Amazon: Everything you wanted to know about its algorithm and innovation
Amazon is such a tech colossus that it can’t help but make headlines. It’s the world’s largest retailer, by market cap. And its chief, Jeff Bezos, is becoming the world’s richest man.
Behind Amazon’s success is its algorithm and innovation.
Since Amazon’s early years, Computer Society researchers have been studying the e-commerce juggernaut and the strategy behind its recommender system, the tool that has changed the way the world shops online.
Our studies into Amazon rank among the most read content in our oeuvre of research and articles, which amounts to more than 650,000 in total, assembled in the Computer Society Digital Library. In fact, a search on “Amazon” yields 14,194 hits in our database.
Here are the highlights of the peer-reviewed research and other articles into Amazon, beginning with a report this year that is now our most popular piece of any content in almost three years.
A recommendation algorithm like no other
The beauty of Amazon’s algorithm is how it uses far less data space, by up to three orders of magnitude.
“Nearly two decades ago, Amazon.com launched recommendations to millions of customers over millions of items, helping people discover what they might not have found on their own.
“Since then, the original algorithm has spread over most of the Web, been tweaked to help people find videos to watch or news to read, been challenged by other algorithms and other techniques, and been adapted to improve diversity and discovery, recency, time-sensitive or sequential items, and many other problems,” writes Microsoft data scientist Greg Linden and Amazon recommender system expert Brent Smith in a 2017 study. Linden used to work at Amazon.
The result has been an algorithm that “scales to hundreds of millions of users and tens of millions of items without sampling or other techniques that can reduce the quality of the recommendations,” Linden and Smith writes in in their “Two Decades of Recommender Systems at Amazon.com” in IEEE Internet Computing magazine.
A “Test of Time” winner
Linden and Smith were working at Amazon.com along with Jeremy York in 2003, and they together authored a study that was just declared a “Test of Time” winner by IEEE Internet Computing magazine in its 20th anniversary issue, in May-June 2017.
Their article highlighted how Amazon broke tradition when it began using its algorithm in the 1990s. Until then, algorithms were “user-based,” and they recommended the next purchase based on what people with similar interests and purchase patterns were finding.
Instead, Amazon devised an algorithm that began looking at items themselves. It scopes recommendations through the user’s purchased or rated items and pairs them to similar items, using metrics and composing a list of recommendations. That algorithm is called “item-based collaborative filtering.”
Our online shopping hasn’t been the same since.
Amazon Web Service is “the most trusted provider of cloud computing which not only provides the excellent cloud security but also provides excellent cloud services,” according to researchers at ITM University in Gurgaon, India, who in their 2015 paper urged cloud computing security as a core operation and not an add on operation.
Not just a bookseller
Amazon isn’t just the leader in infrastructure as a service (IaaS). The firm has decided to become a “software company” and bring forward its own portfolio of middleware (and more importantly DevOps tools) to address the platform as a service (PaaS) and tooling innovation gap, even if at the expense of commonality/portability with anyone else, David Bernstein of Cloud Strategy Partners writes.
Diving deeper into the algorithm
Navigating thousands of Amazon reviews before buying a product can be daunting. Chantal Fry and Sukanya Manna, both of California State Polytechnic University, Pomona, leveraged two flat clustering algorithms on Amazon review data: K-means and Peak-searching to perform clustering of product reviews based on topic.
The experimental results show that K-means clustering performs better than Peak-searching clustering in terms of grouping similar reviews based on topics.
Building a better review system
Researchers with the Institute of Information Science, Academia Sinica, investigated the review system of Amazon.com and proposed a Review-credibility and Time-decay Based Ranking (RTBR) approach, which improves the Amazon review system by exploiting the credibility and time-decay of public reviews.
Our privacy, of course
Lizzie Coles-Kemp of Royal Holloway University of London and Giampaolo Bella of DMI, Universita di Catania, Italy write about how Amazon manages the security and privacy of users through the digital identities that they create with the popular web site.
About Michael Martinez
Michael Martinez, the editor of the Computer Society’s Computer.Org website and its social media, has covered technology as well as global events while on the staff at CNN, Tribune Co. (based at the Los Angeles Times), and the Washington Post. He welcomes email feedback, and you can also follow him on LinkedIn.