The run is stitched. Six hundred shirts sit folded on a table in Tirupur, and the only thing between them and your warehouse is a woven size label nobody ordered, still three weeks away. That is what bad raw material planning for apparel looks like in practice — not a dramatic fabric shortage, but a ₹4 trim holding ₹6 lakh of finished goods hostage. Fabric, waste allowances and trims each behave differently, and each needs its own number.
Key takeaways
- Fabric consumption per garment comes from a marker on your actual fabric width — not from a guess or a competitor’s tech pack.
- Cost and buy from a size-weighted average across the order’s size curve, never from the medium.
- Cutting waste, shrinkage and QC rejects are three separate losses. Give each its own percentage and divide by the combined yield.
- Trims are cheap, slow and high-MOQ — they stop more runs than fabric ever does.
- Buy the whole run’s fabric in one dye lot. A mid-run top-up will not colour-match.
Raw material planning for apparel starts with consumption
Every fabric requirement calculation rests on one input: how many metres one garment actually eats. Most founders inherit this number from the unit as a single figure — “1.6 metres, sir” — and never ask where it came from. It came from a marker, and everything downstream inherits the marker’s errors.
What a marker actually tells you
A marker is the cutting map: every pattern piece for a garment, arranged on a rectangle the exact width of your fabric and nested as tightly as the grain lines allow. The cutting master spreads fabric in plies — a lay — lays the marker on top, and cuts through all of them at once. Net consumption is marker length divided by the garments it contains.
Two things follow. Consumption is a property of the fabric width, not of the garment — a figure quoted without a width is meaningless, because the same shirt eats different metres on 58-inch and 44-inch goods. And marker efficiency improves when several sizes share one lay, since small pieces from one size fill gaps left by another. Ask for both numbers: length, and the efficiency percentage the CAD software reported.
Why you cost from a size-weighted average, not the medium
Larger sizes consume more fabric. Obvious — and then ignored at costing time, because the medium is the sample size and the sample is what got costed. If your curve leans towards L and XL, as most menswear curves do, the medium figure quietly under-buys every run.
Here is a 600-piece run of an oversized cotton poplin shirt on 58-inch fabric, with per-size net consumption taken from actual markers:
| Size | Units in run | Net consumption per piece | Fabric needed |
|---|---|---|---|
| XS | 60 | 1.42 m | 85.2 m |
| S | 120 | 1.50 m | 180.0 m |
| M | 180 | 1.58 m | 284.4 m |
| L | 150 | 1.68 m | 252.0 m |
| XL | 90 | 1.78 m | 160.2 m |
| Total | 600 | 1.603 m (weighted avg) | 961.8 m |
The size-weighted average is 1.603 m. The medium is 1.58 m. A gap of 23 millimetres per garment sounds like nothing — across 600 pieces it is 13.8 metres, roughly nine shirts’ worth of fabric you did not buy and did not cost. Store consumption per size in the product’s bill of materials and let the weighted average fall out of the size curve you actually ordered.
The waste allowance stack
Net consumption is what the garment contains. It is not what you buy. Between the roll arriving and the packed shirt leaving, fabric disappears in three distinct ways — and each deserves its own percentage, because each has a different cause and a different fix.
1. Cutting and marker waste
End-of-lay remnants, the strip lost to the selvedge, the last metre of a roll too short to lay, splices at roll joins, and the negative space the marker could not nest away. Typically 3–5% on simple wovens and knits, and higher on stripes, checks and directional prints, where every piece must be matched and cannot be flipped — a checked shirt can push past 10% on its own. This one is a property of your pattern and your cutting master’s discipline.
2. Shrinkage
Natural fibres relax when washed. If the fabric is pre-washed before cutting — which is what you want on cotton, so the finished garment does not shrink on the customer — you buy a metre and cut something less than a metre. Cotton commonly loses 3–8% in length; heavier knits and loomstate fabrics can lose more. Ask the mill for the shrinkage test on your specific quality, and if they cannot produce one, wash a two-metre swatch yourself before committing. Shrinkage is measured, not assumed.
3. QC rejects
Stitching defects, oil marks, shade variation between panels, and fabric faults that only appear after cutting. Budget 2–3% for a unit you know, more for a first run with a new partner. Rejects differ from the other two: they consume finished garments, so they burn trims and stitching cost as well. That makes them the most expensive percentage in the stack, and the one worth attacking through inline checks at each production stage.
Now the reason a single lazy 10% fails. A flat allowance hides which loss is which, so nothing ever gets fixed — you cannot negotiate a cutting-waste problem with a mill or a shrinkage problem with a stitching unit. And the arithmetic itself is usually wrong.
The formula for a production run
Waste percentages are losses on output, so you divide by yield rather than multiplying by a markup. Losing 4% of your input means 100 metres of usable output needs 104.2 metres bought, not 104.
Applied to the 600-shirt run, with 4% cutting waste, 5% residual shrinkage and a 2.5% reject rate:
Two comparisons make the point. Multiplying instead of dividing — 961.8 × 1.115 — gives 1,072 m, about 10 metres short. The lazy flat 10% gives 1,058 m, roughly 24 metres short, or 15 garments you cannot cut. Neither shortfall announces itself until the cutting table runs out on the last size.
