Return Fees vs. Free Returns: What Ecommerce Brands Should Actually Optimize

Return fees are charges ecommerce brands may apply when customers send merchandise back, and they can reduce how often shoppers return items while also raising complaints, shrinking average order value, and pushing a buyer toward a competitor’s checkout instead of yours. For ecommerce operators deciding how to structure a returns policy, the real question isn’t whether to charge for returns at all; it’s which resolution, refund, exchange, keep-it credit, verification, resale, or physical return, actually protects both the customer relationship and the margin on that order.

Return fees are not a returns strategy on their own. They are one economic lever inside a much larger and very costly retail function, so the better decision is which resolution creates the best outcome for the customer and the merchant, not whether returns should be free or paid as a blanket rule. That’s what the operator data below focuses on: how charging for returns changes behavior, when alternatives like exchanges or keep-it credits work better, where fraud controls help or hurt, how return routing affects cost, and how to balance customer experience with sales and profitability.

Here’s what the operator data below actually shows:

• Charging for returns changes behavior. Merchant data cited by NRF shows lower overall return rates and higher exchange rates among brands that charge for at least one return option, alongside more complaints, lost customers, lower average order value, and lower sales.
• Kulfi Beauty treats exchanges and shade corrections as the first move, and lets customers keep low-value items outright rather than shipping them back.
• Carve Designs prices refunds and exchanges differently on purpose, and pairs that pricing with fit guidance that has measurably changed its cohort return rates.
• Fraud is real, at roughly 9% of returns industrywide, but blanket friction built to stop it tends to punish loyal customers more than it stops bad actors.
• The cheapest return is usually the one that never has to travel back to a warehouse at all.
• Return fees are one lever in a larger economic decision, not a substitute for one.

Return fees and restocking fees work, but they also create a commercial cost

U.S. retail returns were projected to reach $849.9 billion in 2025, with an estimated 19.3% of online purchases sent back, according to NRF and Happy Returns research. At that scale, even small shifts in return rate move real dollars, which is why fees keep coming up in board meetings.

The same NRF research shows why brands can’t treat fees as a free lever. Eighty-two percent of consumers say free returns are an important consideration when deciding where to shop, and 71% say they are less likely to shop with a retailer again after a poor return experience.

NRF’s merchant-side data adds the other half. Seventy-two percent of merchants surveyed charged for at least one return option, and the reported effects cut both ways, reinforcing how an exceptional returns program can be a loyalty driver as much as a cost center.

Reported positive effects after charging	Reported negative effects after charging
53% lower overall return rates	47% more customer complaints
52% increased exchange rates	37% lost customers over fees
Fees recouped some revenue and shifted behavior toward exchanges	34% lower average order value
Some shoppers chose a free alternative return method instead	24% lower sales

Read plainly, that table isn’t an argument for or against fees. It’s evidence that a return fee is a behavior-shaping tool with a measurable upside and a measurable commercial risk attached to the same decision. Fraud sits inside this picture too: NRF puts fraudulent returns at roughly 9% of the total, a benchmark worth knowing before deciding how much friction a policy needs (more below). None of this makes free returns the automatically safer default either; Cahoot has covered why free returns are no longer the sacred, unconditional expectation they were during the pandemic-era ecommerce boom, and has also detailed the rising financial and environmental cost of free returns. Fees change behavior in measurable ways, and an operator who treats that data as directional, not moral, makes better decisions than one who treats fees as either a betrayal or a free win.

Kulfi shows why some low-value returns should not come back at all

Kulfi Beauty’s approach starts before a return gets requested. Speaking on Cahoot’s Ugly Talk series, Kulfi’s Gabrielle Kerins described the brand’s first line of defense as an exchange or shade correction, not a refund.

For returns under $50, Kerins said Kulfi goes further: the customer keeps the item, and Kulfi deducts a processing fee rather than paying to ship the product back. Customers are sometimes encouraged to pass the item to a friend or sibling while Kulfi helps them find a better match. That guidance came from Kerins onstage, not Kulfi’s published policy; the brand’s current public FAQ lists a separate $6.95 return processing fee deducted from the refund, described as a way to partially recover shipping and processing costs, on Kulfi’s FAQ page. By comparison, H&M standardized a $3.99 mail return fee for all customers in 2025.

