What a BSR Actually Means

In one sentence: a BSR is the position of a listing in a constantly updated sales-speed ladder. Rank 1 is the fastest seller in that list right now, rank 50,000 sits 49,999 places behind it, and every listing with any sales history gets a number.

Amazon's Best Sellers Rank help page is the place to check the official wording, and it is worth reading directly. Sellers and third-party guides generally describe the rank as tied to recent and historical sales and refreshed roughly hourly, but Amazon does not publish an exact per-listing schedule, so treat "hourly" as a working assumption.

Two things follow from that definition. First, a BSR carries no unit count. It tells you position, not volume. Second, it is relative: your rank falls when other books speed up, even if your own sales stay flat.

That is why a rank of 40,000 in one week can mean a different number of copies than a rank of 40,000 in another week. The ladder is the same shape, but the rungs are spaced by everyone else's behavior.

Where to Find the BSR on a Book Page

Scroll to the "Product details" block on the book's detail page. You will see a line reading "Best Sellers Rank" followed by a number in a store-wide list, and underneath it one or more smaller lines with category ranks.

Those lines are different things, and mixing them up is the single most common mistake in niche research.

The example numbers are illustrative, not pulled from real listings. The pattern holds regardless: a book can be number 3 in a small subcategory while its overall rank is in the hundreds of thousands, which means it sells a handful of copies a month.

Paperbacks and ebooks are ranked in different lists. Never plug a Kindle rank into a paperback estimate, or the reverse.

How the Rank Moves and Why It Jumps

Amazon does not publish the formula. Sellers and rank-tracking tools report that recent sales count for more than old ones, that the number reacts within hours, and that consistent sales tend to hold a rank better than a single spike. How Amazon weights a Kindle Unlimited borrow against a sale is not known outside Amazon.

Treat those as observations, not documented rules. Amazon can change them without notice.

Why one sale can change a rank dramatically

Deep in the ladder, thousands of books sit at almost the same sales speed. As far as anyone outside Amazon can tell, a single sale can move a book by a very large jump, for example from 600,000 to 250,000 (an illustrative figure, not a measured one), and the rank then drifts back up as the sale ages. A book that sells once every few days will therefore show a saw-tooth rank: sharp drops right after a purchase, slow decay in between.

Why a great rank can vanish overnight

Near the top, the opposite happens. Sellers report that books around rank 500 must keep selling through the day to hold it, and a promotion ending can push them to 5,000 within a day or two. A screenshot of a good rank proves the sales were strong at that moment, not that they will continue.

New releases get distorted early ranks

Sellers commonly report a temporary boost for brand-new listings, so a new title can show a rank better than its steady demand supports. Amazon has not confirmed the boost or how long it lasts. For competitor research, a working cut-off of 90 days is a sensible habit: set aside anything younger, and treat that number as a conservative choice rather than a documented period.

What does not affect it directly

As far as anyone outside Amazon can tell, reviews, star ratings, page count and cover quality do not enter the ranking directly. They affect conversion, which affects sales, which affects rank. The causal chain matters when you diagnose a slipping book.

Five Common Misreadings of a Rank

Rank to Sales: A Working Table for Paperbacks

There is no official conversion. Every rank-to-sales tool fits a curve to data its makers collected, and Amazon has never confirmed one. The table below is a rough scale for US paperbacks in the main Books list, unverified against Amazon data. Use it to sort listings into tiers, then run the same ranks through a dedicated calculator, such as the free BSR-to-sales calculator linked at the top, and compare.

The 30-day column is just the daily figure multiplied by 30, so 8 a day is 240 and 0.1 a day is 3. The top and bottom rows are the shakiest: tools can differ a lot at very good ranks, so check the 1,000 row especially against a second calculator before you rely on it.

Estimates are least reliable at ranks better than about 1,000 and worse than about 200,000, where the curve is steep and the data are thin. Sellers commonly see disagreements of around a factor of two between tools. The band in between, roughly 1,000 to 200,000, is where these estimates are most trustworthy.

Ebooks sell differently

Tools use separate paperback and ebook curves, and their figures for the same rank number can differ, so the two lists are not interchangeable. Amazon documents neither curve. Convert a Kindle rank with an ebook curve, a paperback rank with a paperback curve, and check the result against a second tool.

Other stores are their own ladders

Amazon.com, Amazon.co.uk and Amazon.de each keep separate lists. Because each marketplace has its own list and its own volume, the same rank number does not imply the same copies per day in the UK, Germany or the US. Use a curve built for that marketplace, or a calculator that lets you pick one.

The Formula Behind the Curve

Over a short stretch of ranks, a fitted curve often looks like a power law: sales per day = A divided by rank raised to B, with A and B fitted to data. The working table above cannot be described by one A and one B, though. It is piecewise, and the effective exponent steepens as rank worsens.

Read the table in tenfold steps. Going from 1,000 to 10,000 divides sales by five (an exponent near 0.7), and from 10,000 to 100,000 divides them by eight (near 0.9). Between 100,000 and 500,000 the table drops from about 1 a day to about 0.1, an exponent nearer 1.4. A constant-exponent formula would cut sales by the same factor every decade, which is why simple "rank divided by X" tricks mislead, and why calculators use segmented curves or steeper tails.

