portable. The specific things that remain legally required:

• Clear pre-transaction disclosure of recurring-charge terms, frequency, and amount.
• Affirmative consent to the recurring charge separately from other terms.
• Online cancellation available if signup was online, using the same medium.
• Cancellation must be 'cost-effective, timely, and easy-to-use' (CA language).
• Acknowledgment of the cancellation mechanism in a retainable format.

What's probably still coming back: a re-promulgated version of Click-to-Cancel with the procedural defects cured. The FTC signaled in its February 2026 advance notice of proposed rulemaking that some version of the Rule is likely in 2026–2027. Draft your policy for the California standard now and you won't need to redraft when that happens.

If you accept Visa, Mastercard, or other major payment networks, chargebacks are a core operational reality. A well-drafted return policy is almost the entire representment defense — it's the document your acquirer's dispute team will submit, the document the issuer's team will examine, and the document the card network will rule on. A policy that doesn't map cleanly to the specific reason codes you're likely to see is a policy that loses representments.

The six most common chargeback reason codes for online merchants, and what your policy should do for each:

Visa 13.1 / Mastercard 4855 — Merchandise or Services Not Received. The cardholder claims they never got what they paid for. Cardholder time limit: 120 days (540 days for delayed-delivery services like travel). Merchant response: 30 days (Visa), 45 days (Mastercard). Your policy needs to state the estimated delivery timeline, provide tracking information by default, and include a clear process for reporting non-delivery. Your representment evidence is tracking data plus your policy's stated process; if the cardholder didn't follow the process before disputing, that's a defense.

Visa 13.2 / Mastercard 4841 — Cancelled Recurring Transaction. The cardholder claims they were charged for a subscription after cancelling. This is the single highest-volume chargeback reason for any subscription business. Your policy needs to: (1) clearly describe the cancellation mechanism, (2) state that cancellation is effective at the end of the current billing period (not immediately refundable for the current cycle), (3) provide a confirmation that the cancellation was processed, and (4) maintain logs of cancellation attempts. Your representment evidence is the absence of a cancellation request in your logs, plus your disclosed policy. If your policy's cancellation mechanism is 'email us' and the customer emailed you, and you can't show what happened next, you lose.

Visa 13.3 — Not as Described or Defective Merchandise/Services. The cardholder claims the product or service was materially different from its description. Your policy should include an explicit defective-item process that says: contact us first, we'll replace or repair or refund, and specify the documentation required (photos, description). If the cardholder didn't contact you first, your representment package includes the absence of that contact plus your policy's stated process. This is also where product-page copy matters: if your policy says 'not as described' is defined by your product description, your product description becomes the standard the card network applies.

Visa 13.5 — Misrepresentation. Broader than 13.3 — covers situations where the overall transaction was misrepresented (including refund-policy misrepresentation). A vague policy that doesn't actually disclose the terms is evidence of misrepresentation; a specific policy that the cardholder received before purchase is evidence against it.

Visa 13.6 — Credit Not Processed. The cardholder claims they were due a refund but never received it. Your policy needs to specify the refund timeline (e.g., 'within 14 business days of return receipt') and the method (original payment method). Your representment evidence is your refund record and the timeline stated in your policy.

Visa 12.4 / 12.5 — Incorrect Amount / Duplicate Processing. Your policy should reference what to do about billing errors (contact us, we'll investigate promptly) and commit to correcting duplicate charges or pricing errors on notification. Representment requires your transaction records plus evidence the cardholder didn't follow the policy's error-reporting process.

Key operational practices your policy should reference, even if they're not the main subject:

• Keep transaction authentication records (3-D Secure authentication, AVS, CVV) for representment of fraud-related disputes.
• Maintain Order Insight / Compelling Evidence data (IP address, device ID, shipping address, login history) for Visa 10.4 compelling evidence 3.0.
• Use Ethoca or Verifi alerts to resolve disputes pre-chargeback when possible.
• Keep your dispute ratio below Visa's 0.9% (or Mastercard's 1.5% ECM threshold) to avoid monitoring programs.

