The Seven Keyword Boxes, and What Each One Is For

Amazon gives you seven backend keyword boxes for every title, 50 characters each, and almost no instruction on how they are weighted. The official KDP help page confirms the count, the character limit, and a short list of banned terms, then stops (Amazon KDP Help). Amazon KDP keywords are not a hidden ranking algorithm waiting to be decoded. They are a narrow, documented set of fields, and everything else you have read about amazon kdp keywords, the box order, the supposed hierarchy, the ranking tricks, is inference dressed up as fact.

This piece separates the two. You will see what Amazon actually states, a working method for assigning each box a job, two published worked examples, and a way to test any of it on your own title.

Amazon KDP keywords work by adding search surface beyond what your title and subtitle already cover, letting Amazon’s engine match your book against phrases a reader types that your visible metadata never touches. The boxes function alongside category selection and title metadata rather than replacing them. The gap between what Amazon documents and what actually moves rank is real, so this article draws a clear line between the two throughout. The sections ahead cover Amazon’s own documentation, a box-role method you can test, two published examples with full reasoning, and a protocol for checking any of it against your own listing.

Key Takeaways

• Fifty characters per box: Amazon states each of the seven fields accepts up to 50 characters, which forces phrases rather than word lists.
• Box roles are inference: The conversion, reach, and category-anchor framing below is a working method, and Amazon has published nothing that confirms it.
• Repetition buys nothing: Amazon combines your keywords with words already in your title and metadata, so repeating a title word spends a slot on coverage you already have.
• Keywords raise views, not sales: The one substantial published study found a significant increase in discovery views with no measurable effect on sales.
• Test one box at a time: Record rank for a phrase, change a single box, allow for indexing delay, then recheck.

What Amazon Actually Documents About Amazon KDP Keywords

Amazon’s documented facts about amazon kdp keywords are narrow: seven keyword fields per title, up to 50 characters each, set independently per format in the Book Details step, and editable later without re-uploading the manuscript (Amazon KDP Help). That is the entire confirmed specification.

Its guidance also lists what to leave out: other authors’ names, book titles you do not own, misleading or subjective claims, terms tied to programs such as Kindle Unlimited or KDP Select, and pricing or promotional language. Break these rules and Amazon can reject or suppress the listing rather than simply ignore the term.

According to Amazon’s own documentation (Amazon KDP Help), its search engine combines the words you enter with words already found in your title and other metadata. That single sentence is the basis for the redundancy argument later in this article, and it is one of the few mechanical facts Amazon confirms about how the boxes function.

What Amazon does not publish matters just as much: relative weighting between the boxes, whether box order affects matching, how long re-indexing takes after an edit, or how backend keywords interact with category placement. Say this plainly rather than filling the gap with confident third-party claims.

Amazon documents the shape of the seven keyword boxes and almost nothing about how they are ranked, so any advice about box order or weighting should be treated as a hypothesis you can test. The 50-character limit is worth noticing as a design choice: it fits a natural search phrase of roughly five to seven words and cannot hold a comma-separated tag dump without truncating the part that matters.

A Working Method for Amazon KDP Keywords: Conversion, Reach, and Category Anchors

Label this framing honestly before using it: it is a practitioner method for assigning roles to amazon kdp keywords, not a documented Amazon behavior, as illustrated in the box-by-box reasoning in the Perivane sample report for Agatha Christie’s The Murder at the Vicarage. Amazon has never published anything supporting box-by-box role assignment. The method earns its place because it stops you from writing seven near-identical phrases, not because Amazon reads the boxes in any particular order.

The method itself is simple: assign each box a job before you write anything in it, then judge the phrase against that job rather than against how it sounds on its own. Working this way forces coverage across different levels of buyer intent, instead of seven variations on one popular term you will never rank for.

These seven boxes do not carry different weight as far as anyone can prove. Assigning them different jobs is a discipline for the author, and its value is that it produces seven genuinely different phrases.

There is a real trade-off worth naming. Reach phrases may bring views that never convert, a pattern consistent with the eBOUND finding of more views without measurable sales change. More search surface is not the same thing as more buyers.

Boxes 1 and 2, Conversion Terms

These hold the phrases a reader who already decided to buy something like your book would type.

