MIAMI – Although best known for its video encoding software, Sorenson Media wants the television ecosystem to think holistically by using smart-TV data to inform every second of the viewing experience.

Two of the company’s newest products are a measurement solution for local broadcasters and addressability for linear TV, according to VP of Advertising James Shears.

Using automatic content recognition, “We understand what’s on the screen and we can actually measure what people are doing on the TV,” Shears says in this interview at the recent Beet Retreat Miami 2017. “It’s not panel based, it’s not leveraging set-top box data. It’s just doing something that’s not been done in market before.”

Use cases for local broadcasters include informing their promotional activity and bringing insights to newscasts, both based on the entire life cycle of viewers. “During the newscast or local broadcast, did they leave the channel when a particular story was coming on or did they come back and stay and were heavily engaged? Maybe they need to push happier stories or do something they hadn’t really thought about.”

This represents a more holistic approach “where there has been a hunger for measurement that’s a little more informed and a little more second by second,” Shears says.

With addressable, Sorenson Media “takes out the traditional and starts understanding how you can leverage the digital ecosystem to do something that has been done in the linear TV side, but maybe not at the speed with which we can perform. The key for our platform is that we’re essentially creating an ecosystem.”

Because of everything it sees on TV screens, the company says it’s in a good position to “think about the stream itself” and improve viewers’ experiences. That could involve shrinking the 18 minutes of commercial time per hour and other ways “we can think about breaking the paradigm.”

This video was produced at the Beet Retreat Miami, 2017 presented by Videology along with Alphonso and 605. For more videos from the event, please visit this page.