Keywords

Introduction

In recent years, the growing popularity of services such as Netflix, Amazon Prime Video, and Disney+—along with hundreds of other national and regional subscription video-on-demand (SVOD) platforms that provide access to online libraries of film and television content—has raised complex challenges for policymakers. Established policy approaches in a range of areas including audiovisual licensing, classification, censorship, and local production support are now being disrupted as governments grapple with the “Netflix effect” and its implications for national markets and institutions (Lobato 2019; Kostovska et al. 2020; García Leiva and Albornoz 2021). Meanwhile, consumption practices are also changing as the algorithmically curated interfaces of SVOD services invite audiences to discover content in new ways. In particular, the use of personalised recommendation and other algorithmic filtering techniques has prompted discussion of how SVODs manage the visibility of different kinds of content and whether these discovery environments require a policy response.

This chapter explores how discoverability has emerged as a topic of debate, specifically in relation to SVOD services, and how this is connected to other precedents in audiovisual law and policy such as prominence regulation. We examine how key territories are responding to the challenge of ensuring discoverability of nationally significant content, including public-service broadcaster (PSB) content, local content, and minority-language content. We also reflect on the many tensions inherent in this area of policy—which exists at the interface of media and platform regulation—and consider some of the normative questions raised when governments seek to intervene in audiences’ content choices.

To begin, let us offer a hypothetical example to illustrate what is at stake in this topic. Consider the following scenario: a viewer comes home from work and switches on her Samsung smart TV with the aim of watching the movie Yesterday, which has been enthusiastically recommended by a friend. She opens her Amazon Prime Video app but cannot find the movie in the catalog, due to incorrectly spelling the title in the search bar. She then closes Amazon Prime Video and opens Netflix instead, entering the same search term. Her query generates a screen full of recommended titles including an original Netflix series with a strong female lead (The Queen’s Gambit) which grabs her interest. She selects The Queen’s Gambit and watches several episodes. Exiting the app, she notices that The Queen’s Gambit is now promoted on her smart TV’s home screen, alongside other content recommendations generated by Samsung and its commercial partners. In contrast, recent shows that she has watched on her local PSB do not appear in this recommended row, because these PSBs do not have commercial agreements with Samsung. The following day, when she switches on her smart TV, she notices that both her TV home screen and the Netflix app are suggesting other Netflix originals with strong female leads, including Emily in Paris and Cable Girls. These various recommendations lead her to view and enjoy further Netflix original series of which she was previously unaware.

This hypothetical example illustrates the power of interface design elements, especially personalised recommendations, to guide users in their content choices and thus to shape their media experiences. Within a short space of time, our user’s actions generated a dense, multi-layered discovery environment—a ‘dynamically unfolding, personalized architecture of choice’ (McKelvey and Hunt 2019: 2; citing Yeung 2016)—comprising interconnected recommendation and search functions across multiple apps and the smart TV operating system. Importantly, there is no singular “algorithmic logic” at work here; instead, we find a more complex amalgam of discovery mechanisms that interact to serve particular commercial objectives. These mechanisms include:

  • service-level discovery (the recommender systems, browsing categories, and search engines of Netflix and Amazon Prime Video)

  • hardware-level discovery (the appearance of particular titles on the smart TV home screen after the user has exited the app)

  • active search (querying a specific title)

  • passive or accidental discovery (clicking on a recommended title following an unsuccessful search query)

  • organic search results (relevant to the user’s query)

  • prioritized results (strategic “push” promotion of particular content such as The Queen’s Gambit)

  • personalised elements tailored specifically to the user’s data profile, and

  • universal elements that appear to all users of the SVOD service.

