# Kadaxis

> Kadaxis helps publishers connect books with readers through Amazon keyword research, AI discovery intelligence, and practical AI advice. Founded by Chris Sim.

Canonical site: https://kadaxis.com
XML sitemap: https://kadaxis.com/sitemap.xml
Full machine-readable site text: https://kadaxis.com/llms-full.txt

## Current pages

- [Book discovery intelligence for publishers](https://kadaxis.com/): Kadaxis helps publishers connect books with readers through Amazon keyword research, AI discovery intelligence, and practical AI advice.
- [Amazon keyword research for publishers](https://kadaxis.com/keywords): Amazon keyword research, backlist optimization, and search visibility reporting for publishers, with title-specific recommendations for existing ONIX workflows.
- [AI discovery (GEO) for publishers](https://kadaxis.com/ai-discovery): Find out which reader questions lead AI answers to your book, which titles and sources appear instead, and what to change in the book's description and pages.
- [AI consulting for publishing leaders](https://kadaxis.com/advisory): Practical AI advice for publishing leaders considering workflows, internal tools, retrieval over their content, and vendor decisions.
- [About Chris Sim and Kadaxis](https://kadaxis.com/about): Chris Sim founded Kadaxis to help publishers connect books with readers through keyword research, metadata intelligence, and AI discovery.
- [Clients](https://kadaxis.com/work): Kadaxis has worked with major trade publishers, large independents, and specialty presses on metadata and search visibility.
- [How we work](https://kadaxis.com/approach): How Kadaxis works with publishers on Amazon keywords, AI discovery, title-specific research, and recommendations for existing publishing workflows.
- [Research on book discovery](https://kadaxis.com/notes): Kadaxis research on how readers find books in Amazon search and AI tools, with practical questions for publishing teams.
- [Contact](https://kadaxis.com/contact): Get in touch with Kadaxis about book discovery keywords, GEO and AI in publishing. Questions and project enquiries are welcome.
- [BISG book discovery webinar and research](https://kadaxis.com/bisg): Research accompanying Chris Sim's BISG webinar on SEO, GEO, and keywords, November 19, 2026, with links to Kadaxis studies on Amazon and AI discovery.
- [Book discovery brief for publishing leaders](https://kadaxis.com/brief): A short introduction to Kadaxis's keyword research, AI discovery reports, deliverables, and experience for publishing leaders and their teams.
- [Privacy Policy](https://kadaxis.com/privacy): Kadaxis privacy policy for contact forms, newsletter subscriptions, hosting logs, and privacy rights.

