Abstract
Netflix’s status as a personalised service has been central to its business proposition and brand. However, recent changes to include community-based metrics within the user interface – such as the 2020 addition of a national top 10 feature – denote a shift in corporate strategy from personalisation to communal discovery. This article uses a critical communications and media industry studies approach to consider both the data being produced by the top 10 ranking and the broader industrial function of the list, especially within a longer history of audience measurement.
Keywords
Introduction
Netflix’s status as a personalised service has been central to its business proposition and brand. While it has always been somewhat social, in that your programming suggestions are based on the recommendations of content-based algorithms and collaborative filtering that considers what you and others have watched, there has never been any real sense of other people in the surface of the user interface. This is a space, the rhetoric tells us, tailored to your unique tastes and interests. Netflix ‘claims to continually refine its formulae so that it will grow in proficiency as it seeks to offer tailored, if not bespoke, programming to each subscriber, rather than figuring audience “needs” in the abstract aggregate or as a public’ (Shimpach, 2020: 3–4). However, recent changes to make communal processes of discovery more visible within the Netflix user interface denote a shift in corporate strategy. The 2020 addition of a national top 10 feature is particularly significant. If Netflix is capable of ‘knowing’ the audience in ways that traditional television industries cannot, what can we learn from the service’s inclusion of albeit black box measurements of ‘popularity’? What does the insertion of this feature suggest about what subscribers want and how is Netflix exploiting that?
In trying to understand the logics and limitations of the Netflix top 10 feature, I tracked the Australian overall top 10 list at the same time every day for 1 year. Following critical communications and media industry studies approaches, this data was contextualised with industrial discourses surrounding the arrival of the feature, such as Netflix’s own statements and disclosures, and reactions from key stakeholders in trade press articles. This article considers what can be discerned from the data being produced by the top 10 rankings and questions the broader industrial function of the metric altogether, especially within a longer history of audience measurement.
On face value, reading the top 10 data indicates an appetite for content that is not necessarily recently produced, but recently added to the catalogue. This raises interesting questions for Netflix about the value of commissioning new content as compared to licensing existing titles. The top 10 data also confirm some interesting trends in Netflix consumption in Australia: namely, a preference for US content and a penchant for titles branded as Netflix Originals. The top 10 feature is otherwise problematically repetitive and lacks a methodological transparency that can tell us anything more concrete about viewer behaviour.
There is quite a bit that we can ascertain, however, about what the development of the metric suggests about Netflix’s evolving engagement strategies. I posit that the addition of a top 10 feature acts as an acknowledgement by the subscription video-on-demand (SVOD) platform of the pervasive communality inherent in what has become an increasingly solitary consumption ritual. This is a shift that challenges Netflix’s corporate positioning as a global service by highlighting national consumption. It reflects a return to the use of an albeit opaque ratings system, one perhaps perceived as ‘purer’ than recommendation algorithms. Finally, Netflix exploits this performative socialisation to shape viewing in the face of an abundance of choice and cyclically build hype around titles, to ensure that subscribers feel compelled to keep up with what’s ‘popular’.
This article draws from and builds upon a body of analysis that has considered SVOD recommender systems and their effects (as will be considered below) and the corporate positioning and branding of these platforms (Rios and Scarlata, 2018; Wayne, 2018; Wayne and Uribe Sandoval, 2021). It employs as a foundation recent considerations of how to study user interfaces (Johnson, 2019; Kelly, 2021) and SVOD catalogues (Albornoz and Garcia Leiva, 2022; Lobato and Scarlata, 2019). Netflix’s incorporation of a ‘ratings mentality’ (Bourdieu, 1998: 70) emphasises the continued relevance of historical studies that established the political and economic dimension to media measurement systems.
This analysis begins by tracing the introduction of the top 10 feature: describing what it is, historicising its development and roll-out, reviewing company justifications and media responses to its addition, and considering what we know about how the list is calculated. I then reflect on what we can take away from this contested and controversial metric via critical readings of (1) the data it provides, and (2) the industrial context which draws value from the mechanism. This article concludes with reflections on what the top 10 feature suggests about the enduring sociality of television in transition.
