Publishing field guide
Use comparable books, Amazon suggestions and problem language to choose phrases that describe the book naturally.
Reviewed by Stefan Mitrović, Founder of Automateed · Updated July 16, 2026
60-second summary
Kindle keyword research means finding the exact phrases buyers type into Amazon and choosing seven that truthfully describe your book. The method: harvest candidates from Amazon’s autocomplete, competitor titles and reader reviews; compare demand with a research tool; then fill the seven slots with distinct, natural phrases — no repetition, no stuffing, no misleading terms. Metadata describes; it does not bait.
Real product steps
The workflow combines the store’s own signals with the platform’s research tools, ending in seven slots you can defend.
Workflow map
01
Harvest from the search box
Type your topic into Amazon’s search slowly and record every autocomplete suggestion — these are real queries, ranked by Amazon itself.
02
Mine competitor titles and reviews
The top ten books in your niche encode working keywords in their titles and subtitles; their reviews contain the problem phrased in buyer language. Collect both.
03
Compare candidates in the research tool
Run the list through the KDP keyword research tool to weigh relative demand and spot phrases with interest you had not considered.
04
Let AI draft, then diversify
The KDP metadata assistant proposes seven keywords from the manuscript. Edit for coverage: seven distinct angles (audience, problem, format, occasion) beat seven synonyms.
05
Verify by searching
Search each final phrase on Amazon and look at the results honestly: would your book belong on that page? If not, the keyword is borrowed traffic that will bounce.
Type your topic into Amazon’s search slowly and record every autocomplete suggestion — these are real queries, ranked by Amazon itself.
The top ten books in your niche encode working keywords in their titles and subtitles; their reviews contain the problem phrased in buyer language. Collect both.
Run the list through the KDP keyword research tool to weigh relative demand and spot phrases with interest you had not considered.
The KDP metadata assistant proposes seven keywords from the manuscript. Edit for coverage: seven distinct angles (audience, problem, format, occasion) beat seven synonyms.
Search each final phrase on Amazon and look at the results honestly: would your book belong on that page? If not, the keyword is borrowed traffic that will bounce.
Every step above describes the current Automateed interface — open a free preview and follow along with your own project.
Create a free previewThe full guide
Your searchable surface on Amazon is bigger than the seven slots: title, subtitle, author name and the slots all index. The practical division: put your one primary phrase naturally in the subtitle, use the slots for the six-to-seven next angles, and never spend slots repeating words already in the title — Amazon indexes those anyway, so repetition wastes inventory.
New books cannot win “romance” or “productivity” — but “enemies to lovers hockey romance” and “productivity for ADHD adults” are winnable shelves where intent is precise and conversion is high. Long-tail phrases have smaller volume and dramatically better fit; a first book’s keyword set should lean almost entirely long-tail, graduating to broader terms as reviews and rank accumulate.
Amazon’s metadata guidelines prohibit misleading terms: other authors’ names, competitor titles, “free” or “bestseller” claims, and anything the book is not. Beyond policy, misleading keywords self-destruct — wrong-intent clicks convert poorly and depress the rank the trick was meant to inflate. The compliant habit is also the effective one: describe the book in the customer’s words.
Decisions that change the result
Before you open any tool, write 8–12 buyer questions that match how people shop for your genre. Make them specific enough that they could be typed into a search box: “what to read if…”, “books for people who…”, “workbook for…”, “guide to…”, “stories with…”. This matters because the seven metadata slots work best when each phrase captures a different slice of intent, not the same idea in different wording.
Turn those questions into candidate phrases by removing words that don’t change what the customer wants. For example, “books that help me learn to meditate” can become “meditation basics for beginners” or “meditation workbook for beginners” depending on what your manuscript actually offers. The goal is to describe the experience your book delivers, not to force-fit a market label you like.
When you type your topic into Amazon, don’t use “perfect” terms first. Use the messy, partial phrase a buyer would start with. Try variations like “how to…”, “best… for…”, “beginner…”, “guide…”, and “worksheets…”. Record the autocomplete suggestions exactly, including the parts you would normally remove, because buyers search for those chunks.
After you collect candidates, sort them into three piles: (1) clear fit with your book’s premise, (2) maybe fit depending on the angle, and (3) mismatch. This prevents you from accidentally choosing phrases that describe a different promise (for instance, a “starter guide” vs. a “deep dive,” or a “healthy eating plan” vs. a “meal prep plan”). That mismatch is one of the most common reasons a keyword set underperforms even when the phrase looks popular.
Comparable titles are useful because they reveal how successful authors compress reader intent into short strings: a genre label plus an audience plus a payoff. You’re not copying their exact wording; you’re translating their pattern into phrases that truthfully describe your own manuscript.
Look at both the title/subtitle and the review language. Title/subtitle shows what the market already recognizes, while reviews show what the reader struggles with and how they describe the outcome. When you translate review language, keep your phrasing natural and consistent with what’s actually in your book chapters, worksheets, or story setup.
Worked example
You published a 5-week “language learning” workbook for absolute beginners learning conversational phrases for travel. The manuscript includes guided practice pages, role-play prompts, and an answer key. It is not a full grammar textbook and not a mobile app.
