Fabric MOQ Per Color vs Per Style (2026): 43 Data Points on Dye-Lot Minimums, Colorway Multiplication, and Your True Minimum Order

Key Takeaways

1. 300 pieces per style/colour — a real quoted floor, and 100 red + 100 blue + 100 black does not satisfy it (OMJ Apparel, T1).
2. Per colourway, not per style — how factory minimums are actually applied; one style in four colours is four minimums (Cord Apparel, T3-consensus: described the same way across independent sourcing guides, with no primary dataset measuring it).
3. 10,000 lbs per colour — the minimum order for solution-dyed yarn (Meridian Mill House, T1).
4. 150 lbs per colour — the equivalent minimum for package-dyed yarn, a roughly 67x difference driven purely by dyeing method (Meridian Mill House, T1).
5. About 20 kg per colour — the yarn dye-lot minimum knitwear brands hit at the mill (GUOOU Fashion, T1).
6. 500–1,000 metres per colour — the ordinary fabric-mill minimum, and what forces garment factories to raise their own MOQs (Argus Apparel, T3-consensus: a general market pattern repeated across sourcing guides, not a measured benchmark).
7. No minimum vs. 5,000 metres per specification — what separates in-stock greige from custom greige at the same mill (Weaverine, T1).
8. 38–40% higher cost per kg — to dye the same fibre on 20–25 kg machines instead of 300–600 kg machines (da Silva et al., MDPI Processes, 2025, T1).
9. 28–32 L/kg vs 20–22 L/kg — water for a dark shade versus a light one on the same rope-dyeing machine (Brazzoli S.r.l., T1).
10. 43.96% loss ratio — among China’s 2,036 above-designated-size printing and dyeing enterprises in Q1 2026, on a 2.08% revenue profit margin (CNTAC / National Bureau of Statistics of China, T1).
11. 78.2% of total dyeing cost — is water, the dominant cost in a real 109,000 kg/day plant’s optimised model, and every extra colour is another batch that consumes it fresh (da Silva et al., MDPI Processes, 2025, T1).

1. The Multiplication Trap: One Style, Four Colours, Four Minimums

A 100-piece minimum is not a 100-piece order. OMJ Apparel spells out the arithmetic most quote sheets leave implicit: at a 300-piece per style/colour floor you must buy 300 red shirts or 300 blue shirts — 100 red plus 100 blue plus 100 black is not 300. The unit of the minimum is the colourway, not the design.

The published floors range wildly, and every one of them is quoted per colour: 25 pieces at Wearlets, 50 at YouTricot, 80–100 at Bryden Apparel, 300 at Lin Sweater. Our own tiers sit in the same shape — 100 pcs per colour for dresses, sportswear and plus-size on stock fabric, 300 pcs per colour for knitwear on stock yarn, and 1,000 pcs where the fabric or yarn has to be dyed to order. We charge per colour like everyone else. The difference is where the per-colour floor sits: 100, not 300–500.

Cord Apparel describes the mechanic in the same terms — a separate minimum for each colourway, so one style in four colours costs materially more than one style in one colour. That framing is consistent across independent sourcing guides but is not measured in any primary dataset; no institutional body publishes garment MOQ benchmarks at all, which is why the evidence on this page is drawn from factories’ and mills’ own published terms rather than from an industry survey.

The multiplier is what catches founders out. WavePLM puts the average collection at four colourways per style — a single-source editorial benchmark with no measured study behind it, so treat it as a planning assumption rather than a statistic. Four colourways against a 100 pcs/colour floor is a 400-piece buy for one design. That is four times the number in the business plan, and the plan was not wrong about the price per piece. It was wrong about the unit.

“The minimum you were quoted was never per style. It was per colour, and you are about to order four of them.”

Every floor in this table is the factory’s own published figure for its own programme. They illustrate what a small brand can find in the market; they are not an industry average — and none of them is a per-style number, however it was phrased in the first email.

Floors also move by garment type long before colour enters the picture: a woven dress, a fleece hoodie and a fully-fashioned sweater do not share a minimum. For the category-by-category picture underneath this one, see our MOQ benchmarks by garment type.

