Introduction

The developing topic of prominence and discoverability of content on video streaming services is of growing interest to both media research and media policy in Europe and beyond. In the European Union, the promotion of European works has extended from content quotas to requirements for the prominence of said works in online catalogs, a process fueling ongoing debates over content prioritization and cultural diversity.⁴ Analyzing and assessing the prominence and discoverability of audiovisual works in streaming catalogs present numerous challenges, including the volume and complexity of data and metadata used to organize offerings, the lack of public archives on publishing practices, the difficulty of assessing routes to individual consumption, limited transparency and accountability regarding personalized recommendations and algorithmic filtering, and self-preferencing exercised by powerful players.⁵

As on-demand content circulation evolves, media policy research, media industries studies, and specific subfields–such as television, film, and production studies–are increasingly integrating or entirely shifting to interface analyses, often in combination with catalog analyses and audience research.⁶ The sheer volume of titles and associated metadata in streaming services has accelerated the adoption of digital methods for data collection, visualization, and analysis. While automated analyses offer invaluable insights, they often depend on standardized fields to enable consistent cross-case comparisons. This contribution argues that a close reading and thorough contextualization of case studies prior to (or in combination with) automation are essential to ensure accuracy, relevance, and comparability.

To this end, we present a set of reflections based on an exploratory study conducted on four subscription-based streaming services in two countries and propose a series of methodological considerations to be integrated in future assessments of prominence and discoverability tools and practices. The study was conducted in two phases on the following US-based services: Netflix, Prime Video, Apple TV+, and Disney+ in Spain and Belgium. In the first phase, we mapped all European and domestic works added to the catalogs throughout the month of October 2023 and examined their placement on the interface the week after they were added. In the second phase, we traced the placement of a set of European titles recently added to the Spanish catalogs during the month of November 2024. We also explored, in detail, additional features that could contribute to the discoverability of European works, such as special collections and the search function.

The findings were closely analyzed in light of each service’s content strategies, consumption affordances, interface features, and market context. The limited but focused sample facilitated manual data collection and cleaning and allowed us to take a close, analytical approach to the subject matter, providing rich industrial context and attention to key differences between the respective services.

Thus, we argue for more “close contact for context” before or in combination with setting up automated data collection and comparative analyses to assess the prominence and discoverability of European works in on-demand catalogs. We strongly believe that a deeper understanding of the streaming services as individual cases, their goals, their unique features, and their specific contexts–be it cultural, regulatory, or market-related–needs to be integrated in catalog and interface analyses to account for variations in curation, prioritization, and strategy. In this vein, we also propose the concept of a title’s “discovery arc” to underscore the importance of going beyond static snapshots of user interfaces at specific moments, to trace the spatiotemporal patterns of prominence. A title’s dynamic placement over time informs us not only of their potential popularity in a specific cultural/market setting but also of publishing and curation strategies employed by individual services.

Catalogs, Interfaces, and Paths to Content Discovery

The notions of prominence and discoverability are not new to social sciences, although they constitute a fairly new territory in audiovisual research, having only been addressed in the last decade. Whereas prominence involves the prioritization of certain content/services over others in time and/or space on a given interface, discoverability refers to the likelihood of coming across such content/services, also serendipitously.⁷ To illustrate, content located in the hero-board of a streaming service’s homepage benefits from enhanced prominence; original/exclusive titles that are presented through larger tiles are prioritized by the service; while discoverability applies to content found by users through various (other) means, including personalized recommendations, scrolling, or searching. Thus, although prominence may contribute to discoverability, the discovery process does not depend exclusively on a prominent placement. Discoverability and prominence should be understood not only as outcomes of prioritization processes, defined as the range of design and algorithmic decisions that positively discriminate and promote content,⁸ but also as concepts that are connected to long-standing concerns over power imbalances in media circulation.⁹

A valuable body of existing media research has already focused on video streaming services and publishing practices. As documented by Lobato and colleagues,¹⁰ research focused on the notions of content and exposure diversity¹¹ has developed and applied various forms of catalog and interface analyses. Television scheduling studies, notably by Bruun and colleagues,¹² have applied similar approaches to researching publishing strategies on streaming interfaces. Challenges related to prioritization, personalization, and archiving have also been interrogated by Kelly and Sørensen.¹³ Although their work differs from ours in the use of automated, quantitative methods, their focus on longitudinal analyses of on-demand interfaces also informed our development of the “discovery arc” concept. By combining quantitative and qualitative approaches to conduct a close reading, we focus on content origins and genre, categorization logics, and the multi-territory market dynamics and publishing strategies of four streaming services.

