Accounting For Return Rate Drag In An RPV-To-CM Forecast For Fashion And Apparel
A 4% RPV lift on an apparel store rarely translates to a 4% CM lift. Here's how to apply the return-rate haircut so your forecast holds up after refunds, reverse logistics, and restocking write-offs.
Quick answer
On an apparel store running 20–40% return rates, a gross RPV lift typically lands as 60–80c on the euro of contribution margin after refunds, reverse logistics, and restocking write-offs. Haircut the RPV win by (1 − return_rate) for revenue, then subtract reverse-logistics cost per returned order and the write-off share of returned units before applying gross margin.
Return-rate drag in an RPV-to-CM forecast (apparel)
The CM forecast adjustment that converts a gross RPV lift into net contribution margin after apparel return economics.
Return-rate drag is the gap between a test's measured RPV (revenue per visitor) lift at checkout and the contribution margin (CM) that lift actually generates once apparel returns are processed. For fashion and apparel, where category return rates sit between 20% and 40% and footwear/swim run higher, the gap is large enough that ignoring it routinely overstates forecasted CM by 25–40%. The haircut combines four inputs: the category return rate, the reverse-logistics cost per returned order, the restocking/regrade/write-off mix on returned units, and any behavioral effects (bracketing, serial returners) that the change itself induces.
If you forecast in revenue, you can stop reading. If you forecast in contribution margin — which is the only number that pays for your media budget — return-rate drag is the single biggest reason apparel test forecasts miss.
The mechanics are simple, but the inputs are category-specific. A 4% RPV lift on a dress retailer with 32% returns and €11 reverse-logistics per returned order does not behave the same as a 4% lift on a t-shirt brand with 18% returns and €4 reverse-logistics.
Why apparel return rates change the math
Apparel return rates run 20–40% because size-and-fit uncertainty drives most of the basket. Customers order two sizes, keep one, and the second unit ships back at your cost. Footwear and swim push higher — often 35–50% — because fit tolerance is even tighter.
Crucially, the return rate is not just a revenue haircut. Every returned order incurs reverse-logistics cost (pickup, inbound shipping, inspection, repackaging) that hits CM directly, and a meaningful share of returned units cannot be resold at full price — they get regraded, outletted, or written off entirely.
The over-promise pattern
A growth team forecasts "+€480k annual CM from a 4% RPV lift" on a €5M store at 35% gross margin. They forgot the 28% return rate, the €9 reverse-logistics per return, and the 8% write-off on returned units. Real number: ~€280k. The variance gets blamed on the test six months later — when it was a forecasting error from day one.
The four-input haircut
Net revenue per visitor = gross RPV × (1 − return_rate). That gets you to the revenue you actually keep. It's the easy part.
From net revenue, subtract reverse-logistics cost per returned order multiplied by returned-order count. Then haircut the inventory line for the write-off share — typically 5–12% of returned units on full-price fashion, higher on swim and seasonal. What remains, multiplied by gross margin, is your true CM.
A change that lifts conversion by adding more sizes to cart (a fit-confidence widget, a size recommender, or any nudge that encourages bracketing) inflates the return rate input itself. Model that second-order effect separately or your post-launch CM will trail your pre-launch forecast by a wide margin.
Category benchmarks for the haircut inputs
Typical return economics by apparel sub-category (online, EU/UK)
| Sub-category | Return rate | Reverse-logistics per return | Write-off share of returned units | Gross RPV → CM retention |
|---|---|---|---|---|
| T-shirts & basics | 15–22% | €4–7 | 3–6% | 78–84% |
| Denim & trousers | 25–35% | €7–10 | 5–9% | 65–72% |
| Dresses & occasion | 28–40% | €8–12 | 7–12% | 58–68% |
| Outerwear | 20–28% | €10–15 | 4–8% | 68–76% |
| Footwear | 30–45% | €9–13 | 6–10% | 55–65% |
| Swimwear | 35–50% | €6–9 | 10–18% | 48–60% |
| Lingerie & intimates | 12–20% | €4–6 | 15–25% | 62–72% |
Use these as defaults only if you don't have your own data. The "CM retention" column is the rough share of a gross RPV lift that survives to CM once you apply that category's return rate, reverse-logistics cost, and write-off share at a typical 50–55% gross margin.
