More Thoughts on Metadata and the Second Screen

I helped to host a webinar on metadata today that was produced by MESA and Rovi, and a huge segment of the questions asked were about second screen and social TV.

What kind of metadata is needed to drive a good UX (User eXperience)?  What kind of metadata is required to support better advertising or commerce?  Who provides this kind of metadata?  What are the examples where this is done well in the marketplace today?  How much metadata should the content creator or app developer try to capture?

Let's back up a bit and "normalize" all of our experience in this space.
I like to think of metadata in the video space falling into 3 categories: technical (frame rate, frame size, commercial break, bit rate, etc), descriptive (summary, actors, director, reviews) and contextual (what objects are in the scene, what is happening at that point).  Clearly, the first is a requirement to deliver a quality video service no matter the channel of delivery, the second is critical to search and recommendation, and the third is critical for any higher value experiences (better commerce, contextual advertising, discovery of new content for a consumer).

One the first (technical metadata), there are a ton of providers out there working on this problem, and for most consumers and business partners out there, this problem is largely solved.  This allows you to bring up the right quality choices when watching NetFlix, Hulu or Vudu and ensures you are matching the right codecs with the right devies (and DRM).

The second (descriptive metadata) is technically solved, but is mostly a user interface and scale problem these days.  In other words, the ability to put together a summary of a film or TV show combined with the actors/cast, a price point, and an availability date by country is a pretty standard capability for nearly all video delivery solutions.  Doing that in many countries across many different business models, devices, and millions if not billions of consumers is the scale problem (or cost problem).  Doing it in a manner that allows the consumer to easily find (search) what he/she is looking for (EPG or classic search box) is mostly a UI/UX problem--meaning we know how to deliver the result, but often deliver it in such a confusing manner that the consumer doesn't use the function (or it takes 20 minutes to do so).

So let's focus on the hardest, but most valuable set: contextual metadata.

This is exactly what is required for an enriching content experience on the second screen, a better way to engage audiences in advertising, a more effective way to engage consumers in commerce, and, I believe, the best way to power true Discovery use cases for consumers (vs. search or simple recommendation).

What kind of metadata is needed to drive a good UX (User eXperience)?  What kind of metadata is required to support better advertising or commerce?  Capturing the information that describes the scene is critical to drive these experiences.  A simple example:  you are watching "Risky Business" with Tom Cruise.  He puts on his new wayfarer sunglasses in the final scene and gives his final line of the movie.  With a synchronized experience, we can: a) provide a factoid describing how that scene created a new pop trend for that style of sungalsses in 1986, b) show a Ray-Ban brand advertisement on your tablet, inviting you to click thru to see their new line of reto wayfarer glasses, c) drop you into a store front to buy those exact sunglasses or similar ones, d) show thumbnails of other scenes where Tom Cruise does his signature sunglass scene (TopGun, Mission Impossible, etc), allowing you to put those movies in your queue, rent them for later, etc.

Who provides this kind of metadata?  This space is relatively new.  Digitalsmiths is probably the most comprehensive in terms of breadth of titles and algorithms (different data captured) applied.  RCDb has been doing this for a more limited set of titles and is moving into the metadata syndication space (helping content owners themselves get this rich metadata to third party app developers).  Rovi, Gracenote, TMS, FYI and Redbee are all developing their capability or licensing it from others (see the second screen ecosystem discussion in this blog).

What are the examples where this is done well in the marketplace today?  A few good apps to checkout are Fanhattan and BuddyTV (for deep, rich metadata) and TVplus (for synchronized metadata experiences).

How much metadata should the content creator or app developer try to capture?  This is unfortunately a function of cost.  Since you don't know the use case ahead of time (commerce, advertising, better UX, etc), you need to capture as much as you can upfront and normalize it for the array of use cases you foresee in the near term, and then leverage third party services like those mentioned above to supplement or re-tag in an automated fashion in the future.

This is probably the most important element to getting a cost efficient and scalable UX in place for second screen and social TV yet the least understood in our market place.

Good luck.

www.chuckparker.tv