Kerins also treated the fee itself as a live experiment. Processing returns commonly costs about $10 to $30 per item, which helps explain why brands test deducted fees on low-value orders. A modest, competitively priced fee increase generated little pushback, but she was clear the brand would revisit it if feedback suggested the fee had become a real barrier to a customer’s first purchase.

Not every return Kulfi sees is a customer preference problem. Repeated return feedback on one lip product surfaced a pattern: packaging that had passed quality control behaved badly at certain temperatures, causing leaks. Kulfi used that data to repackage the product rather than assuming shoppers were simply changing their minds, a reminder that return reason data is a quality control signal, not just customer friction.

Carve Designs prices refunds differently from exchanges

Carve Designs’ public return policy draws a clean line between the two outcomes. A refund carries a $10 return shipping fee deducted from the amount refunded, but the brand allows one free exchange per order, and that fee isn’t charged on an exchange unless the same order also includes an item returned for refund, according to Carve’s returns and exchanges policy. Typical online return fees often fall in the $4 to $12 range for mail-in returns. By contrast, percentage-based restocking fees can run higher; Best Buy may charge a 15% restocking fee for opened items. The structure rewards the outcome Carve wants more of, an exchange that keeps revenue in the business, without waiving the cost of the one it wants less of, a refund that sends inventory and cash back out.

Pricing isn’t the only lever Carve pulls. Per a Komar event deck presented by Jay Harris at Ugly Talk NYC, roughly 20% of Carve shoppers opt into the brand’s proprietary swim fit quiz before buying, and the deck reports average order value rose and returns fell among that group, without attaching a specific reduction percentage to the quiz alone. The same deck shows a broader cohort shift: before Carve’s combined fit and purchase-path work, new-customer return rates ran around 35% and repeat-customer rates around 25%; after that work, the same cohorts fell to roughly 20% and 12%. That change illustrates how ecommerce return rates directly affect profit margins. That’s better pre-purchase guidance paired with a return policy that has real economic teeth, not one feature working alone.

The lesson isn’t “add a quiz.” Carve’s fee structure recovers cost and nudges customers toward exchanges, but the larger cohort-level improvement came from reducing wrong-size and wrong-fit purchases before they ever shipped.

The cheapest return is often the one the brand prevents

Apparel and footwear carry some of the highest return rates in ecommerce: fit can’t be verified until the product is on the customer’s body. Coresight Research estimated the U.S. online apparel return rate at 24.4%, with size and fit cited by 53% of surveyed brands and retailers as a top reason. Broader analyses of the rise in ecommerce return rates echo those drivers. The same Coresight research, sponsored by sizing-technology vendor 3DLOOK, estimated 2023 online apparel returns at roughly $38 billion, with about $25.1 billion in processing costs attached. Those numbers are apparel-specific; a beauty brand like Kulfi won’t see the same rate, and comparing return rates across categories without adjusting for that is a common mistake. Fees also tend to be higher for large or bulky items because return logistics get more expensive as size and weight increase.

What does travel across categories is the economics of where a return gets processed. McKinsey’s research on apparel returns management found the difference between a retailer’s least and most expensive return channel averaged $5 to $6 per unit, and that in-store processing could save up to 18 days compared with warehouse processing, improving the odds an item resells at full price. That helps explain why a retailer may set separate charges for different costs, and why a restocking fee can vary by product category; the true cost to process an e-commerce return can run $10 to $35.

Accurate product descriptions, clear sizing guidance, and basic quality control belong in the same conversation as return fees. Kulfi’s repackaging fix and Carve’s fit quiz are both prevention plays: they reduce the number of returns that ever need a fee policy applied. For the full accounting of what a return costs in labor, shipping, and lost inventory value, see Cahoot’s breakdown of the hidden economics of a return.