These discovery elements interact in complex ways which cannot be reduced to a catch-all term such as “algorithmic”. Discoverability is more complex than this, because it comprises both human and machine-generated decisions. Strategic objectives underlying the discoverability architecture of SVOD services and the connected TV platforms in which they operate range from subscriber satisfaction, retention and data capture through to promotion of original content and lucrative commercial partnership agreements (such as pre-installation and integration deals between Netflix and TV manufacturers). The policy questions for regulators are therefore multiple and interrelated. Who has power in the above scenario—the user, the SVOD service or the smart TV? To what extent do familiar policy concepts such as media access, choice, selection, and diversity make sense here? What combination of “push” and “pull” characterises these discovery environments? These are some of the challenging questions that underwrite the current policy discussion about discoverability, to which we now turn our attention.

Defining Discoverability

Discoverability refers to the ‘likelihood of discovery’ of particular content within a digital interface, and how this is shaped by ‘industry dynamics, strategies, negotiations and curation’ (Mazzoli and Tambini 2020: 12). In the case of SVOD services, there are two key considerations:

  1. 1.

    how SVOD services present particular content within their home screens, search, and recommendations (e.g., how easy or difficult it is to discover certain titles in your app, and the relative prioritisation of certain titles—such as originals—over others); and

  2. 2.

    how connected TV devices such as smart TVs present SVOD apps and content within their own interfaces (e.g., how easy or difficult it is to find SVOD apps and content in your device, as opposed to other video services).

The first consideration, discoverability within SVOD apps, is an area of policy concern for those countries where there is a policy tradition of prioritising local, national or regional content. The growing take-up of SVOD services in many nations means that certain forms of screen content that have traditionally been of special policy interest—such as national cinema and television, minority- or majority-language content, and documentaries—are no longer guaranteed the visibility that they would have received in the broadcast schedule. A key question for regulators is therefore how to enhance discoverability of this public-interest content within SVOD services, and what obligations on services might be necessary to achieve this aim.

The second consideration, discoverability of SVOD apps and content, is a policy problem in a different sense. This kind of discoverability refers to the competition between different kinds of video services, both public and private, for prime “real estate” within the connected TV device interface—e.g., the largest, most visible tiles on the device’s home screen. This is often referred to as prominence, a concept with a long history within British and European media law, which we discuss in more detail below. While their regulatory contexts are distinct, these two terms (discoverability and prominence) cannot be easily disentangled. From a user perspective the core issue is essentially the same: how the discovery environment is constructed, what content is available, and the differential visibility of this content.

Current design norms mean that discoverability is constructed through a range of different mechanisms including home screens, recommendation rows, content carousels, search results, autoplay trailers, remote control buttons, “previously watched” reminders, and push notifications (Ofcom 2019; Ofcom and MTM 2019). The cumulative effect of these is to make certain content and/or apps more or less visible than others, with the implication that user attention is guided—strongly or weakly—towards particular options. Discoverability in digital interfaces is therefore ‘a new and evolving locus of media circulation power’ (Hesmondhalgh and Lotz 2020: 393) because it holds the potential ‘to direct audiences toward certain kinds of experience and content, and therefore away from others’ (p. 388). Numerous stakeholders naturally have an interest in discoverability and its regulation, including national governments, which prefer particular news sources and linguistic content to be more discoverable than others, for political, industrial or civic reasons; public-service broadcasters, which seek to retain the national prominence hitherto afforded by broadcast spectrum allocation; and civil society groups, which may seek prioritisation of public-interest content including minority-language, community, independent, or culturally diverse programming.

As a policy concern, discoverability is not limited to video; the topic has a larger historical resonance that needs to be considered. For example, debates about the sequencing of stories in newspapers, the selective display of media goods in retail stores, and the filtering of web search—all fundamentally matters of “discoverability”—have long been a feature of media, consumer and internet policy (Grimmelmann 2011; McKelvey and Hunt 2019; Herbert and Johnson 2019). Similarly, today’s debates about fake news, misinformation and disinformation debates can also be described as discoverability problems, in the sense that they involve algorithmic and human amplification of particular voices over others (Gillespie 2018; Noble 2018). Seen from this perspective, discoverability is more than a matter of interface design. At its core, it is about the politics of visibility in media distribution (Garnham 1990; Cubitt 2005; Lobato 2012; Braun 2015).

