The Power Dynamics Behind the Netflix Algorithm

Currently, Netflix holds considerable power in the market as one of the biggest streaming platforms. Netflix is an American company that emerged on the market in 1997 and that, with 300 million active subscriptions, is the world leader for series and movie streaming. Being the leader in the market, Netflix and its algorithm pose themselves as the mediating platform between film and series makers and the audience, gaining a position of power over deciding what to show and to whom.

Netflix suggests a catalog of films and series to its users according to the subscribers’ preferences, using an algorithm that determines what content a user might be interested in. However, if the algorithm is often accurate, it fails to show many interesting productions that remain hidden on the platform. This algorithm might aid many producers but penalizes others who invest in their content to be published on the platform, but never receive a proper income. There are many reasons behind this, considering that algorithms are far from perfect and still fail to account for the complexity of the human mind. On the contrary, the current algorithm seems to suggest a standardized set of series and films to the consumer. Additionally, it is not entirely unthinkable to me that there might be political and social reasons that determine show suggestions and highlight specific productions included on the platform.

To understand the power dynamics behind the Netflix algorithm, and how it affects film and series production and the content topics, I explore the question: How does the Netflix algorithm affect views of films and series?

How can we begin to understand the power dynamics of the Netflix algorithm?

A foundational perspective to understand the power dynamics that shape the Netflix algorithm is McRobbie’s [1] analysis of the basis for the capitalization of art and artists. According to McRobbie, the condition of modern art is the result of a task force established by governments to change the shape of modern education and provide alumni with creative and artistic skills. Teaching “how to be an artist,” in other words, has favored the capitalization of the industry, where artists are now associated with businessmen and entrepreneurs, and whose skills are treated as marketable assets.

In the case of Netflix, the platform’s algorithm was developed through the joint work of developers who, in 2006, participated in the Netflix Prize to develop a successful algorithm that allowed Netflix users to be paired with the content that is most relevant to them. Creating a mathematical algorithm has sparked some criticism because drafting an algorithm based on cultural preferences gives culture a quantitative attribute, understanding culture through mathematical data that often fails to depict it accurately. [2] Additionally, Khoo [3] notices that algorithms are responsible “for their implications and real-world consequences, including discrimination and the reproduction of existing power structures.” In other words, algorithmic systems accelerate social inequality and discrimination due to their strong power in decision-making within the platforms in which they exist. It is therefore important to notice that what Netflix suggests is the product of complex mathematical datasets, which often fail to depict the complex shades of human preferences and behavior, and which can be biased by gender and race, as evidence suggests. [3] On this matter, it is essential to cite what has been observed by writer and podcaster Stacia L. Brown (cited by Khoo) who has noticed that Netflix was showing Black users a significantly higher number of shows featuring black cast, highlighting the platform’s attempt at making use of sensitive information to select the audience of shows.

Netflix's algorithm is not the only tool that the platform uses to attract subscribers. As declared by a former data scientist at Netflix, the platform often encourages new film and series production to conform to specific standards that are proven to receive positive audience responses. [2] This massive influence over the production of the platform series shows an extreme capitalization of culture and themes of audience interest, leaving little space for independent and divergent productions.

A Case Study of Netflix’s Top 5

To explore this issue through a real-world case, I am taking a closer look at Netflix’s section “Top 10 programmes in [country name] today,” focusing on the top five in Denmark.

First, it is important to notice that all ten titles are Netflix productions or productions of which the streaming rights are held solely by Netflix. This aspect might suggest interest in Netflix suggesting titles produced by the platform itself, creating a closed circle between production and consumption, and making it harder for independent productions to climb up the content views ladder.

Zooming in on the top five, we find five different Netflix series. I will list them in order of appreciation, with Adolescence in first place, followed by Caught, The Residence, Gold and Greed, and Million Dollar Secret. 

Adolescence is a British miniseries about the murder of a teenage girl. The show explores important themes such as gender and self-image, misogyny, and the educational system's failure. Caught is an Argentinian thriller miniseries about a journalist investigating the disappearance of a teenage girl. The Residence is a comedic mystery miniseries set in the White House during a state dinner. The series follows detective Cordelia Cupp as she investigates a murder, blending elements of political intrigue and humor. Gold & Greed: The Hunt for Fenn’s Treasure is a documentary series about the real-life search for a hidden chest of gold in the Rocky Mountains. Million Dollar Secret is a reality competition where twelve contestants live in a luxury estate, unaware that one secretly holds a $1 million prize.

Among the five series, only The Residence features a black actress in the leading role. In Adolescence, actors with a different ethnicity than white are featured (Ashley Walters, Fatima Bojang), but it is important to notice that they are not shown in the show's preview, despite their crucial presence on the show (Ashley Walters is in two out of four episodes of the series). Similarly, Caught and Million Dollar Secret feature some actors with Latinx or black backgrounds, but the actors shown in the promotional content are all light-skinned. Finally, Gold & Greed features almost only white people. These four examples resonate with what Khoo calls “promotional whitewashing,” which consists of showing series that feature mostly white actors without suggesting productions that entirely feature other ethnicities.

As Khoo points out, “Netflix produces original films and television shows based on viewing data collected from audiences.” Therefore, the platform’s algorithm simply exacerbates the power structures and social inequalities that exist in modern society.

As said, the algorithm is personalized by the platform, both according to single user preference (e.g., previously liked series, similar content) but also according to the broad interest of the so-called “taste communities,” which include Netflix users with similar tastes. [3] Another method that Netflix uses to track user behavior is by understanding the time content has been played by the same platform and how many times it has been paused or watched halfway. [2] Having this data is important for the platform’s sponsorship of new content, which, if it does not perform, can be taken down from the start page and be available only through the search function.

In light of my study, considering that the creative industry is capitalized, it can help us understand why it has become normalized that Netflix produces and decides what movies or series to show. Culture, including art, has become measurable and quantifiable through a certain amount of money, whereby artists become “human capital” similar to the working class, which was traditionally considered so. [1]

For content to be available on Netflix’s start page and sponsored, it must represent what the majority wants to watch. In my case, the equality between male and female targeting content is respected, while ethnic equality is lacking. As mentioned, one out of five productions features a black woman as the main character, and even in this case, the majority of the cast is white. This aspect is explained by the fact that Netflix produces what people want to see, as it earns from productions, too. Being featured in the algorithm becomes difficult if certain themes are touched, and the whole idea of cultural and social diversity falls towards homogenization of the content, which is made to resemble the viewer, not to challenge or enrich them with new things. 

As highlighted by Hallinan and Striphas [2], the algorithm influences more than 75% of people’s choice of watching, making it crucial for movie and series creators to be featured among the sponsored content of the platform. As Khoo noticed, only 20% of subscribers select a movie through the platform’s search function. This aspect shows that, for an independent production to become big on Netflix, it is important to make noise elsewhere and attract users and subscribers who will search for the series or movie. This can represent a way to tackle the issue of a movie’s fate, which is decided by Netflix’s algorithm.

Social Impact Analysis

There are many challenges to analyzing such a complex and broad phenomenon as the predictiveness of the Netflix algorithm. Firstly, it is important to recognize that drawing conclusions based on such a small sample as the one I analyzed cannot lead to a full and cohesive theory. Instead, what we can draw from my research is a starting point for further analysis. Secondly, since Netflix is a big multinational company with substantial lawsuits and meticulous decision-making, it is quite hard to make allegations about its algorithm. It is possible to get an idea of what Netflix is trying to do, but it’s difficult to prove certain things. For example, if it is true that black audiences are targeted with content featuring predominantly black casts, their voices must not go unheard. It is, however, difficult to demonstrate that such strategies are in place, especially after Netflix’s statements in which they declared that their algorithm does not make use of data on ethnicity.

Another issue, and probably the more crucial barrier for this Netflix analysis, is that Netflix is a private company. Given this instance, it is clear that the platform will do everything economically convenient for them, including sponsoring the content they produce. It would result in a financial loss if Netflix sponsored content other than the one in which they invested resources. On the other hand, it would be ethical for Netflix, the leading streaming platform on the market, to give equal opportunities to independent movies by adding a section on their start page where they sponsor some titles and try to offer diverse content.

Policymakers should encourage the promotion of socially diverse content by ensuring that algorithms are trained to show content that depicts a wider range of identities by featuring people from diverse racial, ability, and gender backgrounds as main characters. Besides, the platform should safeguard the user's data, ensuring that sensitive content such as personal ethnicity is not a key element for the platform, which can or cannot determine if content is shown. Policymakers should also ensure that the conditions of the precarity of the artist are not heightened by allowing independent artists and producers to be featured in the sponsored content, avoiding the inevitable (and unequal due to available capital) conflict of interest between Netflix and independent productions. In other words, Netflix should serve more as a discovery platform rather than a capitalistic movie-making industry, which owns and produces most of the content shown.

Conclusion

Theories and case studies have shown that Netflix tends to show a certain type of content, which mostly features white people and has fewer non-white actors. The platform has available content with casts of actors with different backgrounds, but—as in my case study—it tends to remain hidden from certain audiences. Since my selection is extremely limited, no generalizable conclusion can be drawn, nor can further speculations be made. What stands out, however, is that my case study observations are coherent with previous analyses of the phenomenon.

My study has shed light on the fact that Netflix tends to put itself in a relevant position in its productions, penalizing the work of independent artists who want to use the platform to grow their audience. However, due to the capitalization of culture, an issue widely debated in the art industry, it is obvious that Netflix will sponsor their productions, in which the platform has invested money.

In conclusion, this short analysis has shown how Netflix contributes to the societal logic of inequality, accentuating the disparity between different identities. It is important for algorithms to be trained differently so that they do not penalize movie and series makers and do not discriminate. Users, too, must become more aware of these logics, and expand their consumption patterns beyond algorithm suggestions by digging deeper into the platform content, searching for new content.

Written by Ilaria Mariani.

Cover photo by Atul Vinayak.

References

[1] McRobbie, A. (2016). Be creative: Making a living in the new culture industries. John Wiley & Sons. 60-89

[2] Hallinan, B., & Striphas, T. (2016). Recommended for you: The Netflix Prize and the production of algorithmic culture. New media & society, 18(1), 117-137. DOI: https://doi.org/10.1177/1461444814538646

[3] Khoo, O. (2023). Picturing diversity: Netflix’s inclusion strategy and the Netflix recommender algorithm (NRA). Television & New Media, 24(3), 281-297. DOI: https://doi.org/10.1177/15274764221102864