Keywords

1 Introduction

In 2016, a consortium of academic research labs and a major IPTV provider in Portugal embarked on a multi-year project to create a new Interactive TV (iTV) platform. This project, christened UltraTV [1], was quite ambitious, targeting several research topics, including a disruptive new User Interface (UI). However, one of the most interesting aspects was the focus in fusing linear and non-linear content, from a multichannel video programming distributor (MVPD) with over-the-top (OTT) sources, like Netflix, but also user-generated content (UGC) from platforms like YouTube and even audiovisual content from social networks, like Facebook Video, all in a seamless interface. UltraTV content unification research sought to understand if the users were pleased with the multi-app approach, that started to carry over from the mobile world into the connected TV world [2] or if they were receptive to a much more integrated way to consume their preferred video content. At the time, and despite the constraints of this research project, it was found that a majority of users would prefer an integrated approach [3].

This paper aims to revisit the concept of content unification, raised by the UltraTV project, and compare the most important features of that project with a set of relevant offerings in the current market, to understand how the industry aligned with those concepts that have the potential to keep the TV set as the preferred device to watch audiovisual content at home [4, 5]. For this assessment, the four most popular streaming media devices (SMD), excluding SmartTVs, in North America were selected, namely: Roku, Apple TV, Amazon Fire TV, and Chromecast/Google TV [6].

With a focus on understanding how UltraTV and these four SMDs cater to the concept of content integration while addressing the three main content watching activities, discovering, viewing and following content, five topics were selected for a hands-on analysis:

  • The general UI concept for content browse and discovery on the platform;

  • The extent of content integration, either built directly into the platform or by using an app-based approach;

  • The support for transversal search with multiple content sources;

  • The way the content viewing is handled, either through deep linking to the assets or by reproducing content in an integrated player at the platform level;

  • The approach for multiple user-profiles and how profile switching is handled at the platform level and inside the apps.

The first three topics of analysis are associated with the “discovering content” activity. In the initial topic, the main UI model for browsing content inside the platform will be evaluated, since browsing is usually the way most users select something to watch. Next, in the case that content is not surfaced in the main UI because the SMD is app-based and thus mostly siloed, the aim will be to understand how content from different sources could be presented to the user. That is, identify if the platform has some kind of aggregator app [2], for the user to browse different catalogs in an integrated way. Lastly, and being “search” the main alternative to “browse”, understand where the platform does support a search functionality and whether it can simultaneously search for content in multiple catalogs.

On the “content viewing” activity the main focus will be to understand if there is an integrated player at the platform level, at the aggregator app level or if the watching always happened inside separated players on the 3rd party apps. In the latter case, the analysis will also address the ability of the platform to deep link directly to content.

For the “following content” activity the aim was to spot the SMDs approach to such concepts as “following a series”, “continue watching”, “watch lists”, “favorite lists” or recommendations, usually perceived as the personalization features of a platform. Most importantly, the analysis will address how those functionalities are integrated into the platform and if there is support for distinct user-profiles. Additionally, the multiple profiles feature will be analyzed to access if it extends from the platform level and reaches the apps that also have this functionality. That is, evaluate if when the profile is changed at the platform level this alteration extends to the integrated 3rd party apps or sources like YouTube and Netflix.

This paper starts by presenting, in Sect. 2, how the UltraTV project addressed each of these topics, to better contextualize what was the idea behind each feature. It then proceeds to present the evaluation of each SMD platform and a comparison table in Sect. 3. The paper finish with a brief conclusion on the way the current trend aligned with the approach proposed on the UltraTV project or, on the contrary if the evolution, shaped by the correspondent business models, followed a completely different path.

2 Content Integration on the UltraTV Project

The UltraTV project fully embraced content unification. With that in mind, it was designed from the beginning to fusion linear and on-demand content, from traditional TV and internet-based OTT providers, but also user-generated content (UGC) from platforms like YouTube and social network videos from Facebook Video. In that way, its approach to browse and discovery puts all content at the same level, in a UI that positions a personalized linear TV selection front-center, surrounded to the right with two personalized playlists, one for on-demand traditional TV content and other for internet content sources. The UI allowed also a more traditional organization style, by having broadcast content categorized by genre in the columns to the left of these central personalized lists. OTT content is found in the columns on the right side arranged by providerFootnote 1 (see Fig. 1).