Politics and Policy in Discoverability

The topic of discoverability is naturally challenging for liberal-democratic media policy because it appears to involve an intervention—algorithmic or otherwise—into individuals’ private consumption. From the perspective of core liberal values such as freedom of choice, such interventions are potentially problematic. However, it is important to note that forms of prioritisation have long been present in national media systems. Public-service media, local content, indigenous content, majority-language content and minority-language content have all been afforded special treatment on public-interest grounds. Enhanced discoverability has historically been achieved through a range of measures, including local content quotas and the establishment of specific channels for minority, multicultural and indigenous content. The policy question at the heart of this issue is therefore not about whether or not prioritization should occur. Rather, it is about who makes the decision to prioritize (the state or platforms); what principles guide such prioritization; and how transparent and contestable such decisions are.

In the present age of “internet-distributed television” (Lotz 2017), the technological conditions that resulted in easy discoverability of PSB content in broadcast television—capacity constraints of broadcast spectrum—are weakening. Internet distribution is mostly governed privately through commercial agreements, terms of service, and software design. Regulators must now decide whether, and how, to extend these earlier traditions of enhanced discoverability into this new environment. Having partly displaced broadcast and pay-TV channels as major audiovisual distributors, should SVODs be subject to the same expectations and obligations that previously applied to those services?

The tradition of prominence regulation in European and British broadcast law is of particular relevance here (García Leiva 2020). Prominence rules govern electronic programme guide (EPG) design and channel numbering, with the goal of maintaining a privileged position in the public consciousness for PSBs and other national institutions. In the United Kingdom, BBC One is allocated the first channel number (1), BBC Two the second (2), and so on. This tradition of prominence rules—which have long considered screen interfaces an appropriate surface for media policy—explains why many countries are hesitant to allow platforms alone to decide these matters. Current debates about discoverability build on these historical foundations, although the concept of prominence appears to be giving way to a wider notion of discoverability that can include interfaces such as voice search, and the diverse ‘routes to content’ (Johnson et al. 2020) on digital platforms. As we now examine, there is no clear consensus on whether and how to extend prominence traditions to encompass SVOD services and connected TV platforms. Different countries are pursuing different models, in line with their own national circumstances.

Regulatory Developments in the European Union, Canada and Britain

Presently, the European Union, United Kingdom and Canada are the key jurisdictions in which SVOD discoverability has been most extensively debated. In each case, specific measures for video services are being implemented or considered. Policy objectives across these territories vary, with emphasis falling on discoverability of different content types including European content, national content, subnational content, and public-service broadcaster content. We now consider each territory and the particular ways in which discoverability surfaces as a policy concern.

European Union

Discoverability debates in Europe take place against a historical backdrop of prominence rules for audiovisual services, an area of policy which has been subject to ‘a vague and heterogeneous implementation’ by different EU member states (García Leiva 2020: 9). Prominence rules have been in place since the 1990s, with an initial focus on the layout of EPGs (van der Sloot 2012; García Leiva 2020).

The 2018 revision of the Audiovisual Media Services Directive (AVMSD) (European Parliament and Council of the European Union 2018) expanded the concept of prominence to include the promotion of European content within online video services. Responding to concerns about US-based SVODs flooding European screens with Hollywood content, the 2018 AVMSD revision introduced new obligations for major SVOD services including Netflix and Amazon Prime Video. Two measures came into effect: a minimum catalog quota of 30% European content, and a requirement to provide “sufficient prominence” for these European titles. The Directive does not mandate specific prominence measures but instead notes several possibilities:

Prominence involves promoting European works through facilitating access to such works. Prominence can be ensured through various means such as a dedicated section for European works that is accessible from the service homepage, the possibility to search for European works in the search tool available as part of that service, the use of European works in campaigns of that service or a minimum percentage of European works promoted from that service's catalogue, for example by using banners or similar tools. (AVMSD 2018 revision, recital 35).

The implementation and enforcement of these prominence rules will be a matter for EU member states. At the time of writing, member states ‘are still in the process of adopting their national prominence frameworks and there are significant differences in the implementations of the AVMS Directive’ (Mazzoli and Tambini 2020: 18–19). Guidance on prominence measures is being prepared by the European Regulators Group for Audiovisual Media Services, which advises the European Commission.

In the meantime, some member states have developed approaches to prominence and discoverability that go beyond the minimum standards laid out in the AVMSD. For example, Germany’s revised Rundfunkstaatsvertrag (Interstate Broadcasting Treaty) specifies a general principle of non-discrimination, such that content cannot be unreasonably hidden, along with an additional provision for positive prioritisation of PSB content and other commercial ‘programmes that contribute to plurality’ (Mazzoli and Tambini 2020: 21).

In summary, the European model of prominence rules—comprising EU-wide minimum standards in the AVMSD, plus additional requirements imposed by member states depending on national circumstance—suggests one possible regulatory template. The European model is not without its problems, of course, notably the inherent power imbalances within the AVMSD’s category of European content (which favours the largest EU nations over smaller nations, due to industry scale and output). Nonetheless, the revised AVMSD offers the most advanced regulatory template currently available and is closely watched by policymakers in other nations for this reason.

Canada

The Trudeau government in Canada has also taken steps towards enshrining discoverability as a central element within audiovisual policy. Current Canadian cultural policy now refers explicitly to the discoverability of national content as a policy objective, not only to its creation and funding. Particular emphasis is placed on discoverability of national content (“Cancon”), including Québécois content.

This policy direction has been building for some time. In 2016 the Canadian Radio-television and Telecommunications Commission (CRTC) held a Discoverability Summit in Toronto. In 2018, it undertook a public consultation about the future of television and released a report (Harnessing Change: The Future of Programming Distribution in Canada) which enshrined promotion and discoverability as objectives of national policy, observing that ‘shifting focus from production alone to include the promotion and discoverability of content will be essential to ensure a vibrant domestic market in the future’ (Canadian Radio-television and Telecommunications Commission 2018, n.p.). A review of Canada’s Broadcasting Act—Canada's Communications Future: Time to Act—was then completed in 2020. Prompted in part by the rapid take-up of Netflix in Canada, the review recommended that major curated video platforms be subject to new obligations including catalogue, discoverability, and transparency requirements, paving the way for new “Netflix laws” to make U.S. streamers more accountable to Canadian cultural policy (Broadcasting and Telecommunications Legislative Review Panel 2020). Nine months later, the Trudeau government tabled draft legislation (Bill C-10) to amend the Broadcasting Act. Bill C-10 sought to introduce a new legislative category of ‘online undertakings’ to apply to hitherto unregulated OTT services, and empowers the CRTC to make orders imposing conditions on ‘the presentation of programs for selection by the public, including the discoverability of Canadian programs’ on such services (Parliament of Canada 2020). At the time of writing, the Bill has not yet been passed into law.

By potentially empowering the CRTC to regulate SVOD content and discoverability, the Canadian approach signals a determination to bring discoverability under the auspices of national cultural policy. In practice, much will depend on the passage of Bill C-10 through Parliament and how the CRTC interprets its principles (and, in the case of Cancon catalog requirements, on the public hearings that will be part of that decision-making process). Key decisions about the details of discoverability regulation—such as whether SVODs need to include a Canadian content row or a minimum proportion of recommended Cancon titles, et cetera—will be deferred to the CRTC.

United Kingdom

While the focus in Canada and Europe has been on the presentation of national or regional content within SVOD services, in Britain the debate has played out a little differently. The key issue has been the relative prominence of SVODs versus PSB services in connected TV devices (i.e., how discoverable BBC, ITV, Channel 4 and Channel 5 content is, in relation to Netflix or Amazon content). In recent years a chorus of voices including PSBs, the UK media regulator Ofcom, and various civil society groups have called for updated prominence rules to ensure PSB content is not “crowded out” in smart TV home screens and other digital interfaces, and can therefore compete with what former BBC Director-General Tony Hall described as the ‘the huge gas giants of the US—the Netflixes, Apples, and Amazons’ (Hall 2017).

British PSBs have long campaigned on this issue, which they see as vital to their future survival. In 2018 Hall warned that the UK was ‘sleepwalking towards a world in which children and young people barely encounter PSB content’ (Hall 2018). Representatives from Channel 4 have stated that reforming ‘prominence is the single biggest thing we need to do to safeguard PSB in this country’ (Milton 2020). After a detailed review, Ofcom released a set of recommendations in 2019 for a new framework that ‘safeguards the discoverability of PSB linear channels on the homepage’ (Ofcom 2019: 37). The aim of these proposals was to ensure that BBC and other PSB apps are guaranteed prime position on the home screens of all connected TV devices sold in the UK, including TVs and game consoles. Recognising the new importance of voice search and recommendation to discovery, Ofcom also recommended that ‘the new prominence framework’s definition of PSB on-demand services includes disaggregated PSB content (e.g. in recommendation and search results) because these routes to content are likely to become more important to viewers over time’ (40). Later that year, a House of Lords Committee Inquiry and Report (Public service broadcasting: As vital as ever) endorsed Ofcom’s findings, calling for a new, legislated ‘prominence framework in line with Ofcom’s recommendations’ (House of Lords Select Committee on Communications and Digital 2019: 60). At the time of writing, no such legislation has yet been introduced into parliament. It therefore remains to be seen how much of this policy agenda will make its way into legislation in the United Kingdom.

Regulatory Options

These ongoing debates in Europe, Canada and the United Kingdom reflect a growing awareness of the importance of discoverability and prominence in media, audiovisual and internet regulation. Elsewhere, debate about these issues is also brewing. In Colombia, regulators have decreed that SVODs must ‘make Colombian works easily available and clearly identifiable in their catalogues’ (García Leiva 2020). The Australian government is requiring SVOD services to report on local content provision and is considering introducing discoverability rules to enhance local content discovery in SVOD interfaces (Australian Communications and Media Authority and Screen Australia 2020).

Clearly, there is no one-size-fits-all solution because the policy objectives underlying these various proposals differ from country to country. In Canada and Australia, for example, the emphasis is on visibility of local content in SVODs. In the EU, the AVMSD revisions focus on visibility of regional (European) content. In the UK, the debate focuses on discoverability of PSBs in relation to SVODs.

Importantly, all the various policy approaches discussed above recognise the need for regulatory flexibility. Regulators are rightly wary of introducing measures that will be difficult to implement or risk rapid obsolescence due to changes in technology and business models. None of the aforementioned proposals in the UK, Canada or EU have established specific discoverability requirements in legislation; instead, they lay out general principles and examples and leave the finer details for regulators to decide, often in consultation with industry. This seems the most appropriate model for effective regulation, although it relies on the capacity of media regulators in each country to develop and enforce appropriate measures.

Reading across these three case studies, we see that there are a range of possible policy mechanisms which vary in their costs, operational implications, and degree of controversy (Table 11.1). At one end of the spectrum we find some “easy options” such as requiring metadata labelling of local content to enhance searchability (e.g., when a user searches for national content). Encouraging SVOD services to maintain consistent and detailed metadata on titles is relatively uncontroversial and benefits all parties. Our research in the Australian context has consistently found that although national labelling in metadata varies enormously between services, good quality metadata makes a significant difference to discoverability of national content via search (Lobato and Scarlata 2019, 2020). Other relatively easy options to enhance priority content discoverability include dedicated landing pages and curated content collections, which can be achieved fairly easily by most SVOD services.

