Amazon sales estimator: how far to trust the number, and the free ways to check it

How an Amazon sales estimator works

Amazon does not publish unit sales for any listing it does not own. So every tool that shows "1,240 sales a month" is working backwards from something Amazon does publish.

From Best Sellers Rank to monthly units

Best Sellers Rank is Amazon's ranking of a product against everything else in its category, based mostly on recent sales. Rank 1 sold the most recently. Rank 50,000 sold far less. The estimator takes that rank, looks up the category, and reads a unit figure off a curve the tool's makers fitted against listings whose real sales they could observe, usually their own users' stores.

The curve is steep at the top and flat at the bottom. Moving from rank 100 to rank 200 in a big category can halve the estimate. Moving from rank 80,000 to rank 90,000 barely changes it. That shape matters later, because it means small rank errors near the top produce big unit errors, and large rank differences in the long tail produce almost none.

Why every tool gives a different answer

Put the same ASIN into three tools and you will usually get three numbers, because each one fitted its own curve, on its own sample of stores, at its own point in time, and each one decides differently how to smooth a rank that Amazon updates roughly every hour.

A few things push the estimates apart:

• Rank is a snapshot. A listing that sold 40 units on a Tuesday afternoon can jump thousands of places and fall back by the weekend.
• Categories are not equal. Rank 5,000 in Home & Kitchen and rank 5,000 in a small sub-category describe very different sales volumes.
• Variations share a rank. A parent listing with 12 sizes may show one rank for all of them, and the tool has to guess how the sales split.
• Seasonality skews the fit. A curve fitted in November will read a rank as more sales than a curve fitted in March.

How accurate is an Amazon sales estimator?

Good enough to size a market, and too loose to plan a purchase order around.

Where estimates hold up

An Amazon sales estimator is good at orders of magnitude. It will reliably separate a listing selling around 30 units a month from one selling around 3,000. For a first pass through a shortlist, that is the question you actually need answered: is there demand here at all, and roughly how much?

Large, stable categories give the tool more observed stores to fit its curve against, so estimates there are tighter. Products with one main variation are easier to read too. And a range across the top 10 listings tells you more than any single row.

Where they break

They break in exactly the situations where a new seller is most tempted to rely on them. Small sub-categories have fewer data points behind the curve. Brand-new listings have a rank but no history. Products that spike on a single deal day can look like steady sellers for a week afterwards. And a listing with many variations can have its sales counted once for the whole family, which makes each individual size look stronger than it is.

A workable habit is to treat any single estimate as plus or minus half. If a tool says 600 units, plan as if it could be 300 or 900, and ask whether the product still works at the low end.

Free Amazon sales estimator checks that come from Amazon itself

You do not need a paid tool to sanity-check a number, and many "free Amazon sales estimator" tools turn out to be a paid estimator with a daily lookup limit. Amazon itself exposes more than most sellers use, and one free chart fills the rest.

Amazon's "bought in past month" figure

On many search results and listing pages, Amazon now shows a line such as "100+ bought in past month" or "1K+ bought in past month". It is first-party data: Amazon counting its own orders. It is rounded down to a threshold, so "1K+" can mean 1,000 or 4,000, and not every listing shows it. But when it appears, it is the best free check on an estimate there is. If a tool says a listing sells 2,500 a month and Amazon says "200+", believe Amazon.

Keepa's rank history chart

Keepa charts Best Sellers Rank over time, and the basic chart is free in its browser extension. What you read here is the shape of the line. A rank line that stays flat for 12 months is steady demand, one that dips every December is seasonal, and one with a single deep spike was probably a promotion. Each drop in rank usually means at least one sale, so a line that moves many times a day is selling often.

Product Opportunity Explorer

If you have a Seller Central account, Product Opportunity Explorer shows niche-level demand from Amazon's own data, including search volume and units sold over recent periods. It works at the level of a niche, not a single listing, which makes it a strong check on whether the market is as big as the estimate suggests. The free product research tools guide covers how to combine it with the other free sources.

Reading a sales estimate in a real product decision

Here is how the numbers fit together on a made-up example: a silicone baking mat selling at $19.99, where you are looking at the first page of results.

The estimate range is wide on purpose, and the useful part is the bottom of it, because a new seller realistically competes with something like the fourth-best listing. None of the rows tells you whether you can take sales from those listings. Demand and winnability are different questions, and a sales estimate only answers the first. To size the money side, run the low-end unit figure through a profitability estimate with Amazon's 15% referral fee and your real FBA fee included.

What SellRadar does instead of a sales estimate

SellRadar deliberately leaves sales estimates out. We think a precise-looking unit figure invites more confidence than a rank-based model supports, and several other tools already do estimates well. If filtering a database by estimated sales is your workflow, one of those tools is the better fit for that job.

It handles the step after that. You type a product, we read the top 20 organic listings on the marketplace you choose, including the "bought in past month" figure wherever Amazon publishes it, and return a LAUNCH, WATCH or SKIP verdict with the reasoning. The methodology page sets out exactly which signals go into it. The free tier gives you 3 full verdicts a month with no card, which is enough to test the product your estimate flagged before you spend on it.

Frequently asked questions