Key Takeaways
- Return volume is not a staffing ratio because inspection effort varies by product, condition, and disposition.
- Forecast intake, physical processing, customer contact, refund administration, and exceptions as separate workloads.
- Refund speed depends on both operational processing and the payment network, so teams should timestamp each stage.
- Small businesses should use their own units per labor hour and exception rate instead of borrowing a large-retailer headcount.
- A faster, clearer return can protect repeat purchase intent, but fraud and policy exceptions still need controlled review.
Returns work is easy to underestimate because the customer sees one request while the business may see several tasks. A returned item can require authorization, a label or drop-off code, carrier tracking, receipt, inspection, disposition, a refund, inventory updates, and customer follow-up. Some products go straight back to stock. Others need testing, cleaning, repackaging, vendor credit, liquidation, recycling, or fraud review.
The ecommerce returns processing staffing statistics below come from a 2025 U.S. consumer and large-merchant study, federal wage data, vendor research, and a named logistics case study. They measure different populations. None supplies a universal employees-per-thousand-returns ratio, so the staffing section shows how to build one from a store's own handling data.
For related workflows, see the ecommerce virtual assistant guide and ecommerce virtual assistant services.
Returns staffing benchmarks at a glance
| Measure | Published finding | Scope and caveat |
|---|---|---|
| Online return rate | 19.3% of online sales | NRF and Happy Returns 2025 estimate; sales value is not the same as units or orders |
| All-retail return rate | 15.8% of annual sales | 2025 estimate from large U.S. merchants |
| Holiday return rate | 17% of holiday sales | Merchant expectation for the 2025 winter holiday season |
| Seasonal staffing | 43% planned to hire seasonal returns staff | Surveyed U.S. merchants had more than $500 million in revenue |
| Third-party logistics | 49% planned more focus on 3PL partners | Holiday returns and fraud management plans among surveyed merchants |
| Processing cost | 30% of item price on average | Optoro 2024 consumer report; a vendor estimate, not a public accounting standard |
| Instant refund preference | 76% were more likely to choose instant refund or exchange | U.S. consumers who made an online return in the prior year |
| Retention risk | 71% were less likely to shop again after a poor return experience | Same 2025 consumer study |
| Refund initiation | Instant at a Return Bar | Happy Returns service rule; bank posting may take up to seven business days |
| Dock-to-stock example | 48 hours | UPS case study for Overstock, not a cross-industry benchmark |
These numbers answer different questions. Return rates estimate sales exposure. The seasonal hiring figure describes a planned response among large retailers. The cost percentage combines more than payroll, while the 48-hour result comes from one outsourced operation. A useful staffing plan keeps those definitions separate.
Online returns create more work than the headline rate shows
NRF and Happy Returns projected $849.9 billion in U.S. retail returns for 2025. Merchants estimated that returns would equal 15.8% of annual retail sales, while the online estimate was 19.3% of online sales. The gap matters to ecommerce teams, but neither percentage says how many parcels will arrive or how many minutes each one will take.
A sales-value return rate can rise because shoppers return a small number of expensive products. A unit return rate can rise because they return many low-cost items. Staffing responds to units and tasks more directly than to dollars. Cost and recovery analysis still needs the sales-value view.
The 2025 research used two surveys conducted during summer 2025. The consumer survey included 2,006 U.S. consumers who had returned an online purchase in the prior 12 months. The merchant survey included 358 ecommerce professionals at U.S. companies with more than $500 million in revenue. That large-company threshold makes the findings useful for market context, but a startup should not copy the reported operating choices as if they came from peer businesses.
Seasonal hiring follows the return wave
Among the large merchants in the NRF and Happy Returns study, 43% planned to hire seasonal staff to handle returns. Another 49% planned to put more focus on third-party logistics partners. Retailers expected 17% of holiday sales to be returned.
The timing differs from outbound fulfillment. Holiday sales create an immediate pick-and-pack peak, but returns build after delivery and can continue through an extended return window. A staffing calendar based only on order date can release temporary labor before the return queue reaches its high point.
Plan the return wave by cohort. For each sales week, apply the store's observed return-request curve and transit time. Then add the likely receipt date and processing service level. The result is a weekly intake forecast rather than one annual percentage.
Processing labor depends on the disposition path
There is no defensible public benchmark that says every ecommerce return takes the same number of labor minutes. Work content changes by category. An unopened standardized item may need a scan and shelf move. An electronic device may need serial-number verification, data removal, functional testing, accessory checks, and secure handling. Apparel can require condition grading, steaming, folding, and repackaging.
The UPS customer story for Overstock describes a dedicated operation that handled 32,000 SKUs and moved returns from dock to stock in 48 hours. UPS says staff decided whether goods went back to the supplier, returned to stock, went to charity, or entered recycling or disposal. The case reported more than 99.9% on-time processing and throughput accuracy, plus $2 million in cost savings over 18 months.
That result is evidence that a defined disposition process can improve speed. It is not a general small-business target. The case does not publish return units, paid labor hours, staffing count, wage mix, or the calculation behind each saving. Without those inputs, it cannot produce a comparable units-per-hour benchmark.
For internal planning, record touch time by disposition:
| Work step | Volume driver | Labor measure |
|---|---|---|
| Authorization and label support | Return requests | Active minutes per request |
| Receiving and identification | Parcels and units received | Units per paid hour |
| Inspection and grading | Units by product class | Minutes per unit by class |
| Refurbish or repackage | Units approved for recovery | Minutes per recovered unit |
| Refund administration | Approved refunds and exceptions | Minutes per transaction |
| Customer follow-up | Contacts and repeat contacts | Active minutes per case |
| Fraud or policy review | Flagged cases | Review minutes and approval rate |
Measure productive work separately from scheduled hours. Training, meetings, breaks, absence, equipment delays, and uneven arrivals reduce the share of a shift available for processing.
Reverse-logistics cost is broader than wages
Optoro's 2024 Returns Unwrapped report says processing a return costs retailers 30% of the item's price on average. Its separate 2024 State of Returns report expressed the estimate as $26.50 per $100 in returned merchandise. These vendor-published estimates are close, but they should not be combined into a more precise figure.
Processing cost can include carrier transportation, receiving labor, inspection, storage, repackaging, markdown, liquidation fees, disposal, customer service, and lost resale value. A labor budget covers only part of that list. It can also miss the cost of delay: seasonal or fast-depreciating inventory may lose recovery value while it waits in the queue.
Use the BLS wage series as a cost input, not a return-processor salary benchmark. In May 2025 the U.S. Occupational Employment and Wage Statistics program reported a median hourly wage of $17.53 for stockers and order fillers. Returns teams can include receiving workers, customer service representatives, technicians, fraud analysts, and supervisors, so one occupation does not represent the whole team. Employer payroll taxes, benefits, facilities, systems, and management also sit outside the published wage.
Refund speed has two clocks
Refund speed begins with the retailer's decision and ends when the customer can see the funds. Those are not the same event. Happy Returns says participating retailers initiate refunds instantly at a Return Bar, while the refunded amount may take up to seven business days to appear, depending on the financial institution.
A dashboard should therefore keep separate timestamps for request approval, carrier acceptance or drop-off, warehouse receipt, inspection completion, refund initiation, and confirmed customer notification. Support staff can explain the stage without promising control over bank posting time.
The 2025 NRF and Happy Returns survey found that 76% of consumers were more likely to choose a return option with an instant refund or exchange. Fast initiation can improve the experience, but the store must set fraud and value thresholds before releasing funds ahead of inspection.
Return experience affects customer retention
In the same survey, 71% of consumers said a poor return experience made them less likely to shop with that retailer again. Four in five said they would tell friends and family about the negative experience. These are stated intentions from people who had completed an online return, not observed repeat-purchase records.
The distinction matters. Survey responses show customer risk, while a merchant's cohort data shows actual retention. Compare repeat purchase rates for customers with no return, a return completed within target, and a late or disputed return. Control for product category, acquisition channel, order value, and customer tenure where the sample allows it.
Staffing can affect retention through queue time and communication. A well-staffed team acknowledges the request, keeps status current, and resolves exceptions before customers have to ask again. A larger team will not fix an unclear policy, poor product information, or a slow physical network on its own.
