In the industry, Amazon's search is known as A9, a name inherited from A9.com, Amazon's search-technology subsidiary. According to Wikipedia, Amazon took the A9.com site down in 2019 after inquiries from The Wall Street Journal for a report on its search algorithm. Amazon doesn't use that name in its SEO guide and doesn't document its current system. It does explain that it matches what customers search for against listing data such as the title, description and backend keywords. In 2024, Amazon researchers published COSMO, a system that infers shopping intent from searches and purchases. According to that paper, it's deployed in search applications such as navigation. For that feature, which the paper calls relatively minor on the search page, A/B tests on about 10% of US traffic showed a 0.7% relative increase in product sales within that segment. The paper doesn't say how much weight it carries in the order of results.
The term "A10" circulates in forums and industry blogs. In Amazon's seller forums, for example, one seller describes it as weighing seller authority, product relevance and external traffic more heavily, and asks for reliable documentation. We didn't find that name on the Amazon pages we reviewed, apart from a seller's question in its forums. There, an Amazon community manager wrote that the company doesn't publish official documentation about its search algorithms. This glossary treats "A10" as a label used outside Amazon's documentation, not as something Amazon states.
What Amazon's guide does say is that listing quality and account health can contribute to rankings, that adjusting price can improve ranking and that keyword stuffing can hurt it. Best Sellers Rank indicates how well a product sells within its category. Amazon doesn't publish how signals are weighted. The idea that CTR, conversion rate or reviews outweigh keywords is an observation by third parties, not something Amazon states.