TV as a medium is undergoing two notable trends - dispersion and atomisation.
Dispersion is the evolution of TV from single screen at-a- time viewing into a rich, distributed viewing experience across multiple screens.
Atomization is the transition from linear to non-linear storytelling with substantial user control of the media consumption process.
Without a technology solution, these two trends have the potential to adversely affect the economics of TV in increased content costs, increased user experience (UX) complexity, and therefore decreased user participation.
This paper proposes a media services architecture that enables the delivery of richer multi-screen media experiences, while still maintaining the coherence of the experience and the cost of service delivery.
We share a practical experience in running such a system over a large, globally diverse media corpus.
The working system supports a Social Electronic Program Guide (EPG) user experience leveraging both current and emerging (Wearables and IoT) device platforms.
The ‘Viewer’s Choice’ Challenge
TV as a medium is undergoing two notable changes - dispersion and atomization.
Dispersion is the evolution of TV into a distributed experience across multiple screens, and atomization the componentization of parts of a TV experience into modular media segments that can be reassembled into derivative media.
These trends change the nature of media viewing and impact what it means to personalize the viewing experience.
As the Google multi-screen study points out , dispersion is more than a simplistic use of different screens for one-at-a-time video streaming.
The sophisticated use of multiple screens sequentially and simultaneously introduces two new UX elements – divided attention and new lean-in levels.
With multiple screens involved, there is both a movement of user attention across screens, as well as the potential for a user to be simultaneously paying partial attention to multiple screens.
In addition to divided attention across screens, the user experience on any single screen has become more intricate as TV UX platforms include support for tiling, transparent overlays and other advanced user interfaces.
For instance, the humble EPG has gone from a rectangular grid on TV, to a tablet interface with multiple levels of click-thru screens, each with a rich, interactive interface.
This means that the minimalistic two-level attention model of lean-back vs. lean-forward needs to incorporate multiple lean-in levels that lie in between those extremes.
As an example of lean-in models, an EPG user could be in navigating to a known program, open to passive discovery, binge recording using the EPG as a tool, exploring the Twitter conversation for a single show, or looking for applications (apps) associated with the show as an overlay experience to TV watching.
Atomization is the transition from linear to non-linear storytelling (and non-linear viewing), where the viewer customizes the what (entry point, content granularity and perspective), in addition to the how (devices, duration, interactivity) of the viewing experience.
As an example of entry points, the choice of user experience could be a highlight reel based on a particular player or character as a seed.
Alternatively, the user could pivot from a car chase scene to all movies with scenes of similar mood, movement and genre.
With an increasing number of content publishers also publishing companion apps [2,3], atomization also means the viewer mixes modalities – interacting with flexible mixes of video (e.g. an NFL game) and associated apps (e.g. a Fantasy App).
Personalization in a Viewer’s Choice World
A user experience with this many choices can be overwhelming without personalization.
That said, traditional personalization (e.g. collaborative filtering) works for linear viewing but is too heavy-handed for a complex, layered viewing model such as the above.
In a viewer’s choice world, personalization needs to transition from prescriptive recommendations of entire media items to lightweight conversations and exploratory hints that the user chooses to opt-in to (or not).
The conversation itself may pivot from content titles, to actors, moods, modalities (raw content vs. apps), pricing (free vs. paid), timing (immediate view vs. cached for future viewing), and a growing number of new dimensions and entry points.
To support frequent and lightweight forms of discovery, two services used as building blocks for personalization are proposed.
The first uses media analytics to enable content providers to create dynamic metadata driven experiences.
This enables new assemblies of derivative content to be presented to end users as substantially new experiences, without substantially higher editorial cost.
The second is a social fusion service that manages multi-device content delivery and ‘artful user interruption’.
This service maps the available user attention to content delivery on screens with varying capabilities and affordances.