The 2020 Amazon Sales Rank Chart Is Out of Date — Here's What Still Works

What a Sales Rank Chart From Any Year Actually Is

Sales rank (BSR) is a snapshot of how fast a product is selling compared to everything else in its category. Amazon updates it roughly hourly and weights recent sales more heavily than older ones — one sale on a slow-moving product can move its rank more than one sale on a fast seller does, which is why two products with the same weekly units can show very different ranks.

A 'sales rank chart' translates that rank into a percentile: it tells you how many units in a category you'd need to beat to sit in the top 1%, top 5%, and so on. Amazon has never published these percentile tables itself. Every chart you've seen, including any from 2020, is a third-party estimate built from category product counts scraped off Amazon's category pages — and Amazon made that data much harder to pull in January 2018. Nothing since then, in any year, is an official figure.

Why the 2020 Version Specifically Will Mislead You Now

Category catalogs don't stay still. New sellers add SKUs every week, so the total product count behind each percentile keeps climbing, which means the rank number required to stay in the 'top 1%' keeps climbing too. A cutoff calculated in 2020 undercounts today's catalog by several years of new listings — the exact gap isn't published anywhere, because nobody kept a clean, audited archive of these estimates across years to diff against.

You can see the underlying instability without needing the old file at all: pull two current charts side by side and the same category won't agree. One live table lists Home & Kitchen at roughly 170 million total US products; another live table, updated the same year, lists the same category at roughly 106 million. Neither is wrong exactly — they're scraping different snapshots with different methods. One of those tables also has duplicate category rows and categories showing zero products for entire marketplaces, which is a visible sign of scraping noise, not malice. A 2020 chart carries that same measurement noise plus four-plus years of uncorrected catalog growth stacked on top.

How to Read the Chart Correctly: A Worked Example

The math behind any of these charts is simple once you have a current total: cutoff rank = category total × percentile. Take Home & Kitchen from a recent US chart: total listed products around 169,935,612, with a top 1% cutoff at 1,660,470 and a top 10% cutoff at 16,604,704.

So a product ranked 1,200,000 in Home & Kitchen today sits inside the top 1% — its rank is below the 1,660,470 cutoff. A product ranked 3,000,000 misses top 1% but still clears top 3% (cutoff 4,981,411 in the same table). The cutoff isn't a fixed constant you memorize; it's derived from that table's total at that moment, which is exactly why a number from 2020 stops matching once the total behind it has moved.

• Category — Beauty & Personal Care, total ~14,366,820: top 1% cutoff 144,903
• Category — Grocery & Gourmet Food, total ~3,761,366: top 1% cutoff 38,239
• Category — Toys & Games, total ~9,125,836: top 1% cutoff 90,720

When the Chart Disagrees With Itself

Two current sources giving two different totals for the same category is normal, not a sign either one is broken. Don't average them and don't treat either as gospel — use them as order-of-magnitude signals. If a rank cutoff you're relying on suddenly looks wrong (a product's rank barely moves despite a real sales change, or jumps oddly with no change at all), the likely cause isn't the chart — it's that Amazon can silently split or remerge category IDs, and the hourly update has lag built in.

The fix when the number looks bad or contradictory: stop trusting a single chart cell and pull the actual ASIN's rank history over several weeks from a rank tracker. A trend line on the product you actually care about beats a percentile table every time, because the table is a snapshot of the whole category and the trend line is a record of the one thing you're trying to judge.

The Mistake People Make With Sales Rank — Including Ad Teams

The classic mistake, the one these charts exist to prevent, is sourcing decisions made off a stale cutoff: assuming a rank of 50,000 still means top 1% in a category when the category has grown past the point where that's true. That's a sourcing-side error and the charts above fix it if you keep the total current.

There's a second version of this mistake on the advertising side, and we've made it ourselves: pointing at an improved sales rank as proof a DSP or sponsored campaign worked. Rank moves for reasons that have nothing to do with any specific ad — competitor price changes, seasonality, organic search shifts — and it can't isolate what your media spend actually caused. Last-click attribution has the same blind spot; it can't prove incrementality, only correlate it. We reconcile DSP and sponsored activity in Amazon Marketing Cloud instead, using holdouts and matched controls to see what spend actually added rather than watching rank bounce. Across 30 of our advertisers in July 2026, that measured portfolio delivered 6.04x return on ad spend — measured across the whole book, not cherry-picked from the best line item.

Where This Leaves You

Whether you ever run Amazon DSP or not, the same rule applies to both problems in this article: don't trust a static number you can't re-derive. reMKTR runs Amazon DSP as a managed service across 109 live advertiser seats and reconciles results in Amazon Marketing Cloud rather than reading rank or last-click as proof of anything — because a snapshot, whether it's a sales rank percentile from 2020 or a post-click conversion count from last week, only tells you what happened, not what your money caused.