Your policy is the thesis statement; your operational records are the evidence; your systems are the infrastructure. All three need to line up. A policy that claims 'refunds processed within 14 days' and a system that takes 21 days loses representments and invites regulatory attention independently.

Selling digital goods internationally is where refund policies most often break. The rules are not intuitive, they vary by jurisdiction, and the failure mode is expensive — a policy that looked compliant for years can be invalidated by one thing the EU Court of Justice decided a month ago. Here's the 2026 state of play.

The core tension: digital content, once downloaded or streamed, cannot meaningfully be 'returned.' The consumer has the file or has consumed the performance. But consumer-protection regimes do not accept 'digital = non-refundable' as a principle. Instead, they've built specific frameworks that let you exclude digital content from general withdrawal rights if — and only if — you meet specific procedural conditions.

EU Consumer Rights Directive Article 16(m). The right of withdrawal does not apply to 'the supply of digital content which is not supplied on a tangible medium if the performance has begun with the consumer's prior express consent and his acknowledgment that he thereby loses his right of withdrawal.' Three conditions: (1) express prior consent to beginning performance, (2) acknowledgment of losing the withdrawal right, (3) the performance has actually begun (not just payment — actual access or download).

UK Consumer Contracts Regulations 2013, Regulation 37(1). Mirrors Article 16(m). Same three conditions.

What counts as 'express prior consent' and 'acknowledgment'. This is where most merchants fail. The Zalando case (ECJ C-332/14) and subsequent guidance make clear that:

• A pre-ticked checkbox does not satisfy express consent. The consumer must affirmatively check the box.
• Acceptance of general terms of service does not satisfy express consent or acknowledgment for this specific purpose.
• The acknowledgment must be specific to the withdrawal right — a general 'I accept the terms' is not enough.
• The consent must be documented in a durable medium (the order confirmation email is typically where this lives).

The practical implementation that works:

1. At checkout, before payment, display a specific unchecked checkbox with language along the lines of: 'I consent to immediate access to [Product Name] and I acknowledge that by doing so I lose my 14-day right of withdrawal / cooling-off right for this digital content.'
2. Require the customer to check the box before enabling the purchase button.
3. In the order confirmation email, include a durable-medium record of the consent: 'Your order includes immediate access to [Product Name]. At checkout you consented to immediate access and acknowledged loss of the 14-day withdrawal right for this digital content.'
4. Your refund policy references this consent process explicitly and sets out that digital content accessed after consent is non-refundable except for defects or material misrepresentation.

The alternative: the percent-accessed threshold. Instead of an absolute waiver, you can commit to partial refunds based on how much of the content the customer accessed. This is the model that works well for courses and info products: 'If you access less than 25% of the content, you can request a full refund within 14 days.' The threshold must be disclosed at checkout and must be verifiable through your system — if you don't actually track access, a percent threshold is not defensible. This model is friendlier to customers (reduces chargebacks) and cleaner to administer if your platform has access analytics.

US treatment. No federal statute requires specific express-consent structures for digital content. But state consumer-protection statutes still apply, and a 'no refund on digital goods' policy that wasn't disclosed before purchase is a UDAP violation in most states. FTC Act §5 applies if the disclosure pattern is unfair or deceptive. The safest approach for US customers is still the express-consent checkbox — it's not technically required, but it's the pattern that best documents your compliance with disclosure requirements and also serves as representment evidence in chargebacks.

Australia. The ACL's consumer guarantees apply to digital content the same way they apply to physical goods. You cannot waive the guarantees regardless of consent language. If the digital content has a major failure — doesn't work, doesn't match description, causes damage — the consumer has the right to a refund, replacement, or compensation regardless of whether they accessed the content. Your policy can limit change-of-mind refunds for digital goods, but it cannot limit ACL remedies for failed content.

