OEM clothing manufacturing for fashion brands: a practical guide

A first order proves capability. A reorder proves consistency, and consistency is what protects your size chart, your returns rate and your customer's willingness to buy the same size twice. Most of what goes wrong in OEM at scale is not a manufacturing failure; it is an asset-management failure, where something that should have been retained and reused was re-created instead.

This article assumes you have already chosen OEM and want to run it well. If you are still establishing what the model is, start with what OEM clothing manufacturing is, and for the comparison against ODM see the difference between the two.

Quick answer: what does running OEM well require?

Four retained standards and one calendar habit. Retain the fit-approved pattern, the approved colour swatch, the pre-production sample and the packing specification. Build the production calendar backwards from your on-sale date with deliberate float. Together these make a reorder a repeat rather than a re-creation, which is the difference between a size chart that holds and one that drifts.

None of the four is expensive to keep. All four are expensive to lose, and the loss is usually silent — you discover it a season later, in the returns data.

Why reorders fail: the asset-not-expense problem

Brands routinely treat the pattern as a cost incurred once, in the same category as the sample fee. It is better understood as an asset, in the same category as the design. A pattern that has been fitted and approved on a body carries information that cost real money to acquire, and re-creating it next season is paying for the same information twice — usually with a worse result, because the second version is fitted against a memory rather than a garment.

The same logic applies to colour and to packing. An approved swatch is a physical standard that a reorder can be measured against. Without it, shade approval becomes a conversation about a photograph, and shade conversations that rely on photographs end in disagreement roughly as often as they end in agreement.

Building the calendar backwards from your sell-through

A production calendar that starts from today and runs forwards will end late, because it has no room in it. Build it backwards from the date the goods need to be on sale, and add the blocks in reverse order.

1. On-sale or delivery date. The fixed point everything else is measured against.
2. Shipping and clearance. Freight transit, customs and inland delivery to your warehouse.
3. Bulk production. Roughly four to six weeks after sample approval for our production.
4. Sampling. Around seven to ten days from tech pack approval, plus the time your comments take to come back.
5. Development. Pattern, construction decisions and fabric — longer on a first style, shorter once a block exists.
6. Float. Deliberate, protected, and not available for anything else.

The float is the part brands delete when they are late, and it is the part that absorbs the delays which are statistically certain to occur. A calendar without float is not tight; it is unhedged.

Where the calendar time actually disappears

• Samples that arrive before a weekend. Two or three days lost before anyone looks at them, repeatedly, across a season.
• Comments arriving in fragments. Five messages over four days cost more calendar time than the comments themselves.
• A fabric change after sampling. It restarts the development clock for that style and invalidates the fit work done against the old cloth.
• A colour added late. A new colourway needs its own fabric, its own cut and its own shade approval, and it arrives at the point where the schedule has least room.
• Approval deferred rather than given. A sample that has been “looked at” but not decided stops the style completely.

Note that every one of these is a decision-timing problem rather than a capacity problem. Factories are rarely the reason a season runs late; they are usually the messenger.

Specification discipline is the whole game

At first order, a good specification gets you a correct garment. At tenth order, it is the only thing preventing silent drift. Four elements carry most of the weight.

Point of measure with tolerances

The measurement at each point, for each size, with an acceptable range at each point. “About 120cm” cannot be checked, only disputed. A tolerance converts a subjective argument into a measurement, and measurement is the only quality control that scales.

Construction described from the inside

Lining, interlining, stay tape, boning, facing, closure specification and hem method. These are where interpretation happens, because they are invisible in the photograph that usually accompanies a reorder. A specification that describes only the outside invites a different inside.

Fabric named by behaviour

Weight, stretch, recovery and drape, not only a trade name. The same name covers a range of cloths, and two of them can produce visibly different garments from an identical pattern. Where a specific article matters, name it and keep a swatch; where it does not, describe what the cloth must do.

