Every founder has the same reaction the first time bulk lands.

You lay a garment on the table, pull out the tape, and it doesn’t hit the number on your spec.

The chest is 1 cm wide. Body length is short.

You start drafting the angry email. Hold the email.

In cut-and-sew manufacturing, no two garments are ever dimensionally identical, not yours, not Uniqlo’s, not anyone’s.

Fabric moves. Cutting shifts. Wash pulls.

The question is never “Is it exactly on spec?”

It’s “Is it inside the band we agreed on?

That band is your measurement tolerance, and once you understand it, you stop panicking and start managing.

This guide is the version we wish every new brand would read before their first bulk.

It’s built to be referenced, not padded with five short chapters you can jump to.

01. What tolerance means and why variance is normal:

A measurement tolerance is the acceptable range above or below a target measurement that you’ll accept from the factory.

It’s written with a ± sign against each point of measurement.

If your chest spec is 56 cm with a ±1 cm tolerance, any garment measuring 55-57 cm across the chest passes.

Point of measure (POM):

A specific, named place a garment gets measured: chest 2.5 cm below the armhole, body length from the high-point shoulder, sleeve length, and shoulder width.

Every POM gets its own target and its own tolerance. You never apply one blanket ± to the whole garment.

Here’s the mental shift that ends the panic.

A tolerance isn’t the factory admitting it’s sloppy.

It’s an acknowledgement that a garment is soft goods cut from a living material and sewn by hand, not a machined metal part.

Setting a tolerance of ±0.1cm would be technically “tighter,” but it would blow your reject rate sky-high and add cost for a difference nobody can feel.

Set it too loose at ±2 cm, and your medium starts overlapping your large.

The craft is picking the right band per point.

The Drift Budget:

Think of your tolerance as a drift budget, a fixed amount of variance each garment is allowed to accumulate on its way from pattern to packed piece.

Every stage of production spends a little of it.

The fabric relaxes and stretches some.

The spreader adds tension and spends some time.

The wash pulls and spins some.

Tolerance is the total you can afford before the garment falls out of spec.

One more distinction that trips founders up: sampling tolerance is tighter than production tolerance.

Your approved fit sample was made slowly by your most experienced sample room, one garment at a time.

Bulk is made fast, across many operators, in hundreds of pieces an hour.

Expecting bulk to hold the sample’s exact numbers is expecting a sprint to match a stroll.

The tolerance is what bridges the two worlds honestly.

If you haven’t yet nailed down what “approved” even means for a sample, that decision comes first.

See our guides on reading and approving a fit sample and giving sample feedback that actually gets actioned.

02. Setting tolerances that protect fit:

Setting tolerances is a series of small decisions, not one global number. Four rules do most of the work.

Rule 1: Tighter on the points that decide fit

Split your POMs into two groups.

Critical fit points: chest, waist, hip, shoulder, and neck opening get tight tolerances because a shopper feels a wrong chest instantly.

Secondary points like hem sweep, pocket placement, and hood height can be breathed a little wider, because small variance there rarely affects wearability.

Blanketing one ± across both wastes your drift budget where it doesn’t matter and starves it where it does.

Rule 2: Knits get more room than wovens

Woven fabric is stable and holds a smaller tolerance.

Knit fabric stretches, twists, and shrinks more, especially in length, so it needs a more generous band.

Ask a jersey tee to hold a woven-shirt tolerance, and you’ll reject good garments for behaving like knit.

As a rough industry benchmark, casual/loose pieces often sit around ±1–2.5 cm on body measurements, and fitted pieces tighten toward ±0.6–1.3 cm, but treat those as starting points, not gospel truth.

Point of measurement Code Knit ± Woven ± Type
Chest (2.5 cm below armpit) CHST 1.0 cm 0.6 cm Critical
Body length (from HPS) CBL 1.5 cm 1.0 cm Critical
Shoulder width SHLD 0.6 cm 0.6 cm Critical
Sleeve length SL 1.0 cm 0.6 cm Critical
Bottom sweep SWEEP 1.5 cm 1.0 cm Secondary

Rule 3: Never let tolerance exceed your grade rule

This is the one that quietly wrecks size runs.

