Online womenswear generates returns at a rate that would be commercially fatal in most other retail categories. The cause is a sizing system that gives shoppers no way to predict fit.
Sizing carries no fixed measurements
Women's sizes are ordinal labels rather than dimensions. A given number corresponds to a different set of body measurements at every brand, and often between ranges within one brand.
Menswear retains measurement-based sizing in several key categories, including shirt necks and trouser waists, so a shopper can predict fit without trying the garment on.
Without that anchor, the only reliable way to establish fit is to have the garment in hand, which is precisely what a return exists to allow.
Bracketing is a rational response
Shoppers order two or three sizes of the same item intending to keep one. From the shopper's perspective this converts an uncertain purchase into a certain one.
Free returns make the strategy costless to the individual, so the behaviour spreads as soon as a retailer offers the policy.
The consequence is that a large share of a retailer's despatched units were never intended to be kept, which distorts every operational figure downstream.
Handling a return costs more than shipping it
The visible cost is the return postage. The larger cost is labour, since each item must be inspected, assessed, repackaged and put back into stock.
Items returned after a style has passed its selling window go to markdown or clearance regardless of condition, losing most of their margin.
Garments that cannot be resold as new are handled again through liquidation channels, adding cost to an item that has already lost its value.
The costs get built into prices
Retailers set margins knowing a predictable share of units will come back. That expected loss is priced into every unit sold in the category.
Which means shoppers who order one size and keep it subsidise those who bracket, and the policy quietly transfers cost between customers.
Several retailers have responded by charging for returns or restricting serial returners, both of which trade sales volume against the cost of processing.
Better fit data is the only structural fix
Detailed garment measurements, model dimensions and fit feedback from previous buyers reduce uncertainty more than any change to the returns policy.
Virtual fitting tools address the same problem, though they depend on accurate garment data, which many suppliers do not produce consistently.
Until sizing is anchored to measurements a shopper can compare against a tape measure, returns will remain a structural feature of the category rather than a fixable inefficiency.