← All posts

KDP keyword research guide 2026

KDP Keyword Research Guide 2026: Rank & Sell More Books

9 June 2026

Most KDP authors spend hours on their manuscript and twenty minutes on keywords. This KDP keyword research guide 2026 fixes that — with a systematic process grounded in how Amazon's A9 algorithm actually ranks books, not how most tutorials describe it.

If you've published on Kindle Direct Publishing and your book isn't getting organic traffic, the problem is almost always keyword selection. Not your cover. Not your description. Keywords — specifically, the wrong ones in the wrong fields.

Key Takeaways

  • Amazon's 7 backend keyword fields are not interchangeable — how you use them determines whether A9 indexes your book for discovery or ignores it entirely.
  • High-volume keywords actively hurt new books. A title with fewer than 50 reviews cannot compete for "self-help books" — but it can rank for "morning routine journal for anxious overthinkers."
  • Discovery keywords and conversion keywords serve different jobs. Conflating them is the single most common KDP keyword mistake.
  • Amazon Autocomplete is a real-time demand signal. The A-to-Z method extracts keyword variations most authors never find.
  • Your seed list should come from reader language, not author language. The reader searches "how to stop worrying at night" — not "anxiety self-help nonfiction."

Why KDP keyword research is different in 2026 (and why most authors get it wrong)

Amazon is not Google. That distinction matters more than most KDP tutorials acknowledge. Google ranks pages based on backlinks, domain authority, and content depth. Amazon's A9 algorithm ranks books based on sales velocity, conversion rate, and keyword relevance — in roughly that order. The implication: a book that converts well on a low-competition keyword will outrank a better-reviewed book targeting a saturated one.

Most keyword advice written for KDP authors is borrowed from general SEO practice. It tells you to find high-volume keywords and optimise for them. That works on Google, where a new page can slowly accumulate authority. It fails on Amazon, where a new book with no sales history and no reviews gets buried behind titles with 300+ reviews the moment it targets a competitive term.

How Amazon's A9 algorithm uses your 7 keyword fields

KDP gives you seven backend keyword fields, each accepting up to 50 characters. Amazon's indexing system reads all seven. It also reads your title, subtitle, series name, and author name for keyword signals — but the backend fields are the primary lever you control post-publication without triggering a full re-review.

Three things to understand about how A9 processes these fields:

  • Order within a field matters less than presence. "journal anxiety morning routine" and "morning routine anxiety journal" will index for the same terms. Don't waste characters on word order — use them for additional keywords instead.
  • Repetition wastes space. If "anxiety journal" appears in your title, putting it in a backend field adds no indexing benefit. Amazon already has it. Use that field for a different keyword cluster.
  • Phrases don't need to be intact. Amazon indexes individual words and combinations across fields. You don't need to write "morning routine journal for anxious overthinkers" as a single phrase — the component words will combine in the index.

The practical result: your 7 fields × 50 characters = 350 characters of indexable real estate. Used correctly, that covers 15–25 distinct keyword concepts. Used incorrectly (repeating title keywords, writing full phrases instead of unique terms), you might cover 4–6.

The difference between discovery keywords and conversion keywords

Discovery keywords get your book in front of readers. Conversion keywords make those readers click and buy. They are not the same list, and treating them as one is where most KDP keyword strategy breaks down.

A discovery keyword is broad enough to generate impressions. "Anxiety journal" generates impressions. But a reader searching "anxiety journal" sees hundreds of results. Your conversion rate on that term — if you rank at all — will be low unless your cover, title, and reviews are already strong.

A conversion keyword is specific enough that the reader who types it is already close to buying. "Anxiety journal with prompts for adults" is a conversion keyword. The search volume is lower. The competition is lower. The reader who types it knows exactly what they want — and if your book matches, they buy.

New books need conversion keywords first. Build sales velocity on specific, low-competition terms. That velocity improves your overall rank, which eventually lets you compete for broader discovery terms.

