How many categories and keywords does KDP give me?

Up to three categories, chosen from Amazon's browse tree, and seven keyword fields of up to 50 characters each. The category and keyword forms appear for each format — an ebook and its paperback are categorised separately, in trees that are not identical.

What are the rules for Amazon keywords?

Amazon prohibits the manipulative uses outright: unrelated or misleading terms, other authors' names, other authors' books, sales-rank claims such as 'bestselling', promotional terms such as 'free', and brands you do not own or are not authorised to use. Its guidance then asks you to avoid duplication and time-sensitive words — not because a reviewer objects to them, but because repeated or stale terms do no search work. Enforcement runs on automation with human reviewers behind it, and a removal comes with a notice and an appeal route.

Do keywords put my book in categories?

Yes, and Amazon says so directly. Your selected categories together with your keywords help Amazon decide how to shelve the book, and its own advice for a shelf the picker does not offer is to signal it with keywords. Some subcategories go further and require designated search keywords before a book can sit there at all.

Should I repeat the genre in every keyword field?

No. The categories already carry the genre; the keyword fields should say what the categories cannot — the tropes, the settings, the character types, the audience — in the words readers actually search. Seven copies of 'fantasy' is one category's worth of information spent seven times.

Amazon categories and keywords

Every book on Amazon sits on shelves, and two of the levers over those shelves are set by hand: the category picks and the keyword fields. They are not the only levers — the title is Amazon’s most-used search attribute, and sales history, availability and listing age all weigh in — but they are the only ones you can rewrite at will, after publication, for free. What follows is what Amazon documents about them, plus how the parts behave in practice.

What a category is

A category is a browsing shelf — one node in Amazon’s vast browse tree, from Books → Literature & Fiction → Genre Fiction → Historical on down.

Two consequences make categories worth more effort than a click:

The ranks are relative, and the competition differs. A book can be ranked on up to three Best Seller category lists, and Amazon compares each book’s customer activity with the other books on the shelf — recent and historical activity both, weighted toward the recent. The same level of sales can rank high on a quiet, precise shelf and vanish on a broad one. What differs between shelves is the competition, not a published threshold.

The shelf is where browsers live. Search captures readers who know what they want; categories mostly capture readers who are browsing. It is not the only discovery path — recommendations and Amazon’s own merchandising run alongside it — but it is the one you can aim.

You may pick up to three categories per format. Amazon can and does adjust placement, but the picks are respected more often than forum lore claims — the lore exists because the browsing tree is deeper than the picker’s list, which brings us to the keywords.

What a keyword is not

A keyword field is not a tag, a hashtag, or a field the reader sees. There are seven fields, each up to 50 characters, and they hold search terms: language that describes the book to Amazon’s systems in ways the title, subtitle and description do not already carry. Amazon’s own advice on what belongs here is a good checklist — settings, character types, character roles, plot themes, story tone.

Three behaviours follow from what the field is:

Repetition is waste. The categories already state the genre. “Fantasy novel”, “epic fantasy book” and “fantasy series” spend three fields saying what the shelf already says. Amazon’s guidance is explicit that keywords should carry information not already represented in the other metadata — and while the indexing of the description is not documented in detail, the direction is clear: the fields exist to add unclaimed terms.

Reader language beats industry language. Readers search in their own vocabulary — the tropes, the setting, the audience. “Enemies to lovers”, “cozy mystery cats”, “slow burn romantasy”, “regency romance widower” — the phrases readers type, not the words the trade uses to shelve.

Seven fields, seven angles. Amazon does not document how terms inside a field combine, and it warns against quotation marks — words in an unquoted phrase can match searches individually. So spend the fields on different reader intents — who the book is for, what happens in it, where it is set, what mood it has — rather than engineering exact matches inside one.

The rules, and the practices

Amazon’s keyword rules come in two strengths, and an article titled “the rules” should not blur them.

Prohibited outright — the manipulative uses that get metadata removed: unrelated or misleading terms, references to other authors or other authors’ books, sales-rank claims (“bestselling”), promotional language (“free”, “sale”), HTML, and brands you do not own or are not authorised to use.

Advised against — the terms that merely do no work: words already carried by the title or the categories, time-sensitive language (“new”, “available now”), generic one-word terms, and spelling variants or misspellings.

Enforcement is automated, with dedicated human reviewers behind it, and a removal comes with a notice and an appeal route. The safe posture is the same under either rule: describe the book, in reader language, with nothing in the fields you would not say to a librarian.

How the two levers connect

This is the part most self-publishing lore gets wrong, in both directions: the mechanism is neither secret nor undocumented. Amazon states plainly that your chosen categories, together with your keywords, help it decide how to shelve the book — and its own advice, when the shelf you want is not in the picker, is to signal that shelf with keywords.

For some subcategories the keyword is not a signal but a requirement. Parts of Amazon’s children’s category and its LGBT categories are gated on designated search keywords, listed in KDP’s own help pages; a book that does not carry them is not considered for those shelves at all.

The practical recipe:

Pick the three shelves the picker offers that genuinely fit the book. Relevance first: irrelevant categorisation is the one thing Amazon says it will override. Among shelves that legitimately fit, favour precision — a book can rank where the competition is thinner.
Spend one or two keyword fields on shelf signals for a more precise category that the picker does not list.
Spend the rest on reader language — settings, character types, roles, themes, tone — the additive search terms.
Revisit after launch. Categories are editable — changes generally take effect within about 72 hours, with rank recalculations lagging behind. Move a book because another shelf describes its readers better, not because a chart looks easier. Metadata is a living surface; updating a published book covers the mechanics.

What this is worth

A sober paragraph, because this topic attracts the most inflated claims in self-publishing.

Categories and keywords do not sell a book that has not earned it: they place the book where its real performance can compound — the chart flags, the also-boughts, the browsing traffic. A strong book in weak metadata sells a fraction of what it should; weak metadata cannot make a weak book sell. The levers are real, and they are multiplier levers, not engine levers.

They are also, uniquely, free to fix. Cover, prose and price all cost something to revisit; the keyword fields can be changed as often as you like.

The category and keyword forms reward one habit: describing the book, everywhere, in the language of its readers, and never the same way twice. It is the cheapest optimisation in publishing and the most reversible.

Nueel keeps the metadata you need at the upload — the description, the series information, and suggested categories and keywords drawn from what the book actually contains — alongside the book, so the same answers are ready to enter consistently when you publish.

Sources

Common questions

How many categories and keywords does KDP give me?: Up to three categories, chosen from Amazon's browse tree, and seven keyword fields of up to 50 characters each. The category and keyword forms appear for each format — an ebook and its paperback are categorised separately, in trees that are not identical.
What are the rules for Amazon keywords?: Amazon prohibits the manipulative uses outright: unrelated or misleading terms, other authors' names, other authors' books, sales-rank claims such as 'bestselling', promotional terms such as 'free', and brands you do not own or are not authorised to use. Its guidance then asks you to avoid duplication and time-sensitive words — not because a reviewer objects to them, but because repeated or stale terms do no search work. Enforcement runs on automation with human reviewers behind it, and a removal comes with a notice and an appeal route.
Do keywords put my book in categories?: Yes, and Amazon says so directly. Your selected categories together with your keywords help Amazon decide how to shelve the book, and its own advice for a shelf the picker does not offer is to signal it with keywords. Some subcategories go further and require designated search keywords before a book can sit there at all.
Should I repeat the genre in every keyword field?: No. The categories already carry the genre; the keyword fields should say what the categories cannot — the tropes, the settings, the character types, the audience — in the words readers actually search. Seven copies of 'fantasy' is one category's worth of information spent seven times.