Our generator produces all four components: the policy language, the checkbox text, the order-confirmation durable-medium language, and the ACL carve-out. They work together — none of them works alone.

The operational parameters — how long a customer has to return, who pays shipping, whether you charge a restocking fee — are where most policies are either needlessly generous or unnecessarily restrictive. Here's how to think about each, informed by what actually survives chargeback representment and regulatory scrutiny.

Return window length. The default should be 30 days for general US e-commerce. Why: it matches California (§1723) and New York (§218-aa) defaults; it's competitive without being generous; it aligns with common card-network chargeback windows for cardholder disputes (120 days, so you want returns closed long before the chargeback window expires). Going shorter than 30 days for US sales creates friction without meaningful cost savings; going longer (Zappos' 365 days, REI's 1 year, IKEA's 365 days) is a marketing investment. For SaaS and digital, 14 days matches the EU/UK statutory minimum and is defensible internationally. For services, 14 days from contract conclusion is the standard. The window starts from the day of receipt for physical goods, not the day of order — this matters because shipping time otherwise eats into the customer's effective return period.

Return shipping costs. Three viable models. First, customer pays return shipping for change-of-mind returns. This is the default for most physical e-commerce and is compliant across all jurisdictions (with carve-outs for defective items). Second, merchant pays return shipping. This is a strong marketing position (Zappos, Bonobos, Warby Parker all use it) but has operational cost implications. Third, split policy: customer pays for change-of-mind, merchant pays for defective or not-as-described. This is the compliance-optimal approach — it mirrors what EU CRD, UK CCR, and ACL all require for defective-item returns, and it creates clear customer expectations.

Critical jurisdiction notes on shipping: EU and UK law both require that the basic standard-delivery cost be refunded on statutory withdrawal returns (but supplements for expedited delivery don't have to be refunded). Your policy needs to distinguish between 'basic shipping' (refunded) and 'premium shipping supplements' (not refunded). ACL requires the merchant to pay return shipping for any item with a major failure, and for minor failures where the item is large, heavy, or difficult to move (e.g., a refrigerator). A policy that says 'customer always pays return shipping' is per-se non-compliant in Australia.

Restocking fees. The controversial one. Restocking fees are generally allowed in US states if clearly disclosed before purchase, typically 10-25% of the item price. But they come with significant carve-outs:

• No restocking fee for defective, damaged, or not-as-described items.
• No restocking fee for EU statutory withdrawal returns.
• No restocking fee for UK CCR cooling-off returns.
• No restocking fee for ACL major-failure returns.
• No restocking fee for returns due to your shipping error or mistake.

The sweet spot for restocking fees is 10-15% on change-of-mind returns of items that genuinely require significant handling to return to inventory (configured electronics, custom-cut materials, opened beauty products). 25%+ invites regulatory attention and chargebacks; 0% is fine for items that don't actually cost you anything to restock. The fee should be disclosed at checkout (near the order button, not three clicks deep) and referenced in the order confirmation email.

Shipping refunds on the original purchase. When you refund a product, three models: (1) refund only the product cost, not shipping; (2) refund product cost plus basic standard shipping; (3) refund everything the customer paid including premium shipping. Model 2 is the compliance-optimal default — it matches what EU/UK law requires for statutory returns and is customer-friendly enough for change-of-mind returns. Model 1 is allowed in most US states if disclosed but creates more customer friction than it saves money for. Model 3 is a marketing choice.

Diminished-value deductions. EU CRD and UK CCR both allow you to deduct for diminished value if the customer handled the goods beyond what's necessary to examine the item (the 'shop inspection' standard). This is useful for items that were obviously used — a returned laptop with hundreds of hours of usage, an opened cosmetic product with half the contents gone. But the deduction has to be justified and documented, and 'we think the customer used it' is not enough. The policy should reference the standard; the actual deduction decisions are operational.

