Sage Journals: Discover world-class research

Abstract

Using the media industry studies approach, this article provides a history of the industrial discourses surrounding Netflix’s audience data. From Netflix’s entry into the streaming market in 2007 until late-2018, the company did not publicize information about viewership. During this time, executives’ public discussions of proprietary data are understood in relation to multiple organizational goals: differentiating the streaming platform from the traditional television industry, denigrating traditional television industry practices, and deflecting criticism. In late-2018, the company began selectively publishing viewership numbers for a small number of original titles to highlight the popularity of the platform’s original content. Although the company maintains its anti-transparency policies, the shift toward selective data releases has significant implications regarding Netflix’s relationship with the traditional television industry. This analysis concludes with a discussion of streaming audience data that situates in the emerging realities of ‘popular’ television in the context the medium’s broader transformations and continuities.

Keywords

media industries Netflix platforms streaming television video on-demand

Introduction

With more than 200 million subscribers in more than 190 countries, Netflix is the largest subscription video on-demand platform in the world. Yet, in television’s streaming age, the notion of popularity has become increasingly ambiguous. On one hand, platforms like Netflix collect and analyze audience data with detail and precision that was unimaginable 20 years ago (Hallinan and Striphas, 2016). On the other hand, these subscriber-supported platforms have little economic incentive to publicize their audience data (Lotz, 2017). This juxtaposition of advanced data capabilities with strict anti-transparency policies is frustrating for industry observers and media scholars alike. However, the conditions that produce this juxtaposition remain fluid. Television is (and always has been) a medium in transition (Uricchio, 2008). Beginning with this recognition, this research uses a variety of publicly available secondary materials to construct a history of the industrial discourses surrounding Netflix’s audience data.

The analysis starts with a brief discussion of the significance of audience data for US commercial television, some background regarding Netflix’s relationship to traditional television networks, and a description of the media industry studies methods employed in this research. The following sections detail Netflix’s public discourse related to its own audience data and television audience data more generally. Between 2007 and 2018, the company enacted strict anti-transparency policies. As a result, executives were often asked to comment on audience data that would not be released. 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. In 2019, however, the platform began releasing viewership data for a very limited number of ‘Netflix original’ titles. The public discussions related to these selective data releases primarily highlight the popularity of the platform’s original content which differs from earlier discourses that served a variety of organizational goals.

Although the release of any audience data is a significant departure, Netflix’s longstanding refusal to comment on viewers’ geographic location and its unwillingness to publicize data for most of its content library reflects the reality that anti-transparency policies are still very much the norm. Yet, in contrast to the company’s preferred image of itself as a revolutionary force in television, the shift from using public discussions of audience data to position itself outside of and superior to traditional networks to using audience data as proof of popularity is unmistakably a shift toward traditional industry practice. Nonetheless, Netflix’s ongoing resistance to transparency alongside the lack of comparative metrics for both traditional networks and other video streaming services has made it increasingly difficult to define popular television. Ultimately, then, 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.

Audience measurement and the television industry

Historically, ratings have been centrally important for the American commercial television industry. These ratings, produced by the supposedly independent Nielsen Company, function as currency in the economic transactions between advertisers and networks. Derived from viewing data collected from a ‘statistically representative’ sample of households using a variety of methods including time diaries and monitoring devices known as ‘people meters’, Nielsen ratings establish the cost of advertising spots during a given program or time of day on a given network. Rather than reflecting actual numbers of viewers, however, ratings ‘function as a social-discursive-economic convention, a streamlined currency that suits the needs and interests of various factions of the television industry’ (Hessler, 2021: 3). In his seminal analysis of television executives during the network era, Gitlin (1983) notes that virtually all industry activity was oriented toward producing ratings success. Even as the growth of cable and satellite television increasingly led to the pursuit of niche rather than general audiences, Nielsen ratings remained the central metric of success in a rapidly changing industry.

The only segment of the US commercial television industry less beholden to the ratings system is premium cable. Subscriber-supported networks like HBO and Showtime do not feature traditional commercial advertising. Instead, they generate revenue from viewers who pay an additional monthly fee (a premium) for access to these channels and from carriage fees paid by pay-television providers for the right to include these channels in their bundle. As a result, the definitions of success associated with these networks are largely divorced from the ratings system and advertising revenue. Instead, success is typically conceptualized in more abstract notions like critical acclaim, ‘buzz’, and ‘quality’. The network most closely associated with such alternative discourses of success is HBO which introduced its famous marketing slogan ‘It’s Not TV’ in 1996. Since then, the network has cultivated a brand identity as the home of quality television that drew on a wide range of its programming but was primarily centered on the shift toward producing adult, edgy, authored, and high-budget original drama series. While HBO’s brand was initially constructed through the promotional efforts of the network itself, and then increasingly depended on signature shows to stand in for the network, it also increasingly relied on critical acclaim and industry awards to support its claim to be the home for creative talent. As scholars have noted, however, HBO’s marketing strategy relies upon the long-standing marginalization of television and its audiences (McCabe and Akass, 2007). HBO is not only different from traditional television; it is superior to traditional television. Specifically, the network invokes the rhetoric of ‘quality TV’ to assert the medium’s progress and place it in the same cultural category as cinema or literature. This rhetoric has been very effective. Original series produced by premium cable networks have accrued more prestige that any other type of programming in the post-network era and helped redefine successful television in cultural and commercial terms.

Once it started producing original series in 2011, Netflix largely followed HBO’s model to establish its brand identity. For example, the company uses the discourse of quality TV in relation to technological progress and emerging modes of audience engagement. In working to define its ‘programming against traditional television’, Tryon (2015) argues that Netflix echoes HBO branding strategies by reconceptualizing ‘streaming as a more engaging form of television, one that exists on a technological and cultural cutting edge’ (106). Yet, however much Netflix relies on the discourses premium cable networks traditionally used to distance themselves from advertiser-supported networks, streaming technology places the service outside of the traditional television industry in ways that were simply not available to HBO or Showtime. In particular, the company often emphasizes binge-viewing as a mode of audience behavior that improves upon traditional television’s liveness and linear scheduling. Given the moral hazards historically associated with television and the contemporary moral panics associated with binge-viewing (De Keere et al., 2021), Netflix has much to gain through its association with binge-able programming.

As a distributor of television content with no need for traditional television distribution systems, Netflix’s decision to move into original programming brought them into direct competition with both subscriber- and advertiser-supported networks. Although founder and CEO Reed Hastings explained the company’s original programming strategy in terms of economic efficiency (Hastings and Wells, 2012), other executives were more transparent about the company’s competitive intentions. In a 2013 interview with GQ magazine, then-chief content officer now-co-CEO Ted Sarandos asserted that Netflix’s primary goal was ‘to become HBO faster than HBO can become us’ (Hass, 2013). This organizational strategy was confirmed later that year when the company released a document titled ‘Netflix Long Term View’ which began with the statement, ‘Over the coming decades and across the world, Internet TV will replace linear TV’ (Netflix, 2013a). Previous scholarship argues that Netflix’s approach to industry practice is usefully understood through the prism of the company’s competition with traditional networks. Regarding the policy of refusing requests to include other network’s promotional branding when subscribers watch licensed content, for example, Wayne (2018) notes that the platform’s user interface obscured other television brands to better position the service as the audience’s primary point of identification (735). Similarly, Netflix’s application of the label ‘original’ to the content distributed by the platform is much more liberal than traditional television networks. As Lotz and Havens (2016) note, much of what Netflix promotes as original is more accurately described as exclusive in a particular market. Building on such scholarship, this article examines Netflix’s public discussions of audience data and ratings in the context of these competitive dynamics.

Methods

To understand the industrial discourses related to proprietary streaming audience data, this research examines a variety of publicly available secondary data from sources including trade press articles, press releases, trade and popular press interviews, videos of industry roundtables, promotional appearances, and more than 10 years of Netflix’s quarterly earnings call transcripts and letters to shareholders (2010–2020). Regarding trade publications, searches for material related to Netflix’s discussion of audience data were conducted in prominent outlets such as Variety, The Hollywood Reporter, Deadline, and The Wrap. Although media industry trade publications and materials are a commonly used data source in media and communication research (Kosterich and Napoli, 2016), the use of these sources is hardly straightforward. As many scholars note, it is necessary to approach trade press material critically and to be cognizant of the inherent limitations associated with such data (Lotz, 2018; Perren, 2015). Nonetheless, in conjunction with other materials, the trade press offers substantive insight into industry dynamics and provides useful perspectives regarding the development of institutional processes.

Given this focus on audience data and related issues in the context of streaming industry discourse, a variety of marketing materials produced by Netflix have not been included in this analysis. Between 2015 and 2017, for example, the company introduced multiple marketing campaigns that included data-like audience numbers to bolster its image as a provider of highly desired content. In September 2015, for example, a Netflix press release accompanied by infographics proclaimed that the service knew precisely how many episodes of a given show a subscriber needed to watch before they became ‘hooked’ on that show (Netflix, 2015). The campaign included video promos featuring well-known television personality Bill Nye the Science Guy attesting to the scientific nature of Netflix’s methodology and the addictive properties of its programming. As industry observers quickly noted, however, the presentation of ‘data’ was fundamentally misleading as the marketing materials made no distinction between hugely popular shows like The Walking Dead (2010-) and critically acclaimed shows with niche audiences like Mad Men (2007–2015) (Van Der Werff, 2015). In the following years, similar campaigns used data-like numbers to promote various forms of streaming enabled consumption including the ‘binge scale’, ‘Netflix cheating’, and ‘binge racing’ (Netflix, 2016, 2017a, 2017b). As there was little industry discourse related to the marketing campaigns structured around such data-like numbers, this material is not addressed here. Netflix’s various other public uses of data, like the use of social media data examined by McNutt (2020), are not addressed in this analysis for the same reasons.

To analyze the collected data, this research employs the methodological approach variously described as media historiography, media industry studies, or trade press analysis (Corrigan, 2018; Lotz, 2018; Perren, 2015). This approach uses elements of the production culture (Caldwell, 2008) perspective to examine common meanings and values expressed by media executives. In addition to the use of trade press analysis as a means to access general television industry discourse, this analysis also follows Burroughs’ (2019) approach to understanding Netflix’s particular articulation of streaming discourse. In combining industry and discourse analysis, Burroughs uses notions of streaming to highlight moments of continuity and rupture characteristic of the ongoing negotiations between new and old media. In a similar fashion, this analysis identifies significant themes within Netflix executives’ public discussions of audience data and television ratings.

The findings generated by this analysis are presented in two sections. The first section addresses the time from Netflix’s entry into the streaming market in 2007 until late-2018 when the only official information about the platform’s audience was the number of domestic (U.S.) subscribers and the number of global subscribers (after 2010) released each fiscal quarter. During this period, executives simply declined to discuss the characteristics of the service’s subscribers or their viewing behaviors. When asked about the viewership for House of Cards (2013–2018), Netflix’s most successful early original series, Sarandos characteristically responded, ‘We’ve said publicly, and we’ll continue to say publicly, that we are not releasing viewing numbers’ (Netflix, 2013b). According to multiple trade press reports, secrecy was also internal practice as this information was withheld from high profile showrunners. Yet, it is precisely in the context of such strict anti-transparency policies that Netflix’s public discourses about viewership and ratings take on increased significance. They provide a window into the platform’s view of itself and its relations with the television industry.

The second section examines Netflix’s public discussions of audience data between late-2018 and mid-2020. During this period, the company selectively published viewership numbers for a small number of original titles while executives promised more transparency. The incorporation of selective data releases into Netflix’s broader anti-transparency policy has multiple implications. In one sense, efforts to demonstrate popularity with audience data is a traditional industry practice. In another sense, however, Netflix’s use of non-standard viewership metrics reflects the distance between the world’s most popular video streaming service and the US commercial television industry. This article concludes with a consideration of streaming audience data in relation to the emerging realities of ‘popular’ television.

Anti-transparency data policy and the practice of discussing without releasing

Prior to 2019, Netflix’s discussions of audience data can be understood in relation to three overlapping organizational goals: to differentiate the service from the traditional television industry, to denigrate traditional television industry practices, and to deflect criticism. First, when discussing audience data, Netflix executives frequently distinguish the service from traditional networks with claims that some industry norm is ‘irrelevant’ or ‘insignificant’ in the context of streaming television. When addressing ratings, for example, executives consistently state that traditional metrics of success are incompatible with the company’s subscriber-supported model. Discussing the platform’s anti-transparency policies in 2014, Sarandos said, ‘We don’t plan on releasing ratings since it’s irrelevant to us, we don’t sell advertising, we don’t negotiate carriage fees, there’s no reason for us to give a specific number of how shows are performing’ (Quintanilla, 2014). Sarandos also differentiates Netflix from networks with regards to the importance of traditionally valued audience segments. He explained, ‘Eighteen-to-forty-nine-year-old viewing is so insignificant to us. I can’t even tell you how many eighteen-to-forty-nine members we have. We don’t track them. It’s an advertising-driven demographic that means nothing to us’ (Villarreal, 2016). These comments are part of the company’s broader efforts to redefine the characteristics of successful television series in the context of transnational platforms and global audiences.

Netflix’s data capabilities further differentiate the service from traditional television. Rather than relying on demographic data generated by audience samples, digital technology allows the company to collect behavioral data for all subscribers’ activities on the platform. According to Sarandos, ‘With streaming, we have insight into every second of the viewing experience. I know what you have tried and what you have turned off. I know at what point you turned it off. It’s very sophisticated’ (Sarandos, 2012). This data allows Netflix executives to claim that the company ‘knows’ the audience in ways traditional networks do not. Discussing the relationship between the platform’s data capabilities and the company’s programming decisions, Cindy Holland, former vice president of Original Series, claims that Netflix targets content to specific ‘taste communities’ rather than demographic groups (Lynch, 2018). As a result, the service differs from traditional networks as Netflix executives like Holland ‘program to suit [the audience’s] tastes’ rather than their own as she implies other executives do.

It should be noted, however, that Netflix’s discourse regarding the relationship between audience data and executive programming decisions has changed over time. For example, in 2012, executives framed the multi-season commitment to House of Cards as a data-driven decision. According to then-Chief Communications Officer Jonathan Friedland (who was fired in 2018 for his repeated use of racially insensitive language in the workplace), the heavy upfront investment in the Kevin Spacey led political drama was justified by the platform’s audience data. He explained, ‘We know what people watch on Netflix and we’re able with a high degree of confidence to understand how big a likely audience is for a given show based on people’s viewing habits’ (Baldwin, 2012). By 2015, however, executives were openly acknowledging the limits of data-driven programming. At the Sundance Film Festival that year, Sarandos explained to media scholar Tim Wu that Netflix’s programming decisions were made based on both data and industry experience. The executive explained, ‘In practice, it’s probably a seventy-thirty mix. Seventy is the data, and thirty is judgment. But the thirty needs to be on top, if that makes sense’ (Wu, 2015). This shift away from data-driven decision making continued and, in 2019, Sarandos was quite explicit about the limited value of audience data. Regarding the choice to green light some shows and not others, he said, ‘The data doesn’t help you on anything in that process. Picking content and working with the creative community is a very human function’ (Hayes, 2019). Although the company’s framing of its programming decisions became more traditional in recent years, public discussions of audience data do not always differentiate the company from its network competitors in neutral terms.

Second, Netflix executives use discussions of viewership and ratings to explicitly denigrate longstanding television industry norms. This is especially clear in reference to the Nielsen ratings system. For example, in 2013, Sarandos told reporters, ‘Overnight ratings work against quality on television. That’s why I don’t want to adopt the one convention that I think is anti-quality’ (Chmielewski, 2013). He made similar comments a year later saying, ‘Maybe it has been necessary for the business of entertainment, but it’s been terrible for the creative side of television’ (Gardner, 2014). In response to industry calls for greater transparency, Sarandos made the same point in 2016 saying:

If we turn it into a weekly arms race by doing box scores of every live-plus-3 or live-plus-7, I think it’s going to have the same result that it has for television, which has been remarkably negative in terms of the quality of shows. (Villarreal, 2016)

Of course, anti-ratings discourse is hardly new. In 2005, for example, HBO’s then-CEO Chris Albrecht observed, ‘Other networks aim to make things popular and then maybe work on making them good. We [at HBO] say, “We gotta [sic] be good, and wouldn’t it be nice if it were popular”’ (Andersen, 2005). Yet, what separates Netflix’s statements about Nielsen ratings from earlier industry criticism is the totality of the critique.

In contrast with Sarandos’ concerns regarding the quality of television content, other executives claim the demographic categories that serve as the basis of the ratings system have little value. Describing the platform’s ‘mountain of data’, Todd Yellin, VP of product innovation notes, ‘That mountain is composed of two things. Garbage is 99 percent of that mountain. Gold is one percent . . . Geography, age, and gender? We put that in the garbage heap. Where you live is not that important’ (Barrett, 2016). In addition to criticizing the industry’s reliance on ratings data, Netflix has also criticized Nielsen directly. In 2017, the ratings giant began producing ratings for streaming content without Netflix’s participation and shortly thereafter publicly announced some viewing estimates related to the second season of Stranger Things (2016–) (Maglio, 2017). A spokesperson for the streaming platform responded:

Nielsen only measures a fraction of [Netflix] members’ viewing. For example, they don’t measure mobile viewing. Our series are global and member viewing patterns vary greatly (what device they watch on, where they’re watching, what they are watching), making it very difficult to model, even if they had accurate samples. (O’Reilly, 2017)

Beyond the criticism of Nielsen’s methodology, the broader criticism of audience sampling here is particularly notable. Of course, before the widespread adoption of digital technology, sampling was the only economically feasible means of generating audience data for the industry.

Third, Netflix also references audience data to deflect criticism. Here, the case of Marco Polo is particularly instructive. The first ten-episode season of this big-budget historical epic cost a reported ninety million dollars and was widely understood to reflect Netflix’s global ambitions (Steel, 2014). Billed as the streamer’s response to the success of HBO’s Game of Thrones (2011–2019), the show was savaged by American television critics who found it to be a ‘disappointment’ and ‘dramatically inert’ (Genzlinger, 2014; Van Der Werff, 2014). A month after Marco Polo’s premier and just 2 weeks after the second season renewal was announced, CEO Reed Hastings told an interviewer the show was ‘a massive success’ that had ‘been super popular with the audience’ (Stenovec, 2015). The second season, with a budget of more than 100 million dollars, premiered in July 2016. Discussing Marco Polo’s performance later that year, Sarandos told an industry roundtable that the show was ‘hugely popular all throughout Asia and Europe’ (The Hollywood Reporter, 2016). Given Netflix’s subscriber-based economic model, he explained, the series’ lack of popularity with American audiences and its harsh reception by television critics was ‘really irrelevant because it’s doing exactly what it was supposed to do’. In these instances, the practice of discussing but not releasing audience data allows Netflix to deflect the criticism most networks face regarding an underperforming big-budget project. Instead of being forced to answer for low ratings and unmet expectations, anti-transparency policies ensure that executives can claim success or redefine the parameters of success without be challenged. Of course, claiming success and succeeding are two different things. In December 2016, Netflix canceled Marco Polo after just two seasons taking a reported $200 million loss on the project (Goldberg, 2016).

Selective data releases and the promise of greater transparency

Netflix publicized specific audience data related to a specific title for the first time on December 28, 2018, after more than a decade of ignoring industry demands and refusing to comment on press speculation. The official Twitter account for Netflix Films (@NetflixFilm) claimed that more than forty-five million subscriber accounts viewed the Netflix original film Bird Box (2018) during the seven days following its premier (NetflixFilm, 2018). After initially refusing to comment on the metrics associated with this data, the next day, a company spokesman explained that a ‘view’ of Bird Box was defined as a subscriber account watching 70% of the total running time including credits but that this definition was specific to this title and ‘should not be taken as a metric for all Netflix content’ (Spangler, 2018). The reasons behind the company’s decision to publish viewership numbers for this particular film and the decision to begin selective data releases more broadly was never disclosed. Nonetheless, it seems likely that the decision to first publish audience data to demonstrate the popularity an original film rather than a television series was related to the consistently poor reviews Netflix films had received from critics.

Less than a month later, in January 2019, the company publicized ‘original’ series viewership data for the first time. In the quarterly shareholder letter, Netflix estimated that the recently launched series You (2018–) and Sex Education (2019–) were both on pace to be seen by more than forty million households during the first 4 weeks following their premiere (Netflix, 2019a). According to press reports, viewership for series was defined as a subscriber account watching at least seventy percent of a single episode of a given show (Porter, 2019a). As surprising as the disclosure of viewing data was, the company’s justification for doing so was perhaps more surprising. In an interview following the earnings report, Sarandos explained the company’s decision to disclose viewing data in traditional industry terms. Claiming this audience data as ‘cultural metrics’, he asserted that eighty million Netflix views ‘means exactly the same thing as eighty million plus people buying a movie ticket to seeing it or eighty million households watching a TV show’ (Netflix Investor Relations, 2019). For subscribers, he continued, this data tells them that they are ‘in the zeitgeist . . . watching the programming that the rest of the world is loving at the same time’. Given the platform’s brand identity grounded in around individualized viewing experiences and more than a decade of insistence that streaming data bore no relation to television ratings, Sarandos’ references to collective audience experience and traditional industry metrics of success represents a significant departure for Netflix.

Nonetheless, the television industry’s response to Netflix’s newfound willingness to publicize audience data was highly critical. At the Television Critics Association winter press tour, FX CEO John Landgraf told reporters, ‘Netflix is not telling you the full story’ (De Moraes, 2019). In particular, he took exception to Netflix’s methodology noting that the company did not use the industry standard of average audience measurement which would have provided an ‘an apples to apples comparison of viewership’. Instead, Landgraf claimed, Netflix publicized grossly inflated viewership numbers for You and Sex Education with the goal of presenting the service’s original content as significantly more popular than it is. He also accused Netflix of inflating perceptions of overall viewership overall by ‘reporting their singles like home runs’. In the months and years that followed, Landgraf’s claims regarding Netflix’s use of selective data release to increase public perception of the service’s popularity became common wisdom within the industry, the trade press, and even the popular press.

Several months later, Netflix executives began promising more transparency. In April 2019, Sarandos told investors that the company planned to begin releasing ‘more specific and granular data’ with the goal of becoming ‘more fully transparent about what people are watching on Netflix around the world’ (Alexander, 2019). At the same time, the company announced plans to test a new ‘Top 10’ feature that would allow subscribers to see which shows and movies were currently the most popular across the service. In July, as the platform announced that 40.7 million member accounts had watched at least part of Stranger Things’ third season using the same methodology as earlier in the year (Porter, 2019b). In October, Netflix released lists of its most viewed movies and series for the previous twelve months (Koblin, 2019). The same month, a letter dated to the previous July became public in which Netflix (2019b) told a UK parliamentary committee, ‘Transparency is important to us . . . To this end we recently began to share metrics more consistently with UK directors and producers about their individual shows and films’. In addition to the previously publicized seventy percent metric, the letter mentions two not previously public metrics: ‘starters’ and ‘completers’. Starters are defined as households that watch 2 minutes of a film or one episode. Completers include households that watch ninety percent of a film or season of a series during the first 7 and 28 days it is available. Netflix claimed that these ‘two metrics will give our creative partners a broader understanding of how members engage with their title from start to finish’. Yet, no substantive audience data was included in the letter. In December, executives were still promising more transparency. At an industry event, Scott Stuber, the head of original films at Netflix, told reporters, ‘You’ll see more numbers from us, more transparency, more articulation of what’s working and not . . . So, we’re definitely headed in that direction as a company’ (Vary, 2019).

However, the promised transparency had not materialized as 2019 ended. Instead, Netflix released a single year-end top ten list including both shows and movies produced with a new metric for audience size. According to this new definition, viewership is ‘the number of accounts choosing to watch at least two minutes of a series, movie or special during its first 28 days on Netflix in 2019’ (Patches, 2019). The company explained that this new metric is ‘a far more pure way to measure popularity rather than 70% metric we’ve used in the past (which discriminates against longer form content)’. Purity aside, less than a month later, Netflix disclosed the practical benefits of the new 2-minute view. As the company explained in its quarterly letter to shareholders, ‘The new metric is about 35% higher on average than the prior metric. For example, 45 [million] member households chose to watch Our Planet [(2019)] under the new metric vs. 33 [million] under the prior metric’ (Netflix, 2020). Not surprisingly, the industry’s response was again highly critical. One unnamed network executive told The Hollywood Reporter that this move is ‘obviously to goose [Netflix’s] numbers’ while another wondered why a company like Netflix ‘who trades in “viewership” isn’t held to the same standard of independent verification as the rest of us’ (Porter, 2020). This criticism had no discernable impact and, in February 2020, Netflix introduced a new daily top-ten feature on its user interface which allows subscribers to see that day’s most popular content in their country. Yet, a ranked list of the most watched titles on a given day provides very little information without knowledge of the differences related to each title and the overall percentage of viewing represented. As such, it seems that Netflix publicizes approximately the same amount of substantive audience data through selective releases as it did when the company refused to publicly comment on viewing numbers.

Conclusion

Using the framework of media industry studies, this article examines industrial discourses related to Netflix’s audience data under conditions shaped by the company’s anti-transparency policy. Until late-2018, executives used discussions of proprietary audience data to position the platform outside of the traditional television industry, to critique traditional industry players, and to deflect criticism. During this period, the company presented the popularity of its original series as axiomatic. As Sarandos explained in 2014, ‘You know in the culture that Orange Is The New Black [(2013-2019)] and House Of Cards are enormous hits’ (Patten, 2014). This discourse, enabled by the platform’s subscriber-based revenue model, supported Netflix’s broader efforts to position streaming as linear television’s inevitable replacement. Although these dynamics remain largely unchanged, by early 2019, selective releases of audience data for what are presumably the service’s most popular series became standard practice. Although Hastings has described this willingness to publicize any data as part of the company’s efforts to ‘grow up a little’ (Goldbart, 2019), Netflix’s use of non-standard metrics remains a point of contention within the industry. Setting aside questions of whether selective data releases represent a move toward meaningful transparency (they don’t), this research has several significant implications for scholars concerned with video streaming services.

First, what Netflix says and doesn’t say about its audience data and ratings more generally must be understood within the context of the television industry. Even as the company seeks to position itself as an outsider, its executives work to promote the company as a streaming television platform and television content producer. In this sense, changes within Netflix’s public discussions of its audience data are indicative of the company’s continuing to desire to have it both ways: be recognized as a significant player in the traditional television industry while maintaining the ability to eschew any traditional industry norms deemed inconvenient. Of course, this dynamic extends to issues beyond audience data. When facing criticism for cancelling nearly twenty shows featuring leads that were female, a person of color, or a member of the LGBTQ+ community, Netflix relied on a similar rhetorical strategy. Defending the company’s programming decisions at an industry panel, Bela Bajaria, Global Head of TV, claimed that the large number of first season cancellations is misleads as continuing series ‘have a renewal rate of 67%, which is industry standard’ (White, 2020). Yet at the same panel, Sarandos positioned Netflix outside of the traditional industry asserting that the company’s recent cancelations were incorrectly ‘measured against the old way of doing things’. In this instance, however, Netflix’s claims can be verified as renewals and cancellations are public. By contrast, the platform’s audience data remains proprietary allowing the company to redefine successful television on their own terms.

Second, the industrial discourse related to Netflix’s audience data is of course situated within the broader transitions from old media to new media defined by ongoing negotiations between established and emerging practices. Needless to say, Netflix is not the only firm managing the tensions between old and new media. As conglomerate-backed video streaming services like Disney+ and HBO Max attempt to catch up with Netflix both in terms of subscribers and cultural influence, it seems likely that discussions of audience data will become more common without necessarily becoming more substantive. In October 2020, for example, Amazon claimed that ‘tens of millions’ of viewers watched Sacha Baron Cohen’s comedy Borat Subsequent Moviefilm over the course of its opening weekend (Spangler, 2020). Such vague claims are certainly not limited to global streaming services. In December 2020, Movistar+, the most popular SVOD in the Spanish market which is a subsidiary of the country’s dominant pay-television provider (Wayne and Castro, 2020), announced that the film While the War Goes On was the service’s most watched movie of the year (Cine & Tele, 2020). Coincidentally, this particular title happens to be the platform’s first original film. Given the different content libraries and target audiences of global and national streaming platforms, it is possible that each might well come to conceptualize success and discuss popularity in distinctive ways.

Third, this research highlights the ways in which Netflix’s public discussions of its audience data further limit possibilities for collective understandings of popular television in the age of streaming. The company’s anti-transparency policies ensure that connections between popularity and viewer behavior are articulated on its own terms. Netflix content is popular because Netflix says so. For observers inside and outside of the industry interested in a better quantitative understanding of the global SVOD audiences, these policies are particularly frustrating. Executives, if they were willing, could provide much more substantive information regarding subscribers and their viewing behavior. If, however, one seeks to develop a qualitative understanding of Netflix audiences and their relationships with specific original series, then the situation is more complicated. Like traditional television institutions, Netflix continues to face the problem that the ‘television audience’ is a fictional construct that will always refuse definitive representation, discursively or otherwise. From this perspective, the industrial discourses surrounding SVOD audience data are not merely a restatement of conventional industry wisdom. They are a re-articulation of industry lore ‘within the slippery mediated terrain of post-network television and digital distribution’ (Burroughs, 2019: 4). As such, even substantive and sustained data transparency could not meaningful address ‘the profound, structural uncertainty about the audience’ which Ang (1991) characterizes as ‘the core predicament of the television industry’ (43).

Looking ahead, it seems likely that the industry discourses related to SVOD audience data will continue to reflect and rely upon anti-transparency policies. It is difficult to imagine how market dynamics might develop in ways that will incentivize data transparency for subscriber-supported platforms. For Netflix, Disney+, HBO Max, and others, the ability to claim that some series is popular and have those claims go unchallenged is a mechanism of generating value in multiple contexts. As marketing, the perception that a service is the exclusive home to popular content drives subscriber acquisition. Similarly, the absence of standardized audience metrics allows each SVOD to define popularity in the most advantageous terms which, as Netflix has shown, becomes a useful means of demonstrating value to shareholders and others in the investor classes. However these companies come to use selective data releases for their own ends, these industrial practices will certainly challenge the dominant conception of popular television based in the idea that popularity is a function of audience size. Yet, there is little reason to approach the increasingly unknowable realities of popular television as a radical break from the past. As Katz (1996) observed nearly 25 years ago, ‘With the rapid multiplication of channels, television has all but ceased to function as a shared public space’ (22). The US commercial television industry has never been interested in accurate and transparent audience data for itself or the public. To whatever degree the Nielsen ratings system effectively provided a shared currency for the television industry, it provided the public with an inaccurate image of American audiences as the company ignored many well-known inadequacies ranging from methodological problems associated with time diaries to the practice of under-sampling viewers of color. Along these lines, future television industry scholarship can mostly usefully address the streaming age by seeking to understand contemporary practices and discourses as both extensions of and points of rupture with the medium’s past.

It would, however, be premature to assume that this state of affairs is a fait accompli. As streaming television diversifies, this emerging reality where the idea of popular television, divorced from actual viewer behavior, functions as a marketing ploy or is produced with gameable metrics to satisfy investors might well be limited to the largest services in the SVOD market. In 2020, for example, three services (Izzy, ChaiFlicks, and Jewzy) entered the SVOD market targeting audiences interested in Jewish/Israeli film and television content. For niche services such as these attempting to build their subscriber base from the same potential audience, the benefits of substantive data transparency could be significant in relation to brand differentiation, content acquisition, and efforts to raise capital. Furthermore, the streaming market is no longer exclusively populated by services using subscriber-supported economic models. Amazon’s investments in IMDb TV, the company’s no-cost advertiser-supported video on-demand (AVOD) service (initially branded as Freedive), reflect growing efforts to attract the advertising budgets once committed to linear television. For AVODs like the Roku Channel, Peacock, Pluto, Tubi, and Crackle, the relationship between dollars and eyeballs will likely be more straightforward than it is for SVODs. Ultimately, understanding the role of streaming audience data in this increasingly complex landscape will necessitate that scholars broaden their focus beyond the largest and most familiar platforms to begin addressing the many emerging realities of ‘popular’ television.

Footnotes

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Michael L Wayne

References

Alexander

(2019) Netflix execs say they’ll finally start releasing viewership data soon. In: The Verge. Available at: https://www.theverge.com/2019/4/17/18412309/netflix-transparency-data-viewers-ratings-subscribers-bird-box-numbers (accessed 16 September 2020).

Andersen

(2005) How HBO could get back its Mojo. New York Magazine, 8 August. Available at: https://nymag.com/nymetro/news/columns/imperialcity/12199/ (accessed 8 September 2020).

Ang

(1991) Desperately Seeking the Audience. New York, NY: Routledge.

Baldwin

(2012) Netflix Gambles on big data to become the HBO of streaming. Wired, 29 November. Available at: https://www.wired.com/2012/11/netflix-data-gamble/ (accessed 8 September 2020).

Barrett

(2016) Netflix’s Grand, Daring, Maybe Crazy Plan to Conquer the World. Wired, 27 March. Available at: https://www.wired.com/2016/03/netflixs-grand-maybe-crazy-plan-conquer-world/ (accessed 9 September 2020).

Burroughs

(2019) House of Netflix: Streaming media and digital lore. Popular Communication 17(1): 1–17.

Caldwell

(2008) Production Culture: Industrial Reflexivity and Critical Practice in Film and Television. Durham, NC: Duke University Press.

Chmielewski

(2013) Ted Sarandos upends Hollywood with Netflix revolution. Los Angeles Times, 25 August. Available at: https://www.latimes.com/entertainment/envelope/cotown/la-et-netflix-ted-sarandos-20130825-dto-htmlstory.html (accessed 8 September 2020).

Cine & Tele (2020) ‘While the War Lassts’, the most viewed film of the year in Movistar. In: Cine & Tele. Available at: https://www.cineytele.com/2020/12/28/mientras-dure-la-guerra-pelicula-mas-vista-del-ano-en-movistar/ (accessed 29 December 2020).

10.

Corrigan

(2018) Making implicit methods explicit: Trade press analysis in the political economy of communication. International Journal of Communication 12: 2751–2772.

11.

De Keere

Thunnissen

Kuipers

(2021) Defusing moral panic: Legitimizing binge-watching as manageable, high-quality, middle-class hedonism. Media, Culture & Society 43(4): 629–647.

12.

de Moraes

(2019) FX Chief John Landgraf Smacks down Netflix big ratings “myth” – TCA. In: Deadline. Available at: https://deadline.com/2019/02/john-landgraf-smacks-down-netflix-ratings-myth-tca-1202549106/ (accessed 28 September 2020).

13.

Gardner

(2014) Netflix’s Ted Sarandos says TV ratings discussions are ‘negative’. In: The Hollywood Reporter. Available at: https://www.hollywoodreporter.com/news/netflixs-ted-sarandos-says-tv-754890 (accessed 8 September 2020).

14.

Genzlinger

(2014) O.K., the women can kill, but they have to be nude. The New York Times, 11 December. Available at: https://www.nytimes.com/2014/12/12/arts/television/marco-polo-fur-and-armor-drama-from-netflix.html (accessed 15 September 2020).

15.

Gitlin

(1983) Inside Prime Time. New York, NY: Pantheon Books.

16.

Goldbart

(2019) Netflix aiming to be more open about data, says Reed Hastings. In: Screen Daily. Available at: https://www.screendaily.com/news/netflix-aiming-to-be-more-open-about-data-says-reed-hastings/5143077.article (accessed 29 September 2020).

17.

Goldberg

(2016) ‘Marco Polo’ canceled at Netflix after two seasons. Hollywood reporter. In: The Hollywood Reporter. Available at: https://www.hollywoodreporter.com/live-feed/marco-polo-canceled-at-netflix-two-seasons-955561 (accessed 15 September 2020).

18.

Hallinan

Striphas

(2016) Recommended for you: The Netflix Prize and the production of algorithmic culture. New Media & Society 18(1): 117–137.

19.

Hass

(2013) Reed hastings on arrested development, house of cards, and the future of Netflix. In: GQ. Available at: http://www.gq.com/story/netflix-founder-reed-hastings-house-of-cards-arrested-development (accessed 5 March 2017).

20.

Hastings

Wells

(2012) Investor Letter Q1 2012. 23 April. Available at: https://s22.q4cdn.com/959853165/files/doc_financials/quarterly_reports/2012/q1/Investor-Letter-Q1-2012.pdf (accessed 6 November 2019).

21.

Hayes

(2019) Netflix’s Ted sarandos weighs in on streaming wars, agency production, big tech breakups, M&A OUTLOOK – Update. In: Deadline. Available at: https://deadline.com/2019/06/netflix-ted-sarandos-weighs-in-on-streaming-wars-agency-production-big-tech-breakups-ma-outlook-1202636595/ (accessed 8 September 2020).

22.

Hessler

(2021) Peoplemeter technologies and the biometric turn in audience measurement. Television & New Media 22(4): 400–419.

23.

Katz

(1996) And deliver us from segmentation. The Annals of the American Academy of Political and Social Science 546(1): 22–33.

24.

Koblin

(2019) Netflix’s top 10 original movies and TV shows, according to Netflix. The New York Times, 17 October. Available at: https://www.nytimes.com/2019/10/17/business/media/netflix-top-ten-movies-tv-shows.html (accessed 16 September 2020).

25.

Kosterich

Napoli

(2016) Reconfiguring the audience commodity: The institutionalization of social TV analytics as market information regime. Television & New Media 17(3): 254–271.

26.

Lotz

(2017) Portals: A Treatise on Internet-Distributed Television. Ann Arbor, MI: Maize Books.

27.

Lotz

(2018) The post-network era: 2000s–present. In: Bodroghkozy

(ed.) A Companion to the History of American Broadcasting. Oxford, UK: Wiley Blackwell, pp.153–167.

28.

Lotz

Havens

(2016) Original or exclusive? Shifts in television financing and distribution shift meanings. In: Antenna: Responses to Media and Culture. Available at: http://blog.commarts.wisc.edu/2016/01/01/original-or-exclusive-shifts-in-television-financing-and-distribution-shift-meanings/ (accessed 29 October 2016).

29.

Lynch

(2018) Netflix thrives by programming to ‘taste communities,’ not demographics. In: AdWeek. Available at: https://www.adweek.com/tv-video/netflix-thrives-by-programming-to-taste-communities-not-demographics/ (accessed 15 September 2020).

30.

Maglio

(2017) ‘Stranger things’: 15.8 million watched season 2 premiere in 1st weekend. In: The Wrap. Available at: https://www.thewrap.com/stranger-things-2-tv-ratings-netflix-nielsen/ (accessed 9 September 2020).

31.

McCabe

Akass

(2007) Sex, swearing and respectability: Courting controversy, HBO’s original programming and producing quality TV. In: McCabe

Akass

(eds) Quality TV: Contemporary American Television and Beyond. London: I. B. Tauris, pp.62–76.

32.

McNutt

(2020) Starstuck: Streaming, celebrity, and the state of the “Netflix Star”. In: Cultural Learnings. Available at: https://cultural-learnings.com/2020/09/25/starstuck-streaming-celebrity-and-the-state-of-the-netflix-star/ (accessed 7 October 2020).

33.

Netflix (2013a) Netflix Long Term View. 24 April. Available at: http://docplayer.net/19416326-Netflix-long-term-view.html (accessed 6 November 2019).

34.

Netflix (2013b) Q2 2013 Netflix, Inc. earnings conference call transcript. Available at: https://s22.q4cdn.com/959853165/files/doc_financials/quarterly_reports/2013/q2/NFLX-Transcript-2013-07-22T22_00.pdf (accessed 27 January 2020).

35.

Netflix (2015) Do you know when you were hooked? Netflix Does. Press Release. Los Gatos, CA. Available at: https://about.netflix.com/en/news/do-you-know-when-you-were-hooked-netflix-does (accessed 6 November 2019).

36.

Netflix (2016) Netflix & Binge: New Binge scale reveals TV series we devour and those we savor. Press Release. Available at: https://media.netflix.com/en/press-releases/netflix-binge-new-binge-scale-reveals-tv-series-we-devour-and-those-we-savor-1 (accessed 28 February 2019).

37.

Netflix (2017a) Netflix cheating is on the rise globally and shows no signs of stopping. Press Release. Available at: https://www.prnewswire.com/news-releases/netflix-cheating-is-on-the-rise-globally-and-shows-no-signs-of-stopping-300406051.html#:~:text=All%20Products-,Netflix%20Cheating%20is%20on%20the%20Rise%20Globally%20and%20Shows%20No,Netflix%2C%20Inc (accessed 1 March 2019).

38.

Netflix (2017b) Ready, set, Binge: More than 8 million viewers ‘binge race’ their favorite series. Press Release. Available at: https://about.netflix.com/en/news/ready-set-binge-more-than-8-million-viewers-binge-race-their-favorite-series (accessed 1 March 2019).

39.

Netflix (2019a) Q4 2018 Letter to Shareholders. Letter. Available at: https://s22.q4cdn.com/959853165/files/doc_financials/quarterly_reports/2018/q4/01/FINAL-Q4-18-Shareholder-Letter.pdf (accessed 27 January 2020).

40.

Netflix (2019b) Supplementary written evidence (PSB0069). Available at: http://data.parliament.uk/writtenevidence/committeeevidence.svc/evidencedocument/communications-committee/public-service-broadcasting-in-the-age-of-video-on-demand/written/104558.html (accessed 23 December 2020).

41.

Netflix (2020) Q4 2019 letter to Shareholders. Letter. Available at: https://s22.q4cdn.com/959853165/files/doc_financials/2019/q4/FINAL-Q4-19-Shareholder-Letter.pdf (accessed 27 January 2020).

42.

Netflix Investor Relations (2019) Netflix Q4 2018 earnings interview video. In: YouTube. Available at: https://www.youtube.com/watch?v=mYsKCfWV0fs (accessed 27 January 2020).

43.

NetflixFilm (2018) Twitter post. In: Twitter. Available at: https://twitter.com/NetflixFilm/status/1078735051406204928 (accessed 27 January 2020).

44.

O’Reilly

(2017) CMO today: CNN to launch digital subscriptions; Nielsen’s ‘stranger things 2’ ratings; advertising escapes tax reform. Wall Street Journal, 3 November. Available at: https://www.wsj.com/articles/cmo-today-cnn-to-launch-digital-subscriptions-nielsens-stranger-things-2-ratings-advertising-escapes-tax-reform-1509710449 (accessed 9 September 2020).

45.

Patches

(2019) Netflix says these are its most popular movies and shows of 2019. In: Polygon. Available at: https://www.polygon.com/2019/12/30/21042995/netflix-top-10-movies-shows-2019-viewers (accessed 16 September 2020).

46.

Patten

(2014) HRTS: Netflix keeping its metrics to itself, ted sarandos says. In: Deadline. Available at: https://deadline.com/2014/02/netflix-metrics-staying-private-ted-sarandos-says-hrts-678316/ (accessed 15 September 2020).

47.

Perren

(2015) The trick of the trades: Media industry studies and the American comic book industry. In: Banks

Conor

Mayer

(eds) Production Studies, The Sequel! London: Routledge, pp.247–257.

48.

Porter

(2019a) Netflix reveals viewership numbers for ‘you,’ ‘sex education’ and more. In: The Hollywood Reporter. Available at: https://www.hollywoodreporter.com/live-feed/netflixs-you-track-40-million-viewers-1177025 (accessed 16 September 2020).

49.

Porter

(2019b) ‘Stranger things 3’ racks up record viewership, according to Netflix. In: The Hollywood Reporter. Available at: https://www.hollywoodreporter.com/live-feed/stranger-things-racks-up-record-viewership-netflix-1223028 (accessed 16 September 2020).

50.

Porter

(2020) Netflix’s new ratings math sacrifices clarity for flashy numbers. In: The Hollywood Reporter. Available at: https://www.hollywoodreporter.com/live-feed/netflixs-new-ratings-math-sacrifices-clarity-flashy-numbers-1272214 (accessed 23 September 2020).

51.

Quintanilla

(2014) The programmer’s summit, HRTS luncheon recap. In: Hollywood Radio & Television Society. Available at: https://hrts.org/the-programmers-summit-hrts-luncheon-recap/ (accessed 31 August 2020).

52.

Sarandos

(2012) Carsey-Wolf Center at UC Santa Barbara. Available at: https://www.carseywolf.ucsb.edu/wp-content/uploads/2018/02/Interview_Ted-Sarandos.pdf (accessed 31 August 2020).

53.

Spangler

(2018) ‘Bird box’ viewed by 45 million Netflix members in first week, company says. In: Variety. Available at: https://variety.com/2018/digital/news/bird-box-netflix-45-million-views-first-week-1203096379/ (accessed 16 September 2020).

54.

Spangler

(2020) ‘Borat 2’ drew ‘tens of millions’ of viewers over opening weekend, amazon says. In: Variety. Available at: https://variety.com/2020/digital/news/borat-2-viewership-numbers-amazon-opening-weekend-mulan-1234816491/ (accessed 23 December 2020).

55.

Steel

(2014) How to build an empire, the Netflix way. The New York Times, 29 November. Available at: https://www.nytimes.com/2014/11/30/business/media/how-to-build-an-empire-the-netflix-way-.html (accessed 15 September 2020).

56.

Stenovec

(2015) Netflix CEO says ‘marco polo’ is a hit, and we’ll never know otherwise. In: The Huffington Post. Available at: https://www.huffpost.com/entry/netflix-reed-hastings-marco-polo_n_6525776 (accessed 15 September 2020).

57.

The Hollywood Reporter (2016) TV Executives Roundtable with Ted Sarandos, Bonnie Hammer and More. Available at: https://www.hollywoodreporter.com/video/watch-thr-s-full-tv-918281 (accessed 7 August 2019).

58.

Tryon

(2015) TV got better: Netflix’s original programming strategies and the on-demand television transition. Media Industries Journal 2(2): 104–116.

59.

Uricchio

(2008) Television’s first seventy-five years: The interpretive flexibility of a medium in transition. In: Kolker

(ed.) The Oxford Handbook of Film and Media Studies. Oxford, UK: Oxford University Press, pp.286–305.

60.

Van Der Werff

(2014) We watched all of Netflix’s Marco Polo. You . . . don’t have to. In: Vox. Available at: https://www.vox.com/2014/12/15/7392517/marco-polo-review-netflix (accessed 15 September 2020).

61.

Van Der Werff

(2015) The big flaw in Netflix’s research into when it has you hooked. In: Vox. Available at: http://www.vox.com/2015/9/24/9392933/netflix-chart-hooked (accessed 7 March 2017).

62.

Vary

(2019) Netflix film chief Scott Stuber says streamer planning to release more audience numbers. In: Variety. Available at: https://variety.com/2019/film/news/netflix-viewership-numbers-1203426670/ (accessed 16 September 2020).

63.

Villarreal

(2016) It’s still ‘Netflix and no ratings’ – Ted Sarandos rejects a numbers ‘arms race’. Los Angeles Times, 17 January. Available at: https://www.latimes.com/entertainment/la-et-ct-sarandos-tca-netflix-20160117-story.html (accessed 14 September 2020).

64.

Wayne

(2018) Netflix, amazon, and branded television content in subscription video on-demand portals. Media, Culture & Society 40(5): 725–741.

65.

Wayne

Castro

(2020) SVOD global expansion in cross-national comparative perspective: Netflix in Israel and Spain. Television & New Media. Epub ahead of print 4 June 2020. DOI: 10.1177/1527476420926496.

66.

White

(2020) Netflix reveals data around show cancellations as Bela Bajaria & Ted Sarandos discuss new strategy. In: Deadline. Available at: https://deadline.com/2020/11/netflix-bela-bajaria-cancellations-ted-sarandos-67-percent-1234613085/ (accessed 29 December 2020).

67.

(2015) Netflix’s secret special algorithm is a human. The New Yorker, 27 January. Available at: https://www.newyorker.com/business/currency/hollywoods-big-data-big-deal (accessed 8 September 2020).

Netflix audience data,streaming industry discourse,and the emerging realities of ‘popular’ television

Abstract

Keywords

Introduction

Audience measurement and the television industry

Methods

Anti-transparency data policy and the practice of discussing without releasing

Selective data releases and the promise of greater transparency

Conclusion

Footnotes

Funding

ORCID iD

References