A sample arrives, feels right in the hand, and gets approved. Six weeks later the bulk shipment feels different — thinner coating, looser cuff, a different shade. Nothing in the sample approval process caught it, because the approval was based on feel rather than measurement.
Samples are evidence. This is how to collect evidence that will still mean something when 20,000 pairs arrive.
Before you open the box
Record four things the moment the parcel arrives:
- The exact model code and specification sheet the sample claims to represent.
- Size marked on the sample and whether it matches your request.
- Batch or lot reference, if present.
- The date the sample was made, or at least the month.
If the sample carries no model code and no size marking, you are not evaluating a product — you are evaluating a hope. Ask for a marked sample.
The measurement pass
Do this before any wear test, because it is objective and repeatable.
| Check | How | Why it matters |
|---|---|---|
| Total length | Measure from cuff edge to middle fingertip, flat | Length varies between factories even at the same nominal size |
| Palm width | Across the knuckles, flat, unstretched | Predicts fit complaints more reliably than the size label |
| Pair weight | Weigh three pairs, record the spread | A wide spread means inconsistent coating pickup |
| Coating weight | Weigh before and after removing coating if you can | The single best proxy for coating thickness |
| Cuff length and elastic recovery | Stretch and release five times | Weak recovery means the cuff will open during use |
| Colour | Compare against your reference under daylight, not office light | Colour drift is the most common bulk complaint |
Write the numbers down. If samples from two suppliers are compared only by hand feel, the comparison is an opinion.
The functional pass
Coating adhesion. Bend the glove sharply at the knuckle 20 times, then look for separation between coating and liner. Lift the coating edge with a fingernail at the cuff. Coating that peels at the edge will peel in use.
Grip, wet and dry. Pick up a steel bar, a glass panel and a plastic sheet, dry and then wetted. Record which surfaces fail. Most grip complaints are wet-grip complaints, and a sample that fails wet grip on the target surface will not improve in bulk.
Abrasion. Rub the palm area on 80-grit paper or concrete for a fixed number of strokes under hand pressure. Compare coating loss between candidate samples. This is not a lab test, but it ranks samples reliably.
Dexterity. Pick up a coin, a screw and a cable tie. Time it. Gloves that fail basic dexterity get taken off on site, which is the most expensive outcome of all — a glove that is not worn protects nobody.
Puncture and cut claims. If the sample claims cut resistance, check that the marking on the glove matches the specification. A cut claim with no marking on the product is a claim, not a specification.
Heat and chemical claims. If the glove is sold for heat or chemical exposure, ask for the test report for that exact model. A generic factory certificate does not transfer to a new model.
The wear trial
Two people, two weeks, real work. Not five people for one day — that produces anecdotes, not data.
Give the gloves to the two people doing the harshest version of the job. Ask them to report:
- Where the glove fails first (coating wear, seam split, cuff failure, colour transfer)
- How many days or shifts it lasted
- Whether they would keep wearing it if it were their own money
The “where it fails first” answer is the one that decides your specification. It tells you which property to spend money on — and which to stop paying for.
Red flags that predict a bulk problem
- No model code or size marking on the sample
- Pairs within one sample bag differ in weight, shade or cuff length
- Coating smell is strong and chemical — often under-cured coating
- Coating edge lifts at the cuff straight out of the bag
- The sample is “the same as” a product you cannot get a specification sheet for
- Size runs small across the board — a common way to hide a cheaper liner
- The supplier cannot say which line or factory made the sample
What to put in writing
Before approving, send one page back to the supplier stating:
- The exact specification approved (model code, gauge, liner, coating, coating weight or pickup, cuff, marking, packaging)
- The sample reference number and the date it was approved
- The acceptance criteria for the bulk order — including the AQL level and the inspection stage
- What happens if bulk deviates from the approved sample
That page is what protects you later. Without it, every bulk dispute becomes a memory contest.
Frequently asked questions
How long should a sample evaluation take? The measurement and functional passes take an afternoon. The wear trial should run at least one to two weeks of real use — faster trials hide the failure modes that matter.
Should I pay for samples? Paying for samples is normal and often a good sign, because it means the sample is made for you rather than pulled from stock. Ask whether the charge offsets against the first order.
One sample or several? Ask for two or three pairs of each candidate in the size your users actually wear. One pair cannot show you consistency, and consistency is what you are buying in bulk.
What if only one of my two candidate samples is properly marked? Treat the unmarked one as unverified. A supplier who does not mark samples usually does not mark cartons either, which becomes a receiving problem at the warehouse.
Related: what the EN 388 marking actually tells you, and how we handle private label marking and packing.