If the total tolerated variance on a point is bigger than the grade rule, the amount that POM grows between sizes, your sizes can overlap.

You end up with an XL that legally passes spec but measures smaller than a large that also passed.

Keep each critical-point tolerance comfortably below its grade rule, and the sizes stay in order.

Rule 4: Measure relaxed, always

A garment measured straight off a fitting or pulled taut reads longer than it is.

Lay it flat on a hard surface, smooth it without stretching, let it settle, then measure.

Half the “out of tolerance” panics we see are actually measuring-method errors; the piece was fine, but the tape wasn’t.

03. The Five Drift Sources: How Fabric and Process Cause It?

Variance isn’t random. It enters at five predictable stations.

Name them, and you can point at exactly which one spent too much of the Drift Budget.

  1. Fabric. The biggest source, by far. Different dye lots and mill batches behave differently; the same quality can shrink 2% in one lot and 4% in the next. Knit fabric arrives on the roll under tension from knitting, dyeing, and winding, and it will relax once that tension is released, whether you plan for it or not.

  2. Spreading tension. If fabric is pulled tight as it’s laid on the cutting table, it snaps back after the marker is cut, and the pieces come out undersized. Good spreading lays fabric at zero pull; sensors hold alignment to a fraction of a centimeter.

  3. Cutting. A cut goes through a tall stack of plies at once. The top ply and the bottom ply can differ. That’s why cutters pull pieces from the top, middle, and bottom of the bundle and check them against the hard pattern before anything moves to sewing.

  4. Sewing. Seam allowances, operator habits, and machine tension vary. A team stitching thousands of pieces introduces more spread than the one person who made your sample; this is why bulk drifts from a sample even when the fabric is identical.

  5. Wash & finishing. Garment wash, tumble dry, and any heat pull dimensions further. This stage is the one most likely to blow a length tolerance on knits, which shrink more in length than width.

How a good factory pre-spends the budget on purpose:

The professional move is to control drift before cutting, not discover it after.

That means two things.

First, relaxation: opening the rolls and letting the fabric rest so its tension dissipates before it’s spread.

Second, shrinkage testing: cutting swatches, running them through the real bulk wash, measuring the shrink in length and width, and building that percentage into the production pattern so the finished garment lands on spec after it shrinks.

In the US, the reference wash-test method is AATCC 135; ISO 6330 is its international counterpart.

For knits, cut-panel laundry (washing panels before assembly) removes most of the guesswork.

Red flag: if you tour a factory and see fabric rolls going straight from delivery to the spreading table with no relaxation area in sight, that’s a drift problem waiting to happen. It’s one of the checks worth making in person.

The deeper mechanics of choosing and testing fabric for stability sit in our fabric selection guide.

04. Catching out-of-tolerance pieces at QC:

Tolerances only protect you if someone measures against them.

Quality control turns your ± numbers into pass/fail decisions at two moments.

In-line vs. end-line:

In-line checks happen during production right after cutting (pieces from the top, middle, and bottom of the bundle against the hard pattern) and mid-sewing.

Catching drift here means fixing a setting, not scrapping a thousand finished garments.

End-line checks happen on finished, washed, pressed pieces before packing.

A factory that only checks at the end line is checking too late. Both matter.

How measurement checks fit the AQL framework:

Finished-garment inspection uses AQL (Acceptance Quality Limit), the international sampling standard defined in ISO 2859-1 (equivalent to ANSI/ASQ Z1.4).

Instead of measuring every piece, an inspector pulls a statistically-sized sample based on the lot size and an inspection level, then counts defects against agreed accept/reject numbers.

Defects fall into three classes, each with its own limit:

Class What it means Common apparel AQL
Critical Safety or compliance failure 0
Major Affects function or appearance; likely a return 2.5
Minor Small cosmetic issue 4.0

Most apparel runs use General Inspection Level II with major 2.5 / minor 4.0 as the default.

An out-of-tolerance measurement is typically logged as a defect within this system, often a major one, since a garment that measures wrong affects fit and drives returns.

