Tech pack vs reference sample: what does a factory need?

This is the comparison. What a pack contains is here, whether you can work without one is here, and what a photograph carries is here.

Quick answer: what each one carries

The complementarity is almost exact: what one carries is largely what the other cannot. That is the argument for sending both rather than choosing.

What only a garment can tell you

A reference sample is a set of measurements, a construction and a fabric that have already been resolved into a physical object. Three things come from it that no document conveys.

• Ease. How much room the garment has relative to the body is a felt quantity. It is written down as a difference between two numbers, and it is understood instantly by wearing the thing.
• Drape in the real cloth. A photograph shows how a fabric looked under studio light. A garment shows what the cloth does after an hour of wear — where it creases, how it recovers, how the skirt moves.
• Seam behaviour. Where a seam sits comfortably, which ones press flat, which ones pull. This is the information that most reduces sample rounds, and it exists only in a made garment.

A sample is also measured directly, which removes an entire category of error. Nobody has to interpret what “bodice length” means when they can put a tape on the garment.

What only a document can tell you

• Tolerances. A garment is one correct instance. It cannot say how much variation is acceptable, and without that, inspection has nothing to check against.
• Invisible construction. A sample can be taken apart, and doing so destroys it. A pack states what the inside is across the whole run.
• The size range. A sample is one size. Only a chart with per-size values can be graded from without inventing the increments.
• Trim specification. A sample has one zip. Production needs four hundred of a specified length, finish and pull type.
• Packing and labelling. Not carried by a garment at all, and decisive for whether the goods can be sold where you intend.
• Permanence. A sample gets worn, lent and lost. A dated document survives to the second order.

The best brief uses both

In practice, the strongest starting point is a reference garment plus a short document that records the numbers and the intent. The garment carries what is hard to write down; the document carries what is hard to see and impossible to inspect against.

1. Send the garment, with a note saying what fits well about it and what you want changed.
2. Add the measurements you can take from it, at least at the key points.
3. State the fabric you intend to use, since a different cloth changes the fit behaviour.
4. State the construction standard — lining, closure, finish — since these are invisible in a garment too.
5. State the size range and the per-size intent, because the sample is one size.
6. State packing, labelling and destination, before freight rather than at it.

Manufacturing perspective: what we do with each

A reference garment and a specification are read together, and each corrects the other. The garment is measured and the numbers compared with your chart; where they differ, the difference is usually a definition — you measured the neckline flat and we measured it on the curve. Finding that out on day one costs a message.

• The garment is used to establish ease and proportion. That is the part no document can supply.
• The chart is used to establish the numbers and to define what the sample will be approved against.
• The construction section is used to build the inside, which the garment cannot show without being destroyed.
• The trim and packing sections are used to book materials, which is where the calendar actually gets decided.

When only one arrives, the work does not stop — it moves to assumptions. A garment with no chart means the numbers are inferred from one size, which is defensible for a single-size order and unreliable across a range. A chart with no garment means the ease and drape are inferred from description, which is where most sample rounds come from.

Practical example: two briefs, two outcomes

Brief B did not send more. It wrote down what Brief A left for the factory to infer, and the difference shows up in the second order rather than the first.

Common mistakes

• Sending a sample instead of numbers. A garment is one correct instance, not a tolerance range.
• Changing the fabric but keeping the sample's measurements. The cloth is part of the fit, so the fit work has to be repeated.
• Not saying what you like about the sample. Without that, the factory protects the wrong elements.
• Assuming a sample shows the construction. It shows the outside; the inside is a decision somebody still has to make.
• Sending a sample without a size range. One size cannot be graded from without inventing the increments.
• Not getting the sample back. Unreturned references are the most common reason a reorder starts from scratch.

What to send us

Luxudress has manufactured womenswear in Guangzhou since 1995, producing over 100 new styles a month. Either asset is a workable start; both is a shorter development.

1. Send the garment if you have one, with a note on what to keep and what to change.
2. Send the numbers you have. We will measure the garment and reconcile the two, and tell you where a definition differs.
3. Tell us the fabric, because it changes the fit work on a reference sample.
4. Tell us the size range so the chart can be built across it rather than inferred.
5. We will return the chart we worked to, so it can be approved rather than assumed.
6. We retain the approved standard, which is what makes your reorder a repeat.

If you are putting a brief together, send what you have and we will tell you what is missing. Product development covers the stage either side of this, and writing a tech pack covers the document if you would rather start there.

Frequently asked questions

Which does a factory prefer, a pack or a sample?

A sample is more useful for the first pattern and a pack is essential for everything after it. If only one can be sent, a sample gets a better first garment and a pack gets a more repeatable one. Most first orders would benefit more from the sample; most reorders need the pack.

Can you work from just a sample?

Yes, and it is a common starting point. What we will do is measure it and return the numbers, because a garment is one size and a run needs a range. The measurements become the chart, and the chart becomes what the sample is approved against.

Will you take the sample apart?

Not without asking, and generally not on a reference the brand expects back. We can usually infer enough construction from the outside plus a description of the interior to reverse it responsibly. Where a garment is genuinely ambiguous, we will say so and ask.

Can I send a sample but use a different fabric?

Yes, and the fit work has to be redone because cloth is part of the fit. The sample still tells us proportion, construction and assembly; the measurements have to be re-established against the new material, because a heavier or more fluid cloth behaves differently on the same pattern.

Should I buy a competitor's garment as a reference?

It is common practice and it comes with a boundary. A reference gives us proportion and construction logic to work from; we do not reproduce another brand's product as a replica. If the intent is to develop something in the same family rather than a copy, say so at the start and it is straightforward.

Do I get the sample back?

Yes, if you ask for it and arrange the return. Keeping a reference garment with your own records is worth doing, because it is what lets a style move between suppliers later without being rebuilt from a description.

Conclusion

A reference sample carries what is hard to write down — ease, drape and seam behaviour. A tech pack carries what is hard to see and impossible to inspect against — tolerances, the size range, the invisible construction and the packing requirement.

Send both if you can, and say what you like about the sample and what you intend to change. Where only one is available, know which questions it leaves open, and expect those to be answered by assumption rather than by you.