From Amazon autocomplete and competitor book listings, you collect candidates such as: “beginner travel language workbook”, “conversation practice for beginners”, “phrasebook for travelers workbook”, “learn to speak for beginners”, “daily speaking practice workbook”, and “absolute beginner conversation guide”. You also find some nearby-but-wrong suggestions like “advanced fluency” and “grammar rules textbook” which don’t match your content.
You pick seven that cover distinct buyer angles while staying truthful: 1) “travel language workbook for beginners” (audience + format), 2) “conversation practice for absolute beginners” (problem + level), 3) “phrase practice for travelers” (use case/occasion), 4) “learn to speak guide for beginners” (promise/outcome), 5) “daily speaking practice workbook” (routine/structure), 6) “role-play prompts language practice” (what’s inside), 7) “beginner speaking worksheets” (format detail). You avoid repeating “beginner” too many times by swapping one slot to “absolute beginners” and another to “travelers” so each phrase still reads naturally.
You decide your subtitle will include “travel language workbook for beginners” in a readable way. Then you use the remaining six slots for other angles. This prevents you from wasting slots on the exact same phrase that already appears in the subtitle.
For each final phrase, you search Amazon and check whether the first page results look like workbook-style books with conversation practice for beginners, not advanced materials or grammar textbooks. If you find that “learn to speak guide for beginners” returns mostly grammar-focused books, you replace that slot with something that better matches the workbook and practice pages, such as “beginner speaking worksheets” (keeping the total at seven).
A strong set doesn’t require maximum breadth. It requires seven non-overlapping ways to describe your book so the right reader finds it, and the search results page looks like a place your book genuinely belongs.
Avoidable mistakes
A broad phrase may have lots of searches, but if it implies a different experience than your manuscript provides, your clicks will likely come from uninterested readers. The quick fix is not “more keywords,” it’s changing the phrasing so it matches your actual deliverable (workbook pages, prompts, practice routines, or story tropes).
Beginners often choose multiple phrases that all boil down to the same thing (example: “beginner guide,” “beginner book,” “easy guide”). Amazon indexes components of phrases, so this usually buys repetition instead of coverage. Better: allocate slots to different angles like audience, format, occasion, and the kind of practice or story promise inside the book.
Some metadata entries are prohibited or considered misleading when they point to a party you didn’t create. Even when you’re tempted by a recognizable label, the safe path is to use intent phrases that describe your book’s subject, audience, and content style rather than other publishers or authors.
Tools can rank and suggest, but they can’t fully reflect how Amazon’s live results currently group books. If a phrase consistently surfaces a different type of product, replacing that slot is usually more effective than leaving it and hoping. Verification is what turns “a plausible phrase” into “a match for this marketplace page.”
Keep the workflow moving
Quality gate
Run these checks against the actual manuscript, files and reader journey before publishing.
Keywords match the book
No prohibited or misleading terms
Title stays readable
Results are measured after publication
Continue the exact workflow
Editorial note
This page is a practical workflow, not a promise of sales, ranking, publishing approval or a specific reader outcome. Platform rules and professional requirements should be checked at the point of use.
Questions specific to Kindle Keyword Research for Beginners
Seven slots, each accepting a phrase. Automateed’s metadata assistant drafts all seven from the manuscript for you to edit.
Phrases — buyers search in phrases, and multi-word slots index for their components too. “Meal prep for beginners” outworks “cooking.”
No — names and brands you do not own are prohibited metadata. Compete for their readers with accurate genre and trope phrases instead.
Yes — title and subtitle are the strongest indexed fields. Put the primary phrase in the subtitle naturally and save slots for other angles.
Review after 60–90 days of data. Change one or two slots at a time so results are attributable; constant churn teaches nothing.
Cramming repeated or irrelevant terms — policy-restricted and commercially useless, since wrong-intent clicks bounce and hurt conversion metrics.
Reviews of comparable books. Readers describe the problem and the payoff in exactly the words other buyers will search.
They confirm real queries, ranked roughly by popularity — the best free demand signal. The research tool adds comparative weight between candidates.
Each edition has its own metadata; keep keyword sets consistent unless the editions serve genuinely different searches.
Seven synonyms of one broad term. Spend the slots on seven different angles — audience, problem, subgenre, occasion, format — that all remain true.
Use your manuscript’s strongest differentiator to break the tie. If both phrases say “beginner conversation practice,” choose the one that highlights what makes your book distinct (for example, “worksheets,” “answer key,” or “role-play prompts”). Then re-check by searching the phrase: pick the one where the results look most like your book’s format, not just its topic.
Usually you don’t need both. Choose one form that reads best in the slot and keeps the set diverse. Since Amazon indexes phrase components, you’ll often capture the related searches without dedicating another slot to a tiny wording change. Your seven entries should each add a different intent angle.
Explore next
Learn how to move from brief to outline, chapter drafts and human revision without accepting generic output.
Open guideUnderstand reflowable and fixed layouts, what Kindle accepts and why one master file rarely serves every channel perfectly.
Open guideLearn trim sizes, margins, bleed, fonts, front matter, EPUB checks and the difference between digital and print interiors.
Open guideUse your own topic
Review the outline, visual direction and available chapters before deciding whether to continue the full project.