Second Source Clothing Manufacturing: How to Split a Style Without Losing the Fit

Dual sourcing is normal in this industry and it is usually discussed badly. Suppliers present it as a betrayal, brands present it as prudent risk management, and both sides avoid the only question that matters: if this style is produced somewhere else, what has to cross the gap for the result to be the same garment?

The gap is larger than most buyers assume. A style is not a tech pack. It is a tech pack plus a pattern that has been corrected over several sample rounds, plus a set of decisions that were made at the machine and never written down, plus a material source, plus an inspection standard, plus the accumulated knowledge of what went wrong the first time. Some of that travels easily, some of it travels only if someone deliberately packages it, and some of it will be lost no matter what you do.

Knowing which parts are which is what makes a second source work. It also makes the relationship with the first source easier, because a brand that understands what duplication actually costs is a brand that can be honest about it.

Why Brands Second-Source

The four reasons to duplicate a style are not equivalent, and they lead to different amounts of work because they demand different things from the second supplier.

The distinction that matters is completeness. A second source for capacity is producing the identical garment and therefore needs the whole file. A second source for a customer requirement may be producing a variant, in which case the file can be narrower — but a variant that is only partly documented is a variant that will drift.

It is worth being clear-eyed about the price reason in particular. A second source brought in purely to create a comparison point, without a genuine intention to place volume, tends to produce a quotation that is not comparable and a relationship that is not real. If the goal is a price benchmark, that is better achieved by understanding the cost structure than by running a competitive sample — the structure of a garment's cost is covered in how a costing sheet is built.

What Actually Has to Travel

A second-source handover is not a document. It is a package, and it has pieces that buyers routinely leave behind.

• The corrected pattern, not the original. The pattern that was used in production has been corrected since it was drafted, and the corrected version is the one that produces the fit. This is the single most commonly missing item.
• The approved sample, physically. Not photographs of it. A second supplier has to hold a physical reference piece to compare against, and one reference is rarely enough across a size range.
• The measurement sheet as approved, with the tolerances that were agreed and the points that are measured, not a generic table.
• The materials, with sources and references. Fabric, lining, interfacing, thread, trims, closures, components. Each has a mill, a reference and a lot history, and each is a place the second version can differ without anyone intending it.
• The construction details that were decided in the sample room. Machine settings on the decisive operations, seam finishes, the order of assembly in the areas where the order matters, and the pressing method.
• The inspection standard. What was checked, at what point, against what tolerance, and what was accepted.
• The known problems. Every style has at least one operation that caused trouble on the first run. That knowledge is the most valuable thing in the package and the thing most likely to be withheld.

The last item deserves emphasis because it is countercultural. Brands worry that telling a new supplier about a problem invites that supplier to price the problem in or to blame the previous factory. What it actually does is prevent the second supplier from rediscovering the same problem at your expense. A handover that omits the difficult operations is a handover that guarantees the second run repeats the first run's hardest month.

There is also a version-control question that is easy to skip. Once two sites are producing the same style, any change has to be communicated to both, and any change made at one site has to be reflected in the file that the other site holds. Without that discipline, the two versions diverge one small correction at a time — and the divergence will not be visible until a customer receives an order that mixes them.

The package itself should therefore be a controlled document rather than a folder. That means a version, a date and a named owner, and it means that the reference sample at each site is labelled with the version it corresponds to. A brand that has two sites and three versions of the file in circulation does not have a duplicated style; it has two different styles and a slow-moving argument about which one is correct.

The Three Failure Modes

Duplicated styles fail in three recognisable ways, and each points at a different part of the handover.

Pattern drift

The second version fits differently, usually by a small amount that is obvious on a body and invisible on a hanger. It happens when the second supplier drafts from the tech pack rather than working from the corrected production pattern, or when the pattern travels as a paper set that is copied with small distortions, or when the grading is redone rather than transferred. The remedy is to send the production pattern in a form the second site can use directly, and to require a fit sample at the extreme sizes rather than only at the middle.

Material difference

The garment is correctly made from the wrong cloth. It happens when the fabric is described by its general type rather than by its mill and reference, when a substitute is accepted as close enough, or when the fabric is right but a component — an interfacing, a lining, a zip, a trim — is not. Material difference is the hardest failure to argue about because the second garment is not defective; it is simply a different product. The remedy is to name every material by source and reference, and to require the second supplier to submit the actual materials for approval rather than describing them.

Technique difference

The garment is made from the right pattern and the right cloth and still does not look the same, because a construction decision was made differently. A seam finished another way, a lining attached in a different order, a hem built differently, a pressing method changed. These differences are individually defensible and collectively visible. The remedy is a construction sheet that specifies the decisive operations rather than only the measurements — which is the same material a first-run sample approval record should contain. What that record covers is set out in what a factory needs in a tech pack and a reference sample.

