Run a small paid batch and read five signals: conversion rate against the category, rating trend, return rate, customer acquisition cost trajectory, and repeat purchase where the category allows it. Fit exists when conversion holds while acquisition cost falls. Reviews and revenue alone tell you what you paid for, not what you found.
The short version
- Revenue is not evidence. Anyone can buy revenue on Amazon. The question is what it costs to buy the next unit.
- Conversion rate is the primary signal, measured against the category, not against your expectations.
- Rating trend beats rating average. The direction tells you what the last hundred buyers thought.
- Return rate is the cheapest lie detector you have. It reports the gap between the listing and the product.
- Fit is confirmed when acquisition cost declines while conversion holds. One without the other is not fit.
What the buy side taught me about this
Before Flapen I ran data and technology at BRANDED and Moonshot Brands, two large Amazon aggregators. The job put me on the other side of this question, because an acquirer is paying today for revenue that has to survive without the founder, without their launch budget, and often without their promotional tricks.
What held value was boringly consistent. Brands with strong conversion against their category, a rating that was flat or rising, and a modest return rate kept earning after the deal closed. Brands whose revenue rested on heavy promotion or a temporary advertising push did not, and their numbers looked identical in a summary spreadsheet. That is the whole discipline of testing fit. You are trying to see, early and cheaply, which of those two things you have.
Three ways to test, compared
| Small validation batch | Pre inventory demand probe | Full launch and observe | |
|---|---|---|---|
| What it is | Around 200 units live on Amazon with real advertising | Paid traffic to an offer before committing stock | A full purchase order and a complete launch |
| Typical cost | $5,000 to $10,000 | Lower, mostly media spend | $8,000 to $15,000 and up |
| Time to a readable answer | Weeks | Days to weeks | A quarter or more |
| What it proves | Conversion, rating, return rate, acquisition cost | Interest and price sensitivity | Everything, expensively |
| What it cannot prove | Long term repeat behavior | Whether the physical product satisfies | Nothing you could not have learned sooner |
| Best for | Almost every first product | Categories with unusual claims or price bands | Brands with capital and an existing catalog |
The decision rule: run the validation batch unless your product is unusual enough that you doubt anyone wants it at your price, in which case probe demand first, then run the validation batch anyway. The full launch is not a test. It is a bet you place after the test.
We test up to four products inside one validation envelope for this reason. Four cheap answers beat one expensive opinion, and the market tends to pick a different favorite than the team does.
Reading the five signals
Conversion rate. Compare it to the category, because a 12 percent conversion rate is excellent in one segment and poor in another. This is the signal that leads all the others, and a weak conversion rate is not a traffic problem waiting for more budget.
Rating trend. Look at the direction over the last four weeks rather than the lifetime average. A 4.4 that is falling and a 4.4 that is climbing are two completely different products.
Return rate. Track it from week two and read the reasons. Returns are the market telling you precisely where the listing and the unit disagree.
Acquisition cost trajectory. Not the level, the slope. Early advertising is expensive by design. What matters is whether the cost per acquired order is falling as rank and reviews accumulate. Flat or rising over six to eight weeks is the clearest negative signal on this list.
Repeat purchase. Only meaningful in consumable categories, and enormously meaningful there. If nobody buys twice in a category where people should, the product is not the one they wanted.
What most agencies will not tell you
Ask an agency to prove product market fit and most will show you a revenue chart. What most agencies will not tell you is that revenue was the easiest of your metrics for them to influence and the least informative one you own.
Differentiation is where this becomes concrete. Real fit usually comes from fixing something the category gets wrong, which you find in competitor negative reviews and in the gap between what buyers rate highly and what they complain about. It almost never comes from invention. An agency that cannot show you the review analysis behind their recommendation has skipped the step that decides everything and gone straight to executing a launch, because executing is what they get paid for.
Related answers
- How to hit product-market fit on Amazon in 3 months
- Is dropshipping to Amazon worth it
- Keyword research workflow for Amazon newbies
- What does a good Amazon account audit include
- Amazon launch services: the complete guide
If you want a second read on whether your current numbers show fit, the audit is free at Flapen.
