Does a Smaller Order Make an Underwear Supplier Less Reliable? Order Size and Batch Consistency

Short answer: a small order is not what makes an underwear supplier unreliable — a merged production run is. Below one dye lot, your units are usually combined with other buyers’ work, so the shade, the fabric batch and the inspection sample all describe a batch you do not control. The trade-off is real, but it sits in consistency, not in price.

Buyers usually frame order size as a negotiation: how low will they go, what does the small quantity cost per piece. That framing hides the actual mechanism. When a programme is too small to fill a dye bath or a cutting lay on its own, the factory does not produce a small version of your order — it produces your order inside somebody else’s. What you receive is then a subset of a batch whose parameters were set by another customer’s specification.

This page sets out what changes at that boundary, how far the standard inspection rules still protect you once it happens, and how to write an order so the boundary is visible before the fabric is dyed rather than after the cartons are opened.

It covers only what happens to batch control at that boundary — the sampling lot, the dye lot and the clauses that make either one enforceable. How small an order can be is set out in our order-size guide, the cost of low volume in the small-batch cost economics page, and the category minimums with their cost breakdown on the MOQ page. This page does not re-price any of them.

What actually changes when your order is merged into another run?

Six things change at once, and only one of them is visible on an invoice.

1. The dye bath is not yours. Colour on piece-dyed knit fabric is the result of dye exhaustion inside a specific bath: liquor volume, salt and alkali dosing, temperature ramp and hold time, then washing and fixing. A dyehouse running 400 kg and a dyehouse running 900 kg of the same recipe do not land on precisely the same shade, because exhaustion rate depends on the ratio of dye liquor to fibre and on how evenly the bath circulates through the load. If your 300 pieces ride in a 900 kg bath, your shade is the bath’s shade, not your lab dip’s shade.

2. The greige batch is not yours. Cotton yarn absorbency varies with fibre maturity, spinning batch and preparation. Two greige rolls knitted from different yarn batches take up the same dye differently even under an identical recipe. Merge two buyers into one bath and the mill will typically blend greige to keep the bath even — which means your units and the other buyer’s units were never separated to begin with.

3. The cutting lay is not yours. A marker is cut across many layers. Small quantities are laid into a combined marker with another style, which is efficient and also the point at which size-to-size trim and fabric direction stop being controlled per order.

4. The inspection lot is not a lot. This is the technical consequence buyers miss, and it is covered in the next section.

5. The test report is not yours. Fibre-content and colourfastness results are issued against a submitted sample or a production batch. A report covering the merged batch tells you the batch passed; it does not tell you that your units were in the sample that was tested.

6. The repeat has no anchor. Your second order has to match something. If the first order was a subset of a merged bath, the reference shade that gets approved is a shade nobody can reproduce on purpose, because the conditions that produced it belonged to another customer’s specification.

Why does a merged run break the sampling plan?

ISO 2859-1, the attribute sampling standard nearly every inspection company works to, defines a lot as a collection of units produced under essentially the same conditions. That definition is not housekeeping — the entire acceptance arithmetic assumes it. AQL tables tell you how many defective units you may find in a sample of n drawn from one lot. Draw the sample from a mixture of two production conditions and the mathematics no longer describes what you are looking at: you can pass a batch whose average is acceptable while your own units are concentrated in the defective tail.

The practical version: if the factory reports “we inspected the batch, 4 defects in 125”, ask which batch. If the answer is a combined run, the sample of 125 covers several customers’ units and your 300 pieces may contribute fewer than 30 of them.

The second effect is arithmetic and it runs the other way. Smaller lots consume a larger share of themselves in inspection, and the acceptance number scales with sample size — so a launch quantity is judged on a much smaller absolute sample than a bulk quantity:

Read the last column together with the fourth. A 300-piece launch order is inspected far more intensively in percentage terms — 16.7% of units — yet it is accepted with up to 3 defective units out of 50. Your entire first season’s stock is 300 pieces, so 3 defects is 1% of everything you own. On a 5,000-piece order, 10 defects is 0.2%. The same AQL, a five-fold difference in exposure, and no supplier is at fault: that is what an acceptance sampling plan is designed to do.

