Why Made-to-Order Could Change Fashion's Inventory Problem

22 octobre 2026

Chart: published estimates of unsold fashion: 2% never reach market (Ellen MacArthur Foundation), 4-9% destroyed before use and 21% unsold on average (European Environment Agency)

VAZZARI FUTURE OF FASHION RESEARCH — 001

By Michael Vazzari, Vazzari Couture · Published 22 October 2026

Fashion has a timing problem.

Most clothing is made months before anyone decides to buy it. The industry produces against a forecast, then spends the season correcting the forecast: markdowns, outlets, warehouses and, at the end of the line, destruction.

Made-to-order flips the sequence. The order comes first. The garment comes second.

This paper looks at what the published evidence actually says about fashion's unsold stock, where the numbers disagree, and what made-to-order can and cannot fix.

KEY FINDINGS

Unsold stock is real, but poorly measured. Published estimates range from 2% of clothing material input lost as garments that never reach market (global, 2015) to an average 21% of textile products remaining unsold in European research literature, because they measure different things. [1][2]
Destruction is the visible tip. The European Environment Agency estimates that 4–9% of textile products put on the European market are destroyed before use: 264,000 to 594,000 tonnes a year. [2]
The scale is measured in billions of items. BoF and McKinsey estimate the industry produced 2.5 to 5 billion items of excess stock in 2023, worth $70 billion to $140 billion in sales. [3]
Policy has moved. Since 19 July 2026, large companies in the EU may no longer destroy unsold clothes, clothing accessories and footwear. [5]
Made-to-order removes the forecast from the finished garment. It does not remove returns, upstream materials or higher unit costs. [2][4]

01. THE PROBLEM IS NOT DEMAND. IT IS SEQUENCE.

Global clothing production approximately doubled in the 15 years to 2015, while clothing utilisation, the average number of times a garment is worn before it ceases to be used, fell by 36%. [1] The Ellen MacArthur Foundation estimates that more than USD 500 billion of value is lost every year through under-used clothing and the lack of recycling, and that less than 1% of textiles produced for clothing is recycled into new clothes. [1]

That is the macro picture. The inventory problem sits inside it.

In a stock-based model, a brand commits to styles, colours, sizes and quantities before the market has spoken. The European Environment Agency puts the logic plainly: producing large quantities and many product types is cheaper per unit, which "leads to the overproduction of product types, styles, colours and sizes", and overproduction and destruction is "deemed acceptable in the textiles industry and even makes economic sense." [2]

The garment is not the failure. The guess is.

02. WHAT THE NUMBERS SAY, AND WHAT THEY DON'T

Ask how much fashion goes unsold and you will find very different answers. Not because someone is wrong, but because each figure counts something different.

Bar chart comparing three published estimates: 2% of clothing material input lost as garments that never reach market (Ellen MacArthur Foundation), 4 to 9% of textile products destroyed before use in Europe, and an average 21% of textile products unsold in research literature (European Environment Agency)

2%: the share of clothing material input sent to landfill or incineration from garments that are produced but never make it to market (global, 2015). [1]
4–9%: the share of textile products put on the European market that are destroyed before use. [2]
21%: the average share of textile products that remain unsold, as found in the research literature reviewed by the EEA. Some of these are returned items. [2]

The EEA is explicit about the weakness of the data: few companies report the volume of unsold goods, and the numbers that exist come from companies' own reporting and lack transparency. It also judges its own destruction share likely to be an underestimate. [2]

A frequently repeated figure, 2.5 to 5 billion unsold garments, does have a verifiable source: the BoF–McKinsey State of Fashion 2025 estimate of excess stock produced in 2023. [3] Note the wording: excess stock, not necessarily stock that is never sold or destroyed. Its own range is two-fold. Other circulating claims, such as "a third of all clothing is never sold", appear in industry and media sources but we could not trace them to a primary dataset, so we do not use them here.

03. THE SCALE OF THE GUESS

Bar chart: about 50 billion clothing units sold in 2000 and more than 100 billion in 2015 (Ellen MacArthur Foundation), compared with an estimated 2.5 to 5 billion excess items produced in 2023 (BoF and McKinsey)

Two calculations put the estimates in proportion (see Methodology).

The value of one excess item. Dividing the BoF–McKinsey value range by its volume range gives the same answer at both ends: about $28 in sales value per excess item ($70bn ÷ 2.5bn; $140bn ÷ 5bn). [3]

The size of the European market implied by the EEA. If 264,000 tonnes is 4% and 594,000 tonnes is 9%, both ends imply roughly 6.6 million tonnes of textile products on the European market. The EEA's two figures are internally consistent. [2] The EEA's separate finding that around one fifth of the 21% unsold share is destroyed works out at about 4%, the bottom of its destruction range. [2]

Billions of items, tens of billions of dollars, hundreds of thousands of tonnes destroyed. All of it made before a single buyer said yes.

