How to Source Apparel Direct from Factories

In this cluster

This is the sourcing operations pillar in the Yarnstick knowledge base, for a US brand founder or ops lead deciding how to get product made overseas. It covers the models you can buy manufacturing under, how to vet a factory, what a cost sheet is made of, and the sequence from tech pack to delivered bulk. Minimum order economics sit separately in MOQ economics and how to move your minimum, because minimums are a fabric and dye-lot problem rather than a partner-selection one.

One warning up front: direct factory sourcing is not automatically better. It is a decision to insource a job someone else was doing, and it pays only when your volume covers the cost of doing that job well.

Terms buyers use for this, and what each one signals

Before the models, a naming note, because this market has no single word for the company that makes your garments and the phrasing people use is a genuine signal. A supplier that answers to "sewing factory" and one that answers to "full-package supplier" are often different businesses with different minimums and different quotes.

One trade, several names. What each phrasing usually tells you about the supplier behind it.

Two consequences worth carrying into every conversation. Two clothing factories can quote the same style on entirely different bases — one CMT, one full-package, one FOB, one FCA — and the quotes are not comparable until you have forced them onto the same footing. And regional phrasings behave the same way: a search for apparel manufacturers in LA returns contractors, full-package houses and sample rooms mixed together, three businesses that do different jobs, unpicked in Los Angeles clothing manufacturers, MOQs and SB 62.

Adjacent searches, and where they belong

One line each, with the honest destination.

The four ways a US brand can buy apparel manufacturing

Every arrangement in this industry is a variation on four models, and the difference between them is not price. It is who holds the factory relationship, who carries material risk, and who is accountable when the goods are wrong.

The four models, compared on what you contract for and what it costs you. Fee ranges are industry ranges rather than published rates, since almost no intermediary publishes one.

Two terms do a lot of work here. FOB is an Incoterm: the seller pays through loading on the vessel at the origin port, and risk passes on board. CMT is a commercial arrangement describing who buys the materials, not an Incoterm. You can buy CMT and ship FOB. The two get used as alternatives, and they are not.

One practical note from the International Chamber of Commerce: FOB, CFR and CIF were written for bulk cargo, and the ICC recommends FCA for containerised freight. Apparel is close to 100% containerised, so the industry habit of quoting "FOB Chittagong" is technically a misuse of the rule. See Incoterms for apparel brands.

White label, private label and cut-and-sew are three levels of development, and CMT is something else entirely

Suppliers use all four terms interchangeably, and that is where quotes stop being comparable. Three of them describe how much of the product is developed for you. The fourth describes who buys the materials. They are different axes, and you pick one from each.

The development axis, from least to most bespoke:

• White label — an existing finished product the factory already makes, sold under your brand with your label swapped in. No specification change, lowest MOQ, no differentiation: a competitor can buy the same body next week.
• Private label — the factory's existing block or base product adapted to your specification: your fabric, colour, trims, labels and minor fit changes. Moderate MOQ, limited differentiation in fit.
• Cut and sew — developed from your own pattern and tech pack from scratch. Full control over fit and construction, highest development cost, highest MOQ, longest lead time.

The commercial axis is separate:

• CMT (cut, make, trim) — you own and supply the fabric and trims, and the factory charges only for cutting, making and trimming. The opposite pole is FOB or full-package, where the factory sources the materials and quotes you a finished garment price.

CMT is not a synonym for cut and sew, however often the two are used that way. You can buy a white-label body on CMT terms by consigning the fabric, and you can buy a fully cut-and-sew style on full-package terms and never touch a mill. Confirm both answers before you compare two quotes, because a CMT quote excludes the largest cost line in the garment.

Most brands start white or private label and move to cut and sew on hero styles once volume justifies pattern development. The reverse order is expensive. The two terms are compared on minimum, exclusivity, development cost and what you can legally claim in private label vs white label, compared line by line. If the honest answer for your volume is a decorated stock body rather than a manufactured one, that market is covered in blanks and wholesale apparel suppliers and custom and decorated apparel.

