How Much Yarn Does a Sweater Actually Need? A Worked GSM and Consumption Calculation
A knitwear quote usually arrives as a single number per piece. Behind it sits a yarn consumption estimate that almost nobody shows the buyer, and on a first order that estimate is where the margin quietly disappears. Get it 8% low on a 50-piece run and you are short about 1.8 kg of yarn, roughly $50 at $28/kg. The yarn is the cheap part. The second dye lot, the shade mismatch against the first batch, and the week added to the delivery date are not.
This is a working calculation, not a pricing formula. Replace every input below with a measured swatch from your own supplier; the point is to check whether a quote is physically plausible before you approve a sample.
Why area times GSM does not work here
For a cut-and-sew woven garment, consumption is close to marker area times fabric weight: you buy meters of cloth and cut shapes out of it.
A fully-fashioned knit panel has no bolt of fabric behind it. Yarn goes straight into interlocking loops, and the amount used is set by loop length, not by panel area. Two same-size panels at the same gauge can differ in weight by 30% if one is knitted with a longer loop. Shorter loops make a denser, heavier, less extensible cloth; longer ones make a looser cloth that can sag. Loop length is a machine setting, so consumption is a production decision before it is a costing one.
So GSM is an output of the knitting parameters, not a number you can safely specify alone. Asking for "240 gsm at 12GG in 2/28 Nm" may give the factory three mutually unachievable constraints. Ask instead for the loop length and stitch density that produce 240 gsm.
The four inputs
| Input | What it means | Typical range | Where it comes from |
|---|---|---|---|
| Gauge (GG) | Needles per inch on the bed | 3GG–18GG | Machine assignment |
| Yarn count (Nm) | Meters per gram, including ply and ends | 2/28 Nm, 2/48 Nm | Cone label or invoice |
| Stitch density | Wales and courses per cm on a washed swatch | 4–8 per cm each | Measured swatch |
| Loop length | Yarn consumed per loop | 0.45–1.2 cm | Stitch cam setting |
| Panel area | Knitted area of all panels, including ribs | 1.2–1.6 m² at size M | Graded marker |
One notation trap matters. 2/28 Nm means two plies of 28 Nm, so the resultant count is 28 / 2 = 14 Nm, or 14 meters per gram. Reading it as 28 m/g halves the answer. Two ends, common at 12GG when one is too fine, doubles the length again. Write the full notation: "2/28 Nm, 2 ends", never "28s wool".
A worked example: 12GG, 2/28 Nm, two ends
Take a size-M pullover in single jersey, knitted flat at 12GG.
wales per cm = 12 needles/inch / 2.54 = 4.72
courses per cm = measured on washed swatch = 6.50
loops per cm2 = 4.72 x 6.50 = 30.7
loop length = stitch cam setting = 0.62 cm
ends = 2
yarn per cm2 = 30.7 x 0.62 cm x 2 = 38.1 cm = 0.381 m
resultant count = 2/28 Nm = 14 m/g
weight per cm2 = 0.381 m / 14 m/g = 0.0272 g
GSM = 0.0272 x 10,000 = 272 g/m2
In one line:
GSM = (wales/cm x courses/cm x loop length cm x ends / resultant Nm) x 10,000
That 272 g/m² is a plausible mid-weight 12GG jersey. A size-M pullover with front, back, two sleeves and ribs has a knitted area near 1.35 m², so the fabric in one garment weighs about 367 g. Ribs, collars and plackets are denser and add 10–15%; use 12%. Knitting waste, linking losses and cone changes add 3–5%; use 4%.
| Step | Calculation | Result |
|---|---|---|
| Body fabric | 272 g/m² × 1.35 m² | 367 g |
| Rib / trim allowance | × 1.12 | 411 g |
| Process waste | × 1.04 | 427 g |
| Per garment | — | ≈ 0.43 kg |
| 50-piece MOQ | × 50 | 21.4 kg |
| 3% spare yarn | × 1.03 | 22.0 kg |
| Yarn cost at $28/kg | 0.43 × $28 | ≈ $12.04 per piece |
The 3% spare covers the repair loop, tension adjustments and finishing damage. Order 22 kg, not 21.4 kg: yarn is dyed by lot, and a lot split across two dyeings is a shade claim waiting to happen.
