Is an AliExpress sample worth buying? Use a weighted pass/fail scorecard and a simple break-even sum to decide before you import a product into your store.
Winning products, one-click import to Shopify and orders placed with suppliers.
Try AliDrop for $17-day free trial · 4.8 ★ on Shopify · Cancel anytimeAn AliExpress sample order is worth the money only if you score it instead of glancing at it. The usual routine is to unbox it, decide it "looks fine" and import anyway, until a refund request lands because the blue is really teal, the zip sticks, or the parcel arrived a month late. Below you get a weighted scorecard, a break-even sum that tells you whether a product is worth sampling at all, and a plan for the day a sample fails.
Is a sample worth paying for? Do the break-even sum first
A sample costs the product price, the shipping and a few weeks of waiting. It earns that back only by keeping a bad product out of your store. So the useful question is how many orders it takes before the refunds you avoid are bigger than what you spent.
The sum has four inputs: the chance the product is bad, the refund rate a bad product would produce, what one refund costs you, and how many orders you expect. The 20%, 35%, 30% and $4.50 figures below are placeholders. Replace them with your own numbers before you decide anything.
| Input | Base run | High-risk run |
|---|---|---|
| Sample price plus shipping ($11.00 + $1.50) | $12.50 | $12.50 |
| Chance this kind of listing turns out bad | 20% | 35% |
| Refund rate on a bad product | 30% of orders | 30% of orders |
| Cost of one refund (supplier cost $11.00 + $4.50 handling and fees, nothing recovered) | $15.50 | $15.50 |
| Expected loss per order from a bad product | $0.93 | $1.63 |
| Break-even orders ($12.50 divided by the loss per order) | 14 | 8 |
| Gross expected loss avoided over 60 orders | $55.80 | $97.65 |
| Net after the $12.50 sample | $43.30 | $85.15 |
Order the sample when the orders you realistically expect in the next 60 to 90 days are at least double the break-even number. The doubling is a margin for the soft input. The chance of a bad product is a guess, and if the true figure is half of your 20%, break-even moves from 14 orders to about 27. At 60 expected orders the sample still pays. At 20 expected orders you are under double the 14, so skip it and treat your first real order as the test.
New, unreviewed listings and the cheapest seller in a category deserve the high-risk column. A seller with thousands of steady orders deserves the base one.
The sum rests on three shaky assumptions, so know them before you trust the answer. It treats a refund as money you cannot get back from the supplier, so if a marketplace dispute often returns your cost, the case for a sample shrinks. It assumes the sample catches the bad product, which is not guaranteed. And it applies one 30% refund rate to every bad product while leaving out lost margin and return shipping.
When to skip the sample and when to insist on one
Risk overrides the math. A $3 silicone phone grip that fails is a small problem. Wireless earbuds that fail can cost you a chargeback and a bad review, even if break-even looks far away.
Cheap items still go through the sum, with a refund cost scaled down to match. Take a $4 item with $3 shipping and about $5.50 per refund (the unit plus $1.50 handling, since many buyers keep a cheap item). A $7 sample at 20% and 30% loses $0.33 per order, so it breaks even near 21 orders. At that price a sample pays off from about 40 orders.
| Product type | Sample? | Reason |
|---|---|---|
| Landed cost under $5, simple build, no safety angle | Let the sum decide | Skip below about 40 expected orders and treat the first real orders as the test |
| Fragile or liquid items | Recommended | Packaging has to survive the trip |
| Anything with a battery or charger | Mandatory | Function and safety cannot be judged from photos |
| Skincare and cosmetics | Mandatory | Label, texture, smell and leaks |
| Clothing and shoes | Mandatory | Size charts differ from supplier to supplier |
| Products you will brand or private-label | Mandatory | You are putting your name on the result |
Cosmetics carry a specific warning, and this one is US-specific. Apart from color additives, cosmetics do not need FDA approval before they go on the market, but manufacturers are legally responsible for the safety and labeling of their products, and some ingredients are banned or restricted. The FDA's page on prohibited and restricted cosmetic ingredients is where to start, and the EU and the UK have their own rules. A sample will not prove a formula is safe. It shows labeling: whether a label exists, whether it lists ingredients and whether the jar arrives sealed.