From a media policy perspective, the themes are timely due to extensive debates in Europe generated by the implementation of the prominence provisions in the 2018 Audiovisual Media Services Directive (AVMSD),¹⁴ and the discussions leading up to its potential revision, in 2026. According to the European Regulators Group for Audiovisual Media Services (ERGA),¹⁵ most used prominence tools by media service providers were dedicated sections on the homepage, and the promotion of European works in promotional campaigns and in catalogs, for instance via banners.

Content prioritization and exposure diversity analyses do not determine a causal link between the position of any specific content/service in a catalog/app/site and its actual consumption. Studying prominence mechanisms sheds only partial light on discoverability processes, as enhancing visibility does not guarantee discovery. Moreover, what is made prominent by players can be highly dependent on different uses and personalization. This leads to a variety of research barriers, especially in terms of methodology. Our study aims to fill some of these gaps by proposing three sets of qualitative considerations to support researchers in setting up comparative analyses, after the methodological explanation that follows.

Methodological Design

To reach the study goal, we explored a number of quantitative and qualitative approaches to assess the prominence and discoverability of a number of titles added to four streaming services in Spain and Belgium, during October 2023 and November 2024. For the purposes of our study, new accounts were created in October 2023 on Netflix, Prime Video, Disney+, and Apple TV+, using internet browsers on laptops for all services. The follow-up study in November 2024 was conducted only on the Spanish catalogs, as this provided a wider pool of domestic works whose “discovery arc” could be traced over a 4-week period. The accounts were used for no other purpose, and no content was actually watched. Throughout the study, we worked with multiple language profiles, namely, Spanish and English in Spain, and Dutch, French, and English in Belgium.

For the mapping, we collected data from JustWatch.com on all the titles added to the four catalogs in the respective markets, prior to making observations on their placement and visibility. We focused on European and domestic works, following the definition used in the AVMSD (including European co-productions with the United States) and the Lumière VoD database. We collected additional data from both JustWatch and IMDb, including title, release year, format (film, series, animation, documentary, entertainment, etc.), country(ies) of origin, language(s), and number of episodes and seasons added.

In October 2023, twice per week, we searched for all the European works added to the catalogs in the previous week, noting their exact location (e.g., hero-board, rows, and intermediary banners) and adding observations on the labeling of the respective rows and tiles. We recorded our searches by taking screenshots of the full landing page and the two successive pages that loaded after scrolling twice to the right. In 2024, we identified between five and fifteen European and domestic works added to each catalog (depending on the service) in the last 10 days of October and traced their placement and “discovery arc” throughout the month of November. We also experimented with the search tool on each service and explored the special collections or categories dedicated to European/domestic works, where available. This was conducted at the end of each study month to limit personalization and avoid influencing the placement and positioning of titles.

Building on existing knowledge in media industries studies and supplemented by our study, we propose three sets of considerations relevant to conducting similar comparative analyses that refer to the object of study and its context. Although the following sections focus primarily on qualitative observations and considerations, our approach does combine findings drawn from quantitative mappings. The catalogs were the starting point of our exploration and comparison, and they provide essential data on catalog size, proportion of European titles, and catalog structure.

Contextualizing the Study: Market Dynamics and Comparative Designs

Previous investigations have pointed out significant differences in service provision and content offerings across markets,¹⁶ ownership structures,¹⁷ and regulatory frameworks.¹⁸ Understanding market characteristics and the dynamics between them is essential not only for conducting comparative research but also for shaping the strategies of transnational players in each market, regarding catalog offerings, investment in original/exclusive content, collaborations with domestic players/creators, language localization practices, and content prioritization.

In the European context, imbalances between market sizes and geographical characteristics have persisted through the years, arguably exacerbated by the strategies of US-based streamers.¹⁹ To this end, Spain and Belgium are interesting settings for an exploratory study because they present a mix of differences and similarities that can illustrate existing patterns within the European audiovisual industry.²⁰ Whereas they differ in market size (large v. small) and the level of audiovisual production (substantial v. comparatively limited), they share linguistic diversity and subnational regulatory and policy interventions.