Table 11.1 Mechanisms for enhancing discoverability
Full size table

The more challenging measures are those that require changes to system design, especially search and recommendation algorithms, or disclosure of commercially sensitive information. Such changes are fiercely resisted by SVODs and connected TV platforms. Submissions to the UK House of Lords Public Service Broadcasting (PSB) in the Age of Video on Demand Inquiry (2019) by Samsung, LG and Sky argued that introducing prominence/discoverability requirements will degrade the user experience, adversely affect interface personalization, and undermine system integrity. These companies claim that heavy-handed discoverability regulation will impose unreasonable costs for industry and will dampen future innovation in service and interface design.

For example, the industry alliance Digital Europe—which represents technology and consumer electronics firms—argues that ‘device compliance requirements must be light touch and not prescribe how CE [Conformité Européenne] manufacturers design their UIs, which advanced features must be included, nor excessively define performance capability’ (Digital Europe 2016: 1). Netflix has also argued against regulatory intervention in its recommender system. In its submissions to government inquiries Netflix emphasises the operational integrity of its recommender, which ‘provides a personal experience that allows members to discover the most pleasing titles based on their personal preferences’ (2017: 2), and warns against government intervention in these systems. It argues that Netflix’s own discoverability design (including its recommendation algorithms) is the best way to ensure unbiased and effective matching of content with consumer tastes.

Challenges and Pitfalls

The challenges inherent in imposing national and regional laws on global technology firms should not be underestimated (Flew et al. 2016). Major global SVODs such as Netflix and Amazon Prime Video design their systems at global scale and have little time for national regulation that departs from industry norms of product and software development. Likewise, commercial agreements about pre-installation and prominence of apps on smart TVs are typically negotiated on a global basis between manufacturers and SVOD services. For example, a particular SVOD might pay Samsung a certain amount of money to have their app pre-installed and highly discoverable on all Samsung TV sets for a certain period (Ofcom and MTM 2019). For these companies, the prospect of particular nation-states introducing discoverability or prominence rules that might undermine these agreements is highly undesirable. Depending on the compliance costs, they may simply ignore national regulations.

Another option is for firms to withdraw from particular markets where regulatory burdens are too high. Google—whose Android TV is one of the major platforms that would likely be impacted by renovated PSB prominence rules—has employed this negotiating tactic of “play by our rules or lose our services” several times already: in Spain, with the withdrawal of Google News in 2014; in Denmark, with the removal of Danish music from YouTube following a dispute with collecting societies in 2020; and in Australia, with the threatened withdrawal of Google Search, in response to the government’s mandatory news media bargaining code for digital platforms, in 2021.

While large jurisdictions have an obvious advantage here over smaller nations striking out on their own, the prospects for ensuring compliance by the major SVOD services and connected TV platforms are by no means certain. There is no guarantee that discoverability rules introduced anywhere in the world will be enforceable in any straightforward or consistent way. Yet the alternative—doing nothing—is also unappealing to many countries. Given the rapid migration of audiences from linear services to SVODs and the widespread take-up of connected TV hardware, governments realise that they cannot simply opt out of the discoverability space without compromising longstanding media and cultural policy objectives.

Other Considerations in Discoverability Regulation

As the discoverability debate evolves, it will be important to look beyond the loudest voices (such as PSBs and the major SVODs) and consult other constituencies whose views have so far not been widely heard. For example, one group to consider is content creators. Do film directors, for example, welcome the prospect of special treatment for their content in SVOD interfaces, or do they prefer to roll the dice and see it succeed (or fail) on its own strength within the SVOD catalogue or platform? In the case of small-nation creators, do they want their content to be included in special “local content rows” or similar—or would this be an unwelcome form of ghettoisation that might turn off viewers? How do they feel about their work being labelled as “local content” as opposed to “premium originals”? These are some of the delicate considerations that need to be factored into decision-making.