04. WHAT MADE-TO-ORDER ACTUALLY CHANGES

McKinsey described the shift in 2019 as products being "pulled" into the market by actual demand rather than "pushed" by forecasts. On-demand production, it noted, requires lower capital investment and leads to smaller inventories and more flexibility. [4]

In a pure made-to-order model, the garment does not exist until it is bought. That changes three things:

No finished-goods forecast. There is no season-end pile of the wrong size in the wrong colour.
No clearance cycle. Without unsold finished stock, there is less to mark down, warehouse or destroy.
Design can move faster than stock. A brand can release more ideas without committing to large volumes of each.

For a sector where the EEA says overproduction "even makes economic sense", made-to-order changes the economics at the point that matters: before production. [2]

05. WHAT MADE-TO-ORDER DOES NOT SOLVE

A credible case has to state the limits.

Higher unit costs. McKinsey notes that production costs are generally higher with smaller batch sizes, as can be transport costs. [4]
Returns still happen. The EEA estimates the average return rate for clothing bought online in Europe at 20%, and that 22–43% of returned online clothing ends up destroyed. Made-to-order does not remove a sizing mistake. [2]
Inventory moves upstream. Fabric, trims and blank garments still have to exist somewhere before an order can be fulfilled. Made-to-order shrinks finished-goods risk; it does not abolish supply-chain stock.
It does not fix the overall volume. Clothing is still under-used: utilisation has fallen 36% in 15 years. [1] Making exactly what is ordered is not the same as people buying less.

06. REGULATION IS MOVING IN THE SAME DIRECTION

From 19 July 2026, large companies across the EU are prohibited from destroying unsold clothes, clothing accessories and footwear under the Ecodesign for Sustainable Products Regulation. Medium-sized companies follow in 2030; small and micro businesses are exempt. Businesses relying on exemptions must provide proof and publish annual reports on what they discard. [5]

When destruction stops being an exit, overproduction becomes a cost that has to be carried. Producing to demand starts to look less like an ethical preference and more like operational logic.

The pressure is wider than Europe's rules. UNEP notes that the fashion sector is considered responsible for between 2% and 8% of global greenhouse gas emissions. [6] In the UK, WRAP reports that 1.42 million tonnes of new textile products were consumed in 2022. [7] A peer-reviewed review in Nature Reviews Earth & Environment estimates the industry produces over 92 million tonnes of waste a year and calls for a deceleration of manufacturing. [8]

07. HOW VAZZARI WORKS

VAZZARI pieces are made to order. As each product page states, every piece is "made to order for you, which reduces overproduction."

It is a design decision as much as a production one. Without the pressure to sell through a fixed quantity, each design can exist as an idea first and a garment only when someone chooses it. We wrote about that system in VAZZARI: 232023, and about the storytelling it allows in Vazzari: The First Brand to Create Outfit Sagas.

We do not claim zero waste. The limits in section 05 apply to us too.

See a made-to-order VAZZARI piece →

08. WHERE THIS GOES

The stock-based model was built for a world of seasons, buyers and long lead times. The data suggests that world is generating its own costs, financially, environmentally and now legally.

Made-to-order will not replace all fashion production. But it changes the question a brand asks. Not "how many will we sell?" but "what will we make when someone asks?"

That connects to a wider shift we explored in Why Fashion Is Becoming More Individual: when clothing is chosen as identity, producing it in advance, in bulk, for nobody in particular makes less sense.

The debate: If destruction is banned and forecasting is still a guess, should the next generation of fashion be produced only after it is chosen?

METHODOLOGY

All figures are taken from the sources listed below and were checked against the source documents on 21 October 2026. VAZZARI's own calculations are:

Sales value per excess item: $70bn ÷ 2.5bn items = $28; $140bn ÷ 5bn items = $28. Inputs: BoF–McKinsey 2023 excess-stock estimate. [3] This assumes the low value goes with the low volume and the high value with the high volume, as the source presents them as ranges.
Implied European textile market: 264,000 t ÷ 0.04 = 6.6 Mt; 594,000 t ÷ 0.09 = 6.6 Mt. Inputs: EEA destruction estimate. [2] This checks internal consistency only; it is not an independent market estimate.
Destroyed share of unsold stock: 21% × 1/5 = 4.2% (rounded to about 4%). Inputs: EEA average unsold share and "around one fifth" destroyed. [2]

The charts show published figures only. Chart 1 compares estimates that use different definitions, geographies and years; they are not additive. Chart 2 shows EMF's approximate published unit values (from its Figure 1, based on Euromonitor data) next to the BoF–McKinsey 2023 estimate for scale only, not as a share. The excess-stock estimate and the unsold-share estimates come from different methods and should not be combined.