Fraud needs targeted friction, not a worse policy for everyone

Fraud is a real cost, but a smaller share of returns than most operators assume. NRF puts the market benchmark at roughly 9% of all returns reported as fraudulent, a useful anchor when one fraud story starts to drive an entire policy.

George Bova, also speaking on Ugly Talk, described a wholesale customer, a restaurant, that had used roughly 40% of a bottle of hand sanitizer before returning it for a refund. That’s abuse a brand can act on directly: a specific customer, a specific pattern, a consumption level that makes “changed my mind” implausible.

Bova also described the failure mode on the other side. One brand, trying to stop that kind of abuse, started requiring records, receipts, serial numbers, and other proof before processing any return. Legitimate refunds slowed down, and negative reviews followed. Friction applied evenly across every customer, instead of targeted at accounts and patterns that actually look like abuse, taxes loyal customers most while doing the least to stop the volume it was meant to reduce. It also should not be applied when the issue is a defective product, since charging restocking fees on defective items is illegal in most regions.

Policy is only one part of the economics, routing and recovery speed matter

A return fee changes what a customer does before shipping an item back. It does nothing to change what happens once that item arrives, and treating those as one cost center is how brands miss real savings. Many retailers waive return fees for in-store returns even when mailed returns cost more.

Blue Yonder’s research found 30% of surveyed retailers had implemented flexible return shipping charges or restocking fees that vary by reason, and 63% said charges always or sometimes vary by reason. The industry is moving away from one flat fee and toward routing decisions based on why an item is coming back.

The gap shows up in the workflow: forward fulfillment is typically three touches, pick, pack, ship. A reverse apparel path in Komar’s event framework at Ugly Talk NYC can run up to seven: receive, inspect, steam, re-tag, re-poly, re-slot, or liquidate. A brand can shrink that path without touching its refund policy, by routing eligible items around those steps instead of charging customers more.

Carve’s numbers illustrate this. Per the same Komar event materials, Redo was attributed roughly $250,000 in return-freight savings for Carve in one year, called onstage hundreds of thousands of dollars, without shortening the return window or adding a restocking fee. That’s one brand’s reported result, disclosed as Komar and Jay Harris’s event material rather than audited data, but it shows pricing and routing are separate levers.

This is the layer where Cahoot operates: an end-to-end fulfillment operations suite built around saving every penny a returns process doesn’t need to spend, a claim backed by fulfillment customer reviews highlighting lower shipping costs and better efficiency. Cahoot’s Peer-to-Peer Returns recovers value from eligible returned items before unnecessary warehouse processing and reverse logistics, building on the same peer-to-peer fulfillment model described in Cahoot’s overview of peer-to-peer as the future of order fulfillment. When a return starts, eligible items can be verified and matched against new demand; if a buyer orders during that resale window, the item ships directly to them instead of completing a warehouse cycle first. If no match exists, the item follows the standard workflow. Amazon charges return fees unless shoppers use label-free drop-off options. That changes routing economics, not policy harshness, part of the shift away from treating a warehouse as the only place a return can go. See how Cahoot’s Peer-to-Peer Returns can reduce unnecessary reverse-logistics cost on eligible returns.

Use a decision model, not a blanket return rule

Kulfi keeps low-value items rather than shipping them back; Carve charges a flat fee on refunds but not exchanges; Cahoot’s routing model changes what happens after a return starts rather than what a customer pays upfront. Each decision gets made at the level of the individual return, not as a blanket rule for every order.

A practical version of that decision looks like this:

• Resale value: What can this item resell for, after reverse shipping, handling, and cycle time?
• Who caused it: Did the customer change their mind, or did the brand cause it through the wrong item, damage, lateness, poor fit or shade guidance, or a defect? When the brand caused the problem, a fair approach also accounts for region-specific legal regulations that may govern what a seller can charge.
• Exchange potential: Would an exchange solve the problem and preserve more revenue than a refund?
• Fraud signal: Is there real evidence of abuse justifying targeted verification, or would friction just slow a legitimate customer?
• Routing need: Does this need to travel back to a warehouse, or is there an eligible route that avoids reverse logistics costs the brand doesn’t need to pay?