Apparel product development: what the brand owns and what the factory owns

Sourcing is the second half of a job whose first half is apparel product development. Brands routinely contract the second half and assume the first arrives free with it. It does not, and the gap is where most first programmes lose their calendar.

The division that works is simple. The brand owns intent and specification: the fabric platform, the colour count, the graded points of measure, the tolerances, the colour standards and the version log. The factory owns feasibility and execution: the pattern, the marker, the standard minutes, the line, and what tolerance its process can actually hold. A brand that hands over a mood board has not outsourced clothing product development — it has asked the factory to make design decisions and agreed in advance to accept them.

That first half is a real cash line before a unit is sold. A tech pack takes 1–2 weeks, a first tee sample runs $200–350 and a hoodie $300–600, two to four rounds are normal, and roughly 60% of physical samples never reach production. Budget development separately from bulk, or the bulk budget absorbs it and the range shrinks mid-season.

Two structural points follow. Colour count is a costing decision made during development, not a design decision made at the end, because minimums are per colourway per fabric. And once you are running more than a handful of styles a season, the coordination itself becomes the constraint — which is where sourcing software and fashion PLM start to earn their cost, and where the development sequence in starting an apparel brand, from plan to first order is worth reading before you contact a supplier.

How to vet a garment factory: what to ask and what to demand in writing

Factory selection is the highest-leverage decision in the programme, and the one most often made on a WhatsApp reply and a price. Structure it.

Questions that reveal how a factory actually operates

Ask these and listen for specificity rather than enthusiasm:

1. What categories do you run most, and for which brands at my volume?
2. Which operations are in-house — knitting, dyeing, printing, embroidery, washing — and which are subcontracted, to whom?
3. What is your MOQ per colourway per fabric, not per style?
4. What SMV and what line efficiency sit behind your CM quote?
5. Do you nominate the mill, or will you work with a mill I nominate?
6. What is your PP sample lead time, and will you hold the cut until PP approval in writing?
7. Which third-party inspection firms have you worked with, and will you accept an inspection I book?
8. Where do Diwali, Eid, Tết or Chinese New Year land in my delivery window?

The subcontracting question matters more than it sounds. Undisclosed subcontracting is where quality failures and forced-labour exposure both hide, and CBP's UFLPA enforcement follows the input rather than the entity you contracted with.

Documents to demand before you develop, not after

• Business registration and export licence in the legal entity name that will invoice you.
• A current social compliance audit — amfori BSCI, SMETA, or an RSC report for Bangladesh — with the site address matching the entity.
• Product and chemistry certifications matched to the claims you plan to make: GOTS for organic, OEKO-TEX Standard 100, GRS for recycled content.
• A written subcontracting disclosure list.
• Mill and trim supplier names for your fabric platform, which the traceability file will need anyway.
• Recent third-party test reports for the fabric you are buying, not a generic sample.

What a certification proves, and what it does not

A certificate proves a defined scope was assessed at a specific site on a specific date by a specific body. It does not prove the goods in your carton fall inside that scope, that the audit was unannounced, that the certificate is still valid, or that the subcontractor who did your embroidery was covered. Certificates are a screening tool and an input to a traceability file, not evidence of admissibility — the UFLPA rebuttable presumption is overcome with clear and convincing supply chain evidence, not with a logo. What each scheme does and does not certify is set out in apparel certifications and sustainability claims.

What a factory cost sheet is made of, line by line

If you take one thing from this page, take this: ask for the quote broken out. A lump FOB price is not negotiable, because you cannot see which line is wrong.

Cost sheet anatomy as a share of FOB, typical export factory.

Worked FOB build: a 180 gsm cotton jersey crew tee

Every line in a $4.30 FOB tee, at finished dyed fabric of ~$8.15/kg and a CM rate of $0.08 per minute.