Fig. 1.
figure 1

UltraTV main approach to browse and discovery

Full size image

When the UltraTV project was conceived, it anticipated an app-based approach as the way to integrate additional features into the system. The idea was to make it simpler to integrate non-content related capabilities, like e-health and e-learning features. However, since the inception, the vision of adding more content or content sources to the platform was to be a fully integrated one, that would not need the addition of specialized apps.

However, to be able to incorporate, in a meaningful way, distinct content sources, the project needed to ingest and process all their metadata: titles, synopses, directors, actors, airing dates, seasons, episodes and so on. The project had access to the basic linear and on-demand content metadata provided by the MVPD. Yet, this metadata is traditionally segregated between what is usually called EPG listings for linear, and VOD or catalog metadata for on-demand content. These two data sources were reconciled by linking the linear programs with their catch-up counterparts and normal VOD assets. Netflix catalog was integrated in a similar way, creating a unified database. Understandably, this approach was not followed in the integration of YouTube and Facebook assets, given the sheer dimension of such catalogs. For these two services the approach was to integrate only a small subset of their top categories, that is, the categories that are usually highlighted in a non-personalized way on the main UI of these UGC platforms.

This unified metadata database was used to implement the cross-origin content columns on UltraTV browse UI. However, the multi-source search functionality, besides using the database, also performed online video searches in YouTube and Facebook Video and merged the different results on-the-fly, to allow the presentation of a unified search experience.

UltraTV content consumption was designed around the concept of an integrated player experience, and in that way detached itself from the usual app-based platforms, that make the user bounce from app to app in order to be able to watch different content. In the project context, this approach was successful except for playing Netflix assets. In that case, a deep linking solution was employed. Having almost all content reproduced directly inside the UltraTV main interface created an immersive experience, that blended the different content sources in the same UI. The integrated player nature of UltraTV also allowed the system to present cross-origin recommendations directly in the player UI, and to collect analytics on the consumption of 3rd party OTT content.

UltraTV approach to multi-profiles was quite advanced for the time, having concepts like group profiles to better address the communal aspect of watching TV on a big screen. An innovative feature was the notion of account linking, where it allowed a user to connect his UltraTV profile to their Netflix, YouTube and Facebook accounts. That way when a user changes his/her profile, the system automatically refreshes itself and fetches personalized content, not only from the UltraTV recommender but also from the different OTT and UGC providers [7].

3 Assessment of Four Popular SMDs in North America

3.1 Roku

The Roku platform started as a project to bring Netflix to the big screen (at a time when it was just available on web browsers) and evolved to represent the most popular SMD platform in North America [6], with its multiple models of inexpensive STBs (Set-top Boxes), streaming sticks and software for connected TVs. In Roku, the general UI concept is one of simplicity, with the main menu on the left and a grid of icons to the right of it, see Fig. 2a). Since the beginning Roku tried to make things simpler for the big screen TV user, that is why their apps are called “Channels”, to keep an easy mental link to a traditional TV set. Roku partners and producers create these “Channels” and publish them through a store, called “Channel Store”. However, since these “Channels” behave just like apps in other platforms, in the sense that they carry not only distinct content, but also have dissimilar UIs and UXs, the whole Roku platform can be considered app-based.

Even with an app-based approach, Roku still has some content aggregation functionalities. One of these is the “Roku Channel”, an app created by Roku itself, that aggregates content curated and licensed by Roku from different partners, both linear and on-demand, see Fig. 2b). This makes Roku not only a platform owner but also an OTT operator itself. For instance, the “Roku Channel” also exists as an app for Android, iOS and even Amazon Fire TV. While inside this app, the content is reproduced in an integrated player and there is no deep linking, even for content from partners that also have standalone apps in the platform, like HBO or Showtime. In that respect, the “Roku Channel” is not really aggregating content surfaced by the different sources that are available in the Roku platform. In reality, it is just like any other content app and has a business model based on advertising for the content that it makes available for free. However, the “Roku Channel” also allows the user to subscribe inside de app to additional premium content, from providers like Showtime, AMC or Starz, creating an additional revenue stream to Roku. This addition of premium subscription content, curated and billed by Roku on the “Roku Channel” to the free ad-supported content, makes this platform looks more and more like an OTT MVPD [2], Fig. 2c). The major streaming platform absent from such integration is Netflix.

Fig. 2.
figure 2

Roku UI

Full size image

One of the more unifying features on the Roku platform is, however, its cross-app search. Roku works with the app creators, so they provide Roku with metadata feeds for the content available inside their respective apps, Fig. 2d). If the same content is present in more than one source, the search interface presents distinct alternatives for the user to choose. However, the playing itself, when initiated thru the search function, is then handled in each of the separated apps via deep linking and not played in an immersive way like in the “Roku Channel”.

An innovative personalization feature leveraged by these metadata feeds is a “follow” functionality integrated on the platform. For instance, when using the transversal search, a user can follow a movie, a series or even a person, and be notified on the “My Feed” menu item about changes on these topics, like new episodes available or price changes.

What Roku lacks is any feature related to profiles. A Roku account is needed for the initial setup, but changing the account requires a factory reset of the device. The apps that allow multiple profiles still support them, but on an app-by-app basis, like Netflix or YouTube, not as a platform profile. Also, there are limited personalized content recommendations on the platform, instead promoted and highlighted content seems to be mostly curated.

3.2 Apple TV

Apple TV evolved from being a “hobby” for Apple [8], to be the 2nd most popular SMD platform in North America [6]. With the launch of the 4th generation in 2015, this small STB got a UI redesign paired with a brand new remote, that still today is not consensual. The main STB UI has a big area on top, called “top shelf”, providing a way for the apps that have a privileged spot on the first row to promote their contents or functionalities. Below there are rows of app icons, or app groups, much like in an iPhone or iPad. Besides the “top shelf” that itself is app-driven, there is no content surfaced on the main Apple TV interface, so this platform is clearly app-based, see Fig. 3a).

Content aggregation really happens inside what Apple calls the “TV” app. On this app, which mimics the main Apple TV UI, the user can browse for content provided by Apple itself, from its “Apple TV+” OTT platform or made available by partner apps or other content sources, see Fig. 3b). This content populates the main “TV” app UI and special thematic sections with areas dedicated, for example, to sports (including live broadcasts) or kids.

Similar to Roku, Apple gets metadata about the content provided by other apps thru a server-to-server feed integration [9]. However, instead of using this metadata feed only for a cross-app search, Apple uses it to make the “TV” app a showcase of the content the user can access on the platform, behaving in this respect as a content discovery application [2]. Most of this content, when activated is then reproduced inside the respective apps via deep linking directly to the content. A nice feature, resulting from the server-to-server feed integration is that the user does not need to have all the apps installed for the content to be showcased and highlighted on the “TV” app. However, when a user tries to play content that does not have the correspondent app installed, he/she is redirected to the App Store for the download and installation.

Fig. 3.
figure 3

Apple TV UI

Full size image

But, more than that, Apple establishes partnerships with 3rd party content producers and OTT platforms to extend this discovery behavior into a full fledge content aggregation platform. Inside the “TV” app, in a feature called “Channels”, there are subscriptions to additional content, which the viewer can use to expand their “TV” app catalog, see Fig. 3c). This makes Apple and the “TV” app also some sort of new OTT MVPD, not much different than a traditional cable operator, and similar to the “Roku Channel” [10]. But, besides the business aspect, “… unlike discovery apps, aggregator apps enable users to watch content from other providers within the interface of their app” [2: 175]. This is a crucial improvement from the UX perspective and implies a big departure from the app hopping approach typical of this platform. However, as in Roku case, Netflix is notoriously absent from that deep integration.

Apple TV has two separated search functions, at the platform level, there is a “Search” app that, as the name suggests, does a cross-app search, but is not restricted to audiovisual content, searching also in iTunes catalog for music albums and songs, or in the App Store for apps and games. But inside the “TV” app there is also a search function, that in this case is focused on searching for movies, shows, cast and crew from multiple content sources, Fig. 3d).

In terms of personalization functions, with all the information that Apple collects from the content consumed inside the “TV” app, and from the deep linking partners that share back to Apple information about assets watched inside their apps, the “TV” app can present a fully integrated “Continue Watch” list.

Recent versions of the Apple TV platform do have the notion of multiple user-profiles, however, the current materialization is mostly limited to the Apple ecosystem. For instance, when the user changes profiles, the set of installed apps changes automatically on the main UI. Similarly, the signed-in user on Apple services (iTunes, Apple TV+, etc.) also changes. Inside the “TV” app, the “Continue Watch” rail and the curated sections also reflect the usage made by the different profiles. However, there is no linking between the Apple TV profile and the profiles inside other apps, e.g., changing from profile A to profile B on the Apple TV UI will have no impact on the selected profile or account on the Disney+ app or the YouTube app.

3.3 Amazon Fire TV

Amazon Fire TV is a diversified line of SMDs made by Amazon. They are little powerful devices a bit more expensive than Roku. The general UI concept for the main screen in Fire TV is split into 3 areas, see Fig. 4a). It has a top main menu, followed by a big area that when selected, automatically starts playing a trailer, a promotional video, or a static banner. Below that promotional area, there are rows of items, that have different functions depending on the usage given to the SMD.

Fig. 4.
figure 4

Amazon Fire TV UI

Full size image

Since the underlying software of Fire TV is based on Android, its main approach is app-based. The system gets a lot of Amazon apps pre-installed, side by side to the usual Netflix and YouTube apps. Throughout the platform, it is easy to see that Amazon content and products have a very prominent position, and this is leveraged, mostly by pushing Amazon Prime Video content up and center. Nevertheless, the rows of tiles in the main screen and in the different tabs from the top menu surface a mix of promoted content and apps from Amazon and its partners. This means that even with an app-based platform there is a lot of content scattered all over the UI, implying that there is a strong connection between Fire TV and content discovery and consumption.

But Amazon approach to content discovery goes beyond the main Fire TV UI, it actually has two additional aggregation apps. Similar to the “Roku Channel”, where the user can watch free ad-based content, Amazon provides an IMDb TV app, see Fig. 4b) [11], and like with the “Roku Channel” this app is also available outside the Fire TV ecosystem. Content available in the IMDb TV app is reproduced in an integrated player, transversal to the platform, and movies and series not watched to the end are surfaced to a special “Recent” rail, on the main Fire TV UI, that works as a cross-app “Continue Watch” functionality. The same happens with Prime Video content watched on the Prime Video app. However, this does not extend to other apps like Netflix or YouTube, in that case just the app icon appears on the “Recent” rail, not the content itself. The same applies to the actual content playing, where apps outside of the Amazon ecosystem reproduce content thru deep linking and not in the same integrated player that the apps from Amazon use.

The major app Amazon has on the platform is the Prime Video app. It is similar to the “TV” app in the Apple TV, working also as an aggregator app, providing under a subscription not only content produced and licensed by Amazon as an OTT platform but also allowing users to top up this subscription with additional partner contents, from producers like HBO, Showtime, etc. Just like Apple, Amazon calls these additional content sources “Channels”, see Fig. 4c), and this also transforms Amazon, much like Roku and Apple, into a kind of OTT MVPD. Again, like in the other two cases, Netflix is absent from this bundling.

Amazon Fire TV, like the other SMDs, uses metadata feeds to surface content from the different apps and partners. That metadata is used by the main UI but also on the search function, see Fig. 4d). Since Amazon has many partners, and is business-driven, when content is selected either from the search function or from the multiple discovery options a wide range of watching options are presented to the user, either to subscribe, rent or buy. Again, similar to Roku and Apple, these metadata feeds allow Fire TV to index movies and series from apps not yet installed and to promote their installation on-the-fly.

Where the search function stands out is on the relations established between the different content. Using very detailed data from IMDb, the system cross-references actors and directors to create links between the different movies and series. This deep metadata integration extends to the player itself, in a functionality called X-Ray. In some content this allows the user to get information about the actors and characters in a scene, along with other snippets of information like the music track used.

On the personalization features, the main functionality is the, already alluded, cross-app “Recent” rail on the main interface Home screen. The IMDb app has a “Continue Watching” feature and the Prime Video app a similar one called “Watch next”, and both share some of the items. However, since the platform has no notion of multiple profiles and even the Prime Video app only got user-profiles recently [12], there is no account linking, with each application managing its own profiles.

3.4 Chromecast and Google TV

Google has the 4th most popular SMD platform in North America [6]. However, this could be a little misleading because most of these devices are Chromecast devices, little dongles, sold directly by Google, and allowing only content reproduction, without a discovery interface and no remote control, making comparisons difficult. Content browse and discovery must be conducted on a PC or mobile device and then sent to be played on the Chromecast device (by “casting” it). Parallelly, Google also develops an Android-based TV software platform, aptly named Android TV, that is used by some hardware partners to create SMDs (NVidia Shield or Xiaomi Mi Box) and SmartTVs (Sony or Hisense). This hardware and software blend creates a full-featured platform, mostly app-based, that is purchasable directly by end-users but also used by IPTV operators as a basis for their offerings.

