Apparel is the category where the order is not the sale. A customer orders a dress in two sizes and a top in three colors, keeps one item, and returns the rest; the brand paid to ship five items and will pay to receive four back, inspect them, repackage three, mark down one that came back creased, and refund the customer in full. Return rates of 20 to 40 percent are normal in online apparel; the margin the brand reported when the order shipped is not the margin it has, and the gap is usually the difference between a profitable brand and a growing one that is not.
What does a return actually cost?
Fully loaded, per return. Outbound shipping on the returned item — paid once already; return shipping, which the brand pays under free-return policies and the customer pays under paid ones; receiving and inspection labor; repackaging and the new packaging; restocking into inventory; the refund's payment-processing cost, which is often not refunded by the processor; customer-service time; and the recovery loss — the share of returned units that cannot be resold at full price because they are damaged, worn, out of season, or in a broken size run, times the markdown or write-off on those units.
The number in most brands. US$12 to US$25 per return before recovery loss, and US$18 to US$40 after it, on an item that may have sold for US$68; and the return produces no revenue — the original sale was reversed.
Recovery rate by category. The share of returned units restocked as new, by product type; knitwear and lightweight fabrics recover worse than denim and outerwear; items returned after the season's end recover at markdown; the recovery rate by category is the input to the loaded cost and is measured, not assumed.
What is bracketing, and what does it do to orders?
Bracketing is ordering multiple sizes or colors with the intent to return most of them. Encouraged by free shipping, free returns, and uncertain sizing; it inflates gross orders, average order value, and conversion, and produces returns at rates of 50 to 70 percent on bracketed orders; a brand whose dashboard shows average order value rising may be seeing bracketing rising.
Measuring it. Orders with multiple sizes of the same style; orders where the return share exceeds a threshold; the bracketing rate by customer cohort and by product; and the contribution of bracketed orders after returns, which is routinely negative.
Reducing it without losing the sale. Size guides and fit tools that work, fit feedback from past returns surfaced on the product page, reviews with size context, a size-exchange flow that is easier than a return, and policy signals — a return fee above a threshold, or free returns only on the first item — tested against conversion.
How are returns by product, size, and reason the fix list?
Returns by style, size, and color, with reason codes. Too small, too large, fit not as expected, quality, damaged, not as pictured, changed mind, arrived late; captured at the return and read monthly; a style with a 45 percent return rate driven by "too small" is a grading problem; one driven by "not as pictured" is a photography and color problem; one driven by "quality" is a supplier problem; each has an owner and a fix.
The size curve and the return curve together. Returns by size against the size curve of what was bought and sold show whether the brand is cutting the wrong sizes, grading inconsistently across styles, or sizing differently from the customer's expectation; the fix feeds the next season's grading and buy.
Returns as a product-development input. The styles that return above the category average are reviewed before they are reordered; a style with high sales and high returns is sometimes a style to fix and sometimes a style to drop, and net margin after returns is the deciding number.
Why is the return policy a margin decision?
The policy sets the return rate and the conversion rate at once. Free returns raise conversion and raise returns; paid returns lower both; a return window of 14, 30, or 60 days, exchanges-only on sale items, a restocking fee on certain categories, and store credit versus refund — each moves conversion, returns, and loaded cost, and the net effect is modeled, not assumed.
Modeling the policy. The brand's own data on conversion and return rate under the current policy; tests of policy variants on a segment — a return fee for orders below a threshold, free exchanges with paid refunds — with conversion, return rate, loaded return cost, and net margin per order measured; the policy chosen for net margin, with the customer-experience cost weighed explicitly.
Serial returners. A small share of customers generate a large share of returns and often a negative lifetime contribution; identifying them, applying policy limits where the terms allow, and excluding them from acquisition lookalikes is a margin program most brands have not run.
What is net margin after returns by style?
Shipped margin less loaded return cost, by style. Net revenue after refunds; product cost on units kept; outbound shipping and fulfillment on all units shipped; the loaded return cost on units returned including recovery loss; marketing attributed to the style; the result by style, ranked, with the return rate beside it.
What it usually shows. The best-selling style by shipped revenue is often in the bottom third after returns, because it is the one most bracketed and most returned; a style with modest sales and a 12 percent return rate often out-earns it; the ranking reshapes the buy, the marketing, and the product-development priorities.
How do returns fit the cash and inventory picture?
Returns are inventory arriving back out of season. Units returned in October from a September order land in inventory as the season ends; their recovery is at markdown; the inventory plan carries the expected return flow by style and the recovery value, and the markdown reserve covers it.
Refund timing and cash. Refunds are issued within days of receipt; a brand with a 30 percent return rate on a US$1.2 million month will refund roughly US$360,000 over the following weeks; the 13-week forecast carries refunds as a disbursement by week from the return curve, not as a reduction of revenue in the month of sale.
The reserve for returns on the balance sheet. Revenue recognized on shipment is overstated by the returns to come; a returns reserve from the brand's own return curve, recorded at sale and released as returns arrive, keeps reported margin honest and prevents the month-end surprise.
Why is this CFO work?
The loaded return cost by category requires logistics, labor, processing, and recovery data joined per return; bracketing measurement requires order-level analysis; returns by product and reason are a data discipline that feeds product development; the policy model is a conversion and margin test; net margin after returns by style is the ranking that decides the buy; and the returns reserve and the refund cash curve are balance-sheet and forecast work. The merchandising and product teams own the fit, the photography, and the assortment; the CFO owns the economics of the return.
Worked example
An online apparel brand, US$18.3 million net revenue after refunds, 31 percent return rate, free shipping and free returns.
| Metric | Found |
|---|---|
| Return rate — overall; dresses; knitwear; denim | 31%; 42%; 36%; 19% |
| Loaded cost per return (before recovery loss) | US$16.80 |
| Recovery rate — restocked as new | 71% overall; 58% knitwear |
| Loaded cost per return including recovery loss | US$27.40 |
| Annual loaded return cost | ≈ US$2.6 million |
| Bracketing — orders with multiple sizes of one style; their return rate | 23% of orders; 64% |
| Top reason codes | Too small 34%; fit not as expected 22%; changed mind 18%; not as pictured 11% |
| Best-selling dress — shipped margin vs. net margin after returns | 61% vs. 14% (48% return rate, bracketed heavily) |
| Mid-volume blazer — shipped vs. net after returns | 58% vs. 44% (13% return rate) |
| Returns reserve on the balance sheet | None; revenue recognized at shipment |
| Policy test (segment, 8 weeks) | Conversion | Return rate | Net margin / order |
|---|---|---|---|
| Current: free shipping, free returns | 2.9% | 31% | US$9.40 |
| Free exchanges; US$6 return fee for refunds | 2.7% | 24% | US$13.10 |
| Return fee on orders with 3+ sizes of one style | 2.85% | 26% | US$12.20 |
Illustrative figures for a hypothetical organization; not a client's data.
The brand's hero dress drove the marketing and lost most of its margin to bracketing and a "too small" reason code that traced to a grading error across two sizes; the blazer nobody promoted was its most profitable style after returns. The plan: the dress re-graded and its fit guide rebuilt; free exchanges with a refund fee adopted after the test; a serial-returner program covering 2 percent of customers who generated 14 percent of returns; returns by reason reviewed monthly with product development; a returns reserve recorded at sale; and the next season's buy ranked on net margin after returns. Return rate modeled to 24 percent; loaded return cost down about US$700,000; reported margin within two points of the final for the first time.