Streaming Media East '15: Big Data Baffles Many Video Providers
When it comes to data, “the trouble is now not capturing it...it’s actually making it actionable” that is still presenting many video providers with problems, according to Joe Inzerillo, EVP and CTO of Major League Baseball Advanced Media. That seemed to be the consensus among the panelists in the “Dissecting Big Data: Trends in Video Consumption and Behavior” session at Streaming Media East.
“We’re sitting on about eight years of data. Our use of that data has been very hard,” says Jim O’Neill, principal analyst at Ooyala. Analytics, O’Neill says, “are our bread and butter” but that doesn’t mean making sense of data is any less tricky for his company. And he isn’t alone. He says that he talked with a representative of TiVo recently who said that company has 10 years-worth of data that they are only starting to act on.
Two themes emerged during the discussion about the successful use of data. Human involvement remains important, and so does asking the right questions.
Michael Dube, streaming media manager at NPR, says “NPR, over the last few years, has been quietly figuring out how we want to increase engagement…” Data played a big role in that, but so did human curation. In about four years, NPR managed to increase the average time spent with its content from just about 4 minutes per session to anywhere between 40 and 60 minutes. And now Dube says, “We’re neck-deep in the middle of formulating an overall video plan for the organization.”
Like Dube and NPR, Inzerillo warns against relying too heavily on data. For instance, MLB found that the average time viewers were watching clips on the homepage was about 12 seconds. When the company decided to do some A/B testing using 12-second clips it found that such a short timeframe made for really bad clips. While the data may have showed how long people were watching clips, it wasn’t necessarily telling why they were tuning out.
Asking the right questions and using human intuition and decision-making is important when it comes to working with data, especially for millennials.
Dube told a story about his onboarding process at NPR. He watched an interview of 12 NPR listeners between the ages of 18 and 30. Only one knew NPR was available on the radio. Many of those listeners access content through the NPR One app which provides a more personal experience based on the listener does and doesn’t like.
While statistics may show that millennials only watch about 90 seconds of video, the panel warned against taking that at face value and simply producing 90-second videos. “Millenials just have different taste, and when you give them what they want, you see the stickiness,” Inzerillo says.
Moderator Ade Adeosum, VP of digital enterprise analytics at comScore, says the main difference between millenials and other groups is that their tolerance level is much different. If publishers don't capture their attention quickly millenials will not stick around. For this reason it’s important to use data not only to find out how long users are watching for, but to also discover why they tune out and what keeps them around. In the end, big data is useless if publishers don’t ask it the right questions and unearth the right information.
Related Articles
If you've got lots of video, you've got lots of data, and the best way to maximize the value of that data is with a "strategy first" approach
08 Apr 2015
For media and entertainment companies still storing data in digital silos, that's no way to move forward. Discover the benefits of content storage management solutions.
09 Jul 2014
It's the era of big data, and online video publishers are embracing this trend in a big way. Learn what deep analytics have to offer the streaming video industry.
13 Aug 2013