From Sketch to First Sample: A Custom Apparel Manufacturer's 5-Step OEM/ODM Playbook

Manufacturer's View: Why Growing Brands Must Treat Sample Development Like Engineering

A growing streetwear and light-athletic brand had the sketch, found a manufacturer, and assumed "send a picture and the factory makes a sample." The first sample came back — sleeves 2cm short, pocket misaligned, color darker than the design — three rounds to barely acceptable, and the planned launch slipped six weeks. The problem is not the factory's craft; it is treating sample development as "make one piece" instead of "build a replicable production line."

Below is the real five-step process from the production floor: what the factory confirms at each step, what the brand must provide upfront, and why we insist on "no sealed sample, no reorder."

For growing brands, this process is worth more than it is for the big names. A major brand has a dedicated merchandiser and redundant budget — one misstep gets absorbed. A growing brand is often one founder wearing the buyer hat, with thin budget and a narrow window; one extra sampling round can miss the entire season. So the smaller the team, the more you should treat sample development like engineering — use process certainty to hedge against volume uncertainty.

Factory Floor Lesson: How a Missing Tech Pack Sank the Timeline

A recently funded DTC brand eager to launch sent only a front rendering and said "make it feel like the picture, you pick the fabric." First sample: narrow fit, lighter GSM than expected — revise. Second: pocket position off — revise. Third: color finally right, but only ten days left to the planned launch, the production pool was full, and it slipped to next month. Three sampling rounds plus one scheduling slip ate the entire fall and winter window.

The root cause: step one never happened. No tech pack, so the factory can only guess; guess wrong and it reworks, occupying the sampling slot; once the sampling slot is occupied, bulk scheduling collapses. This loss is common among growing brands — not a weak factory, but a missing document at step one.

Step 1: Lock the Tech Pack (How a Manufacturer Reads It)

A tech pack is not a design render; it is the "construction drawing" for the factory: size spec with tolerances, fabric composition and GSM, trims (zipper / rib / wash label), construction notes (stitching, topstitch, print position), and target hand-feel. Our first move on receiving it is not cutting — it is a manufacturability review: which processes our dual-factory system can do in-house, which need outsourcing, which fight the GSM. This step blocks 80% of later rework before the cutting table.

The element most often missing from a tech pack is tolerance. ±1cm versus ±0.5cm on a sleeve are different things: the former is easy to schedule but inconsistent across the batch; the latter is consistent but demands better equipment and sewing. We nail every tolerance with the brand during review, so the sample passes and the bulk does not drift.

Step 2: NDA and Lab-Dip Color Approval

Before formal sampling, both sides sign an NDA; the brand's design and pattern enter a controlled state, and the factory cannot leak or give them to a third party to copy. This matters especially for growing brands — your pattern is your moat; no NDA, no development is a hard rule in our dual-factory system.

The most underestimated part of color approval is the dye-lot difference. Same formula, different dye lots show visible deviation, especially deep blue, wine red, and forest green. So we first issue a lab-dip (small lab sample) for the brand to seal a "benchmark color," and every bulk lot is checked against that benchmark card. Dyeing without a sealed color means the whole batch comes out uneven — only re-dye or scrap.

Step 3: Proto / Size-Set / PP Samples

Samples are not one-shot; they advance in three stages:

Proto sample: first validate pattern and structure, usually grey cloth or a near fabric — fast and cheap, purpose is to lock the "shape."

Size set: using the confirmed fabric, make one of every size (XS–XXL and so on) to check the size curve's gradient is reasonable — brands often skip this and end with M fitting but XXL running short.

PP sample (pre-production): using bulk fabric, bulk trims, and bulk sewing — a "standard piece" as the reference for reorders. Only after the PP passes do we talk bulk.

Three sampling stages look costly and slow, but they split risk at minimum cost: each stage solves one class of problem, so "shape wrong" and "color wrong" never mix into one undecipherable mess where no one knows what to fix.

Step 4: Pre-Shrinkage and Wash Validation

A good-looking sample is not enough; it must "still look good after washing." Before sealing we run pre-shrinkage and wash tests: measure shrinkage (cotton knit norm 3–5%) and back-calculate it into the pattern; also check post-wash pilling, color bleed, and deformation. This directly decides how the bulk wash label reads and whether buyers complain after one wash.

For light outdoor clothing (outdoor clothing) — "active without stuffiness, holds shape after wash" — pre-shrink validation is especially critical. Our fabric lab issues a shrinkage and post-wash report the brand can request during a factory audit; it speaks louder than any verbal promise about whether a manufacturer is reliable.

Step 5: Seal the Golden Sample (Ready to Reorder)

Once the first four steps pass, seal one golden sample: signed, archived, filed. It is the "ruler" for all future bulk — every bulk piece must match the golden sample's shape, color, and construction. Once sealed, reorder is "copy against the ruler," no longer dependent on one sewing master's feel; quality stays stable and lead time becomes estimable.

With the five steps done, a sample from sketch to re-orderable typically takes 7–10 days in our in-house pattern room; once in the bulk pool, 25–35 days to lock. Run the process well and a growing brand tests a new structure with 100 pieces, reorders once data confirms — the steadiest line-expansion rhythm. Seal the golden sample and every subsequent order is "copy against the ruler," and lead time goes from "luck" to "estimable" — which is exactly why we would rather spend two weeks at sampling than lose a month at bulk.

Common Questions from Brand Owners