Why chasing high-volume keywords kills new book sales

Here is the mechanism, stated plainly: Amazon shows readers the books most likely to result in a purchase. It determines "most likely" by looking at historical conversion rates for a given keyword. A book with 400 reviews and a 4.6-star rating has a proven conversion rate. Your new book has none. A9 will not surface your book for a high-volume keyword when a better-converting option exists — and for competitive terms, a better-converting option always exists.

Targeting "self-help books" as a new author is not ambitious. It is invisible. You will not appear in the results, which means zero impressions, zero clicks, zero sales from that keyword. The keyword does nothing for you except occupy a backend field that could have held a viable term.

The threshold varies by category, but as a working rule: if the top 10 results for a keyword average more than 100 reviews, a new book cannot compete for organic placement on that term. Look for keyword targets where the top 10 average fewer than 50 reviews. That is where new books can rank.

Step 1 — Build your seed keyword list before touching any tool

Every keyword tool — including Pubscout's Keyword Research feature — requires a starting input. The quality of your seed list determines the quality of everything downstream. Most authors start with genre labels. That is the wrong starting point.

The reader-first framework: think in problems, not genres

Genre labels are author language. "Cozy mystery," "dark fantasy romance," "productivity nonfiction" — these are how authors categorise their work. They are not how readers search for it.

Readers search in problems, moods, and situations:

  • "Books to read when you can't sleep"
  • "Fantasy romance with slow burn"
  • "How to stop procrastinating book"
  • "Mystery series set in small towns"

Before opening any tool, write down 10 answers to this question: What problem, mood, or situation is my reader in when they need this book? These answers become your seed keywords. They will surface keyword variations that genre-first thinking never reaches.

For a nonfiction book on building habits, genre-first thinking produces: "habit formation book," "productivity nonfiction," "self-improvement." Reader-first thinking produces: "how to build habits that stick," "why I can't stick to a routine," "daily habits for anxious people," "habit tracker journal for beginners." The second list has lower competition and higher purchase intent on every term.

Mining Amazon browse categories for hidden keyword signals

Amazon's browse category sidebar is an underused keyword source. When you search a broad term on Amazon Books and look at the left-hand category refinements, you are seeing Amazon's own taxonomy for how readers navigate to books like yours. These category labels are keyword signals.

Search "anxiety" in Amazon Books. The sidebar shows refinements: "Anxiety & Phobias," "Stress Management," "Mental Health," "Mindfulness & Meditation." Each of these is a validated keyword cluster — Amazon built the navigation around them because readers use them to find books.

Go one level deeper. Click into "Anxiety & Phobias" and look at the sub-refinements and the titles that appear. The words in those titles — especially in subtitles, where authors deliberately place keywords — are telling you what terms are working in that niche right now.

Pull 10–15 titles from the top results. Write down every word that appears in their subtitles. You will see patterns. Those patterns are your keyword clusters.

Using Amazon Autocomplete systematically (the A-to-Z method)

Amazon Autocomplete shows you real search queries from real readers in real time. It is the most direct demand signal available — more current than any third-party keyword database, because it reflects what readers are searching today.

Most authors use it casually: type a seed keyword, look at the dropdown, pick one or two suggestions. The A-to-Z method is more systematic.

Take one seed keyword — say, "anxiety journal." Type it into the Amazon search bar, then add a space and the letter A. Write down every autocomplete suggestion. Then replace A with B. Then C. Continue through Z. Do the same with numbers 1–9.

This process surfaces keyword variations that would never appear in a standard autocomplete scan. "Anxiety journal aesthetic," "anxiety journal blank pages," "anxiety journal CBT," "anxiety journal daily prompts" — each of these is a real search query with real volume. Many will have low competition because most authors never find them.

For a single seed keyword, the A-to-Z method typically produces 40–80 candidate keywords. Not all will be relevant. Filter for relevance, then move to evaluation.

