How Do You Choose Amazon KDP Keywords and Categories for an eBook?

Imagine a reader wants a suspense novel about a missing woman in a small coastal town.

They do not know your title.

They do not know your name.

They open Amazon and type:

small town mystery missing woman

Your book could be an excellent match.

Amazon still needs enough accurate information to connect that search with your title.

This is where KDP keyword research and category selection matter.

Keywords help describe specific elements of a book that readers may search for. Categories place the title on broader digital shelves where readers browse by genre or subject.

Current KDP guidance allows authors to enter up to seven keywords or short phrases for a book and select up to three categories based on the book’s primary marketplace and audience. Amazon recommends choosing terms and categories that accurately represent the book instead of trying to manipulate visibility through unrelated metadata.

Good discoverability starts with relevance.

Not with trying to find the easiest shelf to rank on.

Keywords and Categories Perform Different Jobs

Authors often treat the two as interchangeable.

They are not.

Backend Keywords

These are the keyword terms entered in the KDP book-details interface.

Readers do not see them displayed as a keyword list on the product page.

They help Amazon understand additional relevant search language connected to the title.

Visible Book-Copy Keywords

These are natural words appearing in customer-facing material such as:

They should exist because they accurately communicate the book, not because the author is trying to repeat a search phrase unnaturally.

Advertising Keywords

Amazon Ads keywords belong to paid campaigns.

They determine which shopper searches may make an advertisement eligible to appear. Amazon Ads currently distinguishes shopping queries from advertiser-selected keywords and supports keyword targeting independently of KDP’s seven metadata keyword fields.

A keyword that works in advertising does not automatically belong in KDP metadata.

Different systems have different rules.

Start With Reader Language

Do not begin keyword research by asking:

Which words describe my book?

Ask:

What might a reader type when looking for a book like mine?

The difference is subtle but important.

An author may describe a romance as:

A deeply emotional exploration of rediscovered connection.

A reader may search:

second chance small town romance

Both describe the same book.

Only one sounds like realistic search behavior.

Build an initial list around:

Genre
Subgenre
Character type
Setting
Trope
Theme
Reader problem
Skill
Tone
Subject

Amazon itself recommends thinking like a customer and identifies useful keyword types such as setting, character types, character roles, themes, and tone.

Use Amazon Autocomplete as a Research Signal

Open the relevant Amazon marketplace and begin typing a phrase.

Suppose you type:

small town romance

Autocomplete may reveal additional language readers commonly encounter or search around that idea.

Do not automatically copy every suggestion.

Evaluate whether the phrase accurately describes your book.

KDP currently recommends searching Amazon with potential keywords before publication and examining the suggestions shown in the search field. If the resulting books are irrelevant to yours, reconsider the keyword.

Autocomplete is therefore a research tool.

It is not permission to use every popular phrase you see.

Study Comparable Books for Positioning

Look at books serving the same reader.

Study:

You are not copying another author’s metadata.

You are learning the vocabulary of the market.

Suppose ten comparable novels repeatedly describe themselves using:

while your draft metadata says only:

contemporary fiction

Your language may be too broad.

Comparable books help reveal how readers and publishers frame the subgenre.

Broad Keywords Usually Need More Context

Consider:

romance

It describes millions of books.

Compare:

second chance coastal romance

The second phrase provides more information.

Likewise:

Broad: productivity

More specific: productivity for remote managers

Broad: thriller

More specific: psychological thriller family secrets

Specific terms often give Amazon clearer context and can align more closely with reader intent.

Do not make a phrase so obscure that nobody would naturally use it.

The objective is specificity with relevance.

Do Not Waste Fields Repeating Existing Metadata

KDP currently advises authors to avoid information already covered elsewhere in book metadata, including title or contributor details, and also recommends avoiding terms already represented by selected categories.

Suppose the book title is:

The Remote Manager’s Productivity System

Repeating:

remote manager productivity system

in several keyword fields may add little new information.

Additional terms could potentially describe:

Use available fields to expand Amazon’s understanding of the book.

Not to repeat the same phrase seven times.

KDP Keyword Examples by Genre

The examples below illustrate research logic, not guaranteed ranking terms.

Romance

Imagine a contemporary romance featuring two divorced former partners returning to a lakeside hometown.

Possible research directions:

A writer using romance tropes for writers during manuscript development may already have accurate trope language available.