Answer those honestly, and the right resolution usually becomes obvious without a company-wide policy debate. The future of returns isn’t free returns or paid returns. It’s economically intelligent returns, priced and routed based on what a specific return actually costs and recovers, not on an ideology about fees.

Brands matching resolution to individual customer history and segment, rather than just return type, are getting into personalization territory beyond what a single fee policy can do. Customer history can include loyalty status, since members are often exempt from return fees. Cahoot covers that ground in its guide to individualized ecommerce return policies; the fundamentals of an ecommerce return policy are worth reviewing before layering fees, exchanges, or segment logic on top.

Measure return economics by cohort, SKU, recovery, and cycle time

A blended return rate hides more than it reveals. Komar’s operator framework, presented at Ugly Talk NYC, breaks that single number into measurements that actually point to a decision.

• Return rate split by cohort, new versus repeat customers, the way Carve’s shift from 35% to 20% among new customers and 25% to 12% among repeat customers played out differently.
• Return rate ranked by SKU rather than one blended average, since a handful of products usually drive most of the returns.
• Recovery value per unit measured net of cycle time, since an item that resells in three weeks is a different outcome than one resold in three days.
• Reverse cost per unit compared against forward cost per unit, the same comparison that makes the seven-touch reverse path visible.
• The share of returns the brand itself caused, wrong item, damage, lateness, fit or shade guidance, or a defect, the category Kulfi’s leaky packaging story falls into.

That last measurement matters more than it usually gets credit for. A brand that assumes every return is a customer decision will keep adjusting fee policy to influence behavior, when the data might actually point at a packaging defect or a sizing chart that needs updating. Reading return reason data as an operational signal, not just a satisfaction metric, turns returns from a cost center into a source of product improvement.

Frequently Asked Questions

Should ecommerce brands offer free returns or charge return fees?

Charging for returns is legitimate, but not automatically right. Many major retailers adjust policies during the holiday season, such as Amazon allowing returns until January 31, 2026 and Best Buy extending returns until January 15, 2026 for holiday purchases. NRF’s merchant data shows fees can lower return rates and increase exchange rates, while also raising complaints and losing customers over the fee. The better question is whether a fee fits a specific category and customer base, not whether fees are universally good or bad.

Do return fees reduce return rates?

Yes. Per NRF’s merchant survey, 53% of merchants that charged for at least one return option reported lower overall return rates, and 52% reported increased exchange rates, alongside more complaints and lost customers.

Can return fees hurt sales or customer loyalty?

They can. NRF data found merchants who charged fees also reported 34% lower average order value, 24% lower sales, and 37% of customers lost over the fee. Separately, 71% of consumers say they’re less likely to shop again with a retailer after a poor return experience.

When should a brand offer free exchanges but charge for refunds?

This works well when a brand wants to preserve revenue and keep the customer in the product, as Carve Designs does with one free exchange per order alongside a return fee on refunds. It fits apparel and footwear well, where the return is often a fit or shade problem an exchange can solve.

When does a keep-it refund make economic sense?

A keep-it resolution, where the customer keeps the item and the brand deducts a fee from the refund, makes sense when the item’s value is too low for reverse shipping, inspection, and restocking to be worth recovering it. Kulfi applies this logic to low-value returns.

How should ecommerce brands decide which returns within return windows should go back to a warehouse?

A return should go to a warehouse when no faster or cheaper eligible recovery route exists, such as resale to a new buyer during a defined window, local processing, or keep-it resolutions sometimes called returnless refunds. When a match exists, brands recover value without the full receive-inspect-restock cycle and may also avoid charges tied to a prepaid label or return shipping label by directing the shopper to a designated location or to a person for handoff; when it doesn’t, the standard workflow applies.

• Category: Uncategorized
• Tag: Order Fulfillment, Returns, Strategy