This is also where how to find keywords for KDP books becomes a systematic process rather than guesswork. The A-to-Z method, combined with category mining, gives you a raw list of 100+ candidates before you've opened a single paid tool. That raw list is what you bring into keyword evaluation.

Step 2 — Evaluate keyword viability before you commit

A long keyword list is not a keyword strategy. You need to evaluate each candidate against two criteria: demand (is anyone searching for this?) and competition (can a new book rank for it?).

Reading BSR as a demand proxy

Amazon does not publish search volume data for book keywords. What it does publish is BSR — Best Sellers Rank — for every title. BSR is a real-time signal of sales velocity. A lower BSR means more recent sales.

When you search a keyword and look at the BSR of the top 10 results, you are reading the demand for that keyword. If the top 10 books all have BSRs below 50,000 in the Kindle Store, the keyword has enough demand to sustain sales. If the top results have BSRs above 500,000, the keyword has low demand — readers aren't searching it, or they aren't buying when they do.

The viable range varies by category, but as a working benchmark: for a keyword to be worth targeting, at least 3 of the top 10 results should have a Kindle Store BSR below 100,000. That indicates consistent purchase activity on that term.

Assessing the review barrier

Review count is the competition signal that matters most for new books. It is a proxy for how much social proof the existing results carry — and therefore how hard it is for a new title to earn clicks against them.

For each keyword candidate, look at the top 10 results and note the review counts. Calculate the average. If the average is above 100 reviews, the keyword is competitive for a new book. If the average is below 50, it is viable. If the average is below 25, it is an opportunity.

This is the core of what Pubscout's Niche Intelligence tool surfaces automatically — review count distribution, BSR range, and estimated sales across the top results for any keyword or niche. Running this analysis manually is possible but slow. At scale, across dozens of keyword candidates, it becomes the bottleneck in your research process.

The keyword viability matrix

Plot each keyword candidate on two axes: demand (BSR signal) and competition (average review count). Four quadrants emerge:

  • High demand, low competition: target immediately. These are rare but they exist, especially in emerging sub-niches.
  • High demand, high competition: long-term targets. Build toward these after you have reviews and sales history.
  • Low demand, low competition: low priority. You can rank, but there's not enough traffic to matter.
  • Low demand, high competition: avoid entirely. No upside.

For a new book, your launch keyword set should be 80% high-demand/low-competition and 20% high-demand/high-competition. The first group builds velocity. The second group is where you want to be in 6 months.

Step 3 — Place keywords where Amazon actually reads them

Keyword placement on KDP is not just about the 7 backend fields. Amazon indexes keywords from multiple fields, and the weight it assigns to each differs. Understanding the hierarchy changes how you allocate your best keywords.

The keyword hierarchy: title, subtitle, backend

Amazon gives the strongest indexing weight to your title and subtitle. A keyword in your subtitle will rank faster and more reliably than the same keyword in a backend field. This is not a minor difference — it is significant enough to change your titling strategy.

For nonfiction, this is straightforward: your subtitle is keyword real estate. "The Anxiety Journal: Daily Prompts and CBT Exercises for Overthinkers" contains five distinct keyword concepts in the subtitle alone. Each one is indexed.

For fiction, it is more nuanced. Genre signals in the series name and subtitle ("A Small-Town Cozy Mystery," "A Dark Fantasy Romance") function as keywords. Readers search these phrases. Amazon indexes them. A fiction author who ignores the subtitle is leaving indexing weight on the table.

The backend fields handle everything your title and subtitle cannot accommodate. Use them for:

  • Keyword variations you couldn't fit in the title
  • Reader-language phrases ("books for anxious overthinkers")
  • Complementary topic terms that expand your discoverability surface
  • Audience descriptors ("for teens," "for beginners," "for couples")

What not to put in your backend keyword fields

Amazon's KDP guidelines prohibit certain keyword types, and violating them can suppress your listing. Beyond the guidelines, some keyword choices are simply wasteful: