The Complete Guide to Apparel Sourcing

Two things decide most of what an apparel order costs and when it lands, and neither of them is the garment factory. Lead time in apparel is mostly fabric. Cost in apparel is mostly material consumption. Which means almost every choice that looks like a design decision — a heavier cloth, a second colourway, a contrast facing, an extra two centimetres of sleeve — is a cost decision and a calendar decision wearing design clothes.

This guide follows a garment programme in the order it happens: tech pack, materials, mill and factory, minimums, sampling, fit, costing, quality, compliance, production and freight. It is written for brands ordering in the hundreds to the low tens of thousands of units, which is the range where material minimums bite hardest.

Why apparel is not like sourcing a moulded product

Three structural differences shape everything else. The material is bought before the product exists, from a different company than the one that sews it — a mill knits or weaves and dyes, a garment factory cuts and sews, so your order sits on two suppliers with two schedules, two minimums and two definitions of "confirmed". You are buying a size range, not an item: a style that fits beautifully in a medium and badly in an XL is not 80% right, it is a returns problem concentrated at both ends. And cloth is soft, so a specification cannot be a set of dimensions — it has to be measurements with tolerances, at named points, on a garment relaxed in a stated way. For the general shape of a sourcing programme, the product sourcing guide walks the same lifecycle across all categories.

The tech pack, and what an incomplete one costs

Tech pack

The complete manufacturing specification for one style: construction, measurements with tolerances, the bill of materials, colourways, labelling, packing and the quality standard the garment will be judged against. It is the document a factory quotes from, samples from and is held to.

Every question a tech pack leaves open gets answered by somebody on the factory floor, in the way easiest for them, and you find out which way when the sample arrives. That is how a small ambiguity becomes a three-week delay: an unanswered question costs a sample round, and a sample round costs a courier leg each way plus a slot in someone else’s sample room.

• What a production-ready tech pack carries
• Flat sketches front and back with every seam, stitch and closure visible, plus detail views for anything unusual.
• A construction page: seam types, stitch density, thread specification, seam allowances, topstitch positions.
• A points-of-measure list with the base-size measurement, a tolerance and a measuring method for every point.
• The graded size chart across the full range you intend to sell.
• A bill of materials naming every component with supplier, article reference, composition, weight, width, colour and consumption.
• Colourways, each with a colour standard the mill can match to — not a screen colour.
• Trim and label placement, including care and composition labelling for the markets you sell into.
• Artwork at production scale, dimensioned from a fixed reference point; and packing — fold, polybag, hangtag, carton assortment, ratio, marking.
• The quality standard: inspection level, acceptance rule, defect classification, and any testing required.

The bill of materials is the part most often left thin, and the part that drives the cost. A line reading "cotton jersey, black" is not a specification — it is a request for the cheapest cotton jersey the factory can find, and it will be a different cloth in the second order. A usable line names composition, weight, finished width, finish, and either a mill and article number or a physical reference the mill can match.

Fabric: where the lead time actually lives

When a garment order is late, the fabric is usually why — not because mills are unreliable, but because the fabric leg of the calendar is long, sequential and outside the garment factory’s control. Yarn has to be available. The cloth has to be knitted or woven, dyed to an approved colour in a lot large enough for the dye house to run, then finished, inspected and shipped to the factory. Only then does cutting start. Which is why "the factory says four weeks" is almost never the answer to "when can I have it": four weeks is the sewing window, and the date that matters is when bulk fabric reaches the cutting floor. That answer belongs to the mill.

Fabric MOQ and garment MOQ are different numbers

This surprises new brands more than anything else in apparel. A garment factory quotes a minimum in pieces per colour per style — below which a line changeover costs more than the order earns. A mill quotes metres or kilograms per colour per quality — below which a dye lot is not worth running. Different economics, and they do not agree.

Both numbers are examples, not quotations. The arithmetic is the point: divide the mill minimum by your consumption per unit and compare it with the factory minimum, before you commit to a colour range. A four-colour launch at 300 pieces per colour is a 1,200-piece order to the factory and a four-dye-lot order to the mill.

Three honest ways out. Choose a stock-service quality the mill holds in the colour you want, so no dye lot is needed. Reduce colourways and put the volume behind fewer lots. Or accept the residual — buy the minimum, cut what you need, carry the leftover against a repeat you have actually planned. A short-lot surcharge is real too, but it lands in the fabric price and therefore in the largest line of your cost sheet.

Colour is approved separately from fit

Colour runs on its own approval track and its own clock. For solids the mill submits lab dips — small swatches dyed to your standard. For prints it is a strike-off on the actual base cloth; for yarn-dyed wovens a handloom or knit-down. Each round is a real turn of the dye or print process, so the number of rounds your colour takes is a schedule item, not an administrative one. Two details save argument later: state the light source colour is assessed under, and agree a shade-banding expectation, because dyeing varies between lots and across the width of a roll.

