“Free returns 30 days” versus “Returns at your expense within 14 days”. The difference looks marketing. It isn’t. A return policy is a measurable conversion lever whose ROI can be calculated with reasonable precision — and which, poorly thought through, can cost far more than its generosity earns back.
The stakes in 2026 are sharper than before for two reasons: Google has made the return policy a ranking signal in Shopping and AI Overviews (via hasMerchantReturnPolicy), and return fraud has professionalised (open door to “try-and-return”, “wardrobing”, and substitution returns). This article breaks down the real ROI calculation and gives the parameters of a policy that converts without bleeding margins.
Why the return policy is a measurable conversion lever
Three mechanisms stack their effects:
- Reduction of perceived risk. An undecided visitor converts more easily knowing they can send the item back at no cost. Elasticity depends on average basket and category: high on fashion (+5 to +15 percentage points of conversion), moderate on beauty (+2 to +5), low on technical goods where free returns are expected but rarely used.
- Lift on average basket. When returns are free, the visitor adds the second size or the second colour more readily — especially in fashion. Average basket typically lifts +8 to +20% on fashion/footwear stores.
- SEO and AEO signal. Since 2024, Google has displayed a “Free returns X days” badge in Shopping and AI Overviews based on hasMerchantReturnPolicy markup. Effect on organic CTR: +10 to +25% observed on comparable product pages.
On the cost side, two main lines:
- Logistic return cost: return label, transport, quality control, repackaging, restocking. On average £6 to £15 depending on volume and carrier partner, more for bulky items.
- “Unsellable” cost: item received back but no longer sellable at full price — soiled, opened, out of season. Depending on the category, 5 to 30% of fashion returns and 1 to 10% of appliance returns.
The ROI calculation: worked example on a UK fashion PrestaShop store
Take a fashion PrestaShop store with monthly revenue of £70,000, average basket £55, conversion rate 2.2%, current return rate 18% with paid returns.
Hypothesis: move to free returns 30 days.
Expected effects (cautious range):
- Conversion: +6 relative points (from 2.2% to 2.33%) → about +£4,200 monthly revenue.
- Average basket: +12% (from £55 to £62) → multiplier effect, roughly +£9,300 additional.
- Organic CTR on product pages: +15%, translating to about +£1,000 additional revenue via SEO.
- Total monthly gross gain: about +£14,500.
- Return rate: rises from 18% to 28% (free-return effect). Returns volume: 28% × (£70K + £14.5K) = ~£24K in returned goods.
- Additional logistic cost: £24K × ~12% of value (label + handling + QC) = ~£2,900.
- Additional unsellable cost: £24K × ~10% unsellable at full price, sold with 30% mark-down = ~£700 loss.
- Total monthly cost: about £3,600.
Net monthly ROI: 14,500 − 3,600 = +£10,900. Over a year, ~£130K of additional margin from a single policy decision.
This example is deliberately favourable (fashion = a price-sensitive category). On technical goods, the calculation is less clear-cut: conversion gains little, returns become a net cost. That is why the calculation must be done per category, not globally.
The parameters of a policy that converts
Duration
Psychology matters more than duration itself. 30 days has become the UK norm, 60 days is a differentiator, 100 days (used by some fashion pure players) is a strong argument. Beyond that, the marginal effect tapers off.
Legally in the UK, the statutory minimum is 14 days under the Consumer Contracts (Information, Cancellation and Additional Charges) Regulations 2013 for distance purchases. Any policy above that is a commercial choice.
Free returns
Three variants depending on context:
- Fully free returns: prepaid label provided. Best for conversion, worst for margins. Reserve for categories with strong price-sensitive conversion (fashion, footwear, eyewear, underwear).
- Reduced-fee returns: “£3.99 return fee”. Reduces “try-and-return” friction without seriously degrading conversion. A healthy compromise for the majority of merchants.
- Customer-paid returns: default. Acceptable on technical goods where the buyer knows what they’re buying, but compensate with a strong argument (free exchange, for instance).
Return methods
Drop-off point (Evri ParcelShop, DPD Pickup, Royal Mail Local Customer Service Point) > in-store return > home pick-up > customer-arranged send-back. The free drop-off point is the 2026 standard in UK fashion. Home pick-up is a premium service that may be paid or reserved for loyal customers.
Exchange preferred over refund
Mechanically, an exchange (size change, colour change, alternative item) costs much less than a refund, because the revenue is not lost. Promoting “Free exchange” before “Free refund” on the returns page redirects a meaningful share of returns to exchange — average margin gain of 15 to 25% on the returns perimeter.
Targeted anti-fraud restrictions
For fashion/luxury stores, certain rules limit “wardrobing” (buy-wear-return):
- Visible anti-return tags on the garment (hang tag that must remain attached).
- No returns on sale items below a certain threshold (legally grey, check with a solicitor).
- Blacklist serial returners (return rate >70%, more than 3 consecutive returns).
Technical implementation on PrestaShop
The native PrestaShop Returns module is minimalist: it generates a return number but handles neither automatic labels, nor accounting workflow, nor carrier integration, nor return reasons statistics.
A complete return workflow on PrestaShop 8 or 9 implies:
- Customer return request form with reason selection (size, colour, defect, other), optional photos, choice between exchange and refund.
- Automatic carrier return label generation (Royal Mail Tracked Returns API, DPD, Evri, UPS).
- Back-office workflow: parcel reception, quality control, decision (acceptance, refusal, mark-down), restocking or disposal.
- Automatic credit note generation with correct operation order (VAT, shipping, bank fees).
- Statistics by reason, by category, by customer, to identify problem products.
At DataFirefly, the dfproductreturn module covers this complete workflow on PrestaShop 8 and 9, with ERP integration for merchants who run a multi-channel back-office.
Measuring what really matters
The right metrics to track on a 3 to 6-month horizon after implementing a policy:
- Conversion rate per category before/after. Must rise on price-sensitive categories.
- Return rate per category. Will rise, that’s normal. Monitor that it stays within the projected range (the ROI calculation is sensitive to this variable).
- Exchange / refund ratio. Ideally >30% in exchange.
- Average return cost (logistics + unsellable). Track monthly to spot drift (carrier, fraud).
- Top 20 products with abnormal return rate. Often a hidden product defect, a bad size guide, a misleading photo. Targeted action is more effective than “tighten the policy”.
The trap of policies that are too generous
A fashion store that abruptly moves from paid returns to 100-day free returns without logistic preparation creates a problem. The return spike arrives 30 to 45 days after the change, and if the back-office is not sized, refunds drag, Trustpilot reviews collapse, and the initial marketing gain is wiped out.
The right sequence: pilot on one category for 2 months first, measure, size the back-office (people, carrier integration, buffer stock), then generalise progressively.
Conclusion: a return policy is a product, not a legal text
Too many merchants treat the return policy as a Terms page drafted once and forgotten. It is a product: it has a target (the price-sensitive category), a price (the additional cost), an ROI to calculate, and a performance measurement over time. Well-designed, it is one of the best UX investments for a fashion or beauty merchant in 2026 — and a signal that tips Google Shopping and AI Overviews.
Poorly designed, it is a gift to fraudsters and a recurring expense. The difference is in the calculation, not in the intention.
