Amazon return policy 2026: refunds, claims and seller deadlines
Amazon customer refunds, FBM SAFE-T claims and FBA inventory reimbursements follow different rules. First-scan refunds can happen before inspection. Track the case type, trigger event and applicable deadline instead of applying one clock to every return.
The customer has been refunded, the warehouse has not finished inspection, and the team is searching for a single Amazon return deadline. There is no single clock that answers all three problems.
For sellers and support teams, separate the customer’s refund from a seller reimbursement claim. Then identify whether the order was seller-fulfilled or fulfilled by Amazon. This prevents a valid rule from being applied to the wrong workflow.
Reviewed September 6, 2026 for Amazon US. Check the current notice and eligibility requirements for the individual case before acting.
Do FBM sellers have four days before every refund?
No. Amazon’s refund-processing announcement and staff clarification give sellers four calendar days after receiving relevant returns to process a refund from January 26, 2026. Refund at First Scan remains in place, so eligible cases can be refunded before warehouse inspection.
Treat the receipt event and refund status as separate records. Before issuing money manually, check whether Amazon has already refunded the buyer. A parcel arriving at the warehouse does not mean the refund is still awaiting your decision.
Keep weekend coverage in mind where the applicable period uses calendar days. Set an internal reminder earlier than the platform deadline and label it as your operational target, not an Amazon rule.
Which claim process applies after a refund?
FBM SAFE-T and FBA inventory reimbursement are different processes. Do not route them into a generic “claim within 60 days” queue.
Amazon’s US SAFE-T announcement sets a 30-day filing period from February 16, 2026, with the trigger depending on return receipt and refund timing; lost-in-transit cases have a separate trigger. Eligibility still matters. Review the announcement and the case record rather than assuming every refunded order qualifies.
Record the refund date, carrier events, warehouse receipt and evidence of any discrepancy. If the item returned differs from what was shipped, the person preparing a claim needs the relevant records, not a general note saying “customer abuse.”
Why can an FBA reimbursement be below the sale price?
For inventory lost or damaged before a customer order, Amazon moved to a manufacturing-cost basis. The official timing update set March 31, 2025 as the effective date. Amazon’s policy explanation distinguishes manufacturing cost from selling price and other expenses.
A lower reimbursement therefore does not by itself prove a processing mistake. Check the event type and cost basis before challenging the amount. Keep the relevant cost evidence with the inventory or finance owner, rather than asking frontline support to reconstruct it.
Amazon also publishes different FBA claim windows by event type. Confirm the specific route for warehouse loss, customer returns or removals. Do not generalize one window across them.
What should your return tracker contain?
Use a record that supports the next decision:
| Field | Why it matters |
|---|---|
| Order and fulfillment type | Distinguishes FBM from FBA work |
| Case or claim type | Identifies the applicable process |
| Trigger event and timestamp | Determines which clock has started |
| Refund status | Prevents duplicate action |
| Deadline and official rule | Keeps the instruction auditable |
| Evidence and owner | Makes the next action possible |
This is a suggested operations checklist, not a replacement for Seller Central. Check discrepancies against the platform record and preserve the relevant notification.
How can support and the warehouse avoid working at cross-purposes?
Agree how receipt and condition findings reach the case owner. Preserve relevant packaging and condition evidence before disposal when appropriate. Make an exception visible if a parcel cannot be matched to an order.
Use responsive support handoffs for customer updates and return-cause analysis for recurring operational problems. Start by reviewing one recent refund and tracing which events happened before the team noticed them. Fix that visibility gap before adding more reminders to the wrong deadline.