How to source a clothing manufacturer for a new fashion brand

This is the sequence: what to do in what order, and which steps a first-time buyer usually does too early. Where factories are found and how they are verified is finding a reliable factory, the evaluation criteria are choosing a manufacturer, and the questions to ask are what to ask a manufacturer.

The reason to follow an order rather than work on all of it at once is that each step makes the next one cheaper. A new brand that specifies the product before contacting anyone will get comparable answers from the first message. One that starts with a list of suppliers will spend a fortnight being politely ignored.

Quick answer: the sequence

Steps one and two are the ones new brands skip, and they are the ones that determine whether steps four onwards are useful. A brand with a specification can filter suppliers in an afternoon; a brand without one can only collect quotations.

Step 1: specify before you search

A specification is not a tech pack. It is enough detail that a factory can answer a question about your garment rather than about garments in general. For a first enquiry that means the silhouette, the fabric, the sizes, the finish standard and the date.

The reason this comes first is mechanical. Manufacturers are not ranked by quality in any list you can look up — they are matched to garments. The same factory can be the best possible choice for a structured satin gown and a poor one for a soft knit. Until you know what you are making, you cannot tell which is which.

Step 2: settle the fabric

Fabric is the single largest reason a first quote turns out to be wrong. It decides drape, construction method, availability and cost, and a change to it after sampling invalidates the fit work and restarts development for that style.

• An article and a swatch is the ideal — costing and booking both become accurate.
• Behaviour described — weight, drape, stretch, recovery — is the practical middle ground.
• Naming a fabric characteristic (“satin”, “crepe”) is close to no information, because both names cover a wide range of cloths.

If you are new to fabric, physically handling a few options before writing the brief will change what you ask for. Swatch books from a mill or a fabric market cost very little and settle questions that no amount of description resolves.

Step 3: build a long list from two channels

Use two channels rather than one, because they have different failure modes and the overlap is informative. A platform gives breadth and is easy to game; a fair or a referral gives depth and is hard to scale. Coming from both means the names that appear in both are worth a closer look.

The list does not need to be long. Thirty names is plenty, and it should take an afternoon rather than a fortnight. What matters is that you resist the temptation to keep adding — every additional name is another reply to read and another follow-up, and the quality of the list matters more than its length.

Step 4: filter with one technical question

This is the step that does the most work for the least effort. Send every name on the list the same short message: the category, the construction, the fabric, the quantity, the date — and one technical question about your own style.

A question such as “how would you keep this cowl from collapsing?” is close to impossible to answer convincingly without production knowledge. That is why it filters so well: it cannot be bluffed, and the replies sort themselves.

What to do when nobody answers well

If the replies all come back vague, the problem is usually the message rather than the market. Three things cause it, and all three are fixable in an afternoon.

• No specification. Without a fabric, a size range and a finish standard there is nothing concrete to answer, so a supplier can only describe itself.
• No question. A brief that asks for a price invites a price. One that asks how a specific construction would be solved invites a technical answer.
• Sent to the wrong tier. Both large exporters and small workshops answer slowly to enquiries they expect nothing from. Match the size of the supplier to the size of your order.

There is a fourth possibility worth naming: the list was built from a single channel, and that channel is populated by intermediaries. If every reply sounds the same, the replies are probably coming from the same kind of company. Adding a second channel — a fair, a referral, or a brand whose product resembles yours — usually breaks the pattern.

Step 5: verify the survivors

By this point you should have a handful of names. Verification is a set of small independent checks: the address, a recorded walk-through of the floor, who makes the pattern, and a reference in a similar category. The contradictions between those answers are more informative than the confirmations, especially on a first enquiry where everyone is on their best behaviour.

Step 6: sample two

Sampling is the most expensive part of sourcing, for the factory as well as for you, so it belongs at the end of the process rather than the start. Two is the right number: enough to compare, few enough to keep the cost proportionate.

• Judge the sample against your specification, not against the other sample. The question is which one meets the brief.
• Ask what they would change. A supplier with a view on your construction read it carefully.
• Notice how they handled the ambiguity. Every brief has one, and the response to it is the best available prediction of the whole relationship.

