Returns management is the fulfillment flow run backward, and most sellers design it never: outbound gets a process, inbound gets a corner of the bench. Multichannel raises the stakes because every platform brings its own returns contract, and every mishandled return quietly corrupts the one stock pool all your channels are promising from.
The platform contracts, briefly
Each channel binds you differently, Amazon’s returns policies (and FBA’s automated handling), eBay’s returns settings interacting with seller standing, Etsy’s shop-policy-driven returns, and your own store’s stated terms. The details shift; the operational conclusion does not: you cannot run per-platform processes. You run ONE internal flow strict enough to satisfy the strictest platform, and map each channel’s paperwork onto it.
The internal flow: four steps, recorded
1. Receive. Returns get receiving discipline: logged against the original order on arrival, same day. The unlogged return box is where shrinkage-shaped mysteries are born.
2. Inspect. Sellable as new, sellable as open-box/used (where the channel permits), repairable, or write-off. One person’s honest judgment, thirty seconds, photo when it matters (damage claims, dispute evidence).
3. Disposition. Each verdict has a lane: restock to the pool, relist in a condition-appropriate channel (open-box on eBay is a legitimate lane), repair queue, or documented write-off. Deciding lanes in advance is what makes step two fast.
4. Restock, into the SYNCED pool. The unit re-enters the same pool every channel sells from, immediately, so availability reflects it everywhere in seconds. A returned unit sitting uncounted is a self-inflicted stockout; one restocked without inspection is a future complaint wearing a fresh label.
The money mechanics
- Refund timing versus restock timing are different events, refund per platform rules, restock only after inspection. Systems that couple them force bad choices.
- Fees behave asymmetrically: some platform fees return on refund, some do not, Etsy’s stack and marketplace commissions each have their own rules. High-return categories should price the asymmetry into channel markups.
- Return shipping ownership is a policy lever per channel; whoever pays, the label cost belongs in your returns-rate math.
Reading returns as data
A returns log with reason codes is free product research:
- Reason clusters by SKU: “not as described” clusters mean listing copy or photos lie somewhere, fix the listing, not the customer.
- Damage clusters point at packaging or a carrier lane.
- Size/fit clusters (apparel) justify size-guide work and pre-empt the next hundred returns.
- Serial returners exist; platforms have abuse processes, use them with documentation rather than policy tantrums.
Returns rate per SKU also belongs in reorder math: a 20 percent return rate on a SKU effectively raises its available supply and lowers its true velocity, both worth knowing before the next purchase order.
Common questions
How fast must a return re-enter availability?
Same day as inspection. The gap between receive and restock is inventory that exists physically but not commercially, waste in slow motion.
Should returned units go back to the same channel they came from?
Sellable-as-new units return to the shared pool (all channels). Condition-downgraded units go to the channel that sells that condition honestly, open-box lanes exist for this.
How do FBA returns fit?
FBA returns land in Amazon’s process and back into FBA supply or removal orders, keep that pool separate from your merchant pool, as always.
What returns rate is normal?
Category-dependent (apparel high, hardware low). Your trend matters more than the benchmark: a rising rate on a stable catalog is a specific problem with a start date, find it in the reason codes.
Close the loop, keep the counts
Returns restocked into one synced pool, visible on every channel in seconds, with the sync history that keeps drift impossible. From $49/month with unlimited orders. See pricing.