Order it the way a customer would
Pick the same variant you plan to sell, the same shipping option your store will rely on and a delivery address in your main market. A sample sent by the fastest paid method to your own desk says little about what a customer abroad receives.
If the unit price is low, order two. Shopify's guide to purchasing product samples notes that comparing several samples shows shipping and quality consistency, and that if every box from a supplier arrives damaged, a larger order may arrive the same way. For a product with several colors, one unit per main color is plenty.
Before the parcel arrives, save a screenshot of the listing, the order page and the tracking page, and write down the order date. You need it for the delivery-time check and for any dispute. Keep the product as an unpublished draft in your Shopify admin until the score is in, so nothing goes live on a hunch.
The three marketplaces differ in how much you can ask. On AliExpress, order the exact listing and variant like any customer. Alibaba suppliers tend to handle samples by negotiation, so confirm the fee, lead time and bulk specification in chat and in writing before you pay. On Temu, screenshot the listing and its return policy (in the marketplace's help pages and on your order page) before you buy, and lean harder on your scorecard. On any of them, sellers can edit photos or swap materials after your sample was bought. Treat a passing sample as proof for that seller and that listing at that moment, nothing wider.
The weighted scorecard: 10 checks, 30 points
Score every check 0, 1 or 2, then multiply by its weight. Function and safety count triple because a failure there ends in refunds or worse. Listing match counts double. The rest count once. Do checks 1, 2, 4, 5, 6, 7 and 10 at unboxing, with both units on the table if you bought two, and take photos as you go. Run the function test over a week, and fill in 8 and 9 from your order records.
| Check (weight) | How to test it | Score 2 / 1 / 0 |
|---|---|---|
| 1. Listing match (x2) | Hold the item beside the listing photos in daylight; check color, size and variant | Identical / small difference / wrong item or variant |
| 2. Build and size (x1) | Check seams and joints; weigh and measure it against the listing | Solid and accurate / minor flaws / cheap or well off the listing |
| 3. Function (x3) | Use it for the job it is sold for, for about a week | Works as promised / works with a flaw / fails at its main job |
| 4. Safety and labels (x3) | Look for the marks your market expects (CE or UKCA in Europe and the UK, FCC on US electronics), ingredient lists, warnings and a plug that fits; a mark is a reason to relax only a little | Complete and legible / one gap / missing, blurry or wrong plug |
| 5. Packaging and paperwork (x1) | Photograph the box before opening; look for crush and leaks, and for invoices or flyers that show prices | Intact and clean / scuffed but fine / damaged contents or price paperwork inside |
| 6. Included parts (x1) | Count every part against the listing; look for instructions | All present / minor part missing / key part missing |
| 7. Finish and smell (x1) | Check paint, stitching, stickers and any chemical smell | Clean / slight issue / strong smell or rough finish |
| 8. Delivery time (x1) | Compare order date and arrival date with the quoted estimate | Within quote / up to a week late / more than a week late |
| 9. Tracking (x1) | Review the tracking history for gaps and sudden status jumps | Continuous / one gap / silent for long stretches |
| 10. Second unit (x1) | Compare the second unit with the first; with one unit, score this check 1 | Identical / small variation / noticeably different |
Read the result with this ladder, top rule first, so the first rule that fits wins.
- Drop the product: a zero on check 3 or 4, or a total under 20, however cheap it looked.
- Send it back to the seller with specifics, or test a second seller: a zero on check 1 or 6, or a total of 20 to 24. A supplier can sometimes fix a wrong variant or a missing part.
- Pass: a total of 25 or more with no zero on checks 1, 3, 4 or 6.
How to run the tougher tests
Function needs a proper trial, and the right trial depends on the product. Load a bag with a few kilos and carry it for a day. Charge a gadget from empty to full and compare the time with the listing. Wash a garment once following its own care label. Fill a bottle, close it and leave it on its side in a bag for an hour.
