TL;DR
- Returns are roughly 15.8% of all US retail sales and 19.3% of online sales in 2025.
- The grading model (A/B/C/D) determines what happens next: restock, refurb, liquidate, or destroy. Each path has a different recovery curve.
- The customer-facing return experience drives repeat purchase. Friction at return time costs more than the return itself.
- RMA portals, label generation, and inspection workflows are now table stakes. The differentiation is in disposition speed and recovery rate.
How big returns really are.
The 2025 NRF Retail Returns Landscape report puts US returns at $849.9 billion for the year, or roughly 15.8% of total annual sales.[1] That is a small reduction from $890 billion in 2024 (16.9% rate) but still represents one of the largest cost centers most retailers manage.[2] Online return rates run higher than store rates; the same NRF analysis puts ecommerce returns at 19.3% of sales in 2025.[3]
Category matters more than channel. Apparel sees the highest rates by a wide margin, often 25% to 40% depending on brand and style mix, driven by sizing variability and the try-it-on nature of the purchase.[10] Electronics run lower at 8% to 10%, but with higher per-unit cost and a higher fraud rate because resale value is meaningful.[4] Footwear and accessories sit in the middle at 12% to 17%.[4]
Approximate 2025 return rates by category
| Category | Typical return rate | Notes |
|---|---|---|
| Apparel | 25% to 40% | Sizing and fit drive most returns |
| Footwear | 15% to 20% | Sizing and style |
| Accessories | 10% to 15% | Color and fit issues |
| Beauty | 5% to 10% | Often non-returnable for hygiene reasons |
| Consumer electronics | 8% to 12% | Functionality and buyer's remorse |
| Home goods | 8% to 14% | Damage in transit common |
Fraud is a meaningful share of the bill. The NRF estimates 9% of returns in 2025 are fraudulent, with practices like overstated quantities (71%), empty-box returns (65%), and counterfeit decoy returns (64%) accounting for the bulk of the abuse.[1] Retailers responded by investing in AI fraud detection, with 85% reporting some form of AI-driven fraud screening in 2025.[1]
Grading: A, B, C, D, and what happens to each.
A return is not finished when it arrives at the warehouse. It is finished when somebody decides what to do with it. That decision starts with grading. The standard four-tier model (A, B, C, D) sorts returned units by condition and routes each grade to a different disposition.
Grading and typical disposition
| Grade | Condition | Disposition | Typical recovery |
|---|---|---|---|
| A | Like new, original packaging intact, never opened | Restock to inventory | 85% to 100% of cost |
| B | Open box, all components, minor cosmetic wear | Refurb light, sell as open box | 50% to 75% of cost |
| C | Functional but visible wear, missing accessories | Liquidate to wholesale | 20% to 40% of cost |
| D | Non-functional, damaged, missing core components | Recycle, destroy, or scrap | 0% to 10% of cost |
The grade ratio differs by category. Apparel returns are mostly grade A or B (often a change of mind, never worn or worn once). Electronics returns skew lower (grade B or C) because the customer often opened the box, used the item, and decided against it. Furniture and home goods returns frequently arrive grade C or D because of damage in transit.
A meaningful share of returned merchandise (10% to 25% by industry estimates) cannot return to inventory at full price.[5] That is the recoverable lost value, and it is the prize that a working grading process protects. A grade-A unit incorrectly classified as grade C goes to liquidation and recovers 30% of cost instead of 95%. A grade-C unit incorrectly classified as grade A gets restocked, ships to a customer, and comes back as a complaint.
Restock or liquidate: the decision tree.
Each returned unit faces a sequence of questions. The cleanest version of the decision tree looks like this.
- 1
Is it within the return window?
If yes, proceed. If no, route to a supervisor for exception handling. Most policies allow occasional exceptions for high-value customers.
- 2
Is the product physically present and matches the RMA?
Open the box, inspect, scan. Empty box returns and wrong-item returns get flagged for fraud review.
- 3
What grade is the unit?
Apply the grading SOP: A through D based on condition, packaging, and completeness.
- 4
Can the item be restocked?
Grade A returns to inventory. Grade B may go through light refurb, then to an open-box channel.
- 5
If not restockable, can it be refurbished?
If yes and the math works, route to refurb. If the refurb cost exceeds the recovery delta, liquidate instead.
