To build strong KDP backend keywords, list the phrases a reader would type, remove anything KDP bans, and sort the rest into seven slots. First, use three slots for targeted long-tail phrases. Next, use two for category-related terms. Finally, use the last two for extra descriptive words. Then search Amazon to check each phrase, and review the slots as your book sells.
Every self-published author gets the same seven boxes for KDP backend keywords. However, most fill them with single words like “mystery” or “romance.” That wastes space, because thousands of other books use the same words.
Long-tail phrases solve that problem for your KDP backend keywords. In short, they are longer, more specific search strings that describe exactly who your book is for. This guide shows you how to turn them into seven well-planned back-end search keyword slots.
First, KDP backend keywords are search terms readers never see. You add them in your book details, and Amazon uses them to help match your book to searches. According to KDP’s official keyword guide, relevant keywords can boost your placement in search results. They work along with other factors, such as sales history and Amazon Best Sellers Rank.
Here is what KDP officially says you can do:
Also note that keywords are only one signal. Kindlepreneur’s 120-book experiment found that a phrase in the title or subtitle ranked about 37% higher on average than the same phrase in a keyword box. That is a third-party test, so treat it as a direction, not a rule. Still, it shows why a clear title and subtitle matter.
This is where advice gets confusing. KDP’s own page says single words can work better than phrases, and specific words work better than general ones. It also warns against quotation marks. Its example explains why: if you enter “complex suspenseful whodunit,” only people who type all those words find you. Without quotes, customers can search any of those words and still find your book.
So what does a long-tail slot really mean? In fact, it means a short string of specific words in natural order, with no quotation marks. Each word still counts. However, the phrase also shows Amazon the exact combination you care about.
Testing supports a mix. In Kindlepreneur’s 120-book experiment, filling more of each box indexed books for more terms. At the same time, books with a specific phrase tended to rank higher for it. The author’s conclusion was to combine both tactics. That experiment used 120 volunteer books, and its author notes the sample is not ideal. Use it as guidance and verify on your own book.
| Source | Claim type | How to treat it |
|---|---|---|
| KDP Help | Official rules and best practices | Follow it. Re-check it, because KDP updates its pages. |
| Kindlepreneur experiment | Third-party test on 120 books | Useful direction, but the sample is small and the author sells keyword software. |
| Your own results | Search results and sales on your book | Most relevant. Change one thing at a time and record it. |
Also, a plan keeps you from filling boxes at random. This layout follows the approach in Kindlepreneur’s 120-book experiment, adjusted to match KDP’s own rules.
| Slots | Job | What goes in | Example type |
|---|---|---|---|
| 1 to 3 | Targeted long-tail phrases | Specific phrases a real reader would type. These carry most of your search intent. | Subgenre + character + setting |
| 4 to 5 | Category anchors | Related terms that match the categories you chose, without repeating the category names. | Setting, tone, plot theme |
| 6 to 7 | Topic-rich fill | Extra descriptive words that use leftover characters and widen what you can be found for. | Reader pain points or story elements |
However, one adjustment matters. KDP says to avoid words already in your category names. So anchor terms should be related ideas, such as setting or tone, not the category name itself. That fits KDP’s list of useful keyword types: setting, character types, character roles, plot themes, and story tone.
First, think like a reader when you plan your KDP backend keywords. Write down every way someone might describe the book they want. Use the five types KDP suggests as a starting grid:
Also, nonfiction authors can swap these for reader problems, desired outcomes, and audience types. For instance, “meal prep for busy nurses” names both a problem and a reader.
Next, gather more candidates for your KDP backend keywords than you need. Aim for 40 to 60 candidates. Use these free methods:
AI cannot see Amazon’s search data, so confirm every phrase on Amazon itself. Human review matters here, as our AI chatbot vs live chat comparison explains for customer support.
Now cut hard. Remove anything on KDP’s avoid list. KDP says its list is not exhaustive, and all keywords must follow its Terms and Conditions. See also KDP’s metadata guidelines.
| Do not use | Why, per KDP |
|---|---|
| Words already in your title or contributor names | KDP says to avoid information covered elsewhere in your metadata. |
| Words already in your category names | KDP lists these as keywords to avoid. |
| Quality claims such as “best novel ever” | Subjective claims are on the avoid list. |
| Time-sensitive words: “new,” “on sale,” “available now” | These go out of date and are on the avoid list. |
| Generic words such as “book” | Information common to most items in the category adds nothing. |
| Spelling errors, or spacing, plural, and capitalization variants | KDP says variants are unnecessary. |
| Other authors’ names or brands you do not own | KDP calls this misrepresentative and has zero tolerance for misleading metadata. |
| Quotation marks, HTML tags, or program names like “Kindle Unlimited” | All are named on the avoid list. |
Next, keep only phrases that truly describe your book. Misleading keywords can hurt readers and your account. After that, drop near-duplicates. For example, keep “village mystery” but drop “village mysteries.”
