In short
Amazon KDP gives you seven keyword fields of up to 50 characters each, and three categories. The keywords tell Amazon's search engine which reader queries your book should answer; the categories decide which bestseller lists it can appear on. Never repeat words already in your title or author name. Amazon indexes those automatically.
Amazon is a search engine that sells books
Readers do not browse Amazon. They search it. They type "cosy mystery small town bookshop" or "how to start a podcast" into a box and buy from what comes back.
Which means your book's visibility is not decided by its quality. It is decided by whether Amazon's index can match it to the words a reader typed. And that index is built from metadata you supply, in about ten minutes, on a screen most authors rush through to get to the publish button.
This is the highest-leverage, lowest-cost work in self-publishing. It is free, it takes an afternoon to do properly, and it is reversible at any time.
Seven fields, not seven words
KDP gives you seven keyword fields, each accepting up to 50 characters. This is the first thing people get wrong: they treat them as seven single words.
They are seven phrases. You have roughly 350 characters of search real estate, and a field holding one word wastes forty of them.
| Field | Wasteful | Effective |
|---|---|---|
| 1 | mystery | cozy mystery small town bookshop |
| 2 | detective | amateur sleuth female detective series |
| 3 | murder | light murder mystery no gore |
| 4 | books | book club mystery for women |
| 5 | fiction | english village whodunit |
| 6 | crime | clean read mystery cat sidekick |
| 7 | novel | first in series cozy crime |
The left column is what most authors type. The right column is what readers actually search.
What a good keyword phrase looks like
A useful keyword phrase has three properties:
- A reader would plausibly type it. Say it out loud. If it does not sound like something a human would search for, it is not a keyword. It is a description.
- It is specific enough to have low competition. "Fantasy" is a wall you will never climb. "Slow burn enemies to lovers fae romance" is a door.
- It describes the reading experience, not just the plot. Readers search by mood and by promise. "Clean read", "slow burn", "no cliffhanger", "light-hearted", "gripping page turner". These are among the most-searched terms in fiction and almost nobody uses them.
Also worth covering, if they apply honestly:
- Setting. "Scottish highlands", "1920s New York", "space station".
- Character type. "Single dad", "retired detective", "reluctant hero".
- Comparable titles or authors. Permitted only where genuinely comparable, and never where it would mislead a buyer.
- Reader occasion. "Book club pick", "beach read", "gift for dad".
What never to put in a keyword field
Every one of these wastes a slot, and some risk a metadata violation.
- Words already in your title, subtitle, series name or author name. Amazon indexes those fields automatically. Repeating them buys you nothing and costs you a slot.
- Your own book title. Same reason. Ranking for your own title is already handled.
- Category names. Your categories are set separately. Do not spend a keyword field on "romance".
- Other authors' names or trademarked titles, unless the comparison is genuine and non-misleading. Amazon actively enforces this, and the penalty is title suppression.
- Subjective claims. "Bestselling", "award-winning", "best book ever". Not searchable, and against metadata guidelines.
- Time-sensitive claims. "New", "2026", "latest".
- Misspellings. Amazon handles these. You do not need to.
- Repeated words across fields. Once Amazon has indexed a word, a second appearance adds nothing.
Keyword stuffing is a real risk, not a theoretical one. Cramming unrelated search terms into your title or subtitle to catch traffic violates Amazon's metadata guidelines and gets titles suppressed. The keyword fields exist so you do not have to do that: use them, and leave your title clean.
How to find phrases readers actually type
You do not need a paid tool. Amazon will tell you for free.
-
Use the Amazon search bar as a keyword tool
Start typing a phrase in the Kindle store and read the autocomplete suggestions. Those are real, high-volume queries from real shoppers. This is the single best free keyword source in existence, and it is sitting in front of everyone.
-
Work through the alphabet
Type your genre plus "a", then "b", then "c". Each letter surfaces a different set of real searches. Collect everything that fits your book.
-
Read your competitors' product pages
Look at how the top 20 books in your subcategory describe themselves. The recurring phrases in their descriptions are the vocabulary your shared readers use.
-
Mine your reviews, and theirs
Readers describe books in reviews using exactly the language they search with. "I wanted something light with no swearing" is a keyword phrase wearing a disguise.
