Architecture behind Netflix

Architecture behind Netflix
Reed Hastings, CEO Netflix, and other top officials with the cast of Netflix Originals

Global video-streaming service ups the game by deploying cutting-edge technology to enrich viewing sensibilities


Sadiq Shaban

Published: Sun 22 May 2016, 2:19 PM

Last updated: Sun 22 May 2016, 4:26 PM

Paris - Netflix, the global provider of streaming movies and TV series, deploys some of the most fascinating and advanced technology behind the scenes to up the stakes of viewer experience. Earlier this year the service was rolled out in the UAE and 130 new countries around the world, bringing the total to over 190 countries.
Driven by a number of factors like algorithms the Netflix experience includes personalised ranking, page generation, search, similarity, ratings and other such indices. The way these algorithms operate globally, paves way for an improved ability to connect audiences worldwide with shows and content of their choice.
Built upon the bedrock of solid data, Netflix has been able to take the user experience to another level, making it richer and more tailored. Noted Reed Hastings, CEO, Netflix, "We do a really deep data analysis at Netflix to find how much a new programme would be viewed, and therefore how much budget we should put behind it."
By putting a premium on its recommendation software, the algorithms that gather viewing preferences and tastes, Netflix is able to suggest the perfect tag from its catalog. That is how, for instance, suggestions for action or comedy flicks on the platform are based. A similar principle applies to each category of suggested films/shows.
"Netflix aims to float the right options toward the viewer, based on their preferences, interests, and watching history, so that not even a moment that they browse goes wasted. As a company we work very hard to better our services. Since our global launch, we have attempted to make it even better," Netflix head honcho told Khaleej Times.
The architecture behind Netflix's technology, primarily it's A/B testing method, has been a game changer. It has helped the service to present different materials to different users - to see what is more popular. Granular in its scope, little details like what images people see when they check a particular title are factored in to enrich the interface.
Todd Yellin, Netflix's head of product innovation explained, "A typical member scrolls through 40 titles before logging out. From the thousands of titles available, Netflix has to priortise a few dozen on the top of each page. That is the main reason why your Netflix page will look different to that of your neighbours. We aim to put in front of you what you prefer."
With a back-end that works overtime to ensure a smooth viewing experience, Netflix monitors how many devices its members watch the content on, what time of day they catch a show, what day of week, how long they watch a particular programme for, do they abandon a programme for good or do they come back to watch it later?
"We have a number of take-aways from the company's A/B testing. Importantly we let data determine the design changes in Netflix. Gone are the days of asking people what they like to watch. At Netflix we have a better sense for our user base by looking at what people clicked on. This helps us bubble up recommended content, rather than labeling it. This is content served up as never before," Yellin explained. 


More news from TECH