Getting to know the top 10 feature
The top 10 feature is a unique category row available on all Netflix interfaces, regardless of device. Updated daily, it ranks titles based on popularity within a user’s country (ascertained by the geolocation of their IP address). While the position of the row will vary depending on how relevant the feature is to each user (Johnson, 2020), it can be prominently located at the top of the homepage above all other category rows, directly below the main carousel (see Figure 1). The user can access separate lists of the top 10 movies or top 10 TV shows when they click into those format tabs from the main menu, but the homepage top 10 list combines both film and television content. Titles that make any of these lists are also tagged with a small red top 10 badge in the top right-hand corner of their tile, wherever they appear on the platform (see Figure 2). Engagement with the metric is also facilitated by the ‘New & popular’ menu tab available in browser and smart TV interfaces (see Figure 1), and via direct-to-user emails and in-app notifications that ask subscribers, ‘What are people watching in your area?’ Netflix Australia homepage (web browser), September 2021. Netflix Australia category rows and top 10-badged titles, September 2021.

While the top 10 lists are now a familiar part of the Netflix experience, they are in fact a relatively recent addition to the user interface. The company began quietly testing the value of the feature in 2018. Content rows like ‘Popular on Netflix’ and ‘Trending’ might have seemed objective to subscribers, but such categories are ‘actually personalized content that also happens to be popular’ (Braimah as cited by Adalian, 2021). The Netflix product team tried de-personalising the ‘Popular on Netflix’ category in a handful of markets, feeding subscribers a row of shows which actually were the most popular and including the member’s country name in the category (for example, ‘Popular in the United States’). This experiment piqued enough interest in members to warrant pursuing it further (Adalian, 2021). The introduction of a top 10 list was first publicly flagged by Netflix in its Q1 letter to shareholders in April 2019, when it announced intentions to trial the feature in the United Kingdom in Q2 (Netflix, 2019a). The company introduced the lists for beta testing with users in the UK and Mexico from May. Following positive feedback, Netflix launched the top 10 feature internationally in February 2020.
This development can be seen as both a reaction to a larger, ongoing debate about transparency of streaming data and a response to subscriber interest in popular content. Mike Wayne (2021) has chronicled Netflix’s late-2018 shift in strategy, away from a consistent anti-transparency policy and towards selective data releases. In April 2019, the SVOD’s then-chief content officer Ted Sarandos promised to start rolling out ‘more specific and granular data and reporting’ (Patten, 2019). The top 10 lists were perceived as a tangible response to these promises to investors and industry (Lawler, 2019). As The New York Times noted, this ‘decision to go public with even some of its data was something of an about-face for Netflix, which once claimed it never wanted to get involved in the weekly Hollywood competition based on ratings or box office receipts’ (Koblin, 2019). In a call with investors in October 2019 Sarandos explained: One of the things you saw, we’ve launched in the U.K. and we’re looking to expand presentation of the top 10, so that people can come to Netflix and see the top 10 most popular things in different categories. Once again, I think that one way that people choose content is by popularity. It’s not the only way, and it’s not the only way we want people to. But if they want to use that as a tool to guide their decision-making, we want to help them do that. So publishing that top 10 that refreshes every 24 h is one way that we’re helping out on the consumer side (Netflix, 2019b).
The corporate rhetoric overwhelmingly established this as feature that would benefit subscribers. As the feature rolled out globally, Netflix Director of Product Innovation, Cameron Johnson, said: ‘When you watch a great movie or TV show, you share it with family and friends, or talk about it at work, so other people can enjoy it too. We hope these top 10 lists will help create more of these shared moments, while also helping all of us find something to watch more quickly and easily’ (Johnson, 2020).
What Netflix certainly provoked was excitement from a sector that had long been kept in the dark. The news media and trade presses quickly clung to the top 10 lists as ‘the most information Netflix has given the public about how “well” its shows are going’ (Tassi, 2020a), attributing the function with the ability to ‘take an already popular show like (Tiger King) and make it snowball’ (Tassi, 2020b). Variety described it as a ‘step toward greater transparency’ (Clarke, 2019). Within weeks, services such as FlixPatrol started to aggregate the cross-nation data the lists provided into something ‘approximating a cumulative audience count’ (Lotz, 2022).
However, traditional media analysts were reticent to take these metrics at face value, with the top 10 feature attracting significant scrutiny from observers who have questioned the underlying methodology used to determine the titles that appear in the list. It was initially the case that in compiling these lists, Netflix considered the ‘most-watched individual season of a show, film or special (regardless of when it launched). “Watched” [meant] members finished at least 70% of one episode’ (Netflix UK, 2019). However, Netflix’s rankings of its most-favoured content in the US and UK, published in Deadline and Variety in December 2019 (Andreeva, 2019; Donnelly, 2019; White and Kanter, 2019), interpreted ‘popularity’ somewhat more generously. These rankings were ‘based on the number of accounts that have watched at least 2 minutes of a movie, TV series or a special – both original and acquired – during its first 28 days on Netflix in 2019. For series that air multiple seasons in one calendar year, only the most popular season is counted’ (Andreeva, 2019). Netflix justified this approach as being ‘long enough to indicate the choice was intentional’, and the methodology as being similar to the BBC iPlayer in their rankings based on ‘requests’ for the title, ‘most popular’ articles on the New York Times which include those who opened the articles, and YouTube view counts. This way, short and long titles are treated equally, leveling the playing field for all types of our content including interactive content, which has no fixed length (Netflix, 2020a: 4).
The new ‘2-min constitutes a view’ methodology conveniently results in viewership that is about 35% higher on average than the previously used 70% completion metric (Netflix, 2020a), which explains why Netflix has continued to sporadically share some viewing figures – using this methodology – in its quarterly reports or via industry publications in a sparse and shrewd way, ever since. As the top 10 mechanism was rolled out internationally, Netflix confirmed via Variety (Spangler, 2020) that the feature had also adopted this new viewership-tracking methodology: that is, tallying the accounts that had watched at least 2 min of a title over the previous 24 hours. The service’s 2021 shift to report on hours viewed for its titles (Netflix, 2021) has ultimately trickled into how the top 10 lists are collated as well (De Rosso, 2021).
The various limitations of this data over time have also been noted by media scholars. On its methodological approach, Wayne (2021: 10) has questioned whether a tally of ‘starters’ rather than ‘completers’ really paints a reliable picture of ‘what people are watching in your area’. Lotz has gone further to describe the Netflix top 10 lists as ‘fairly dumb data’: These lists offer only ordinal indications, and simply rank series with no sense of the intervals among the rankings or way to discern whether the most viewed show on Tuesday was viewed by fewer than the lowest ranked show on Saturday. Consequently, the data cannot be aggregated into anything more than daily, relatively incomparable, snapshots (Lotz, 2020).
Lotz also points out that it is not clear how much Netflix viewing is represented by top 10 titles, and judiciously queries the potential blending of series and movies within the overall feature. After all, we do not know how the viewing of multiple episodes of the same series is handled and represented over time here, alongside the consumption of more than 2 minutes of a single film. While the instinctive response of the media to this data is to excitedly point to the titles that feature most prominently in the row, determining what actually is most popular is not as simple as it seems. Is popularity of a title to be best ascertained according to the time it spends consecutively in the list or according to an accumulation of days if it appears recurringly? Similarly, is the success of individual seasons of a show to be compiled in service of the entire series? Is the key indicator of success specific placement within the feature itself – should we commend titles that retain the #1 spot for the longest period? These questions remain unanswered.
Clearly, the Netflix top 10 feature is contested and controversial. It is not possible to gain a meaningful picture of regional Netflix consumption from a top 10 list alone. What we can take from it is much more partial and contingent. However, this is not without value. Despite its limitations, there is still much we can infer from the top 10 lists. As critical media scholars we must consider what Netflix is trying to achieve here. What is the platform seeking to reveal or obscure? What can we learn not just about Netflix viewing practices in a particular region, but about global engagement and satisfaction with the highly personalised and algorithmically curated service? Answering these questions involves critically reading (1) the (albeit limited) data offered by the top 10 feature and (2) the industrial function and context of the top 10 feature itself.
