How “free” are our choices in the Big Data era? The example of Netflix
Written by Evangelos Farmakidis *
We have just finished the last episode of the new season of our favorite Netflix series and we have decided to go for a walk. Before we turn off the TV and get ready, another series is stirring our interest. It appeared in the trending and in the “choices for you” and it happens to be the kind of series we like. After a quick look we decide to watch only the Trailer to make sure it is to our liking.
Indeed it is! Going-out is canceled and a new episode marathon (Binge-Watching) begins.
How free are our choices in the era of Big Data?
Did we really want to stay at home and watch a new series or did we want to go out for a walk?
Maybe according to an artistic manner our will is significantly influenced and our decisions guided?
We will try to give a short, simple and understandable answer to the above questions.
The Netflix example is certainly not accidental. It is not as a coincidence the arrival of the new proposed interesting series on the screen of our television.
Netflix today counts 137 million subscribers in 190 countries and owes much of its success to Big Data. The data analysis is a practice the company has implemented since the early years of its establishment, when streaming service was unavailable and Netflix provided its customers exclusively with services such as sending DVDs to their home by regular mail. Studying the preferences of its clients, it proposed films that might be of their interest.
In doing so, it wanted to increase its income while simultaneously facing up the problem that arose whenever a film won an Oscar or a famous film critic wrote an enthusiastic review for a film: the demand for that film was booming and as a result Netflix can’t cope with the rising demand and on the other hand older movies were not chosen by the customers so the company loses more revenue.
So it had to find a way to turn its clients into less famous or older movies. For this reason, it developed a prediction algorithm, which called Cinematch, to propose to its users new movies based on other users’ preferences.
Later, in October 2006, wanting to improve the algorithm’s performance, it launched an open competition, the Netflix Prize. The team, which would be able to improve the algorithm’s results to a satisfactory degree would win a $ 1 million cash prize. That competition drew the attention of the entire global community, brought together more than 40,000 teams of experts (in the fields of mathematics, statistics, information technology, etc.) from 183 different countries around the world. For that purpose, researchers got access to ratings and reviews 500,000 users of Netflix.
It took three years to achieve the desired result and the prize was finally awarded on 21 September 2009. The winning algorithm was the result of a consortium of 4 teams called BellKor’s Pragmatic Chaos and improved the results of the existing algorithm by 10.06%. Today, the result according to which we receive proposals, exceeded at 85% success rate.
Keeping us busy with a constant stream of suggestions, Netflix manages to renew our subscription every month. If there were not such proposals, it is likely that after the end of the last season of our favorite series, we would cancel our subscription, at least until the new season of our favorite serie is released.
A typical example of Netflix’s personalized commercial practice is the following: To promote its – perhaps most famous – series, which established the company, House of Cards, different trailers with different versions of the same series were shot, but aimed at a different audience groups, depending on their preferences. Thus, each trailer featured a different trailer for exactly the same series, according to their preferences. For example, drama lovers saw a more dramatic version of the series, while adventure lovers saw a more adventurous version of the series etc.
Netflix today process various data from its users, such as age, gender, geographic location, information about their computer or other devices used to access the service, programs they has watched since their registration, days and times associated, the history of their searches, and even the way they scrolled while browsing. It still records every time they pause, go back to re-watch a scene, or pass a boring scene.
In science today, there is no universally accepted definition for Big Data. However, we can say that “Big Data” defines data, regardless of type, that have the following key characteristics: excessive volume, high variety, and high collection speed – even in real time – from multiple sources.
Data mining is the process whereby one obtains useful information through the proper processing of “raw”, unclassified, complex, and large volumes of data previously collected and stored on huge and vast databases.
The informations extracted from data is a powerful “weapon” in the hands of Marketers, who use them to promote products or even design new ones.
Today’s data, during the time of the 4th Industrial Revolution, have the same value as oil for the 2nd Industrial Revolution and steam for the 1st (The world’s most valuable resource is no longer oil, but data ).