Stress the product a little beyond normal use, since some buyers will. A printed CE, UKCA or FCC mark can be self-declared or faked, so it never proves compliance. Batteries are shipped under extra carrier and customs rules, so ask the supplier and your shipping carrier what applies before you list anything battery-powered.
You cannot repack the item in dropshipping, since the supplier ships it straight to your buyer. A poor packaging result is a supplier problem you can only solve by switching supplier. Look inside the parcel too. A packing slip with the price printed on it, or a flyer from another shop, is a reason to mark the check down.
A worked example: a portable pet water bottle
Say you are testing a leak-proof pet water bottle. The sample arrives after 22 days against a quoted 15 to 20. The color and size match the photos, so listing match scores 2, which counts 4 points. Build scores 2. Laid on its side in a bag, it weeps a few drops around the cap within the hour. A leak-proof bottle that leaks has failed at its main job, so function scores 0. The cap carries no material marking, so safety and labels score 1, which counts 3 points. The box is intact for 2 points. There are no instructions, which scores 1 on included parts. Finish and smell score 2. Delivery is slightly late for 1. Tracking is continuous for 2. The second unit has slightly different thread on the cap for 1.
That adds up to 18 out of 30. The zero on function alone makes it a drop, and so does the total. Now picture the leak going live. If that bottle sold 60 orders and 30% came back as leak refunds, that is 18 refunds at $15.50 each, or $279, against the $12.50 you spent on the sample. That is the worst case, though. Across all the products you might test, the average loss from skipping samples is much smaller: 60 x $0.93 = $55.80 at a 20% bad chance. Either way the sample paid for itself. Next, find another seller of the same bottle and score it.
What to do when the sample fails
A failed sample cost you one unit instead of a run of refunds. Photograph the problem, keep the packaging and open a dispute on the order if the item does not match the listing. The deadline to do that is shown on your own order page and depends on the order, so check it the day the parcel arrives rather than trusting a number you read online, this article included. For how the tracking countdown interacts with disputes, the guide to tracking AliExpress shipments walks through it.
Name the listing claim, describe what you received, attach the photos and say what outcome you want, for example a full refund or a partial one. If the product already sold a few orders before you spotted the fault, you can refund customers from the Orders page of your Shopify admin; Shopify's help page on refunds explains the steps for full and partial refunds.
Keep the failed unit and your notes as a reference for screening the next seller. For checks that go beyond the product itself, such as seller history and communication, work through the supplier vetting steps before you spend on another sample.
After a pass: keep the sample as your benchmark
When a customer says the strap feels thin, you compare it with the strap in front of you. When the supplier changes, you can tell whether the product did too. A passing sample proves one unit, not every batch, so keep watching your refund reasons.
Order a fresh sample whenever something material changes: a new supplier, a sudden price drop, a different listing photo set or a run of reviews mentioning a fault that your sample never had. Score the new unit from scratch.
Sampling every product is slow, so it helps to start from a shorter, pre-screened list. Before a product reaches the catalog, AliDrop checks it for real demand, a reliable supplier, fast shipping and a healthy margin, and a 7-day free trial lets you look through it yourself.
Sample order FAQ
Should I order a sample for every product I add?
No. Run the break-even sum and look at the risk. A simple item with a landed cost under about $5 and no safety angle is usually fine to test with your first real orders unless you expect around 40 or more. Batteries, skincare, clothing and anything you brand get a sample every time, and fragile or liquid items usually should too.
Can I use photos of my own sample in my store?
Yes, with care. Your own photos show what customers will actually receive, which also helps if you ever need to dispute an order. Only use them if the product you sell is the same item from the same seller, and ask before copying a supplier's or brand's images, since they may be copyrighted.
How many units should I order?
Two is a sensible default when the unit price is low, say under about $15, since a second unit reveals whether quality is consistent. Add one unit per main color or size you intend to sell, but stop there until the product proves itself.
Does a good sample still protect me if I switch to another seller?
No. A sample only describes that seller, that listing and that batch. If you change seller, or the listing changes noticeably, order and score a new sample before you import it.