Market dynamics and content strategies inform the investment made by services in licensed versus original/exclusive offerings. These titles are powerful forms of branding and distinctiveness, likely to benefit from higher visibility on the interface through positioning, labeling, appearance (e.g., larger tiles), or “nudges” (e.g., auto-play).²¹ Thus, the amount and focus of investment made in original European and/or domestic content can determine the extent to which these titles are prominent and therefore discoverable. Netflix is the company that has invested most in European content–and in a wider diversity of markets–followed by Prime Video, Apple TV+, and Disney+, which have focused primarily on the Big European Five: the United Kingdom, France, Spain, Germany, and Italy.²² According to the European Audiovisual Observatory,²³ Spain has received the most investment in original content in Europe from transnational streamers, after the United Kingdom, while Belgium did not make the top ten countries by streamer spending on original content in 2023. The markets also differ in the weight of players contributing to original content production. This is led by public broadcasters in Belgium (56 percent), with transnational streamers accounting for only 12 percent. The opposite is true for Spain, where streamers accounted for 53 percent of the spending, while the public broadcaster accounted for only 11 percent.

In this context, it was not surprising to find that, among the European titles added to the catalogs in October–November 2023 and 2024, a large number of the ones visibly displayed on the interfaces were labeled as Originals/Exclusives. Even more important, only Originals/Exclusives were promoted in the hero-board, across all services (see Figure 1 for examples). These were generally new seasons of an already popular series, Spanish originals/exclusives in Spain, and UK originals, with the US–UK co-production The Kardashians season 4 (2022–) among the most prominent titles on Disney+ in October 2023. Thus, interface analysis can strongly benefit from data on streamers’ investment strategies across markets, and tracing developments over time, where, for example, the emergence of Central-Eastern European markets as growing semi-peripheral production hubs²⁴ constitutes informative cases to investigate.

Figure 1

Examples of European Original/Exclusive Titles Present in the Hero-Board in October 2023 (Left) and November 2024 (Right).

Licensing deals can also influence the availability and discoverability of titles even within the same country. In Belgium, Prime Video and Netflix present slightly distinct offerings based on the language settings of the users’ profiles, due to the market divide between the French- and the Dutch-speaking communities that operate as separate media markets. In our study, different titles were available in the Dutch-language profile of the same user than in the French- or English-language ones. This highlights the importance of understanding market dynamics and making informed choices regarding automated data collection to account for how language localization and content availability can influence the discoverability of certain titles, even within the same catalog.

National and regional regulations are also key differentiators between European countries, as policy frameworks regarding the discoverability and prominence of European works remain fragmented across markets. Thus, it is essential to note what measures streaming services are required to take, in order to determine whether the evaluations can be compared. For example, not all European regulators specify the existence of dedicated sections/collections on the homepage, or the promotion of European works through banners, according to the ERGA reports.

Understanding the Cases: Industry Practices and Business Models

In teasing out the distinctive features of video streaming services, Lotz proposes four structuring characteristics: geographic reach, library specificity, library ownership, and corporate ownership, imagined as reels on a slot machine, each combination leading to different norms, behaviors, and priorities.²⁵ Johnson also indicates the importance of determining the business origins of services, distinguishing among TV natives, online natives, and content natives.²⁶ For example, services operated by media providers–either online natives (Netflix) or established media conglomerates (Disney+)–have distinct models from the ones potentially used as aggregators or entry-points to wider ecosystems, including hardware and software (Apple TV+), or other services/channels (Prime Video).²⁷ This not only sets different strategic trajectories for the services but also informs their catalog composition and interface architecture in different ways.

Strategies related to the composition of catalogs are essential to the way content is presented and curated on the services. Catalog sizes in Belgium and Spain differed considerably (see Table 1), and catalog strategies also determined wide disparities in the volume and type of content added weekly/monthly to each catalog. This influences not only catalog sizes but also how dynamic the presentation of recently added titles is. Services that add a limited number of new titles, such as Apple TV+, primarily rely on new episodes of ongoing series. This generally keeps respective titles consistently visible on the landing page for weeks in a row, for all users. Services that add a more significant number of new titles each week, such as Prime Video and Netflix, rely more strongly on personalized recommendations for discoverability and prominence.

