we write about the things we build and the things we consume
vrm: social tv? forget the 2nd screen apps, it's all about generic data
In my last post on Viewer Relationship Management (VRM) I described some of the stronger opportunities for broadcasters to use data more effectively. Social data is high up this list.
all about the "2nd screen"?
I just checked on my Twitter and Facebook friends in Zeebox. 3% of the 772 people I follow on Twitter and 6% of my 170 Facebook friends are using the app. Of these, 100% are either working in media, or closely connected to the Zeebox team. More than 80% have previously, or are currently working on 2nd screen projects for a broadcaster. In short, this is still a very niche audience.
There’s plenty of breathless surveys about, proclaiming huge reach for 2nd screen activities well in excess of 50%. The Guardian recently published a somewhat breathless survey roundup.
2nd screen affects much, much less than 11% of tv viewing
The reality is that most of this involvement is just unconnected browsing and people communicating about their lives in general, which may or may not be connected to the flickering box in the corner. While surveys focus on 2nd screen reach, it’s clear that hours spent with second screen experiences are still tiny.
Earlier today @Macedoines pulled the killer stat:
.@adrideo According to
<a href="https://twitter.com/ofcom">ofcom</a> people spend 27 minutes per day per person accessing the Internet from home. So 11% at most(average).</p>— Nicholas Barr (Macedoines) October 30, 2012
It’s a great one to remember: even if we assume that ALL home internet is 2nd screen activity, 2nd screen is much, much less than 11% of TV time.
This is a killer blow for TV-specific 2nd screen apps, because social only really works at scale, when a large number of your friends get involved.
meanwhile, generic data
While TV-specific 2nd screen application use is tiny, there’s still plenty of activity around TV on the social networks. This is the same generic data you could collect about people’s discussion on news stories, or the weather, or consumer goods brands.
Twitter is sometimes perceived as a niche medium, but we’re now seeing increased uptake, especially among a younger audience. The realtime nature of Twitter is a major plus for broadcast, which is still a predominantly live experience.
On Twitter, major TV shows get hundreds of thousands of tweets or more. Because people are responding in realtime, a lot of analysis is possible by matching the tweets to the timeline of the show. We do a lot of that here, using our global index of video and audio data, Atlas.
Plenty of people are doing simple analysis of tweet rates and volumes, evaluating the parts of shows that resonate, and looking at individual topics and characters. We’ve been taking this a couple of steps further than just counting. Our focus is on:
- analysing meaning, picking out the popular phrases
- understanding what the audience for a show is also doing, beyond just watching this show
These techniques have huge potential to increase understanding of how audiences respond to shows, and more importantly to understand the audiences better.
Over on Facebook, the vast majority of activity is less realtime, but the platform reaches a much broader audience. "Friction free" logging of TV activity has not taken off as in the music industry, but Facebook contains a rich set of likes and posts around TV shows. We find that most users have between 200 and 2,000 TV likes in their social graph.
Facebook applications can immediately use this data to build very valuable services, and without waiting until a majority of a user’s friends join the application.
The important things on Facebook are:
- Crafting a compelling reason for users to add your application, and give permission for you to use this data. As always, there has to be a very strong value exchange in the eyes of the user.
- Retrieving and processing the mass of data fast enough to give users near-instant feedback. This is no easy technical challenge, and has tested many in our talented engineering team.
- Turning the relatively unstructured data from Facebook into meaningful information. Atlas has some strong Facebook services that help here.
- Making sense of the mass of information in the social graph. Not all friends are equal, and not all likes and comments indicate a positive view of a show. But with careful analysis and product design these issues can readily be overcome.
We’re currently helping several clients with these challenges, and building valuable services on top of our data analysis systems.
getting into the detail—care to join us?
We’ve been working in this area for more than five years, well before the company was conceived, and pushing the envelope of what’s possible much further. We would love to work with in finding novel ways to make the most of social data, as part of your wider VRM strategy. We’re ready to get into even more detail, and we serve the best cakes of Bloomsbury. Shall we talk?