People’s viewing habits are changing. The popularity of streaming television is rising, while the reach of broadcast TV has shown a slow, but steady decline.1 Consequently, the video-on-demand (VoD) services of British Public Service Broadcasters (PSBs) are being compared to subscription-video-on-demand (SVoD) services like Netflix or Disney+, who dominate the streaming platform economy.2 But unlike SVoD, broadcaster video-on-demand (BVoD) services are not guided solely by market competition, prompting us to investigate whether the public service values of linear broadcast television are being replicated on BVoD services.
At the ‘Building an Evidence Base for Public Service Media and Audiovisual Policymaking’ (Building an Evidence Base for PSM) project, we are developing a systematic, computational approach to understand what kinds of programmes are available, visible and discoverable on the VoD services of PSBs. We focus on ‘genres at risk’, such as programming about arts, religion and international issues and cultures, which are central to the social and cultural mission of public service broadcasting.
Our work is one of the first attempts to measure and monitor the provisions of these programmes on the VoD services of PSBs. It constitutes a timely response to the 2024 Media Act, which marks the biggest change to UK public service media (PSM) legislation in two decades.3 Specifically, the Media Act introduces prominence legislation designed to ensure that PSBs’ VoD services distributed on connected platforms (such as smart TVs) are easy to find. For PSB VoD services to meet this requirement, they must make a significant contribution to the relevant PSB’s remit, and the programmes included to fulfil this remit (‘public service remit content’) must be readily discoverable and promoted within the PSB app.4 At the same time, the Media Act replaced specific genre requirements previously set out in the Communications Act 2003 (c. 21) s. 264(6),5 that required PSBs to provide programming on subjects such as science, education, arts and culture, as well as for religious and international programming, with a requirement for the UK PSB system as a whole to provide access to ‘an appropriate range of genres’.6 This raised concerns that the provision of certain ‘genres at risk’, which tend to be provided almost solely by PSBs and are central to fulfilling the social, cultural and civic function of PSM, would decline and prompted calls for Ofcom to monitor the genre provision of PSBs on their linear and VoD services.7
There is no industry-wide or regulatory standard for assessing whether public service remit content is appropriately offered to viewers on VOD services
Both changes introduced by the Media Act demand new methods to monitor the extent to which PSBs are fulfilling their remits on their VoD services in terms of the range of genres that they provide access to and how prominent and discoverable these genres are. We are developing computational methods to evaluate the availability of ‘genres at risk’ in the UK PSBs’ VoD catalogues, and their prominence on their VoD interfaces. We track changes on the interface over time and across four accounts personalised for different content preferences. This enables us to monitor what kinds of programmes occupy the most prominent positions on the interface, and how this is affected by personalisation. By analysing programme metadata such as genres, titles or descriptions, and testing the platforms’ search function, we also assess programme discoverability across PSBs in the UK.
Evaluating PSM provision on VoD services
The shift from broadcast television to VoD has fundamentally altered how audiences discover content. Where broadcast television relies on scheduled programming and curated channels, VoD services combine editorial curation with user-specific personalisation and algorithmic recommendation, tracking what users have watched and trying to gauge what they might like next. Unlike broadcast television, VoD services aren’t already playing a programme upon being switched on. Rather, the user encounters an interface offering a catalogue of programmes ordered, typically, through a combination of human and algorithmic curation. Some programmes are higher up on the page, some programme thumbnails are bigger than others, and most available programmes need to be navigated to through the user interface. The design of the user interface, therefore, can exert influence over which content is more or less likely to be selected by users.8
User interface design, combined with people’s changed viewing habits, impacts their likelihood of discovering new content. Consequently, researchers, regulators and the PSBs themselves are grappling to understand how PSBs should fulfil their remits to ‘inform, educate and entertain’ in the era of streaming television.9 Editorial decisions are accompanied by algorithms and take place across more than twenty rails, which can be of various sizes and move across the homepage, each containing anywhere between four and fourteen programmes, as well as hundreds of programmes stored in the VoD catalogue, accessible through search, an A-Z list or genre categories. The functionality and design of VoD platforms introduce additional aspects that influence whether a programme is readily accessible for a viewer. Thus, there are more variables shaping how people can access and find a diverse range of content that explores shared and societally important issues. While the Media Act aims to address these changes to the provision of PSM, it introduces new challenges of monitoring whether and how PSBs are fulfilling their remits on VoD. If PSBs are to remain accountable to and trustworthy for the public whose interests they serve, regulators and the industry at large need a systematic approach to evaluate whether public service remit content is prominent and easily discoverable to audiences.