Interpolating between table rows on a log scale is a fair way to estimate. For a book at rank 30,000, sit between 10,000 at 8 a day and 100,000 at 1 a day: roughly 3 a day is a sensible guess, and a calculator will likely agree within a factor of two.

Worked Example: A 6 x 9 Puzzle Book at Rank 52,000

Assume a competitor's 6 x 9 puzzle book is priced at $9.99 and you recorded its overall rank over seven mornings.

Sorted low to high the readings are 38,000; 41,000; 45,000; 52,000; 60,000; 120,000; 210,000. The middle value, the median, is 52,000. Use the median, not the average, because the two outliers on Wednesday and Friday would drag an average up to about 80,000 and understate the book.

At rank 52,000 the table interpolation gives about 2 copies a day, or roughly 60 a month. Now the royalty: on Amazon marketplaces the paperback royalty is 60 percent of the list price minus the printing cost. Amazon has adjusted its royalty terms over the years, so as of September 2026 check the current rate and any low-price bands on KDP's own royalty help page before you rely on this.

Printing cost depends on page count, ink type and trim size. Assume $3.00 for this example, which is a stand-in for the figure KDP's printing-cost page would give you for your own specs. Then:

• 60 percent of $9.99 is $5.994
• $5.994 minus $3.00 printing cost is $2.994, about $2.99 per copy
• 60 copies a month times $2.994 is about $180 a month

So a comfortable single title in this slot earns something in the region of $180 a month gross royalty for the competitor, before ad spend and taxes. Sales made through Expanded Distribution are paid at a lower percentage, so a title with a lot of wholesale sales earns less per copy.

Notice how quickly the answer changes with the assumption. Halving the daily sales from 2 to 1 (rank 52,000 versus about 100,000) halves the income, from about $180 to about $90 (30 times $2.994 is $89.82). A rank around 500,000, roughly ten times worse, falls to about 3 copies a month at most: 3 times $2.994 is about $9, which is pocket change.

Validate a Niche With BSR in Six Steps

A rank is most useful as a comparison across a set of competitors. The routine below turns single readings into a go or no-go decision.

Here is how a result might look for ten hypothetical competing titles, again illustrative rather than real listings.

Six of the ten sit under 150,000, and five of them under 100,000. The median of the ten values is 117,500 (the midpoint of 95,000 and 140,000). That reads as a niche with real but modest volume: attractive if you can publish a clearly better product, unattractive if you would only be the eleventh near-identical listing.

The data cover only what is currently visible. They tell you nothing about how many books tried this niche and vanished.

The routine has a real cost. Ten titles over seven days is about 70 readings a week, and a rank-history tracker can save you the daily logging by charting rank over time automatically. Octozia's calculator does the rank-to-sales conversion but does not track ranks over time, so you still need a tracker or the spreadsheet for the readings.

Limits: What a Rank Cannot Tell You

Rank-based estimates cover only copies sold on Amazon. They ignore several things that decide whether a niche is worth your time.

• Profit. A book at rank 20,000 with a $6.99 price and high printing cost can earn less per copy than a rank 60,000 book at $12.99. Estimate royalty per copy separately.
• Sales channels. Copies sold through Expanded Distribution or bookstores are not visible in the rank and are paid at a different rate.
• Read-through and free borrows. Kindle Unlimited page reads earn through the KENP mechanism, which is separate from unit sales.
• Trend and seasonality. A holiday puzzle book can rank high in December and near zero in March. Compare year-round averages before committing.
• Ad-driven ranks. A competitor buying ad traffic can hold a rank that paid promotion is propping up.

Printing cost is a good example of a gap worth closing early. Standard trim and large trim sizes use different fee tables on KDP, and 8.5 x 11 puzzle books fall in the large-trim group. Look up the cost for your exact trim, page count and ink type on KDP's printing-cost page rather than borrowing a number from a blog post. The free printing-cost and royalty tools among the KDP calculators can run the sum for you, though the printing-cost tool models regular trim only.

If you are comparing where to print and distribute, the KDP versus IngramSpark comparison covers channel differences that a BSR alone hides.

Using the Estimate Before You Publish

Set a decision rule before you look at a single rank, so the data cannot be bent to fit hope. One usable rule for a low-content or puzzle book: proceed only if at least four of ten comparable titles have a median rank under 100,000, and your expected royalty per copy clears your minimum after printing cost.

That rule is a personal threshold, not an industry standard. Tighten or loosen it based on how much work each book costs you. Sellers who produce interiors quickly can accept thinner niches, because the cost of a miss is small. If you build interiors with a tool, remember that the export format matters for your workflow; the KDP Creator exports a ZIP of numbered 300-DPI PNG page images with a manifest rather than a finished PDF, so you assemble the interior yourself.

For book layout ideas that feed the same decision, see the guide on KDP interior templates. For the wider set of publishing topics, the KDP publishing category collects the rest of the guides.

After you publish, keep tracking your own rank. The same medians you took from competitors work on your book, and they show whether a launch tactic actually lifted normal demand or only produced a one-day spike.

FAQ