Gracenote’s Wheeler And Prasad On The Future Of Personalization, Discovery And Context
In the first of our Innovator Spotlight interviews from our new Special Report, “All Grown Up: FAST Comes Of Age”, we sat down with two of Gracenote’s leading lights to discuss everything from the need for better discovery on FAST to ways to solve the transparency issue. Check it out and download the free report.
“Overall, content discovery is a problem on CTV because there are so many different applications, so many different business models, and constant merging and changing of media empires,” says Trent Wheeler, Head of Innovation at Gracenote whom we interviewed with Kanishk Prasad, VP, Product. “It’s very hard for consumers to understand where the content is that they’re looking for.
Kanishk Prasad (L) and Trent Wheeler (R)
ALAN WOLK (AW): Discovery is still one of the biggest complaints in FAST and CTV overall. People say they can never find what they want because there’s just too much. How does personalization actually help?
TRENT WHEELER (TW): Personalization can help, but I think reimagining CTV UIs could have more impact. FAST tried to solve discovery with the EPG, because the presentation is familiar and people know how to use them. But in FAST the main problem is brand. What do these FAST brands mean? If a channel is called “Gen X TV,” what content can the potential viewer expect? Something like “Pickleball TV” is straightforward. But many FAST channels don’t have brands that clearly articulate the content they deliver.
And because FAST channels aren’t all organized the same way, any FAST channel can feature a broad range of content. If we don’t improve organization and get more clear media brands, then search has to work really, really well. I do think we’ll start to see real progress around AI-based conversational discovery—more natural language, and less “please turn to this app and then look for this thing.” And then, hoping that it all works.
AW: FAST platforms like to frame things as “linear channels vs on-demand,” but in reality everybody has both. From a UX standpoint, how should personalization help people navigate between those?
TW: We want TV to be lean-back as much as possible. The EPG works for lean-back because it’s a grid, it’s easy to read, easy to understand. But when content is spread across a ton of apps—and sports matches are “all over the place” and change a lot—we need a more content-centric discovery experience.
To do that, we have to be less afraid to ask consumers questions about what they want. The number one thing that controls what someone watches when they sit down is: how much time do they have? If they’ve got two hours, maybe it’s a movie. Maybe they just want the TV on as noise while they walk around the house. But no TV asks those core use cases.
So personalization can’t just be the algorithmic part—“can we make better recommendations.” The industry has been doing recommendations for a decade-plus and overall, they’re not great.
We need the other side of personalization, which is the person: give them more agency and control, ask more upfront questions, have more conversation. We built an app ecosystem that looks like a mobile phone. Nobody thinks a phone is a lean-back device.
AW: A lot of home screens now blend “what the algorithm thinks” with paid placements—sometimes it’s basically a billboard. How do you strike the right balance?
TW: That’s a balance the device platform needs to make based on what its users are looking for. Frankly, I think we’re a little too far in on paid. And in many cases, platforms have in-house content, so they’re paying themselves.
I think stronger voice and conversational interaction will help us cut through consumer frustration. Over the next two to three years, I expect rapid changes. And when discovery becomes more agent-driven, you’re going to have to rank content options inside a response: your in-house catalog, your partner catalog, and what the viewer is actually asking for. Consumers will work with the platforms that help them find stuff in the best possible way—you’ve got to strike that balance.
AW: Do you think we’ll see personalized “channels”—an algorithm-like feed that’s unique to each viewer?
TW: It’s absolutely a use case: “I want my TV to play things that I like while I’m doing things around the house.” But TV is different than Spotify playlists or TikTok. Those are three- to four-minute chunks. If you don’t like something, you wait it out or hit next. With a 30- or 60-minute show, people don’t have the same acceptance.
So I don’t think it’s the future of TV as the dominant use case. The dominant driver is still time available, and that requires selection up front. But yes—comfort TV, familiar episodes, a mix of news—those “keep playing” experiences will exist. I just don’t think they’ll be more than about 20% of the overall CTV use case.
AW: The lack of transparency has been a huge issue for FAST. Why are advertisers so concerned about where their ad ran if they know they reached their demo?
Kanishk Prasad (KP): Advertisers care about where their ads ran because context shapes how a message is received. On FAST channels, which are often lean-back environments, content like sports and other “water-cooler” programming, naturally captures more attention than other program types. Additionally, identity solutions for CTV advertising are imperfect since the television is a shared household device. Unlike mobile, it’s hard to know exactly which household member is watching the big screen at any given time. Content, however, can be a strong indicator of who is watching. For example, if the intended target for an ad is the dad of the house, but there were multiple impressions within The Summer I Turned Pretty, it is more likely the teenage daughter was exposed to the messaging.
AW: How can we make advertising more transparent?
KP: I feel like Gracenote already has a head start as the power behind consumer search and discovery on the TV screen. We’ve spent decades building relationships with content producers and distributors to create a single source of truth for entertainment metadata and content IDs which are leveraged across major platforms, streaming apps, networks, and pay-TV providers. We are currently working on bringing that high-fidelity data into the programmatic bidstream, to make decisioning on content signals a reality. In addition to activation, our data enables reporting down to the show level. This gives agencies and brands a crystal clear view of the content that each impression ran against. It is a big step forward in the world of CTV, and one that advertisers have been asking for for some time. More transparency will give the ecosystem more confidence in CTV which will drive more investment and propel the virtuous cycle forward.
AW: Why is contextual targeting becoming so important for advertisers on FAST?
KP: Contextual targeting is especially important on FAST channels because it adds scale and transparency where audience signals alone can be lacking. It helps advertisers align their message and creative with viewer mindset, tone and attention in the moment - factors that strongly influence recall and resonance. Context also provides the most granular brand safety and suitability solution, giving advertisers confidence that their messaging won’t appear within content that could be harmful to their brand. Ultimately, it achieves scalable and meaningful exposure by anchoring ads in environments that naturally reinforce the message.
