The honest answer is a structure, not a number
Ask five people what a t-shirt costs to make and you will get five numbers, all of them true and almost none of them comparable. The reason is not secrecy. A t-shirt’s price is the sum of decisions that have not been made yet: how heavy the fabric is, how many colours, how many pieces, what is excluded, and what the cotton price was the week the fabric was booked.
So this page does the thing a per-piece figure cannot. It sets out where the money goes, which line is worth arguing about, and how to read a quotation you have already been sent.
The one structural fact worth memorising
Materials and direct labour together account for roughly 55–65% of FOB price, and fabric alone is normally the largest single element of it [0][1].
That single fact reorganises how most buyers approach suppliers. The factory’s margin, the negotiation, the relationship and the concessions all live in the remaining minority of the number. If you want the price to move, you change the specification — the fabric weight, the dye route, the print, the packing — because those are the lines with money in them. You cannot negotiate your way to a materially cheaper garment while leaving the fabric unchanged, and buyers who try are the ones who receive a cheaper garment than they thought they ordered.
Where the money goes, line by line
The first table above is an illustrative split for a plain 180gsm cotton crew-neck at 5,000 pieces. It is constructed, not measured, and the method is stated here so you can disagree with it usefully: fabric consumption is calculated from the fabric weight and an assumed cutting waste, labour is estimated from the notified minimum wage and a standard minute value, and the remaining shares are allocated to land inside the sourced 55–65% band for materials plus labour.
Use it as a checklist. When a quote arrives, the six lines should be answerable:
- Fabric — weight, composition, consumption, and the price per kilogram used.
- Trims — rib, thread, labels, interlining, and which of them carry variable data.
- Cut, make, trim — the labour content, usually expressed as a standard minute value.
- Dyeing and finishing — the route, the colour depth, and whether your colour shares a dye batch.
- Packaging — the specification, which decides both cost and shipping volume.
- Overhead and margin — what is left, and the line to leave alone.
A quote that cannot be broken into these six is not a quote. It is an offer, and it will be renegotiated at whichever point you have least room to refuse.
Why the quantity changes the price, and not evenly
The second table above walks the quantity ladder. The important non-linearity is at the bottom and the top.
At the bottom, under roughly 300 pieces, you are not buying production — you are buying sampling with the same paperwork. Fabric minimums and dyeing minimums are fixed costs, and dividing them by 300 pieces produces a number that looks like price gouging and is actually arithmetic.
At the top, above 20,000 pieces, cost improves but exposure grows faster. A 20,000-piece order is inventory risk, and the difference between a garment that sells and one that does not now costs the same as a small house. The buyers who lose money at scale are rarely the ones who got a bad unit price.
The cost lines nobody quotes, and everybody pays
The third table is the most useful one on this page. Freight, duty, inspection, labelling, testing and pattern development are real costs that sit outside FOB, and they are the usual reason a landed-cost model that looked healthy in a spreadsheet does not survive contact with a port.
Two of them deserve specific caution:
Duty. The US position on Pakistani goods has moved repeatedly since the 29% reciprocal tariff announced in April 2025, in both directions and at short notice. Any landed-cost model carrying a duty rate from a previous quarter is carrying a number with no relationship to what you will pay. Check the current rate before you price, not after.
Testing. Destination-market rules on fibre content, care labelling, flammability and restricted substances fall on the importer. A failed test on finished goods is the expensive version of discovering a requirement, because the goods already exist and are already paid for.
What to do with this
Run the arithmetic for your own garment rather than adopting ours. You need four numbers: fabric consumption in kilograms, fabric price per kilogram, the standard minute value, and the quantity. Everything else follows.
Then take the quote you have and see whether it can answer the six lines. If it cannot, you have learned something more valuable than the price — you have learned what you would be agreeing to. If it can, compare it against another factory’s six-line quote, and the cheapest number will stop being mysterious.
Worked examples
Worked example: the arithmetic of a fabric price move
A plain crew-neck consumes roughly 0.22 kg of cotton jersey per garment once cutting waste is counted. At that consumption, every $1.00/kg move in fabric price changes the fabric element of one piece by about $0.22. On a 5,000-piece order that is $1,100 — which is why fabric price, not labour negotiation, is the line to understand. Labour moves the other way: using a notified minimum wage of PKR 40,000 on a 26-day month, a one-hour change in the standard minute value of a garment at roughly PKR 192 per hour is worth about PKR 192 per piece at one hour, and a factory with a 20-minute standard makes that PKR 64. Both figures are arithmetic from the sourced rates above, not quotations — and they show which lever is worth pulling.
Worked example: why 22% cheaper usually is not
Three quotes for a 1,000-piece tee arrive at $3.10, $3.95 and $4.20. The lowest excludes the neck rib as a separate component, assumes a two-colour print against a brief that specified four, and prices polybag packing where the brief specified a printed carton. Reconciled to the brief, it lands slightly above the middle quote. This is not usually deception: quotes are built from what a factory reads in a brief, and a brief that says "printed tee, 1,000 pieces" genuinely can be read three ways. The fix is procedural — one specification, sent to every factory, with the quote required to be broken into fabric, trims, labour, finishing, packing and margin. Once every quote uses the same lines, the comparison is arithmetic.