SummaryÂ
Tech pack mistakes are the leading cause of sample-vs-bulk mismatches in apparel production, not factory error. The most costly and most common: missing points of measure, no stated tolerance, color specified by name instead of Pantone TCX, no stitch type or SPI, an incomplete bill of materials, artwork submitted without placement measurements, missing grading rules, no care instructions, and no formal sealed sample process. Each causes a specific, predictable failure, and each has a specific, cheap fix.
“Like a Gildan But Heavier” Is Not a Spec
Here’s a sentence that shows up constantly in first-time tech packs: “like a Gildan but heavier.” It feels descriptive to the person writing it. To a factory, it’s a guess dressed up as an instruction, heavier by how much, in what fabric weight, with what hand feel? The factory fills in the blank with its own assumption, the sample comes back wrong, and the founder is furious at a factory that built exactly what it was actually given.
This is the pattern behind nearly every sample-vs-bulk mismatch: not incompetence on the factory floor, but a tech pack that never specified the thing the founder had in their head. Each of the nine mistakes below carries a real cost, usually a resampling fee, and is always lost weeks, and each has a specific, inexpensive fix. Here they are, one at a time.
Mistake 1: Missing Points of Measure
What gets submitted: A flat sketch with a couple of headline numbers, “chest 20 inches,” and little else.
What the factory builds: A garment matching the one dimension provided, with every other measurement left to the pattern maker’s own judgment.
What it costs: A full resample, plus the fee that typically comes with it. A comprehensive tech pack needs roughly 15–25 points of measure (POMs), chest, waist, hip, shoulder width, sleeve length, neck opening, front and back length, and more, each tied to a labeled reference number. Skip most of them, and you’ll approve a sample that “looks right” on a hanger, then discover in bulk that a dimension nobody specified drifted in a direction nobody caught.
The correct spec: A complete POM list, each with a reference number, tied to a labeled diagram showing exactly where on the garment that measurement is taken. This becomes the backbone everything else in the document builds on.
Mistake 2: No Tolerance Stated
What gets submitted: Exact measurements with no plus-or-minus range attached, “chest: 20 inches,” full stop.
What the factory builds: A garment that technically satisfies the spec at 21.5 inches, because without a stated tolerance, there’s no defined line marking where a measurement becomes a defect.
What it costs: Exactly the scenario in the brief behind this article, a garment a full 1.5 inches off spec, and the factory is technically correct, because nothing on the tech pack said it couldn’t be. No tolerance means no contractual basis to reject the garment at delivery, and the resample or rework cost lands on you either way.
The correct spec: Standard tolerance for casual wear typically runs ±0.5 to ±0.75 inches for chest, waist, and similar wide body measurements, tightening to ±0.25 to ±0.5 inches for critical fit points and fitted garments. Set it too tight and you’ll reject acceptable garments unnecessarily; leave it off entirely and you have no grounds to reject a bad one. State a tolerance next to every single POM, not just the headline measurements.
Mistake 3: Color Named Instead of Coded
What gets submitted: “Forest green,” “navy,” “off-white”, descriptive, confident, and completely unspecific.
What the factory builds: Their own honest interpretation of that name, since “forest green” describes a family of shades rather than one exact color.
What it costs: A bulk dye run in a shade that’s recognizably in the right family and unmistakably wrong sitting next to your marketing photos, one of the most common and most avoidable mismatches in the entire production process, and one that usually means eating the cost of a dye correction or living with the mismatch across the whole run.
The correct spec: A specific Pantone TCX code for every color in the garment, body fabric, trims, thread, labels. Pantone TCX exists specifically to remove ambiguity from color communication across a global supply chain; a name is an opinion, a TCX code is a coordinate nobody can misread.
Mistake 4: No Stitch Type or SPI Specified
What gets submitted: No mention anywhere of how seams should actually be constructed.
What the factory builds: Whatever stitch construction is standard for that garment type at that specific factory, which may not match the durability, stretch, or finish your design actually needs.
What it costs: A seam that puckers on a stretch fabric because it was closed with a rigid lockstitch instead of a stretch-appropriate construction, or a seam that fails earlier than it should because SPI (stitches per inch) was never specified and the factory defaulted to a looser setting.
The correct spec: Name the actual construction callout for each seam using ISO 4915 stitch type numbers, a 301 lockstitch for a flat, non-stretch seam; 401–403 chainstitch variants where more give is needed; 504/516 overlock for edge-finishing on knits; 605–607 coverstitch for hems on stretch fabric. Pair each with a specified SPI, since stitch density directly affects both seam strength and how much a seam can stretch without failing.
Mistake 5: An Incomplete Bill of Materials
What gets submitted: Fabric specified, and little else named.
What the factory builds: A garment where the fabric matches your spec, and everything else thread color, label placement, drawcord tips, and zipper pulls gets sourced from whatever the factory has readily on hand.
What it costs: A finished garment where the small details clash: a visibly mismatched thread color, generic drawcord ends where you expected branded ones, and a woven label sitting in the wrong spot because nobody ever specified where it should go. small individually, expensive collectively across an entire bulk run.
The correct spec: A complete bill of materials (BOM) listing every component, fabric, thread (color and type), main label, care label, hangtag, drawcord and tips, zippers, buttons, and any other trim, each with supplier, composition, color code, and placement noted. If it touches the finished garment, it belongs on the BOM. A physical trim card, carrying actual swatches of every trim item, gives the factory something concrete to match against rather than a written description alone.
Mistake 6: Artwork at the Wrong Resolution or Without Placement Measurements
What gets submitted: A low-resolution logo pulled off a website and dropped onto a mockup with no stated size or position.
What the factory builds: Artwork scaled and placed by eye, at whatever resolution survived the file transfer, frequently visibly pixelated once printed or embroidered at true garment size.
What it costs: A print or stitch-out that looked sharp on a screen and rough on fabric, sitting slightly off-center because “roughly there” was the only placement instruction anyone actually gave.
The correct spec: Vector artwork wherever possible, or a high-resolution raster file at true print size and resolution, never a screenshot pulled from a webpage. Placement needs exact measurements from a fixed reference point: “centered horizontally, 3 inches below the collar seam,” not “on the chest.” If your design involves custom embroidery, stitch count and thread colour matter just as much as placement, and both belong on the same spec sheet rather than left to a digitizer’s interpretation.
Mistake 7: No Grading Rules
What gets submitted: A spec sheet built for one size only, almost always a medium.
What the factory builds: Every other size in the run, scaled by the factory’s own default assumptions about how much a garment should grow or shrink between sizes.
What it costs: A brand’s medium that fits exactly right while the small and the XL both quietly drift, a complaint that surfaces slowly, size by size, as customers outside your own sample size start reporting fit issues you never caught, because you were almost certainly only checking the sample in your own base size.
The correct spec: Grading rules: the specific increment each POM changes by between adjacent sizes, specified across the entire size range, not just the base size. Bodies don’t scale uniformly, so different points of measure need different increments, documented explicitly rather than left to a factory’s best guess.
Mistake 8: No Care Instructions
What gets submitted: Nothing about washing, drying, or ironing anywhere on the document.
What the factory builds: A garment with no care label at all, or a generic one that doesn’t actually reflect the specific fabric and finish used to make it.
What it costs: Customer complaints and returns from garments damaged by care instructions that were never actually verified against the real fabric composition, a cheap, entirely avoidable problem that still quietly erodes trust in the product.
The correct spec: Explicit wash, dry, iron, and bleach instructions on the tech pack, checked against your actual fabric composition and finish, not copied from a similar-looking product without verifying it applies.
Mistake 9: No Sealed Sample Process
What gets submitted: Verbal or emailed approval of a sample photo, with nothing physically preserved anywhere as a reference.
What the factory builds: Bulk production based on whatever the factory believes was approved, which can drift subtly from what you actually saw and signed off on weeks earlier.
What it costs: A dispute with no fixed reference point. Without a physically preserved, tagged, and dated sealed sample, “this doesn’t match what I approved” is an argument that’s genuinely difficult to win, because there’s no physical object either side can point to and compare against.
The correct spec: A formal sealed sample step built into your process, a physical garment, signed off by both sides, preserved untouched as the reference for every unit in the bulk run and for any future reorder.
What These Nine Mistakes Actually Cost, Added Up
Individually, each of these looks like a small oversight. Stacked together across a real order, they compound into the two costs founders report most often: a re-sampling fee, typically the cost of producing and shipping a corrected sample, sometimes several hundred dollars depending on complexity, and lost weeks, since every failed sample round restarts the approval clock rather than simply continuing it. A tech pack with even three or four of these nine gaps can realistically add a full sample cycle to a project, which on a tight launch timeline is often more costly than the resampling fee itself.
Catch All Nine Before You Submit
Every one of these mistakes shares the same root cause: a gap between what a founder actually means and what a factory can act on without guessing. “Like a Gildan but heavier” is a feeling. A GSM figure, a Pantone TCX code, and a complete POM list are instructions, and only instructions reliably survive the trip from your head to a cutting table on the other side of the world.
Run your own document against this list before you send it: complete POMs with tolerances, coded colors, named stitch types with SPI, a full BOM including every small trim, print-ready artwork with placement measurements, complete grading rules across the whole size range, verified care instructions, and a sealed sample step built into your approval process. Wherever any of the nine is missing, that’s precisely where your sample is likely to go wrong first.
If building this correctly on a first attempt feels like a lot to get right alone, that’s exactly the gap our tech pack service exists to close, turning a rough concept into a document a factory can execute without filling in blanks on your behalf. It’s worth having in hand before any cut-and-sew manufacturer’s project begins, since these nine points are the actual input a factory depends on to get a sample right the first time, rather than the second or third.
Don’t find out which of these nine mistakes is sitting in your tech pack after the sample fails. Get it checked first.
Frequently Asked Questions
What is a tech pack in clothing manufacturing?
A tech pack is the complete technical document a brand sends a manufacturer to produce a garment, covering measurements, materials, construction, colors, artwork placement, and size grading. It’s the primary reference a factory builds from and the main source of sample-vs-bulk mismatches when it’s incomplete.
Do I need a tech pack to get clothes manufactured?
Technically a factory can attempt production from a rough sketch or reference image, but without a proper tech pack, every gap you leave open gets filled by the factory’s own assumption, which is exactly how a vague concept turns into a sample that doesn’t match what you actually pictured. A complete tech pack is the professional standard for communicating a design accurately.
What must be included in a tech pack?
At minimum: a flat sketch, complete points of measure with tolerances across every size, a full bill of materials, named stitch types and SPI, Pantone TCX color codes, print-ready artwork with placement measurements, grading rules, and verified care instructions.
What is a point of measure?
A point of measure (POM) is a specific, labeled location on a garment where a measurement is taken, such as chest width, sleeve length, neck opening, and similar dimensions. A comprehensive tech pack typically specifies 15–25 POMs per style, each tied to a reference number and a diagram showing exactly where the measurement applies.
What tolerance is acceptable in garment production?
It depends on the specific measurement and garment type. Standard tolerance for casual wear typically runs ±0.5 to ±0.75 inches for wide body measurements like chest and waist, tightening to ±0.25 to ±0.5 inches for critical fit points and fitted garments. Every point of measure needs a stated tolerance; without one, a factory has no agreed standard it can be held to.
