Menu
Feature

Dynamically Scheduling TV in Households

Written by

We recently submitted a white paper to CHI 2013, and to our delight we have been accepted for the conference this year.

In our design projects, we are conscious that TV scheduling is at risk of being compromised in the move to On Demand (OD) TV and automated recommendations. Services such as Netflix have been defined by their breadth of offering, and disrupt traditional scheduling models [1]. In this landscape, live broadcasters are in a unique position to leverage their experience of providing "quality over quantity".

Research suggests [2] that TV watching patterns are still in parity with traditional scheduling. Users value conversations created by small, curated selections of content [3]. Recommendations systems are at risk of being skewed and abused by household members sharing an account.

We propose using recognition technology to observe and augment individual behaviours within family units in order to prevent this.

Problem

TV is a social experience [4]. Social networks are used by TV services [5] to tailor recommendations. Twitter’s research [6] illustrates the uncanny parity between social conversation and TV schedules, demonstrating how important this direction will be in the future. In the scramble for social network analysis, the oldest social network of them all, the household, could be ignored. Research and products such as Microsoft Kinect [7] and Samsung Smart TVs [8] demonstrate facial and skeletal recognition in use. Using such technology, we can obtain live data on close social networks by recognising who is in the room.

CHI: Audiences

The role TV scheduling plays in defining a live broadcaster such as ITV in the UK.

Scheduling acumen in the TV Guide and Electronic Programme Guide (EPG) [9] has been effective in forming user’s understanding of channel brands by visualising the broadcaster’s priorities and interests.

Traditional live broadcasters recommend content via curated TV schedules to the masses. E-commerce sites such as Amazon use recommendations based on behaviours and browsing patterns per user using a rich layer sales data to augment this. Social networks such as Facebook and Twitter have pioneered the concept of mutual interests within social networks.

CHI: Cable Cutters

The migration of some users from traditional cable and terrestrial to over-the-top services.

Our position is that the advent of OD should not be a watermark where traditional scheduling is moved aside for automated recommendations. Live broadcasters should bring their knowledge of scheduling and combine them with technology to create the next generation of the household TV experience.

Design Solution

Live broadcasters should focus on playback of their content on as many devices as possible, even if it requires a simplification of the core feature set of the product. This ubiquity can then be harnessed to more accurately analyse viewing patterns across devices.

Cameras within devices such as Xbox Kinect allow applications to detect [10] who is watching what content [11], combined with when and where it was watched. Devices such as mobile phones can gather personal viewing data and feed it back into the family account. Using this information, live broadcasters can learn about a household’s viewing habits on a person- by-person basis, using strong and weak signals.

We propose combining approaches of data mining and TV scheduling in order to meet traditional Key Performance Indicators (KPIs) at the same time as creating products that visibly respond to emergent behaviours. We wish to investigate how social networks within the home can be exploited to provide users with recommendations based on their family’s activity. We predict that by observing individual's habits, we will be able to ascertain in what context they enjoy or avoid shows. This will give broadcasters a unique opportunity to present alternative popular content recommendations.

Finding commonality in a student household

Fariq, Alex and Sarah are students living in a shared house. They share a TV in the living room, but also watch TV individually on their own devices.

  • Fariq regularly watches 10 O’Clock Live in his room, a daily topical comedy show featuring David Mitchell, a British comedian and actor.
  • Alex is a big Peep Show fan, a sitcom featuring David Mitchell. He watches it on his smartphone regularly when he’s studying.
  • Sarah loves The Big Fat Quiz of the Year and Alternative Election Night, both topical shows often featuring David Mitchell.

The household decide to sit down on a Friday evening and watch TV together. T the product detects that all three students are in the room. Wife Swap, a reality show, is currently showing live.

CHI: Concept

The product detects who is in a room and recommends content based on an intersection between all users’ personal viewing habits.

The product knows that none of the three users watch reality shows, and looks for an alternative to Wife Swap. The TV finds a intersection between their tastes in David Mitchell. An episode of Alan Carr: Chatty Man, a weekly talk show, where David Mitchell featured is found in the archive.

A new series of Alan Carr: Chatty Man and the broadcaster have indicated to the product that they are interested in gathering new viewers for the show. This mimics a human TV scheduler repositioning content in a different timeslot to obtain a new audience.

Sarah comments that she recognises David Mitchell from previous shows and enjoys the show, and Fariq and Alex are pleased to watch something with David Mitchell.

When households have differing interests

Kim is a 16 year old who lives with her parents and two younger sisters. Every Saturday, the family sits down to watch TV for the evening.

When Kim wants to watch TV alone, she chooses The Only Way Is Essex, a scripted reality show, on her laptop as her parents don’t let her younger siblings watch it. The broadcaster’s recommendation engine has learned that she watches gossip-based shows such as The Only Way Is Essex and Take Me Out when alone. It has never observed the family watching either of these shows on the TV. The product infers that the family do not enjoy watching gossipy shows, but that Kim is a fan.

It is Saturday evening and the family are watching Dancing on Ice, a talent show. Kim goes to her bedroom, turns on her tablet and looks for something to watch by herself. The product knows that it is Saturday evening and that the broadcaster needs to push primetime shows with minimum viewership guarantees. It also knows that her family are watching Dancing On Ice on the TV. Take Me Out is an important primetime show for Saturday viewing figures and is tagged as ‘gossip’. The product recommends Take Me Out to Kim instead of Dancing On Ice.

Next Saturday, the family are out for the evening and Kim is home alone. She turns on the TV, and the TV detects that she is alone. Dancing On Ice is about to start, but the product recommends that she catches up on Take Me Out instead.

Consideration

There are privacy and security implications to such technology. Geerts, David, and Dirk De Grooff [12] propose giving users controls over actions and system settings and privacy guarantees. The increasing adoption and comfort with technology such as Xbox Kinect, plus the benefits it brings, could bring this approach into the mainstream mindset.

Conclusion

Accounting for social viewing behaviour within a household as part of a recommendations engine could create a more powerful and trustworthy system.

We propose that leveraging this information against time-based scheduling could be valuable to broadcasters. This will allow them to evolve their scheduling model to compete with over-the-top providers and content aggregators.

The importance of ‘household’ to broadcasters is an area that will need ingenuity and technology to retain knowledge of as we move into a world of connected media devices.

References

  1. The mistake Netflix is making now
  2. Telescope: A look at the nation’s changing viewing habits from TV licensing
  3. Schwartz, Barry. The paradox of choice: Why more is less. Harper Perennial, 2005.
  4. Wang, Qi, et al. "Digitality and materiality of new media: online TV watching in china." Proceedings of the 2012 ACM annual conference on Human Factors in Computing Systems. ACM, 2012.
  5. Using Facebook Connect on Netflix
  6. Tune In With Twitter
  7. Leyvand, Tommer, et al. "Kinect identity: Technology and experience." Computer 44.4 (2011): 94-96.
  8. Experience the Wonder of Samsung's New Smart TVs
  9. Vinayagamoorthy, Vinoba, et al. "Researching the user experience for connected tv: a case study." Proceedings of the 2012 ACM annual conference extended abstracts on Human Factors in Computing Systems Extended Abstracts. ACM, 2012.
  10. Yin, Qi, Xiaoou Tang, and Jian Sun. "An associate- predict model for face recognition." Computer Vision and Pattern Recognition (CVPR), 2011 IEEE Conference on. IEEE, 2011.
  11. Perez, Kathryn Stone, Alex Aben-athar Kipman, and Andrew John Fuller. "CONTENT DISTRIBUTION REGULATION BY VIEWING USER." U.S. Patent No. 20,120,278,904. 1 Nov. 2012.
  12. Geerts, David, and Dirk De Grooff. "Supporting the social uses of television: sociability heuristics for social TV." Proceedings of the 27th international conference on Human factors in computing systems. ACM, 2009.

Ready to talk to us? Interested in visiting?
Contact hello@ostmodern.co.uk or
call us on +44 (0)203 215 0041