Reading the top 10 data
Let us begin with what can be learned from the top 10 data, because despite its restrictions and silences, a long-term collection and analysis of the metric reveals some noteworthy insights into Netflix viewing. In trying to understand the logics and limitations of the lists, I collected data from the Australian overall top 10 rankings over the course of a year (11 March 2020–11 March 2021). While the list would be updated at different times each day, every night at 10p.m. (Australian Eastern Standard Time) I opened the Netflix app and recorded the following data points: title, format (series/film), branding as a Netflix Original, country of origin (as per Netflix, or where unavailable as per IMDb), year of release (as per Netflix, though this proved problematic with some errors identified and requiring confirmation via IMDb), and genre (as per Netflix’s categorisation, the first listed only). I also noted the date titles were added (or returned) to the Netflix Australia catalogue (using third party aggregate sites such as https://anz.newonnetflix.info/ and independent media summaries of Netflix press releases via sources such as TechRadar and Gizmodo), as well as the resulting sequential and cumulative days a title spent in the overall top 10 list. I inadvertently adopted the systemic approach to video-on-demand interface analysis later expounded by JP Kelly (2021), which he described as akin to Franco Moretti’s ‘distant reading’ (2013). This methodology was used to ‘identify large-scale patterns and derive new insights… that would not be possible through the type of ad hoc observation and/or close textual analysis that has characterised previous studies of these now ubiquitous interfaces’ (Kelly, 2021: 265–266).
370 unique titles rotated through the list over the course of the year. These skewed slightly in favour of films (57% films vs 43% series/stand-up specials), but television content spent longer in the top 10 overall (which is unsurprising, given the methodological approach considered above). A variety of genres were represented, though assessing these further was complicated by Netflix’s practice of categorising some content by its provenance or source material, rather than genre: in the Australian catalogue, The Crown’s genre is ‘British’, The Old Guard a ‘Movie Based on Books’, and so on. Through this experiment I was able to confirm several Netflix viewing trends.
First, there was a close correlation between top 10 titles and content that was recently added to Netflix. 88% of the unique titles that were featured in the top 10 list made their debut in the feature within just 4 days of their addition to the Australian catalogue. Some titles, like the Chris Hemsworth Netflix Original film Extraction, made it in (and in this case to the top spot) within just hours of its release. Of course, this makes sense: audiences want to watch new releases, something they have not seen before. But, interestingly, many of the films and shows that made it to the overall top 10 list were not recently produced or exclusive to Netflix, but rather just recently added to the local service. For example, the 1999 John Travolta thriller The General’s Daughter, Vin Diesel’s 2003 action movie A Man Apart, and the Cher-led Burlesque (2010) all appeared in the overall top 10 list for several days each, within 1–4 days of their addition to the Australian catalogue. Nielsen (2021), the audience measurement company and primary source of TV ratings information in the US, found that in 2020 audiences were turning to nostalgic content, particularly comedies, for comfort during the pandemic, and I saw this play out in the data. Overall, nearly 18% of top 10 titles over the period considered were produced prior to 2015. However, for the most part, the appearance of old American broadcast dramas and sitcoms, such as Prison Break, The Blacklist, Friends and The Big Bang Theory in the top 10 feature occurred almost immediately after a new season or episode dropped. I attribute the sudden popularity of these older titles to the recency with which they had been licensed to the Netflix service in Australia.
While there were a few exceptions to this new-to-service rule, these could understandably be found in seasonal offerings (Christmas films like The Grinch and The Christmas Chronicles) or suddenly culturally relevant titles (such as Pandemic: How to Prevent an Outbreak), kids and family movies (such as Cocomelon) during the school holidays, and the return of a Netflix Original title like The Kissing Booth to the top 10 list in the week that its sequel was released. This points to Netflix’s evident need to keep subscribers interested with a constantly revolving library of new and old titles. The importance of new additions as well as new commissions rationalises the giant deals SVODs have done for library rights to legacy shows (Horton, 2019) and confirms that, despite its focus on industry-leading original content production, Netflix is still capitalising on the ‘long tail’ (Anderson, 2007) of existing familiar content. This finding also confirms that the interface is extremely influential in user actions. Marketing and design thinking is achieving its aim of making the catalogue feel dynamic and attractive, even though there is a lot of old stuff in there. New-release promotion via the Recently Added and Trending rows and carousels (and perhaps ‘New to Netflix this month’ advertising campaigns via social media) is clearly effective, as the top 10 list has become an echo chamber of these features.
Another key finding from the top 10 data substantiates very early claims about the lists (Maas and Baysinger, 2020): they verify the popularity of Netflix Original-branded content. There was never less than at least four Netflix Originals in the overall top 10 list on any given day over the period considered. Of the total number of unique titles that rotated through the top 10, 211 (57%) were Originals. These branded commissions and exclusive acquisitions are core to the service’s business model (Afilipoaie et al., 2021; Wayne, 2018) and promoted heavily in the New Release row, in the promotional carousel on the homepage, in email and notification marketing and in paid advertising. Data from the top 10 is evidence that all of this is working. But just as my first key finding substantiated, Australian audiences were engaging the most with branded Originals that were recently added to the catalogue. 96% (202 titles) of the Netflix Originals in the top 10 list were 2020–2021 first releases or new season/episode drops.
Finally, as is common in smaller English-speaking nations like Australia, a preference for American content was also evident in the top 10 list. For the entire year-long period of analysis, titles produced in the United States or with an American production partner made up at least half of the overall top 10 list. In fact, for most of the year at least 80% or more of the top-ranked titles hailed from the United States. That is, in these instances only one or two shows or movies were not American. Only six unique local (Australian) titles rotated through the feature. While much English-language content from Canada and the United Kingdom, and numerous titles from France, Germany, Japan and South Korea (among others) also made the homepage top 10 list, the data suggests that Australians are using Netflix to consume predominantly American content. While the success of Australian-produced content in other Netflix markets remains unclear, this effectively undermines the significant moves that the service has recently made to increase its global production footprint and produce local content. While Netflix executives have maintained that local audiences want to watch local content (Scarlata et al., 2021), this data paints a different picture. Netflix is clearly more local in some markets than others (Lotz et al., 2022).
Evidently, a critical reading of 1 year of Netflix’s overall top 10 data confirms several viewing trends that we may have only previously suspected or assumed. Looking at the top 10 data is a productive exercise in that it has shown the close relationship between subscribers and new to service content, Netflix Originals and – notable from an Australian perspective – an engagement with the SVOD for the purposes of accessing content from the United States. But there are many omissions and silences in the top 10 that challenge its use as an index of consumption generally. In particular, over the period of analysis, I identified a practice occurring within the metric that compels an analysis of the wider role of the mechanism itself. Despite apparently being updated daily, I noted that the top 10 list regularly remained stagnant for days at a time. For an incredible 97 days (26% of the year), I recorded no change in the overall top 10 list from the previous day. To be clear, this inactivity was dispersed over the entire year, with the list sitting unchanged one or 2 days each week. For an additional 29 days over the period examined, the order of the titles within the list shifted slightly, but the films and series remained the same. While a certain amount of movement of titles within the top 10 is understandable – especially down the list over time – I could discern no clear reason for incidents of daily duplication. Repetition did not occur on a particular day of the week or after a substantial drop of new titles to the catalogue. This finding further complicates the reliability of the overall top 10 list: is Netflix intentionally creating a cycle of anticipation and participation here? Why does the service even need to try and derive value from the use of familiar market information regimes like ratings if its algorithm knows each subscriber so well? Having analysed the top 10 data, let us now move on to a more conceptual analysis of the industrial functions of the top 10 feature itself and consider what it indicates about Netflix’s evolving strategic objectives.