As such, Netflix knows which of its programs to recommend us, but it also supports the production of new programs based on the preferences and habits of its users. Knowing exactly what its users prefer, it produces programs that are almost a success before they even turn around.
The use of these marketing methods is by no means reprehensible, neither it is, of course, the intention of the writer to condemn Big Data, which prove to be very useful in many areas of our life beyond commercial activity, such as in Medical Science.
On the other hand, the benefits for the informed consumer are numerous, as he is given the opportunity to make the right choices which will be to his liking and suit his needs while saving time and money.
The European Union has already recognized the value of personal data since the mid-1990s and has set up a specific legislative framework to facilitate their free flow and to protect the residents of its member states. Its last major legislative initiative is the adoption of the General Data Protection Regulation or more commonly known as GDPR. It should be noted that the new ePrivacy Regulation is expected to address, including other and the processing of personal data in electronic communications.
The importance of big data for the modern economics and the science of Marketing is unquestionable. After all, as Dan Zarrella has rightly pointed out, “Marketing without data is like driving a car with your eyes closed”.
However, consumers do need to be aware in order these practices to serve their interests , not by influencing their will, manipulating their decisions and defining their lifestyles.
- Mareike Jenner, (2018), Netflix and the Re-invention of Television. Palgrave Macmillan.
- Pant V., Yu E., (2018), Conceptual Modeling to Support the “Larger Goal” Pivot – An Example from Netflix. In: Buchmann R., Karagiannis D., Kirikova M. (eds) The Practice of Enterprise Modeling. PoEM 2018. Lecture Notes in Business Information Processing, vol 335. Springer, Cham.
- Kai-Ingo Voigt, OanaBuliga, Kathrin Michl, (2017), Entertainment on Demand: The Case of Netflix. In: Business Model Pioneers, Springer International Publishing.
- Jenkins J., (2017), Netflix. In: Schintler L., McNeely C. (eds) Encyclopedia of Big Data, Springer, Cham
- Roberts R., (2017), Live TV, Netflix, Amazon, the Universe!In: Mastering Media with the Raspberry Pi. Apress, Berkeley, CA.
- McDonald K. & Smith-Rowsey D., (2016), The Netflix effect: Technology and entertainment in the 21st century. London: Bloomsbury Academic.
- Amatriain X., Basilico J., (2015), Recommender Systems in Industry: A Netflix Case Study. In: Ricci F., Rokach L., Shapira B. (eds) Recommender Systems Handbook. Springer, Boston, MA.
- Mary J. Cronin, (2014), Netflix Switches Channels. In: Top Down Innovation, Springer International Publishing.
- Keating, Gina, (2012), Netflixed: The Epic Battle for America’s Eyeballs. Portfolio/ Penguin.
- Robert M. Bell, Yehuda Koren& Chris Volinsky, (2010), All Together Now: A Perspective on the Netflix Prize. CHANCE, 23:1, 24-29.
- S. Finlay, (2014), Predictive Analytics, Data Mining and Big Data, Palgrave Macmillan UK.
- Min Chen, Shiwen Mao, Yin Zhang, Victor CM Leung, (2014), Big Data: Related Technologies, Challenges and Future Prospects, Springer International Publishing.
- Hrushikesha Mohanty, PrachetBhuyan, Deepak Chenthati, (2015), Big Data: A Primer, Springer India.
* Evangelos Farmakidis is a member of Homo Digitalis, trainee lawyer, graduate of Master of Science in “Law and Informatics” of the Department of Applied Informatics, of the University of Macedonia and the Law School, Democritus University of Thrace, postgraduate student of Criminal Law and Forensic Sciences of Law School, Democritus University of Thrace, holder of a Diploma in Social Economy and Social Entrepreneurship and Accredited Ombudsman of the Ministry of Justice, Transparency and Human Rights.