Why Do Apparel Measurements Drift Between Batches?
14 Customization & Manufacturing
A size tolerance is the amount of variation a garment is allowed to show against its published measurement chart, and it is the reason two shirts cut from the same pattern can differ by a centimetre or two. Measurements drift between batches because fabric shrinks and relaxes at different rates, because spreading and cutting apply different tension from one run to the next, and because sewing and finishing add their own movement. Understanding where that drift comes from is what separates a controlled tolerance from an excuse for a defect.
Key Takeaways
- Batch variation starts with the fabric roll, not the sewing line: shrinkage, elasticity and recovery differ between lots.
- Spreading tension, cut placement and seam allowance control how much of that variation reaches the finished piece.
- Heat applied in printing and finishing can shrink or extend a garment after it has been measured.
- How a garment is laid out for measurement can change the number by more than the tolerance itself.
- No single global tolerance table applies to every fabric, so the standard has to be defined per product.
Where the Drift Comes From in the First Place
Drift is a chain rather than a single fault. Each stage of production takes a garment that was on-measurement and moves it slightly, and the finished piece carries the sum of those movements.
Fabric lots differ in shrinkage rate, elasticity, thickness, structure and recovery. Knitted fabrics are especially sensitive to how far they are stretched, how long they are left to relax and whether they return to their original state before cutting.
On the line, spreading tension, the placement of each cut piece, seam allowance width and the way fabric is fed through the machine all contribute. Different operators and different machines add their own variation to the same pattern.
Finishing is the last contributor and the one sellers notice late. Washing, drying, heat setting and the heat applied during printing can shrink, extend or locally distort a garment after the panels were already measured and approved.
The Six Places a Garment Moves on the Way to the Box
| Stage | What changes | Why it matters for tolerance |
|---|---|---|
| Fabric lot | Shrinkage rate, elasticity, thickness, recovery | Two lots of the same fabric weight can shrink differently after the same wash |
| Relaxation | Knit fabric relaxes after unrolling | Cutting too early locks in a measurement that later changes |
| Spreading and cutting | Tension in the lay, placement of each piece | Over-stretched fabric produces pieces that shrink back after cutting |
| Sewing | Seam allowance width, fabric feeding | Small seam differences accumulate across several seams |
| Heat processes | Shrinkage or extension during printing and pressing | A print run can move a measurement after QC has already passed it |
| Measurement | Flat or hanging, zipper closed or open, relaxed or stretched | The reading itself changes by more than the tolerance in soft areas |
Why Knits Drift More Than Wovens
A knit structure is built from loops, so it can stretch and recover. That is what makes a cotton t-shirt comfortable and it is also why a cotton t-shirt is harder to hold to a tight tolerance than a woven shirt.
Our pattern team sees the largest movement in garments that combine a soft knit with a stretch element, because the fabric keeps responding to handling. A woven fabric with a stable structure has far less freedom to move once it is cut.
Blends behave between the two extremes. A fit block built on one blend will not automatically transfer to another, which is why a new fabric needs its own tolerance check rather than an inherited one.
The Measuring Method Changes the Number
Two people can measure the same shirt and report different numbers without either being wrong. A garment laid flat and a garment hung on a form will not agree. A knit that is allowed to relax naturally reads differently from one pulled straight along the tape.
The definition of the start and end point matters as much as the tape. Chest measured pit to pit below the armhole is a different figure from chest measured across the widest point, and both are legitimate if the chart says which one is meant.
Whether a zipper is closed, whether a collar is flattened, and how much tension is applied all shift the reading. Soft and elastic areas show this most clearly, which is why a tolerance standard is only meaningful when the measurement method travels with it.
This is where a published standard helps and where it stops short. ISO 18890:2018 describes how garment dimensions should be measured, so it unifies the measuring points and the method. It is not a table that declares how many centimetres of variation are acceptable on every part of every garment.
What a Tolerance Standard Can and Cannot Do
A workable tolerance standard defines the range in advance, keeps the process inside that range and states what happens when a piece falls outside it. A standard that is written after a complaint is a negotiation, not a control.
| Question | Traditional garment production | Print on demand |
|---|---|---|
| Typical size tolerance | About plus or minus 0.5 inch | About plus or minus 1 inch, roughly 2 to 3 cm |
| Production model | Large standardised runs | Made after the order, highly customised, small batches |
| Main source of variation | Process control across a long run | Fabric lots, hand work and per order configuration |
| What counts as a defect | Deviation beyond the agreed limit | Deviation beyond plus or minus 1 inch, or a clear construction fault |
The practical threshold for a POD garment is stated in the refund policy: a size deviation beyond plus or minus 1 inch can trigger a remake or refund, while a deviation inside that range is treated as normal variation rather than a defect.