Check the basis on that fabric line before you compare it to anything. Mills quote fabric either as finished dyed fabric per kilogram, as above, or as yarn plus a separate conversion charge for knitting, dyeing and finishing — so a $5.50/kg quote and an $8.15/kg quote can describe exactly the same cloth. Always confirm which basis a quote is on and restate every mill on the same one, because a cost sheet built on a yarn-basis number quietly omits conversion and understates your largest line. The mechanics are in reading a fabric spec: GSM, yarn count and spinning, the mill-side process in fabric sourcing for brands, mill by mill, typical weights by garment in the GSM fabric weight chart, and the yarn upgrade that moves this line most in combed ring-spun cotton and what the upgrade costs.

Worked FOB build: a 320 gsm brushed fleece hoodie

The same build on a hoodie carrying about 1.1 kg of fleece, reconciled to a $10.20 FOB.

Read the two together and the structure is obvious. A tee is a fabric problem: jersey is 51% of its FOB, at the top of the 35–50% band and a little above it on a fabric-heavy basic, so a $0.50/kg move in yarn price or a 10% miss on marker efficiency is worth more than any labour negotiation you will win. Chasing CM on a knit basic is chasing the wrong 19%. The hoodie is the same shape, harder: fabric is 59% of FOB there, because 1.1 kg of fleece is four times the cloth in a tee.

Costing from SMV: the minute method, and why line efficiency breaks quotes

The way to sanity-check a CM quote is to rebuild it from minutes.

CM cost = SMV × cost-per-minute ÷ line efficiency

SMV is the standard minute value, the engineered time to sew one garment. Cost-per-minute (CPM) is total daily factory cost divided by total productive minutes per day, so it includes wages, overhead, rent, power and machine depreciation, not just direct labour. Line efficiency is the share of standard minutes the line actually delivers.

Standard minute values by garment, indicative of an engineered line.

Now run the tee at $0.08 per minute:

• At a theoretical 100% efficiency: 8.5 × $0.08 = $0.68
• At a realistic 75% efficiency: 8.5 × $0.08 ÷ 0.75 = $0.91
• At 60% efficiency, common on a new style or a new line: 8.5 × $0.08 ÷ 0.60 = $1.13

That is a 66% swing in the labour line, and it is invisible in a lump quote. Line efficiency is the hidden variable. Cost at 60–75%, not 100%. Note too that the cost sheet above carried 10 minutes for a tee where the published SMV benchmark says 8.5: factories often cost at achievable rather than engineered minutes, which is legitimate and exactly why you ask for the minute count and the assumed efficiency separately.

Indicative CMT price benchmarks in USD per piece, by country.

Use these as a smell test on a quote, not a target. CM ranking no longer decides sourcing country, because the duty layer swamps it — see India vs Vietnam vs Bangladesh vs China.

The development-to-delivery sequence, stage by stage

The full sample sequence runs proto, fit, size set, salesman, garment performance test, PP, sealed, TOP and shipment sample, detailed in the sampling calendar from proto to TOP. At programme level the calendar looks like this.

Lead-time build for a first order at a new factory. The first six rows are the PO-to-FOB build; ocean transit and clearance sit on top of them and take you to PO-to-DC.

Totals, and always say which basis you are quoting. PO to FOB — ex-factory, goods on the vessel — a first order runs 12–20 weeks globally and 10–16 weeks from India or Bangladesh, and a repeat order compresses to 8–12 weeks because sampling drops out. PO to your US distribution centre adds ocean transit and clearance and runs 15–20 weeks from India and 18–23 weeks from Bangladesh. Planning against the FOB number and receiving against the DC number is the single most common calendar error in a first programme. Two gates matter more than the rest: no bulk fabric is cut before the PP sample is approved and sealed, and no colour is approved outside a light booth against a current Pantone TCX standard. Where in the process the colour goes in — yarn, piece or finished garment — changes both that calendar and the minimum, and is compared in garment dye vs piece dye vs yarn dye.

Development is a real cash line: a first tee sample runs $200–350 and a hoodie $300–600, two to four rounds are normal, and roughly 60% of physical samples never reach production.