Another constituency in these policy debates is the audience itself. Do audiences want particular content types to be prioritised over others? How do they feel about institutions—whether platforms or governments—intervening in their content choices? There is surprisingly little empirical research on audience attitudes to discoverability and prioritisation, with the effect that these attitudes are not well understood. Empirical audience research in the UK and US points to a remarkably diverse array of discovery practices among audiences, suggesting that legacy promotion (including word of mouth and recommendation from friends) is often as consequential as algorithmic recommendation (Johnson et al. 2020; Frey in press).

Policy debates about discoverability must also take into account the larger ethical and sociological dimensions of the topic. Regulation of discoverability is inherently controversial because it appears to involve intervention into the private content choices of citizens and consumers. Constant reflection is therefore required on the rationales and mechanisms for such intervention. This includes acknowledging the ideological tensions inherent in the idea of regulating discoverability, as well as the frictions between different policy objectives.

For example, García Leiva (2020: 11) notes the inherent contradiction between cultural policy and competition policy objectives, observing that regulation cannot ‘effectively guarantee prominence for certain contents without colliding with other objectives (notably those related to competition), nor, on the contrary, could a light-touch approach be possible without putting a strain on other principles (notably the protection of diversity)’. As García Leiva observes, there are risks on both sides. A pure laissez-faire model is also likely to be unsatisfactory over the long term, because the kinds of nationally significant content accorded special value in liberal-democratic broadcast regulation—including local, minority and diverse content—may not secure the same protection afforded by prior broadcast laws. At the other extreme, heavy-handed intervention into discoverability has many risks. It may end up annoying consumers, adding extra costs for industry, and will inevitably involve “picking winners” among services or content, in ways that may clash with prevalent values of choice and freedom in the online environment, and raising the twin spectres of paternalism and propaganda.

The political context around discoverability regulation is crucial. So far, the territories that have most actively developed discoverability policy for SVODs and other video services are all liberal democracies: the European Union, Canada, and United Kingdom. However, we cannot take a liberal, pluralist cosmology for granted. There is, of course, a long tradition among illiberal states of enhancing discoverability of national propaganda. Hence the outcome of policy interventions into discoverability is likely to be determined by the social and political context in which such policies are developed. Future discoverability policy may conceivably serve nationalist rather than localist objectives or may aid propagandists rather than PSBs (categories which are not always mutually exclusive). This is not, in itself, an argument against policy intervention—because the costs of inaction may be just as great. It does, however, remind us that rationales for and risks of intervention are not stable from country to country. Policy mechanisms must also be rigorously defined to minimise future abuse and scope-creep.

This noted, discoverability should not be conceived as a zero-sum contest between state and market. Both state and market are capable of positive and negative “discrimination”, for good or ill. Nor should we begin with a romantic idea of the sovereign user as existing outside the distribution system, because any system is always-already constructed by the range of available choices, and personal taste develops in co-dependency with these options and choices. Hence policy debates should reject the purist notion of either the state or the platform “biasing” personal choice, but instead should proceed from the understanding that our choices can never be disentangled from the underlying systems—both human and algorithmic—that construct the range of available options.

As legal scholar James Grimmelmann (2011) argued in relation to search engines, efficient distribution of information always requires some kind of discrimination and thus some degree of enhanced or degraded discoverability. The notion that any algorithm—or any media distribution environment—can be fully free of “bias” is inherently problematic. This is, perhaps, one area where policy debates may benefit from the insights of media, cultural and communication experts for whom such cultural contradictions are a core business.

In conclusion, all forms of discoverability policy—and indeed, all discoverability design features—involve some kind of discriminatory intervention into the realm of consumption. Yet this realm of consumption cannot exist—indeed, is not conceivable—outside of discrimination. The question is therefore not about whether discrimination should occur, but by whom and according to what principles such decisions should be made. In the case of SVODs, we would conclude that maintaining some limited capacity for state intervention in discoverability is essential, precisely because these environments have already been—and will continue to be—organised to serve commercial purposes and not merely the personal preferences of users.