Amazon product research for beginners: a method that actually works

What product research is actually for

Product research has one job: to stop you spending money on a product that cannot work. It is a filter, not a discovery engine. Its success case is usually a no.

This matters because it sets expectations correctly. Beginners often approach research hoping to find a hidden goldmine, get frustrated when everything looks competitive, and eventually pick something anyway because they have run out of patience. That sequence is how most first products fail. The alternative is to accept that you will reject most ideas, that rejecting them cheaply is the point, and that the one that survives a genuinely strict process is worth the wait.

Step 1 — Get candidates (cheaply)

Candidates are abundant. Do not spend money here. Sources, roughly in order of how likely they are to produce something not everyone else has already seen:

• Something you already understand. A hobby, your job, a problem you have had. You will spot a lazy listing in your own domain instantly. This is a real edge and it is free.
• Amazon's Product Opportunity Explorer. In Seller Central, free, first-party demand data. Far less picked-over than third-party databases because fewer beginners know it exists.
• Three-star reviews on established products. If four hundred people complain the handle snaps, that is a product brief, not a review.
• Walking the category tree. Slow, manual, four or five levels deep — which is exactly why it surfaces things the filter-users never reach.

Aim for ten to fifteen candidates before you start judging any of them. Judging while you gather makes you fall in love with the third idea.

Step 2 — Kill the obvious no's in two minutes each

Before any real analysis, eliminate anything that fails a structural check:

• Gated or restricted categories. Some need approval you will not get as a new seller. Check before you research further.
• Anything with safety, electrical or ingestible complications. Supplements, batteries, children's products, anything touching skin or food. Not impossible, but a bad first product.
• Branded or patented items. If it is somebody's protected design, you are not making a version of it.
• Large, heavy or fragile. Fulfilment fees scale with size and weight, and breakage generates returns. Your first product should be small and light.
• Highly seasonal. If eighty percent of demand is in six weeks, your first year is a single high-stakes bet with no recovery time.

This usually removes about half the list, and it takes two minutes per candidate. Do it before the work that takes thirty.

Step 3 — Read the competition (the step people skip)

This is the actual research and it is the part beginner guides usually reduce to "check the review count". Here is what to do instead.

1. Search the product in an incognito window, on the marketplace you would sell in. Your logged-in results are personalised and misleading.
2. Ignore Sponsored placements. They are bought positions, not the organic field. Counting them makes every page look unwinnable.
3. Open the top 20 organic listings.
4. For each, note the review count, rating, seller name, price, and an honest one-to-five score for listing quality.
5. Look at the shape of the review counts, not the average. Flat means open. Two giants and eighteen small players means the top is closed and the middle is not.
6. Count how many sellers are generic versus recognisable brands. A page of generics is an opening for someone willing to build a real brand.
7. Count the listings you could genuinely produce a better version of. Be strict.
8. Read the three-star reviews on the top two. The recurring complaint is what your version should fix.

The decision rule: if fewer than three of the twenty look beatable, stop. That is a successful research outcome — you have just saved an inventory order.

Step 4 — Check the economics before you fall in love

A product you can win is worthless if there is no margin in winning it. Do this arithmetic explicitly, on paper, before sourcing.

1. Take the median price of the page — not the top listing's price, and not what you hope to charge.
2. Run that price through Amazon's free FBA revenue calculator with realistic dimensions and weight. It gives you the referral and fulfilment fees from Amazon's own schedule.
3. Subtract your landed unit cost: the unit price plus shipping, duties and inspection. Not the factory quote alone.
4. Subtract an honest figure for PPC. New listings need advertising to be seen at all, and it is frequently the largest line after cost of goods in the first months.
5. What remains is your margin. If it is thin at the median price, you are planning to compete below where the market has already settled.

Beginners routinely skip step four or use the best-case price. It is the most common way a product that "looked great" turns out to lose money per unit at scale.

Step 5 — Verify before you commit

1. Check price and rank history in Keepa. Is the price stable or was it discounted when you looked? Is demand seasonal? Has Amazon itself sold on this listing?
2. Order samples from two or three suppliers. Never order inventory from a supplier whose sample you have not held.
3. Re-check the page a few weeks later. Markets move. A page that looked open in March can have three new entrants by May.

Only after all five steps should money go into inventory. It feels slow. The alternative is faster and considerably more expensive.

The six mistakes that cost beginners the most

What to buy, and when

Nothing, at first. Seriously. Your first few candidates should be researched with Seller Central's free tools, an incognito window and the FBA revenue calculator. That teaches you what the signals mean — which is what lets you use a paid tool well later instead of trusting its number blindly.

When free stops being enough, it will be for one of two specific reasons: you are screening more candidates than free tiers allow, or you need alerting on markets you are watching. Buy for the reason you actually hit.

SellRadar sits at the decision step — you type a product, it reads the top 20 organic listings and returns LAUNCH, WATCH or SKIP with the reasoning and a per-listing weakness map, written for your selling model. Three a month are free with no card, which is enough to run your real shortlist. It does not do keyword research, listings, PPC or sourcing; for those you will eventually want a suite, but not yet.

When you are ready to go deeper than a beginner's overview, the full research method covers the same ground at working depth, the validation checklist turns it into something you can run line by line, and validating a specific idea walks one product end to end.

Frequently asked questions