Reading the top 10 mechanism
In 2017 Netflix claimed that more than 80% of the television series that people watched were being discovered through the platform’s recommendation system (Plummer, 2017). As a result, much time and rigour has been spent examining the defining role of Netflix’s proprietary recommendation algorithm and its potential impact on cultural decision-making. This has been reasonably dominated by pertinent technologically-determinist commentary on the datafication of identity, algorithmic culture and the mathematisation of taste (see Alexander, 2016; Beer, 2013; Cheney-Lippold, 2017; Dourish, 2016; Hallinan and Striphas, 2016; Striphas, 2015). Until recently, little attention had been paid to the actual user experience of and response to these algorithms. This is unsurprising – practically speaking, this engagement is difficult to reliably capture en masse and SVODs are often black boxes when it comes to sharing this data.
However, while Netflix’s recommendations are still key for the service, new audience surveys have uncovered widespread ambivalence about the overall credibility of algorithmic recommender systems (Frey, 2021). Studies have also identified frustrations about the inaccuracy and repetition of algorithms, lamentations about their limiting or worrying nature and user concerns that they are no longer in control of their exploration (Johnson et al., 2020). These findings have clarified that of the many factors that combine to inform a user’s content selection, word of mouth or recommendations from friends – that is, knowing and wanting to conform to ‘what everyone else is watching’ – contributes enormously to the Netflix experience. In fact, market researchers Conviva identified word of mouth as the top source for streaming content discovery, driving an estimated 59% of decisions to watch content for the first time: ‘Tapping into word-of-mouth marketing strategies to spark interest and discussion is critical for streaming content providers to gain widespread market awareness and supercharge streaming content discovery’ (Conviva, 2021: 13). This is exactly what the Netflix top 10 feature, with its return to a ‘ratings mentality’ (Bourdieu, 1998: 70), has endeavoured to do.
The feature mimics empirical ratings services and perhaps, because ‘these metrics are built by aggregating what ordinary people say or do, they seem trustworthy’ (Webster, 2014: 77). Sharing this data (as opaque and problematic as it may be) contributes to defining ‘the fact of viewing television not as the activity of an isolated individual, but as a shared, common, social activity’ (Bourdon and Méadel, 2014: 14). The top 10 lists also mark a move away from the rhetoric that seeks to position Netflix as global, renationalising the service by highlighting regional consumption and asking, ‘What are people watching in your area?’ With the feature, Netflix openly acknowledges and responds to the pervasive sociality and locality inherent in what has become an increasingly solitary consumption ritual. But it also tries to exploit it as well, to cyclically build hype around titles, compel viewing to counter ‘choice fatigue’ (Ellis, 2000: 169) and troll subscribers into ‘keeping up’ with what’s popular.
As James G. Webster (2014: 76) considered in The Marketplace of Attention, traditional ratings services, what sociologists have called ‘market information regimes’, are highly constructed in that ‘the information they depend on is never completely neutral… the way it’s made can affect the very thing it’s supposed to measure.’ It is important to note that while Netflix’s viewing data is more complex and arguably more accurate than any of the inadequate or limited measurements systems to date, the SVOD is using a mechanism that mimics familiar market information regimes and their apparent trustworthiness. In reality, the methodology and accuracy of the top 10 feature is entirely private and not an industry-agreed upon regime.
There has long been a political and economic dimension to media measurement systems. Ratings are not just a medium through which producers ‘perceive the dynamics of their market’ (Kosterich and Napoli, 2016: 255): these mechanisms also play a vital role themselves in impacting, perhaps even perpetuating, audience engagement as well. They are ‘socio-technical mechanisms’ that are effective for various actors (Bourdon and Méadel, 2011). There is value in making ratings systems available to the public not just to reflect popularity, but create it: Aggregating and reporting what visitors to a website have chosen or what members of a social network recommend introduces powerful signals about social desirability for those who follow. Humans are prone toward herding, and seeing what other people are doing can trigger stampedes... Although popularity is hardly a foolproof guide to what is of greatest importance or highest quality, it seems clear that recommending things that are popular drives traffic and further enhances popularity. Reducing oceans of data to a simple head count and reporting it can inflate the final tally (Webster, 2014: 91; 94).