Two further points from the same standard matter for repeat business. Normal inspection moves to tightened when 2 out of 5 consecutive lots are not accepted on original inspection, and returns to normal only after 5 consecutive lots are accepted. If the cumulative number of lots not accepted while on tightened inspection reaches five, the standard calls for inspection to be suspended until the supplier shows corrective action. A supplier who cannot tell you which state your programme is in is not running the switching rules at all — and the switching rules are the part of ISO 2859-1 that actually improves a supply relationship over time.

What does the repeat order expose that the first order hides?

First orders are judged against a lab dip. Repeat orders are judged against the previous bulk, and that is a far harder test, because the previous bulk is a physical object with a measured colour, a measured shrinkage and a measured hand feel — all of which have to be reproduced.

Shade. Instrumental colour difference is calculated to ISO 105-J03, which defines the ΔE_cmc(l:c) computation, normally ΔE_cmc(2:1) for flat textiles. The standard does not set a pass or fail number, because the tolerance is a commercial agreement between buyer and supplier. What is common practice is a tighter tolerance piece-to-piece inside a lot (around ΔE_cmc(2:1) 1.0) and a looser one lot-to-lot (around 1.5 to 2.0), always assessed under a defined illuminant. Write the illuminant down. Two shades can sit inside tolerance under D65 daylight and separate visibly under the warm white of a shop floor — metamerism, and a frequent reason a shade that passed under one light is rejected under another.

Colourfastness. Fastness is measured to ISO 105-C06 for laundering, ISO 105-E04 for perspiration and ISO 105-X12 for rubbing, and graded against the grey scales of ISO 105-A02 (change in colour) and ISO 105-A03 (staining). Fastness results belong to a dye recipe and a fibre, not to a purchase order. If your repeat is dyed in a bath with a different recipe optimisation, the earlier fastness result does not carry over even though the shade looks identical.

Dimensional change. Shrinkage is measured by washing to a defined procedure under ISO 6330 and measuring the change to ISO 5077. Common contract tolerances for cotton single jersey sit at no more than about 5% in length and 4% in width after three washes — an industry practice figure, not a legal one, and it should always be written as a test method plus a number rather than a single percentage. A merged lot can pass shrinkage on the batch average while a single size band — usually the largest, where the fabric is under most tension in cutting — sits outside it.

The pattern is the same in all three: the measurement is only meaningful against a stated lot. Accept a merged run and you have measurements without a lot.

Which risks actually land on the buyer?

Merging is not dishonest and it is not rare; it is normal commercial practice below a certain quantity. What matters is whether the risk is disclosed and priced into the buying decision. These are the six that show up most often, with the mechanism that produces them and the clause that removes them.

What does 2,000 pieces per style and colour actually buy?

For a custom OEM programme, 2,000 pieces is not a sales threshold and it is not a negotiating position. It is the smallest quantity at which the fixed legs of a programme stop being shared with somebody else. Those legs are physical:

• Dye bath minimum. A dyeing machine has a minimum liquor volume below which the load does not circulate evenly. Fill it with too little fabric and the shade goes streaky; the bath is economic only above a certain weight of fabric.
• Cutting lay. A marker is cut through a stack of layers. Too few garments per size and the marker is combined with another style, which is where size-to-size trim control is lost.
• Print setup. Every colour in a print is a separate screen or plate with its own registration setup. Below a certain run length the setup is carried by the next order in the queue.
• Elastic and waistband weaving. Jacquard waistbands are woven to order; the loom setup is per design and per width.
• Yarn dyeing, for yarn-dyed stripes. The stripe colours are dyed as yarn before knitting, so the minimum is set by the yarn dyeing batch, not by the garment quantity.
• Packing changeover. Polybag, label, hangtag and carton configuration are per SKU. More SKUs at a lower quantity per SKU means more changeovers than production.

One way the arithmetic can resolve, purely as an illustration: 2,000 pieces split across five sizes at 400 each, with 400 per size cut as 40 layers of a 10-piece marker for that size. Real markers normally nest several sizes and depend on fabric width, so this is not a lay plan we would quote; the only point is that the quantity divides into whole lays rather than into remainders that get merged. Once the quantity is fixed, the calendar is the other half of the programme — the reverse-planning figures sit on the production timeline page, and the size ratio that determines how many lays you need is on the size conversion chart.

The same arithmetic explains the other category figures. Children’s underwear carries more size breaks in the same range, so the minimum rises to 3,600 pieces. Sleepwear is quoted in sets at 600, which is 1,200 pieces, because a set is a top and a bottom. Where a category minimum looks arbitrary, the size-break count usually explains it.

