Big Data Analytics, House of Cards and Future of Television Creation & Consumption
There are some posts, posted earlier on this blog, which talk briefly on how Netflix used Big Data Analytics to come up with the theme, plot, characters, actors and other elements of their original series, House of Cards. (Post link1, link2). This post dives deeper into how Big Data Analytics and Algorithmic Programming will shape the future of T.V. Series. This partially ties with the discussion in class in Session 6 regarding how consumer behavior is studied, modeled and influenced.
The television consumption industry is changing at a very brisk pace. People now expect television to fit into their own schedules and it is predicted that weekly scheduled television broadcasts will soon be eclipsed by on-demand entertainment. (Leber, 2013).
There has been a lot of buzz and stir in Data Analytics and Television Media circles since the creation of House of Cards. Netflix knew that its original series would be a grand hit based on the data of viewing habits of its 33 million users (Leber, 2013). Based on these tagging and recommendation system Netflix could understand
- how viewers were enticed by political dramas (like British version of House of Cards)
- how large proportion of viewers liked watching actor Kevin Spacey in the direction of David Fincher
- how different trailers attracted the attention of different users.
and use this data for the creation of the series, which was made available all at once. Netflix was able to leverage its long lasting relationship with customers and using that data of their viewing habits they were able to place their bets and produce ‘House of Cards’.
How much and what kind of data does Netflix deal with?
To put the movie viewing habits of viewers in perspective, the number of people watching movies streamed online was higher than those watching them on physical DVDs in 2013-14 (Sandvine, 2014). Also a third of data buffered during peak hours was due to movies streamed online.(Sandvine, 2014). Netflix mines the data of around 30 million ‘plays’ a day, including when their users pause, rewind and fast forward, 3 million searches, 4 million ratings by Netflix subscribers as well as the devices on and time of day when shows are watched.(GigaOm, 2012). In addition, TV shows and movies on Netflix are annotated with hundreds of tags inserted by viewers to describe the genre, the tone, the action and the talent, among plethora of other things. While in the past, those tags were used to recommend other shows by Netflix from a long list of titles on the service, now Netflix is using them in the creation of original content because it knows what people want. (Carr, 2013). While looking back at the show created and application of data analytics, begs to ask a more fundamental question, ‘Can creativity be automated’?.
Steiner, in 2012, in his article titled the same, started with the premise that creativity can’t be copied by machines as creativity is believed to be the product of mysterious processes within the right side of the human brain but later tells that complex algorithms are moving into creative fields and proving that in some of these pursuits and humans can be displaced. This is already observed in the case of music composers and creative writing. He also mentions that David Cope, a professor emeritus at UC Santa Cruz, believes that technology is almost there in not only replicating and reusing good music but creating new music. (Steiner, 2012)
Cope has been weaving thousands of lines of LISP code into music-making algorithms for 30 years. At first this produced crude music unfit for public performance, but now his algorithms compose music that imitates masters like Johann Sebastian Bach so well that people can’t always tell the difference. Cope feeds music to machine-learning algorithms that create new compositions by changing and building on patterns it finds in existing music. But his latest algorithm, which he’s dubbed Annie, takes programmed creativity yet a step further. She decides on the musical patterns, the criteria, and ultimately, the path she takes to making music. “I have no idea what she’s going to do sometimes,” Cope says. “She surprises me as much as anybody.”
Implications on Television Content Creation and Consumption
As companies like Google, Netflix, and Amazon, which know more details about our detailed watching habits, start to become more dominant forces in the creation of original programming, they could start also shaping creative decisions of directors and writers as well (Leber, 2013). In the years to follow all aspects of television content making including screenplay may be written to meet the whims of data-driven media streaming companies. Also, consumption, as mentioned earlier in the article, is believed to entirely move from broadcast schedules to on – demand viewing.
IBM defines ‘Big data analytics’ as
the use of advanced analytic techniques against very large, diverse data sets that include different types such as structured/unstructured and streaming/batch, and different sizes from terabytes to zettabytes. It has typically has one or more of the following characteristics – high volume, high velocity, or high variety. Big data comes from sensors, devices, video/audio, networks, log files, transactional applications, web, and social media – much of it generated in real time and in a very large scale. Analyzing big data allows analysts, researchers, and business users to make better and faster decisions using data that was previously inaccessible or unusable.
- Leber, J., 2013. “House of Cards” and Our Future of Algorithmic Programming [Online] Available at: http://www.technologyreview.com/view/511771/house-of-cards-and-our-future-of-algorithmic-programming/ [Accessed 17 10 2014].
- Sandvine, 2014. Global Internet Phenomena. Available at: https://www.sandvine.com/downloads/general/global-internet-phenomena/2014/1h-2014-global-internet-phenomena-report.pdf/ [Accessed 17 10 2014].