Trims are cheap, small, and the usual reason a shipment waits

Trims are a rounding error in the cost sheet and a first-order risk in the calendar. Cutting cannot finish into an empty trim store, and a garment cannot be packed without its labels. The list is longer than most brands expect: thread, interlining, zips and pullers, buttons, rivets, snaps, drawcords, elastic, tapes, care and composition labels, size and brand labels, hangtags, polybags, cartons.

Two decisions matter. Nominated or factory-sourced? Nominating buys consistency across factories and seasons and makes you responsible for that supplier’s lead time; letting the factory source buys their relationships and prices and makes trim consistency something you inspect for. Either is defensible; not deciding is not. And which single trim has no second source? A custom puller, a woven label at a specific loom width, a zip matched to your fabric — anything with one supplier is a single point of failure on your critical path.

Mills and factories: who you are actually buying from

The word "factory" hides three very different arrangements, and which one you are in changes your minimums, your price, your working capital and how much coordination you personally own.

This is why the mill is often the real constraint, not the garment factory. Sewing capacity is comparatively fungible: a competent factory can usually find a line. A specific cloth, in a specific weight and finish, dyed to a specific colour, at a lot size you can afford, from a mill with a slot — that is the narrow part of the funnel. A shortlist of ten factories that cannot buy your fabric at your quantity is a shortlist of zero.

Validating either kind of supplier is the same discipline: the legal entity, the site, the capability at that site rather than at the group, the capacity in your season, the audit and testing history. The factory vetting guide covers the process; how to vet an overseas manufacturer covers what each document does and does not prove.

The sampling ladder

Samples are not one thing. Each rung answers a different question, and skipping a rung means the question it answered gets answered by your customers.

1. Proto sample

   Made from whatever comparable cloth is available, to test whether pattern and construction produce the garment in the sketch. Judge shape and proportion, not colour or precise fit — the fabric is wrong on purpose.

2. Fit sample

   The base size in correct or close-substitute fabric, on a consistent form or fit model. Measurements and tolerances settle here. More than one round is normal; planning for one round means being late.

3. Salesman or photo sample

   Correct fabric, colour and trims, made to sell from and shoot rather than on the production line. A commercial asset, not a manufacturing approval.

4. Pre-production sample

   The one that matters most: made in the factory that will run the order, from bulk fabric and actual bulk trims, on the equipment that will produce it.

5. Golden sample

   The approved reference, sealed and held by both sides, against which every inspection is judged. Without one, "acceptable" is an opinion — see what a golden sample is.

6. Top-of-production sample

   Pulled from the first real output and checked against the golden sample before the run continues. It catches the drift between a careful pre-production piece and what a line produces at speed.

Most factories charge for samples and many credit the charge against a placed order; some charge for the pattern separately. Ask in writing before the first round, and treat courier time between rounds as calendar — on a multi-round fit it often exceeds the sewing time.

Fit and grading across the size range

Fit is approved on one size. Production is sold in all of them. Grading takes the approved base-size pattern out to the full range, and it is a technical decision with commercial consequences rather than a spreadsheet operation. A uniform grade — the same increment at every point for every size — fits the middle and distorts at the ends, because bodies do not scale linearly: armhole depth, neck width, shoulder slope and rise do not grow at the rate the chest does. The symptom is not a bad review; it is returns concentrated in the smallest and largest sizes while the middle sells clean.

• A graded size chart across the full range with the base size identified — not a base measurement and a note saying "grade as usual".
• A tolerance per point of measure, tighter where a customer notices and looser where cloth relaxation dominates, plus a stated measuring method for each.
• Fit checks on at least the smallest and largest sizes, not only the base — those two are exactly what the grade tests.
• Agreement on how the fabric relaxes and shrinks after wash and finishing, since the tolerance has to absorb it.

Costing a garment from consumption, not from a guess

A garment price is built, not quoted. Every line in the build-up traces back to a number in the tech pack, which is why the tech pack is a costing document as much as a manufacturing one. The single most important input is consumption.

Consumption

The quantity of a material a single finished garment uses, including the cloth lost in cutting. It comes from laying the graded pattern into a marker at the fabric’s usable width and measuring the result — not from measuring the garment and adding a margin.

Material cost per unit = consumption per unit × material price + yield allowance

Consumption depends on the pattern, the fabric’s usable width, the size ratio in the marker, and whether the lay is one-way. A large print repeat, a nap forcing a single direction, or a stripe that must match at the seam all push it up before anyone has changed the design.

That is the practical heart of apparel costing. Narrowing a fabric from 150 cm to 140 cm, or matching a check across a front seam, moves consumption by a few percent — and because material is the largest line in the sheet, a few percent there outweighs most of what anyone will find by negotiating labour.

Read the Share of FOB column before the value column. Here the first line alone is 44% of the price and materials together are over half, so a 10% move in fabric price changes FOB by more than a 10% move in the cut-and-make rate. Your proportions will differ by product; the method of building the sheet does not.

Two lines deserve suspicion whenever a cost sheet arrives. A round consumption figure usually means nobody has made a marker, so the number will be corrected upwards once someone does. And a flat yield allowance across every fabric ignores that slippery, stretchy and directional cloths waste different amounts.