Our research addresses a pressing gap: there is no industry-wide or regulatory standard for assessing whether public service remit content is appropriately offered to viewers on VOD services. The Building an Evidence Base for PSM project arose from close cooperation with the International Broadcasting Trust (IBT), the Sandford St Martin Trust and Campaign for the Arts. These three charities monitor how PSBs fulfil their role of informing, entertaining and educating the public through international, religious and arts programming respectively. While previously it was possible to analyse broadcast television schedules to determine whether and how such programming was offered, monitoring PSM provision now requires methods to evaluate VoD services.
It is vital to understand what ‘readily discoverable and promoted programming’10 means in the platformised viewing ecosystem. While all programmes present on the platform are technically available should the user click on them, they are not offered equally. A programme located at the top of the homepage gets watched many more times than one on the thirtieth page of the A-Z catalogue. Even the homepage itself contains more visible placements at the top and less visible ones further down, to which users must scroll downwards, or to the right, to view. The amount of time the user is willing to spend selecting a programme is not unlimited either. Research has shown that a viewer will flick through and scroll for a finite amount of time before either coming across a programme to watch, or not watching anything at all on that particular VoD platform. 11
Therefore, individual parts of the interface hold different significance in user journeys to discover new content. In addition to simply identifying whether a programme is present somewhere on a VoD service, it is essential to identify where exactly it is present, how big its thumbnail is and whether it is searchable.
Methodological and technical challenges
In the Building an Evidence Base for PSM project we are developing computational methods to evaluate how PSBs fulfil their civic duty to provide an ‘appropriate range of genres’. We have been evaluating all four of the main UK PSBs’ VoD services – BBC iPlayer, ITVX, Channel 4 and 5. We analysed the VoD catalogues to map the availability of ‘genres at risk’ and tested the search functionality on each VoD service to understand programme discoverability. Currently, in the last methodological stages of the project, we focus on how programmes are visually represented and ordered on the homepage to understand prominence.
The first challenge in assessing the provision of ‘genres at risk’ on the UK PSBs’ VoD services was to develop methods to identify specific kinds of programmes. The particular genres we are focusing on are not always represented in the metadata or categories used by the PSBs themselves, and a framework for locating ‘genres at risk’ is yet to be established. We piloted different computational and mixed methods approaches to tackle the size of the VoD catalogues, encountering technical shortcomings and inconsistencies that could obstruct future inquiries into genre availability.
Inconsistent categorisation
The most significant one for monitoring the availability of programmes and the range of genres offered is an inconsistent categorisation of programmes into the existing genre categories on the VoDs. Unlike SVoD platforms such as Netflix, PSB VoDs do not employ an elaborate system of categorising their programmes into thousands of genres.12 Although we knew we would not be able to navigate to our programmes of interest by relying solely on the handful of genres used by the PSBs’ VoD services, we found that the broad genres currently used on the PSBs’ VoDs could not be relied on to identify our genres of interest.
Unless organisations have in-house expertise in coding, platform analysis and media studies, they cannot monitor programmes on VOD platforms
Our catalogue analysis revealed how overused the factual/documentary genre category is on PSBs’ VoDs (e.g., ‘factual’ on iPlayer, or ‘lifestyle & documentaries’ on ITVX) and that programmes are not always categorised according to their commissioning purpose. For instance, ‘Growing up Jewish’, a programme following young Jews through religious and cultural practices such as bat and bar mitzvah that aired on BBC 1, was assigned to two iPlayer genres – ‘factual’ and ‘religion and ethics’. This programme was commissioned by the BBC’s Religious Commissioning Unit. Another recent programme, also commissioned by the unit, is ‘The Holy Land and Us – Our Untold Stories’, exploring Palestinian and Jewish perspectives on the founding of the state of Israel. This programme was assigned to iPlayer’s ‘religion & ethics’ genre only, even though all other webscraped metadata pointed to its geopolitical nature. This inconsistent categorisation, that seems to refer neither to programme commissioning nor description, complicates efforts to analyse the genre metadata within the VoD services at scale.
An absence of data-sharing
The significant differences in technical functionality and structure between the individual VoD platforms posed another challenge we encountered in gathering information about the current offering of ‘genres at risk’. Each VoD service organises programmes and programme metadata differently. As webscraping tools interact with distinctive elements of webpages, we’ve had to build two custom scripts per VoD – one to acquire catalogue information to monitor programme availability and another one for the homepage to monitor programme prominence. This creates an entry barrier: unless organisations have in-house expertise in coding, platform analysis, and media studies, they cannot monitor programmes on VOD platforms.
These technical barriers to data access make PSBs’ cooperation crucial for data sharing. Our charity partners, who monitor ‘genres at risk’ programming through broadcast TV schedules, are unable to do so on VoD platforms. Currently there are no established data sharing practices. While Channel 4 has been willing to share programme metadata when directly approached, new standards for cooperation and systematic data transparency are necessary to maintain the existing quality of academic and policy research.
Vision
The biggest industry-wide contribution of the Building an Evidence Base for PSM project will be a knowledge base, outlining what data is available on the VoDs and which methods can be used to collect and process it. This is one of the very first efforts to establish methods of independent scrutiny in response to the recent legislative changes in the UK. With streaming television rising in popularity and prominence legislation being debated and implemented in other countries, these approaches will and should continue if public service broadcasters are to remain accountable to the public. It is for this reason that we don’t see the methodological and technical barriers we’ve encountered only as operational difficulties. The resistance of broadcasters towards data sharing raises questions about accountability. If they are to meet their public mission, they must allow for independent evaluation of their services.
We are currently in the last methodological stage of the project, devising a prominence scoring framework for the VoD homepage interface. In the absence of a standard industry-wide framework, our prominence scores are informed by media and interface studies,13 and our exploratory personalisation research. Using our four different accounts for each of the four British public broadcasters, we conduct daily data scraping to score programme prominence based on the position of a programme on the homepage of the interface, programme’s position within a carousel, and the visual prominence of a programme’s thumbnail. Our scoring system will account for the distinctness of these carousels, as user interactions with them differ from the patterns seen on the rest of the homepage.14 We rely on the theoretical and empirical findings of previous research, our own exploratory analysis and short-term aggregated user-interaction data confidentially shared by the BBC and catalogue data shared by Channel 4. Other PSBs have not responded to our requests to share data, comment on our findings or cooperate in this research. This makes our efforts to develop a robust and industry-wide measure or framework more challenging. If we are to rely on broadcasters to report on their fulfilment of public service obligations, we risk creating a system whereby they effectively assess their own performance. This lack of independent oversight could be seen as antithetical to the principles of public service and could undermine public trust.
To move forward, we recommend the development of a national framework for assessing VOD programming prominence, co-created by academic researchers, PSBs, charities and regulators such as Ofcom. This framework should include:
- Clear methods for identifying ‘genres at risk’
- Standardised metadata practices across platforms
- Transparency around interface curation
- Data access for public institutions and publicly funded research initiatives
and then