If you genuinely need fewer pieces, what are the honest options?

Merging is not the only way to buy small, and three of the four alternatives below avoid a new dye bath entirely — which is the point.

Buy a running shade. Take a colour and fabric that is already in production for another programme and specify against the running lot. You lose the exact shade you wanted and gain a lot you can actually inspect; the shade is by definition reproducible because it already exists in bulk.

Cut the number of SKUs, not the number of pieces. One colour across all sizes at 2,000 pieces is a dedicated lot. Three colours at 700 each is three merged runs. Buyers routinely solve a quantity problem by spreading it thinner across colours, which is exactly backwards.

Repeat an existing style. A repeat needs no new marker, no new print setup and no new waistband development. The minimum for a repeat is lower than the minimum for a new style, because most of the fixed legs are already paid.

Buy sample sets as samples. Salesman samples and development samples are a separate stream with their own pricing and their own purpose. They are not a production lot and should not be received, inspected or sold as one.

What does not work is asking for a smaller production run at the standard price and assuming the difference is margin. The difference is a shared bath, and the cost of sharing it appears later, in shade, in cartons and in an inspection report that cannot be enforced.

How do you write it into the order so it is enforceable?

Five fields turn a verbal assurance into something you can reject against, and they belong in every custom OEM purchase order — not in an email after the deposit.

1. Lot identity. One lot per style, colour and size range, with a lot number assigned before dyeing and repeated on the packing list, the carton label and the inspection report. Where the quantity exceeds one dye bath, require each bath to be declared as a numbered sub-lot rather than the whole order being called one lot.
2. Dye lot on the paperwork. The dye lot number belongs on the packing list next to the quantity, and on the carton label. Without it, a shade complaint six weeks later cannot be traced to a bath.
3. Inspection declaration. Lot size, inspection level, AQL, sample size, Ac/Re and defect classification (critical, major, minor) stated in the report. A report with a defect count and no lot size is not an acceptance decision.
4. Colour tolerance. Method, figure and illuminant: ΔE_cmc(2:1) of 1.0 piece to piece and 1.5 lot to lot, assessed under D65. Without a number, “write the illuminant down” gives you nothing to reject against.
5. Remedy for a merged lot. State what happens if the goods arrive from a lot the supplier did not declare. The workable version is re-inspection at the supplier’s cost against the declared lot, with the option to reject.

Sample clause, for adaptation: “Each colour and style in this order shall be produced as one lot, dyed in one bath, identified by a lot number stated on the packing list and carton label; where the quantity exceeds the capacity of one bath, each bath shall be declared as a numbered sub-lot. The pre-shipment inspection shall be declared to ISO 2859-1 with lot size, general inspection level II, AQL 2.5 for major defects, critical defects not permitted (c = 0, where any critical nonconformity rejects the lot), sample size and Ac/Re. Colour shall be assessed to ISO 105-J03 at ΔE_cmc(2:1) of 1.0 or less piece to piece and 1.5 or less lot to lot under illuminant D65. Goods received from an undeclared merged lot may be re-inspected at the supplier’s cost and rejected.”

Adapt the AQL values to your own programme. What must not be adapted is the lot declaration, because every other clause in the inspection chain depends on it.

What should you check before you accept a smaller first order?

• Ask for the lot number before the dyeing is scheduled, not after the goods are packed.
• Confirm the colour will be dyed in one bath with no other buyer’s units, and get that in writing.
• Ask how many size breaks the quantity produces. If the answer is “whatever you want”, the marker will be merged.
• Check that the quantity divides into whole cutting lays rather than leaving a remainder.
• Request the previous repeat’s dye lot number and its ΔE against the approved shade. No record means no repeat control.
• Confirm the fastness and shrinkage reports are issued against a lot of your quantity, not a combined batch.
• Ask which ISO 2859-1 state your programme is in: normal, tightened or suspended.
• Verify the packing list will carry lot number, dye lot and per-carton assortment.
• Photograph the dye lot tag on the fabric roll at inspection, and keep it with the inspection report.
• If the answer to any of the above is a verbal assurance, price the risk yourself before you place the order.

How do you tell a reliable supplier from a cheap one on this specific point?

Every supplier will say they do not merge. The difference is whether they can produce the record. Five questions separate the two, and each has an answer that can be checked rather than believed.

“What was the dye lot number on my last order?” A supplier with lot control answers from a system. A supplier without it answers with a date.