• Specific intent: Full phrases naming the subgenre plus the defining hook.
• Reader vocabulary: The words a shopper uses, not trade or catalogue vocabulary.
• Narrow beats broad: A phrase you can plausibly rank for beats a head term dominated by bestsellers.

Boxes 3 to 5, Reach Terms

Use these boxes to widen the net into adjacent and mid-tail searches you would not otherwise appear in.

• Adjacent subgenres: Categories a likely reader also browses.
• Occasion and audience: Who is buying and why, where that is genuinely accurate.
• Mid-tail length: Three to six words, specific enough to have limited competition.

Boxes 6 and 7, Category Anchors

Anchor these final two boxes to phrasing associated with one browse category you have deliberately chosen, matched against the classification work covered in our guide to BISAC and Thema codes.

• One target category: Anchor to a single destination rather than hedging across several.
• Category language: Phrasing that mirrors how that shelf is described on Amazon.
• Unverified mechanism: Amazon does not document keyword-to-category effects, so treat anchoring as a test, not a lever.

Why Repeating a Word Across Boxes Wastes a Slot

According to Amazon KDP Help (Amazon KDP Help), its search engine combines the words you enter in the keyword boxes with words it already finds in your title, subtitle, and series metadata. That combination is the cause: every word already present in your visible metadata is already indexed before you type anything. The effect is that repeating a title word in a keyword box adds no new match surface, it only duplicates ground already covered. The result is fewer of your 350 available characters spent on new search terms, which narrows the range of phrases your book can be found under.

Seven boxes at 50 characters is 350 characters of total surface area. Every repeated word reduces the number of distinct terms Amazon has to match your book against, which is the entire value of the boxes in the first place.

An audit is simple to run, as the Perivane report for The Murder at the Vicarage demonstrates: write out the title, subtitle, series name, and author name, list every distinct word across them, then strike those words from your draft keyword phrases and see how many boxes still say something new.

Be honest about the limits of this claim. Because Amazon does not publish its matching rules, nobody can say repetition actively harms ranking. The safer claim is narrower: repetition adds nothing you did not already have, so it carries an opportunity cost rather than a penalty.

Redundancy in the keyword boxes is not tested to hurt. It is simply tested to be unnecessary, because Amazon already indexes your title alongside your keywords. Watch for four common mistakes: single-word tags, comma-separated lists that truncate at 50 characters, quotation marks that may force a narrower literal match, and leaving boxes empty when Amazon has already given you the space.

Two Worked Examples From the Perivane Sample Reports

Two full keyword reports, one fiction and one non-fiction, are published in full at perivane.com. For each, the value is in the stated rationale behind each box, not just the phrase itself.

The fiction example, Agatha Christie’s The Murder at the Vicarage, walks boxes 1 through 7 in order: a conversion phrase built on the classic detective subgenre and village setting, reach phrases moving into adjacent cosy and golden-age territory, and the final two boxes anchored to a single chosen browse category. Each choice comes with a stated reason drawn from how a mystery reader actually searches.

J.C. Ryle’s Expository Thoughts on Matthew, the non-fiction example, follows the same structure but separates a different pair of vocabularies. A reader searching for a commentary uses different words from a reader browsing devotional writing, and the report keeps those in distinct conversion and reach boxes rather than blending them.

The contrast between the two reports is instructive. The fiction report leans on trope and mood language, the non-fiction report leans on function words such as commentary, study, and devotional. Both carefully avoid words the title already carries.

What makes a keyword report useful is the rationale, because the phrase tells you what to type and the rationale tells you what to change when it fails.

How to Test Amazon KDP Keywords Box Roles on Your Own Title

Nothing in this article about the roles of individual amazon kdp keywords boxes is published Amazon behavior. Treat it as a hypothesis, and test it against your own listing before trusting it.

• Baseline first: For each phrase you care about, search Amazon as a customer in a logged-out or private window, and record where your book appears, or that it does not appear at all. Record the date.
• Change one box: Amazon lets you edit keywords from the Book Details page without re-uploading the manuscript, so isolate the variable. Changing three boxes at once tells you nothing about which one moved.
• Allow for indexing delay: Amazon does not publish a re-indexing time. Wait several days at minimum before drawing conclusions; a null result at 24 hours means nothing.
• Recheck the same way: Same phrase, same logged-out conditions, same recording sheet.

