Small-business outsourcing research

How long do customer returns take to reach a refund or exchange decision?

A protocol for measuring the wait between a customer return and a recorded refund, exchange, or store credit decision.

How long do customer returns take to reach a refund or exchange decision?

Research question and scope

This protocol asks how long a customer return waits before the shop records a refund, an exchange, or a store credit decision. The unit is one return, not one customer and not one item. The study measures the share decided inside the shop's stated target, the median wait, and the returns that remain open. It does not judge whether the decision was the right one for the business.

Set the scope to one shop, one return counter, and one written return rule. Record the target time, the approval levels, and the date the rule took effect. If the shop changes its payment tool during the study, keep the original process for returns already open and analyze later returns separately. A new tool can change the decision speed without changing staff behavior.

Define the population as every return recorded during the window, including returns that were refused and returns that were later reversed. State any exclusion, such as an internal warranty repair that is not a customer return, and apply it to every return. Test the definition on ten closed returns before collection begins. The reader should be able to rebuild the population from the return log.

Definitions that make the measure reproducible

Define the clock start as the time the return was recorded at the counter, not the time the customer first mentioned the problem. Define the clock stop as the time a refund, exchange, or credit was recorded with a date. A note that the return was received is not a decision. Record both the return time and the decision time so the wait can be rebuilt.

Define a decision as one of three outcomes: money returned, a replacement item issued, or a credit recorded with a balance. Treat an approval given but not yet processed as still open. Treat a return moved to a repair or warranty path as a new case with its own start time. A definition that stops at the conversation rather than the record will understate the wait.

Evidence standard and data dictionary

For each return, capture the return identifier, sale identifier, customer, item, amount, reason, received time, decision time, outcome, approver, and staff member. Add the payment method and whether the customer received a written confirmation. Keep the original sale and the return entry linked. Record any return that was refused and the reason given to the customer.

Capture the reason for a delay when it is documented: a missing approval, an item sent for inspection, a missing receipt, or a payment tool outage. Do not infer a reason from an empty field. Mark unknown reasons as unknown and report the count. A guessed reason hides the step that actually needs repair.

Sampling and collection

Collect every return in the window rather than a sample, because the rule applies to each return. If the shop runs several counters or locations, collect each separately and record the group. A combined rate can hide one counter that never decides on the same day. Report the number of returns and the value behind every percentage.

Draw times from the return log and the payment record, not from a spreadsheet maintained after the fact. If either record is incomplete, say so and reduce the claim. Record the export date, the filters, and any return that could not be matched to a sale. A missing sale link is itself a finding about the routine.

Analysis and reporting

Report the share decided inside the target, the share decided later, and the share still open at the end of the window. Show the median and range of hours from return to decision. Present the results by counter, outcome, and return reason. Use the same denominator for every rate.

Separate a return that waited for an approval from one that waited because nobody processed the payment. An approval gap is a different finding from an unread queue. Report both because they lead to different fixes. A shop may accept a day of waiting for a high-value return but should not accept a return that nobody owns.

Limits of the evidence

The return log can show when the decision was recorded, but it cannot show whether the customer was told. A recorded refund is not a received refund. The protocol also cannot capture a return handled at the counter and never entered in the log. State that limitation before presenting the rate.

A small shop may have too few returns to support a stable trend. Describe the return count and the item mix instead of ranking staff. Do not compare a shop that inspects every item with one that accepts the customer's word. Report the return rule and the approval levels beside the rate.

What to change

Give the result to the owner and the counter lead and ask them to choose one observable change. The change might be an approval threshold written at the counter, a daily open-return check, a named return owner per shift, or a decision message sent to the customer. Record the owner, the effective date, and the measure that will show whether the change worked. Repeat the protocol with the same definitions.

If the on-time rate rises but refused returns also rise, the team may be deciding faster by refusing more requests. If both fall but credits grow, the shop may be issuing credit to close the clock. Report the movements together before declaring success. Close the review when the owner accepts the finding and the next collection date is set.

Sources

Frequently asked questions

Does a favorable sample prove the routine caused the result?

No. The design is observational and supports a bounded operating decision, not a universal causal claim.

What should an owner review first?

Start with high-consequence defects, missing denominators, exceptions, and cases that required the owner to reconstruct the evidence.

Related research