How to Self-Publish a Book on Amazon KDP

We may earn a commission from some links on this site at no cost to you.

Self-publishing on Amazon KDP is a catalog business. You publish books, Amazon prints each paperback the moment someone orders it, the Kindle store delivers the ebook instantly, and you collect monthly royalties from search traffic you never have to chase. You can start for under $1,000, sometimes close to zero if you write the book yourself. The catch is pace: most self-published books sell few copies, money arrives months in, and Amazon pays royalties about sixty days after the month they are earned.

None of that is a reason to skip it. It is a reason to work the method, which is exactly what our KDP self-publishing course teaches. This guide runs the same steps in order: royalty math, shelf research, production, packaging, launch, and the catalog.

Learn the royalty arithmetic first

Most people write the book and then discover what a sale pays. Reverse that, because every decision downstream depends on four pieces of arithmetic, all covered in Royalty arithmetic before you write a word.

Ebooks pay 70 percent royalty inside a price band, which since July 2026 runs $2.99 to $12.99 on Amazon.com, minus a delivery fee Amazon says averages about six cents. Outside the band you get a flat 35 percent. A $4.99 ebook earns about $3.43 per sale. A $13.99 ebook earns about $4.90, and here is the trap: at $12.99 you earn roughly $9.03, and at $13.00 you earn about $4.55. One dollar of list price nearly halves your take. Price inside the band unless you have a specific reason not to.

Paperbacks pay 60 percent of list price minus printing cost, with a cliff of their own. Amazon's own example: a 300-page black-and-white paperback at $9.99 earns $1.39. The same book priced $9.98 slips into the 50 percent rate and earns $0.39. One cent, one dollar. Printing costs are a published table: $2.30 flat for 24 to 110 pages, then $1.00 plus 1.2 cents per page, which is why a 150-page journal costs $2.80 to print and earns $1.99 at a $7.99 price.

Kindle Unlimited pays per page read, at a rate that has run roughly $0.004 to $0.005 per page, so a 300-page novel read cover to cover earns $1.20 to $1.50. That sounds tiny until you notice genre fiction readers binge, and page reads can rival or beat sales revenue for fiction in the program. Audiobooks pay better rates through ACX but production can run four figures per book for a narrator, so treat audio as a later-stage move for books that already sell.

Nobody retires on one sale. A $4.99 ebook makes about $3.40, a low-content journal about $2, a page-read novel about a dollar-forty per full read. The operators who earn real money here run catalogs: tens of books, each found by a different search, each contributing a trickle.

Research the shelf, not the market

Research in this business is shelf reading, and Research the shelf before you write the book turns it into a five-minute scan. Search your candidate phrase on Amazon and read the first two pages like a scout report: titles, review counts, publication dates, cover quality at thumbnail size.

You are answering three questions. Is money moving, meaning at least three comparable books with a strong sales rank, under roughly 500 on Amazon's best-seller list? Is there a gap, like old covers or generic titles on proven demand? And can you be credible on this shelf, because shoppers smell padding and reviewers say so permanently?

The move that turns a crowded shelf into an open one is sub-niching. Not "meal prep" but "meal prep for shift workers." Not "grief journal" but "grief journal for widowers." Every sub-niche word does two jobs at once: it tells that specific reader the book is for them, and it places you in a less contested corner of search.

For low-content books, add the gift flip. Nobody wakes up wanting a journal, but people search "gifts for" constantly. In one widely taught example from this lane, "hedgehog journal" drew about forty Google searches a month while "hedgehog gifts" drew around 2,900. Build the journal, title it for the gift searcher.

Produce the book: write it, hire it, or assemble it

Producing the book lays out three roads. Write it yourself: free, slow, and the best book when the expertise is really yours. Hire a ghostwriter: full nonfiction budgets taught in this space run roughly $1,200 to $2,000 including cover and formatting, and you commission a paid writing test before signing anyone for a whole book. Or assemble it: the low-content road, where you build a clean interior once and produce covers per niche.

Whatever the road, budget for editing. Proofreading runs about two cents per word and copyediting about 2.7 cents, per Reedsy's 2025 marketplace data, so a 30,000-word manuscript costs roughly $600 to proofread. Skipping editing is the most expensive decision in this business, because it shows up in one-star reviews that never come down.

The AI rules deserve exactness, because this is where accounts die. Amazon requires you to disclose AI-generated content at publishing, meaning content created by an AI tool from your prompts. AI-assisted content, where you wrote the material and used AI to brainstorm or edit, requires no disclosure. Lying on that disclosure is an account-level risk. And per the US Copyright Office, fully AI-generated text is not copyrightable, so the "AI does the work" fantasy builds a catalog with no protectable content in it. Author the book, let AI assist, disclose what the machine generated, edit until the voice is yours.

Package it so the right shopper clicks

Production was never the moat, since anyone can produce a book now. The moat is packaging: covers in the $150 to $250 budget tier, titles with the sub-niche words in them, and keyword choices that put you on the right shelf. The course's covers, titles, and keywords lesson is the fight that decides whether a finished manuscript sells.

Launch with reviews and small ads

First readers matter because a book with no reviews converts badly. Then ads, run as the course teaches in Ads on a small budget: one auto campaign at a modest daily budget in the fifteen-to-thirty-dollar range operators teach, bids near the floor at twenty to forty cents, plus a manual keyword campaign and a manual product campaign on comparable books. Then the hardest part: leave it alone for one to two weeks, and let a campaign spend around fifty dollars before judging it.

The governing number is ACOS, ad spend divided by the revenue it generated. Your break-even is royalty divided by list price: a book at $12.97 earning a $5.56 royalty breaks even at about 43 percent. Fund the first month's ads, $100 to $300, as tuition. Most of it will not come back, and if that money is rent, do not spend it.

Plan for the lag, then build the catalog

Amazon pays royalties monthly, roughly sixty days after the month they were earned. March's sales arrive near the end of May. Ad spend is real-time while royalties are not, so money you can float for ninety days is the honest definition of a year-one ad budget.

Then the part that decides everything. A single book faces a cold problem: low rank, no visibility, no momentum. A catalog changes the mechanics, since each book is a separate entry point into a different search, books in the same niche cross-sell on the also-bought row, and your production costs fall with practice. The honest first-year numbers from What the money actually looks like: one documented author sold 11,000 copies and kept about $3,000 of profit after ads and production. A publishing-industry estimate puts median income near $12,749 for authors who treat it as a business with three or more books, and you should read that as a ceiling-shaped data point, not a promise. One low-content operator reported about $100 across the first three months, and later months in the high four figures after more than a year of building.

The unit of progress is not "my book." It is "my next book." The publishers who quit almost always quit after book one, before any compounding could start.

Who this fits

It fits patient people who like research and repetition, working alongside income rather than instead of it, since back-loaded money and empty runways do not mix. Our runway planning course handles that split. If you are choosing between this and the other under-$1,000 businesses, the startup quiz sorts them by budget and temperament. And if you want the full method with worksheets, the KDP course is free, start to finish.