In practice a second source usually fails a little in all three at once, because the same handover gap produces pattern, material and technique differences together. Attacking them in that order — get the pattern right, then the materials, then the construction — is the sequence that converges fastest.

Splitting an Order Without Splitting the Quality

A brand that has decided to run a style at two sites still has to decide how the volume is divided, and the division affects how visible the difference is.

• A split by size is usually the least visible. Each site produces a portion of the size range, so no customer receives two units from different sites in the same size.
• A split by colour or fabric is similarly contained, provided the colour is consistent between the two — which is a dye lot question rather than a quality one.
• A split by destination keeps the divergence inside a market, which is useful where returns are handled regionally.
• A split by time is the most visible. Two production slots at two sites means two waves of the same style reaching customers at different moments, and the difference between them will be noticed.
• A split by channel is often the right answer. Wholesale and direct-to-consumer orders can come from different sites with less risk of a customer seeing both.

The one division to avoid is a split that mixes sites inside a single shipment without the brand knowing which unit came from where. If a customer complaint arrives, the ability to trace a unit back to a site is what allows the problem to be investigated rather than only corrected. On any duplicated style, unit-level traceability is worth requiring as a matter of policy, not as an exception.

It is also worth deciding in advance which site is the reference. Where two sites produce the same style, one of them should be the standard against which the other is judged — usually the site where the style was developed, because that is where the corrected pattern and the tacit knowledge live. Naming that explicitly prevents a drift where both sites gradually converge on their own version and neither matches the approved sample.

Running Two Suppliers Before You Need To

The most reliable version of dual sourcing is the one that is already working before it is needed, and it costs something to maintain.

• Qualify the second site with a real but small run, produced from the full file, at a size that is representative rather than convenient.
• Keep the qualification current. A site that passed a trial two years ago has different staff, different equipment condition and possibly a different factory layout.
• Refresh the reference sample periodically so the comparison piece is still representative of what the first site is currently producing.
• Maintain the file at both sites. A qualification that is not updated whenever the style changes is not a live qualification.
• Decide the switch criteria in advance. A contingency that is only invoked under pressure tends to be invoked badly.
• Accept that maintaining a qualified alternative has a cost, and that the cost is insurance rather than waste.

Brands that have been through a production interruption usually conclude that this maintenance was worth paying for, and brands that have not usually conclude that it is unnecessary. Both conclusions are reasonable given the information each has, which is why the decision is better framed as a risk tolerance question than as a procurement optimisation.

Where the decision goes the other way — a brand that chooses a single site and invests in that relationship instead — the investment looks different but is not smaller. It means a deeper level of development cooperation, more visibility into the supplier's capacity, and a stronger understanding of the supplier's own risk. Both strategies are coherent; a brand that does neither is the one that gets caught.

The Cost of Duplication

Duplicating a style is not free, and the cost is mostly in work rather than in money.

1. A second development sample round. The second site has to prove it can make the garment, which means sampling, fitting and approving — work the first site has already done and is not repeating.
2. Material approval at the second site. Even with named references, the second site has to source, submit and have the materials approved.
3. A full-size fit check. The extremes of the range have to be fitted again, because the first site's fit conclusions do not transfer automatically.
4. Inspection at a new site. A new site is an unknown site until it has demonstrated that its inspection produces the same reading as the standard.
5. Ongoing file maintenance. Two sites receiving the same changes doubles the communication and the risk of one site missing an update.
6. A smaller effective order at the first site. This is the cost buyers most often overlook — splitting volume changes the economics of the remaining volume, sometimes in ways that affect the price of everything else.

The last point is the one that makes dual sourcing a genuine trade rather than an obvious win. Volume is one of the inputs to price, and a style split across two sites may attract a higher price at both than the same volume at one. That is not a reason to avoid dual sourcing; it is a reason to model it rather than assume it is free. Which is the same discipline that applies to any production decision: understand the cost of the option, then choose it or not.

There is also an inspection consequence. A brand running the same style at two sites needs the inspection regime to be the same at both, which means the same standard, the same checkpoints and the same acceptance criteria. If the two sites are inspected differently, the quality difference customers perceive will be a difference in inspection rather than a difference in production. How that regime is structured is covered in inline and final inspection.

The decision to duplicate should therefore be made with the whole cost in view, including the cost of maintaining the second qualification. Where that total is smaller than the exposure it removes, the duplication is straightforwardly worth doing. Where it is not, the honest conclusion is that the brand is buying optionality rather than saving money — which is a legitimate reason to do it, provided that is the actual reason and not a rationalisation after the fact.

Keeping the Relationship When You Diversify

This section is written from the supplier's side, which is an unusual position for a supplier to write from, so it is worth being direct about the incentives. We would prefer to make all of a customer's volume. A customer who splits it is a customer who is managing risk, and the honest response to that is to make the split work rather than to argue against it.

• Say it rather than hiding it. A brand that duplicates a style quietly creates a supplier who eventually discovers it, and the discovery is worse than the announcement.
• Be specific about what stays. “We are prototyping elsewhere” is not the same as “we are moving a third of this style”. Suppliers plan on the second statement.
• Ask for the file rather than extracting it. A factory that is asked to help build a handover package will usually do it; a factory that discovers its pattern has been photographed will not.
• Expect a price consequence and discuss it. Reduced volume at the same site is a real change to the supplier's situation and a fair one to raise.
• Keep the relationship's other work intact. Brands that move one style and continue developing the rest with the same supplier usually keep the good terms and the tacit knowledge.

There is a version of this that goes wrong for reasons of dignity rather than economics. A brand that asks a supplier to prepare a handover gets a good handover; a brand that withholds the intention and extracts what it needs gets a technically complete package with the difficult operations left out. The difference is invisible until the second run, which is the worst time to discover it.

A Second-Source File Checklist

The following is the package we would send, and the package we would want to receive. It is not a long list, and every line on it has been the cause of a duplicated style going wrong at some point.

Most of that package already exists inside the first factory, in some form. Producing it as a coherent document is a deliberate act rather than a by-product of normal production, which is why it is worth asking for it explicitly, in advance, as a deliverable — and worth asking for it at the start of a relationship rather than in the middle of one.

Not every supplier will produce that package readily, and the response to a refusal is worth thinking through before you need it. A supplier who cannot produce a construction sheet does not have one, which is useful information about how the style is currently being made. A supplier who will not is making a commercial choice. In either case the fallback is to build the file from what you already own — the approved sample, the measurement sheet and the material list — and to have the second site measure and document the reference garment itself. That is slower and less complete, but it is a real option, and it is one good reason to keep your own copy of every approved sample and every approved measurement sheet.

FAQ

Will my manufacturer be offended if I ask to duplicate a style?

Some will, and that is information about them. A manufacturer that treats a second source as a personal matter is a manufacturer that is likely to treat other commercial questions the same way. A good supplier will ask what is driving the decision and then help make it work.

Can I demand the pattern? Who owns it?

Ownership is a contractual question and should be settled in writing before development begins rather than at the point of duplication. Where the brand owns the pattern, the handover is straightforward. Where it does not, the practical answer is usually a production pattern that can be used elsewhere rather than the drafting files. How that is structured is covered in protecting a design in a manufacturing relationship.

How closely does the second version have to match?

As closely as the customer experience requires, which is a business decision rather than a technical one. On a technical product with a strong fit expectation, the tolerance is effectively zero. On a seasonal style with a short life, a small difference may be acceptable. The mistake is not defining it and then discovering the tolerance from the returns.

Should I second-source from the very first order?

Usually not. A brand needs a settled version of the style before there is anything to duplicate, and duplicating a style that is still changing produces two moving targets. The natural moment is once the style has run at least once in bulk and the pattern has stopped moving.

How long does a handover take?

It depends on the style and on how much of the file exists already, so it is not worth quoting a period in the abstract. What determines it is the number of sample rounds the second site needs before its fit matches, which is why the package above is the thing that shortens it.

Will a supplier help me build the second-source package?

A supplier who is told honestly, and who has an interest in keeping the rest of the relationship, generally will. Ask for it as a deliverable, name what you need, and be clear about what volume stays. What suppliers resist is not duplication — it is duplication they discover.

Can I second-source just one style and keep the rest?

Yes, and that is the common pattern. It keeps the tacit knowledge and the working relationship intact for the rest of the range while giving the brand a tested alternative on the style where the risk or the volume matters most.

Working With Luxudress

Luxudress is the factory-direct front end for womenswear production across production facilities in Guangzhou and Dongguan, covering development, sampling, material sourcing, manufacturing, inspection and export packing.

We are a second source for some of the brands we work with, and a first source for others, and we will tell you honestly which arrangement suits the volume you are describing. On a handover we will produce the package above against a style developed elsewhere, run the fit checks at the extremes of the range, and tell you which operations are likely to give you trouble rather than which ones are easy.

If you are considering duplicating a style, send the tech pack and the approved sample. You can also read how to compare manufacturers before placing an order, browse the style library to see constructions we have already developed, or order from ready stock by the piece to examine our finishing on a real garment before committing to a production run.

Request a quote and we will return a costed plan for the style as specified.