2026 dataset · 11 marketplaces · a perpetual licence, not a subscription See pricing →

About us

A tool built by people who had to make these decisions themselves.

AMZ Verdict came out of a practical problem rather than a product plan: deciding whether an Amazon product is worth launching means reading fee tables that run to dozens of pages per marketplace, and getting one rate wrong changes the answer completely.

What it actually took

The work behind this is not the interface. It is 544 category referral rates read out of Amazon’s published sources across 11 marketplaces, with the cliff and marginal tier structures distinguished by Amazon’s own wording; the fulfilment, storage, returns and exit-fee tables for each region; VAT and duty per market; and the classification forks where a product’s sub-type changes the answer.

Then the harder half: deciding what those numbers mean. Seven checks, the bands that separate a warning from a failure, the grouping of failures by root cause so one problem is not counted four times, and the rule that an unpriceable line can never produce a clearance. Those are judgement calls, they were argued out one at a time, and they are the part a spreadsheet cannot borrow.

The practices we hold ourselves to

  • Third-party summaries are never authority. Every figure is read from Amazon’s own published page or rate card. Several widely repeated third-party numbers were found to be wrong.
  • Every table is dated and rated. Source, verification date and a declared confidence, so you can see how far to trust each line rather than trusting all of them equally.
  • Assumptions are shown, never buried. If a figure was assumed rather than known, it says so on the line it affects.
  • A rule with nothing enforcing it is a rule already broken. The conventions this product runs on are asserted in an automated test suite, not remembered — including the ones about how it is allowed to talk to you.
  • It never predicts your outcome. It names the mechanism and what would change it. Nobody can honestly forecast what a product will earn.

Who advises it

Giovanni Papini advises AMZ Verdict and the judgement in it draws on his work. He has spent over ten years in e-commerce, digital marketing and Amazon marketplace management for established fast-moving consumer goods brands, across both the 3P seller FBA and 1P vendor sides — which is why this tool judges both rather than picking one.

He is the author of Selling on Amazon 1P Vendor and 3P Seller — Marketplace Wisdom for Professionals, ISBN 9798264313998, published 2025-09-07 — and the frameworks set out there are among the sources this software is built on. The acronym glossary inside the tool is taken from its glossary chapter verbatim rather than paraphrased.

4.9 out of 5 from 24 ratings on Amazon, as of 2026-08-29

The book has been reviewed and featured by illuminem, ESSEC Alumni, Professional Luxury.

What we will not build

A connection to your Amazon account. It would be the obvious next feature and it is permanently off the table — partly because live-data tools already do that job well, and partly because not needing an account is what lets an answer exist before the business does. It is also what lets an AI assistant call this on your behalf.

We also will not sell your data, because we never receive it. Your products, costs and margins are calculated in your browser and stay there.

Talk to us

Questions about the data, a figure you think is wrong, or a marketplace you want calibrated — [email protected]. Corrections are genuinely welcome; every table carries a confidence rating precisely because none of them is beyond challenge.