Amazon Keyword Research for Authors: How to Pick the Seven Strings That Matter
By Abdul Moeed
Amazon gives you seven keyword strings when you upload a book to KDP, and this guide from Nova Publishers shows how to research them, pick strings that drive real discovery and choose your categories. Each string can be up to 50 characters. That is your entire on platform SEO budget. Most indie authors fill these in during a tired evening at the end of a long production cycle. The result is books that never get found by the readers most likely to love them.
What Amazon's keyword field actually does
Two things, both important. First, the strings help Amazon match your book to the searches and niche subcategories where readers look for it. Second, the strings shape what Amazon's recommendation algorithm associates your book with, which determines who sees it in "customers also bought" panels and in keyword search results.
Strings do not literally have to be searched. They have to convey what your book is about, what genre, what audience, what reader emotion, and what comparable titles. The algorithm reads them as a thematic signal more than as a search query.
Step one: study the comp shelf
Find five books that are doing well in your exact subgenre. Open each on Amazon. Scroll to the "Look for specific items by" or "Print length / Language / Publisher" section, then study what categories Amazon has slotted that book into. Look at the bestseller rank and which subcategories the book appears in. This tells you what categories your book could plausibly compete in.
Step two: find the search terms readers actually type
Three tools cover most of the research work. Publisher Rocket is the standard paid tool, designed specifically for KDP keyword research and Amazon ad targeting. KDSpy is a lighter Chrome extension that surfaces keyword competition data on any Amazon search results page. Amazon's own search bar autocomplete is free and underrated. Type the start of a phrase a reader might use and read the suggestions Amazon offers. Those suggestions are real reader queries.
Build a list of 50 to 100 candidate keywords. Examples for a cozy mystery: "cozy mystery series with cat", "small town murder mystery", "female detective amateur sleuth", "cozy mystery box set", "library mystery clean", "book club mystery 2026".
Step three: pick seven strings that combine reach and specificity
Each of your seven Amazon keyword strings should pack two or three concepts that go together. Not one keyword per string. Three.
Examples for the cozy mystery:
- String 1: "cozy mystery series with cat amateur sleuth"
- String 2: "small town murder mystery clean"
- String 3: "female detective librarian sleuth"
- String 4: "book club cozy mystery small town library"
- String 5: "witty mystery series first book"
- String 6: "golden age style cozy puzzle mystery"
- String 7: "clean mystery no graphic violence"
Each string covers a different axis: the trope (cat sleuth), the setting (small town), the protagonist (librarian), the reader scenario (book club), the tone (witty), the style (golden-age puzzle), the content rating (clean). Together they tell Amazon's algorithm what kind of book this is and who should see it.
Step four: pick categories with care
KDP lets you choose up to three Amazon categories for each format when you set up the book. Since 2023 you pick them directly in KDP; the old advice to email support for ten extra categories no longer applies. Keep KDP's keyword rules in mind too: no other authors' names, and no time-sensitive words like "new" or "on sale".
Pick categories where your book genuinely belongs (do not pick Literary Fiction for a thriller because it looks less crowded; readers who land there won't buy it, and the book will sink). Within that, favour one or two smaller subcategories where the top books aren't selling thousands a day, so a new release has room to be seen.
Step five: re research every six months
Reader search behavior moves. Genres heat up and cool. New tropes go viral on BookTok. The seven strings that worked in spring 2026 may not be the right seven in fall 2026. Set a calendar reminder. Re research. Update through KDP.
Bonus: Amazon ad keywords
Amazon ad keyword targeting uses a different keyword field, the one inside the Amazon Advertising console. Targets here can be book ASINs (your direct competitors) or search terms. Start with the comp ASINs you found in step one. Add the search terms from step two that have decent volume but moderate competition. Pause anything that does not generate at least one click per dollar within 30 days.
The mistake that costs the most
Most indie authors leave at least three of the seven strings as single keywords ("thriller", "romance", "science fiction"). One word strings are wasted real estate. Use all 50 characters of every string. Pack two or three concepts. Treat each string like a tiny ad headline that has to convey what makes this book different.
Publishing with Nova Publishers? Keywords, categories and the rest of your metadata are set up in your own retailer accounts as part of every package. See package details.