Set expectations around what this test can show. It measures appearance and position for a search phrase, which is discovery. It should not be expected to move sales, since the one substantial published study, covering 600 Canadian ebooks over a single quarter using publisher-supplied ONIX keyword fields rather than KDP’s seven boxes, found increased views with no measurable sales effect. Its authors note that converting views to purchases needs further research.

One confound is worth naming directly. Rank moves for reasons unrelated to your edit, so a single test on a single title is weak evidence on its own, and repeating the exercise across titles builds a stronger picture.

Why the Seven Boxes Matter

Few discovery levers are as fully within an author’s control as these seven boxes, changeable at will without waiting on a retailer or an algorithm update. Filling them well is slow, manual work: seven distinct phrases, checked against your own title for overlap, checked against live Amazon search results, per format, per book. Done once and reviewed occasionally, they widen the range of searches your book can answer.

Conclusion

Amazon’s rules governing amazon kdp keywords stay narrow: the KDP Help page documents seven boxes at 50 characters each and prohibits certain terms (Amazon KDP Help). Everything about box roles, order, and weighting is inference, and the conversion, reach, and anchor framing offered here is a working method, presented as exactly that.

Writing seven genuinely different phrases is the core practice that holds regardless: keep them clear of words your title already carries, and test changes one box at a time with a recorded baseline. Pair this work with a clear subtitle and a full pass through your publishing checklist, since keywords work alongside metadata rather than instead of it.

Doing this properly for one title means drafting and rejecting a couple dozen candidate phrases, checking each against live search results, then repeating the exercise per format. It is a few hours of unglamorous work per book, and it is the reason most authors leave boxes half-filled.

The two sample reports discussed above, showing the seven boxes with the full reasoning for each, are published in full at perivane.com.

Frequently Asked Questions

What are Amazon KDP keywords?

Amazon KDP keywords are the seven backend search phrases entered during the Book Details step of publishing on Kindle Direct Publishing. Each of the seven boxes holds up to 50 characters, and Amazon indexes them for search matching, but shoppers never see the text on the book’s product page.

How many keyword boxes does KDP give each title?

KDP provides seven separate keyword boxes per title, each accepting up to 50 characters. Amazon’s own KDP Help page confirms this count and character limit but does not explain how the seven boxes are weighted or ranked relative to one another.

Does the order of the seven keyword boxes matter?

Amazon has never published anything confirming that box order affects ranking or matching. Assigning roles like conversion, reach, and category anchor to specific boxes is a practitioner method for organizing distinct phrases, not a documented Amazon behavior, so treat any claims about order as untested hypotheses.

Should I repeat words from my book title in the keyword boxes?

No. Amazon’s search engine already combines your keywords with words found in your title, subtitle, and series metadata, so repeating a title word in a keyword box adds no new match surface. It wastes part of your 350 total available characters on coverage you already have.

Do KDP keywords actually increase book sales?

Not directly, based on available evidence. A 2019 eBOUND Canada and Ontario Creates study of 600 Canadian ebooks found that added keywords produced a significant increase in discovery views but no measurable effect on sales. That study tested ONIX fields, not KDP’s seven boxes, so results are suggestive rather than conclusive.

How can I test whether a keyword change is working?

Record your book’s rank for a target phrase in a logged-out search, change only one keyword box, wait several days to allow for indexing delay, then recheck under identical conditions. Changing multiple boxes at once makes it impossible to isolate what actually moved your rank.

Part of the guide: Getting Found: Discovery in Three Places at Once

Sources

• Amazon KDP Help – Official documentation on the seven keyword fields: count, character limits and prohibited content
• eBOUND Canada / Ontario Creates – 2019 study: added keywords produced a significant increase in views on Amazon and no effect on sales. Tested publisher-supplied ONIX keyword fields, a different mechanism from KDP keyword boxes, across 600 Canadian ebooks over one quarter
• Perivane – Full sample keyword reports for The Murder at the Vicarage and Expository Thoughts on Matthew