Ecommerce Fulfillment
E-commerce Returns Statistics 2026

On this page
The latest public evidence does not produce one universal 2026 returns rate. It does show why returns remain an operating issue: U.S. online sales are large, parcel networks handle billions of shipments, and returned goods need a controlled path back into inventory, resale, repair, liquidation, or disposal.
E-commerce returns statistics in 2026
The most recent U.S. Census Bureau quarterly estimate available on 2026-08-02 is for the first quarter of 2026. Seasonally adjusted retail e-commerce sales were $326.7 billion, up 2.7% from the fourth quarter of 2025 and 9.8% from the first quarter of 2025. E-commerce represented 16.9% of total retail sales. The Census release schedule placed the second-quarter report on August 18, 2026, so it was not treated as available evidence here.
That sales figure is market context, not a return count. The return estimates below come from separate studies and keep their stated scope.
| Measure | Latest published figure | Period | What it measures |
|---|---|---|---|
| U.S. retail e-commerce sales | $326.7 billion | 1Q 2026 | Seasonally adjusted retail e-commerce sales |
| E-commerce share of retail sales | 16.9% | 1Q 2026 | Share of total retail sales |
| Estimated online sales returned | 19.3% | 2025 forecast | NRF and Happy Returns estimate |
| Estimated total retail returns | $849.9 billion | 2025 forecast | NRF and Happy Returns estimate |
| Appriss total retail returns | $706 billion | 2025 estimate | Appriss benchmark scope |
| U.S. parcel shipments | 23.1 billion | 2025 | Pitney Bowes index |
What the return estimates measure
NRF and Happy Returns projected $849.9 billion in total retail returns for 2025 and estimated that 19.3% of online sales would be returned. The study also reported 15.8% of annual retail sales returned overall and said 9% of all returns were fraudulent. These are retailer survey estimates, not a Census series and not a direct count of packages entering reverse logistics.
Appriss Retail's 2026 Total Retail Loss Benchmark reported $706 billion in returns during 2025 and estimated that 14.2% of that amount, or $100 billion, was preventable loss from returns fraud and abuse. Appriss says its benchmark combines a consumer survey, internal client analytics, and industry data. Its figure and NRF's figure should not be averaged because their methods and constructs differ.
The narrower conclusion is useful: returns are large enough to affect inventory, customer service, transportation, and warehouse control. The evidence does not support a claim that every retailer has a 19.3% return rate or that every returned dollar becomes reverse-logistics cost.
The sales base and parcel network
Pitney Bowes reports 23.1 billion U.S. parcels in 2025, up 3.3% year over year. Its index covers business-to-business, business-to-consumer, consumer-to-business, and consumer-consigned shipments up to 70 pounds. Consumer-to-business shipments matter to returns, but the index does not identify which parcels were returns.
USPS reported 6.8 billion shipping and package pieces in fiscal year 2025. Its annual report gives a detailed total of 6.837 billion pieces, down from 7.252 billion in fiscal 2024. The report includes Parcel Return Service within its parcel categories. USPS volume is a carrier measure, not a national returns measure, and it cannot be added to Pitney Bowes without double counting.
A return may use a label, drop-off point, carrier scan, consolidation route, receiving dock, inspection station, and refund workflow. A shipment count cannot tell an executive how many touches a returned unit requires.
Labor and facility implications
BLS reported 15.4597 million retail-trade employees and 6.6037 million transportation-and-warehousing employees in June 2026, seasonally adjusted. Retail trade employment fell by 7,500 over the month, while transportation and warehousing employment rose by 2,300. These broad series do not isolate returns workers or establish staffing for one operation.
GAO's 2024 review of warehouse and last-mile delivery safety found an estimated serious-injury rate of almost four in 100 full-time workers in those work groups in 2022. GAO reported that serious-injury rates rose 20% for warehouse workers and 23% for last-mile delivery workers from 2018 to 2022. The report is not a returns study, but it matters when returns add lifting, sorting, repacking, walking, and exception handling to a physical process.
Measure the work directly: units received, inspection minutes, repack minutes, exception minutes, storage days, and the share requiring a second decision. External statistics do not supply a staffing ratio.
Reverse logistics is a set of decisions
UPS describes reverse logistics as more than a transportation leg and reports that business-to-business and commercial returns accounted for 35% of reverse-logistics revenue in 2024. Its 2025 Form 10-K reports that UPS logistics revenue declined in 2025 as the mix of forwarding and logistics activity changed. Neither source gives a universal cost per returned item.
FedEx's 2025 e-commerce trends research says consumers expect reliable delivery, proactive updates, proof of delivery, and frictionless returns and resolutions. Deloitte's June 2026 post-purchase survey also treats returns and exchanges as part of the experience after checkout. These sources describe expectations and survey findings, not a causal model for repeat purchase or a required service level.
| Control point | Evidence to record | Operational question |
|---|---|---|
| Authorization | Order, reason, policy, channel | Is the return eligible under the stated rule? |
| Inbound receipt | Carrier event, package ID, date | Did the expected unit arrive? |
| Inspection | SKU, condition, quantity, required images | What condition did the operation receive? |
| Disposition | Restock, repair, resale, liquidation, disposal | Which approved path applies? |
| Refund or exchange | Decision, amount, timestamp, owner | Was the customer outcome released under policy? |
| Inventory update | Location, status, adjustment reference | Can another person trace the stock record? |
The inventory exception workflow is a useful control model for returned units that do not fit the routine rule. Keep return decisions, inventory adjustments, and refund authority separate when the process requires different owners.