Step 7: the first order

A first order should be technically modest and fully specified. Its purpose is to test the process — the communication, the tolerance discipline, the inspection — rather than to prove the supplier can execute your hardest style. A new brand that opens with its most ambitious design risk everything.

Keep the first order narrow. Fewer styles, fewer colours, sizes you can actually sell. Our material minimum is 300 pieces per design with 100–200 per colour, and the per-colour figure reflects the mill's dye batch — which is why reducing colourways is usually the most effective lever a new brand has.

Manufacturing perspective: what signals a new brand

Factories read new brands quickly and mostly sympathetically. Three signals change how an enquiry is handled.

• A complete brief gets a real answer. It also tells the factory you intend to be a workable customer, which affects how much effort goes into the response.
• Narrow first orders are welcome. A small, well-specified run is easier to sequence than a large, vague one, and many factories will take it to establish a relationship.
• Decisive brands get priority. The single fastest way to lose a production slot is to leave a sample unapproved. Approval timing is the variable that most affects when work can actually be scheduled.

The three mistakes that cost the most

• Contacting suppliers before specifying. It makes every reply incomparable and turns a week of work into a spreadsheet of unrelated numbers.
• Sourcing on price alone. The lowest quote is usually the least specified garment, and the gap reappears as freight, rework or returns.
• Testing with the hardest style. A first order is a test of process; using your most demanding design makes a normal learning curve look like a failure.

The rest of the common errors — spreading quantity across too many colours, deciding the fabric after sampling, leaving the packing specification until shipping — are variations on the same problem: something that is cheap to decide early being decided late.

How we work with new brands

Luxudress has manufactured womenswear in Guangzhou since 1995, producing over 100 new styles a month. First orders from new brands are a normal part of our work, and the most useful thing we can do is tell you what is still open in your brief.

1. We reply within 24 hours, with technical answers rather than an acknowledgement.
2. We name the gaps. A thin brief gets questions, not a cautious price.
3. We tell you when a style is a poor first order. Some categories are better attempted once the relationship works.
4. We quote a narrow first run seriously. A small order is a reasonable way to start.
5. We say what will make it late. Fabric and decision timing, in that order, almost every time.
6. We retain the approved standard. So the second order is a repeat rather than a new project.

If you are starting out, send us the design and the fabric you are considering. Working with startups explains how we handle first-time buyers, and our production process shows the stages you would be buying.

Frequently asked questions

I have no manufacturing experience. Where do I start?

With the garment, not the supplier. Write down the silhouette, the fabric, the sizes, the finish and the date. You will not get every detail right, and that is fine — the point is to have enough that a factory can answer a specific question about your product rather than a general one about manufacturing.

How many manufacturers should I contact?

Ten to thirty names on the initial list, filtered down to three or four. Sending the same short brief with one technical question to all of them produces a usable shortlist in about a week, without spending anything on samples.

Should I choose the cheapest manufacturer?

The cheapest quote is usually the least specified garment, so the comparison is not what it appears. Compare total landed cost and the construction actually being offered, and weight capability above price — a factory that has made your category before is not learning on your order.

What kind of style should my first order be?

Technically modest and close to something the factory already makes. Sequins, hand embellishment and complex structure are not first-order categories; they are where a new relationship is tested hardest. Choose the style with the fewest open questions.

Can I start with a very small order?

Often, if the styles are similar. Material minimums still apply, because fabric is dyed by the batch — our standard is 300 pieces per design with 100–200 per colour. The effective way to go smaller is to reduce colourways rather than pieces, and to keep the styles close enough that they share fabrics and construction.

How long should sourcing take?

A fortnight from a settled specification to a defensible shortlist, and a little longer if you sample two candidates. Sourcing that runs for two months is usually not searching for a better factory; it is avoiding a decision about the garment.

Do I need a sourcing agent?

Only if language or verification is genuinely blocking you. An agent cannot supply the production opinion that makes a factory good, because that is not theirs to give. If you do use one, ask who pays them and whether they contract or introduce.

Conclusion

Sourcing is a sequence, and its order matters more than its effort. Specify, settle the fabric, build a list from two channels, filter with one technical question, verify what survives, sample two, and place a modest first order.

The new brands that struggle are almost never the ones that asked too few manufacturers. They are the ones that started talking to factories before they could describe the garment, and so spent a month collecting answers to questions they had not yet asked.