Two thirds of refunded orders wrote to you first
64% of refunds on five UK marketplaces had a customer conversation before the money moved. That conversation is the window in which a replacement panel beats a return, and the number tells you whether you are using it.
One number that says where the leak is
Take every refunded order and ask one question: did the customer write to you before the refund? If most did, you had a chance to offer something other than money back, and the refund rate is a reply-time problem. If most did not, the refunds are happening in the operator’s own systems, and the fix is upstream: packaging, carrier, listing.
Across five UK marketplaces and 90 days the answer was 64%. Two thirds of the money went out after a conversation the seller was part of.
Wrote first, by marketplace
Marketplace E is the interesting row. Its customers almost never write before a refund, because that operator runs returns through its own portal and its own support team. The seller learns about the refund from the order data, days later, with a reason code and nothing else. On that marketplace the refund rate is a product and logistics number only; no reply speed will move it.
What the refunds were for
“Change of mind” dominates where the customer writes first, and “item returned” dominates where they do not. That is the same fact seen from two sides: a customer who talks to you tells you why; a customer who uses the portal ticks a box.
The window, and what fits in it
The conversation before a refund is short. In this data the median gap between the first customer message and the refund was under three days, and on Marketplace B a quarter of refunds followed within 24 hours of the message. Whatever you can offer has to be offered in the first reply.
- Damage or missing part: a replacement component, sent by post, with a photo request in the same message. Cheapest outcome by a wide margin. The sums.
- Delivery failed: a rebooked slot with the carrier’s tracking quoted, before the customer asks for their money.
- Change of mind before dispatch: let it go; a cancellation before shipping is the cheap refund.
- Change of mind after delivery: a partial refund to keep the item is often cheaper than collection plus write-down, and some operators let you offer it in the thread.
The silent third
36% of refunds happened with no conversation. Nothing in the inbox will change those. What changes them is upstream: the packaging that survives a two-man courier, the listing image that matches the colour, the delivery promise the carrier can keep, and the operator’s returns policy you agreed to. Track the silent share per marketplace; if it grows, the operator changed its returns flow, not your customers.
- 64% of refunded orders had a conversation first; that is the addressable share.
- It ranges from 5% to 79% by marketplace, depending on whether the operator routes returns to the seller or to its own portal.
- The window is short: median under three days; the offer has to be in the first reply.
- The silent third is a packaging, carrier, listing and policy problem, not an inbox problem.
Questions people ask
How many refunds could have been avoided with a faster reply?
Not all of them, but the addressable set is the refunds where the customer wrote first. In 90 days of data from one UK seller that was 64% of refunded orders. Within that group, damage and missing-parts refunds are the ones a replacement part or partial refund most often resolves; change-of-mind refunds mostly are not avoidable.
Why do some marketplace refunds happen without any customer message?
Because the operator refunds under its own policy: returns initiated through the marketplace's returns portal, failed deliveries returned to depot, or cancellations before dispatch. On operators that route returns through their own flow, the seller sees the refund in the order data but never sees a message. Those refunds cannot be steered, only prevented upstream.