I have a confession.

My name is Tom. I’m thirty something and I’m an addict.

Every Saturday morning, I sit down and watch an episode or two of South Korean game-theory, celebrity-reality show The Genius. It’s an uncontrollably enjoyable programme of trust, betrayal, whiplash editing, sudden cliffhanger moments and brilliant reaction shots, all set to a collage of pop music and on-screen cartoon indicators of extreme emotion. The participants are charming, highly intelligent, amusing and charismatic.

The Genius felt like something from the future when I first saw it. There are an awful lot of ‘did you see the bit where…’ moments, and Shazam is never far away from my phone thumb. There’s a minor cult of viewers across the internet who like to share their favourite moments, and it’s good fun to re-watch with likeminded friends on a mobile or tablet when we meet up.

I’m sure it’s some people’s idea of hell, but I love it. I’m afraid to say I watch this programme on the internet, in my dressing gown and pyjamas, while eating cereal and lying on a sofa. It’s not an edifying picture but it does kickstart my weekend.

My point is we all know viewing habits are changing; and content discovery is having to keep up too.

With linear viewing, time and channel were the two critical pieces of metadata by which viewers would find content they wanted to see. Other ways of viewing, especially on demand services, have meant that genre, viewing history, and the viewing habits of people like you have become increasingly important ways in which we discover and access our media content.

This data was typically available from scheduling and programme information available ahead of broadcast time. Now the schedule and programme information is only the start of the process.

One way we can enrich the data is with captions. As captions are effectively a timed-text transcript of the spoken content of a programme, adding other ‘events’ into a programme file is an easy task for the person creating those captions.

But what do we mean by ‘events’?

At IBC we’ll be showcasing how the addition of location, product, scene, mood and character information at the captioning stage can significantly transform the viewing experience.

Got an hour or so and want to relive all of Tyrion’s moments from Game of Thrones? Having the character information available for each scene means that the viewer can easily access them by just clicking on his character profile.

Only got 30 minutes and want to narrow it down? Add a ‘confrontational’ mood indicator to give you just clips where he’s at his cutting best.

Got a particular favourite clip from your personalised reel? Share it and comment on it via social media.

Love that track? Including music information means you can quickly and easily see what a song is, as it’s playing, and save that information for later in your personal profile, or buy it right there and then.

Product and location data is excellent for shows like The Apprentice. You can tag restaurants, shops, health spas – whatever is relevant – and allow the viewer to see where the locations are when they appear, and even book a ticket, table or seat, while still watching the programme.

And it’s ideal for my favourite programme, The Genius, where I can get a line-up of all Sunggyu’s winning moments, or Sangmin’s meltdowns for each series.

This process doesn’t stop when the programme airs. We’re exploring how we can enrich the live captioning process by logging key event metadata for sporting events, and ensuring every news story and video is easily retrievable by ensuring we keep its metadata.

After broadcast, we use Natural Language Processing to dive deeper into programme content, pulling out keyword summaries, and finding connections between what you’re watching, and other media and advertising content.

I said earlier that we know viewing habits are changing. But, there is still a lot of discussion about the detail. For those keen to focus on the cutting edge of mobile and tablet on demand viewing, it’s very easy to ignore the fact that the majority of people still sit down to watch television together. As the recent Ericsson ConsumerLab report showed, one major factor is social. We enjoy sharing our viewing experience with others. And in a fragmented on demand landscape, that social experience is harder to come by.

The penumbra of tweets and retweets, gifs, +1s, Facebook conversations and Tumblr shares that distinguishes the talked about from the not talked about is facilitated and encouraged by this simple process of putting in additional metadata at a comparatively inexpensive stage by people who are close to the media. It makes it easier to find the things you’ve seen that you’ve enjoyed most. And it encourages re-watching and social sharing.

Even if you are watching in your pyjamas.

Tom Wootton, Senior Product Manager, Access Services