The Two Ways to Cost a Return

Both are honest, they answer different questions, and mixing them up is why returns estimates vary so wildly between brands.

On a $72 order with $22 of landed cost, $6.50 of outbound shipping, $8.20 of return postage, $4.50 of handling, a 2.9% processing fee the processor keeps, and 65% of the goods value recovered:

• Out-of-pocket loss: about $29. You are that much worse off than if the order had never been placed.
• Swing versus a kept sale: about $70. That is 98% of the order value — the profit you did not make plus the cost of taking the item back.

Use out-of-pocket for cash and P&L. Use the swing when you are deciding what a return prevention project is worth, because that is what each avoided return actually returns to you.

Where the Money Actually Goes

Return shipping is the most visible cost and rarely the largest. On the worked example:

Raising recovery beats cutting postage. At 65% recovery you write off 35% of landed cost on every return. Getting recovery to 85% saves $4.40 a return here — more than most brands ever negotiate off a return label.

What Returns Do to Break-Even ROAS

This is the connection that makes returns a paid media problem rather than an operations one.

A $68 order, $22 landed, $6.50 fulfilment, 2.9% processing:

• Before returns: contribution margin 55.2%, break-even ROAS 1.81x
• At an 18.5% return rate: contribution margin 37.1%, break-even ROAS 2.69x

That is 18 points of margin and nearly a full point of ROAS, and none of it appears anywhere in the ad platform. A media buyer hitting 2.2x against a target built on pre-return margin believes they are profitable and is not.

The blended figure is straightforward: contribution on a kept order, weighted by the share kept, minus the out-of-pocket loss on a returned one, weighted by the return rate. Mode 3 above does it on your numbers, and the break-even ROAS calculator and contribution margin calculator use the same definitions.

Units or Revenue?

Track both. The gap between them is the useful part.

On the worked example the unit return rate is 18.5% and the revenue return rate is 19.6%, because returned orders average $72 against a $68 AOV. That gap says your more expensive orders come back more often, so any planning done on averages understates the damage.

When the two are close, returns are spread evenly across price points and either measure will do. When they diverge sharply, the problem is concentrated — and it is worth finding out in which lines before doing anything site-wide.

What Actually Reduces Returns

In rough order of how much they move the number for a typical DTC brand:

• Sizing guidance that reflects your actual fit — not a generic chart. Fit feedback from previous buyers, model measurements alongside the size worn, and honest "runs small" flags.
• Imagery that matches what arrives. Colour accuracy, scale references, and texture. A product that photographs better than it looks converts well and comes straight back.
• Recovery rate on returned stock. Not prevention, but it changes the cost of every return you do get, and it is entirely within your control.
• Returns-reason data you actually read. "Wrong size" and "not as described" point at completely different fixes, and most brands collect the field without ever analysing it.
• Delivery speed and condition. Late and damaged arrivals convert into returns at a much higher rate than the product deserves.

The number in mode 4 is the point of all this: if getting from 18.5% to 14% is worth $149,000 a year, that is what you can justify spending to achieve it.

What This Calculator Cannot Tell You

• It uses one average order. Return rates vary enormously by product line — in apparel, by style. A site-wide figure will overspend on the categories that come back and underspend on the ones that do not. Run it per line where you can.
• It does not model the conversion trade-off. Free returns raise both conversion rate and return rate. Making returns harder cuts the return rate and usually cuts sales too. This calculator prices one side of that; only a test prices the other.
• It does not account for repeat purchase. A customer who returns once and buys again for years is not the same as one who returns and leaves. Some return cost is a retention investment.
• It assumes a single recovery rate. In reality returned stock splits across A-stock, discount, liquidation and write-off, and the mix shifts with season and speed of processing. Whatever comes back as A-stock re-enters sellable inventory, which is why a high return rate should feed into your reorder point rather than being treated as a separate problem.
• It does not separate fraud or wardrobing. Serial returners behave differently from genuine sizing problems, and the response is a policy question rather than a margin one.