Netflix’s top 10 lists ostensibly counter algorithm apathy and cultivate audience agency. But in using them, subscribers are also operating – as Dallas W. (Smythe, 1977: 1) famously argued – as a ‘peculiar commodity’ (Smythe, 1977) whose participation is a form of free work for the SVOD, which is selling this data not to advertisers but back to other subscribers to promote even more engagement. This is not new. Eileen R. Meehan (1990) has written extensively about the ‘commodity audience’ and the ways in which ‘economic self-interest restricts and reformulates measurement techniques’ (118), and Ien Ang (1991) emphasised the ways ratings and measurements have been used by commercial and public broadcasting institutions as an instrument of discipline and surveillance. More recently, Fenwick McKelvey and Robert Hunt (2019: 4) assessed how platform vectors ‘encourage users to become part of the process of content discoverability’, and Apple’s App Store rankings and Spotify’s charts were described as ‘aimed at driving up usage’ (Spangler, 2021). However, this is new for Netflix, which has claimed not to need these traditional market information regimes in the past. If ratings metrics are aligned to the strategies of a company, as Webster (2014) put forth, these lists represent a significant shift for Netflix, away from paradigms of hyper-personalised algorithmic echo chambers towards the exploitation of older new-release broadcast models and the social dynamics of viewing that accompanied them.
This supports critical scholarship that has viewed the video-on-demand interface as a ‘site of ideological power and contestation, where their design can determine a user’s behaviour and/or limit their choices’ (Kelly, 2021: 266). The top 10 list is emblematic of how Netflix has tried to ‘downplay the control of the computer that sits behind the interface’ and ‘create an illusion of abundance and plenty… minimise interactivity while creating an illusion of user agency… [and] orient user behaviour towards viewing’ (Johnson, 2019: 112–113). The data presented appears as a convenient and objective collation of the consumption of other viewers, there to help, inform and socialise the user, rather than guide them. But while this is an evidently effective strategy, it is one of which scholars and subscribers should be wary. The Netflix top 10 feature operates as a ‘visible and compelling interactive scripted space… [which] draws attention away from the relatively unscrutinised aspects of the networks and code that makes [Netflix] function’ (Chamberlain, 2011: 243).
Conclusion
In November 2021, Netflix announced a major expansion of its top 10 lists, with the launch of a dedicated website (top10.netflix.com) with weekly country and global lists of the most popular titles (Films (English), TV (English), Films (Non-English) and TV (Non-English)), based on hours viewed. The site also features a list of the platform’s most popular films and programmes of all time, based on the total hours viewed in a title’s first 28 days on Netflix. The SVOD has engaged independent accounting firm EY to audit its metrics and their report will be published in 2022. According to Netflix VP of Content Strategy, Planning and Analysis, Pablo Perez De Rosso (2021), this step is a direct response to the feedback Netflix has received about its metrics over the years and aims to ‘help fans discover new stories and join new conversations’.
The enduring presence of the Netflix top 10 lists and the addition of this new standalone website confirms that subscribers have found value in the feature and continue to engage with it. This analysis of the overall top 10 list over an extended period has demonstrated that there is a close relationship between audiences and new-to-service content, Netflix Originals and titles from the United States. But a reading of the role of the mechanism itself raises arguably more interesting questions about recent shifts in the SVOD’s corporate strategy, namely its need to respond to algorithm apathy and cater to the sociality inherent in the video consumption experience. Netflix may be ‘monopolising the ability to define popular television in the context of global streaming platforms’, but it has not entirely ‘divorced popularity from observable viewer behaviour’ (Wayne and Uribe Sandoval, 2021: 14). Instead, the facilitation of visible communal discovery via an (at present) incontestable top 10 feature is first and foremost helping Netflix exploit the labour of its subscribers.
Footnotes
Acknowledgements
The author thanks Ramon Lobato, Amanda D. Lotz and Stuart Cunningham for their generous editorial input.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was funded by the Australian Government through the Australian Research Council (project DP190100978).
