Pick products by market first, product second. We only enter niches doing at least $2 million a year, growing, with a visible rating gap between demand and the quality of what is on offer. Score every candidate on size, trajectory, returns, competition, and differentiation evidence from negative reviews, then validate with a small test run.
The short version
- Markets make winners, not ideas. The niche's economics decide your ceiling before you design anything.
- $2 million a year is the floor. Below that, there is not enough revenue to capture profitably after acquisition costs.
- The rating gap is the opportunity signal. Strong demand plus mediocre reviews equals room to enter.
- Differentiation comes from competitor complaints. Read their one-star reviews, never brainstorm in a vacuum.
- Validation beats conviction. A 200-unit test run answers questions no spreadsheet can.
Why product picking works backwards for most sellers
The usual sequence runs idea, then product, then a market check to confirm what the seller already wants to believe. That order guarantees motivated reasoning. The mechanism that actually produces winners runs the other way: find a market with proven demand and visible dissatisfaction, then build the product that resolves the dissatisfaction.
Demand you can measure. Dissatisfaction you can read, because buyers document it in reviews with remarkable precision. When a niche shows strong sales velocity and a cluster of products rated in the middling range, the gap between what people buy and what they wish they had bought is your entry point. No gap, no entry, regardless of how much you like the idea.
The scorecard
Score any candidate niche across six weighted criteria. Anything scoring low on the first two is disqualified outright, whatever the rest says.
| Criterion | Weight | What a strong score looks like |
|---|---|---|
| Market size | 25 | At least $2 million per year in the niche |
| Trajectory | 20 | Growing demand over multiple years, not a spike |
| Rating gap | 20 | High-volume listings with clearly disappointed buyers |
| Return rate | 15 | Category norms that will not eat your margin |
| Competitive structure | 10 | No single brand owning the whole first page |
| Differentiation evidence | 10 | Specific, fixable complaints in negative reviews |
Two of these deserve expansion.
The size floor exists because customer acquisition is never free. In a $500,000 niche, even a heroic share leaves too little revenue to recover launch spend, which is why we hold the $2 million line at Flapen no matter how charming the product is. Small markets punish good execution almost as hard as bad execution.
Return rate is the silent killer. A category running high returns takes your margin twice, once in refunds and once in the review damage that raises your acquisition cost. Check it before falling in love.
Behind those six headline criteria, our research teams work through 90+ data points per market, covering segment dynamics, price band structure, and growth quality. You do not need all ninety to start. You need the discipline of scoring before wanting.
From score to shelf: the validation gate
A high score earns a small bet, not a big one. Phase one for us is about 200 units and $5,000 to $10,000, and we will run up to four candidate products through that gate at once. The test answers three questions the research cannot: does the rating hold above the niche average, does the conversion rate support the price point, and is the real cost of acquiring a customer inside the margin.
Only products that clear all three graduate to scale. The rest are cut without ceremony, and cutting them is a win, because the alternative is discovering the same answer with a full container of stock. Capital requirements for the full journey, from first order to established listing, are laid out in our FBA launch roadmap.
What product gurus will not tell you
The winning product lists sold across this industry are lagging indicators. By the time a product appears on one, its window is closing under the weight of every other buyer of the same list. Real opportunities are found in market data and review text, which do not scale to an audience, which is why nobody sells them.
Second, differentiation by invention is a trap. Adding a novel feature nobody asked for feels like strategy and tests like noise. Every differentiation decision we make comes from the negative reviews of incumbent products, because a documented complaint is pre-validated demand for the fix.
Third, most niches fail scrutiny. If your process approves one candidate in three, the process is broken. Across a serious screen, rejection is the default outcome, and the willingness to keep rejecting is what separates portfolios that compound from portfolios that limp.
Related answers
- Validate product ideas before launching on Amazon
- Saturated niches to stay away from
- Best products to sell on Amazon with small budget
- Product idea generation methods without tools
- Amazon seller roadmaps and capital: the complete guide
When you want the scoring and the validation run by people who do it weekly, talk to Flapen.
