For the past decade, figuring out how to adapt traditional TV businesses to new consumer habits has been one of the industry’s most widely debated and discussed issues.
This isn’t easy or cheap. For decades, broadcasters, cable networks and pay TV operators relied on highly specialized, proprietary technologies that typically required many months if not years to deploy. That approach produced very high quality video and extremely reliable systems, but it left the TV industry flat-footed as it faced rapidly changing consumer technologies over the last 15 years. In response, networks and operators began searching for infrastructure that would allow them to launch new services and channels in days or weeks, not years.
One example of the imperatives facing programmers is the seemingly obscure issue of metadata.
Metadata, or information associated with a specific piece of video, is important because it allows programmers to use granular detail to better personalize the delivery of content for a specific consumer. But adding metadata to shows has traditionally been a labor-intensive task, requiring people to manually input many details about a program.
Discovering Machine Learning
Over the last year, Discovery Inc. embarked on an ambitious initiative to use machine learning and artificial intelligence to add more metadata to the video in its library, executive vice president of technology strategy and operations Brinton Miller said.
At the end of the process, humans still have to look at the content to confirm that the machine’s identification of a Corvette is in fact correct, but Miller said the system is now 80% to 90% accurate.
“We haven’t yet seen the fruits of that labor, processing years and years of back content, but you will see it next year and it will be very important in the future,” he said. “We’re living in a world where you are increasingly programing for the individual and you need control over your metadata to be able to deliver those kinds of personalized services.”
This project was the latest move in Discovery’s larger strategy to create a much more flexible infrastructure by moving operations into the cloud and unifying its broadcast and digital operations. As part of that effort, Discovery has moved the playout of about 300 channels into cloud-based services run by Amazon Web Services (AWS). That means the programmer doesn’t have to build a new facility to launch channels.
“There are still a lot of complexities to launching a channel or a digital product, but heavy infrastructure is no longer the roadblock,” Miller said. “We are at the point where launching a new feed to an affiliate or a digital product is down to minutes, not months.”
In the TV provider sphere, companies such as Comcast already make heavy use of cloud-based systems for network DVR functionality and advanced video platforms like X1, making it much easier and faster to update user interfaces, roll out new features, add new content or serve up personalized shows.
That has helped boost video-on-demand usage and viewing, Comcast Cable vice president of entertainment services Daniel Spinosa said. In the third quarter of 2018, Comcast customers watched more than 1.6 billion hours of content via VOD, the X1 platform’s OTT apps or the Stream platform, up 23% from 2017.
To further improve the user experience, operators are looking to expand edge computing, CableLabs president and CEO Phil McKinney said. “Rather than having a big centralized cloud operation a thousand miles away, edge computing allows you to take a lot of things and put them closer to the consumer to deliver a very responsive and low latency experience.”
The Gig Economy
Operators have also been working to increase broadband speeds to enable many new digital services. “In 2016 only 4% of homes served by cable operators have access to 1 [Gigabit per second] speeds,” McKinney said. “Today, more than 70% of those homes have access to 1 Gpbs speeds.”
The DOCSIS 3.1 technologies that have played a key role in those faster speeds can also be scaled up to 10 Gpbs, McKinney noted, and the specification can handle full duplex broadband speeds of 10 Gpbs both upstream and downstream.
Equally important, the research consortium has done work on “coherent optics” that will boost the speeds of the fiber cable operators already have in the ground.
“It means that the fiber already in the ground can get a 5 times capacity increase right off the bat, and there is a road map for coherent optics to take those improvements to 10 times or even 100 times faster,” he said. “It means we take the cost model of needing to dig up ground and add strands of fiber to address the need for more capacity out of the equation.”
These improved services helped the largest cable operators add 2,165,000 new broadband subscribers in the first three quarters of 2018, which positions them to profit from the shift to streaming media, according to Leichtman Research Group.
The next section will explore some of the tech trends executives will follow at this year’s CES.
More from Viewer Watch 2019: