Exploring Netflix myths: Towards more media industry studies and empirical research in studying video-on-demand

Abstract

Using Netflix as a lens, this article identifies and unpacks three central interrelated myths – binge-watching, on-demand, and big data – surrounding global video-on-demand services. These myths are problematic because they make certain ideas about these services seem natural and self-evident, restricting our understanding of their role in culture and society. Moreover, these services provide little transparency and data access to evaluate their claims leading to a call for more media industry studies and empirical research. This article points to valuable avenues of inquiry, emerging from recent studies, that can serve as a source of inspiration for future research in this field.

During the past decade, video-on-demand (VoD) services have been positioning and repositioning themselves discursively against other media, while making claims as to their revolutionary character. Most often they do so in contrast to the broadcasting logics of television, emphasizing elements such as their data-driven decision making and the predictive power of their recommender system (Burroughs, 2018; Van Es, 2023; Wayne, 2022). As such, they have been understood as disruptive to established industry practices. Meanwhile, a shroud of secrecy envelops the operations of their data-driven systems and the evolving practices associated with them. This environment has proven fertile soil for the construction, perpetuation, and propagation of narratives about global VoDs. It allows them to define the terms under which they are understood. More specifically, these narratives centre on binge watching (Turner, 2021), on-demand (Van Es, 2023) and big data (Frey, 2021; Van Es, 2023; Wayne, 2022) and are presented as statements of fact that do not require further scrutiny. They constitute myths in the Barthesian sense and have, too often, been uncritically reproduced by media and academics. They are problematic as they tend to oversimplify how we think about these media. As such, they limit our understanding of how culture is produced, circulated, and consumed on VoDs, and Netflix in specific.

To facilitate critical engagement with VoDs, it is crucial that these myths are evaluated. To do so, this paper argues, requires the field amps up research from a media industry studies perspective and conducts more empirical studies. Confronting these myths is not to deny that ‘streaming logics’ (Burroughs, 2018) are reshaping broadcasting logics (and vice versa). However, as transformations unfold, their actual newness is proven far more complex, gradual, and non-linear. In other words, complexity and contradictions in these transformations are overlooked. While acknowledging that Netflix is a ‘zebra amongst horses’ (Lotz, 2021: 195), following Burroughs (2018), the VoD is used as a lens in this article due to its central role in producing and circulating discourses about the video streaming industry at large.

This article first examines the discursive positioning of Netflix and the confusion and ambiguity it has generated over the nature of the service. Academics, for instance, have struggled with the question of whether or not Netflix is still television (Jenner, 2018; Lobato, 2017, 2019; Lotz, 2022). It then explores three interrelated myths central to Netflix’s identity that have been produced and perpetuated in this process of positioning: binge-watching, on-demand, and big data. Subsequently, it advocates for more media industry studies and empirical research as important ‘checks and balances’ to examine these myths and ensure that Netflix cannot solely define the terms under which we understand its role in culture and society. This then leads to a discussion of the challenges posed by opacity and restricted access, exacerbated by the processes of datafication and algorithmisation. Subsequently, the article highlights various recent studies offering research directions that could be built and expanded on to get a better understanding of VoDs, also detailing steps that have been taken in my own research. The article also considers recent shifts in the field, noting how Netflix is increasingly resorting to more established industry practices revealing the limitations of these myths. The conclusion then advocates for connecting and elaborating on innovative methods to probe these myths. This article highlights the crucial role of innovative and collaborative research approaches in making these services more transparent and knowable. It aims to encourage constructive scrutiny of the myths surrounding VoDs and by presenting initial efforts made towards this goal, it seeks to inspire further exploration and examination in this field.

Netflix’s discursive positioning

When Netflix entered the online streaming landscape, Mareike Jenner (2018) notes it strategically marketed itself as television with its CEO Hasting claiming they wanted to ‘become HBO faster than HBO can become us’ (Jenner, 2018: 5). In Netflix and the Reinvention of TV (2018), Jenner explores Netflix as ‘dominant challenger to linear television, viewing practices, nationalized media systems and established concepts of what television is’ (Jenner, 2018: 3; original emphasis). However, the positioning of Netflix in the media landscape has been more complex. Ramon Lobato (2019) observes how Netflix combines different technologies and institutions. He argues that Netflix has many faces and that it is ‘a media object that perfomatively enacts its association with these media at different times and for different purposes’ (Lobato, 2019: 43). As he explains, for the government it presents itself as a digital media service to navigate regulatory concerns. In public relations, leveraging consumer familiarity, it postures itself as television while its interface design suggests a cinematic experience. In terms of its recommender system, however, it promotes itself as new media.

Scholars have also been grappling with the question ‘what is Netflix?’ (Lobato, 2019: 19) and have been trying to establish which conceptual frameworks they can use to study it. Perhaps most telling in this regard is how Amanda Lotz (2017) initially insisted that Netflix was internet television. Later, however, she stated that her claim of Netflix as television was an ‘emotional response’ and that ‘Although there is considerable formal similarity in the content commissioned by linear ad-supported channels and SVODs, the nature of this content, the strategic purpose of the content to different services, and the strategies by which they pursue that purpose are critically different’ (Lotz, 2022: 117). Nowadays, she gives preference to describing Netflix as a form of internet-distributed video.

Similarly, scholars (Johnson, 2019: 30; Lobato, 2019: 36; Poell et al., 2022: 6) have underscored the importance of distinguishing Netflix from the concept of platforms. Their work highlights that Netflix lacks direct economic and infrastructural accessibility for third parties. This distinction raises a pertinent question: can user-generated forms of video, which bypass traditional industrial modes of production, be considered television? Johnson (2019) refines the distinction, shifting focus from modes of production to the nature of content acquisition. She points out that content acquisition can fall into two categories: passively acquired, as observed on platforms like YouTube, where users upload content to the platform; or editorially selected, as is the case with Netflix, which produces or licenses content.

In the end, how Netflix is understood depends on the specific perspective from which one is looking (Johnson, 2019: 32). Moreover, as Lobato rightfully points out, ‘what is more important than what we call Netflix is how we think about it’ (Lobato, 2019: 44). For the analysis of Netflix, Lobato advocates combining multiple analytical approaches, merging insights from television studies, digital media studies, and platform perspectives. In this context, the intervention and critical examination of myths become urgent as they have a profound impact on our thinking about VoDs and traditional television, constraining our understanding in significant ways. The three myths discussed in this article have been central to how Netflix has characterized itself over the years.

Reflecting on these myths, the value of television theory is recognized because of the dichotomy it draws between the old and the new. As emphasized by Judith Keilbach and Markus Stauff (2013), the broader significance of television theory to media theory lies in its engagement with an object continually undergoing transformation and, consequently, redefinition. They compare television to an experimental system with ‘a heterogenous constellation of theories, objects, instruments and practices redefining each other constantly’ (Keilbach and Stauff, 2013: 83). In so doing they take seriously the suggestion of Lotz to speak not of television, but televisions (2014: 91) in the plural. As they explain, it is useful to consider how ‘post-network television often re-articulates already existing topics, problematizations, or supposed ‘potentials’ with different emphases and strategies’ (Keilbach and Stauff, 2013: 92). They argue it is useful to focus on ‘problematisation,’ involving debates and strategies around various issues, that help produce the rearrangement of assemblages. In what follows it is proposed that the myths surrounding VoDs centre on problematizations associated with attention, scheduling, and individualization, respectively. In the case of each of these problematisations, VoDs propose a solution.

Fuelling interrelated myths

Before discussing the myths associated in more depth, it is important to explain how the concept myth is understood here. As Roland Barthes explains, ‘Myth does not deny things, on the contrary, its function is to talk about them; simply it purifies them, it makes them innocent, it gives them a natural and eternal justification, it gives them a clarity which is not that of an explanation but that of a statement of fact’ (Barthes, 1993 [1972]: 143). Vincent Mosco (2004: 19) suggests that when technologies are new, a lot of hope is projected onto them even though their real social impact is only realized when they become banal. Referencing Barthes, he states that myths provide ‘[..] “euphoric clarity” by eliminating complexities and contradictions’ (Mosco, 2004: 30). Crucially, these myths don’t concern false beliefs per se, but rather involve a process of naturalization that impedes critical scrutiny. Importantly then, there is a need to examine the ‘facts’ presented by myths and to reclaim the complexities and contradictions they often eliminate.

In the following paragraphs, the three interrelated myths propagated around VoDs, and Netflix in particular, are explored: binge watching, on-demand and big data. Importantly, these myths are highly interconnected and have been disentangled for the sake of clarity. While by no means an exhaustive list of myths related to VoDs, these central myths have persistently circulated and been perpetuated in both popular media and academic discourse without receiving sufficient scrutiny. These myths are partially produced by the industries themselves through what has been termed ‘streaming lore’ (Burroughs, 2018), which encompasses the everyday sensemaking carried out by industry professionals. The ideas circulating within this context contribute to the formation and perpetuation of these myths. As is discussed, the myths around VoDs often reaffirm established myths around traditional broadcast television that Caldwell (1997) has written about: distraction, liveness, and human expertise.

Myth of binge-watching

First, to address the myth of binge-watching connected to the problematization of attention. Binge-watching is made possible because VoDs release entire seasons of content at once. It is, as Jenner states, ‘how content is supposed to be watched on Netflix’ (2018: 109) and has come to stand for all audience consumption practices. Netflix has presented binging as ‘the new normal’ and implied it constitutes a ‘more attentive viewing practice’ (Tyron, 2015: np). This is because releasing seasons all at once ‘conditions the audience to consume in specific ways that elevate the status of the show to “complex”’ (Burroughs, 2018: 9). Binging challenges the temporality of televisual content with its windows of opportunity for viewing, liveness, and schedules. Apart from being presented as a pure and natural way of watching content on Netflix, binge watching underscores the viewers autonomy affording them the flexibility to watch content when they want (Jenner, 2018: 115). It connects to another myth, namely one centred around on-demand, which will be discussed shortly.

In the process of differentiating itself from broadcast television, the myth of distraction has been reproduced as well. As Caldwell (1997) explains, viewers of television have often been seen as inattentive and distracted. In part, he states, this perception derives from the domestic context in which the medium is consumed. Moreover, as he claims, even updates to this theory provide that viewers are often occupied with other activities while watching television. Caldwell advocates that ‘Theorists should not jump to theoretical conclusions just because there is an ironing board in the room’ (1997: 27). This reasoning applies to VoDs too. The availability of entire seasons at once does not necessarily mean that people are predominantly binging on it (or that the content on offer is necessarily more complex).

Myth of on-demand

Second, centred on the problematization of scheduling, Netflix suggests that it offers viewers more control, freeing them from the constraints of broadcast television schedules (Jenner, 2018: 114). As Wayne points out, Netflix ‘Often emphasizes binge-viewing as a mode of audience behaviour that improves upon traditional television’s liveness and linear scheduling’ (2022: 195). This was perhaps particularly evident in 2017 when Netflix, as part of an April fools’ prank, introduced the series Netflix Live (2017) with Will Arnett. In the series, Arnett narrated mundane activities such as microwaving, photocopying, and toasting bread, humorously poking fun at liveness, long seen as a central trait of broadcast television (cf. Van Es and Keilbach, 2018).

Zooming in on traits associated with liveness (e.g. slow motion, direct address, replays, narration), Netflix Live (2017) portrayed broadcast television as an exceptionally boring and inferior form of entertainment. The series was jokingly concluded with its cancellation after the first episode stating that Dave from analytics had miscalculated interest. Herein it also strategically capitalized on its Big Data reputation, which is the third myth that will be discussed. Overall, Netflix positioned itself with the prank as a new and improved form of television (Van Es and Keilbach, 2018).

Here again, in the process of differentiation, an earlier myth around broadcast television is reaffirmed, namely the myth of liveness. As mentioned, liveness has long been seen as one of the defining characteristics of traditional broadcast television. Caldwell refers to the ideology of liveness as a ‘theoretical obsession’ of television studies and a ‘paradigm that simply will not die’ (Caldwell, 1997: 27). He finds the myth to be problematic because the overestimation of the centrality of liveness in television theory has been at the expense of the analysis of other modes of practice and production. While liveness has been used by television as a mark of distinction, he explains, very little programming actually supports this claim.

Myth of Big Data

Third, related to the problematization of individualization, Netflix feeds the myth of big data perhaps best captured in the early buzz surrounding House of Cards (2013-2018) and the VoD’s ability to predict success (Frey, 2021: 109–112). As explored elsewhere (Van Es, 2023) notions of magic and wizardry have long surrounded the service, extolling the belief that data possess the capacity to capture and know all in an objective manner. The Netflix recommender system is portrayed as accurately discerning human tastes providing viewers exactly what they want. In Finn’s words, ‘for Netflix, the brand is algorithmic, the magic computational, and the humans are kept carefully out of the spotlight’ (2017: 94). Netflix consistently underscores the extent to which its content is consumed due to the effectiveness of its recommender system. Burroughs (2018) explains how Netflix guarantees their audiences that they will get content that they love because of their use of data and algorithms. In contrast to Nielsen Media Research, which relies on a representative sample for measuring viewing behaviour, they collect vast amounts of data from all users, asserting a fuller understanding of audiences and their preferences. It is also claimed that with VoDs, demographics of subscribers are no longer important (Lotz, 2022).

Netflix, however, doesn’t disclose much data about consumption nor information about the workings of their algorithms. As Wayne comments, ‘This public discourse about proprietary data provided Netflix opportunities to differentiate itself from its linear competitors, explicitly criticize traditional industry practice, and deflect criticism about programming decisions’ (Wayne, 2022: 194). This also ensures that assertions of success are indisputable, while any shortcomings remain out of sight.

In cultivating its data-driven image, Netflix underscores a distinction from traditional industries, thereby reinforcing an existing myth centred on the human. As claimed elsewhere (Van Es, 2023), Netflix establishes a dichotomy between data and humans, implying that they possess opposing characteristics and epistemic potentials. However, there has been a notable shift in their narrative. While continuing to celebrate the virtues of data and algorithms, Netflix increasingly lays claim to the role of human expertise and creativity in their processes, presenting themselves as both science and art. Nevertheless, the overarching narrative continues to position Hollywood and television studios as dinosaurs. They rely on fallible human intuition that cannot predict or cater to the (individual) preferences of viewers. This remains a central point of contrast for Netflix.

Resisting simplified thinking

They ways in which VoDs are envisioned by the industry (cf. the myths) is reproduced in and by media and circulated also in academic publications. Problematically, harking back to Lobato, it limits how we think about it, our understanding of the role of Netflix in culture and society. Consider, for instance, binging. Graeme Turner (2021) has claimed the concept is not useful in television studies. As he explains:

Having served the industry as a means of branding changes to the practice of delivery, and television studies as a placeholder for the closer analysis of changes in consumption, not only must we define binge-viewing more precisely but we also need to develop new descriptors for the various behaviors and practices which do not fit that definition but which are allowed to be subsumed under the term. (Turner, 2021: 231).

Turner is thus concerned with the fact that we do not know what is happening as people are watching. This, he provides, requires ‘more located and nuanced observational accounts of the evolving “cultures of use” within consumer households to develop more accurate and usable set of terms to describe what is actually happening in the domestic spaces as people watch television’ (Turner, 2021: 229). This is rather like the criticism voiced by Caldwell on glance theory – which describes television viewers as inattentive and distracted. Turner recognizes some work being done on audiences and their consumption practices, but he finds there is also a need for more ‘directed or located observation’ (Turner, 2021: 230) rather than surveys, interviews and focus groups alone.

In Netflix Recommends (2021), Mattias Frey sets out to address the myth of big data. Specifically, his focus centres on misconceptions about the algorithmic recommender system. As he writes, ‘Largely untested and sometimes speculative in nature, they are proving to be out of step with the available empirical evidence. How might we better understand algorithmic systems by not stipulating that their “revolutionary” features blow up long-standing user norms a priori?’ (Frey, 2021: 8). He calls for empirical investigation. By adopting a more evidence-based approach, he suggests, a more nuanced understanding of algorithms can be provided.

More industry studies and empirical research

The three myths created and perpetuated by VoDs – pertaining to binge-watching, on-demand, and big data – prompt a whole host of questions. Some of these issues are now gradually being addressed in scholarly research, but they require much more systematic attention in transdisciplinary collaborations:

• What makes up the library of the VoD catalogue?

• How is content circulated? (interface, workings of recommender system)

• What, how and for how long are users watching content?

• What content is popular/culturally significant?

• How are data used by these services for producing and distributing content?

Instead of sticking to the simplified realities of myths, the field need to tackle these sorts of questions and in doing so, further our understanding of these services and other televisions. In other words, it is important to challenge industry-sourced information. To achieve this, a greater emphasis on media industry studies and empirical research is needed. However, this work is being hampered by issues of opaqueness and access.