If you process more than a few hundred transactions a month, chargebacks will happen. Some you'll lose because the customer has a legitimate complaint; some you'll lose because your policy failed you. Here are the ten most common policy-driven representment failures, in rough order of frequency:

1. Vague cancellation mechanism. Your policy says 'cancel anytime' but doesn't specify how. The customer sends an email; you never respond; they chargeback as Visa 13.2. Your representment package has no cancellation record because there was no structured cancellation process. Fix: your policy must specify exactly how cancellation works, preferably a self-serve mechanism, and your system must log cancellation attempts.

2. Missing pre-purchase disclosure. Your policy is on a '/refund-policy' page linked from the footer. The customer's checkout page says nothing about returns, refunds, or cancellation. In state disclosure regimes (CA, FL, MA, MN), this missing conspicuous disclosure invalidates any restrictive policy. In card-network representment, 'the policy was available on the website' is weaker evidence than 'the policy was shown at checkout.' Fix: short-form policy copy at checkout, with a link to the full policy.

3. Restocking fee on EU withdrawal. Your policy has a 15% restocking fee clause that applies to all change-of-mind returns. An EU customer exercises their 14-day right of withdrawal. You deduct 15%. They file a complaint with their national consumer-protection authority, which refers it to their credit card issuer. You lose the representment because EU CRD doesn't allow restocking fees on statutory withdrawal returns. Fix: carve out statutory returns from restocking fee language.

4. Defective-item process collapsed into standard returns. Your policy treats all returns the same — 14 days, customer pays shipping, no restocking waiver for defects. A customer reports a defective item after 30 days; you refuse the return; they chargeback as Visa 13.3. You lose because the card network expects (and most jurisdictions require) a separate defective-item process with different terms. Fix: dedicated defective-item section with longer window (most regimes: 'reasonable time' from discovery), no shipping charge to customer, no restocking fee.

5. No digital content consent mechanism. You sell digital downloads and have a 'digital products non-refundable' policy. An EU customer downloads a product, requests a refund, you refuse, they complain. You lose because EU CRD requires express consent + acknowledgment — a general policy doesn't satisfy Article 16(m). Fix: checkout checkbox with specific language; order-confirmation durable-medium record.

6. Return window too short for state defaults. You're based in Texas, have a 10-day return window, and sell to California customers. Your CA customers can claim the 30-day default because your conspicuous disclosure requirements failed (short window + online-only disclosure + no signage — CA courts have held that online disclosure can satisfy §1723 but only if it's prominent). Fix: for US multi-state sales, 30 days is the practical minimum and matches the most common state defaults.

7. ACL 'no refund on sale items' language. You copy a US sale-exclusion policy into your Australian storefront. The policy says 'no returns on sale items.' The ACL makes this per-se unlawful for items with a major failure. ACCC enforcement follows. Fix: don't ever use blanket 'no refund' language for AU customers; carve out ACL consumer guarantees explicitly.

8. Inconsistent policy across touchpoints. Your website policy says 14 days. Your order confirmation email says 30 days. Your customer service reps tell people 7 days. In representment, the card network sees conflicting evidence. In class actions, plaintiffs' lawyers love the inconsistency. Fix: one canonical policy, referenced everywhere, with customer-service training to match.

9. No chargeback-deterrent language. Your policy doesn't mention chargebacks at all. Customers default to filing a chargeback for any issue because your policy doesn't signal any alternative process. Fix: 'Contact us first' language in the policy; 'Before you file a chargeback, please email us — most issues resolve same-day' in order confirmation emails.

10. Outdated subscription cancellation language. Your policy still references the FTC Click-to-Cancel Rule as if it were in force (it was vacated July 2025). This signals either that you're not paying attention to regulatory change or that your policy was auto-generated by a tool that hasn't been updated. Neither impression helps in an enforcement action or a press inquiry. Fix: reference the actually-applicable rules (ROSCA, CA auto-renewal, FTC Act §5) and the 19 June 2026 EU cancel-button deadline.