2. Where the Rule Comes From: Yarn Sells by the Kilo, Fabric by the Metre, Both Per Colour

No garment factory invented the per-colour rule. It arrives from upstream, and it arrives in a different unit at every step.

Meridian Mill House publishes the starkest version: solution-dyed yarn carries a 10,000 lb minimum per colour, package-dyed yarn 150 lbs — a roughly 67x gap that has nothing to do with how many sweaters you want and everything to do with how the yarn is coloured. Solution dyeing puts pigment into the polymer before the filament is extruded, so the colour decision is made at the chemical stage and the run has to be enormous to justify a line changeover. Package dyeing colours yarn that already exists, so it can be done in far smaller lots. Same jumper, same shade, two orders of magnitude apart in minimum.

For knitwear the practical number is smaller and still per colour: GUOOU Fashion puts the yarn dye-lot minimum at around 20 kg per colour. That is the figure a knitwear brand actually collides with, and it is why a fifth colourway in a sweater range is a separate purchase from the spinner, not a line item on the same one.

At the fabric level the unit becomes metres. Argus Apparel and Szoneier Fabrics describe the same ordinary mill floor as a general market pattern: 500–1,000 metres per colour, and roughly 500 metres or more per colour for OEM producers. Both figures are repeated across independent sourcing guides rather than measured in any dataset, so read them as the shape of the market rather than as a benchmark. Juyou Textile speaks only for itself, and is more specific: the mill’s own published policy sets a minimum of 1,000 metres for any custom colour. Argus states the consequence plainly — mill minimums force garment manufacturers to raise their own.

Weaverine shows the escape hatch and its price in one line: no minimum at all on in-stock greige, 5,000 metres per specification the moment the greige is custom. Take the colour the mill already holds and the minimum can disappear; ask for one it does not, and the floor arrives with it. That trade is the single most useful lever a small brand has, and it is available before any negotiation starts.

“Your factory did not set the per-colour minimum. The dye lot did, three suppliers upstream.”

Note the unit changes at every tier: pounds and kilos at the yarn stage, metres at the mill, pieces at the factory. A brand comparing quotes in pieces alone cannot see which constraint is actually binding.

Which is why the fabric conversation should happen before the MOQ conversation, not after it. We walk through how fabric sourcing works upstream of the cutting table — stock ranges, mill lead times, and where a colour decision stops being reversible.

3. Inside the Dye House: The Machine Sets the Floor

Dye minimums are a hardware fact before they are a commercial policy. Fong’s ECO-8 holds 80–550 kg per tube and up to 4,400 kg per batch; the THEN AIRFLOW Synergy runs 50–1,500 kg per chamber and scales to 3,600 kg across twelve. A vat sized for hundreds of kilos does not become economic because your order is small.

The liquor ratio is the number that decides how far down a machine can flex. It is the weight of water in the bath per unit weight of goods, and it determines how much water, steam and dye chemistry a run consumes regardless of how full the machine is. Fong’s ALLWIN reaches 1:4 and 34 litres per kg on medium and dark shades; Loris Bellini’s skein cabinets sit at 1:12 to 1:15. A machine at 1:15 is putting three times the water and heat through the same kilo of yarn as one at 1:5, and that gap is what a dye house is protecting when it quotes a per-colour minimum.

Part-loading exists but has limits. Loris Bellini’s PULSAR is explicitly built to run down to 50% of nominal capacity at a near-constant liquor ratio — genuinely useful engineering, and on a 1,500 kg machine it still leaves a 750 kg floor. Smaller equipment starts lower: the ABEP arm machines run from 10 kg to 200 kg per unit, and skein cabinets from 5 kg. That is the band in which sampling and lab dips are possible at all, and it is also why a lab dip is cheap while a bulk colour is not.

The peer-reviewed literature points the same way without giving a clean number. A 2026 Springer Discover Chemistry paper finds that low liquor ratio processes cut both water and energy use against conventional cotton dyeing; the article text sits behind a publisher authentication wall, so only the qualitative finding is used here and the stat is held at Tier 2. The direction is not in dispute; the magnitude is not confirmable from outside the paywall.

None of this equipment is a legacy problem waiting to age out, either. The global textile dyeing machines market is put at USD 6.24 billion in 2026 and is forecast to reach USD 10.27 billion by 2035 (Business Research Insights) — the 2035 figure is a projection, not a measurement. Capital is going into bigger, lower-ratio machines, which lowers cost per kilo at volume and does nothing at all for a 40-kilo colour run.