Transparency around content availability is limited, leading to user frustration and disorientation, as consumption-focused studies have shown.²⁸ Therefore, research on discoverability and prominence must be closely tied to data on available offerings. However, reliable sources remain scarce, and our study identified numerous errors in databases used for this purpose. Sites such as JustWatch, Lumière VOD, and UNOGS.com listed titles absent from indicated catalogs or revised their information shortly after our consultation. Country of origin, genre, and language data also vary across sources like JustWatch, IMDb, and The Movie Database (TMDB), with errors and omissions, particularly in co-productions. This is crucial for studies on content diversity, where accurate listing and correct ordering of co-producing countries help determine the majority production partner. Genres were also inconsistent especially on IMDb, which attributes several (sub)genres to each title. At times, titles either lacked crucial indicators–such as “documentary” or “animation”–or had misleading labels such as “documentary” for game shows or reality TV content.

In line with the previous point, it is also important to consider the policy context, and the close link between the prominence requirements and the quota for European works. In tracing titles added to the catalogs, we often came across independent licensed European titles that never appeared on the interface. As an example, we noted nine short films (10–20 min) from a variety of European countries, added to the Disney+ catalog on the same day in October 2023. These practices–where content that counts toward the 30 percent quota is typically acquired through inexpensive package deals–often involve titles that may be perceived as unappealing to users and therefore not prominently featured on the service. This approach is reminiscent of the “quota quickies” of the 1990s,²⁹ a strategy broadcasters employed to meet the European works requirement under the Television without Frontiers Directive.

Revenue models also shape content strategies and interface layouts. Content is easier to discover in catalogs that show approximately twenty to forty rows on the interface, as in the case of Apple TV+, Disney+, and Netflix, compared to Prime Video’s approximately ninety to one hundred rows (see Figure 2). Thus, content featured at the bottom of the Apple TV+ page could still be more easily discovered than content featured mid-way through the Prime Video page, if we compare the visibility of titles on the respective landing pages. The latter presents a different layout in line with its hybrid revenue model, combining content included in the subscription with a transactional video-on-demand (TVoD) option for titles that are only available to rent or buy. Prime also acts as an aggregator for other channels and services that users can subscribe to in specific markets,³⁰ which are also sometimes featured prominently on the homepage (see Figure 4). Moreover, these external services/channels curate the content presented in their promotional rows/collections, which leads to additional questions over comparative interface (meta)analyses and the individual content strategies behind them.

Figure 2

Screenshots of the Four Interfaces Taken on November 22, 2024, from Left to Right: Apple TV+, Disney+, Netflix, and Prime Video.

More national titles in catalogs, at least in our sample, also clearly resulted in more titles being presented on the landing page. Among the titles added in October 2023, between 15 and 37 percent were European, with great fluctuation between streaming offerings and markets (see Figure 3). Few were domestic, especially in Belgium, where certain services added no Belgian titles in the sample month. Among all the European titles that appeared at least once in the Spanish sample (eighty titles), 34 percent (twenty-seven titles) had Spain as the country of origin. In Belgium, we detected only one title with Belgium as the country of origin, an exclusive 2023 Prime Video film co-produced with France. The study conducted in November 2024 in Spain confirmed the steady visibility of Spanish titles throughout the month, particularly if they were recent releases (2023–2024) of original/exclusive productions.

Figure 3

Number of Titles Added on Each Service per Format and Origin. Own Elaboration Based on JustWatch Data Collected between October 24 and 28, 2023.

This is important to the configuration of comparative studies as the significant differences between country catalogs, in terms of number of titles added but also formats, must be considered in the interpretation of results. Our observations confirmed that serialized content received consistent prominence with each daily/weekly episode added to the catalog in both sample months, as opposed to Netflix’s practice of adding full seasons. The latter’s “discovery arc” mainly depended on their popularity, as discussed in the next section.

Most of the titles we found during both our studies were recent releases (2023–2024) and were located predominantly in the same set of categories of rows and banners. Thus, on all services in both countries, titles were recurrently found in the categories originals/exclusives and novelty (e.g., New to …, Recently added), as well as, to a lesser extent, in popularity (e.g., top ten series in Belgium, trending now, and most watched) and genre/format (e.g., crime series, military, and war movies). Netflix distinguished itself as the only streamer using the origin- or language-related categories as predominant row placements. Localization strategies can thus shape content curation across markets, facilitating the discoverability of domestic/regional content.

“Getting Personal” with the Object

Previous research has confirmed that the organization of interfaces and technological affordances have significant implications for the discoverability and prominence of content and users’ “routes to content.”³¹ Working with catalogs and, specifically, the interfaces of streaming services condition user interactions as “their design can determine a user’s behavior and/or limit their choices.”³² Thus, there are several implications for describing and assessing prominence mechanisms that need to be considered, as interfaces can be seen as sites of ideological power and contestation inextricably linked to the business models and company policies identified in the previous step.

Interface Architectures

Although, at first glance, the interface architectures of the four services seem very similar, it is important to note variances between them that may impact the discoverability and prominence of titles. At the top, all pages have a hero-board, consisting of one or more auto-play trailers, or static banners. Except for Netflix, all other services have a carousel hero-board that moves automatically and can go up to fourteen to fifteen pages on Prime and Disney+ (also called a Super Carousel by Prime³³). This is an important consideration for prominence assessment and one which, we argue, requires further audience research, in order to determine to what extent a hero-board banner situated in the fifteenth position is more easily discoverable than a tile placed lower on the landing page.

Moving further down the page, the number of visible tiles depends not only on the service but also on the device used to access the catalog. Moreover, services do not only differ in number of rows but also in horizontal length of rows, which, on some services, may require a significant number of clicks to navigate entirely. As we conducted our study on a laptop interface, five to six tiles were visible in each row, per page (see Figure 2). The Netflix algorithm places the most strongly recommended titles on the left of each row, unless the system language is set to Arabic or Hebrew, in which case these go from the right to left.³⁴ However, we have limited information regarding the strategies of other services, especially if their own A/B testing would indicate that tiles located closer to the center of the page are more visible/attractive to users.

The presentation of tiles is another key aspect of the interface analysis. The size, shape, and design of tiles significantly influenced the prominence of certain titles over others (see Figure 2). Original productions were generally very visible, either through tiles that are larger or differently shaped, or by having logos/labels to indicate this. In our empirical studies, almost all European titles found in the sample were marked as “Original.” Larger, more prominent tiles were also used for “Top 10” rows and other categories such as “Series to binge” on Disney+ or “Previews for you” on Prime. The visual indicators on Prime tiles became more complicated–and rather confusing–in autumn 2024, when all content available with a Prime subscription received a Prime label, to distinguish it from content available only to rent/buy or on different services entirely (see Figure 4).

Figure 4

Screenshot of the Top (Left) and the Lower Rows (Right) on the Prime Video Homepage on November 27, 2024.

Interfaces also displayed varying levels of dynamism. As we observed changes throughout the study, we noted interfaces that changed their selection of titles every day, even upon refreshing the page, such as that of Netflix. Meanwhile, others were significantly more static, noting limited changes in titles and presentation from one day or even week to another, such as Apple TV+ (see Figure 5). The more dynamic interfaces also changed themes and presentation in line with special events or time of year (e.g., Halloween and Christmas) and also proposed different rows or genre/format categories than previously. In our study, this influenced the visibility and discoverability of some European titles, as more dynamic interfaces often presented more diverse offerings over time, also placing European titles in different category rows. In 2024, Prime Video increased its level of dynamism through the addition of live events, including sports and concerts. These received two rows of their own (“Recently finished” and “Upcoming live events”), but, so far, they do not count toward the European works quota. Another example of enhanced dynamism is the top 10/most watched lists that are updated daily. We also noted that this seems to follow a standard set by Netflix: In October 2023, it was used by three services, but by November 2024, it was present on all four services we studied.

Figure 5

Examples of Differing Levels of Dynamism on the Netflix and Apple TV+ Interfaces. Screenshots Taken on October 6, 2023 (Left), and October 9, 2023 (Right).

In November 2024, we traced the positioning of several European titles that had recently been added to the four catalogs, to determine their “discovery arc” over the 4 sample weeks. The majority of the titles we found in our bi-weekly screenshots started off in novelty categories (e.g., “Recently added” and “New to …”) and rows dedicated to original/exclusive titles, both generally placed at the top of the landing page. As previously mentioned, daily/weekly episode releases ensured that a title remained visible on the homepage for longer. This was less often the case with films or full season releases as the ones that did not reach the popularity categories (e.g., “Most watched,” “Top series in …,” and “Top searches”) in 1–2 weeks after being added to the catalog became significantly less visible.

The Disney+ Original Regreso a la Las Sabinas/Return to Las Sabinas (2024–2025) was an interesting example, as it was the only title with daily episode releases during the sample month (see Figure 6). One of the streamer’s most significant investments in the Spanish market,³⁵ the series is a melodrama made up of seventy 45-minute episodes released daily on Disney+, starting October 11, 2024, and it was featured visibly on the interface during the sample month of November 2024. Not only did it often make the top 10 most watched list, but mid-way through the month, it was also given an entire row for users to catch up with previous episodes (“Catch up on Return to Las Sabinas”).

Figure 6

Screenshots Tracking “Regreso a Las Sabinas” on the Disney+ Interface in November 2024.

In line with our observations, we believe that tracing the placement of titles over time can not only provide valuable insights into content and curation strategies across markets but also be used in connection with consumption studies on the discoverability and popularity of titles across catalogs and interfaces.

Consumption Affordances

Our study made efforts to limit algorithmic personalization of the services, in order to make observations that were as neutral as possible. We know that personalization is shaped not only by user activity and viewing history but also by personal information of the user (e.g., gender and location), devices used, languages, or the time of day. However, the services use different levels of algorithmic curation, as some try to adapt more readily to user tastes than others, which are more static or generic in their presentation. An example of this is Netflix’s strategy to “jump start” recommendations by indicating three liked titles. Netflix also indicates that the titles users engage with more recently will outweigh titles they engaged with in the past, in terms of driving the recommendations system.³⁶ However, the speed with which these adaptations are made may differ between services, in line with our observations on the levels of dynamism in the interface presentation. Thus, it is an important aspect to consider should research compare curation on different personalized accounts across services.

The search tools available are also key to determining content discoverability. In our final search of European content not featured on the interface, at the end of October 2023, a good deal of titles could not be readily found by using the search tools available. Of the four services, Netflix had the most sophisticated and intuitive search tool, leading to the highest return of (correct) results. The search was not limited to the exact title or profile language. Thus, searches could be made not only for titles in different languages but also by content language (“French movies”) or other categories/keywords (e.g., directors or actors). Searches on Netflix generally returned several results alongside the title sought, most probably titles identified by the algorithm as relevant or “similar” to the initial search. Netflix also provides the option to “Browse by Languages” in the top menu, based on original language, dubbing, and subtitles.

Our findings also highlight two important considerations for content discoverability: Both search results and available offerings vary depending on the profile language chosen on a country catalog. On Netflix, this was mainly related to availability determined by licensing deals, as discussed in the previous section. However, on Prime Video and Disney+, this was mainly due to the rather rudimentary search tool available in October 2023, which required the exact title in the language the work was added to the catalog in order to return the correct result. This is particularly problematic for titles in non-hegemonic languages and for multilingual markets because the user cannot always know the language in which the content was added to the service, or its original title. On Disney+, a new search page opened automatically showing predetermined “Collections.” On Apple TV+, results always presented titles in four rows, which include content not necessarily “similar” to the one searched for. By November 2024, the search tools on these services had also improved. On Disney+, searches could be made “by title, character, or genre,” while the Prime results also returned titles in different languages, with several exceptions, but no other search categories or keywords. Apple TV+ also returned correct results in different languages, and even returned several European titles in a top row of “main results” when searching for European series, as well as correct results by genre or actor name.

In line with European policy requirements, many services have also added special collections/pages for European works in their catalogs (see Figure 7). The Netflix catalogs include one “European” (as well as one “Española” in Spain) collection for series and another for films. European content is also widely present on the Netflix homepage, also through dedicated rows that highlight the country of origin and language (e.g., “Películas y series en español” and “Belgian movies and TV”). Prime Video has a section called “Made in Europe” in its top menu, which opens a page with several rows of content.

