Amazon product validation: should you sell this product?
Amazon product validation means checking whether a product sells, and whether it can make you money, before you order stock. LumaiScope’s Product Validator does it from an ASIN or a product link. It scores five signals from the listing’s data and your unit cost, blends them into a score out of 100, applies a few fixed rules and returns GO, WATCH or NO-GO.
This page sets out every signal, weight, threshold and rule the verdict uses, with a worked example, so you can check the maths on any result. It works on 12 Amazon stores (US, Canada, Mexico, Brazil, UK, Germany, France, Italy, Spain, Japan, India and Australia), and the Free plan includes 2 Validator runs a month.
What the verdict checks
Each signal is scored from 0 to 100 on its own. Its weight is its share of the blended score.
Marketplace truth
30% of the scoreIs this a real listing that is selling now?
- Sales rank earns up to 80 points. The points start at 80 and fall in a straight line to 0 at a cut-off rank.
- The cut-off is #100,000 by default. Where we have harvested the top listings in the product's category, it is 8 times that sample's 75th-percentile rank, kept between 25,000 and 500,000, so a category with deeper ranks gets a higher cut-off.
- Reviews earn 1 point per 5 reviews, up to 20 points at 100 reviews.
- A listing with no price or no sales rank can't be scored here.
Profitability
30% of the scoreIs there a margin at your cost?
- Your profit per unit after Amazon's selling and FBA fees for that store (including one month of storage), as a share of the selling price. In stores whose prices include VAT or GST, it is a share of the price without the tax.
- The score is the margin itself: a 30% margin scores 30, a margin of 100% or more scores 100, and a margin of zero or below scores 0.
- Fees are estimated from that store's fee schedule. If you have connected your Amazon seller account for the store, Amazon's own fee quote replaces the estimate.
Demand
20% of the scoreHow much do buyers want it right now?
Three parts, each scored 0 to 100 and weighted 40%, 20% and 40%. A part we can't measure is left out and the others are scaled up to fill its share.
- Sales level (40%): Amazon's "bought in past month" figure on the listing, scored as a percentile against the top listings in its category. Without a category benchmark it goes on a log scale where 100 sales score 40, 1,000 score 60 and 10,000 score 80. A listing without the figure skips this part. It is never estimated from rank, because rank already counts under marketplace truth.
- Search level (20%): Amazon's monthly search volume for a keyword taken from the product's title, on the same log scale.
- Momentum (40%): whether the sales rank is improving. Once we hold 3 daily rank snapshots, the current rank is compared with their average (up to the last 30). A flat rank scores 50, a rank 20% better than the average scores 60, and a rank twice the average scores 0. Before that, the product's weekly sales history (the last 4 weeks against the 8 before) stands in, and failing that, the change in Google searches for its keyword.
Trend signal
10% of the scoreDo two independent sources say demand is rising?
- Source one is Google searches for the product's keyword, the last 3 months against the 3 before. It sets the score: 50 plus the percentage change, so searches up 10% score 60.
- Source two is the product's current sales rank against its recent average, once we hold 7 daily snapshots.
- A source counts as rising at an improvement of 10% or more. When both are rising, a WATCH can be promoted to GO (rule 4 below).
- With no Google search data for the keyword, the signal is left out of the blend. A missing trend reading never holds a verdict back.
Differentiation
10% of the scoreHow hard will it be to stand out?
Checked in this order, and the first match sets the score:
- 500 or more reviews with a rating under 4.3 scores 75: a busy listing that buyers rate poorly leaves room to compete on quality.
- Fewer than 50 reviews scores 65: the listing is lightly contested.
- A rating of 4.6 or higher scores 35: buyers already rate the incumbents highly.
- Anything else scores 50.
Then 15 points come off if Amazon itself sells on the listing, and 10 more if 20 or more sellers are active on it. A listing with no reviews has no reading here.
How the score becomes a verdict
The five scores are blended by weight and rounded to a whole number. A signal that can’t be measured is left out and the remaining weights are scaled up to fill its share, so a missing signal never counts as zero. The score then falls in one of three bands:
| Blended score | Verdict | Shown as |
|---|---|---|
| 65 to 100 | GO | Strong opportunity |
| 40 to 64 | WATCH | Worth watching |
| 0 to 39 | NO-GO | Skip |
Five rules decide the verdict, applied in this order:
- Margin at or below zero: NO-GO, whatever the other signals say.
- Marketplace truth and demand both under 25: NO-GO. A single strong signal can't carry a product that is neither selling nor in demand.
- Otherwise the score's band above sets the verdict.
- Trend promotion: a WATCH becomes GO when both trend sources are rising and the margin is above zero.
- Missing data cap: if marketplace truth, profitability, demand or differentiation couldn't be measured, a would-be GO is held at WATCH, and the result says which input was missing. The trend signal is exempt.
A result also shows a Deal Score, which is Lumai AI’s own 0 to 100 reading of the same data. The verdict comes only from the blended score and the rules above, so the two numbers can differ.
What each verdict means, and what to do next
GO (Strong opportunity)
The measured numbers support selling it at the cost you entered. The result lists what would turn it down: the price, unit cost and total Amazon fees at which your margin reaches zero, and, where rank alone can do it, the sales rank at which the score drops below 65.
Next: Replace the estimates with real figures before you order: a supplier quote for the unit cost, and the product's real size and weight. Then run it again.
WATCH (Worth watching)
Something is promising, but at least one signal is weak or missing. The result lists the targets the product misses: a sales rank below the category median, estimated monthly revenue above the category's median units times its median price, and a review count at the category median. Without a category benchmark the targets are a rank of 15,000, revenue of 3,000 in the store's currency, and 50 reviews. If missing data held it at WATCH, the result says which input.