One dated version of the truth

A single document with a version date, replacing any earlier version explicitly. Specifications distributed across email threads, chat messages and verbal agreements are not specifications. The failure is not that the information is wrong; it is that two people can each be correct and still disagree.

From a manufacturer's perspective: what a brand's specification does to the plan

Inside the factory, the planning problem for a reorder is different from the planning problem for a new style, and the difference is entirely about what information survived. A reorder with a retained pattern, a retained swatch and a retained pre-production sample needs almost no development time. It goes straight into the production plan at a known cost per operation, on a marker that has already been proved efficient.

A reorder without those assets is a new development wearing the same style number. It needs the pattern re-created or re-fitted, the shade re-approved, and the construction re-interpreted — and each of those consumes the calendar time that the brand believed it had already banked. The factory is not being difficult; it is being asked to buy back information it gave away with the last order.

This is why the two questions worth asking a manufacturer at the start of a relationship are not about capacity. They are: will you retain the approved standards between orders, and who holds the approved pattern? Both answers should be yes and in-house. A factory that cannot answer them clearly is a factory that will re-create your garment each season rather than repeat it.

Example: three seasons of the same style

Consider one style carried across three seasons. The garment is not redesigned; only the way it is specified and ordered changes.

Season two is the dangerous one, because nothing obvious is wrong. The garments pass inspection, the measurements are within a normal band, and the difference only appears in aggregate — as a slightly higher return rate and a slightly lower repeat purchase. By season three the size chart has genuinely moved, and the brand is explaining to customers why a medium is not a medium.

Making a reorder actually a repeat

A reorder instruction should be short, because it is a reference to existing standards rather than a new specification. If it is long, something has been lost.

• Name the style and the season it was approved in. Not the season you are ordering in.
• Confirm the retained standards by reference. Pattern version, approved swatch, PP sample, packing specification.
• State only what changes. If the answer is “nothing”, say nothing and expect an identical garment.
• Confirm the fabric position. Stocked, still available, or substituted — and if substituted, re-approved before cutting.
• Repeat the size ratio. Grading depends on it, and ratios often shift between seasons.
• Give the delivery window. So the blocks can be laid against it rather than around it.

If any change is unavoidable, the moment to make it is before sampling rather than after. A fabric substitution made after sampling invalidates the fit work done against the old cloth, which means the reorder carries development time after all.

Scaling without losing the standard

As volume grows, the risk migrates from capability to coordination: more styles, more colours, more sizes, and more people with a partial view of the specification. The defences are unglamorous and they work.

• Freeze the specification at pre-production. Changes after that point cost schedule as well as money.
• Require one written answer. Anything agreed verbally is confirmed in writing the same day.
• Keep the colour count honest. Every added colourway is a new fabric position, a new cut and a new shade approval.
• Consolidate rather than fragment. Two lines in one order usually outperform three styles in three orders, for the same reason minimums behave as they do.
• Retain standards centrally. Physically, in one place, labelled by style and season — not inside an email attachment.

The last point is the one most brands get wrong, and it is the cheapest to fix. A physical swatch library, a set of approved pre-production samples and a versioned specification folder will do more for consistency across a season than any process change at the factory.

Common mistakes established brands make with OEM

• Redrafting a style that already fits. It is the fastest way to lose a size chart, and the savings are illusory.
• Losing the swatch. Shade approval then depends on photographs, which is not approval.
• Changing fabric mid-season. Drape changes, so fit changes, so the size chart changes.
• Adding colourways late. Each arrives with a fabric position and a cutting slot you have not planned for.
• Treating the PP sample as a formality. It is the physical standard for the entire run and every reorder after it.
• Breaking the specification across channels. Two versions of the truth guarantees a dispute in which both parties are right.
• Averaging supplier relationships. Running the same style across two factories splits the retained standards and doubles the drift.

What to prepare before each season

A short checklist, run before the season rather than during it. The value is in doing it early enough that a gap can still be closed without air freight.