That mirrors what we measure on the floor. On a cut and sew program the tolerance is set with the fabric behaviour in mind, then checked against samples before the first production run rather than after.
Limitations: Where a Tolerance Rule Stops Working
A tolerance figure is not a blanket defence. If a garment is visibly the wrong size to the buyer, a number on a policy page does not settle the claim, and the seller is left arguing about a measurement the buyer cannot take.
Tolerance also cannot rescue a construction fault. A missing seam, a pocket sewn at an angle or a hem that is shorter on one side is a defect, not variation, and no tolerance range applies to it.
All-over print garments carry a constraint of their own, because the fabric is printed before it is cut and sewn. Pattern at a seam, an underarm or a side seam may not line up perfectly, and a shift of about half an inch to one inch at a seam is normal rather than a fault. The same process explains why a print position can sit slightly differently from the digital file.
Where the artwork itself carries the risk, the design can be adapted instead of argued over. Key text, logos and symmetrical elements placed on a seam, a neckline, a cuff or a hem are the pieces most likely to look off, so the safer choice is to move important graphics to a large central area such as the chest or the centre of the back.
Finally, tolerance is not a substitute for a size chart. A chart that is precise about measuring points reduces disputes far more than a tolerance clause does, because most complaints start from a buyer measuring in a way the chart never described.
How to Handle an Out-of-Tolerance Claim
Handle the claim with the same method the standard uses. Ask for the measurement, the measuring points and a photo of the tape on the garment, then compare that to the chart rather than to the buyer's expectation.
If the piece is outside the stated range, the fix is a remake or a refund, and the case is straightforward. If it is inside the range, the useful response is to explain the measuring method, because the buyer often measured a different point than the chart specifies.
Either way, the case is logged. Our qc team records every measurement claim against the size and the fabric lot, because a pattern of claims on one size or one fabric is a process signal that belongs in front of the pattern team rather than being absorbed one order at a time.
Setting Buyer Expectations Before the Order
A clear size chart does more work than any policy paragraph. Publishing the measuring method alongside the numbers removes the most common source of disagreement, and a size chart that cuts returns is cheaper than the returns it prevents.
The measuring points themselves are worth spelling out, and the guide to measuring shirt details shows how much difference a defined point makes. Sellers who publish the accepted size range for custom apparel up front get fewer measurement disputes, because the buyer knows what variation to expect before the parcel arrives.
Range planning matters as much as the base size. A fit developed from a base size into a full range keeps proportions consistent as sizes are graded, and the same discipline helps plus size apparel avoid the distortion that appears when a pattern is scaled without re-checking the softer areas.
Fashion fit is a separate question from measurement control, and a custom fit programme solves a different problem than a tolerance standard does. Sellers building a product line should start from a custom apparel range whose charts and measuring points are already documented.
Frequently Asked Questions
Is a 2 to 3 cm difference between batches normal?
Yes. Roughly plus or minus 1 inch, which is about 2 to 3 cm, is the accepted size tolerance for print on demand apparel. Traditional large-scale production normally works to a tighter band, because the fabric lots and the process are more uniform across a run.
Does variation between batches mean the factory is poor?
Not by itself. Fabric shrinks and relaxes, and hand work introduces variation. The quality question is whether the tolerance was defined in advance, whether it is reasonable for the fabric, and whether pieces outside it are caught and corrected.
Why does the same shirt measure differently at home?
Because the measuring method differs. Flat and hanging layouts give different readings, a knit measured under tension reads longer, and chest is often taken at a different point than the chart specifies. Compare against the stated measuring points before treating it as a tolerance failure.
Is there a global tolerance table for all garments?
No. There is no single table that covers every garment, fabric and brand. ISO 18890:2018 covers how garment dimensions are measured, which unifies the method rather than declaring one acceptable range for every part.
Do printing and finishing change the measurement?
They can. Washing, drying, heat setting and the heat used in printing can shrink, extend or locally distort a panel. That is why checking a measurement after finishing matters more than checking it immediately after sewing.
What deviation makes a garment refundable?
A size deviation beyond plus or minus 1 inch, a clear printing error or a serious manufacturing fault can justify a remake or refund. A deviation inside the stated range and a minor colour difference are treated as normal variation.
A size tolerance is useful only when it is defined before production, measured with a stated method and applied consistently after delivery. Sellers who publish the range, the measuring points and the handling process turn the most common POD complaint into a predictable conversation instead of a dispute.
Related Articles
- Fit development from base size to full range
- Average size and design in festive apparel
- Alternatives to Printify compared