Next: Use the targets as a checklist, and run the product again when its numbers move.
NO-GO (Skip)
The numbers don't support selling it as things stand. If the margin rule stopped it, the result shows the break-even price and the highest unit cost that leaves any margin. Otherwise it shows, where the maths allows, the sales rank or the margin the product would need to reach WATCH.
Next: If the margin stopped it, check whether a lower unit cost is realistic. If not, move on to the next product.
What the verdict can’t tell you
- It is a judgment on measured data at the moment you run it, with no guarantee of sales. Ranks, prices and sellers change every day.
- The weights and thresholds are heuristics. They are all on this page so you can decide how far to trust them.
- The margin counts your unit cost and Amazon's fees. Shipping to Amazon, advertising, coupons and returns are left out, so add them to the unit cost you enter if you want them counted.
- A product we have only just started recording has no daily rank history, so its momentum leans on weekly sales history or Google searches until 3 snapshots exist, and the trend signal's rank source needs 7.
- Amazon shows the "bought in past month" figure only once a product sells enough, so slow sellers have no sales-level reading.
- It doesn't check patents, trademarks, brand restrictions, category approval or safety rules. Check those yourself before you order.
A worked example with round numbers
An illustration: the product is made up, and its numbers are round so the arithmetic is easy to follow. We ran them through the Validator’s own scoring code. The product sits on Amazon US, in a category we hold no benchmark for, so the default rank cut-off of #100,000 and the log scale for sales apply.
| Input | Value |
|---|---|
| Sales rank (BSR) | #20,000 |
| Average rank over its last 30 daily snapshots | #25,000 |
| Reviews and rating | 300 reviews, 4.4 stars |
| "Bought in past month" | 1,000 |
| Amazon searches a month for its keyword | 10,000 |
| Google searches for that keyword | Flat: the last 3 months match the 3 before |
| Sellers on the listing | 6, and Amazon is not one of them |
| Margin after Amazon fees, at your unit cost | 30% |
The five scores
| Signal and working | Score |
|---|---|
Marketplace truth (30%) Rank: 80 × (1 − 20,000 ÷ 100,000) = 64. Reviews: 300 ÷ 5 = 60, capped at 20. Total 64 + 20. | 84 |
Profitability (30%) A 30% margin scores 30. | 30 |
Demand (20%) Sales level: 1,000 sales score 60. Search level: 10,000 searches score 80. Momentum: 100 − (20,000 ÷ 25,000) × 50 = 60. Blend: 0.4 × 60 + 0.2 × 80 + 0.4 × 60 = 64. | 64 |
Trend signal (10%) Searches are flat: 50 + 0 = 50. The rank is 20% better than its average, so 1 of the 2 sources is rising. | 50 |
Differentiation (10%) 300 reviews at 4.4 stars match none of the special cases, so 50. No seller deductions. | 50 |
The blend and the rules
0.30 × 84 + 0.30 × 30 + 0.20 × 64 + 0.10 × 50 + 0.10 × 50 = 25.2 + 9 + 12.8 + 5 + 5 = 57
- Margin is 30%, above zero: no hard stop.
- Marketplace truth (84) and demand (64) are not both under 25.
- 57 is in the 40 to 64 band: WATCH.
- Only one trend source is rising: no promotion.
- All four required signals were measured: no cap.
Verdict: WATCH (Worth watching)
What would change it
- Google searches up 10% instead of flat: the trend score becomes 60 and the blend 58, still WATCH. Both trend sources are now rising and the margin is positive, so rule 4 promotes it to GO.
- A better margin with nothing else changed: the blend reaches 65 only once the margin passes 55%.
- A margin of 0% or below: NO-GO under rule 1, whatever the other scores.
Before and after a validation
To check a margin before you spend a run, use the free FBA profit calculator. To turn a rank into monthly sales, use the BSR sales estimator. For the manual method, read the guide to validating product demand. Inside the app, the help article How the Validator scores a product covers the same rules.
Frequently asked questions
What is an Amazon product validation tool?
A tool that checks whether a product is worth selling before you buy stock, from its sales rank, sales, search demand, competition and your margin after fees. LumaiScope's Product Validator returns GO, WATCH or NO-GO and shows the scores and rules behind the call.
How do I know if I should sell a product on Amazon?
Run its ASIN through the Validator with your real unit cost. GO means the measured numbers support it at that cost. WATCH means something is promising but at least one signal is weak or missing. NO-GO means the numbers don't support it as things stand. Check suppliers, patents and category rules yourself as well.
What score does a product need for a GO?
A blended score of 65 or more, a margin above zero, and all four required signals measured. A WATCH (40 to 64) can also become GO when Google searches for its keyword and its own sales rank are both up by 10% or more.
Why can the verdict say WATCH when the Deal Score is high?
The Deal Score at the top of a result is Lumai AI's own 0 to 100 reading of the same data. The verdict comes only from the blended signal scores and the rules on this page, so the two numbers can differ.
Which Amazon stores does the Validator support?
Twelve: US, Canada, Mexico, Brazil, UK, Germany, France, Italy, Spain, Japan, India and Australia. Fees and margin are worked out on each store's own fee schedule and in its own currency.
How many validations do I get for free?
The Free plan includes 2 Validator runs a month. Starter ($14 a month) includes 10, Growth ($29) includes 30 and Pro ($49) has no monthly limit. Running the same product again within 30 minutes of a counted run, for example with a new unit cost, doesn't use another run, up to 10 times.
Get 2 free Validator runs a month
Paste an ASIN from any of the 12 supported stores, enter your unit cost, and see the verdict with the scores behind it. No credit card.
Start free