A buyer approves a sample. Weeks later the bulk arrives, every piece inside the agreed GSM tolerance, every inspection passed. And the buyer still writes to us and says the same thing: this does not feel like my sample.
This is the complaint we actually hear. It is rarely about a GSM number being wrong, because GSM is verified on a machine before any shipment leaves our floor. It is about hand-feel. When a buyer holds the bulk garment next to the sample, something reads as lighter, or softer, or less dense, even when the fabric weight is within a few grams of approval. Most articles about this problem either deny it happens or bury it in shrinkage charts. Nobody explains it in plain terms from inside the factory. So here is that explanation, including the parts that are uncomfortable for a factory to put in writing.
We build this kind of consistency into every order in our knitted apparel program, and the first thing worth saying is that small drift is not a defect. It is how knitted fabric behaves. The question is not whether a factory can eliminate it. No factory can. The question is whether the factory understands it, controls it, and tells you the truth about it before your order, not after.
What GSM does and does not tell you
GSM, grams per square meter, is the number everyone specifies and the number everyone measures. It is necessary. It is also incomplete.
Two pieces of fabric can sit at the same GSM and feel different in the hand. Hand-feel comes from how the yarn is knitted, how tightly the loops are formed, how the fabric is finished, and how it has relaxed. GSM is a weight measurement. It does not capture loop density or surface character. A buyer comparing a sample to a bulk piece is not running a GSM machine in that moment. They are using their hands, and their hands are reading everything the number leaves out.
So when a buyer says the bulk is lighter, the honest first step is to check the GSM internally, which we do. In most cases the number confirms the fabric is within tolerance. That does not mean the buyer is wrong. It means the buyer is feeling something real that GSM alone cannot explain.
Why drift happens between a sample and a bulk run
A sample is small. At Apex a sample run produces one or two rolls of fabric, knitted on machines we keep dedicated to sampling, and dyed together in a single batch. Everything about that run is contained and controlled.
A bulk run is a different physical event. The same fabric specification is knitted across multiple machines at once, because a single machine cannot produce bulk volume in a workable timeframe and we cannot leave the other machines idle. Those machines are set up to the same specification, but bulk machines are calibrated to handle large continuous quantities, and that calibration is not identical to a machine producing two sample rolls. The setup is close. It is not the same setup.
Dyeing adds the larger share of the variation. A sample dyes in one batch. A bulk order produces many rolls from several machines, and those rolls are dyed across multiple batches, not one. The rolls do not all come from the same machine, so they enter the dye-house already carrying the small machine-to-machine differences from knitting, and then each batch is dyed as its own event, with its own loading, its own timing, its own handling. The quantities are large and people are running the process, so small differences enter between one batch and the next. This is not carelessness. It is what happens when a process runs many times instead of once.
A few things that buyers expect to cause drift usually do not, at least not on our floor. The yarn is the same. The poly-cotton ratio is the same. Finishing is the same. The variation that matters comes from the knitting setup across multiple machines and from dyeing across multiple batches. That is where a sample and a bulk run genuinely diverge.
There is also the question of scale itself. Knitting is the foundation of the fabric, and a 100-kilogram run and a 1,000-kilogram run of the same specification do not behave identically on the machines. The larger the run, the more the fabric passes through continuous production rather than a short, contained one, and continuous production introduces its own small movement. A sample is the smallest run a fabric will ever have. Bulk is, by definition, larger. Some distance between them is built into the difference in volume.
The journey of a GSM number from greige to finished fabric
Here is the part most buyers never see, and it explains why the sample and bulk can land on the same final number through slightly different routes.
GSM is measured at several stages, not once. It is checked at the greige stage, which is the raw knitted fabric before any processing. It is checked again before finishing. It is checked after finishing. And it is checked once more after the fabric has relaxed, before cutting.
The greige fabric is deliberately knitted lighter than the target. Our greige sits roughly 20 to 25 percent below the GSM the buyer approved. That is not an error. Finishing is the process that brings the weight up. After greige GSM is measured, the finishing process is set according to that reading, and finishing compacts and settles the fabric so the GSM rises to the approved target. A sample and a bulk run can both reach 350 GSM while their greige readings and finishing adjustments were not perfectly identical. The destination matches. The path varies slightly. The hand can sometimes read that path.
For PC blends specifically, the gap stays tight. A poly-cotton blend goes through double-side dyeing, which means the fabric absorbs dye more thoroughly, and that process pulls the final GSM closer to target. In our experience a PC blend holds within about 3 to 4 percent between sample and bulk. That is tighter than many buyers expect. The only fabric that holds tighter is 100% polyester, which has the least movement of the three common compositions, with cotton-rich blends being the hardest to keep consistent.
Relaxation: the 24-hour step before cutting
After dyeing and finishing are complete, the fabric is not cut right away. It is rested.