• Confirm which styles are being repeated, and retrieve their approved standards
• Check whether the fabrics are still available, and whether any are being discontinued
• Settle the colour count and the colourways before sampling begins
• Confirm the size range and the ratio, including any extension into curve sizes
• Agree the delivery windows for each drop against your on-sale dates
• Confirm the packing and labelling requirements for each sales channel
• Book sample capacity before your peak, not during it
• Identify which styles are genuinely new and need development time rather than a repeat

Frequently asked questions

Why does the same style fit differently two seasons later?

Almost always because the approved pattern, the swatch or the pre-production sample was lost or re-created rather than retained. The second version is fitted against a memory, and small differences accumulate. Retaining the fit-approved pattern and reordering from it is the only reliable defence.

What should a brand ask its factory to keep between orders?

Four things: the fit-approved pattern, the approved colour swatch, the pre-production sample, and the packing specification. Ask for confirmation that all four are retained in-house, and ask to see them at the start of a reorder. All four should be cheap to keep and decisive when they are missing.

How far ahead should an OEM calendar be built?

Work backwards from the on-sale date and include shipping, clearance, bulk, sampling and development, then add float. Sampling runs around seven to ten days from tech pack approval and bulk roughly four to six weeks after sample approval in our production; everything around those figures depends on how quickly decisions are made on your side.

Can I change the fabric on a repeat order?

You can, but it is not a repeat — it is a new development for that style. Changing the cloth changes drape and fit behaviour, so the fit work has to be done again and the size chart should be re-confirmed. Do it before sampling, never after, and treat it as a development cost rather than a substitution.

Should I split one style across two factories?

Generally no. Running the same style in two places splits the retained standards, and the two versions will drift apart because each factory is measuring against its own copy. Splitting capacity across styles is more defensible than splitting a single style across suppliers.

When should the specification be frozen?

At pre-production. Up to that point changes are development and cost time in an expected way. After it, changes invalidate the standard the bulk is being inspected against, which means the inspection itself stops meaning anything.

When should a brand consolidate its supplier base?

When coordination cost starts to exceed any benefit from the spread. More suppliers means more versions of the standard, more shade approvals and more incoming inspection. Consolidating into a factory that holds the blocks and retains the standards usually produces a more consistent garment, not just a simpler process.

Can a brand move OEM production to a new factory?

Yes, and it is much easier when the specification, the approved swatch and the pre-production sample are held by the brand rather than only by the factory. That is one practical reason to keep your own set of retained standards: it is what makes your production portable.

How we run OEM for repeat brands

Most of our work with established brands is repeat production, which makes retention the central part of the relationship rather than an afterthought.

1. We hold the approved standard. Pattern, swatch and pre-production sample are retained against the style, in the building.
2. Reorders start from the retained pattern. Not from a re-draft, unless you ask for one.
3. We confirm what is unchanged before quoting. A repeat should be priced as a repeat.
4. We name anything that has moved. Fabric availability and shade positions are checked at the point you place the order, not at cutting.
5. We inspect against your standard. The pre-production sample from the original order, not a fresh interpretation.
6. We flag the calendar early. If a delivery window is tight, you hear it before the order is confirmed.

Luxudress has manufactured womenswear in Guangzhou since 1995, with pattern, cutting, sewing, finishing and packing under one roof. That structure is what makes retention workable: the pattern maker who approved your fit is in the same building as the line that repeats it, so a reorder comes back matching the sample you approved rather than matching a description of it. OEM manufacturing covers the service in detail. If you are consolidating suppliers or moving a category, send us your size chart and an approved garment and we will tell you what it would take to hold that standard across a season.

Conclusion

OEM at scale is an asset-management discipline. The pattern, the swatch, the pre-production sample and the packing specification are the four things that make a reorder a repeat, and they are cheap to keep and expensive to lose.

Build the calendar backwards with float you protect. Ask your factory to retain the standards and to prove they have them. Then a reorder becomes what it should be — the same garment, produced again — rather than a new development that happens to share a style number.