## Current research

- [Publishers are rewriting book pages as FAQs. Under 1% of the pages ChatGPT cited had one that matched.](https://kadaxis.com/notes/publisher-faqs-under-one-percent): We collected 27,702 ChatGPT answers to reader requests and read the publisher pages behind them. Two pages in a hundred had an FAQ at all, and on the pages that did, the answer’s wording usually matched a passage somewhere else in the text. Put those together and FAQs account for less than 1% of answers returned by ChatGPT.
- [Where AI book recommendations come from](https://kadaxis.com/notes/where-ai-book-recommendations-come-from): Publisher websites appeared in three quarters of the ChatGPT source lists we examined. Our study of 27,702 answers to reader requests looks beyond the familiar website names to the book descriptions, reading guides and other pages that surfaced, and what publishers can learn from them.
- [GEO and AEO for publishers: Matching books to reader needs](https://kadaxis.com/notes/geo-and-aeo-for-publishers-matching-books-to-reader-needs): Reviews show why readers choose a book and how they use it. Connecting those insights with customer demand and ChatGPT’s recommendations helps you decide what to explain in your descriptions, which material to share, and where to publish it so the right readers can discover your books.
- [How books show up in ChatGPT and AI search](https://kadaxis.com/notes/how-books-show-up-in-chatgpt-and-ai-search): To improve a book’s visibility in ChatGPT and AI search, publishers need to understand what informs an answer and which components they can influence.
- [Should Book Keywords Be Single Words or Search Phrases?](https://kadaxis.com/notes/should-book-keywords-be-single-words-or-search-phrases): We conducted a study that found natural reader-search phrases made far better use of limited keyword space than tag-based single-word keywords. Search phrases produce substantially more visibility when searched directly or in expanded search forms.
- [Nearly half of publisher keywords are a single word](https://kadaxis.com/notes/nearly-half-of-publisher-keywords-are-a-single-word): We analyzed 2.4 million keywords across 94,385 books. Single-word keywords make up 44.57% of the total, by far the largest share. A quarter of the books carry no keywords at all. When publishers use the field, they add a median of 26 keywords per book. Effort is not the problem. Shape is.
- [Your book description is a search surface](https://kadaxis.com/notes/your-book-description-is-a-search-surface): Descriptions are written for readers who already found the book. New Kadaxis research shows Amazon also indexes them: phrases appearing nowhere else on a listing surface the book in organic search about 1.6% of the time, roughly 260 times the matched-control rate. The effect is real, small, and strongest for distinctive language.
- [ONIX keywords work. Most keywords that publishers create don't.](https://kadaxis.com/notes/amazon-keywords-work-most-publisher-submitted-keywords-dont): Plenty of people in publishing suspect keywords do not work. Recent Kadaxis research shows why: only about eight percent of publisher-assigned keywords rank in Amazon search. Keywords built for 2,000 books from a recent Kadaxis client ranked forty-two percent of the time.
- [When to put "book" in a keyword](https://kadaxis.com/notes/when-to-put-book-in-a-keyword): The old advice was simple. Never spend keyword space on product descriptors, because the retailer already knows the product is a book. That rule is too crude. Product descriptors carry real retrieval value when they tell the search system what kind of object the shopper wants. They turn into waste only when they repeat what the catalog already supplies or force an unnatural phrase.
- [Keyword portfolio ordering](https://kadaxis.com/notes/keyword-portfolio-ordering): A keyword portfolio is an ordered coverage system. The first keyword is measured against the book's visible metadata. The second, against that metadata plus the first keyword. Every position is judged by what it adds over everything above it, so a lower-scoring keyword can outrank a higher-scoring one when it reaches readers the rest of the portfolio has missed.
- [Keyword portfolio allocation](https://kadaxis.com/notes/keyword-portfolio-allocation): Most keyword strategy is treated as a ranking problem: generate candidates, score them, keep the top N, ship the list. That works at the individual-keyword level but fails at the portfolio level. A keyword buys access to a search doorway: a cluster of reader intent, search results, substitute books, and commercial expectations. Portfolio strategy is deciding which doorways deserve scarce metadata space, which overlap, and which introduce risk.
- [Amazon's warm-start research makes launch metadata a ranking decision](https://kadaxis.com/notes/amazon-warm-start-launch-metadata-ranking-decision): Amazon has described a production search-ranking system that uses substitute products to give new products a warm start before they have sales history of their own. For publishers, the implication is direct: cover, title, description, comp-title language, and category placement help determine the behavioral neighborhood a new book enters at launch. That neighborhood can shape early visibility, sales velocity, and the book's ability to build its own ranking signals.
- [The discoverability funnel](https://kadaxis.com/notes/the-discoverability-funnel): Amazon accounts for roughly half of US print book sales and two-thirds of US ebook sales, and most US consumers now start their online product searches on Amazon rather than Google. The path from a book's metadata to a reader's purchase runs through a funnel with four sequential filters. Most candidate books drop out at one of them.
- [The narrow space where keywords actually work](https://kadaxis.com/notes/the-narrow-space-where-keywords-actually-work): For a book keyword to do anything useful, it has to clear four independent tests: someone has to be searching for it, Amazon has to consider the book relevant to it, it has to add signal the rest of the metadata does not already supply, and the book has to be able to rank for it. Most keyword candidates fail at least one.
- [Not every keyword is doing work](https://kadaxis.com/notes/not-every-keyword-is-doing-work): In our recent analysis, only about eight percent of publisher-assigned book keywords produced a visible match in retailer search. The other ninety-two percent are not wrong; they are mostly repeating signals the rest of the metadata already supplies.
- [Relevance is not the same thing as ranking](https://kadaxis.com/notes/relevance-is-not-the-same-thing-as-ranking): Most keyword strategies treat search visibility as a single problem. It is two problems, in sequence, and confusing them is why so much keyword work fails to move performance.
- [Amazon Keywords Don't Stop at 250 Characters. Publishers Do.](https://kadaxis.com/notes/amazon-keywords-dont-stop-at-250-characters): A direct look at whether Amazon ignores ONIX keyword phrases after 250 characters, and what publishers should actually do with the field.
- [What reader reviews know that catalogues don't](https://kadaxis.com/notes/what-reader-reviews-know-that-catalogues-dont): Catalogues describe books in the trade's language. Readers search in their own. The gap between those two registers is where most of the lost sales live.
- [Why book keywords don't live inside the book](https://kadaxis.com/notes/why-book-keywords-dont-live-inside-the-book): Most keyword tools start with the text of the book. The text has value, but reader search often lives in a different vocabulary: need, outcome, comparison, emotional texture, and intent.
- [What we learned trying to predict bestsellers](https://kadaxis.com/notes/what-we-learned-trying-to-predict-bestsellers): A year of work on bestseller prediction from manuscript text produced a working system and a clear lesson: the model was answering the wrong question, and the right question requires variables the manuscript cannot supply.
- [Why comparable titles fail when derived from book text](https://kadaxis.com/notes/why-comparable-titles-fail-when-derived-from-book-text): Most automated comp-title tools operate on the text of the book. The comparisons readers and the trade actually make operate on something the text cannot fully capture: the reading experience.
- [Different keywords on different formats of the same book](https://kadaxis.com/notes/different-keywords-on-different-formats-of-the-same-book): A common piece of advice is to assign different keyword sets to the hardcover, paperback, ebook, and audiobook editions of the same title to expand the work's reachable search surface. We tested whether this works at scale. The phenomenon exists in specific cases, but our data does not support recommending it as default practice.

## Reading guidance

- Use current service pages to describe available work. Use the contact form for commercial enquiries.
- Research findings describe their measured scope; search visibility is not a measure of sales lift.
- Historical articles remain available on the website, outside this current research selection.