“What is the ΔE between my approved shade and the bulk you shipped?” The number exists or it does not. “It matched visually” is an answer, but it is not a measurement.

“Which AQL and inspection level did you declare, and what was n?” Anyone running ISO 2859-1 properly knows their own sample size without looking it up.

“How many lots were rejected in the last five?” This is the switching-rule question. A supplier who has never had a lot rejected in five years is either exceptional or not recording.

“If I order 700 pieces in three colours, what happens to my shade?” The honest answer describes a shared bath. The dishonest answer promises a dedicated one, and the buyer discovers the difference in the second carton.

This is the practical content of the phrase reliable supplier: not a claim about capability, but the ability to answer a lot-level question with a lot-level record. The 12-point scorecard tests a supplier as a whole; these five questions test one thing only — whether lot records exist. A trading company that manages production across partner factories can hold those records: the certifications belong to the partner factories that hold amfori BSCI and OEKO-TEX Standard 100, while lot control, inspection declaration and repeat shade standards are managed by the sourcing team that places and follows the order. What a buyer should test is whether the record exists, not who signs it.

Does the same logic apply to children’s underwear and sleepwear?

It applies more strictly, for a reason that has nothing to do with commercial risk. Test evidence is lot-specific by construction: a report covering a combined batch is weaker evidence than one covering the lot you received. In the United States, general apparel falls under 16 CFR 1610, children’s sleepwear under 16 CFR 1615 for sizes 0–6X and 16 CFR 1616 for sizes 7–14, and small parts under 16 CFR 1501. Fibre content and care labelling are verified against a sample rather than a batch average: fibre content is governed by 16 CFR Part 303 under the Textile Fiber Products Identification Act and care labelling by 16 CFR Part 423, while in the European Union Regulation (EU) 1007/2011 requires the declared fibre composition to be accurate for the goods.

Children’s underwear carries a minimum of 3,600 pieces per style and colour for the reason set out above — more size breaks in the same size range. Sleepwear programmes are quoted in sets at 600, which is 1,200 pieces. Neither number is a negotiating tactic; both are the point at which the order stops sharing a bath and a lay with somebody else.

For a buyer assembling a first children’s range, the sequencing that works is: fix the sizes and the size ratio first, then the colours, then the quantity — because the quantity that clears the minimum is a function of the size breaks, not the other way round.

What is the one sentence to take to a supplier meeting?

Ask for the lot number and the dye lot before the order is placed. A supplier who can produce them is managing the thing that actually determines whether your second order looks like your first; a supplier who cannot is selling you whatever came out of the bath that week, and no amount of inspection will recover the difference afterwards.

References

• ISO 2859-1 — Sampling procedures for inspection by attributes, acceptance sampling plans, lot definition and switching rules. iso.org
• ISO Online Browsing Platform — ISO 6330 (domestic washing procedures), ISO 5077 (dimensional change), ISO 105 series (colour fastness and colour difference). iso.org/obp
• US eCFR — 16 CFR 423, Textile Fiber Products Identification Act labelling rules. ecfr.gov
• US eCFR — 16 CFR Part 1610, Standard for the Flammability of Clothing Textiles. ecfr.gov
• EUR-Lex — Regulation (EU) 1007/2011 on textile fibre names and labelling. eur-lex.europa.eu
• US FTC — clothing and textiles labelling and advertising guidance. ftc.gov
• OEKO-TEX — STANDARD 100, the certification held by our partner factories. oeko-tex.com
• amfori — BSCI, the social audit held by our partner factories. amfori.org
• ICC — Incoterms 2020, the point at which risk passes. iccwbo.org

Next step

If you already have a supplier and want to test whether lot control exists, work through the supplier verification checklist and the first-order verification sequence. If your programme is the men’s boxer brief line that carries most of our volume, the construction and specification detail sits on the men’s boxer briefs page. To test us on the five questions above, ask the lot question directly — the answer should be a lot number, not a promise.

Published: 2026-11-09 · Xiamen Unitex Trade Co., Ltd. — OEM & ODM underwear sourcing partner

Unitex Underwear (Xiamen Unitex Trade Co., Ltd.) — OEM/ODM underwear sourcing and supply-chain partner in Xiamen, China; in the trade since 2009, incorporated in 2013. Based in the Xiamen–Quanzhou–Jinjiang knitwear cluster, Fujian Province, south-east China.