The Complete Guide to Amazon's A10 Algorithm

What is the Amazon A10 algorithm?

What "A10" actually is (and is not)

"A10" is a term coined by the Amazon seller and agency community, not an official Amazon name. Amazon has never published its product-search ranking algorithm, has never branded a version "A10," and has not confirmed any algorithm "version change." The earlier label "A9" traces to A9.com, a former Amazon search subsidiary, and even that name was never marketed by Amazon as a consumer-facing ranking product.

What "A10" describes is a pattern that sellers and agencies have observed in how the marketplace ranks products today: a shift away from rewarding keyword-heavy listings and toward rewarding products that demonstrably sell and satisfy customers. Treat the term as useful shorthand for current ranking behavior, not as a documented, named system you can find in Seller Central.

This distinction matters because much of the "A10" literature online presents inferred mechanics - exact percentages, weightings, and factor lists - as if Amazon published them. Amazon did not. In this guide, Amazon-confirmed facts are flagged explicitly and separated from practitioner consensus.

What Amazon actually confirms

Amazon publishes very little about ranking, but a few statements are official and worth anchoring to:

• Ranking relates to sales volume. In its official Best Sellers Rank guidance, Amazon states that an item selling well "has a better chance of ranking higher both in organic search and on Amazon Best Sellers lists."
• Recent sales matter more than older sales. Amazon confirms Best Sellers Rank is "calculated using sales volume data" and that "recent sales count more than older sales."
• Best Sellers Rank is not the same as search ranking. Amazon explicitly separates the two and notes BSR is not influenced by page views or customer reviews.
• External traffic is a sanctioned activity. Through the Brand Referral Bonus program, Amazon pays brand-registered sellers a credit for driving off-Amazon traffic that converts, and provides Amazon Attribution to measure it.

Notably, Amazon does not officially claim that external traffic improves search ranking. The Brand Referral Bonus is framed as a fee credit, not a ranking benefit. The "ranking halo" from external traffic is a practitioner inference, not an Amazon statement.

A9 era vs A10 era: what practitioners observe changed

The following is industry consensus, not Amazon-confirmed mechanics. Sellers describe the "A9 era" as heavily weighted toward keyword relevance plus raw sales velocity - meaning a listing packed with the right keywords and pushed hard with ads could climb fast. The "A10 era" is described as more behavioral and quality-sensitive:

• External traffic that converts is treated as evidence of genuine demand, not just on-platform activity.
• Organic, non-sponsored conversions appear to carry more weight than sales driven purely by ad spend - sellers report that ad-bought velocity alone holds rankings less durably than it once did.
• Seller authority - account health, fulfillment reliability, shipping speed, low return and defect rates, and time on the platform - is described as a bigger input.
• On-Amazon engagement - click-through rate from search, dwell behavior, add-to-cart, and conversion - is described as feeding back into placement.

The practical upshot most agencies agree on: keyword stuffing alone no longer ranks. Relevance keywords act as a gate to appear for a query; behavioral and performance signals decide where you place. For a deeper look at the keyword-and-velocity mechanics that defined the earlier era, see our complete guide to the A9 algorithm.

How "A10" came to be: a timeline

The "A10" label grew out of the older "A9" naming as sellers observed Amazon's ranking shift from keyword-and-velocity mechanics toward behavioral and authority signals, and finally toward AI-driven, personalized search.

1. 2003

   A9.com founded

   Amazon spins up A9.com, a search-technology subsidiary. Its name later became the seller-community shorthand "A9" for Amazon's product-search ranking, though Amazon never marketed it as a consumer ranking product.

2. 2008

   The public A9.com search engine shuts down

   A9.com discontinues its public web-search portal, and Amazon's product-search work moves under Amazon internally over the following years. Sellers continue using "A9" to describe ranking behavior.

3. 2020-2021

   Sellers start using "A10"

   Agencies and sellers coin "A10" to describe observed ranking changes - more weight on external traffic, organic conversion, and seller authority. This is a community label; Amazon never announced any "A10" version.

4. 2021

   Brand Referral Bonus launches

   Amazon launches an official program paying brand-registered sellers a credit for driving converting off-Amazon traffic, validating external traffic as a sanctioned growth lever (though without claiming a ranking boost).

5. 2024-2026

   AI-driven, contextual search

   Amazon publicly leans into AI shopping (the Rufus assistant) and contextual retrieval (referenced internally as COSMO), making results more personalized and further loosening the idea of a single fixed ranking formula.

Relevance is the gate, performance is the placement

A useful mental model that holds up against both Amazon's confirmed statements and practitioner observation: Amazon ranking works in two stages. First, relevance determines whether your product is even eligible to show for a search - this is driven by the keywords in your title, bullets, description, and backend search terms matching the shopper's query. Second, performance determines your position among the eligible products - driven by conversion, sales velocity, and customer-satisfaction signals.