Policy, customer data, and fraud controls
The FTC tells online shoppers to check who pays return shipping, the return window, restocking fees, and refund terms before sending an item back. For operators, that points to a data requirement: the policy version used for the transaction needs to remain findable when a dispute or exception is reviewed.
The FTC's 2024 Consumer Sentinel Network Data Book recorded 383,441 reports in the online shopping and negative reviews category, or 5.92% of all reports. This category is broader than returns and includes complaints about online shopping and reviews. It should not be presented as a return-fraud rate.
Appriss's findings on fraud and abuse are not a reason to make every return a fraud investigation. A consistent policy, linked evidence, and an escalation path are more useful than an undocumented judgment at the receiving dock.
What the direction means for operations
The direction is mixed. Census online sales were higher year over year in 1Q 2026. Pitney Bowes parcel volume grew in 2025. USPS shipping and package volume declined in fiscal 2025, while Ground Advantage volume increased. Carrier mix, service choice, and product category can move independently of total retail sales.
Returns may rise because sales grew, product mix changed, a policy widened, or inbound parcels were consolidated differently. The external statistics cannot identify which explanation applies locally.
Use the alternatives to spreadsheet inventory tracking article when the return record needs a reviewable system rather than a shared workbook. The choice of tool does not replace a definition, owner, approval rule, or source record.
Consolidated returns statistics
| Source | Figure | Date or period | Interpretation limit |
|---|---|---|---|
| U.S. Census Bureau | $326.7B e-commerce sales | 1Q 2026 | Sales, not returns |
| U.S. Census Bureau | 16.9% online share | 1Q 2026 | Share of retail sales |
| NRF and Happy Returns | 19.3% online sales returned | 2025 forecast | Survey estimate, not 2026 actual |
| NRF and Happy Returns | $849.9B total returns | 2025 forecast | Different denominator from online-only rate |
| Appriss Retail | $706B returns | 2025 estimate | Benchmark method differs from NRF |
| Appriss Retail | $100B preventable return loss | 2025 estimate | 14.2% of Appriss's return estimate |
| Pitney Bowes | 23.1B parcels | 2025 | Covered U.S. parcel flows |
| USPS | 6.837B shipping and package pieces | FY 2025 | USPS network volume |
| BLS | 15.4597M retail-trade employees | June 2026 | Broad industry employment |
| BLS | 6.6037M transportation and warehousing employees | June 2026 | Broad industry employment |
| GAO | Almost 4 in 100 serious-injury rate | 2022 | Warehouse and last-mile workers |
| FTC | 383,441 online-shopping reports | 2024 | Complaint category, not return count |
Sources
These sources were checked on 2026-08-02. Dates in the article refer to each source's reporting period, not the verification date. No 2Q 2026 Census figure was used because its release schedule placed that report on August 18, 2026.
- U.S. Census Bureau, Quarterly Retail E-Commerce Sales, 1Q 2026 (opens in a new tab)
- U.S. Census Bureau, 2026 retail release schedule (opens in a new tab)
- National Retail Federation, 2025 Retail Returns Landscape (opens in a new tab)
- Appriss Retail, 2026 Total Retail Loss Benchmark Report (opens in a new tab)
- BLS, employment and average weekly earnings by industry, June 2026 (opens in a new tab)
- USPS, total shipping and package volume (opens in a new tab)
- USPS, fiscal 2025 annual report (opens in a new tab)
- Pitney Bowes, Parcel Shipping Index 2026 report (opens in a new tab)
- UPS, 2025 Form 10-K (opens in a new tab)
- UPS, challenges of reverse logistics (opens in a new tab)
- FedEx, 2025 E-Commerce Trends to Watch (opens in a new tab)
- Deloitte, omnichannel post-purchase experience (opens in a new tab)
- FTC, Consumer Sentinel Network Data Book 2024 (opens in a new tab)
- GAO, worker safety and health in warehouse and last-mile delivery (opens in a new tab)
Frequently asked questions
Is 19.3% the 2026 e-commerce return rate?
No. NRF and Happy Returns reported it as a 2025 estimate for online sales. The latest Census e-commerce figure here is for 1Q 2026, and Census does not publish a matching national return-rate series.
Can NRF's $849.9 billion and Appriss's $706 billion be compared?
They can be placed side by side as evidence that the scale is material, but they should not be combined or averaged. Their methods, populations, and definitions differ.
What is the best first operational measure?
Start with received units by channel and elapsed time from first inbound scan to inspection. Add disposition, refund timing, exception age, and recovery value after the definitions are stable.
Related Reading
For a process-level control model, read the article on how to build an inventory exception workflow. The article on alternatives to spreadsheet inventory tracking is also contextual operating guidance, not a substitute for the external statistics above.
Frequently Asked Questions
What is the current e-commerce return rate?
NRF and Happy Returns estimate that 19.3% of online sales will be returned in 2025. That is a forecast for online sales, not a measured 2026 rate, and it should not be blended with studies that use a different denominator.
Why do return estimates differ?
They can cover online sales, all retail sales, selected retailers, or observed client transactions. Period, population, exchanges, and the definition of a return can all change the result.
Does a higher online sales figure prove a higher warehouse return workload?
No. It establishes a larger sales base, not a company-specific return count or staffing requirement. Compare external context with internal receipts, inspection time, disposition, and refund records.
Which returns measures should an operations team track?
Track return authorization volume, received units, days to first scan, days to inspection, disposition rate, refund timing, resale recovery, exception age, and cost by channel. Define each measure before comparison.