How to audit marketplace returns in 6 steps

Returns are not just a customer service metric. They are a profit metric wearing a customer service jacket. Very sneaky. A return removes revenue, adds handling cost, distorts advertising performance and often points to a fixable product, content or fulfillment issue.

Step 1: pull returns at SKU level

Start with SKU, marketplace, date, order value, refund value, return reason and quantity. Channel averages are too polite; they hide the products causing trouble.

returns_audit = sku + marketplace + order_date + refund_date + reason + refund_value + handling_cost

Step 2: normalize reason codes

Amazon, bol, Mirakl and other channels use different labels. Group them into practical buckets: damaged, wrong item, size or fit, not as described, delivery issue, buyer remorse and unknown.

Step 3: calculate return-adjusted margin

For each SKU, subtract refunds, return handling, marketplace fees that are not recovered, fulfillment, product cost and ad spend. Then compare contribution margin before and after returns.

Step 4: connect returns to advertising

A campaign can look efficient before returns and painful after returns. Review return-heavy SKUs with ACOS, ROAS and TACoS. If returns erase margin, lower bids until the product is fixed.

Step 5: diagnose the fix

1. Content mismatch: improve titles, images, size charts and compatibility notes.
2. Quality issue: escalate supplier or packaging checks.
3. Delivery issue: review carrier, promise and warehouse process.
4. Buyer remorse: check price, expectation setting and review themes.

Step 6: build the weekly action list

Common pitfalls

• Auditing returns only monthly, after the margin damage is done.
• Looking at return rate without order value and ad spend.
• Ignoring unknown reason codes instead of improving data capture.
• Letting ads scale before product fixes are complete.

What to check before you call it done

• Top 20 return-cost SKUs identified.
• Return-adjusted contribution margin calculated.
• Ad bids reviewed for high-return SKUs.
• Content, quality or fulfillment owner assigned.
• Follow-up date added to the weekly marketplace review.

Code example: simple return-cost flag

You can start with a simple rule before building a full model. Flag SKUs where return cost is more than 20% of contribution margin or where return-adjusted margin falls below your floor.

if return_cost / contribution_margin_before_returns > 0.20:
    flag = "review urgently"
if contribution_margin_after_returns < margin_floor:
    action = "pause ads until fixed"

Segment by marketplace

Do not assume a SKU behaves the same everywhere. A product may have acceptable returns on Amazon, weak returns on bol and expensive returns on a Mirakl marketplace because delivery promises, customer expectations or content requirements differ. Segmenting by marketplace shows whether the problem is product-level or channel-specific.

Prioritize by money, not noise

Sort the audit by total return cost, not only return percentage. A product with a 4% return rate and high order volume can hurt more than a product with a 20% return rate and tiny sales. Build the action list from euro impact first, then use return percentage to understand severity. This keeps the team focused on leaks that matter instead of chasing the loudest-looking metric.

When two SKUs have similar return cost, prioritize the one with active ad spend, low stock or strategic category importance. The goal is not a perfect audit. The goal is fewer preventable returns and cleaner contribution margin next week.

FAQ

What return metric matters most?

Return cost as a share of contribution margin, not return rate alone.

Should ads stop on high-return products?

Not always, but bids should respect return-adjusted margin.

How often should returns be audited?

Weekly for active ad SKUs and monthly for the wider catalog.

Can content reduce returns?

Yes. Clear specs, sizing, compatibility and expectation-setting often reduce avoidable returns.

How does FiveX help?

FiveX connects returns with profit analytics, advertising and marketplace performance so teams fix the leaks that actually cost money.