Fewer returns, smaller footprint: the sustainability case for try-on

Most sustainability initiatives in fashion cost money and are worth doing anyway. Reducing fit-driven returns is unusual: it removes journeys, removes handling and removes write-offs, and the finance team is on your side from the first conversation. That makes it worth getting right, and worth describing carefully rather than loudly.

A return is not the reverse of a sale

It is a new set of journeys, and usually more of them than people assume. The parcel goes out. The shopper drops it at a collection point or waits for a courier. It travels to a consolidation point, then to a returns-processing facility that is often not the warehouse it shipped from. If the shopper wanted a different size, a replacement ships too.

The legs that matter most are the residential ones. A single parcel collected from a home address is the least efficient movement in the entire chain — low density, unpredictable routing, frequently a failed attempt. Trunk freight between warehouses is comparatively efficient per unit; the last mile, repeated, is not.

Then there is the processing itself: opening, inspecting, steaming or laundering, repackaging in new materials, re-listing. Each of those is labour and consumables that the original sale already paid for once.

What actually happens to returned clothes

Returned garments are graded, and the grade decides the outcome. Some go back to A-grade stock and sell again at full price. Some go to outlet or open-box channels at a reduced margin. Some are sold in bulk to liquidators. Some are written off.

The deciding factor is arithmetic, not intent. If handling a returned item costs more than the item can realistically be resold for, the commercially rational outcome is to stop spending money on it — which is how low-value garments end up in bulk channels or waste streams despite being perfectly wearable. Nobody sets out to destroy stock. The processing cost simply overtakes the residual value, and for a lot of fast-moving product it does so quickly.

Regulators in several markets have started restricting the destruction of unsold clothing, and the direction of travel there is one-way. Whatever the specific rules end up being, a business that generates fewer returned units has fewer of these decisions to make.

Why fit-driven returns are the ones worth attacking

Because they are avoidable in a way that other returns are not. Roughly half of apparel returns are size or fit related, per industry aggregates from the NRF and Narvar. A shopper who changed their mind made a real decision; a shopper who ordered two sizes intending to send one back never wanted two.

That second case compounds. A bracketed order books an extra outbound parcel and a guaranteed return leg at the moment the basket is filled, before anything ships. Removing the reason to bracket removes both journeys at once, which is why the fit half of the returns pie is worth more attention than its share suggests.

The honest limits

Virtual try-on is not an offset and should never be described as one. It does nothing about how a garment was produced, what it was made from, or the impact of the first delivery — which is the delivery that happens whether or not the purchase was a good one.

It also has a footprint of its own. Generating images uses compute, and a shopper who renders twenty looks has consumed real energy. It is small next to a van, but it is not zero, and a company that pretends otherwise invites the scrutiny it deserves.

The credible claim is narrow and defensible: fewer avoidable return journeys, measured as returns avoided against a control. Not a tonnage figure derived from a spreadsheet of industry averages you did not collect. If you cannot show the workings, do not publish the number.

Why shoppers reward the measured version

Vague environmental language has been used so heavily that it now reads as a warning sign. Specificity is the differentiator: a brand that says “fit-related returns on these categories fell by this much over this period, measured this way” is making a claim that can be checked, and claims that can be checked are the ones that survive contact with a sceptical customer.

Practically, that means publishing the return-rate delta rather than a carbon estimate, naming the period and the product set, and saying plainly what you did not measure. Under-claiming a real result is a stronger position than over-claiming a modelled one.

A sensible sequence

1. 01Establish your baseline fit-driven return rate by category, using reason codes rather than a single store-wide average.
2. 02Enable try-on on the categories with the highest fit-driven returns and hold a comparable set as a control.
3. 03Wait for the return window to close before reading anything, then compare cohorts of orders rather than calendar months.
4. 04Report returns avoided as a count, not a carbon figure, and state the measurement method alongside it.
5. 05Only translate that into environmental language once you can show which orders it came from.

None of this is dramatic, which is rather the point. The environmental case for try-on is the same as the commercial one, described honestly: fewer returns, for a reason you can name.