At Apex, fabric relaxes for a minimum of 24 hours before it is cut. The GSM shift from this step is small, two to three percent, but the step is not optional and we do not shorten it. A factory under schedule pressure is tempted to skip relaxation, because 24 hours of fabric resting is 24 hours that feels like lost time. It is not lost time. It is the difference between a garment that holds its shape and a garment that does not. When a sample was rested fully and a bulk run was rushed, the buyer will feel it, and no GSM certificate will explain why.
When the difference is real, and what we do about it
Everything above describes drift that sits inside tolerance. The honest industry tolerance for GSM is around plus or minus 8 percent. Apex commits to plus or minus 5 percent, and our PC blends usually run tighter than that. Inside that band, the fabric is correct, the order is correct, and the conversation with the buyer is an explanation, not an apology.
But it is not always inside the band. We had one order where the bulk came out more than 5 percent off target. The problem was in front of us and we had two choices. Re-knit the entire bulk from zero, which costs time and money, or run the finishing process again to correct the weight. We chose to refinish. After refinishing, the GSM was still slightly above our 5 percent line, very slightly, and pushing it further risked damaging the fabric itself. So we stopped, we told the buyer exactly what had happened and exactly where the fabric stood, and we delivered the order at a discounted price.
That is the line. Inside tolerance, we explain. Outside tolerance, the buyer does not absorb our mistake. We put it in writing, because a buyer who is told the truth once will trust the next ten orders.
There was a recent case that shows the more common version of this. A buyer's T-shirt sample was 210 GSM and the bulk measured 200 GSM. The buyer did not notice the GSM at all. They noticed the hand-feel, comparing the two pieces side by side. We checked internally first, then listed every possible cause, then held a video call and walked the buyer through each one. By the end of the call the buyer agreed with our explanation and was satisfied with the order. The fabric was within tolerance the entire time. What resolved it was not a credit. It was an honest, specific explanation from people who understood their own process.
How Apex controls drift
Apex treats the sample as a process to reproduce, not just a number to hit. Samples run on dedicated sample machines, and when bulk starts, the bulk machines are set up to repeat that exact method, not just matched to the same specification sheet. Anyone can hit a GSM number once. The discipline is reproducing how you hit it. The 24-hour relaxation rule is part of that same discipline, and it is the step a factory under schedule pressure is most tempted to skip.
This matters most for the buyer described earlier, the established brand running the same order across several factories. That buyer gives an identical specification to every factory and expects identical fabric back. It does not work that way. Every factory has its own machines and its own methods. Two factories working from the same tech pack will produce fabric that is close but not interchangeable, because the specification describes the destination and not the route. A factory that understands this will tell you so. A factory that promises perfectly identical results is promising something the material cannot deliver.
What a buyer should ask any factory
You can protect yourself from unpleasant surprises by asking a few direct questions before you place an order. A factory's answers will tell you quickly whether it understands its own process.
First, ask for the GSM tolerance percentage in writing. Not a verbal promise, an actual figure on the quote or the purchase order. The tech pack should always state the GSM you approved and the range you will accept around it. A factory confident in its process will give you a number without hesitation.
Second, ask whether samples are knitted on production machines or on separate sample machines, and if they are separate, how the factory reproduces the sample method when bulk begins. A factory that has thought about this will have a real answer. A factory that has not will become vague.
Third, ask how dyeing is handled across a bulk run, whether it is one batch or several, and what the factory does to keep batches consistent. The answer tells you whether the factory understands where its own variation comes from.
Fourth, ask the factory to define hand-feel acceptance, not only GSM. Since fabric can sit inside the GSM tolerance and still feel different, a buyer who specifies only a number has left the thing they actually care about unspecified. The difference between factories shows up clearly once you understand how PC blend ratios behave on the floor, and an honest factory will talk about hand-feel in concrete terms rather than hiding behind the certificate.
There is one answer that should give you pause. If a factory promises that bulk will be exactly the same GSM as the sample, with no variation at all, that is a red flag. It is not possible. There is always a slight difference, two to three GSM at minimum. A factory that promises otherwise is overpromising, and overpromising on something this basic tells you how the rest of the order will go.
Checking the bulk when it arrives
Test your bulk when you receive it. The more professional route is to arrange an inspection at the factory before the bulk leaves, while there is still room to act. If you are testing GSM yourself after delivery, test at least five pieces and take the measurement from the center of the body, away from seams and edges, where the reading is truest.
What you are checking for is not whether the number is identical to the sample. It will not be. You are checking whether it sits inside the tolerance the factory committed to in writing. If it does, the fabric is correct, and any difference you feel is the normal character of knitted production. If it does not, you should be having a direct conversation, and a factory worth working with will have that conversation honestly.
Sample-to-bulk drift is not a flaw to be hidden. It is the nature of the material. The factories worth your time are the ones that will explain it to you before you order, not the ones that act surprised when you ask.
Faisal Ali, Apex Corporation