Each of these is fixable in the generator in less than five minutes. The hardest part is knowing to look for them, which is why the scorecard flags them automatically.

You do not need a week to do this right. An afternoon is enough for initial publication; ongoing maintenance is thirty minutes a quarter. Here's the sequence that actually works, informed by observing what merchants skip when they try to do this themselves:

1. Pick the business model. Ten minutes. The model template presets the refund methods, exclusions, and jurisdictions that actually fit your operation. Don't try to build your own from scratch — start from a template.
2. Fill in business info. Five minutes. Business name, legal entity, website, support email, return address (for physical goods), effective date. These populate every clause.
3. Select jurisdictions based on where your customers actually are. Five minutes. Look at your last six months of orders, identify the top countries by volume. US is always selected. Add EU, UK, or AU if you have more than a handful of customers from those regions. For US multi-state sellers, always include CA and NY (they have the broadest default rules); add FL if you have significant Florida traffic, MA if you have significant Massachusetts traffic.
4. Set the return window and refund methods. Ten minutes. 30 days for US physical e-commerce; 14 days for digital, SaaS, and services. Full refund + exchange + store credit for physical; prorated refund + cancel-no-refund for SaaS; full refund for digital with access-waiver model. Shipping: customer pays return shipping except for defective items. Restocking fee: 10-15% only if you actually have handling costs to recover; zero otherwise.
5. Configure exclusions. Five minutes. Final sale (if you do clearance), custom/personalized (if you offer customization), perishable (if relevant), hygiene (for intimate goods), digital accessed (for digital products with access-waiver), gift cards, services already rendered (for services), damaged by buyer (always enable).
6. Handle subscriptions and digital access. Five minutes. If you have subscriptions, select the subscription type (monthly, annual, trial-convert). If you sell digital content, choose the access model (waived-on-start for most digital; percent-accessed for courses; fully-refundable for the most generous approach).
7. Check the scorecard. Ten minutes. Aim for 90%+. Each warning has a specific fix; each failure is something you actually have to address. Common warnings: missing return address (add it), restocking fee vs EU conflict (carve out statutory returns), digital consent (enable waived-on-start and plan to add the checkout checkbox).
8. Export and publish. Ten minutes. HTML for your website (typically at /refund-policy/ or /returns/), Markdown for your internal docs, JSON for backup. Update your site footer to link to /refund-policy/. Add short-form copy to your checkout page (from the placement copy the generator produces).
9. Add the checkout disclosure. Ten minutes. This is the step most merchants skip, and it's the one that actually matters for chargeback representment and state disclosure compliance. Use the generator's short-form checkout copy near the order button. For digital products with a waiver, add the consent checkbox.
10. Save the JSON config somewhere you won't lose it. Notion, Google Drive, a password manager's secure notes. Label it 'Return policy config 2026-04-16' or similar. Quarterly updates take ten minutes if you load this; they take an hour if you rewrite from scratch.
11. Set a calendar reminder for quarterly review. Thirty minutes. Load the JSON, rerun the scorecard, check for regulatory changes (the two big 2026 ones to watch: 19 June 2026 EU cancel-button deadline, and any FTC re-promulgation of Click-to-Cancel), export updated version, publish.

Total initial investment: about 80 minutes of focused work for a policy that would cost $300-$1,500 to have drafted by a digital-commerce attorney. Ongoing maintenance: about 2 hours a year. Your actual business will throw curveballs at this — a new product category, a new jurisdiction, a regulatory change — and when it does, the JSON configuration makes the update incremental instead of wholesale.

One closing observation: the return policy is not the most glamorous document on your website, but it's consistently the most scrutinized. Customers who never read your terms of service read your return policy carefully before buying. Card networks examine it in every chargeback. State AGs request it when investigating consumer complaints. The small amount of care it takes to get it right pays back many times over in preserved revenue, reduced disputes, and avoided regulatory exposure.