This is why two listings with identical keywords can rank very differently: the one that converts more shoppers into buyers, ships reliably, and earns repeat purchases out-positions the one that merely mentions the right words. It also explains why stuffing a title with every keyword variation backfires - it can dilute relevance for your primary terms and depress click-through and conversion, the very signals that decide placement.

The external-traffic lever

The single most-discussed "A10" tactic is driving external traffic that converts. The logic, per practitioner consensus, is that a shopper who arrives from Google, a social ad, an email, or an influencer and then buys signals real-world demand that Amazon's own ad inventory cannot fully manufacture.

Here is where the confirmed/inferred line is sharpest. Amazon confirms it wants and rewards external traffic - the Brand Referral Bonus pays you roughly a 10% average credit on qualifying off-Amazon sales, and Amazon Attribution exists specifically to track it. Amazon does not confirm that this traffic lifts your organic ranking. The ranking benefit is something sellers infer from the sales-velocity signal Amazon does acknowledge: external traffic that converts adds to your sales, and sales help ranking. Drive converting traffic to a well-optimized listing, not raw clicks - traffic that bounces can hurt your conversion rate and work against you. Many sellers pair this with Amazon Sponsored Ads to seed early momentum before defending rank organically.

Important caveats on accuracy

Because Amazon does not publish its algorithm, every "A10 ranking factor" list online - including parts of this guide's signals section - is reverse-engineered from seller experiments, agency data, and patents. Treat specific numbers ("reviews under 4.3 stars cut ranking 20-40%," "FBA gets a fixed boost," exact factor percentages) with caution: they are estimates, often repeated without primary evidence, and they vary by category and over time.

Amazon's ranking is also no longer well-described by a single static formula. Recent Amazon search work involves AI-driven, contextual, and personalized retrieval (sellers reference internal initiatives like COSMO and the Rufus shopping assistant), meaning results differ by shopper, query history, and context. "A10" is best understood as a convenient label for "how Amazon ranks now," not a fixed algorithm with knowable settings. For how Amazon's AI layer shapes discovery, see our guide to Amazon Rufus and Bedrock.

A10-era ranking signals

The signals below mix the few Amazon-confirmed inputs with the broader practitioner consensus. Read the second column carefully: it flags what is documented versus inferred, so you do not over-trust a factor Amazon has never stated.

The practical takeaway is that relevance keywords only get you into the running; the signals that decide placement are overwhelmingly about whether shoppers buy, whether you can deliver, and whether the sales are durable rather than ad-propped.

How to optimize for the A10 era

To optimize for A10-era ranking, write listings for humans first, fix the entire conversion path before driving traffic, earn organic and converting external sales, and protect seller authority - then test in your own category rather than trusting fixed factor weightings.

1. Write listings for humans first, with keywords placed naturally

   Relevance keywords get you eligible to appear, but click-through and conversion decide placement. Stuffing keywords dilutes your primary terms and depresses the behavioral signals that actually rank you.

2. Optimize the whole conversion path before driving traffic

   Main image, title, price, bullets, A+ content, and reviews all affect conversion rate - the signal most tied to placement. Sending traffic to a weak listing wastes spend and can lower your conversion rate, hurting rank.

3. Drive converting external traffic and tag it with Amazon Attribution

   External traffic that buys adds to the sales velocity Amazon does acknowledge, and qualifying off-Amazon sales earn the Brand Referral Bonus credit. Attribution tags let you measure which sources actually convert.

4. Earn organic sales, not just ad-bought velocity

   Practitioners report organic conversions hold rankings more durably than sales propped up purely by ad spend. Use ads to seed momentum, then defend rank with organic performance.

5. Protect seller authority: avoid stockouts, ship fast, keep defects low

   Account health, fulfillment reliability, and Buy Box ownership are described as trust inputs. Stockouts can suppress ranking even after you restock, because the system flags supply unreliability.

6. Build genuine reviews to lift conversion, not to game rank

   Reviews raise buyer trust and conversion rate (an indirect ranking help). Amazon says reviews do not affect Best Sellers Rank, so treat them as a conversion lever, not a direct ranking hack - and never buy fake reviews.

A10 myths vs. reality

Few Amazon topics carry as much confidently-stated folklore as "A10." Here are the most common myths and what is actually true.

Frequently asked questions

The bottom line

Related guides

References

1. Amazon - Guide to Amazon sales rank: Best Sellers Rank (BSR) (official)
2. Amazon - Brand Referral Bonus: Drive traffic and boost your bottom line (official)
3. Amazon Seller Central - Brand Referral Bonus help page (official)
4. Semrush - Amazon SEO: Top Strategies to Optimize Your Product Listings
5. Moz - Amazon vs. Google: Decoding the World's Largest E-commerce Search Engine
6. Moz - How to Rank Well in Amazon, the US's Largest Product Search Engine
7. AMZScout - Amazon A9 vs. A10: Top Tips for Boosting Product Rankings
8. Search Engine Journal - The Top Search Engines & AI Search Engines To Know