Every returns process answers three questions: who touches the parcel, who holds the money while the case is open, and who owns the conversation with the customer. The answers determine how fast a refund can be issued, how much control the business keeps over the product and how much operational work lands on a team that is already handling orders. This comparison sets out six models used by Bangladeshi online sellers, the criteria that separate them and the situations where each one is the better choice.
It deliberately does not build a cost model. The cost per return is covered in a separate guide. Here the question is which operating model fits your volume, product mix and tolerance for customer friction.
The six models
- In-house warehouse returns desk. The customer contacts support, receives an address or a pickup instruction, and the warehouse inspects and restocks the item. Refunds are issued by finance after inspection.
- Courier reverse pickup. The business books a reverse pickup through its courier account. The courier collects from the customer and returns the parcel to the warehouse, where the same inspection and refund steps apply.
- Drop-off or collection point network. The customer takes the parcel to a shop, pharmacy or courier point. The parcel moves in bulk to a returns hub, and the customer is refunded after consolidation.
- Software-led returns portal with your own logistics. The customer raises the return in a branded portal, chooses a reason and prints or shows a label. The portal applies policy rules, but the parcel is still collected and inspected by your team.
- Fully outsourced returns operations. A third party receives the parcel, inspects it against your rules, decides the disposition and triggers the refund or exchange. You receive stock and financial updates.
- Marketplace-native process. The marketplace owns the customer relationship and applies its own return window, evidence rules and refund route. The seller has limited ability to change the workflow.
Criteria that decide the comparison
- Customer effort: how many steps, calls and trips are required before the customer receives a decision.
- Inspection control: whether your own team decides condition, resale value and fraud risk.
- Refund speed: the time from a validated return to money reaching the customer, including provider settlement.
- COD and MFS fit: whether the model can handle cash-on-delivery refusals, bKash or Nagad refunds and original-channel rules.
- Inventory accuracy: how quickly returned stock is graded, restocked or written off.
- Operating load: the support, logistics and finance hours the model consumes.
- Data ownership: whether you can see the reason codes, inspection outcomes and customer history.
- Scalability: what breaks first when volume doubles during a campaign or a festive peak.
Comparison at a glance
| Model | Customer effort | Inspection control | Refund speed | COD and MFS fit | Operating load |
|---|---|---|---|---|---|
| In-house warehouse desk | High: customer arranges or follows up | Highest | Depends on warehouse queue | Strong, if cash handling is controlled | High |
| Courier reverse pickup | Medium: one pickup booking | High | Depends on courier return time | Good, but COD needs a separate cash rule | Medium to high |
| Drop-off network | Medium: customer travels to a point | Medium: inspection is delayed | Slow because parcels are consolidated | Good for cash, weaker for prepaid | Medium |
| Software-led portal | Low: self-service request | High if your warehouse inspects | Fast decision, physical return still takes time | Good, with rules per channel | Medium |
| Outsourced returns | Low to medium | Low to medium: depends on the contract | Fast if the provider is authorised to refund | Variable, needs explicit rules | Low |
| Marketplace-native | Lowest for the customer | Low | Set by the marketplace | Handled by the platform | Low, but little control |
Where each model breaks
In-house warehouse desk breaks when the same small team handles receiving, inspection, customer messages and restocking. Parcels sit unopened, refunds age and the customer calls again. It works when volume is modest and one person owns the queue with a daily cut-off.
Courier reverse pickup breaks on failed pickups and address confusion. The customer waits at home, the courier cannot find the address and the case is reopened. Confirm the pickup window, the phone number and the parcel condition before the courier is booked. A reverse pickup also creates a second logistics cost that is easy to overlook.
Drop-off networks reduce failed pickups but add travel for the customer. They suit low-value, non-fragile items in dense urban areas. They are weak for bulky goods, high-value electronics and customers outside the network's coverage.
Software-led portals improve the request and communication trail, but they do not move a parcel. If the warehouse inspection remains manual, the portal only makes the queue more visible. The model works when policy rules are clear, reason codes are enforced and the portal is connected to inventory and finance.
Outsourced returns can be efficient at scale, but the contract decides the outcome. A provider paid per processed parcel has little incentive to reject a weak claim or to flag a counterfeit return. Agree inspection standards, photograph requirements, disposition rules, refund authority, reporting and the process for disputed decisions.
Marketplace-native returns are not a choice if the marketplace sets the rules. The seller's job is to keep evidence, meet the platform deadline and reconcile the deduction. Sellers who also run their own store should not copy the marketplace policy blindly, because the economics and the customer relationship are different.
Choose by context
- In-house desk when volume is low, products need specialist inspection and the warehouse can meet a daily cut-off.
- Courier reverse pickup when customers are spread across cities and the courier has reliable coverage in your delivery areas.
- Drop-off network when the catalogue is low-value and customers are concentrated near collection points.
- Software-led portal when support is drowning in status questions and you want policy enforced before a parcel moves.
- Outsourced returns when returns are a distraction from core operations and you can write a precise service contract.
- Marketplace-native when the platform is your main channel and you have no practical way to override its process.
Score the choice, then pilot it
Weight the criteria for your business. A fashion seller with high return rates should weight customer effort and refund speed heavily. An electronics seller should weight inspection control and fraud evidence. A low-margin seller should weight operating load and inventory recovery.
Run a 30-day pilot in one category or one city. Record return request volume, pickup success rate, days to inspection, days to refund, restock rate, write-off rate and support contacts per return. Compare the pilot with the current process on the same measures. A model that looks cheaper on a proposal often loses on failed pickups and repeated customer contact.
Whichever model you choose, keep the reason code, the inspection evidence and the refund reference in one record. That record is what makes the next comparison possible, and it is the first thing an auditor or a payment provider will ask for.
If you need the workflow designed around your courier, payment and inventory reality, start with Commerce Systems. The pricing page explains how discovery, build and support are scoped. For the cost per return, read the ecommerce returns and refund cost guide.
Sources
- Shopify, Admin GraphQL API: Return object, accessed 17 November 2026.
- WooCommerce, Refunds documentation, accessed 17 November 2026.
- Stripe, Refunds documentation, accessed 17 November 2026.
- Bangladesh Bank, Mobile Financial Services statistics, accessed 17 November 2026.
Responses (0)
No comments yet. Add the first one below.
Leave a response
Your email address will not be published. Required fields are marked *.