The article “How Netflix Reverse Engineered Hollywood”, brought up some very interesting information regarding the popular streaming website. It discussed how thoroughly Netflix has analyzed their movie library, how the system is organized into specific categories, and various other features of the company’s massive movie database. However, the idea that struck me most of all was Netflix’s ultimate purpose for all this complexity: To gain and maintain subscribers by tailoring the webpage to each individual user.
Near the top of my personal Netflix page are categories such as “Top Picks for Jonah”, that are designed to suggest movies or TV shows that might interest me. After browsing these categories, I noticed that only a few titles attracted my attention. Despite the company’s vast database and sorting algorithms, it has only partially drawn my attention with its carefully constructed display. How well can a computer-automated system reflect the tastes of its human users? According to the blog post mentioned in the article, Netflix likes to believe its system performs very well: “Members connect with these rows so well that we measure an increase in member retention by placing the tailored rows higher on the page instead of lower.” Their system may be highly personalized and specific, but it also has to deal with random factors, such as movies that a viewer watches that may not reflect their overall taste in movies and television. Even when Netflix uses human terms and means of organizing to develop these mini-genres, it can only predict our preferences so well.
Nevertheless, I agree with Madrigal in his statement that Netflix has created a truly staggering database and suggestion system. When compared to other personalized internet systems, such as Pandora and Steam (A massive video game hub), Netflix achieves a much closer approximation to my interests. While Pandora may come close to Netflix’s level of specificity in categorizing musical works, Steam uses a much broader system of categorization, and both can only somewhat satisfy my tastes with their recommendations. It is interesting to see how many internet companies value personalization, and how well their computerized systems can service a human user.
Pictured above is my Steam Discovery Queue, a personalized list of computer games that Steam thinks might interest me, similar to Netflix’s “Top Picks for Jonah” category. As you can see the game in the front is some sort of soccer simulation game, which would never be of any interest to me. However, the next two games I may end up trying at some point. The fact that such a random suggestion as “Football Manager 2015” appears in my personal queue shows how a computer-generated preference system has its flaws.