What direct sourcing actually puts on your desk

Removing the intermediary does not remove the work. It moves it. Five jobs land on you.

Quality control. You specify the AQL, book the inspection and decide what to do with a failed lot. Pre-shipment inspection runs $290–360 per man-day with QIMA and $600–1,200 with SGS, Bureau Veritas or Intertek, and rework after a failed inspection typically costs 4–9% of order value. See AQL inspection for brand owners.

Compliance documentation. Fibre content, country of origin, RN or legal name and care instructions are mandatory under the Federal Trade Commission's Textile Act and Care Labeling Rule, and the reasonable-basis requirement attaches to the finished garment, not the shell fabric. On top sits the UFLPA file: yarn-forward traceability, gin and bale documentation, mill declarations. CBP uses documentary review plus isotopic and genetic testing of cotton fibre to test origin claims.

Logistics. Booking, ISF filing 24 hours before lading, drayage, demurrage and detention. An inaccurate or late ISF carries liquidated damages of $5,000 per violation, and the importer of record stays liable even when a broker files it.

Duty and classification. Under 19 CFR 102.21 the cut-and-sew country is almost always the country of origin for apparel, the legal mechanism that made China+1 work. As of 16 August 2026 the Section 301 forced-labor tariffs effective 24 July 2026 add 10% for India, Bangladesh, Cambodia, Indonesia, Pakistan and Sri Lanka and 12.5% for China and Vietnam, on top of MFN. On a cotton knit tee at 16.5% MFN that is roughly 26.5% all-in from India or Bangladesh and 29.0% from Vietnam. Treat those as illustrative of method and verify against the current HTSUS. Full stack: landed cost for apparel imports.

Importer-of-record liability. If you buy FOB or FCA at origin, your brand is normally the importer of record: your bond, your entry, your classification, your penalty exposure. No commercial agreement with a supplier or agent transfers it. Work through it in should your brand be the importer of record.

When direct sourcing is the wrong answer

Stay with an intermediary or a managed platform if any of these are true:

• You are below the volume a tier-1 factory will service. A Bangladeshi tier-1 quotes 5,000–10,000 pieces per style. If your buy is 300, direct access to that factory does not exist at any price.
• You have no one to do the merchandising. Chasing lab dips, resolving fit comments, tracking trim delivery, managing the critical path. If nobody owns it, the calendar slips and the slip costs more than the commission.
• You are entering a country you do not know. Language, time zone, festival calendars and payment norms are learnable, but not on your first order.
• Your assortment is wide and shallow. Many styles at low depth is the worst case, because every style carries its own development, sampling and minimum.
• You cannot carry the working capital. FOB terms usually mean a deposit against fabric booking. Agents sometimes bridge that. Factories generally do not.

The counterpoint, and the reason to revisit the decision annually: what an agent absorbs is coordination work — chasing, converting, re-keying, reconciling. That is unglamorous and repetitive, which is the definition of a workflow that can be automated rather than staffed. The arithmetic is in what a buying house actually costs you.

Where to go next in the sourcing cluster

• MOQ economics and how to move your minimum
• The complete apparel tech pack guide
• The sampling calendar from proto to TOP
• AQL inspection for brand owners
• Colour approval, lab dips and Delta E
• Reading a fabric spec: GSM, yarn count and spinning
• What a buying house actually costs you
• Private label vs white label, compared line by line
• A tech pack template you can build from
• Finding clothing manufacturers near you
• Fabric sourcing for brands, mill by mill
• The GSM fabric weight chart
• Combed ring-spun cotton and what the upgrade costs
• Garment dye vs piece dye vs yarn dye
• Yarnstick glossary

And sideways, where the sourcing decision meets another one: manufacturers by product category when the question is who makes your specific garment, blanks and wholesale apparel suppliers and custom and decorated apparel when the volume does not yet justify manufacturing, starting an apparel brand, from plan to first order for the sequence before this one, apparel certifications and sustainability claims for what you can defensibly say about the goods, and sourcing software and fashion PLM for running it at scale.

Frequently asked questions