These three slots hold the clearest long-tail phrases among your KDP backend keywords. Pick phrases that are relevant, searched for, and realistic to rank for. Write each in natural order, because KDP notes readers search “military science fiction,” not “fiction science military.”
Categories and KDP backend keywords work together. Kindlepreneur’s 120-book experiment reports that Amazon may move a book out of a chosen category if the metadata does not support it. That is a third-party observation, so confirm it for your own listing. Whatever the case, supporting terms help.
Choose terms linked to your categories but not copied from them. For a cozy mystery, that might be a setting such as a village or a tone such as gentle suspense. This keeps you aligned with KDP’s advice to avoid category words.
Use the last two slots of your KDP backend keywords to widen your reach. Add descriptive words that fit your book but that you cannot target one by one. For fiction, think about character roles, story catalysts, and flavor of the genre. For nonfiction, use reader pain points, hoped-for results, and audience types.
Also, count your characters as you go. Because slots are limited, unused space is wasted space. However, relevance always comes first, so never pad a slot with off-topic words.
Finally, enter the phrases. Go to your Bookshelf, click the ellipsis next to your book, choose Edit book details, and add them under Keywords. Then click Save and Continue and publish from the pricing page, as KDP instructs.
Afterward, review your KDP backend keywords on a schedule. Search your top phrases on Amazon and note where your book appears. Change one slot at a time, and write down the date. Results take time and depend on sales and many other factors, so avoid judging too quickly.
This example is fictional and illustrative. It shows structure for an imaginary cozy mystery. Check every phrase on Amazon before you use something similar.
| Slot | Illustrative phrase | Characters | Job |
|---|---|---|---|
| 1 | cozy mystery small town bookshop amateur sleuth | 47 | Targeted phrase |
| 2 | witty widow detective whodunit with a rescue cat | 48 | Targeted phrase |
| 3 | feel-good village mystery with tea shop gossip | 46 | Targeted phrase |
| 4 | gentle suspense rural English village setting | 45 | Category anchor |
| 5 | amateur detective friendship found family clues | 47 | Category anchor |
| 6 | light humor no gore low violence clean read | 43 | Topic-rich fill |
| 7 | retired teacher sleuth old secrets coastal town | 47 | Topic-rich fill |
Notice the pattern in these KDP backend keywords. Each slot is specific, natural, and under 50 characters. None repeats a category name, uses a quality claim, or names another author. The numbers are only for illustration.
Map each slot of your KDP backend keywords to subgenre, character, setting, and tone for sharper matches.
Use reader problems and outcomes, such as a specific audience plus a clear goal.
Keep a shared sheet of KDP backend keywords so each book in the series stays consistent but distinct.
Use one checklist and one review schedule for every title you manage.
High Dreams LLC builds websites, e-commerce systems, and AI workflows that help authors and publishers get found. Tell us what slows you down.
High Dreams LLC is a Colorado-based AI and digital growth agency focused on practical AI for business. The team builds AI chatbots, voice agents, workflow automation, websites, apps, and e-commerce systems. Each project starts with the workflow a business actually needs, and human escalation is built in from the start.
KDP says you can use up to seven keywords or short phrases per book. Keep an eye on the character limit in each text field. Many guides cite 50 characters, so confirm the limit shown on your own form.
Use short strings of specific words in natural order. KDP says specific words work better than general ones and warns against quotation marks. Without quotes, readers can match any word in your phrase.
Yes. KDP says you can change keywords and descriptions as often as you like. Open your Bookshelf, choose Edit book details, update the Keywords section, and publish the change.
KDP advises avoiding information covered elsewhere in your metadata. In Kindlepreneur’s test, repeating words neither helped nor hurt. A sensible approach is to avoid repeating the title itself and use slots for new information.
No. KDP lists the name of an author not associated with your book as a keyword to avoid. It says it has zero tolerance for metadata meant to advertise, promote, or mislead.
Not according to the one public test we found. Kindlepreneur reported about 37% higher average rankings when a phrase sat in the title or subtitle. Treat that as a direction. Use a clear title and use the slots for extra terms.
In short, seven slots are enough when each one has a job. Use three for targeted phrases, two to support your categories, and two to widen your reach. Finally, follow KDP’s rules and keep every phrase honest.
Start with one book this week. Build the list, fill the slots, and review in a month. Then repeat the process for your next title.
KDP rules and Amazon search behavior change often. Details reflect pages retrieved on October 9, 2026. The cozy mystery example is fictional. This article is not legal or publishing advice, and results are not guaranteed.