Recently, Google launched a new product called Chromecast with Google TV [13], which is really Android TV, with a new interface, running also in a small dongle, but this time with a remote control. This is Google attempt to compete in the same market as the other three platforms with a similar offering, and with a strong emphasis on content.

The main UI concept on Google TV is similar to Amazon Fire TV, with a top main menu, a big slideshow area for promotional content and rows of items, see Fig. 5a).

Fig. 5.
figure 5

Google TV UI

Full size image

In this case, Google TV has a first row with personalized recommendations, followed by a row of user apps and then several rows of content arranged by categories or themes making content a first-class citizen, relegating apps to a secondary position. This makes Google TV UI the main way to discover content, and Google promotes it as a hub for accessing all the different streaming platforms in one place. The search function is also well developed and, similarly to the other SMDs, enables cross-app search and can propose alternative ways to purchase or rent the content when available in more than one partner. A feature not easily seen in other SMDs is the ability to search for concepts like “drama movies” or “comedy shows”. Being the owner of YouTube means that, most of the times, when a search does not get to a single piece of content, either because it did not find anything on the catalog, or inversely it finds many items, there is always a set of YouTube items provided alongside the main ones, see Fig. 5d).

The playing of content is almost all done via deep linking, even when the content belongs to the Google ecosystem, like YouTube. A little bit strange at first is that in this platform Google enables some premium content, like Hollywood movies, to be played legally for free, with ads, in YouTube, see Fig. 5b).

Google also proposes personalization as a main feature, by having an automatically generated “For you” tab in the default landing page, that focus the catalog on content for which the user does have installed apps. Additionally, in the “Library” tab there is a “Watchlist” function. The user can populate this list with items he finds browsing the unified catalog. The platform also supports a cross-app “Continue watching” list, with partner apps reporting back to Google the consumptions done inside their apps, see Fig. 5c). There is no profile support yet on Google TV [14]. The personalization only applies to a single master account and there is no linking between this account and the profiles used inside the YouTube or Netflix apps.

3.5 Comparison Table

To help with the comparison between the 5 evaluated platforms and the 5 selected research topics, Table 1 summarizes their different approaches.

Table 1. Comparison between UltraTV and the four SMDs
Full size table

In this systematization it can be seen that the platforms have a strong alignment in the way that they address content unification, being UltraTV the most contrasting one. The concept of user-profiles is the one where there is the main variation between all the platforms and where the biggest future improvements are expected [12, 14].

4 Conclusion

Content unification and aggregation features are gaining strong traction on the four evaluated platforms. For instance, transversal search functions are currently universally available using metadata feeds from different partners. Still, they lack a deeper integration with non-curated content like YouTube (except for Google TV) or with other social networks, that UltraTV promoted.

These four platforms are mostly app-based, and the additional integration and unification are being promoted thru the creation of special apps like “Roku Channel”, Prime Video app and Apple “TV” app. With this strategy, the big players are making those apps silos of deeper integration and personalization, more than endorsing these features to the platform as a whole. For instance, these apps promote an integrated play experience where the platform itself mostly push for deep linking solutions.

On additional personalization features, Fire TV has a “Recent” rail which presents previously watched content and apps. Apple “TV” app goes even further by presenting a transversal “Continue Watching” feature with data shared backed by multiple partners. Google TV, besides also having a “Continue Watching” feature, has a personalized “For You” area that brings recommendations and personalization to the main UI.

Nonetheless, UltraTV although conceived more than three years ago, still had a set of personalization features not broadly available in the market. One of the most notorious was the profile linking, which allowed the platform to automatically expose on the main UI the profiles built into the different partner apps. This made personalization not only a feature of the different apps but also a functionality of the platform itself and made profile changing much easier for the user. Yet, taking into account the progress being made in the contemporary SMDs this will be a deeply welcomed function to foster the TV set as the primary device to watch audiovisual content at home.

An additional research topic, already raised in [2], is that these four SMDs and their respective support platforms are positioning themselves to play the role of MVPDs of the OTT world. They are stockpiling content rights from multiple sources while commissioning exclusive content. They are also working as the bundling partner for additional premium content providers like Showtime and Starz. With that, they are not working just as paywalls, they are creating an immersive UX around a large enough set of content. This could in the future keep the user always inside their walled garden experience, with all its potential harms and benefits.