FOB is also not what the garment costs you. Freight, duty, clearance, insurance and inland delivery sit on top, and they land differently on a light, high-value garment than on a bulky one. The landed cost guide takes the build-up through to a cost per unit in your warehouse.

Quality: what an AQL inspection decides, and what it cannot

Apparel inspection almost always runs on sampling rather than checking every piece. The standard is ISO 2859-1, *Sampling procedures for inspection by attributes — Part 1: Sampling schemes indexed by acceptance quality limit (AQL) for lot-by-lot inspection*. The lot size and inspection level select a sample size; the AQL you nominate selects the acceptance and rejection numbers for that sample. If your quality manual cites a year, note that the 1999 edition was withdrawn on 22 January 2026 and a 2026 edition has replaced it.

An AQL is a sampling and acceptance rule, not a quality target. It does not say a percentage of defective garments is acceptable to your customer; it says how many defects in a sample of a given size are consistent with a lot at that quality level. Buyers commonly classify defects as critical, major and minor and set the rule differently for each, with critical at zero tolerance. Put the classification in the tech pack, because an inspector without a written one will use their own.

What a visual inspection cannot decide: whether the garment fits the size chart across the range, whether the colour will survive washing, whether the fabric will pill, shrink or bleed, and whether the label composition is the composition in the cloth. Those are measurement and laboratory questions, specified separately and run on samples drawn from bulk. The manufacturing quality control guide sets out the full programme, and quality control checks in manufacturing covers what each stage is for.

Compliance: what is law, and what is a scheme you chose

Apparel buyers routinely mix two different kinds of requirement in one sentence, and the confusion is expensive in both directions — treating a voluntary certificate as though it discharged a legal duty, or treating a legal duty as though a certificate would cover it. They are not the same category of thing.

• GOTS label grades are threshold-based: "Organic" needs at least 95% certified organic fibre, "Made with x% organic materials" covers 70–94%, and below 70% a product cannot be GOTS-certified at all.
• GRS can be used business-to-business at 20% recycled content or more; consumer-facing GRS labelling requires at least 50%.
• Textile Exchange has published criteria for a Materials Matter Standard superseding its current recycled standards — released 12 December 2025, effective 31 December 2026, mandatory 31 December 2027. A date to know if you are writing a multi-season claim into artwork or a contract.
• A voluntary certificate is only as strong as its scope, its expiry date and whether the certified entity is the one that made your goods. A legal obligation applies whether or not anyone audited you — see manufacturing compliance for the EU market.

The production calendar, and what sets each duration

A garment calendar is a chain of prerequisites, not a list of durations. The most common planning error is treating the stages as parallel when they are sequential; the second is planning the sewing window carefully and the fabric window optimistically.

Nothing here is a duration, deliberately: durations depend on the mill, the factory, the season and the product, and a number invented here would be worse than no number. The two lines that most often move a delivery date are the two fabric lines, and neither is under the garment factory’s control.

Where apparel programmes usually go wrong

• Planning the colour range before checking the dye-lot minimum. The most common way a viable order becomes an unviable one.
• Approving a fit sample in substitute fabric. Drape, weight and stretch are most of how a garment fits, so that approves the pattern only.
• Costing from a guessed consumption. Until someone has made a marker at the real width, the largest line in the sheet is a placeholder.
• Grading uniformly. The middle of the range will be fine and the ends will come back.
• Letting the tech pack lag the samples. A decision agreed in email and never written into the pack will not survive into the second order.
• Nominating a single-mill fabric with a long lead time, then negotiating hard on the garment price. You have moved the risk to where you have least leverage.
• No golden sample. Every later quality conversation becomes a matter of recollection.

A checklist before you place the order

• What should be true before a purchase order goes out
• The tech pack is complete, with tolerances and a defect classification, and matches what was agreed at the last sample review.
• Consumption per unit comes from a marker at the real usable width, for the real size ratio.
• The fabric minimum divided by consumption is compared against the quantity per colour, and you know who carries any residual.
• Colour is approved on bulk-representative material, under a stated light source, with a shade-band expectation agreed.
• Every trim has a named supplier, a lead time and, where it exists, a second source.
• A pre-production sample was made in the factory that will run the order, from bulk fabric and bulk trims, and the golden sample is sealed.
• The inspection level, acceptance rule and test list are named in the order, not left to the inspector.
• The cost sheet is a build-up you can trace line by line to the BOM, not a single FOB figure.
• The fabric delivery date to the factory is a milestone in the calendar, with an owner.

How Library of Trade works an apparel brief

Library of Trade manages the programme rather than introducing you to a factory. A brief becomes a specification and a bill of materials; materials and upstream suppliers are sourced against it; factories are matched on capability, capacity, price, quality and compliance; samples are developed and approved; production, quality and freight are managed through to delivery. You contract with your suppliers directly — this is not a marketplace and not a directory. Software structures the specification, compares options across suppliers and holds one version of what was approved; the judgements — whether this mill can hold this colour, whether a factory’s capacity claim survives your season, whether a cost sheet is honest — are human work, and we keep it that way.