Amazon KDP Keywords: How to Choose the 7 Keyword Slots (2026)

How Amazon decides which books to show

Amazon's book search works in the same two stages as the rest of the catalog. First it filters to books that are textually relevant to the shopper's query, drawing on your title, subtitle, series name, author name, the seven keyword fields, your categories and — to a lesser degree — your description. Then it ranks those relevant books by how likely they are to sell: click-through rate from the search page, conversion on the detail page, sales velocity, reviews and, for Kindle Unlimited titles, pages read.

For a new book you cannot influence the second stage yet. You have no sales history, no reviews and no read-through. What you can control completely, on day one, is the relevance stage — the exact set of queries your book is even eligible to appear for. If a phrase never appears in your title, subtitle, categories or seven keyword slots, Amazon has very little reason to surface your book for it.

This is why keyword research is the single highest-leverage hour a self-published author spends. Ranking on page one of a phrase with 200 searches a month and eight competing books is far more profitable than being invisible on page nine of a phrase with 20,000 searches and 40,000 competing titles.

The seven keyword slots: rules that actually matter

Each of the seven KDP keyword fields accepts up to 50 characters. They are not tags: each slot should hold a complete search phrase a real reader would type, not a single word. "cozy mystery" is a phrase; "mystery" on its own wastes a slot, because that word is already covered by your title, category or another phrase.

Amazon combines words across your slots and your metadata, so you do not need to repeat a word you already used. If slot 1 is "cozy mystery small town" and slot 2 is "cozy mystery cat detective", the repeated words are dead characters. Write slot 2 as "cat detective amateur sleuth" instead and you cover more ground with the same 50 characters.

• 50 characters per slot, 7 slots — treat each as one full search phrase, not a keyword dump.
• Do not repeat words that already appear in your title, subtitle, series name or author name.
• No commas needed; Amazon reads the field as space-separated terms and commas eat characters.
• Never use other authors' names, trademarked series names or 'bestseller', 'free', 'new release' — these violate KDP metadata guidelines and can get a book suppressed.
• No subjective or unverifiable claims ('best book ever'), no misspellings and no ranking terms like 'top 10'.
• Include reader-intent phrases: audience, occasion, tone, setting and problem solved — 'gift for new dad', 'clean romance no spice', 'ADHD planner for adults'.
• Keywords can be edited any time in KDP; re-check them 30 and 90 days after launch.

How to find KDP keywords readers actually type

Amazon itself is the best free keyword tool for books, because its autocomplete is generated from real shopper queries in the store you are selling in. Open Amazon in an incognito window, set the department to Kindle Store or Books, and start typing your genre. Each suggestion is a phrase real readers search often enough for Amazon to remember it.

Work through the alphabet trick: type your seed phrase followed by a, then b, then c, and note the suggestions each letter reveals. Then validate demand and competition by actually searching the phrase: look at how many results come back, whether the top titles literally contain the phrase, and whether the books on page one look professional and well-reviewed. A phrase where the top ten books have fewer than 50 reviews each is a phrase you can realistically win.

• Use Kindle Store autocomplete, not the all-departments search bar.
• Check 'Customers also bought' and 'Also-boughts' on the three closest competitor books for phrasing you missed.
• Read competitor reviews: the words readers use to praise or criticise a book are the words they search with.
• Check the Kindle Store bestseller lists in your target category for recurring subtitle patterns.
• Note the Best Sellers Rank of the #1 and #10 books in a category — that tells you how many daily sales it takes to rank there.
• Repeat the whole exercise per marketplace: amazon.com, amazon.co.uk and amazon.de all have different autocomplete data.

Categories and BISAC codes: the other half of discoverability

When you publish, KDP asks you to choose up to three browse categories (and in the paperback and hardcover flows, a BISAC subject code). BISAC codes are the standard industry subject classification — for example FIC022070 for 'Fiction / Mystery & Detective / Cozy / Culinary'. Amazon uses your BISAC code and your keywords together to place the book in browse categories, which is what powers the orange bestseller badge.

Some Amazon categories are only reachable through a specific keyword combination on top of a category selection. That is why keyword and category planning should happen at the same time, not in two separate sessions: the right phrase in slot 5 can be what unlocks a low-competition category where 12 sales a day makes you #1.