- GigaOm, 2012. Netflix analyses a lot of data about your viewing habits. Available at: http://gigaom.com/2012/06/14/netflix-analyzes-a-lot-of-data-about-your-viewing-habits/ [Accessed 17 10 2014].
- Carr, D., 2013. Giving Viewers What They Want. Available at: http://www.nytimes.com/2013/02/25/business/media/for-house-of-cards-using-big-data-to-guarantee-its-popularity.html [Accessed 17 10 2014].
- Steiner, C., 2012. Can Creativity Be Automated? MIT Technology Review June 2012.
Very interesting blog post. Now that Netflix is starting offering original content, it will be fascinating to see how their content compares to that of traditional broadcasters. Obviously , these original broadcasters lack the data and the extensive algorithms that Netflix has constructed over the years. However, I wonder if this data driven content can have a certain natural upwards limit. I wonder if past preferences and choices are always the best predictor for future needs, especially for something so intertwined with culture. I would say that film and television are in particular fields where visionaries can have a large impact on what people want to see. As the research of Robert E Kennedy (2002) shows imitative television content has often lower returns than differentiated content, and thus it might be more valuable to introduce new shows. I mean it was Gladiator as first new historical epic that took the money and the accolades, while the imitators such as Alexander and Kingdom of Heaven took large hits, even though they used than popular actors such as Orlando Bloom. How would this system see trends that have not happened yet? Lastly, I assume this data can only be collected from the films and shows that are in their library. The data might show that a lot of people like Kevin Spacey, because a relatively big part of his filmography is in the current Netflix library. Netflix has of course big fluctuations in their third party content, especially now that certain license holders are trying out their own networks. Their projections might thus become slightly biased, and show the preferences based on past available content not the actual preferences of members. Now obviously, Netflix has a very valuable, and probably profitable, value proposition, but it might not be the only option in the future. What do you think?
Kennedy, Robert E. “Strategy Fads and Competitive Convergence: An Empirical Test for Herd Behavior in Prime‐Time Television Programming.” The Journal of Industrial Economics 50.1 (2002): 57-84.
A very well structured comment. Personally, I believe that that data driven content will have a limit as your argued. However the recent research is counter intuitive and pointing in the opposite direction. In the article titled, ‘Can creativity be automated’ author Steiner mentions that these algorithms have also been able to capture the trend apart moving from mere replication. So unless there is something which is absolutely disruptive in the taste of people, if we are able to predict the trend of change then it will quite possible to create content suitable for that.
Also another way of looking at this would be to assist movie creators by telling them about content that is not suitable. Even if some one wants to create differentiated content there can be suggestions on some part of the film/music which definitely will not be liked by major section of audience.
Netflix bias may definitely be a factor for analytics engine in Netflix, but consider what will happen if the likes of Amazon, Google and Facebook start doing this. Amazon owns IMDB and irrespective of any license people will continue voting on IMDB. They have lot of reviews and data and if they can generate insights out of that wont that be valuable for years to come. It is mind boggling to see the ways in which the creativity industry will be influenced by technology.
Let me know your thoughts.
Steiner, C., 2012. Can Creativity Be Automated? MIT Technology Review June 2012.
A very interesting post! And I do agree with most of your arguments! I, however, wanted to touch upon a different, yet, related topic.
So, in the past, publishers and authors had no way of knowing what happens when a reader sits down with a book. Now, e-books are providing data behind the sales figures, revealing not only how many people buy particular books, but how they read them.
The major new players in e-book publishing—Amazon, Apple and Google—can easily track how far readers are getting in books, how long it takes them and even which search terms they use to find books.
For example, Barnes & Noble, which accounts for 25% to 30% of the e-book market through its Nook e-reader, has recently started studying customers’ digital reading behaviour. Data collected from Nooks reveals, for example, how far readers get in particular books, how quickly they read and how readers of particular genres engage with books. They have determined that novels are generally read straight through, and that non-fiction books tend to get dropped earlier. And these are just some examples which in fact are already shaping the types of books that Barnes & Noble sells on its Nook.
The privacy issue that is often raised in this topic is that e-book users should be protected from having their digital reading habits recorded. “There’s a societal ideal that what you read is nobody else’s business”
So it is rather interesting to see how these trends develop in both industries, as although somewhat similar, these two industries still reflect quite different habits of people. Finally, the privacy issue raised above might have a big influence on these developments. What are your thoughts?
Alter, A. (2014). Your E-Book Is Reading You. [online] WSJ. Available at: http://online.wsj.com/articles/SB10001424052702304870304577490950051438304 [Accessed 20 Oct. 2014].