Flat haircut vs SKU-weighted haircut
If the test moves traffic across the whole catalog evenly — a header change, a checkout tweak — a flat category return rate is fine. If the test concentrates lift on specific SKUs (a PDP redesign tested on dresses, a recommender pushing footwear), you need a SKU-weighted haircut or you'll over-credit categories with worse return economics.
The shortcut: pull the top-20 SKUs by incremental units from the test, weight each by its own 90-day return rate, and use that blended rate instead of the store average. The flat-vs-SKU-weighted difference is often 4–7 percentage points of return rate, which moves CM forecasts by tens of thousands of euros at this revenue band.
Behavioral effects most forecasts miss
Bracketing — ordering multiple sizes intending to return some — and serial returners (the ~5% of customers who drive ~30% of return volume) bend the return rate non-linearly with conversion lifts. A change that increases AOV by adding a second size to cart can lift the return rate by 3–5 absolute points on its own.
Free-return policy changes are the other input that quietly resets your forecast. Moving from paid to free returns typically lifts return rates 4–8 points within a quarter; moving the other way compresses them but also depresses conversion. Either way, your historical return rate stops being a valid input the day the policy changes.
Frequently asked questions
For mixed apparel, 28% is a reasonable default; 32–35% if you skew dresses/occasion, 20–22% if you skew basics. Pull the trailing 90 days of refunded order value over gross order value as soon as you can — defaults always overstate or understate by category mix.
Both, in sequence. First haircut gross revenue by (1 − return_rate) to get net revenue. Then subtract reverse-logistics cost and write-off impact. Then apply gross margin to get CM. Applying the haircut only at the CM line double-counts in some cases and under-counts in others.
Sum your 3PL return-processing fees, inbound carrier costs, and a labor estimate for inspection/restocking, then divide by returned-order count for the period. Most apparel operators land at €6–12 per returned order; the per-return reverse-logistics cost guide breaks down the line items if you want to refine it.
Yes if your test touches PDP, size charts, or fit content — size-and-fit returns are 50–70% of apparel return volume and respond to UX changes, while defects and changed-mind returns don't. Bucketing them lets you forecast which test families can actually reduce returns.
Not in apparel. AOV lifts that come from bracketing inflate return rate even if the headline number looks stable in the first weeks. Hold the test live for at least one full return window (typically 30–45 days post-purchase) before locking the return-rate input.
Moving to free returns typically adds 4–8 absolute points to return rate within 60–90 days, with most of the lift in bracketing and serial-returner behavior. Forecast both the conversion lift and the return-rate lift; the net CM impact is often slightly negative for higher-AOV fashion and slightly positive for basics.
Yes — for site-wide changes (checkout, header, search) where the lift distributes proportionally across the catalog, a flat category-average return rate is fine. SKU-weighted is required when the test concentrates on specific categories, like a PDP redesign tested only on dresses or a recommender tuned for footwear.
At least 45 days post-purchase for the last cohort in the test, because most apparel returns land in the 14–35 day window. Closing the forecast at day 14 systematically overstates CM by 30–50% because you've only seen the early-return tail.
5–8% for full-price basics and outerwear, 8–12% for dresses and occasion wear, 10–18% for swim and intimates. Seasonal SKUs returned outside their season can effectively write off 100% — track in-season vs out-of-season returns separately if seasonality is a material part of your mix.
It's the apparel-specific layer on top of the general annual CM forecast methodology. The general framework assumes a constant return rate and a single reverse-logistics line; the apparel version splits return economics by sub-category and adds the behavioral inputs (bracketing, policy changes) that fashion operators can't ignore.
Test ideas before you ship them
Run unlimited A/B tests, attach hypotheses to outcomes, and build a searchable archive of what works — and what doesn't.