-
Test the phrase before you commit
Search each candidate on Amazon. If the results are books like yours, it is a good keyword. If they are unrelated, it is not: regardless of how well it describes your book.
Categories: specificity beats size
Categories decide which bestseller lists your book can appear on. This is not vanity: an orange #1 Best Seller badge on a product page measurably lifts conversion, and it is won in a category, not in the store as a whole.
Which produces a counterintuitive rule: choose the smallest accurate category, not the biggest.
| Broad | Specific | |
|---|---|---|
| Category | Fiction > Mystery | Mystery > Cozy > Culinary |
| Competition | Tens of thousands of titles | Hundreds |
| Realistic rank | Invisible | Top 20 is achievable |
| Bestseller badge | Essentially never | Genuinely possible |
| Reader who finds you | Nobody | Exactly the right one |
A high rank in a small, accurate category beats invisibility in a large one every time. The reader browsing "Cozy > Culinary" is precisely the reader you want; the reader browsing "Mystery" is not browsing at all, because nobody browses a list of forty thousand books.
Choosing your three
KDP lets you select three categories at setup. Use them deliberately:
- One primary: the most accurate description of what the book is.
- One adjacent: a genuine secondary readership. A cosy mystery with a strong romance thread legitimately belongs in a romance subcategory too.
- One narrow: the smallest accurate category you can find, chosen specifically because you can realistically rank in it.
The rule that binds all three: accuracy is not optional. Placing a book in a category it does not belong to is a metadata violation, and it also fails on its own terms: readers who find your book in the wrong place leave one-star reviews saying it was not what they expected.
The description does double duty
Your book description is indexed by Amazon's search engine and read by every human considering the purchase. One field, two jobs, and most authors serve neither.
It is not a synopsis. It is sales copy. Structure that works:
- A hook in the first line. Amazon truncates the description behind a "read more" link on mobile. Everything that matters has to be above the fold.
- The promise. What will the reader feel, learn, or get?
- Enough plot or content to make it concrete, and no more. You are selling the book, not summarising it.
- Social proof, if you have any.
- A clear close. Tell them to buy it.
Natural use of your key phrases here helps the index. Stuffing them does not, and reads as spam to the human who was about to buy.
This is not a one-time decision
Keywords and categories are editable at any point, and republishing takes effect within about 72 hours. Almost nobody revisits them.
A reasonable rhythm:
- At 30 days, look at which search terms actually brought sales, and prune whichever keyword fields did nothing.
- At 90 days, reconsider your categories. If you are nowhere in one of them, it is not working: find a narrower one.
- Whenever the market shifts. Genres develop new vocabulary constantly. The phrase readers use this year may not be the one they used last year.
The authors who treat metadata as a living asset consistently outsell the ones who set it once and never look again: with the same book.
Frequently asked questions
How many keywords does KDP allow?
Seven keyword fields, each holding up to 50 characters. Each field takes a phrase, not a single word, so you have roughly 350 characters of search real estate in total.
Should I repeat my book title in the keywords?
No. Amazon already indexes your title, subtitle, series name, author name and description. Repeating any of those words in a keyword field wastes a slot that could have covered a phrase you are not otherwise ranking for.
How many categories can I choose on KDP?
Three at title setup. Choose narrow, specific subcategories rather than broad ones: ranking visibly in a small category earns a bestseller badge, while sitting invisibly in a huge one earns nothing.
Can I change my keywords after publishing?
Yes, at any time, and you should. Edit the title details in your Bookshelf and republish. Changes typically take effect within 72 hours. Treat keywords as something you iterate on, not something you set once.
What is keyword stuffing on Amazon?
Cramming unrelated or repeated search terms into your title, subtitle or keyword fields to game visibility. It violates Amazon's metadata guidelines and can get a title suppressed or removed. Keyword fields are for phrases a reader would genuinely type, not for lists of every term you can think of.
Keywords are set on the details screen during upload: see the full walkthrough of publishing a book on Amazon. If your metadata has already been flagged, here is what Amazon objects to and why.
Sources
Rates and formulas on this page were checked against Amazon's official Kindle Direct Publishing documentation in September 2026. KDP revises its printing costs and royalty tables periodically, so verify current figures in the KDP Printing Cost & Royalty Calculator before you set a price.