Use only tropes the book actually delivers.

Do not label a romance “enemies to lovers” when the characters have one disagreement and are dating by Chapter Two.

Thriller

Imagine a psychological thriller about a journalist investigating her sister’s disappearance.

Research directions might include:

missing sister thriller
family secrets suspense
investigative journalist mystery
small town psychological thriller
cold case mystery
female protagonist suspense

A fiction character bible can help confirm which character roles, settings, and conflicts are genuinely central enough to use in metadata.

Avoid keywording a minor element simply because it sounds commercially attractive.

Self-Help

Suppose the book helps professionals rebuild confidence after losing a job.

Potential phrases:

Start with the reader’s problem.

Readers are more likely to search for the challenge they are experiencing than for the author’s internal framework name.

Business

A management eBook may explore:

first time manager
remote team leadership
employee accountability
difficult workplace conversations
manager communication
team expectations

Specific problem-based language often communicates more than:

business success

or:

leadership book

Work completed during nonfiction ebook fact checking should ensure the visible description accurately supports any professional or research claims associated with the positioning.

Poetry

A collection may be described through:

The keyword strategy should match the actual collection after poetry ebook formatting and editorial sequencing are complete.

Do not position a mixed collection as visual poetry simply because two poems use unusual spatial formatting.

Educational eBooks

Educational search behavior can be highly problem-oriented.

Examples:

Age and learning level matter.

KDP’s current audience settings also affect Children’s and Teen & Young Adult category eligibility. Amazon says appropriate reading-age information may be necessary for placement in age-specific categories.

Choose Categories for Relevance First

Authors sometimes hear advice such as:

Find the smallest category possible so you can become a bestseller.

This is risky when the category does not genuinely describe the book.

KDP explicitly warns against inaccurate category selections and says it does not tolerate category choices designed to mislead customers. Amazon may change category placement when a category does not match the book’s content.

Choose the three strongest available matches.

For a historical romance, relevant choices might center on:

Forcing the same title into an unrelated low-competition subject undermines the shopping experience.

Niche Categories Can Be Useful When They Are Accurate

Niche does not mean irrelevant.

Suppose you publish a nonfiction guide specifically about managing distributed software-development teams.

A highly relevant management or remote-work subcategory could be more useful than the broadest possible business category.

The decision should ask:

Would a reader browsing this category reasonably expect to find my book here?

If yes, the category may be worth considering.

If not, do not use it simply because fewer competing books appear there.

Categories Differ by Marketplace

A category available on Amazon.com may not exist in exactly the same form elsewhere.

KDP currently states that categories vary by marketplace and format, and Amazon updates its bookstore category structure over time. Authors should focus on their selected primary marketplace when choosing categories.

For a U.S.-focused launch, research the U.S. Kindle Store rather than copying a category path from another country’s Amazon storefront.

Backend Keywords Should Not Contain Promotional Claims

KDP’s current keyword guidance lists several types of terms authors should avoid, including:

Subjective quality claims
Time-sensitive promotional language
Generic words common to nearly every item
Misleading author names
Unauthorized brands
Amazon program names
HTML tags

Examples to avoid include:

best novel ever

new release

available now

Kindle Unlimited

Do not use another author’s name in KDP metadata to make your book appear associated with them when they are not a contributor.

Amazon identifies misleading metadata as a customer-experience violation.

Do Not Create Multiple Variations of the Same Word

Using fields for:

mystery

mysteries

Mystery

mystery book

is inefficient.

KDP advises against unnecessary variants based only on spacing, punctuation, capitalization, or pluralization.

Use the space to describe another useful dimension of the title.

For example:

Each phrase adds new context.

Keywords, Categories, and eBook Descriptions Should Agree

Suppose your backend keywords suggest:

cozy mystery

Your categories say:

Psychological Thrillers

Your ebook description writing emphasizes graphic serial-killer suspense.

Something is wrong.

The book’s:

should collectively communicate a coherent reading experience.

Editorial development comes first.

Developmental editing for ebooks can change a manuscript’s central positioning substantially.

Beta readers for ebooks may also reveal that readers perceive the genre or audience differently from the author.

Finalize the book’s identity before locking in metadata.

Accessibility Is Not a Search Keyword Strategy

Accessible ebook formatting matters because readers should be able to use the book with screen readers and other assistive technology.

