How to order samples from China and assess them
How to order samples from a Chinese supplier, what shipping really costs, what to measure before approving, and why a sample never guarantees the batch.
Ordering a sample from a Chinese supplier costs USD 15 in handling plus the price of the sample and the courier freight, and it is the cheapest expense of the whole import: it decides whether you produce or do not produce before you have paid for the batch. A sample shows that the factory can make the product at that level. It does not show that production will repeat it, and that distance is what the sealed reference sample and the pre-shipment inspection cover.
What a sample proves and what it does not prove
A sample proves capability: that the factory can produce to that standard. It does not prove consistency: that it will repeat it three thousand times. It is usually made more carefully than the batch, sometimes in the sample room and with different tooling. Approving a sample and relaxing is confusing the ceiling with the floor.
That is why a sample is not approved on its own: it is approved against a written specification. Without a document that says what makes it good, you only have a piece you like and nothing with which to reject the batch that arrives later. The gap is closed by the sealed reference sample, which turns the comparison into a fact, and by inspection on finished production, which turns the verdict into data.
Why a sample always costs more per unit than the batch
The unit price of a sample is always higher than that of the batch, sometimes several times higher, and it is not an abuse: it is the cost structure. Never negotiate the batch price starting from the sample price.
| Item | Sample | Batch |
|---|---|---|
| Units | One or a few | At or above the MOQ |
| Unit price | Several times the batch price | Negotiated |
The useful question is not why they charge so much for one piece, but whether that amount is credited against the order. Many deduct it from the first invoice and others do not: asking for it in writing costs nothing.
Sample shipping usually costs more than the sample
On a small sample the courier weighs more on the invoice than the product, because of chargeable weight. The carrier charges the greater of actual weight and volumetric weight, and the volumetric figure comes from dividing the volume of the package by a divisor that each one publishes. With a divisor of 5,000 cm³ per kilo — the one used by the large express carriers — a box of 40 × 30 × 20 cm bills 4.8 kg even if the product weighs 900 grams; some postal services use 6,000, so confirm the divisor in the carrier’s rate.
From that comes the useful rule: do not order one sample per shipment, order all of them and consolidate them. Three samples sent separately pay three times the minimum bracket and three inbound handlings; the same three consolidated in a warehouse in China pay one rate and one clearance, which is the same lever that brings down freight cost from China. MeliPrep receives the sample, verifies actual measurements and weight, photographs it and forwards it consolidated: USD 15 per sample, with sample and courier at cost, the handling fee published in our pricing.
The three questions to ask when ordering a sample
A badly ordered sample answers the wrong question, and that is worse than not having one because it builds confidence with no basis. All three are asked in the same message, before paying.
What is it made of, with grade and reference? “Same as the batch” is not an answer, and “good material” is not one either. You ask for the grade and the composition: whether the cable conductor is copper or copper-clad aluminium, whether the textile is cotton or a blend, whether the plastic is virgin or recycled ABS. A change of grade does not change the appearance and changes everything else.
Did it come from current production or was it made separately for me? A sample made on purpose may have gone through hands and machines the line does not have. If it came from production, ask for the lot and the date.
What is the production tolerance around the sample? The sample is a nominal value, not a specification, and without the figure for how far the batch may deviate there is nothing to inspect against: millimetres in a thickness, grams per square metre in a textile.
What to measure instead of look at
A photo does not carry a measurement, and a good part of the defects that ruin a batch are invisible to the eye: they are carried by the measurement, not by the appearance.
| Category | What is measured | With what |
|---|---|---|
| Cable | Actual conductor cross-section in mm² and actual metal | Gauge or micrometer |
| Textile | Weight in g/m² and shrinkage after washing | Scale over a known area |
| Plastic | Wall thickness in mm and Shore A or D hardness | Gauge and durometer |
| Battery | Actual capacity in mAh or Wh | Tester and full cycles |
Ask for the parameter to appear in the quotation and measure it when the sample arrives. A 0.5 mm² cable and a 0.75 one look identical in a photo, and the difference decides whether the product meets the electrical standard at destination. That is why the sample goes through a warehouse and does not go straight to your home: it is received during the warehouse and consolidation stage, where it is weighed, measured and photographed, and that figure is the one later compared with the batch’s.
How to assess the sample: test it as your customer will use it
The test that counts is not the one the factory would run, but the one your customer will run, and it lasts days, not minutes. A bench check confirms that the product switches on; a week of daily use confirms what decides a return: the coating that peels, the hinge that loosens, the battery that drops on the second day.
If the product is used daily, use it daily for a week before deciding. Whoever orders it treats it better than their customer will.
Test also what the factory does not consider product: packaging, manual and labelling. The box falls in transit and the manual is printed in the wrong language; the mandatory marking in the language of destination is covered by the electronics certification guide for Mexico and Brazil.
The sealed reference sample: what is done with it
The sealed reference sample is a copy signed, dated and kept, against which the batch is compared, and it is the only standard that survives memory. Without it, two people arguing about a defect are arguing about impressions.
It is created by signing and dating the copy and its record sheet, keeping one at origin and one at destination, and taking a set of photographs with a visible ruler. The record sheet carries the recorded measurements — weight, dimensions, thickness, weight per square metre — because the piece alone does not say what was measured.
When to take it matters: the ideal is to seal a unit from the first approved production, not the previous sample. How it is used in sampling and with what limits is on the inspection to AQL criteria page.
The comparison that decides: sample against a bad batch
Suppose an order of USD 10,000 with 3,000 units. These are the two routes.
| Line item | With sample and sealed reference | Without sample |
|---|---|---|
| Outlay before producing | Sample and courier at cost, USD 15 in handling, USD 299 per inspector-day | Nothing |
| When the defect appears | At the factory, before shipping | In your warehouse, with the supplier already paid |
| What can be corrected | The batch, with the balance retained | Replace or sell at a discount |
| Money recovered | High: the balance has not left | Low |
Pre-shipment inspection is around 3% of the value of the goods on an order like this, and the sample is a fraction of that. A bad batch is not discovered at customs: it is discovered in the returns.
When ordering a sample is not worth it
In very low-value accessories with no regulatory exposure, sampling can cost more than the information it provides. If the full order is the size you would absorb without changing your year, and the product is not customised, does not carry your brand, has no battery and no mandatory standard, the sample answers a question you already know.
In that case it is usually better to buy a small trial batch than three samples. A trial batch is real production, so it answers what a sample cannot — how the run behaves — and it comes off the same line. Measuring on receipt is still mandatory.
There are three cases that are never skipped. A product with your brand, because an error forces the tooling to be redone. A product with a battery or that plugs into the mains, because there is a safety and transport test. And a product with a mandatory standard at destination, because certification is carried out on finished product.
Customs: Mexico, Argentina and Brazil treat samples their own way
A sample is not a commercial import, but it is not free of formalities either, and writing “no commercial value” on the package does not exempt it from anything. What determines the treatment is the regime it enters under, the declared value and the origin: it varies, and it has to be confirmed before shipping.
| Country | What applies to an incoming sample | What to confirm |
|---|---|---|
| Mexico | The general rate on imports by courier went from 19% to 33.5% on 15 August 2025 | Whether your shipment falls within a de minimis threshold according to origin |
| Brazil | Samples with no commercial value: simplified declaration without registration in Siscomex (IN SRF 611/2006). With radio, Ato 18.086 of ANATEL (May 2026) releases test units only | Whether your product requires prior homologation and through which channel the test units enter |
| Argentina | INUMUE regime of the courier channel: up to 50 kg per package and USD 3,000 per shipment, for legal persons and with no commercial purpose | That the courier enters the regime code and what tax treatment applies in addition |
In all three countries the entry route is decided with the carrier or the customs broker before shipping, and in two of them the sample regime exists only through the courier channel: a sample travelling inside a consolidated shipment is not cleared as a sample, but as commercial cargo. A sample held up by a badly made declaration costs more time than the sample, and that time decides whether you produce before Chinese New Year (2027: 6 February).
What to leave in writing before paying for the sample
- The specification with material, grade, measurement and tolerance in the parameter that matters to you.
- Whether it came from current production or was made separately, and the lead time with a date.
- What happens if the batch does not match the sealed sample: replacement, credit note or retention of the balance.
- Which address it is sent to: the service’s warehouse in China, not your home address.
A well-ordered sample is a document with a piece inside; a badly ordered one is a piece with a photo. The difference is decided before paying.