Choose specific over broad every time. 'Fiction / Mystery & Detective / Cozy / Culinary' will make you visible and badge-eligible far faster than 'Fiction / Mystery & Detective', where you compete with tens of thousands of titles including traditionally published bestsellers.

• Pick up to 3 categories, and make at least two of them narrow, low-competition sub-categories.
• Match your BISAC code to the actual content — a mismatch triggers reader returns and poor conversion.
• Check the #1 book's Best Sellers Rank in a category before choosing it: a lower rank number means more sales are needed to reach the top.
• Non-fiction: your category should reflect the reader's problem, not your professional discipline.
• Categories can be changed after launch via KDP support or the category selector — test, then adjust.

Title, subtitle and description: where keywords carry the most weight

Your title and subtitle are indexed more heavily than the seven keyword slots, so the most valuable phrase you found in research belongs there — as long as it reads naturally to a human. Fiction titles have less room to manoeuvre, which is what series names and subtitles are for: 'A Willow Creek Cozy Mystery, Book 1' is both branding and a keyword payload.

Non-fiction has far more freedom, and the subtitle is where the search phrase and the promise live together: 'Sourdough for Beginners: A Step-by-Step Guide to Baking Your First Loaf in 7 Days'. The main title carries the topic phrase, the subtitle carries the audience, the format and the outcome.

Your description is indexed too, but its main job is conversion. Lead with a hook in the first two lines (that is all Amazon shows before 'Read more'), use short paragraphs and light HTML formatting, and place your secondary phrases naturally in the body rather than stacking them at the end. A keyword-stuffed description that stops readers from buying will hurt your ranking through low conversion far more than the extra relevance helps.

Rank Up List vs Publisher Rocket for KDP keywords

Publisher Rocket is the best-known dedicated KDP research tool: a one-time desktop purchase that pulls Amazon search volume estimates, competition scores, category ranks and AMS ad keyword lists. It is a research tool — it hands you data, and you still write the title, subtitle, seven slots and description yourself.

Rank Up List approaches the same problem from the other end. You paste your book's topic, genre, audience and KDP-specific fields (BISAC, series, format), and it writes the finished metadata: a title and subtitle within KDP's limits, seven distinct 50-character keyword phrases with no wasted repetition, category suggestions and a structured description you can paste straight into KDP — in English, Italian, French or German, so you can publish the same book across Amazon marketplaces.

They are not mutually exclusive, and for a serious non-fiction launch using both is reasonable: Rocket for hard volume and competition numbers, Rank Up List to turn that research into compliant, ready-to-paste metadata in seconds instead of an afternoon. If you only want one, choose based on whether your bottleneck is finding the data or writing the listing.

• Publisher Rocket: one-time desktop purchase, Amazon-specific volume and competition estimates, AMS keyword export, no writing.
• Rank Up List: subscription with a free tier, writes title, subtitle, 7 keyword slots and description, multilingual, also covers Etsy, Amazon retail, eBay, Shopify and Redbubble.
• Rocket is English-market centric; Rank Up List produces metadata in EN, IT, FR and DE for international Amazon stores.
• Both replace guessing — the mistake is doing neither and filling the seven slots from memory.

The seven mistakes that waste your keyword slots

• Filling the slots with single words instead of full search phrases.
• Repeating words already present in the title, subtitle or series name.
• Using another author's name or a trademarked series to catch their traffic — a metadata guideline violation.
• Choosing only broad categories where new books are invisible.
• Ignoring reader-intent phrases (gift, audience, occasion, tone) that convert far better than genre labels.
• Setting keywords once at launch and never revisiting them as the market shifts.
• Using the same English metadata on amazon.de and amazon.fr instead of localizing it for those readers.

A practical KDP keyword checklist

• Collect 25–40 candidate phrases from Kindle Store autocomplete across your genre and audience.
• Search each one and record result count, top-10 review counts and whether titles contain the phrase.
• Keep the 10–12 phrases with visible demand and beatable competition.
• Put the strongest phrase in the title or subtitle, the second strongest in the series name or subtitle.
• Write seven 50-character slots with zero repeated words between them and no banned terms.
• Choose three categories, at least two narrow, aligned with a matching BISAC code.
• Write a description with the hook in the first two lines and secondary phrases used naturally.
• Re-check rankings and edit keywords at 30 and 90 days after launch.
• Repeat the phrase research per marketplace before publishing translated editions.