Rounding to the roll
You cannot buy 1,081.6 metres. Fabric ships in thaans, and a thaan is whatever length the mill wound — commonly 30–50 m for shirting, but variable enough that mills bill on actual metres received rather than nominal roll length. With an average thaan of 40 m:
Always round up, never down. The 38 spare metres cost roughly ₹7,000 at ₹185/m and buy you a re-cut when a panel goes wrong; the alternative is a top-up in a different dye lot, which is not a fix at all. At ₹185/m the 1,120 m lands at about ₹2,07,000, or ₹345 of fabric per shirt — the number that should feed your costing sheet.
Then write the received metres back against the PO. Thaans are rarely the length the invoice claims, and the gap between ordered, received and consumed metres is the cheapest error check in the whole process. Running material POs through a structured buy-materials and receive-goods stage makes that reconciliation automatic instead of a WhatsApp memory test.
Trims: cheap, slow, and capable of stopping everything
Fabric gets planned because it is expensive. Trims get forgotten because they are not. Buttons, main and care labels, size labels, hangtags, zips, thread, interlining and poly bags might be 5–8% of garment cost — and any one missing means zero shippable units. That asymmetry is the whole problem: low value, long lead time, high minimum order quantity.
Trims planning is arithmetic too, just more of it. The same shirt needs 7 front buttons, 2 cuff buttons and 1 spare stitched inside — 10 per garment, plus breakage:
Do that for every line on the trim card. Three traps catch people repeatedly:
- Size labels follow the size curve. Main and care labels are one design across the run; size labels are not. Order 600 identical ones and you own 600 mediums and a stalled run. Break size-coded trims down the same curve you broke the fabric down by.
- Thread is bought by cone, not by metre. Consumption depends on stitch type, density and seam length, so a shirt can run roughly 150–250 m across all operations. Most units hold thread and bill you for it — but only in shades they already carry. A custom shade means a dyeing lead time you did not plan for.
- Packaging is a trim. Printed poly bags, branded tape, mailer boxes and thank-you cards sit outside the garment BOM in most brands’ heads and inside the critical path in reality. Custom-printed packaging is frequently the longest lead time on the card.
Lead times and MOQ behaviour by category
The table below is a planning framework, not a price list. Lead times and minimums vary by supplier, season, cluster and how much business you already do — Tirupur’s knit ecosystem turns local trims around faster than a Noida woven-export unit sourcing from Delhi. Treat these as ranges to confirm in writing with your own suppliers, then replace them with your measured actuals.
| Material category | Typical lead-time range | How the MOQ usually behaves |
|---|---|---|
| Greige / mill-stock fabric | Days to ~2 weeks | Minimum per quality; often the easiest to buy small |
| Dyed-to-order fabric | ~3–5 weeks | Minimum per shade per lot — the binding constraint on colourways |
| Screen / rotary printed fabric | ~4–6 weeks | Minimum per design and colourway, plus one-time screen charges |
| Digital printed fabric | ~1–3 weeks | Low minimums, higher per-metre price — useful for tests |
| Custom trims (woven labels, branded buttons, custom pullers) | ~3–6 weeks | Minimum per design, often several hundred to a few thousand pieces |
| Stock trims (plain buttons, standard zips, thread) | Days | Sold by gross, box or cone — rounding units, not real barriers |
| Custom-printed packaging | ~2–4 weeks | Minimum per print run; plain stock packaging is near-immediate |
Read that table backwards and you get your order sequence. The longest lead time on the card sets your start date, and it is rarely the fabric — custom trims and printed packaging must be committed at the same meeting where you approve the fabric. Where a minimum exceeds what one run needs, there is real room to negotiate minimums or stagger deliveries.
Dye lots: why a mid-run top-up will not match
Fabric colour is a batch outcome, not a specification. Dyeing depends on the greige lot, water chemistry, temperature, machine load and time, so two batches run to the same recipe on the same machine come out measurably different. Within a lot, shade is consistent. Across lots it is not — and a difference invisible on two swatches is obvious on a shirt whose sleeve came from one lot and body from another.
This is why rounding up is not optional. Buy 1,058 m for a run that needs 1,082 and the shortfall is not solved by ordering 30 more metres — it is solved by re-dyeing a minimum lot at a new lead time, or by shipping garments that do not match. Buy the full requirement, waste included, in one lot up front, and record the lot number against the received material. The same logic covers anything with a batch identity. If it was coloured in a bath, it has a lot.
All of this lives in the BOM — nowhere else
Consumption per size, three waste percentages, roll length, dye lot, ten trim lines with their own minimums and lead times: roughly twenty numbers per style, changing every season. In a founder’s head they vanish when the founder travels. In a WhatsApp thread they are unfindable by season two. In a spreadsheet they are one merged cell from a costing error nobody catches.
A properly structured bill of materials is the only sane home: one record per style, materials broken down per variant, waste allowance held as a field rather than as folklore, every line carrying its own supplier and lead time. From there the arithmetic above stops being manual — order 600 pieces on a size curve and the material requirement, purchase quantities and order-by dates fall out automatically. That is the calculation Honey Shelf runs when it drafts material POs, and the same discipline that turns production planning from a monthly panic into a schedule.
The founders who never miss a launch date are not better at guessing. They wrote the numbers down once, in one place, and let the run plan itself.