Most failed product tests aren't caused by the ads. They fail because the product was never going to work: demand was too thin, the margin couldn't pay for a customer, or the item was too fragile or too risky to ship from overseas.
Validation is the work you do before the test campaign. It's cheap, it takes an hour per product, and it removes most bad candidates before they cost you anything. Here is the checklist we use, in the order we use it.
Step 1: Check that demand is real
Start with the number most sellers look at first: lifetime orders on the AliExpress listing. It is a lagging signal, but it is hard to fake at scale and it tells you whether anyone has bought the product at all.
We read it in tiers:
| Lifetime orders | What it tells you |
|---|---|
| Under 100 | Cold demand. Not enough buyers to judge. Product Score caps the Opportunity Score at 40 here, whatever the other dimensions say. |
| 100 – 500 | Early traction. Worth watching, risky to test. |
| 500+ | Proven demand. Real buyers exist. |
| 2,000+ | Top demand tier. Proven, but check saturation (step 5). |
Two cautions. Order counts are per listing, so a product sold by twenty suppliers may look smaller than it is: check several listings before you call demand cold. And lifetime orders don't show when people bought. Read the most recent reviews: if the newest ones are months old, the demand may already have peaked.
Step 2: Name the problem or the desire in one sentence
If you can't finish the sentence "People buy this because…" in under ten words, a cold audience won't understand it in a three-second ad either.
Good answers are concrete: "…their cable keeps falling behind the desk", "…their dog pulls on the lead". Weak answers are vague: "…it's cool", "…it's a nice gift". Vague products can still sell, but they lean entirely on creative and need a much bigger testing budget.
Step 3: Run the unit economics
This is the step that kills the most products, and the one beginners skip. Before you look at anything else, work out how much you can afford to pay for a customer.
Take a product that costs $8.00 delivered and sells for $24.99. With a 2.9% + $0.30 payment fee and a 3% returns allowance:
- Payment fees: $1.02
- Returns allowance: $0.75
- Contribution before ads: $15.22
That $15.22 is your break-even CAC: the most you can spend on ads to win one order without losing money. Now compare it with what customers typically cost on your channel. Our planning benchmarks are $9.50 for TikTok, $14 for Meta and $17.50 for Google Shopping in the US. A break-even CAC that sits well above the benchmark leaves room for a learning phase. One that sits below it means you are paying to lose money from the first sale.
A quick sanity check before you do the full maths: aim for a retail price of at least 2.5x the supplier cost, inside the $19.99–$39.99 impulse band. The break-even CAC calculator does the full calculation, and our guide to break-even CAC and ROAS explains each line.
Step 4: Check that it can survive the trip
Dropshipping adds 7–20 days of handling by people who don't care about your customer. Ask:
- Weight and size: light, compact items ship cheaply and arrive intact. Bulky items pick up volumetric shipping costs.
- Fragility: glass, ceramics and thin plastics break, and every breakage is a refund plus a reshipment.
- Restricted contents: batteries, liquids, aerosols, powders and magnets often face carrier restrictions.
- Sizing: clothes and shoes bring size-related returns that are expensive to process across borders.
A product that fails here is a dropshipping blocker: it may sell well from a local warehouse but not from a supplier abroad. Product Score caps such products at 55 so a strong demand score can't hide the problem.
Step 5: Look for the risks that end stores
Some products are profitable right up until they get your ad account or payment processor shut down. Rule out:
- Anything that needs health, medical or safety claims to sell
- Products for children or babies, which face strict safety rules in most markets
- Branded, licensed or look-alike items (IP complaints)
- Items that touch skin, food or the body without clear material information
- Weapons, vapes and other restricted categories on ad platforms
Products with serious risk are capped at 50 in Product Score. These caps are our safety gates: a product can't earn its way past them with good scores elsewhere.
Step 6: Gauge the competition
High demand with no competition is rare. High demand with heavy competition is normal, and still workable if you can stand out. Search the product on AliExpress and count near-identical listings, then look it up in the Meta Ad Library and TikTok Creative Center to see how many stores are advertising it right now.
We cover this in detail in how to tell if a dropshipping product is saturated.
Step 7: Sketch the first three seconds of the ad
Before ordering a sample, write down the opening shot of your ad. If it's obvious (a before/after, a satisfying demo, a visible problem solved on screen), the product has creative potential. If the only idea is a product photo on a white background, expect expensive ads.
Putting it together: a validation scorecard
| Check | Pass | Fail |
|---|---|---|
| Demand | 500+ orders across listings, recent reviews | Under 100 orders, or no reviews in months |
| Clear reason to buy | Problem or desire stated in under ten words | "It's cool" |
| Unit economics | Break-even CAC well above your channel benchmark | Break-even CAC near or below it |
| Logistics | Light, robust, no restricted contents | Heavy, fragile, sized, or restricted |
| Risk | No claims, no IP, no regulated category | Any of these |
| Competition | Room to differentiate | Identical listings and identical ads everywhere |
| Creative | Obvious 3-second hook | Nothing to show |
A product doesn't need to ace every row, but a fail on unit economics, logistics or risk should end the conversation. Those are the three that no amount of ad skill fixes.
When you do test: set a kill rule first
Validation lowers the odds of a bad test; it doesn't remove them. Before launching, decide when you will stop. A common rule of thumb is to stop a product once you have spent one to one-and-a-half times its break-even CAC with no sale, or after several days of spend with ad-level signals (click-through rate, cost per add-to-cart) clearly worse than your other tests. Writing the rule down beforehand keeps you from "giving it one more day" with money you've already decided to risk.
Do this faster
Product Score runs steps 1 to 6 for every listing in a research: it scores demand, logistics, risk and competition across 9 dimensions, applies the safety gates, and estimates break-even CAC for your channel and market. Anything that clears the 65+ shortlist threshold is worth the hour of manual validation above. See how scoring works or browse scored products by niche on the winning products pages.