Do not treat accessibility terminology as a discoverability trick unless it accurately describes something a relevant reader needs to know.

Accessibility belongs in the publishing process first.

Metadata should communicate it only where appropriate and truthful.

Lead-Magnet Keywords Serve a Different System

An ebook lead magnet may target search language on:

Those terms do not automatically belong in Amazon backend metadata because the commercial goal is different.

The lead magnet attracts a business prospect.

The Amazon eBook metadata helps shoppers discover a retail product.

Keep the two research processes connected but distinct.

Advertising Keywords Are Not Backend Keywords

This distinction prevents a common mistake.

A backend KDP keyword helps describe the published title to Amazon’s store systems.

An Amazon Ads keyword is used in paid targeting.

Sponsored Products currently supports automatic and manual targeting. Under manual keyword targeting, advertisers can use keyword match types and optimize campaigns according to actual search performance.

Advertising can therefore test a much broader pool of search ideas.

Do not attempt to cram an advertising keyword list into KDP’s seven metadata fields.

Use backend fields for the most meaningful descriptive concepts.

Use advertising systems for campaign targeting and experimentation.

Track Changes Without Assuming Causation

KDP allows authors to update keywords, and category changes can also be submitted after publication. Amazon notes that new category placements can take up to 72 hours to display.

Keep a record of:

Keyword set
Categories
Date changed
Sales
Page reads where relevant
Search visibility observations
Advertising activity
Promotions

Change only a few variables at a time when possible.

A ranking increase after changing one keyword does not prove that keyword caused the increase.

Amazon notes that search placement can also be affected by factors such as sales history and Amazon Best Sellers Rank.

Look for sustained patterns.

Not overnight conclusions.

Practical KDP Keyword Research Workflow

Step 1: Define the Reader

Write down:

Step 2: Generate 30 to 50 Candidate Terms

Think like a reader.

Do not filter too early.

Step 3: Check Amazon Autocomplete

Look for natural search language.

Step 4: Review Comparable Books

Identify recurring terminology and subgenre positioning.

Step 5: Check Relevance

Search each promising phrase.

Ask:

Do the results resemble my book?

Step 6: Remove Redundancy

Delete terms already strongly represented by:

Step 7: Remove Prohibited or Misleading Terms

No unrelated author names, false quality claims, temporary promotions, or unauthorized brands.

Step 8: Select Up to Seven Strong Terms or Phrases

Favor distinct dimensions of the book.

Step 9: Choose Three Accurate Categories

Select the strongest available shelves in the primary marketplace.

Step 10: Record the Starting Metadata

Save the exact keyword and category choices so later tests can be compared meaningfully.

The purpose of KDP keyword research is not to trick Amazon into displaying the book everywhere.

It is to help Amazon understand where the book belongs so the right reader has a better chance of finding it.

Frequently Asked Questions

How Many Keywords Can Authors Add in Amazon KDP?

KDP currently allows authors to enter up to seven keywords or short phrases in the book-details section.

How Many Categories Can an eBook Have in KDP?

Authors can currently select up to three categories during KDP title setup. Available choices depend on the book’s primary marketplace, audience, and format.

Should Authors Use Amazon Autocomplete for Keyword Research?

Yes. KDP itself recommends searching potential terms on Amazon and reviewing autocomplete suggestions. Authors should still use only phrases that accurately describe their books.

Are Amazon Ads Keywords the Same as KDP Backend Keywords?

No. KDP backend keywords are book metadata, while advertising keywords are used for paid campaign targeting. The same research may inspire both, but they operate in separate systems.

Should Authors Choose Irrelevant Categories With Less Competition?

No. Amazon requires categories to accurately represent the book and warns against category choices intended to mislead or manipulate shoppers.

Can Authors Change KDP Keywords and Categories After Publication?

Yes. KDP allows authors to update keywords and submit category changes after publication. Authors should track changes carefully because sales, advertising, promotions, and other factors can also influence visibility.

Author

Benjamin Lucas

Benjamin Lucas is a professional content strategist and publishing writer with expertise in digital content creation, eBook development, and online branding. He regularly writes about publishing trends, business communication, technology, and content marketing, helping readers understand the evolving digital landscape. Through his work with Hire Ebook Writer Now , Benjamin focuses on creating engaging, informative, and professionally crafted content that supports both businesses and aspiring authors.