“A 1,500 kg vat that can part-load to 50% still has a 750 kg floor. That floor is your MOQ, wearing a different unit.”

These are manufacturer-published specifications for equipment installed in commercial dye houses. Machine capacity is the hard constraint; everything a brand negotiates sits on top of it. The one market-sizing row above is a forecast and is labelled as such.

4. The Cost of a Colour: Why a Small Batch Costs More Per Kilo

A peer-reviewed cost model of a working knit-dyeing plant puts a number on the penalty that factories usually only gesture at. Dyeing polyester on the plant’s 20–25 kg machines cost R$2.34 per kg against R$1.67–1.69 per kg on its 300–600 kg machines — 38–40% more for the identical fibre. Same dye, same shade, same building; the only variable is batch size.

Water is why. In the same study water accounted for 78.2% of total process cost, and water scales with the bath rather than with the goods in it. A half-empty machine fills, heats, holds and drains almost the same volume as a full one. Coats Group reports the other side of that equation: moving from an 8:1 to a 4.5:1 liquor ratio saves 44% of the water for the same output. Both numbers say the same thing from opposite ends — the bath is the cost, and small lots buy a full bath for part of a load.

The gap between modern and legacy parameters is the same story at industry scale. The MDPI authors put modern machinery at 35.5 L/kg for cotton, 20 L/kg for polyester and 16 L/kg for polyamide, against 90, 102 and 74 L/kg under conventional parameters. A dye house that has re-equipped has genuinely lowered its cost floor. It has not removed it.

Shade depth stacks on top. Brazzoli’s own datasheet prices a dark shade at 28–32 L/kg and 1.7 kg of steam per kg of fabric, against 20–22 L/kg and 1.1 kg of steam for a light one. A five-colour range that includes two darks is not five equal runs. And each Pantone match costs USD 40–120 in lab dips across 2–4 rounds before a single metre is dyed — a figure OneAim Apparel publishes from its own quote data, with methodology undisclosed, so treat it as directional.

“Small batches are not merely inconvenient for a dye house. They are 38–40% more expensive per kilo, measured.”

The MDPI figures come from a single Brazilian plant processing 109,000 kg/day across 16 machines, so the currency and absolute values are plant-specific. The ratio between small and large machines is the transferable finding.

5. Why the Dye House Will Not Absorb It: Sub-2.2% Margins

Ask why a mill cannot simply eat the cost of a short colour run and the answer is in China’s own industry accounts. In Q1 2026, 895 of 2,036 above-designated-size printing and dyeing enterprises were loss-making — a 43.96% loss ratio, with a revenue profit margin of 2.08% across the sector.

The cost base barely moves when volume does. Revenue fell 2.62% year on year in that quarter while total profit fell 16.10% — profit eroding roughly six times faster than sales. A business at 2% net margin has no room to run a half-empty vat as a favour. When a dye house holds its per-colour minimum, that is not a posture it adopted for the negotiation. It is the difference between the order being worth taking and not.

This is not a story about a shrinking industry. Full-year 2024 was the sector’s best since 2018 at RMB 322.387 billion of revenue and a 5.98% cost-profit rate, on 41.62 billion metres of printed and dyed fabric in the first three quarters alone; China’s total cloth output was 30.7 billion metres in 2025. It is a story about scale economics. The industry is enormous, it runs thin, and the per-colour minimum is the mechanism that keeps a small order from being written at a loss.

The wider dye market tells the same structural story. Mordor Intelligence puts the textile dye market at USD 13.21 billion in 2026 with a forecast USD 16.71 billion by 2031, and Grand View Research forecasts USD 66.19 billion for dyes and pigments overall by 2033, with textiles taking 63.1% of the dyes market. Both 2031 and 2033 numbers are projections. Growth at that rate is volume growth, not margin growth — which is precisely why the per-colour floor is not going away.

Our own supply chain is the practical version of this. Our yarn spinners and dye houses sit within a two-hour radius of Jiaxing, which is how we can price a colour before quoting rather than after — and why our knitwear production at 300 pcs per colour is a stock-yarn number rather than a custom-dye one.