- 6
Liquidate or destroy?
Grade C goes to liquidation channels. Grade D is destroyed, recycled, or donated depending on category and policy.
- 7
Issue refund or credit?
Final disposition triggers the refund. The customer is told the refund is processed within an SLA.
Refurb workflows that actually scale.
Refurbishment is not a single workflow. It is a set of workflows by product type. Apparel refurb might be steam, inspect, and re-bag. Consumer electronics refurb is functional test, factory reset, replace missing accessories, repackage. Furniture refurb might be wipe down, replace a cushion, and re-box for an open-box channel. The labor profile and tooling differ by category.
For brands running their own refurb, the operating discipline is to track refurb recovery per SKU. Some SKUs have a strong open-box market. Others recover more by shipping to a wholesaler at a fraction of cost. Knowing which is which is a quarterly analysis, not a one-time decision. Demand patterns shift, and so does the recovery curve.
What a working RMA portal does for the customer and the warehouse.
A return materials authorization (RMA) portal is the customer's entry point into the return process. The good ones do five things: authenticate the order, ask the reason for return (which feeds product feedback), generate a prepaid label or QR code, set expectations for refund timing, and flag the inbound return so the warehouse expects it.
On the operations side, the RMA flag means a return arriving at the dock can be matched to an order in seconds, which compresses the receiving and inspection cycle. Returns that arrive without an RMA take longer because somebody has to find the order, verify the customer, and confirm eligibility before any other step happens. The portal pays for itself on labor inside the first month at modest volume.
Customer-facing return UX is now a competitive surface. UPS reports that easy returns build loyalty and that a meaningful share of consumers will avoid a brand after a painful return experience.[6] The retailers that lead on returns offer a few features in common: in-person drop-off without printing a label (QR code at a partner location), quick refund on the original payment method, and clear status updates by email or SMS through the return cycle.
“A return that just works is an underrated retention tool. The brands that nail it earn a second purchase. The brands that fight it lose the customer.”
Customer-facing return UX in 2025.
The features that matter to customers in a 2025 return experience are mostly small. A self-service portal with clear eligibility (so the customer knows the answer before asking). A choice between a refund to the original payment method and store credit (often with a small bonus to credit). A label or QR code generated immediately. Optional in-person drop-off at a partner network. Status updates as the return moves through receipt, inspection, and refund processing. None of those features are exotic. The difference between brands that do this well and brands that do this poorly is largely execution.
The trend that drove 2024 and 2025 was the rise of paid returns. Many brands began charging $5 to $10 for return shipping or restocking. The economic logic is sound on the margin. The customer impact varies. For brands with strong loyalty and unique product, paid returns reduced casual abuse without much retention impact. For brands competing on convenience, paid returns hurt repeat purchase rates measurably.[7]
Reducing return rates without strangling sales.
The cheapest return is the one that never happens. The work to reduce return rates lives mostly upstream: better product photography, accurate sizing data, customer reviews on the PDP, real fit guides for apparel, and AR try-on tools where the category supports it. Each of those investments shows up as a small reduction in return rate, and small reductions compound at scale.
On the operations side, packaging that protects the product through the round trip is a quiet driver of recovery. A unit that ships out, gets returned, and arrives back in grade-A condition is restockable. A unit that ships out in inadequate packaging, gets returned, and arrives crushed is grade C or D. The packaging cost difference is usually small. The recovery difference is large.
How Warpspeed runs reverse logistics.
Warpspeed runs returns inside the same operating system as outbound fulfillment, which removes the handoff that often creates inventory drift. Inbound returns match against the original order, scan into the WMS, and route to a grading station. Grading uses a standardized SOP per SKU family, with photo capture stored against the return record. Disposition (restock, refurb, liquidate, destroy) is decided per unit, not per order, so a return with one grade-A item and one grade-C item routes both correctly.
The customer-facing portal generates labels, surfaces status, and triggers refund workflows when disposition completes. For brands that want a third-party return network (in-person drop-off without a printed label), the integration is in place. Reporting covers return rate by SKU, by reason code, by customer cohort, so brands can see where returns concentrate and act on the upstream causes.
Need a returns operation that recovers margin and keeps customers.
Talk to our team about RMA portals, grading SOPs, refurb workflows, and customer UX.