Returns are not a post-purchase problem. They are a pre-margin problem wearing a post-purchase hat. By the time the parcel comes back, the ad click has been paid, the fulfillment cost has happened, the customer service ticket exists and the SKU margin is already sulking in the corner.
Marketplace returns forecasting helps teams predict where return-driven margin leakage will appear before spend and stock are committed. It links profit analytics, returns margin leakage, advertising, marketplace analytics, contribution margin and marketplace profit forecasting.
Why return forecasting belongs in the growth meeting
Most teams review returns after finance closes the month. Growth already moved budget. Inventory already reordered. The campaign already scaled. Forecasting moves the signal earlier. If a SKU has a rising return rate, weak review language, size complaints and a campaign pushing it harder, the problem is visible before the return report lands.
In 2026, this matters because marketplace growth is increasingly paid, seasonal and operationally fragile. CPCs spike around retail events, stock shifts between Amazon, bol, Mirakl and DTC, and consumers expect easy returns. A forecast that uses gross sales but ignores expected refunds, non-recoverable fees, handling cost and resale loss is basically a spreadsheet doing motivational speaking.
| Signal | What it predicts | Action |
|---|---|---|
| Return rate by SKU | Margin leakage risk | Cap ads or adjust price |
| Review themes | Reason for return | Fix content or product detail page |
| Campaign mix | Return-prone demand source | Shift budget away from weak audiences |
| Fulfillment cost | Cost of every failed order | Change stock and shipping rules |
Build the forecast at SKU level
Channel averages are too blunt. Forecast returns per SKU, marketplace, campaign and fulfillment method. Start with historical return rate, then weight the forecast with seasonality, category, reviews, content changes, delivery promise, promotion depth and ad mix.
forecast_kept_units = expected_orders * (1 - forecast_return_rate)
return_adjusted_revenue = gross_revenue - expected_refunds
margin_after_returns = revenue - COGS - fees - fulfillment - return_cost - ad_spendThe important output is kept revenue, not ordered revenue. A campaign that creates €50,000 in attributed sales but loses €8,000 to refunds and handling cost should not be reviewed like a €50,000 win. That would be very tidy and very wrong.
The 2026 returns forecasting framework
| Layer | Inputs | Forecast decision |
|---|---|---|
| Demand | Sessions, conversion, ranking, seasonality | Expected orders and units |
| Return behavior | Historical rate, reason code, category, country | Expected refunds and handling cost |
| Economics | COGS, commission, fulfillment, resale value | Contribution margin after returns |
| Media | ACOS, TACoS, campaign role, CPC trend | Budget guardrails and bid ceilings |
| Operations | Stock cover, delivery promise, packaging, content | Scale, fix or pause |
Segment by reason code, not only rate
A twelve percent return rate can mean several different things. Size or fit returns point to content, measurement or assortment issues. Damaged returns point to packaging, carrier or supplier quality. Wrong item returns point to warehouse controls. Buyer remorse may point to pricing, expectation setting or product-market fit.
Normalize reason codes across marketplaces into practical buckets: damaged, wrong item, size or fit, not as described, compatibility, delivery issue, buyer remorse and unknown. Then forecast those buckets separately for high-volume SKUs. The fix for damaged goods is not the same as the fix for “does not match photo.” One needs operations. The other needs content. Both need an owner.
Connect returns to advertising before scaling
Advertising can amplify return risk by pushing the wrong traffic to the wrong product. Review every promoted SKU with TACoS and ROAS, ACOS, return-adjusted revenue and contribution margin. If returns erase the margin, lower bids or cap budgets until the product, content or fulfillment issue is fixed.
| Signal | Healthy | Action when weak |
|---|---|---|
| High ROAS + low return rate | Demand is likely profitable | Test budget scale |
| High ROAS + high return rate | Revenue may be overstated | Lower bids until root cause is fixed |
| Low ROAS + low return rate | Traffic or conversion issue | Improve targeting, price or content |
| Low ROAS + high return rate | Double margin leak | Pause, diagnose, relaunch carefully |
Use scenarios, not one heroic forecast
Returns are uncertain, so one forecast is too confident. Build a base case from recent return behavior, an upside case where content and QA fixes reduce avoidable returns, and a risk case where promotion traffic or a new marketplace increases returns. If the risk case turns contribution margin negative, the SKU should not receive uncapped advertising or aggressive replenishment. Growth that only works in the happy forecast is not a plan; it is a mood.
Turn the forecast into an operating rhythm
Review the top SKUs by return-adjusted margin, not only by return rate. A small high-margin product with eight percent returns may matter less than a hero SKU where four extra return points erase thousands in margin. Assign actions weekly: content updates, packaging checks, carrier changes, campaign caps, price changes or inventory adjustments.
FAQ
How far ahead should marketplace teams forecast returns?
Use four to eight weeks for operating decisions and a twelve-month view for budgeting, buying and category planning.
Should returns be forecast at category or SKU level?
Start at SKU level for products that drive revenue, ad spend or return cost. Roll up to category only after the SKU model is clean.
How do returns affect advertising targets?
Expected returns lower kept revenue and contribution margin, so break-even ACOS and target ACOS should be stricter for return-heavy products.
What data do you need?
Orders, refunds, reason codes, handling cost, marketplace fees, product cost, ad spend, stock, reviews and a changelog for content or operational fixes.
Can FiveX help with returns forecasting?
Yes. FiveX connects marketplace, advertising and profit data so teams can forecast return-adjusted margin and act before returns eat the plan.
Make returns forecastable, not surprising
Returns will never disappear. But they can stop being a monthly surprise party for your P&L. If you want a clearer view of return-adjusted growth across marketplaces, talk to FiveX. We will help you connect returns, ads, stock and contribution margin into one operating dashboard.