Short version: Amazon SEO in 2026 means optimizing for three layers at once — the A9/A10 engine that indexes and ranks listings, the COSMO AI layer that ranks by shopper intent, and Alexa for Shopping (formerly Rufus), the conversational layer that answers shopping questions directly. Keywords still get you found; click-through rate, conversion rate, and sales velocity decide how high you rank; and structured, complete listing data decides whether AI surfaces recommend you at all.

"Amazon SEO" gets used loosely to mean everything from keyword stuffing to full listing optimization. In practice it's the discipline of making a product findable and rankable across every way Amazon surfaces products — traditional search, personalized recommendations, and now conversational AI answers. This guide is the overview: what's actually running under the hood, what ranks a listing, and where to go deeper on each piece.

The three layers of Amazon search

Amazon search isn't one algorithm — it's three systems stacked on top of each other, and none of them replaced the one below it:

  • A9/A10 — the foundational ranking engine. It indexes your listing's text and attributes, decides which searches you're eligible to appear in, and ranks the candidate pool using performance signals. It's still running; nothing has replaced it. "A10" is a label the seller community uses for the current, more performance-weighted version — Amazon has never officially confirmed that name.
  • COSMO — an AI layer built on a knowledge graph that sits on top of A9/A10. Instead of matching literal keywords, it connects what a shopper searches to what they actually want, so a listing can rank for intent-matched queries even without the exact phrase.
  • Alexa for Shopping — the conversational layer, replacing the standalone Rufus assistant that Amazon retired on May 13, 2026. It answers shopping questions directly in the search bar and generates AI overviews and comparisons, drawing on your listing copy, attributes, reviews, and Q&A — but only where that content is structured clearly enough for it to extract.

The practical takeaway: you're not choosing which layer to optimize for. A9/A10 has to index and rank you before COSMO or Alexa for Shopping can ever surface you. See our guides on what Alexa for Shopping is and how it differs from keyword search for the deeper dive on the AI layers.

What actually ranks a listing

Keyword relevance gets you into the candidate pool. What moves you up within it is mostly commercial performance. Amazon and third-party seller-tool data point to the same core set of signals:

  • Click-through rate (CTR) — whether shoppers pick your listing over the seven or eight others on the same results page. Driven by your main image, title, and price.
  • Conversion rate — whether a click turns into a sale. Widely cited as the single strongest ranking signal once a listing is being shown.
  • Sales velocity — how consistently a product sells over time, weighted by keyword. Steady daily sales carry more weight than a short promotional spike.
  • Reviews and review velocity — rating, count, and recency all feed both ranking and shopper trust.
  • Inventory availability and pricing stability — stockouts and erratic pricing suppress rankings, since Amazon deprioritizes listings it can't reliably fulfill.
  • Account and listing health — order defect rate, return rate, and policy compliance.

This means Amazon SEO isn't a one-time copywriting task. Getting keywords into the right fields earns you visibility; converting on that visibility is what keeps it.

Field-by-field: where the work actually happens

Once you understand the ranking logic, the work is mapping keywords and structured data to specific fields, in priority order:

  1. Keyword research. Find the terms and the intent behind them before writing anything. See Amazon keyword research: a beginner's guide.
  2. Title. Capped at 75 characters as of July 27, 2026 (media exempt) — brand, primary keyword, and the one attribute that matters most. See title optimization & the 75-character limit.
  3. Bullets and Item Highlights. Secondary keywords and use cases, written for a human but readable by COSMO's intent matching.
  4. Backend search terms. Synonyms, variants, and misspellings not already used elsewhere in the listing. See the complete backend keywords guide.
  5. A+ Content. Lifts conversion, but text baked into images isn't indexed — pair it with complete copy elsewhere. See the A+ Content guide.
  6. Images, attributes, reviews, and Q&A. The data Alexa for Shopping and COSMO draw on most directly for AI answers and comparisons.

For the full walkthrough in this order, see how to optimize an Amazon listing: a step-by-step guide.

Example — where one keyword cluster lands

Product: insulated travel mug

Title: brand + "insulated travel mug" (primary keyword)

Item Highlights / bullets: "leak-proof," "fits car cup holder," "keeps coffee hot 6 hours"

Backend terms: "spill proof tumbler," common misspellings

Q&A / A+: "does it fit a standard cup holder?" answered directly — this is what Alexa for Shopping surfaces in a comparison.

A quick-start checklist

  • ☐ Keyword list built and clustered by primary / secondary / long-tail + intent
  • ☐ Title within 75 characters, brand + primary keyword + top attribute
  • ☐ Item Highlights and bullets carry secondary keywords and use cases
  • ☐ Backend search terms filled with terms not used elsewhere
  • ☐ A+ Content built, with key claims also present as indexable text
  • ☐ Product attributes fully filled in (not left blank)
  • ☐ Reviews and Q&A monitored for real shopper language and gaps
  • ☐ Pricing and inventory kept stable; account health monitored

Common Amazon SEO mistakes

  1. Treating it as copywriting only. A perfectly worded listing with a weak conversion rate will still rank poorly.
  2. Keyword stuffing. Amazon indexes each keyword once per field; repeating it wastes space and can hurt clarity, which hurts CTR and conversion.
  3. Leaving attributes blank. Every empty attribute is a question Alexa for Shopping and COSMO can't answer about your product.
  4. Optimizing once and never revisiting. Search language and competitor listings shift; refresh keyword targeting quarterly.

See exactly where your listing is losing visibility

OptimalCentral audits your listing against everything A9/A10, COSMO, and Alexa for Shopping look for, and returns a prioritized fix list — free, in under a minute.

Start free trial →

Start with the 2026 listing audit checklist to see where you stand today, or jump straight to keyword research if you're starting from scratch.

Frequently asked questions

What is Amazon SEO?

Optimizing a product listing — title, bullets, backend search terms, A+ Content, images, reviews and Q&A — so it ranks well in Amazon's search results and gets recommended by its AI layers. Unlike Google SEO, it's judged heavily on commercial performance (CTR, conversion rate, sales velocity), not just keyword relevance.

Is Amazon SEO still just about keywords?

No. Keywords get you indexed by A9/A10 and build the candidate pool, but ranking within that pool is driven more by click-through rate, conversion rate, and sales velocity. COSMO also ranks by shopper intent, not literal keyword matches, so natural, relevant language matters more than keyword density.

Do A9 and A10 still exist, or did COSMO replace them?

A9/A10 is still the underlying ranking engine that indexes listings and produces the candidate set for a search — it hasn't been replaced. COSMO is an AI layer on top of it; Alexa for Shopping (formerly Rufus, retired May 13, 2026) is a further conversational layer on top of both. "A10" is community terminology, not an Amazon-confirmed name.

Where should I start with Amazon SEO?

Start with keyword research, then fix the highest-impact fields first: title (within the 75-character limit), backend search terms, and bullets/Item Highlights. Add A+ Content and complete every product attribute next, since Alexa for Shopping and COSMO both rely on structured attribute data, not just visible copy.