Woven Fabric Tolerances: What GSM and Width Variation Is Normal

A fabric tolerance is the agreed band within which bulk production may vary from the specified value — a 200 GSM target with a tolerance of plus or minus 5 means anything from 195 to 205 GSM conforms. Tolerances exist because woven fabric is not machined metal: yarn lots vary slightly in count, loom tension drifts over thousands of metres, and wet processing shrinks cloth by amounts that can be predicted but never fixed to a decimal. The real quality question is therefore not whether bulk will vary from spec — it will — but whether that variation is bounded by a band both sides put in writing before production began.

Why bulk never lands exactly on spec

The variation starts at the fibre. Cotton is a natural material, and yarn spun in different lots differs marginally in count and moisture content; a slightly finer warp lot lightens every metre woven from it. On the loom, warp tension and pick density drift within normal maintenance limits across a long run. Then finishing changes the cloth physically: desizing, scouring, dyeing and stentering all shrink or set the fabric, and a stenter operator holds width to a target, not to a millimetre.

Stack those stages together and a bulk delivery landing within a few percent of spec is a well-run job. A supplier who says their fabric always lands exactly on spec is describing their paperwork, not their production.

What a written tolerance looks like

There is no single industry-wide number — which is why buyers searching for an "acceptable" GSM variation find contradictory answers. What we can offer is our own practice, which we consider a defensible norm for woven fabrics: on our orders the written tolerances are GSM plus or minus 5 and width plus or minus 3 inches, stated on the order confirmation itself.

Just as important is how conformance is judged — a number without an inspection method is half a tolerance. Our orders name the method: bulk is inspected against an approved retained sample, greige orders follow a header matching procedure, and four-point grading applies when agreed on the order. Whatever numbers you settle with any supplier, insist on that same shape: band, method, reference sample, all in writing.

Property by property

Property	What actually varies	A sensible written approach
GSM	Yarn lot, pick density, finishing shrinkage	A band around target (ours is ±5), measured on conditioned fabric by an agreed method
Width	Stenter setting, relaxation after finishing	A band on width (ours is ±3 inches) — specify whether the band is on cuttable or overall width
Shade	Dye lot chemistry, batch to batch	Continuity judged against the approved lab dip and retained sample
Delivered quantity	Piece lengths never divide evenly into an order	An agreed over/under shipment percentage per order
Weaving quality	Normal loom faults per piece	A named grading system (e.g. four-point) with a points threshold

What your order confirmation should pin down

• Target GSM and its band, with the test method and fabric state — GSM moves through processing, so say whether the figure is greige or finished
• Cuttable width and its band — stated as cuttable, since selvedges do not make garments
• Full construction — yarn counts and ends × picks, not just weight
• Shade reference and how batch-to-batch continuity will be judged
• Inspection method — which sample is the reference, who holds it, and which grading applies
• Over/under shipment allowance

Getting these onto one page is most of what a good fabric brief exists to do.

Red flags

Two outrank all others: a supplier who declines to put tolerances in writing — the variation still happens, you just lose any recourse — and one who promises bulk will "match the sample exactly", a physically impossible claim that usually precedes a dispute. A quote silent on whether width is cuttable or overall belongs on the same list.

The trade-off

Tighter bands than ours are achievable, but they are not free: a very narrow GSM band shrinks the pool of units that can hold it and lengthens development, and that risk is carried somewhere in the job. Looser bands shift the risk to you — at ±10 GSM a 150 GSM shirting can drift visibly between deliveries, and a garment factory feels a two-inch width loss in marker efficiency before anyone weighs the cloth. The honest position is to agree the tightest band your end use genuinely needs, not the tightest that sounds rigorous — then confirm it survives the jump from sampling to bulk.

Tolerances are easiest to agree before an order is signed, which is where a sourcing partner earns its keep — we put the bands, the reference sample and the inspection method on the order itself. Browse woven qualities with full construction data in our Fabric Library, request free swatches or start a development.