“At a 2.08% net margin and a 43.96% loss ratio, a dye house that waives its per-colour minimum is not being generous. It is going under.”

CNTAC figures are published in Chinese and denominated in 亿 (hundred-million) units; the metre and RMB values here are converted to billions. The Q1 2026, full-year 2024 and Q1–Q3 2024 releases are three separate reports. The two market-sizing rows are forecasts and are labelled as such.

6. Cutting the Multiplier: What Fewer Colours Is Actually Worth

The cheapest way to lower a minimum is to stop multiplying it. McKinsey found up to an 8% margin increase for the companies most disciplined about which SKUs they add and cut — research published in June 2020, so read it for the mechanism rather than as a current rate. The mechanism has not moved.

The downside case is well measured. Business of Fashion and McKinsey put 2023’s surplus at 2.5–5 billion garments worth USD 70–140 billion. Source Fashion’s 2025 work with Insider Trends puts annual output at 80–150 billion garments with up to 40% unsold — a figure carried consistently in industry coverage but not confirmable against the publisher’s own site this run, so held at consensus rather than primary. A fourth colourway ordered to satisfy a per-colour floor, rather than because anyone wanted it, lands in that number.

And the exit from over-buying is expensive. First Insight’s consumer survey found 45% of women need to see a markdown of 41% or more before entering a store, and 39% will travel elsewhere for a lower price — figures fielded in late 2016 and released in 2017, so read them as the shape of markdown behaviour rather than a current rate. Meanwhile the colour work itself is billable before anything ships: lab dips and colour-matching surcharges sit among seven below-the-line costs that OneAim Apparel puts at 6–11% of FOB, from its own 2024–2026 quote dataset with methodology undisclosed.

The practical levers are unglamorous, and they all work before the quote arrives:

• Launch three colours, not five. At a 100 pcs/colour floor that is a 200-piece difference on one style — roughly the working-capital gap between a first order that can be funded and one that cannot.
• Take stock fabric or stock yarn where the palette allows it. Weaverine’s no-minimum-on-stock-greige terms are the extreme version; most mills hold something usable in black, navy, ecru and grey.
• Keep custom dyeing for the one hero colour that earns it. One custom shade in a range of four is one 1,000-piece commitment, not four.
• Phase the second colour drop into a repeat order. At roughly 30 days for bulk after PP sample approval, a second drop is a season decision, not a year-one one.

Colour count is only one of the levers on a quote, and it is the one the brand controls outright. The rest — fabric choice, delivery window, payment terms, size ratio — are negotiable in different ways; we set out the levers that actually move a factory’s MOQ separately.

“Every colourway you add is a separate purchase order, a separate dye lot, and a separate chance to be wrong.”

The First Insight and McKinsey figures predate 2026 (2016–17 and 2020 respectively) and are used for mechanism, not for current-year rates. Both are flagged in the methodology below.

Fabric MOQ Per Colour vs Per Style by the Numbers: All 18 Key Data Points (2026)

Methodology and Sources

Every figure in this article was traced to the organisation that measured it and confirmed on that organisation’s own page before inclusion. Where a factory or mill states its own minimum order quantity on its own site, we cite it as that company’s stated policy for its own programme, not as an industry benchmark — the distinction matters on this topic, because no institutional body publishes garment MOQ benchmarks. Industry-wide generalisations (for example “mills require 500–1,000 metres per colour”) are held to a lower tier and marked as such even where four vendors repeat them. Of 77 raw candidates, 22 were dropped at verification as ghost stats or misattributions, a further 24 were cut at assembly for weak sourcing or off-axis scope, and 43 were kept.

Tier breakdown: 43 data points kept across 32 unique sources — 37 Tier 1 (86.05%), 1 Tier 2, 4 Tier 3-consensus and 1 Tier 3-flagged (11.63% combined). Tier 1 is the organisation that measured or published the number as its own; Tier 2 is a reputable source whose underlying text could not be fetched; Tier 3-consensus is a market-wide pattern corroborated across independent sources without a primary measurement behind it; Tier 3-flagged is single-source and carries a visible inline qualifier wherever it appears. No Tier 3-flagged figure appears in the Key Takeaways.

