A9 vs ‘A10’ for KDP: What Changed, What Didn’t, and What Amazon Actually Says

“A9 versus A10” makes a neat headline because it suggests a clean before-and-after story. Old Amazon rewarded keywords. New Amazon rewards engagement, authority, and external traffic. Change the checklist and you are done.

Amazon’s public KDP documentation is not that tidy. There is no published KDP specification showing a dated A9 formula, a dated A10 replacement, and a table of old and new weights. So this article takes a different approach: which common A9/A10 claims are supported, which are plausible but unverified, and which publishing tactics remain sensible either way?

The Quick Answer: Search Evolved, but the Public Formula Never Arrived

Amazon has continuously changed search, recommendations, ads, product pages, and AI shopping. That evolution is real. What KDP authors do not have is an official document saying “A9 used these weights, A10 now uses these weights.” The best comparison is therefore evidence-based: compare old tactics with current KDP rules and current shopping behavior, then keep the tactics that still serve relevance and readers.

A9 and A10: What the Labels Are Useful For

That distinction matters for SEO content too. If you publish a page promising “the 2026 A10 weights,” you create a maintenance problem because the page looks precise without having an authoritative source. A more durable page explains the documented constraints, the observable customer journey, and the tests an author can run on their own books.

“A9” has long been used as shorthand for Amazon search technology. “A10” became a popular marketing label for a newer, broader view of Amazon search. The labels can be useful conversationally, provided they do not turn into invented certainty.

When someone says “A10 rewards external traffic 3x more,” ask for the Amazon source. When someone says “modern Amazon search uses more context than exact keywords,” that is directionally consistent with the way Amazon now describes search and AI shopping, but it still does not give you a KDP scoring formula.

The labels can still be useful as historical shorthand when you are reading older KDP advice. “A9” usually points to an earlier search-optimization mindset centered on metadata relevance and sales performance, while “A10” is often used by marketers to describe a broader belief that Amazon now interprets more customer and off-Amazon signals. The problem begins when shorthand is converted into precise weights, secret factors, or guaranteed tactics with no Amazon source behind them.

That is why this article is deliberately narrower than our current guide to Amazon search and the so-called A10 algorithm. The broader guide explains what authors can verify today; this comparison explains why the old A9-versus-A10 story is a poor foundation for publishing decisions.

What Amazon Says About Search in 2026

KDP says search results can be influenced by factors including past sales history, availability, length of time listed, and popularity. It also tells authors to use relevant keywords and accurate categories, and Amazon’s Read Sample documentation says words inside the book can help return relevant book results. These points favor a coherent book-and-listing strategy. They do not justify fixed weights for “authority,” “engagement,” “CTR,” “external traffic,” or “review velocity.”

Which Old Tactics Still Make Sense?

Exact-match language is not inherently bad. If “large print word search for seniors” accurately describes the product, it can belong in keyword research and may belong naturally in visible copy. The problem begins when you repeat a phrase because a checklist told you density itself earns rank. Relevance is useful; mechanical repetition is not.

The same is true of categories. Older playbooks often treated a tiny category as a badge hack. Current KDP guidance emphasizes accurate categorization and warns against misleading choices. A category should create a plausible browse route for the right reader, not merely offer the lowest apparent competition number.

Relevant keyword research still matters. KDP still provides up to seven keyword fields. Category fit still matters. You still choose up to three categories. Sales still matter. KDP explicitly says sales history can influence search and customer activity drives sales rank.

What has aged badly is the idea that the goal is to repeat exact phrases mechanically everywhere. Amazon gives you dedicated metadata fields, while the public description has to persuade a human reader. Our semantic search and KDP listings guide explains how to separate those jobs.

Which “A10 Tactics” Should You Treat as Hypotheses?

What Actually Changed for Authors

Amazon’s shopping surfaces are also more personalized than the old “one ranking for everyone” mental model suggests. Alexa for Shopping uses customer preferences and shopping history in addition to product knowledge. That means two shoppers can have materially different discovery journeys even when they type similar questions.

For authors, this makes market fit more important, not less. You cannot optimize for every customer. You can make the book unmistakably relevant to a defined audience and give Amazon accurate information that supports that match.

The customer journey is broader. A shopper may arrive through traditional search, Sponsored Products, social media, an email, an Amazon recommendation, or Alexa for Shopping. They can compare products conversationally and encounter AI summaries on search and product pages.

That makes consistency more important. Your metadata, cover, description, reviews, and opening pages should all describe the same book for the same audience. This is a reader-and-product discipline first, not a way to manipulate a hidden score.

The reader journey also became more conversational. Amazon renamed Rufus to Alexa for Shopping in May 2026 and describes the assistant as drawing on its product catalog, reviews, community Q&A, information from across the web, and personalized shopping context. For authors, that strengthens the case for clear, consistent product information, but it still does not establish a special “AI ranking factor” for books.

Measurement improved too. Eligible KDP authors can use Amazon Attribution to compare certain non-Amazon marketing channels with Amazon activity. That gives you a better basis for judging external campaigns than the old advice to assume every off-Amazon click carries a hidden algorithm bonus.

How to Build a Strategy That Survives Algorithm Changes

Build a before-and-after record for meaningful changes. Note the date, what changed, current keyword positions, sales-rank direction, advertising conditions, price, and major promotions. Then resist changing three more things the next morning. The cleaner your test, the more useful your own evidence becomes.

Choose markets where you can compete. Research the language real shoppers use. Make the book and listing clearly relevant to those readers. Track what happens after changes. Protect rankings that already work and improve weaker opportunities deliberately.

That is also why the newer Rank Fuel workflow is book-centered. The current Earnings Outlook combines keyword positions with sales-rank momentum, while Preserve/Improve keyword intentions can be carried into Listing Optimizer. The goal is continuity of evidence, not an A10 score. For the fuller current-search framework, read Amazon’s “A10” Algorithm: What KDP Authors Can Actually Verify.

Frequently Asked Questions About A9 vs A10

The Better Comparison

The useful distinction is not old algorithm versus new algorithm. It is unsupported certainty versus evidence-led publishing.

Keep the parts of classic KDP SEO that still help readers find the right book. Add the newer realities of conversational shopping, better measurement, and book-level performance tracking. Drop the imaginary percentages.