Some programs also run a separate measurement check, measuring the pulled sample point-by-point against the graded spec and flagging every POM that falls outside its ±.

The number that matters: a defect rate label like “AQL 2.5” doesn’t mean “2.5% is fine.” It maps, through the sampling tables, to a specific sample size and a specific accept/reject count for your lot. Agree on those numbers with your factory before production, in writing, so nobody argues about them after bulk lands.

05. Specifying tolerances in your tech pack:

All of this is worthless if the numbers live in your head.

Tolerances have to be written into the tech pack, because the tech pack is the contract your factory measures against.

If a ± isn’t on the page, it isn’t agreed to, and you have no ground to reject it on.

What the graded measurement page needs:

  • Every POM, named and coded. Use standard codes so there’s no ambiguity: HPS (high-point shoulder), CB (center back), SL (sleeve length), and CHST (chest), and pair each with a one-line “how to measure” and, ideally, a technical sketch showing exactly where the tape sits.

  • A target for every size. The full graded run, not just your base size, with the grade rule visible between sizes.

  • A tolerance column. A ± for every POM, at every size. This is the column most first tech packs forget, and its absence is why bulk disputes turn into he-said-she-said.

  • Your own numbers, not a downloaded chart. Copying a generic measurement chart off the internet is the fastest way to get a garment that fits nobody. Tolerances are specific to your fabric, your fit, and your factory.

One scope note so you send the right document: writing the tolerance column is part of building the measurement page, which is one section of the full tech pack.

For the whole document, see our walkthroughs on how to write a tech pack from scratch.

This guide owns the tolerance decision specifically; those own the document around it.

The one thing to remember:

Bulk will never match your sample to the millimeter, and it isn’t supposed to.

Your job isn’t to eliminate variance; it’s to decide how much you can afford at each point, write it into the tech pack as ± your factory agreed to, and check bulk against that band instead of against a fantasy of perfection.

Spend your drift budget on purpose. That’s the whole discipline.

From the manual

This is one chapter of The Complete Streetwear Manufacturing Guide

These chapters cover taking a streetwear brand from idea to shipped units, costing, sourcing, tech packs, QC, freight, and launch. All written from the Sialkot factory floor.

Open The Full Guide

FAQ

Yes. A sample is made slowly by one experienced maker; bulk is made fast across many operators from different fabric lots. Some variance is unavoidable in cut-and-sew production.

What matters is whether each garment falls inside the ± tolerance you agreed for that point of measure, not whether it hits the exact spec number.

As a rough benchmark, knit body measurements often sit around ±1–1.5 cm, and fitted or critical points are tighter.

But the right number depends on your fabric, fit, and price point, so there’s no universal figure; it should be set per point of measure and written into your tech pack rather than copied from a chart.

Knit fabric stretches, twists, and shrinks more than stable woven fabric, particularly in length.

It also moves during cutting and sewing and again in the wash.

Holding a knit to a woven’s tight tolerance would reject good garments for simply behaving like a knit.

It depends on how many and on the AQL you agreed to.

The finished-garment inspection samples the lot and counts defects against accept/reject numbers set before production.

A few out-of-tolerance pieces inside the agreed limit are a pass; enough to exceed the reject number is grounds to reject the lot.

This is exactly why the AQL level and defect limits should be agreed upon in writing up front.

Control drift before cutting: relax the fabric so roll tension dissipates, run a shrinkage test through your real bulk wash, and build the shrink percentage into the production pattern.

Then check in-line pieces from the top, middle, and bottom of each cut bundle against the hard pattern, not just at the end.

On the floor · Sialkot

Written by

Faizan Ahmad

Chief Apparel Technologist & Head of Manufacturing, Gibben Clothing · Sialkot, Pakistan

Faizan leads production at Gibben Clothing, a cut-and-sew streetwear manufacturer in Sialkot, with 8+ years turning raw yarn into retail-ready hoodies, tees, bottoms, jackets, tracksuits, and headwear. He doesn’t just write about clothing; he works the floor, so every guide here is grounded in real fabric behavior, QC standards, and production data from live runs.