How to validate an Amazon product idea before you spend money

Run the cheap tests first

Every validation step costs something — time, money, or both. Ordered correctly, the cheap tests eliminate most ideas before you reach the expensive ones. Ordered badly, you pay for samples on products a thirty-second check would have killed.

Notice that stage five is the first one that costs real money, and it is the last. That is the whole design. If you are ordering samples for an idea whose competitive page you have not read, you have inverted the sequence.

Compressed to its essentials, that table is how to know if a product will sell on Amazon, which is the version to run when you have twenty ideas and ten minutes.

Stage 1 — Structural disqualifiers

Two minutes per idea. You are looking for reasons this cannot work regardless of how good the market is.

• Category gating. Some categories require approval that a new seller account will not receive. Check before anything else.
• Intellectual property. If the design, brand or mechanism is protected, you are not making a version of it.
• Hazmat, batteries, ingestibles, skin contact, children's products. All add compliance, testing or documentation burden. Not impossible; wrong for a first product.
• Size and weight. Fulfilment fees scale with both, and bulky items compound every downstream cost. Small and light is not a preference, it is a margin decision.
• Extreme seasonality. If most of the year's demand lands in six weeks, your first year is one bet with no chance to correct.
• Fragility. Breakage becomes returns, returns become negative reviews, and both arrive before you have review depth to absorb them.

This typically removes half a shortlist for almost no effort.

Private label adds disqualifiers this stage does not cover, and validating a private label product works through them.

Stage 2 — Can you break into this page?

The decisive stage, and the one most often reduced to a glance at review counts. What you are really asking is whether the twenty sellers currently holding page one include enough beatable ones.

1. Search in an incognito window on the marketplace you would sell in — logged-in results are personalised.
2. Skip Sponsored placements. They are bought positions, not the organic field.
3. Open the top 20 organic listings.
4. Record for each: review count, rating, seller name, price, and a strict one-to-five judgement of listing quality.
5. Examine the shape of the review distribution rather than its average.
6. Count generic sellers versus recognisable brands.
7. Count the listings you could produce a visibly better version of within your real budget.

SellRadar automates this stage: it reads the top 20 organic listings, flags what is weak about each one, and returns LAUNCH, WATCH or SKIP with the reasoning, framed for your selling model. The value is not that it is more accurate than doing it carefully by hand — it is that at one minute per idea you will actually run it on the ideas you are unsure about.

Stage 3 — Is there money in winning it?

A page you can enter is worthless if the economics do not clear. Do this explicitly and pessimistically.

1. Use the median price of the page, not the highest listing and not what you hope to charge. The median is where the market has settled.
2. Run it through Amazon's FBA revenue calculator with realistic dimensions and weight — that gives referral and fulfilment fees from Amazon's own schedule rather than an estimate.
3. Subtract landed unit cost: factory price plus freight, duties, and inspection. The factory quote alone is not your cost.
4. Subtract an honest PPC allowance. A new listing is invisible without advertising, and in the first months it is often the largest line after goods.
5. Consider returns in categories where fit or expectation mismatch is common.

What remains is your real margin. If it is thin at the median price, you would be entering a market you can only win by undercutting a level that is already established — which is not a business, it is a countdown.

Stage 4 — What has this market been doing?

A snapshot tells you the shape of today. Five minutes of history tells you whether today is representative.

• Is the price you used normal, or a promotion? Building margin math on a discounted week is a common and expensive error.
• Is demand seasonal? A rank line that spikes annually will make a February snapshot look dead and a November one look extraordinary.
• Has Amazon sold this listing itself? A recurring first-party presence changes the risk profile substantially.
• Which direction is rank moving? A declining category is a worse bet than a small stable one.

Keepa is the standard tool for all four and the basic charts are free. This stage is cheap and it catches the traps that no amount of present-tense analysis can see.

Stage 5 — Samples, and only then

1. Order from two or three suppliers, not one. You are testing the supplier as much as the product.
2. Compare against the incumbent you intend to beat. Buy theirs too — it is the cheapest competitive research available.
3. Check the specific thing the three-star reviews complain about. If the market's recurring complaint is a weak seam, inspect the seam.
4. Re-run the competitive read before committing to inventory. Weeks have passed; new entrants appear.

Only after stage five does a serious inventory order make sense. It feels slow while you are doing it, and it is dramatically cheaper than the alternative.

Write the decision down before you act on it

One habit improves validation more than any additional check, and it costs about five minutes per product.

Before ordering anything, write down why you decided yes — the specific listings you judged beatable, the differentiator you identified, the margin at the median price, and the assumptions the margin depends on. One paragraph is enough.

It does three things that are hard to get any other way:

• It exposes weak reasoning immediately. An argument that feels convincing in your head frequently falls apart when written as a sentence someone else could read.
• It makes the assumptions explicit, which means you notice when one changes. If your case depended on the page's median price and the page discounts, you know your case has changed rather than vaguely feeling uneasy.
• It gives you something to review afterwards. When a product works or does not, the written reasoning tells you which of your judgements were sound and which were optimistic. That feedback loop is how research skill actually improves, and without a record you are reconstructing your past thinking from memory — which reliably flatters it.

Keep them all in one place. After five or six products you will see a pattern in where you are consistently generous, and for most people it is the same place: the count of listings they believed they could out-execute.

Frequently asked questions