Page 70 - AC/E Digital Culture Annual Report

70
additions or new releases. Digitalization and the free circulation of content not only multiplied the number of access points, but also acceler- ated consumption. The work of curating and prescribing contents is more important than ever because users are demanding full customization. And time is of the essence. Netflix, for example, has stated that there is a critical range of time with its service of 60-90 seconds. This is the time the average user takes to decide what to see after entering the platform. Once this time is up, the chances that the user will wind up view- ing content is drastically reduced (Gómez-Uribe; Hunt, 2015).
Linking content with the right user ensures consumer satisfaction and an optimal experience overall on the platform.
For the “many products with a low margin” business approach to be profitable, it is essential for the provider to know its contents and its consumers’ preferences very well. And it must be capable of properly linking the two. This is done using the famous algorithms, mathematical formulas with which these services implement a kind of automated system of recommendations, refined by “machine learning” dynamics. These processes cross check the available content with all of the information the provider has from
its users to make a series of predictions. The objective? To highlight a selection of titles that might interest the user. Linking content with the right user ensures consumer satisfaction and an optimal experience overall on the platform. So much so that the user will want to return.
Anatomy of the content
Curiously, the main competitive advantage for Netflix has been its ability to include the human factor in this entire process, radically changing the recommendation concept as we know it. The company has focused on a type of reverse engineering (Madrigal, 2014), increasing knowl- edge of the content using metadata: technical and artistic information, genre attributes, and
an infinite number of semantic markers. Those responsible for assigning these labels are called “taggers”10, personnel hired by the company
to highlight all relevant information for each available content (such as the fact that it is a story with a happy ending, involves animals, or takes place in a specific period, etc.). This human and subjective veneer derives from an indexation of content never seen before that drives the emotional selection of content. The result has been to go beyond the traditional classification of genre (horror, drama, comedy or family,
etc.) with the more than 80,000 categories11
that Netflix uses to index the content (such as “Dramas featuring a strong female lead”, “Movies for getting over your ex” or “Green dogs and misfits”).
These processes are applied with very similar dynamics to the selection of images with which Netflix presents the content within its platform, which is dynamic and personalized. The aim is to choose the most effective visual stimulus to promote views for each specific target. How? By highlighting on each program mask12 the most appealing aspects for each user profile, such as specific leading actors, scenes, compositions, etc. This process is very useful for programs with very heterogeneous audiences.
Image: NETFLIX INTERFACE. Example of different graphic combinations in which the series Stranger Things may be presented
to users. Source: Netflix Tech Blog.
Spotify is another good example of a company that has invested in going beyond the traditional airtight categories of the music it manages. Its
THE NETFLIX MODEL’S IMPACT ON CULTURAL CONSUMPTION ON THE SCREEN · ELENA NEIRA
Digital Trends in Culture