If you want to predict the future, create it ! How Amazon keeps selling you stuff


The story

Amazon is without doubt one of the world-wide giants of retail e-commerce. With $207 billion turnover generated by the retail operations in the U.S. and abroad, it accounts for nearly 50% of all online sales in the U.S. in 2018.

The Challenge

As a global e-commerce platform, with a never ending range of products from different categories, the main challenge is to show to customers relevant items, quickly. But how to do that on a client-by-client basis thousands of times per second?

35% of @Amazon.com’s revenue is generated by its recommendation engine.


Might need AI and machine learning?

Schedule an AI clinics with our experts


The Solution

Amazon rolled out a recommendation engine that predicts what customers want and are likely to buy. By showing only such items, Amazon is able to create unique and personalised client experiences and dramatically increase conversion rates.

Amazon uses a variety of signals to recommend items including :

  • your browsing history to refresh your shopping "cravings" ;)

  • your purchase history to highlight new versions of products you already bought

  • graph models that identify items related to the ones you purchased

  • customer profiling to suggests items bought by customers similar to you also bought this

The Benefits

Amazon claimed that 35% of their online sales revenue was generated by their recommendation engine in 2015. A huge competitive advantage that is now available to everyone! It was news of last year that Amazon is providing access to the same recommendation engine technology on AWS where everyone can train and use it!