Four figures were reclassified upward at assembly: Bryden Apparel’s 80–100 pcs per colourway, Shijin Fashion’s 50–150 pcs per style, Juyou Textile’s 1,000 m per custom colour and Weaverine’s stock-versus-custom greige terms are each a company’s own published commercial terms, verbatim-confirmed on that company’s own page, which makes the publisher the measurer. Industry-wide claims appearing on those same pages were not upgraded.

One unit correction is worth stating plainly, because it recurs in English-language coverage of Chinese textile data. The Chinese numeral unit 亿 is 10⁸ — one hundred million, or 0.1 billion, not one billion. China’s Q1–Q3 2024 printed and dyed fabric output of 416.22亿 metres is therefore 41.62 billion metres, not 416.22 billion. Two CNTAC export figures for the same period were excluded from this article entirely rather than published under an unresolved magnitude dispute.

Recency notes

• Most of this article’s evidence is not a study with a fieldwork date. It is a factory’s, mill’s or machine maker’s currently published minimum or specification, fetched from that company’s own page during this research. A published minimum is not a historical measurement that ages; it is the price of doing business with that supplier today. Those 25 stats are dated 2026, the year their currency was confirmed. Dated studies, surveys and official statistical releases keep their true year.
• First Insight’s markdown figures (45% of women needing a 41%+ markdown; 39% willing to travel for a lower price) come from a survey of 1,303 US consumers fielded in late 2016 and released in 2017. Used for the shape of markdown behaviour, not as a current rate.
• McKinsey’s “up to 8% margin increase” from SKU rationalisation is from Harnessing the Power of Simplicity in a Complex Consumer-Product Environment, June 2020. An earlier attribution to a different McKinsey article was corrected at verification, and an apparel-specific framing attached to the raw candidate was dropped as unconfirmed.
• Business of Fashion and McKinsey’s excess-garment figure describes 2023 production, published in The State of Fashion 2025. Direct fetch of both publishers was blocked (403 / timeout); the figure is confirmed via the named joint report and nothing contradicts it.
• Source Fashion’s 80–150 billion garments / up to 40% unsold figure is from June 2025 work with Insider Trends. The publisher’s site could not be fetched directly, so it is held at Tier 3-consensus.
• CNTAC’s most recent full-year printing and dyeing accounts are for 2024; the quarterly loss-ratio and margin data are Q1 2026, the most recent published at the time of writing. NBS cloth output is full-year 2025.
• The Springer Discover Chemistry paper on low liquor ratio dyeing sits behind a persistent publisher authentication redirect. The DOI resolves and the journal is real, but the article text could not be fetched, so only the qualitative finding is used and the stat is held at Tier 2. A specific energy figure circulating in search summaries (~32.4% less energy at 1:3 versus 1:10) was not used, as it could not be confirmed against the paper.
• Brazzoli’s Ecologic Plus II figures come from a datasheet carrying a 2024 revision that remains the manufacturer’s current published specification for that machine.
• Three market-sizing figures state a forward-looking forecast and are worded as forecasts in the text: Business Research Insights (textile dyeing machines, USD 6.24bn 2026 to a projected USD 10.27bn by 2035), Mordor Intelligence (textile dye, USD 13.21bn 2026 to a projected USD 16.71bn by 2031) and Grand View Research (dyes and pigments, USD 66.19bn by 2033). None appears in the Key Takeaways, the summary table or the chart.
• OneAim Apparel’s two figures ($40–120 per Pantone match; 6–11% of FOB) are the firm’s own self-attributed quote data with undisclosed methodology. They are kept inside sections that carry that qualifier and are deliberately excluded from the Key Takeaways, which carry no context.

Documented absences

• No institutional body — statistical agency, trade association or academic group — publishes garment MOQ benchmarks by colour or by style. Every per-colour floor on this page is a named company’s own published policy.
• No measured study of average colourways per style exists that we could locate. WavePLM’s four-per-style figure is editorial framing and is flagged as such everywhere it appears.
• No dataset quantifies sell-through against colourway count. Searches for one returned only definitional content on sell-through rates.
• Per-unit cost curves by order quantity are published only by individual vendors as illustrative price bands, with no disclosed methodology. They were excluded here.

Last updated: September 2026. We update this page quarterly.