KDP royalty calculator: what do you keep per sale?

Amazon pays ebook royalties at two rates, and the formulas are public. At 35%, you earn 35% of the list price with no delivery fee: a $2.99 book pays about $1.05. At 70%, you earn 70% of the list price minus a delivery fee based on your file size — Amazon's own example is a $2.99 book with a 2 MB file paying $1.79 after a $0.30 delivery charge.

The delivery fee surprises first-time publishers because it scales with file size. On Amazon.com it runs $0.15 per megabyte, so a lean text-only book loses pocket change while an image-heavy book can lose over a dollar a sale. Compressing images before upload is the highest-leverage hour in this whole topic.

Everything else on this page builds on those two formulas. If you remember only one thing, remember this: 35% means price times 0.35 with no fee, and 70% means price minus delivery, times 0.70.

On Amazon.com, the 70% option starts at a $2.99 list price — books priced below that earn 35%. There is also an upper bound to the eligible band, and the exact ceiling can change, so confirm it on Amazon's list-price page before you lock a price. Outside eligible territories, sales fall back to 35% regardless of price.

Most short ebooks by independent authors live or die on this threshold. Price at $2.99 or above in an eligible marketplace and you roughly double your per-sale royalty versus $0.99 — which is why the pricing recommender pushes short books toward the $2.99 floor instead of bargain pricing.

The table below shows both tiers side by side for a typical 2 MB file on Amazon.com, before any VAT or territory adjustments.

Take a 15,000-word how-to guide listed at $4.99 on Amazon.com with a 2 MB file. At 70%: subtract the $0.30 delivery fee ($2 × $0.15) to get $4.69, then take 70% — $3.28 per sale. At 35% the same book pays $4.99 × 0.35 = $1.75. Choosing the right tier is worth $1.53 every single sale.

Now add volume. At 100 copies a month, the 70% tier pays about $328 while 35% pays $175 — a $153 monthly gap from one pricing decision. At 20 copies a month the gap is still $31, which covers most months of an EarnDraft plan. Small numbers, real money.

Run your own figures through the widget above: change the file size to 8 MB and watch the 70% royalty drop by $0.63 a sale. That single experiment convinces authors to compress images more effectively than any lecture.

Four things move your payout after you've done the basic math. VAT or sales tax in the buyer's country comes off before your percentage. Sales outside 70%-eligible territories pay 35%. If Amazon price-matches your book below your list price, royalties follow the actual sale price. And a bloated file quietly taxes every 70% sale through the delivery fee.

None of these are traps if you expect them. Price with VAT in mind for European buyers, check the territory table if your audience is international, and never upload print-resolution images to an ebook file — 72–150 DPI is plenty for screens.

When in doubt, KDP's own royalty estimator inside the title setup shows your exact per-marketplace payout before you hit publish. Trust that screen over any third-party calculator, including this one.

Multiply your per-sale royalty by honest volume, not launch-week fantasy. A $4.99 book at 70% pays $3.28 a sale: 30 copies a month is about $98, 100 copies is about $328. Compare that against what the book cost to make with the self-publishing cost calculator — most short ebooks break even between 50 and 200 copies.

Be blunt with yourself about where copies come from. Books don't sell while you sleep unless an audience, a search ranking, or an ad budget moves them. EarnDraft is honest about this everywhere: we build the asset fast, but marketing is still your job, and anyone promising passive income without it is selling you something.

The healthy way to read this calculator is backwards: decide what the book must earn, divide by your per-sale royalty, and ask whether you can reach that many buyers. If the answer is no, fix the audience plan before the manuscript.

Then forecast the slope, not the launch spike. Month one is noise (friends, launch promos, curiosity); months two through four reveal the baseline as ratings accumulate and also-boughts settle. Model the year as a quarterly ramp: Q1 discovery, Q2 steady state, Q3 either compounding (if you're publishing book two) or decay (if the catalog sits at one).

Put numbers on the scenarios: 10 copies a month is a $33–$66 hobby at typical short-book royalties, 50 is a $165–$330 side stream, 200 is a $650–$1,300 engine worth protecting. Ask which tier your audience plan credibly reaches, then halve it for the budget and keep the other half as upside.

The forecast's real job is permission: permission to spend on book two when the base case clears costs, and permission to stop polishing book one when even the ceiling doesn't. A forecast you act on beats an accurate one you file away.

Run this five-minute checklist inside KDP: confirm your list price sits inside the 70% band for your marketplaces, check the file size on the upload screen and compress if it's over ~3 MB, read the per-territory royalty preview, and answer the AI-content question truthfully — the disclosure checker walks through it in plain language.

Then zoom out: is Amazon even your best channel? If you hold the audience yourself, direct sales keep more per copy — compare the channels with where to sell before defaulting to exclusivity.

Do all of that and the royalty line on your dashboard will hold no surprises. That quiet confidence at publish time is the whole point of this page.

Ebooks pay the highest percentage and the lowest absolute dollars per unit: ~$3.28 on a $4.99 sale at 70% versus a few dollars on a paperback after print costs — but paperbacks command higher prices and gift purchases, while Kindle Unlimited pays per page read instead of per sale. Same book, three different paychecks shaped by reader behavior, not just math.

Don't pick a format by royalty rate alone. KU favors finishable short books and demands exclusivity; paperback favors gifts, events, and credibility props; ebook-only keeps production near zero and suits testing. Most independents sequence them: ebook first for speed and data, KU or wide second by evidence, paperback once the ebook proves demand.

Whatever the mix, keep one budget that sees all of it — the cost calculator totals production once, and every format's royalties pay down the same number. Format debates get calmer when everyone agrees what 'paid back' means.

Royalties are income where you live, and platforms may withhold a slice before you ever see it — US authors file a W-9 with KDP; many non-US authors face 30% withholding without a treaty claim. None of this is advice, just the nudge: set aside a fixed share of every payout from day one, because the first tax bill arrives exactly when the book feels most profitable.

Records stay simple at this scale: monthly payout screenshots or CSVs, production receipts in one folder, and a one-line ledger per book (cost in, royalties out). That ledger doubles as your break-even tracker and your 'should I write another' evidence — accounting that pays for itself in decisions, not just compliance.

If earnings cross hobby thresholds in your country, talk to a local professional early rather than during an audit. An hour of advice costs less than one reconstructed year, and the question 'when does this become a business' has a different answer everywhere.