Issues of opaqueness and access

More specifically, what is needed to combat these myths has been termed critical media industry studies. This involves a keen interest in the study of ‘microlevel industrial practices’ and laying bare the role of human agents in ‘interpreting, focusing, and redirecting economic forces’ (Havens et al., 2009: 236). It concerns methods that concentrate on scrutinizing industry discourse and practices to reveal the commonsense assumptions that inform their practices. As stated by Caldwell with regard to ethnographic research, it ‘can provide rich insights that speculative theorizing misses’ (Caldwell, 2013: 63–4), yet he also points out that the industry is not very transparent or accessible.

With VoDs it has, however, become even more difficult to understand how culture is produced, circulated, and consumed. More concretely, datafication and algorithmization are processes that have created additional layers of complexity to understanding these services and further cloak the activities of industry. Answering some of the questions about VoDs requires ‘hard’ quantitative data as well. As discussed above, Netflix and many other VoDs limit the release of data about consumption on their services. As Wayne explains,

Whatever shared sense of collective audience-hood was discursively produced by widely accepted albeit thoroughly flawed ratings systems like Nielsen is being replaced by the industrial discourses associated with black-box audience data within which claims of popularity and cultural significance cannot be substantively challenged. (2022: 194)

Moreover, in contrast to published TV guides or cinema schedules, there is still a lot unknown about what makes up the libraries of these VoDs. Also, the how, when, and why of users exposed to certain content on the service, driven by (algorithmic) recommendations, remains unclear.

Within critical algorithm studies, Nick Seaver reflects and tackles some of the challenges involved in this type of research. In Computing Taste (2021), Seaver conducts an ethnography of music recommendation algorithms. Pushing back against the idea of access as a singular event, he proposes to address it as a texture characterized by ‘patterns of disclosure and refusal’ (Seaver, 2021: 15). For instance, he highlights a topology of knowing and not knowing within companies. Nobody knows everything. Furthermore, he cautions against assumptions that secrecy necessarily implies relevance or signals something that needs to be known. Seaver suggests that much of what critical scholars of algorithms want to know – how they are thought about and understood – is accessible by building relationships within the industry.

These lines of inquiry and discussions about the related challenges are not entirely new. For example, in 2017 Lobato pointed out to the difficulties is studying platforms as opposed to television channels. He argued the need for a research agenda round Netflix catalogues to study content diversity and global media flows. Similarly, Lotz (2021) points to challenges in the study of VoDs, particularly around the investigation of cultural power. She highlights the lack of publicly available data about content consumption and emphasizes Netflix’s business model, which is directed at providing individual satisfaction rather than appealing to a broad audience. For interrogating cultural power in VoDs, Lotz suggests ‘raising broader questions about the intersection of texts and industrial practices’ (2021: 896) with the goal of deeper theorization of these services. However, until now academics have tended to target only specific questions and challenges around VoDs. The broader and more fundamental challenge of opacity and access and the ability of a service like Netflix to define itself on their own terms needs to be put on the agenda.

The following section offers a mapping of the different methods and approaches that are already being used to study VoDs that can serve as inspiration. While laudable and interesting efforts, they are only first steps that have been taken that allow us to evaluate the narratives propagated by Netflix. Mirroring the questions raised above about VoDs which stem from the discussed myths, these have a focus on library analysis, content circulation, viewing behaviours and data practices and use. However, as is apparent, these approaches have limitations that require attending to. Moreover, as insightfully noted by Lobato et al. ‘Empirical research on VODs has been siloed till date […] As a result, certain aspects of catalog-interface-audience interaction remain under-researched’ (Lobato et al., 2024: 1342). Getting a better understanding of Netflix is going to require innovative methods centred on interdisciplinary and transdisciplinary collaboration.

Library analysis

To counterbalance assertions of quality, complexity, and notions of control perpetuated in the myths of binge-watching and on-demand, it is crucial to examine critically the library of VoDs catalogues. In pursuit of this goal, library analysis studies are conducted. As mentioned, Lobato made a plea for this area of research several years ago. In the context of Australian Netflix, for instance, Lobato and Scarlata (2017) have conducted a manual content analysis of the amount of Australian content. Here they relied on counting the amount of content tagged Australian and then relied on trade press for the overall amount. By contrast, Lotz et al. (2022) conduct an analysis of 17 different Netflix libraries using large-scale quantitative analysis of proprietary data sets from third-party Ampere Analytics. This company provides metrics on subscriber video on demand services and their TV and film titles by country and service. They provide data on their own terms, not always scrutable to academics. Additionally, the high cost of acquiring these data makes them unaffordable for many researchers. More recently, scholars have turned to scraping the website JustWatch, a VOD metadata aggregator (e.g. Iordache et al., 2023) for information on the availability of movies and shows on streaming platforms. However, this data has not always proven to be entirely complete or reliable.

Content circulation

The myth of big data claims that Netflix knows us better than we know ourselves. To better understand media circulation (which includes content visibility and prominence), VoDs have been accompanied by the emergence of interface analyses in television studies (Johnson, 2019). These studies have looked at matters like how, with the shift from programming in traditional television to the use of adaptive agents in online platforms, the televisual flow has transformed (Uricchio, 2008), distinctions between linear/broadcast and non-linear/on-demand television (Johnson, 2017), and media circulation power (Hesmondhalgh and Lotz, 2020). Additionally, Van Esler (2021) highlights how the Netflix interface provides an illusion of viewer control while steering them towards original content. These studies typically focus on surface-level observations, providing limited understanding of the underlying data flows, algorithms and their effects on content visibility.

Then there have also been attempts at reverse engineering the recommender algorithms. Fatima Gaw (2022), for instance, is interested in in understanding how the Netflix algorithm constructs cultural taste. She offers a socio-technical analysis of the recommender systems, analysing Netflix documents to reverse engineer the algorithm, reflecting on user accidents and controversies with the algorithm and finally using the data from the first two to interrogate the configuration of the algorithm. To uncover how the NRS works and its impact on processes of taste-making, Pajkovic (2022) went in a different direction. He set up three Netflix user profiles with three distinct taste persona and traced them over time. While this work is interesting and important, it remains limited in scope and is highly context-dependent, raising more questions than it answers about how the recommender system works.

Frey (2021), for his part, provides valuable insights into how people engage with recommender systems through empirical studies of user reactions to these systems. More specifically, he conducted the quantitative analysis of two nationally representative surveys (UK and USA) and the qualitative analysis of thirty-four in-depth interviews. To a slightly different end, using the walkthrough method (Light et al., 2018) and in-depth interviews with Netflix users, Varela and Kaun (2019) explored the use and perception of Netflix’s recommendation algorithm. Both studies were important in that they countered techno deterministic ideas of algorithms as all-powerful entities. However, their focus on small samples of users in specific regions limited their capacity to generalize findings more broadly.

Viewing behaviours

Is binge-watching a pure and natural way of watching content on Netflix? What does it even entail? To understand viewing behaviours, including on streaming services, Lotz and McCutcheon (2023), for The Australian Streaming Stories Viewing report, conducted a survey (n = 2060) amongst Australians over the age of 18. It focused on scripted series and movies they conclude, amongst others, that viewing of streaming is less distracted and less personal than often thought. Interested in investigating the motivations, contexts, and affective states surrounding Netflix viewing, Castro et al. (2021) used a browser extension to log interactions and supplemented this ‘objective data’ with questionnaires. This work helped in understanding binge-watching as an individual activity connected to boredom, lasting on average a little over two hours. These insights, however, were based on the data collected from only 11 millennials over a ten-day period.

In a rather different vein, academics (Idiz et al., 2024; Scarlata, 2023) have interrogated the Netflix top 10 list over a period to reflect on ‘popularity.’ However, the Top 10 is less than ideal for research purposes in that it is ‘problematically repetitive and lacks a methodological transparency that can tell us anything more concrete about viewer behavior’ (Scarlata, 2023: np). Only recently, the European Audiovisual Observatory (Grece and Tran, 2023) published a report on SVOD usage in the European Union for the years 2022-3. Their findings, however relied on survey data rather than actual user interactions and covered only a limited selection of works.

Data practices and use

While Netflix boldly asserts the role of data and algorithms in the myth of big data, gaining insights into how VoDs such as Amazon Prime, Disney+ and Netflix influence production processes, industry structures, and content is crucial. To explore this, Rasmussen (2022) conducted interviews with European film and television workers to understand how they interpret and interact with streaming data. She also uses interface ethnography – meaning attending events in which the industry present itself to ‘the public.’ To a similar end, but based on observational participation and interviews, Keilbach and Surma (2022) provide insight into the role of data analysis in the creative process of script development at Videoland in the Netherlands. Both papers reflect on the challenges relating to dealing with Non-Disclosure Agreements signed either by the interviewees or the researchers and the need to navigate industry speak.

Discussion

The studies mentioned above are inspiring research pathways that contribute to exploring myths about VoDs. While important initial steps and offering valuable methodological directions to be further pursued, they have limitations. In addition to de-siloing empirical studies, to gain a deeper understanding of how VoDs operate and how users engage with their content, more interdisciplinary, and transdisciplinary, approaches are needed.

Regarding media industry research, universities are increasingly focused on achieving impact and fostering public engagement, encouraging collaboration with external parties. Collaborative research projects with industry can provide unique access to data, discourses, and practices (Schäfer et al., 2024). At Data School (Utrecht University), for instance, we collaborate with industries on shared areas of interest. Admittedly, this is usually easier with public-facing companies who are willing to be more open and not hide behind non-disclosure agreements. For instance, master students work on thesis projects in collaboration with the German public broadcaster ZDF on recommender systems. They are given access to anonymized interaction data from the VoDs and communicate directly with the data scientists working to improve and maintain the system. Such collaboration presents a dual narrative. On the one hand, it offers opportunities to help shape the digital society. However, on the other hand, it gives rise to tensions around academic integrity, encompassing liberties to publish findings and delineate trajectories of research inquiries. This dynamic raises questions about the role of academics within society.

Another possible path is research involving the utilization of data donations (cf. Araujo et al., 2022). Data donation projects capitalize on the European General Data Protection Regulation (GDPR), that grants citizens the right of access to their personal data from processors such as Netflix. In these projects, researchers encourage individuals to request their data traces from platforms and donate them to academic research. As part of a larger consortium dedicated to establishing sustainable infrastructure to facilitate such projects, the author of this article is conducting a pilot study focused on Netflix. The study received Netflix data donations from 126 subscribers in the Netherlands and aims to advance understanding of how people engage with content on the VoD. By employing computational methods, it can help create empirical insights into phenomena such as binge-watching, taste clusters and content popularity – topics that have captivated media scholarship. Data donations research has several limitations, including challenges in recruiting participants and concerns about sampling bias and representativeness.

While progress is starting to be made, both VoDs and broadcast television continue to transition. Broadcast television has responded to the emergence of VoDs by providing on-demand access to exclusive content online (Johnson, 2017). Nielsen has even initiated a weekly ranking system in 2020, highlighting the most popular VoDs based on minutes consumed. VoDs, for their part, have tempered some of their initially perceived ‘radical’ and ‘disruptive’ features. In 2021, Netflix announced an experimentation with their release format, opting for staggered releases over a month instead of dropping all episodes at once. Netflix, once positioned in discourse as the antithesis of traditional television, has ventured into live programming and introduced a ‘Basic with ads’ subscription tier at a reduced rate in some countries. This tier provides access to a limited library along with some non-skippable advertisements. The evolving strategies of both VoDs and broadcast television challenge the notion of stability in media forms and demonstrate that the trajectory of their development is hardly linear. It supports this article’s argument that the prevalent myths about Netflix oversimplify the company’s identity and its impact on culture and society.

Conclusions

This article has argued the need for more media industry studies and empirical research to understand VoDs. Such research plays a crucial role in examining the myths produced and circulated by VoDs. These myths offer us simple narratives about our media by eliminating complexity. Too often, these myths been uncritically reproduced, and their claims have been taken for granted.

Although current research shows promising progress, further exploration is needed, and several challenges still remain. Most often, as discussed, these relate to heightened issues of opacity and access. For instance, a limitation of interface studies is that they are still rather focused on the front-end, that is the visible tip of the interface. Additionally, retrieving data from commercial companies or scraping it from online websites raises questions about how the dataset was constructed and the extent to which the availability/disclosure of certain data points determines the type of research questions that can be posed. It also requires either computational skills – which are not always common in the humanities and social sciences – or sufficient funding to cover the high purchasing costs.

Addressing these obstacles calls for innovative and collaborative methods that combine both qualitative and quantitative approaches. In efforts to enhance accessibility, the preservation of academic freedom should be guaranteed. This includes establishing sustainable pathways for access to data that avoid reliance on commercial parties. The power imbalance between VoDs and academics in producing knowledge about these services needs to be addressed. To counter their opaque and inaccessible nature, emphasis should shift towards rendering these services more observable (cf. Rieder and Hoffman, 2020) to relevant parties such as journalists, academics, and governmental bodies. This allows, on the one hand, better theorizing and, on the other hand, better regulation and policy making.