Amazon Optimisation
COSMO vs A9/A10: What’s the Difference and Why It Matters for E‑Commerce Visibility
Amazon search is no longer just about matching keywords. It’s becoming more semantic, contextual, and intent-driven—and two of the most important systems shaping this shift are A9/A10 and COSMO. But what exactly are they? How do they interact? And what should brands optimise for?
5 August 2025
8 min read
🔍 What Is the A9/A10 Algorithm?
A9 was Amazon’s original search engine algorithm. It ranked products based on a combination of:
- Keyword relevance (title, bullet points, backend search terms),
- Click-through rates, conversion rates, and sales velocity,
- Availability, price, and Prime eligibility.
A10 is the informal name for its evolved version. While Amazon has never officially released an “A10,” the term is widely used to describe updates that place greater weight on:
- Organic behaviour (e.g. add-to-cart rates, time on page),
- Seller authority and performance,
- External traffic from trusted sources.
TL;DR: A9/A10 = Ranking engine for matching products to queries, based on a mix of relevance, performance, and commercial signals.
🧠 What Is COSMO?
COSMO is not a search engine, it’s a common sense knowledge graph system.
It mines user intentions from millions of search-browse-buy behaviours and creates structured “common sense” relationships between concepts, like:
- “Customers search for ‘winter coat’ to stay warm”
- “Pregnant women often buy slip-resistant shoes”
- “People buy protein shakes after gym workouts”
COSMO uses:
Behavioural signals (search/buy, co-buy),
LLM-generated explanations,
Human curation and instruction-tuned fine-tuning, to create millions of semantic triples (X → usedFor → Y).
TL;DR: COSMO = Intent engine that adds context and common sense to Amazon’s understanding of shopper behaviour.
🚀 Why You Should Care
For Brands:
Optimising for A9/A10 is no longer enough. You need to align your content with shopper intent, what problems your product solves, for whom, and why.
COSMO-like systems power LLMs like ChatGPT and Amazon’s own Rufus assistant. If your content doesn't reflect use cases, benefits, audience, you’re unlikely to be indexed meaningfully by these newer AI search layers.
For Marketers:
Start thinking in semantic layers: go beyond “what does my title say?” to “what questions does my product answer?”.
Use structured content: A+ content, FAQs, videos, and bullet points that explicitly address needs, problems, and uses - because COSMO and LLMs pick up on those.
For Search Optimisers:
Classic SEO = optimise for A9/A10.
AI/LLM visibility = optimise for COSMO-style intent.
Smart brands will do both.
📌 Key Takeaways
A9/A10 = Match and rank based on query, conversion and performance.
COSMO = Understand intent based on behaviour, context, and common sense.
Both shape product discovery, but COSMO is powering the next wave of AI shopping agents like Amazon Rufus, Buy with AI, and ChatGPT plug-ins.
Brands that think beyond keywords— toward needs, problems, and intent will win on both fronts.
Final Thought
Think of A9/A10 as the map, and COSMO as the compass.
The map shows where products are, based on keywords and performance.
The compass tells Amazon which direction a shopper intends to go, even if they don’t use the perfect words.
If your content speaks to that intent, your brand gets found. If not, you're invisible, no matter how strong your keywords are.
A9 was Amazon’s original search engine algorithm. It ranked products based on a combination of:
- Keyword relevance (title, bullet points, backend search terms),
- Click-through rates, conversion rates, and sales velocity,
- Availability, price, and Prime eligibility.
A10 is the informal name for its evolved version. While Amazon has never officially released an “A10,” the term is widely used to describe updates that place greater weight on:
- Organic behaviour (e.g. add-to-cart rates, time on page),
- Seller authority and performance,
- External traffic from trusted sources.
TL;DR: A9/A10 = Ranking engine for matching products to queries, based on a mix of relevance, performance, and commercial signals.
🧠 What Is COSMO?
COSMO is not a search engine, it’s a common sense knowledge graph system.
It mines user intentions from millions of search-browse-buy behaviours and creates structured “common sense” relationships between concepts, like:
- “Customers search for ‘winter coat’ to stay warm”
- “Pregnant women often buy slip-resistant shoes”
- “People buy protein shakes after gym workouts”
COSMO uses:
Behavioural signals (search/buy, co-buy),
LLM-generated explanations,
Human curation and instruction-tuned fine-tuning, to create millions of semantic triples (X → usedFor → Y).
TL;DR: COSMO = Intent engine that adds context and common sense to Amazon’s understanding of shopper behaviour.
🚀 Why You Should Care
For Brands:
Optimising for A9/A10 is no longer enough. You need to align your content with shopper intent, what problems your product solves, for whom, and why.
COSMO-like systems power LLMs like ChatGPT and Amazon’s own Rufus assistant. If your content doesn't reflect use cases, benefits, audience, you’re unlikely to be indexed meaningfully by these newer AI search layers.
For Marketers:
Start thinking in semantic layers: go beyond “what does my title say?” to “what questions does my product answer?”.
Use structured content: A+ content, FAQs, videos, and bullet points that explicitly address needs, problems, and uses - because COSMO and LLMs pick up on those.
For Search Optimisers:
Classic SEO = optimise for A9/A10.
AI/LLM visibility = optimise for COSMO-style intent.
Smart brands will do both.
📌 Key Takeaways
A9/A10 = Match and rank based on query, conversion and performance.
COSMO = Understand intent based on behaviour, context, and common sense.
Both shape product discovery, but COSMO is powering the next wave of AI shopping agents like Amazon Rufus, Buy with AI, and ChatGPT plug-ins.
Brands that think beyond keywords— toward needs, problems, and intent will win on both fronts.
Final Thought
Think of A9/A10 as the map, and COSMO as the compass.
The map shows where products are, based on keywords and performance.
The compass tells Amazon which direction a shopper intends to go, even if they don’t use the perfect words.
If your content speaks to that intent, your brand gets found. If not, you're invisible, no matter how strong your keywords are.