How Netflix revolutionised television programming with Big Data
Netflix’s original business of delivering DVDs to homes on a monthly subscription model was disruptive enough for the times. However, it seems rather staid when one looks at how the company has transformed itself in less than a decade since it commenced streaming of television programming over the Internet.
Equally commendable is that the company is now using all the learnings from the enormous volumes of data provided by its subscriber base of over 50 million users to drive its original content strategy. By most yardsticks, the company been enormously successful with this data-driven approach, so much so that its application of Big Data is of as much interest to businesses as its carefully-commissioned programmes are for its users.
Data – the “killer app”
For any business, the ability to gaze at the proverbial crystal ball and fairly accurately predict the success of a product or service would be priceless. Netflix has become the poster child of accomplishment in this precise realm. The success of the Kevin Spacey-starrer ‘House of Cards’ in the US market was apparently such a foregone conclusion that the company did not even require a pilot of the programme before it was commissioned.
What might seem as a blind punt to the uninitiated was in reality a ‘no-brainer’ decision for Netflix based on the total clarity of the story the company’s data told. By knowing deeply the popularity of the original series in the UK, past consumption of movies and programmes featuring lead actor Kevin Spacey amongst its user base and the acceptance of the genre of content produced by the show’s Executive Producer, David Fincher, Netflix knew they had a winner on their hands much before the cameras began to roll.
The success of House of Cards was not a flash in the pan for Netflix’s original programming. Apparently, the company has already established a success rate of 80%, with the jury still out on one of its programmes, which is more than twice the success rate enjoyed by most television networks with their own programming.
Using the data
According to an article in The New York Times that quotes networking provider Sandvine, “a third of the downloads on the Internet during peak periods on any given day are devoted to streamed movies from the Netflix service. And last year, by some estimates, more people watched movies streamed online than on physical DVDs.”
It is reported that Neflix monitors over 30 million “plays” a day, which includes information such as when users pause, rewind or fast forward, ratings by its subscribers, over 3 million daily searches besides the now seemingly obvious factors like time of day shows are watched or on what devices.
Besides, there are hundreds of tags or meta descriptors that each unit of content is annotated with, which then become useful not only in personalising content or promotions, but also in content planning and commissioning. With over 50 million users and growing, consuming diverse types and volume of content while leaving behind ample data sets that point to their preference, the scale of Big Data that Netflix deals with can be imagined.
The challenges are many, says Chris Pouliot, Director of Analytics at Netflix in this interview, as the company works out what data to collect and how to perform analytics on the Big Data.
“My team does not only personalizations for movies, but we also deal with content demand prediction. Helping our buyer figure out how much do we pay for a piece of content…. The personalization recommendations for helping users find good movies and TV shows. Marketing analytics, how do we optimize our marketing spin. Streaming platform, how do we optimize the user experience once I press play. There’s a wide range of data, so there is a lot of diversity,” he says.
Writing for Wired.com, Phil Simon lends his perspective on Netflix being the ‘quintessential data visualisation’ organisation. Using a presentation by Jeff Magnusson and Charles Smith from the company at a Hadoop Summit as the basis, Phil contends that most organisations would not even know their customers half as well as Netflix does. “Through Big Data and dataviz, Netflix seamlessly delivers mind-boggling personalization to each customer. At the same time, Netflix can easily aggregate data about customers, genres, viewing habits, trends, and just about anything else. Equipped with this data, Netflix can attempt to answer questions that most organizations can’t or won’t even ask.”
An important but less-discussed consideration is the change in sociological dynamics likely as a result of the understanding of users facilitated by Big Data. Citing the example of Michael Jackson’s Thriller, Prof. Markus Geisler seems to indicate that Netflix could be the harbinger for altering sociological dynamics, which could be a significantly profound consequence of intimately knowing users’ interests and behaviours.
How Netflix has leveraged Big Data and analytics to literally reinvent the company for greater success holds a mirror to numerous companies in various other markets.
While a comparable service is yet to make its mark in the Indian market, the rapidly growing bandwidth and proliferation of smartphones indicate that such a day is not too far. In such an eventuality, the potential of using Big Data in servicing this gigantic market is not only tantalising, but may well be an unavoidable necessity.
It’s easy knowing what the audiences wants, rejects, and consume. The biggest indicator is social media chatter. In India, brand monitoring has become the proverbial anvil on which mettle is tested. Not to mention that various companies today offer affordable and insightful data crawling and web scraping solutions. Any company that has the vision to collect and aggregate data from forums, blogs and social media and uses it to improve content and programming is only poised to become the next Netflix and mine the vast market that India is.
Image Credits : onlinebusinessdegree