Figure 7

Screenshots of Collections/Pages for European Works Taken on October 31, 2023, From the Top: Netflix, Prime Video, and Disney+.

However, not all the titles featured were European productions, and it is unclear why titles such as Asteroid City (2023) or American Psycho (2000) were included, other than for their wide popularity or the presence of some in a row called “Discover: Movies dubbed in Spanish.” On the homepage, collections reflect the country of origin on catalogs in Spain but mainly language on the catalogs in Belgium (“Watch in your own language”). In November 2024, a new row was added in the third position entitled “Popular European titles.” Disney+ had a “Made in Europe collection” banner integrated in the hero-board carousel; collections for “Dutch-speaking titles” and “French titles” in Belgium; and “Spanish movies and series” in the main menus and respective landing pages. Apple TV+ presented a minimalistic interface with no upper menu, except for the search tool. No European collections were featured on the homepage, and the ones present were not linked to the origin of the works but to genre or actors.

Conclusion

Based on the observations made during our exploratory research, we conclude by arguing for more “close contact for context” before or in combination with setting up automated data collection and comparative analyses to assess the prominence and discoverability of European works in streaming catalogs. As we have shown, there is still value to be gained from “getting personal” with our research objects, as scholars have not yet found a reliable and practical way of systematically harvesting catalog content and conducting interface analyses. This is particularly complex when conducting comparative case study analyses. Nonetheless, qualitative analyses also benefit from building on quantitative data to further the understanding of catalog sizes, compositions, and dynamics, as a basis for further qualitative interpretations.

To this end, we believe enhanced collaboration could be highly beneficial in bringing scholars with different knowledge and methodological expertise together, as streaming research is further blending research domains. We also applaud efforts to reach a common taxonomy, a set of methods, data sources, and the sharing of datasets through open (institutional) repositories or research networks, which can contribute to broader and more systematic archiving initiatives.

Close contextualization is also necessary to overcome the lack of access to transnational streaming companies, as interviews with industry representatives or on-site observations are nearly impossible to set up. Researchers have already found innovative ways to bypass some of these challenges, by collecting catalog and content data from third-party sources, or analyzing content investment strategies through licensing and commissioning patterns, in lack of financial data.

As already stated by colleagues,³⁷ it is important to conduct research that helps us understand streaming services by interrogating some of the myths they circulate and we, as scholars, potentially perpetuate. As in this study, it became clear that we have certain assumptions and tend to make guesses on the intentions or tactics behind the observations we make. However, we also encountered elements that generated even more research questions. During the November 2024 data collection, a curious finding was the consistent placement of the same title in the Netflix hero-board, on both the English- and the Spanish-language profiles: the US TV series Prison Break (2005–2017). While the rest of the interface was very dynamic and presented a large diversity of content, the title was featured at the top of the landing page in every screenshot we took during the 4 weeks. Although this could be explained by the limited level of personalization we strived for, it is still a confusing finding as it differs from the data collected in October 2023.

This article’s analysis was conducted using blank accounts, although it is important to note that Netflix already prompts some degree of nudging during initial account setup, based on choices made at installation. From the moment a user begins interacting with a service–by watching content, searching for titles, or even making incidental selections–the interface becomes dynamic. Key elements such as title placement, visibility, search suggestions, and category labels are subsequently influenced. While some studies have attempted to address this dynamism through artificial profile training (e.g., reverse-engineering user behavior) or by employing “mystery shopping” methods to map existing or random placements, none–quantitative or otherwise–have convincingly captured the evolving nature of personalized recommendations. This dynamism is crucial to understand if we are ever to measure prominence systematically, particularly for European works, which require clear visibility in increasingly complex and adaptive interfaces. It is also important to consider the potential of personalized recommendations to support the prominence of European works, an avenue that policymakers could explore more extensively, especially if the assessment of existing prominence will evolve in the coming years, hand in hand with possibilities for users to customize recommender systems.

We also believe valuable insights can be obtained by conducting comparisons over time, to determine the evolution of priority-setting at the company level. The streamers underwent some notable changes between the two sample months in terms of presentation and affordances. As the companies are continuously searching for their competitive advantage, they are prioritizing different types of content, fine-tuning their recommender systems and search tools, and adopting various licensing and content strategies.