Managing The Metadata: Q&A With Greg Varhely
Videos of the full interview can be viewed here.
Below is a short excerpt:
CW: You prepare content for multiplatforms. Let’s say that I am a programmer, and I give you my program to place on various media platforms. How do you start?
GV: Assuming it’s going everywhere, we have to create a content package. This package not only includes the video file, it also includes the metadata which has all the descriptions, all the schedule information, the billing information and background IDs. There is also the artwork that goes with it.
As these user interfaces become more sophisticated, you are seeing more and more artwork that helps entice a consumer to watch the content.
You have to build a complete package. The challenge you have is that any time a piece of that package varies for whatever reason, you have to create a new one. An example: If you have a 10-episode series, you are not just creating the 10-content packages that go out into the digital world, you are creating 10 for VOD, 10 for TV Everywhere, 10 for all of the co-electronic sell-throughs: Amazon, Netflix. If you are putting any of this content on mobile, you are creating 10 different packages just for that. And what happens then is that it starts to multiply.
With VOD, you might need 10 for SD and 10 for HD. You may have variances based on MVPDs. Most of them do in fact vary, particularly in the TV Everywhere world, where no one is following any set specifications. Even if the video files are the same, if your metadata is different, if content providers have different schedule patterns -- they give MVPD “A” a 30-day window and MVPD “B” a 60-day window -- you have to create different package versions.
In my last corporate position, a 10-episode series could end up being 200 or 300 different content packages. They all have to be built and distributed and tracked and archived separately.
We are at a place where one piece of content on average was creating five or six different versions. Some of the high-end programs have 20 different versions. It is a huge challenge, [which] is what has prevented a lot of this content from being distributed across all the possible platforms.
CW: When you talk about metadata in the packages, is there some piece of metadata that is included in the package that links the content across all platforms, like an asset ID?
GV: There are efforts going on right now to try and create a standardized asset ID, so that if X studio creates a piece of content and then they syndicate it, that asset ID can carry through.
CW: And it is not erased or eroded across platforms?
GV: No -- that is the point. It is to become a universal UPC code. There is an effort by IDER to create standard program codes. So if you have a series like “Friends,” when it airs on a network and then goes through many syndications, that ID would carry forward, possibly as long as the syndication continues, even over 30 years.
IDER is a great concept, but there are two challenges with it. One is that you have to have content producers, distributors, everyone, buy into it. The second is that it only tracks a core set of the data -- things like descriptions, actors, year of production -- core things that you need. The challenge is that once someone gets ahold of it, it’s the additional metadata that changes.
With “Friends” as an example, the original NBC airing is very different when it moves into syndication. They edit. They add more ad time, which changes the run time. You could cut out a scene that maybe has an actor in it, [which] makes the credits no longer valid. You have to manage that kind of editing against the metadata, and it becomes a constant effort.