A workload-based staffing model
Build headcount from observed work rather than the market return rate alone:
forecast returned units = forecast shipped units × observed unit return rate
base labor hours = sum of forecast units by disposition × observed minutes per unit ÷ 60
support hours = forecast return contacts × active minutes per contact ÷ 60
scheduled hours = (base labor hours + support hours + exception hours) ÷ productive share
If a store forecasts 5,000 returned units and its measured weighted handling time is 7 minutes per unit, physical processing requires 583.3 productive hours. If staff spend 75% of scheduled time on measurable return work, the schedule requires 777.8 paid hours. Those are author calculations that illustrate the formula. They are not published industry benchmarks.
Divide paid hours by the scheduled hours available from one worker during the planning period, then round up only after adding interval coverage and supervisory needs. Run at least three scenarios: normal demand, the expected seasonal peak, and a stress case with higher exceptions or slower processing.
For a small team, cross-training can cover uneven volume without assigning a full-time person to every step. Document which tasks an ecommerce virtual assistant may complete, such as checking eligibility, issuing approved labels, updating status, and reconciling standard refunds. Keep policy overrides, high-value refunds, suspected fraud, and regulated or hazardous goods with an authorized employee.
Metrics to review each week
| Metric | Calculation | Staffing use |
|---|---|---|
| Unit return rate | Returned units divided by shipped units | Forecasts physical intake |
| Requests per 100 orders | Return requests divided by orders, multiplied by 100 | Connects sales forecast to case volume |
| Units per productive hour | Completed units divided by direct work hours | Converts intake into labor |
| Exception rate | Returns requiring manual escalation divided by returns | Sizes senior review capacity |
| Dock-to-disposition time | Disposition timestamp minus receipt timestamp | Shows queue and processing delay |
| Refund initiation time | Refund timestamp minus the defined policy start event | Measures controllable refund speed |
| Contacts per return | Return-related contacts divided by returns | Exposes repeat work caused by weak updates |
| Recovery rate | Recovered value divided by original returned value | Links speed and disposition quality to economics |
| Post-return repeat purchase | Customers who buy again divided by customers with a completed return | Tests actual retention by experience cohort |
Publish the denominator next to every rate. A return rate based on units cannot be compared directly with NRF's sales-value estimates. A refund-time median can also hide a long backlog, so add the 90th percentile and oldest open case.
FAQ
What percentage of ecommerce sales are returned?
NRF and Happy Returns estimated that 19.3% of online sales would be returned in 2025. The estimate uses sales value. A store planning receiving labor should also calculate its own returned units as a share of shipped units.
How many people are needed to process ecommerce returns?
No cited source establishes one universal staffing ratio. Forecast returned units, measure handling minutes by product and disposition, add customer-support and exception work, then divide productive hours by the team's observed productive share and scheduled availability.
How fast should an ecommerce refund be processed?
Track the retailer-controlled initiation time separately from payment posting. Happy Returns says participating retailers initiate a refund instantly at its Return Bars, but the customer's bank may take up to seven business days to post it. A store's promise should match its policy, fraud controls, carrier events, and payment provider.
Do easier returns improve retention?
The 2025 NRF and Happy Returns survey found that 71% of consumers were less likely to shop again after a poor return experience. That measures stated intent. Merchants should verify the effect with repeat-purchase data for their own return cohorts.
Should a small ecommerce business hire seasonal returns staff?
Seasonal labor is one option when the forecasted return wave exceeds trained capacity. In the large-merchant NRF survey, 43% planned seasonal returns hiring and 49% planned more focus on 3PL partners. Small businesses should compare temporary hiring, cross-training, and outsourced capacity against their own peak duration, training time, and exception mix.
Sources
- National Retail Federation, 2025 Retail Returns Landscape, October 15, 2025
- National Retail Federation, 2025 returns press release and survey methodology, October 15, 2025
- Optoro, Returns Unwrapped 2024, 2024
- Optoro, The State of Returns 2024, 2024
- UPS Supply Chain Solutions, Overstock customer story, accessed October 8, 2026
- UPS and Happy Returns service and refund FAQ, accessed October 8, 2026
- U.S. Bureau of Labor Statistics, May 2025 Occupational Employment and Wage Statistics, released May 15, 2026
Tags
Ready to put this into practice?
Book a free 15-min match call
Tell us what role you're filling. We'll match you with a pre-vetted virtual assistant - or tell you honestly if we're not the right fit.
Book a free call →