OEM T-Shirt Manufacturing in China: From Your Tech Pack to an Approved PP Sample

An OEM T-shirt programme is a chain of approvals, and every link is a document. The tech pack defines what will be built, the measurement tolerance defines what counts as correct, the sample stages confirm that the factory can reproduce it, and the pre-production sample sets the standard the bulk order is measured against. Where any link is missing, the gap is filled by someone else’s judgement.

What does OEM require that ODM does not?

Your own pattern, specification and approvals.

In an OEM build the brand owns the design intent and the factory converts it into a pattern, a sample and bulk garments, which means the brand has to supply and approve documents at every stage.

The practical difference is the amount of documentation. Because the brand owns the design, it also owns the accuracy of the measurement chart, the tolerances and the approval decisions. That is more work at the front and considerably more control afterwards, which is why brands usually move to OEM once a style has proven demand rather than starting there.

Which tech pack elements does a factory actually read?

Measurements, construction, fabric, trims and artwork.

The pattern maker reads the measurement chart and the construction notes, while costing reads the fabric, trims and decoration, so a pack that omits any of the five produces an incomplete quotation.

In practice the measurement chart and the fabrication notes are read most closely, because they determine the pattern and the cost. Construction notes cover seam types, hem treatment, neck rib dimensions and shoulder taping, and they are where a factory decides how the garment will be made. Artwork belongs in the pack with placement and size stated, not attached separately in an email, because decoration affects both the cost and the schedule. The test of a usable tech pack is whether a factory can build a sample from it without asking a question that the document should have answered. It is also worth naming the pack with a version number and a date as soon as it is issued, because on a programme that runs for several seasons the difference between version two and version three becomes the difference between replacing a shirt and repairing a relationship.

Why does tolerance prevent disputes?

Because it defines what counts as correct.

A measurement chart without tolerances invites interpretation, and interpretation changes with the person measuring, the fabric and the day.

Tolerance also has a commercial function: it sets the boundary between a defect and normal production variation, which is what an inspection is measured against. A tolerance that is too wide turns the measurement into a suggestion and invites drift between lots; one that is too tight makes the bulk unachievable and produces rejections for variation no customer would notice. Ask the factory which tolerance it can hold on each point of the chart, then set the chart to that answer rather than to an aspiration. Where a measurement is critical to fit — the chest or the body length, for example — a tighter tolerance is justified, while less consequential points can carry a wider one.

Which substitutions should you watch for?

Fabric, neck rib and decoration parameters.

The fabric lot, the neck rib and the print settings are the components most often changed quietly, because all three can be altered without changing how the garment looks when inspected.

A substitute fabric may pass a visual check and behave differently after washing; a rib from a different supplier may look identical and lose recovery; a print applied at a different temperature may appear correct and crack within weeks. The counter-measure is a change-control clause that names those components, requires written approval before any substitution, and states the remedy if an unapproved change is found. Ask every candidate factory what it does when a specified fabric is unavailable; the answer separates partners who stop the line from those who solve the problem and explain it afterwards.

What does each sample prove?

Four gates, each answering a different question.

The development sample tests the concept, the fit sample tests the block, the size set tests the grading and the pre-production sample tests whether the line can reproduce the approved garment.

Each gate also produces a document: the measurement chart from the fit stage, the recorded values from the size set, and the signed approval from the pre-production stage.

Skipping a gate moves its risk into bulk, where it is more expensive to correct. On T-shirts the size set is the gate most often omitted and the one that catches grading errors across the whole range, so it should be produced and measured rather than extrapolated from one size. The pre-production sample is built with production fabric, production trims and production methods, which is why approving it means approving what the line will actually do. Where a factory publishes its development sequence, the stages are usually visible: design layout, custom sample, fabric sourcing, pre-production approval, cutting and mass production, then quality control, packing and shipping. Finetex Global describes the same sequence on its apparel design page, which makes it a useful comparison when a supplier describes its own process in vague terms.

How do minimums and colour count interact?

Colour usually drives the threshold, not quantity.

The minimum is set by the fabric lot, the dye lot, the trims and the line changeover, and each colour adds a lot of its own.

That makes the colour count the most effective lever on a first OEM order. A single-colour programme with a deep size run will always sit lower than a three-colour programme with the same total quantity, because the second and third colours add lots without adding efficiency. Ask the factory which component sets the minimum for your specification and what would change if you consolidated to one colour, then decide whether the range is genuinely better with additional shades on the first order or whether they can wait. Where a brand needs the visual variety, producing one colour first and adding shades on the reorder usually produces both a lower threshold and clearer sell-through data.

Who owns the pattern and the artwork?

Whoever paid, unless the contract says otherwise.

Where the brand funds development, the pattern, the graded nest and the artwork should belong to the brand, with the factory licensed to produce only the orders placed.

This is the clause that determines whether a programme can move later without being rebuilt. Ask explicitly for ownership or a transferable licence, and for the graded nest to be handed over in a usable format, along with the colour separation files for any print. Where the factory retains ownership and licenses the design to you, the licence should state the field of use, the territory and the term, so a reorder three seasons later does not fall outside the agreement. Requesting the file list before sampling is cheaper than discovering the gap afterwards, and a factory that treats the specification as the buyer’s property will hand it over without prompting.

Which quality gates belong in an OEM run?

Incoming fabric, in-line checks and a final inspection.

Five checks cover the points where an OEM programme most often loses consistency, and each corresponds to a defect that appears after the goods have been sold.

• Incoming fabric inspection for weight, shade and hand-feel against the approved standard.
• In-line checks on neck rib tension and shoulder seam finishing.
• Measurement verification across the size set during the run.
• Decoration adhesion checks where the shirt carries a mark.
• Final inspection at a stated acceptance level, with defect classes defined.

Agree the plan in writing before production and apply it to every order.

Ask what happens if the final inspection fails, because the answer should name rework, replacement, deduction or rejection rather than opening a discussion. Where the order value justifies the cost, a third-party inspection before shipment gives an independent record that is useful if a batch underperforms in market.

How should two OEM quotes be compared?

Equalise scope, then compare the making.

Two quotes are comparable only when the fabric, the trim set, the decoration, the packing and the quantity assumptions are identical, and after that the difference is labour minutes, efficiency and risk carried.

Ask each factory to itemise the quotation by the same categories and to state what is excluded, because an excluded sampling charge or an assumed packaging format can make one quote look cheaper without being so. Where the two remain close after equalising scope, the decision usually comes down to capability and communication: which factory can describe its controls, produce a measured size set and answer questions without prompting. Cost discipline in this category is less about finding the lowest quote and more about ensuring the quotes describe the same product.

Sources: Supima — extra-long staple cotton; 16 CFR Part 303 textile labelling; OEKO-TEX STANDARD 100; ISO 9001 quality management; Textile Exchange standards.