We use Google Analytics 4 to see which pages are useful. It loads only if you accept it, and then sets two cookies, _ga and _ga_C372PEK7KT. Nothing is stored until you choose, and nothing it records is used for advertising or sold.
Poor sizing costs fashion brands far more than returns — it distorts every line of the P&L from overproduction to markdown.
TL;DR — Poor sizing is not a customer experience problem. It is a structural P&L problem. Add up overproduction, returns processing, inventory markdown, customer service load and environmental write-down and the total lands somewhere between a fifth and a third of gross revenue for a ready-to-wear brand. That is our own model, and the stack that produces it is set out below so it can be argued with. Most brands measure only the returns line. The other costs remain invisible — until a made-to-order comparison makes them impossible to ignore.
Poor sizing creates costs that most fashion brands do not measure — and therefore cannot manage. Every fashion operator has a returns line in the P&L. Fewer have mapped what sits before it: overproduction locked into a size distribution nobody validated, inventory capital tied to units that will never sell through, customer service agents fielding questions that should never have arisen. Poor sizing is the single largest driver of operational waste in ready-to-wear fashion.
The full cost is distributed across the supply chain in ways that make individual line items look unrelated. Returns processing is visible. The costs upstream and downstream of it are less so. This article maps each one — with numbers.
Definition
Sizing-driven operational cost
Any cost that would not have been incurred if the garment had been produced to the buyer's verified body measurements at the point of order. This includes pre-production overproduction hedges, in-season inventory carrying costs, post-sale returns processing, and post-return markdown or destruction costs. In aggregate, these costs are structurally larger than the returns processing cost alone.
Overproduction by size is the primary upstream cost of poor sizing, and it stems from a single structural problem: brands cannot predict which sizes their customers actually need. Size uncertainty forces brands to hedge production across a broad distribution, typically six to eight size brackets per SKU, committing capital to units that may never find a buyer.
A European jacket brand stocking sizes 46–54 illustrates the distortion clearly. Sales concentrate in the middle of that range — sizes 48–50 — not because those sizes represent where actual customer bodies cluster, but because buyers who fall outside the range have already stopped trying. The brand reads this as demand signal and doubles down on the middle. The extremes keep accumulating as slow-moving or dead stock.
In practice, production across a wide size range means capital committed to units that will sell slowly or not at all. It means warehouse space occupied by slow-moving inventory for months per season. Under the EU Ecodesign for Sustainable Products Regulation 2024/1781, it increasingly means compliance risk for goods manufactured and eventually destroyed — with reporting obligations that make the destruction cost visible in ways it previously was not.
Inventory distortion — the mismatch between which sizes are in stock and which sizes buyers actually want — is the in-period cost of poor sizing. Sizes that fit poorly accumulate as unsold stock while sizes that correspond to the cut's best-fit bodies sell through quickly. The result is a perpetual demand mismatch that no reorder strategy can fully correct.
This distortion is costly in direct terms: capital tied up in stock that is not selling. The volume is large enough to be a line of its own: clothing returns run into the billions each year in the UK market alone, and fit is the reason most often given. That figure captures only the return event — not the overstock that never reached a buyer in the first place.
The indirect costs are larger. Reorder cycles are disrupted when available sizes do not match incoming demand. The product mix in later-season inventory does not reflect customer preference — it reflects which sizes happened to fit the brand's block. Promotions become necessary to clear stock that accumulated for sizing reasons, compressing blended selling price and degrading brand positioning for the following season.
When measuring inventory health across a full season, the most predictive leading indicator of end-of-season markdown depth is not sell-through rate on the best-selling size — it is the spread between the best-selling and worst-selling sizes. Wide spreads signal a sizing distribution problem, not a demand problem.
Sizing questions generate disproportionate customer service volume relative to any other product attribute. Sizing and fit are the leading driver of pre-purchase customer service contacts in fashion, ahead of delivery, availability and product queries combined. Each contact that could have been eliminated by accurate fit information represents pure cost with no revenue upside.
In practice, a single sizing-related customer service journey spans multiple touchpoints. A customer unsure of their size contacts the brand before purchase. If they buy and the garment fits poorly, they contact again with a complaint. They initiate a return — requiring agent time to process. The returned garment requires inspection, then a refund or exchange must be issued. Each of those steps costs agent time, and a returned luxury item costs considerably more to put through that sequence than a standard e-commerce return does.
For luxury brands with high-touch service models, this cost is compounded. A customer service interaction about a size question costs agent time, potentially specialist advisor time, and delays the original transaction while creating no new value. At a house with an average transaction value above €2,000, a 20-minute advisor conversation to resolve a fit complaint represents a meaningful percentage of margin on that sale.
Returned luxury garments face a severity of post-return cost that lower price-point segments do not. The expectation of newness — implicit in every luxury transaction — means a returned garment, even in pristine condition, has been owned. In many market contexts, it cannot ethically or legally be re-sold as new. The brand faces three options, none of them good.
