How Disney Streaming Approaches OTT Personalization
See more videos like this on StreamingMedia.com.
Learn more about OTT personalization at Streaming Media West 2021.
Read the complete transcript of this video:
Dave Lankford: I think there is a value to the human element and being able to curate the content. But how do you scale that to a global audience? How do you scale that to the wide breadth of audiences across multiple ages? And that's where we really lean into personalization and machine learning, as well as experimentation. And on the personalization side, we start developing that consumer profile after about 30 seconds of viewing. Obviously we may have some information as you come in, in terms of the path in which you purchased, the country in which you purchased, the device on which you purchased.
But at the end of the day, we start looking at that signal of what are you watching. and there are a couple of different areas that we do look at. Obviously, there's what you're presenting on that landing screen on that hom experience. That's an important surface for us. What is the content that's presented, what is the order in which it's presented, both top to bottom and left to right. Those are very important vectors, especially depending on the audience. If it's a left-to-right audience in most Western worlds, that top-left corner starts to be part of that experience. If you're looking at right-to-left in Arabic-speaking or Hebrew-speaking areas, you're going to be looking in a different direction. So we understand those patterns of how people are looking at our content and making sure that we're surfacing the most relevant content up top.
But curation comes in play, especially when you have content that you don't have a lot of signal about. There is importance about priming the pump, making sure that you are exposing that content to folks and those stories, because you'll start to see who's clicking on that, who's watching it, who's consuming it all the way through and not. So I do think that you have to understand what you know, and what you don't know, and accept what you don't and find ways around that. And I would say there are other surfaces--the post-play experience, or the up next, when a video is done, what is that next thing you present to them? That's a really important surface for us, as well as if someone's browsing or searching. Sometimes if someone's searching, they're searching for a title, maybe not because that's what they want to watch, but it's like something they want to watch. And being able to use personalization to understand that using knowledge graphs of our content as well, understanding the deep connections from thematic to story, to arcs, to character types, those all are things that we invest in to say, "Let's know more about our content."
I would also say that the context in which you present it--being able to say something's recommended for you or because you watched this--or even other contexts in which you present that is important. The images that you present, the titles that you present, those all provide the way in which people understand, "Why is this being presented to me? And do I agree with that assessment?"
On top of that, it's not a one size fits all. We have a philosophy on my team which is "Yes, And." If you've ever been to improv, and you've heard Del Close, Del Close always said Yes, And." And that why bring that up is, a lot of times when we were first starting, I would hear teams come to present and they would say, "Well, Dave, here's option A, here's option B, here's option C." And where we've gotten to is, "Let's test option A and option B and option C."
We need to have a hypothesis, but in order to personalize and create the best experience, we have to remove our own personal bias as people--or as teams--and say, "Let's experiment across different algos or different contexts or different ways of presenting content so that we're actually letting our consumers speak to us." So again, I think experiment, you can't really talk about personalization and machine learning without talking about experimentation.
Related Articles
Streaming conventional wisdom tells us that targeted, personalized ads are not only what OTT users want--it's what gives streaming advertising a purpose-built precision that broad-brush broadcast TV advertising can't match. But recent data from Publishers Clearing House suggests that streaming consumers don't put the premium on personalization that we thought they did. Is the jury really out on targeted advertising? Publishers Clearing House's Smriti Sharma, NBCUniversal's Matt Farina, and Media Cartographer Evan Shapiro discuss in this clip from Streaming Media Connect 2023.
17 Apr 2023
Expanding into new territories is fraught, but experts from Disney, DistroTV, and BritBox have found a way. Here are their insights into what OTT platforms need to do to make it work.
09 Nov 2021
In spaces like pharma and CPG, we've seen the value of anonymized databases. The same should become true for the entertainment industry, and it should happen sooner rather than later.
14 Oct 2021
Disney Streaming's Rebekah Mueller explains how Disney works to deliver a seamless, friction-less experience when launching content and brands in different international OTT markets in this clip from Streaming Media Connect 2021.
04 Oct 2021
Studios and producers are reaping the rewards of the OTT revolution. It's about time everyone involved in creating our favorite shows and movies gets their fair share.
29 Sep 2021
New operating system delivers personalized search, discovery, and recommendations to individual viewers without the need to switch between profiles.
09 Sep 2021
HBO Max VP Product Management CJ Harvey describes current and emerging OTT premium content personalization strategies at HBO Max and efforts at human curation strategies that go beyond typics recommendation algorithms in this clip from Streaming Media Connect 2021.
07 Apr 2021
HBO's Sarah Lyons explains how HBO approaches algorithm-driven recommendations and human curation in personalization on the HBO Max OTT channel in this clip from a panel at Streaming Media West Connect 2020.
30 Oct 2020
Resonance AI's Randa Minkarah, Streaming Media's Nadine Krefetz, and American Public Television's Garry Fields discuss the role of machine learning in targeted OTT UX customization in this clip from their panel at Streaming Media West Connect 2020.
26 Oct 2020
Nice People at Work's Marc Maycas discusses how data and personalization can help OTT providers meet their users' QoE expectations in this clip from his presentation at Streaming Media West.
17 Apr 2020