Edge Commerce JournalInfrastructure for the storefront era

Oct 28, 2026 - Edge Commerce Journal

What a Return Label Really Costs Per Order

The package came back marked insufficient address. The street was right, the apartment number was one digit off, and by the time support sorted it out the buyer had already asked for a refund instead of a reship.

That is an address validation problem, not a carrier problem. Shopify Help Center documentation covers address collection and validation settings at checkout, and Baymard Institute checkout research keeps finding the same pattern: buyers will fix a flagged field in the moment, but they will not fight a support thread three days later. Related reading: How Edge Personalization Works Without Slowing the Storefront Down.

Price a return as a chain, not a label. The carrier charge is set by dimensional weight as often as by scale weight, inspection and restocking take paid minutes at the bench, and a share of what comes back can only be sold at a markdown or not sold at all. Add those three to the label before you judge whether free returns are a generosity or a slow leak. Most stores that run the full chain find the label was roughly a third of the true cost, and the other two thirds were sitting in operations where nobody had priced them.

Dimensional weight prices the box you wish you had not used

Carriers price many parcels on dimensional weight: the box volume converted to a weight equivalent, billed against the scale weight, whichever is higher. Apparel returns in oversized outbound boxes are the classic case. The garment weighs almost nothing, the box bills like a brick, and the return label inherits the same geometry because the customer reuses the box you sent.

The fix is boring and measurable. A smaller return mailer inside the outbound box, or a poly mailer offered at the returns portal, changes the billed weight on every soft-goods return. Pull 50 recent return labels, compare billed weight to scale weight, and the size of the problem stops being an opinion.

Inspection minutes are a wage bill with a queue attached

Every returned unit gets opened, checked, steamed or cleaned, refolded, rebagged, and scanned back into stock, or routed to a reject pile. Time it honestly at your bench: even a brisk four minutes per unit, at a loaded warehouse wage, is a real per-order cost that never appears next to the label price in any dashboard. Peak season makes it worse, because the queue delays resale while the item is still in season.

Grade strictly at intake and record the grade. A unit graded resellable on day two and a unit graded resellable on day twenty are both called available eventually, but only one of them had a fair chance of selling at full price. The delay is a cost even when no invoice ever shows it.

The markdown line nobody assigns to returns

Some share of returns comes back worn, incomplete, missing tags, or past the point where full price is credible. That share varies wildly by category and by how the policy is worded, which is exactly why you should measure your own instead of borrowing a benchmark. Take one month of returns, count units by outcome, full price resale, markdown, outlet or donation, and write off, and multiply each by its realised price gap.

Stores are often surprised which category carries the loss. It is rarely the cheap basics. It is the high ticket item bought in two sizes with one always coming back, a bracketing pattern the returns data names clearly once outcomes are counted instead of assumed.

Decide the policy from the chain, not the label

Free returns, customer paid labels, and exchange first flows are all defensible. What is not defensible is choosing between them using only the carrier invoice. Build the per-order figure properly: label plus average inspection minutes plus expected markdown loss, divided across all orders shipped, not just the ones returned. That is the number a policy change has to move.

Then test one lever at a time. An exchange first prompt, a smaller return mailer, a stricter grading rule. Each lever moves a different line of the chain, and if you change all three in the same month you will never know which one paid.

Keep reading: How Edge Personalization Works Without Slowing the Storefront Down, Edge Observability: Seeing Storefront Errors Where They Actually Happen, and Image and Font Delivery at the Edge: The Quiet Speed Win cover the nearby cases in this series.

Frequently asked questions

Why is my return label billed for more than the product weighs?

Dimensional weight. Carriers convert box volume into a weight equivalent and bill the higher of that figure and the scale weight. Oversized boxes make light products bill heavily, on the way out and again on the way back.

How do I work out my true cost per return?

Add the label charge, the inspection and restocking minutes at a loaded wage, and the average markdown or write off loss per returned unit. Use your own grading outcomes for a month rather than an industry benchmark, because the markdown share is category specific.

Does offering free returns always lose money?

No. The question is whether the conversion and basket effect covers the full chain cost, not just the label. Stores with low markdown loss and fast restocking can carry free returns comfortably. Stores guessing at the chain cost cannot tell either way.

What is the fastest lever to cut return costs?

Usually the box. A right sized return mailer cuts dimensional weight on every return immediately, with no policy change and no customer friction. Bench timing and grading discipline come next, and policy wording last.