WRAP puts the value of clothing going into UK landfill at around £140 million a year. That is disposal at the end of a garment's life rather than deliberate destruction of stock, but the two meet in the same skip: a returned unit that cannot be resold has taken the full environmental cost of production and delivered none of the use.
Sizing inadequacy's costs do not sit in parallel — they compound. Overproduction creates a larger base of inventory that can generate returns. A higher return volume creates more markdown exposure. Markdown at depth distorts pricing signals for the following season. Distorted pricing leads to production decisions that perpetuate overproduction. Each cycle reinforces the next.
The compounding is why returns reduction alone does not solve the problem. A brand that reduces returns by 10% but keeps the same production distribution will simply shift cost: fewer post-sale costs, but the same pre-sale overproduction. The structural fix requires addressing the origin point — the moment where garment specifications meet buyer measurement data.
When garments are produced to known, verified measurements — as in a made-to-order operation — overproduction by size becomes structurally unnecessary. There is no size distribution to hedge because every unit has a named buyer with verified dimensions. In our own made-to-order operation, where every garment is cut to a measured body, fit returns are rare enough not to be a planning variable. That is the comparison that makes the ready-to-wear numbers look like a choice rather than a constant.
The operational case for made-to-order is not primarily about craft or luxury positioning. It is about eliminating an entire category of P&L cost that ready-to-wear brands have normalized as unavoidable. It is not unavoidable. It is a consequence of producing garments without knowing who will wear them.
Map all the sizing-driven costs against revenue and the total is larger than most finance teams expect, because only one line of it is usually reported. The stack below is our own model for an EU mid-to-premium ready-to-wear brand; each band is a range because each depends on category and price point.
Aggregated, the stack lands between roughly a fifth and a third of gross revenue. The returns line, the one that typically appears in board reporting, is at most half of it. The rest is distributed across production, logistics, and G&A in ways that obscure the root cause.
A made-to-order operation with verified measurement data at point of order eliminates the upstream cost category almost entirely. No size distribution hedge is required. No overproduction by size. No inventory carrying cost on slow-moving sizes because there are no speculative size positions. Customer service volume attributable to fit uncertainty drops to near zero for customers with validated fit profiles.
The residual cost in a made-to-order model is production cost per unit — which is higher per garment than volume production — and a small residual return rate from genuine quality issues or changed requirements. In aggregate the economics favour made-to-order once average order value is high enough for the saved overproduction to outweigh the higher unit cost. Where exactly that threshold sits depends on the category and on what the measurement infrastructure costs to run, but it is well inside the range a premium brand already operates in.
The fit data that makes this possible — the measurement profiles, fit history, and garment specifications — is what the Size Passport infrastructure is designed to capture, maintain, and make portable across garment categories and brands. When a customer's measurements travel with them across orders, the sizing guess is removed from the equation permanently.
Published breakdowns place fit and sizing behind somewhere between 40% and 70% of apparel returns, depending on category and price point, with the proportion at the upper end in high-precision categories such as structured jackets and tailored trousers. It is the largest single return reason in every breakdown of the kind we have seen, ahead of quality defects, delivery damage and buyer's remorse. What it is not is one published number, so the share used in the modelling below is a stated assumption and not a finding.
General overproduction refers to producing more total units than the market will absorb at a given price. Overproduction by size refers specifically to producing units in sizes that the brand's customer base does not actually need — a structural mismatch between the size distribution in production and the size distribution in the real body population. The latter is more common, more preventable, and more directly tied to sizing data quality than total production volume decisions.
Digital size recommendation tools, including AI-based size advisors and fit finders, reduce pre-purchase uncertainty and lower return rates, and the published trials that measure them find the size of the effect depends almost entirely on how good the body data behind them is. They do not, however, address the upstream overproduction problem: they help buyers choose from an existing size range, but they do not eliminate the need to produce speculatively across that range in the first place. The structural solution requires producing to known measurements, not recommending the best available approximation.
A Size Passport is a verified, portable measurement profile linked to a specific individual — containing body measurements, fit preferences, and garment outcome history. When a brand produces a garment against a customer's Size Passport data rather than a generic size, overproduction by size is eliminated for that order, return risk falls to near zero for fit-related reasons, and customer service contacts related to sizing uncertainty do not occur. The cost saving per order is structural, not probabilistic.
Yes, materially. The EU Ecodesign for Sustainable Products Regulation 2024/1781 introduces reporting obligations and eventual restrictions on the destruction of unsold consumer goods, including textiles. As enforcement tightens, the write-down-and-destroy option — currently the default for sizing-driven excess inventory — becomes both more costly and more legally constrained. The regulatory trajectory favors production models that generate less excess, with made-to-order representing the clearest structural response.
Sources
Related concepts