Three ways this goes wrong
Count notation. It accounts for more bad estimates than anything else. A supplier quoting "28 Nm merino" may mean 2/28, 1/28 or 2/48 plied down. Ask for meters per gram from the cone label rather than the shorthand.
Net area versus knitted area. A graded marker gives the finished silhouette. The knitter produces panels with selvedge waste, extra courses at transitions and separate rib blanks linked on later. If your area comes from finished measurements, add 8–12% or the answer will be optimistic by the same amount.
Structure changes. Jacquard floats, intarsia joins and tuck stitches all change the yarn in a loop. A style that is 80% single jersey and 20% double jacquard cannot be costed from one swatch: break it into zones, measure each, and weight the areas. A panel described as "jacquard" can vary 20–30% in consumption depending on float length.
Washing moves the number
Fabric weight is measured on cloth whose dimensions are still settling. After washing and flat drying, length relaxes 2–4% and width 2–5% while yarn mass stays the same. A 5% shrink in each direction removes roughly 10% of the area and raises GSM about 11%, so a 272 g/m² off-machine swatch can finish at 300 g/m² without a gram of extra yarn.
Specify the state, not just the number: "GSM after one wash and flat dry, per the approved care label." If the factory measures off-machine and you measure after washing, a 10% discrepancy looks like a dispute and is only a definition. The same applies to stitch density, since wales and courses taken before relaxation understate the finished density.
What to ask for before the PO
A consumption figure is only as good as the inputs behind it. Ask for these in writing, on the sample approval sheet rather than in an email thread:
- Gauge and machine type, and whether the quoted gauge is needle count or finished wale density.
- Full yarn notation including ply and number of ends, backed by the cone label or a test report.
- Stitch density, wales and courses per cm, measured after the agreed wash cycle.
- Loop length in millimeters, or the cam setting that produces it.
- Structure by zone: body, ribs, plackets, any jacquard or intarsia areas.
- Washed swatch GSM, with the wash and dry method named.
- Graded panel area for every size in the run, plus the rib and trim allowance.
- The waste assumptions: knitting waste, linking loss, and the spare-yarn percentage built into the quote.
If a supplier can give you items 3, 4 and 6, the rest is arithmetic you can verify in a spreadsheet. If they cannot, no amount of negotiation on unit price will tell you whether the quote is real. A broader buyer-side procurement checklist from MySweaterFactory covers the commercial terms around these numbers, but the calculation above is the part worth doing yourself first.
Where it lands in the cost structure
Yarn is usually 35–55% of the FOB cost of a plain knitted sweater, and a larger share on fine-gauge programs where labor per garment is low. That is why a 10% error in consumption can move gross margin by three to five points even though the absolute yarn cost looks modest. On a 50-piece first run the effect is amplified: knitting setup, program writing, dye-lot minimums and lab dips are spread across very few garments, with no volume to absorb a correction.
The schedule hurts more than the cost. A shortfall discovered after knitting starts means re-dyeing, and a dye lot plus fresh lab-dip approval is commonly 7–12 days — a week or two on a lead time already committed to a retail calendar. Estimating before sample approval costs one spreadsheet and is the cheapest schedule insurance available.
None of this replaces a weighed sample. Weigh the panels off the machine, wash them, weigh them again, and compare with the estimate. A 5–8% variance is normal, and the measured number becomes the standard for the production order. The point of running the arithmetic yourself is not to catch a factory out. It is to know which question to ask when a quote moves 15% and the explanation is "yarn price".
Author bio: A knitwear sourcing consultant writing about spec sheets, yarn engineering and production economics for independent apparel brands. Further reading: MySweaterFactory.
Top comments (0)