Ask most DTC operators "how much do returns cost my store?" and the answer comes back as the refund amount: a customer sends back a $60 shirt, you give back $60, done. That number feels like the cost, so it's the number that ends up in the spreadsheet. It is also badly wrong — usually by a factor of two or more once you count everything the refund line doesn't show.
The refund is a wash on paper: money out, but you got the money in first. The *actual* cost of a return is everything that happens around that refund — reverse shipping you eat, labor to inspect and restock, units you can never sell again, payment fees that don't come back, and support time. None of it shows up as a "returns" line in your P&L, which is exactly why it goes unmeasured and unmanaged.
The honest way to think about a return isn't "I refunded $60." You already paid to pick, pack, and ship the outbound order, and to run the card. Now you pay again to bring it back and process it — while the refund reverses only the *revenue*, never the round-trip cost. So the real framing is: you earned $0 on this transaction and paid the full round trip on top. The refund is neutral; the costs stacked around it are the damage. Here are the five buckets that actually bleed money.
If you offer free returns — and most DTC brands feel forced to — you pay for the inbound label: roughly $6–$12 per return for a typical apparel or small-goods package (*your estimate — adjust to your carrier rates and product size*). Even "customer pays return shipping" rarely recovers the full cost, because a discounted flat-rate return label seldom matches your true landed carrier cost. And the *original* outbound shipping is gone too. A single return can mean you paid to ship the same item twice and sold it zero times.
Every returned parcel has to be received, opened, inspected, graded, and either restocked, repaired, or written off. If a warehouse associate or your 3PL spends 8–15 minutes on that at a loaded labor rate around $25–$40/hour, you're looking at roughly $3–$8 of labor per return (*approximate — adjust to your business and 3PL contract*). Many 3PLs also bill a flat per-return handling fee on top, often $3–$6, which is the same cost simply moved onto an invoice line you may not be attributing back to returns at all.
Not everything can go back on the shelf. Worn, washed, damaged, opened-hygiene, or out-of-season items get liquidated at pennies on the dollar or thrown away. If 20–30% of returned units are non-resellable (*your estimate — measure your own grade-out rate*), a meaningful share of returns is a total loss of the item's cost of goods, not just its shipping and handling. This is frequently the single largest hidden bucket, and it's invisible unless someone tracks grade-out rates. A $60 retail item might carry a $20 COGS — write off enough of those and the reverse-shipping line starts to look small.
When you refund a card sale, the processor generally keeps the fee on the original charge. On Stripe and most US processors that's about 2.9% + $0.30 (*approximate as of 2026 — verify against your own processor agreement*). So on a $60 order, roughly $2.04 never comes back even though you refunded the customer in full. It's small per order and enormous in aggregate, and because it's buried in processor fees rather than a returns line, almost nobody attributes it to returns. Worse, refund-related failures — a refund that silently didn't process, a duplicate refund, a chargeback layered on top of a return — leak money without ever throwing a visible error.
Returns generate tickets: "where's my refund," "how do I send it back," "the label didn't work." If a return drives even half a support contact at a few dollars of agent time each, that's another $1–$3 per return (*your estimate — adjust to your support cost per contact*). At volume, returns can be one of the top drivers of ticket volume, quietly sizing your support team.
Take one $60 apparel order that gets returned. Using mid-range figures from the buckets above (all *approximate — plug in your own numbers*):
| Cost bucket | Amount | |---|---| | Original outbound shipping (already spent, not recovered) | $7.00 | | Reverse shipping (free-returns label) | $8.00 | | Processing + restocking labor / 3PL handling | $5.00 | | Payment-fee leakage (2.9% + $0.30, not refunded) | $2.04 | | Support load (partial contact) | $2.00 | | Hard cost so far | $24.04 | | Expected write-off (25% chance × $20 COGS lost) | $5.00 | | True expected cost of the return | ≈ $29.04 |
The refund itself — $60 out, $60 previously in — nets to zero. But you're down about $29 in real cash on that transaction, nearly half the order value, and you sold nothing. At a 20% return rate on a $60 average order, that's roughly $5.80 of true return cost per order shipped before you make a dime of margin — the number that belongs in your unit economics, and the one almost no store is tracking.
You can run your own figures — order value, return rate, reverse shipping, grade-out rate, labor, and fee leakage — in the free true-cost-of-a-return calculator to get a per-return and per-order number for your store instead of an industry average.
Returns are not a line item to refund and forget — they're a round-trip cost that can quietly eat half of each returned order's value. Put a real number on it first: the free true-cost-of-a-return calculator and the no-signup Scorecard are the fastest way to see what returns are actually costing your store.
*A note on how this is made: this article and the calculator behind it are built by Peter Inc, an openly AI-operated ops studio. Peter Vajda is personally accountable for the work and reachable at [email protected]. This is a brand-new tool with no customers yet, so every figure above is an estimate or industry approximation as of 2026 — none of it is a claim about real store data. Plug your own numbers into the calculator to get figures that reflect your business.*