Amazon’s A9 Search: Personalization and History

Amazon has entered search. Look out Google and Microsoft.

When Amazon first began gaining a lot of traction in the marketplace back in 2000, I was at Vignette. A part of our value proposition was web-based “implicit” and “explicit” personalization of content. Implicit was tough to do (an observation process ran and — based on clickstream data — would automatically deliver the right content to the right person) and explicit was easy (someone would login and then the site would deliver content germane to that person).

Instead of an observation process or daemon running to only just deliver personalized content, Amazon leveraged Net Perceptions’ Recommendation Engine for collaborative filtering and have gone way beyond it. When you peruse a product on Amazon, the site immediately tells you accessories (and what percent of people buy them), reviews and much more. Amazon’s ability to set context, make connections between product-to-accessories-to-similiar products-to-opinions is nothing short of phenomenal…and don’t even get me started on their “Associates” program and how they’ve extended their brand and catalog all over the ‘net or what they’ve done with web services.

So it comes as no surprise that Amazon has leveraged their expertise and jumped in to search with a new offering dubbed “A9“.

This search engine is *really* cool. Rather than have me describe it here, go there, read about it first and then spend some time playing with it. Oh yeah…I should mention that here’s yet *another* example of using an Ajax approach to building fast, browser-based capability so the user interface feels like a desktop application.