Beyond Demographics: Why Emotions Are the Future of CTV Targeting

Pete Crofut, Wurl’s VP of Business Development for Agencies and Brands, has always been suspicious of the “right time, right message” promise of the digital advertising industry. Privacy concerns aside, the proxied approach of third-party cookies left a lot of room for errors. Besides, emotions have a stronger influence on consumer behavior, but there was no real way to select actual placements based on it. So when Wurl, one of the world’s largest platforms for FAST channel operators and a subsidiary of AppLovin, asked him to help build BrandDiscovery, he quickly signed on. 

BrandDiscovery allows brands to target their CTV ads based on the emotions of specific scenes within TV shows. Using various AI technologies, BrandDiscovery analyzes a scene’s text, images, and sentiment, determines an emotional score, and then places ads with a similar emotional score.

The result? Stellar reviews, per Matthew Kramer, Head of Brand Investment at Media.Monks, ​​a global, digital-first marketing, advertising, and technology services company. Pete and Matthew sat down with AdMonsters to discuss BrandDiscovery and the potential to redefine the right time and message based on the consumer’s emotional response to the content they’re viewing.

Susie Stulz: What prompted Wurl to build a platform that allows CTV advertisers to choose placement based on emotions?

Peter Crofut: Looking at the history of linear television buying, advertisers have been adept at leveraging TV programming to create emotional resonance, which drives performance. However, there isn’t much control over where to place those ads or an understanding of where those audiences are and how to forge emotional connections.

That led Ron Gutman, Wurl’s CEO, to ask: do emotions have a role in contextual targeting, and if so, how should they drive CTV targeting? Wurl began testing it and quickly realized that it could drive a 2x to 3x lift in performance.

SS: What exactly does BrandDiscovery do?

PC: BrandDiscovery lets advertisers precisely match their ads with the emotion in context of real-time programming across our supply of over 60 billion monthly premium CTV impressions. Every scene has an emotion, and they bring consumers into that emotional journey. They’re also highly captive, watching that content on the biggest screen in the house in a non-skippable environment. The advertiser has the consumer’s attention by default, but is all attention equal?

We believe that there’s positive attention and negative attention. Positive attention occurs when ads match the emotions of the content and resonate with the consumer at the precise moment the ad is shown. Negative attention occurs when there is a mismatch between the emotion of the ad and the consumer’s emotional journey with the content.

In this respect, BrandDiscovery seeks to earn positive rather than negative attention. We want to make it easier for brands and agencies to select at a scene level where their ads appear so that they’re the most relevant to the audience.

SS: How do you go about determining the emotions of each scene?

PC: We’ve trained our generative AI on millions of pieces of content that we analyze according to Plutchik’s wheel of emotions. This helps us mimic what a viewer may feel while watching that scene. Then we apply generative AI to analyze the sentiment of an ad creative and marry the two. If you’re watching a sad show and a happy commercial comes on, it’s an emotional mismatch, and the brand can come across as insensitive.

We also apply this to matching genres to brand suitability and brand safety considerations at the scene level. Does a show about office dynamics fall into the comedy genre or romance genre? Well, it can be both, depending on the scene. We want to get it right for the advertiser.

SS: What is Plutchik’s wheel of emotions?

PC: It’s a system developed by Robert Plutchik in the 1980s. Basically, there are eight primary emotions—anger, fear, sadness, disgust, surprise, anticipation, trust, and joy—along with three different intensities for each. Combined with the three intensities, these emotions cover the entire expanse of human emotions.

SS: And how did you tie the emotional matching of the scene and ad content to performance?

PC: We have a lot of metadata from our FAST channels and a lot of performance data through AppLovin. When we analyzed that data, we noticed that certain contextual features perform better than others.

SS: Media.Monks is where the rubber meets the road. Have you tried BrandDiscovery with any of your clients? If so, what are your impressions?

Matthew Kramer: We aimed to see if this approach could drive KPIs across the entire funnel. CTV is very effective at driving brand metrics, but we wanted to see if we could drive lower funnel metrics such as engagements. BrandDiscovery increased our brand metrics and doubled the engagements by optimizing against emotional resonance. It cut our cost-per-engagement in half.

SS: That’s cool. What type of campaign did you test?

MK: We used BrandDiscovery with a financial services client that wanted to launch a CTV campaign in its five key cities to drive business in the lead up to the holiday shopping season. The client was keen to optimize the campaign’s bottom line, so we wanted to measure the impact of BrandDiscovery on the full funnel. We worked with Wurl on two studies. The first used Kantar data to measure brand metrics, and the second used EDO to look at engagement.

SS: What emotions did you choose to run the campaign against?

MK: Well that’s the beauty of the platform. We let the algorithm do the emotional scoring. We knew going in that the creatives were likely to score high on trust and joyful emotions, but we left it to the AI to take it from there.

SS: The industry always discusses the right audience at the right time. How does BrandDiscovery update that quest?

PC: It pushes our industry to think differently about that. How do we even know what right is? Consumers expect personalization, even if they’re consuming free content in a skippable environment. Using emotions shows the audience that you understand where they are, and that rewards the brands.

SS: I understand the importance of emotions, but in performance-based campaigns, brands also need to target audiences that have a need for what they’re selling. How do you incorporate that need?

PC: All campaigns start with a KPI, and our approach isn’t different. But we also bring plenty of viewership data about the audience for each channel we operate. This data allows us to hone in on where people are in their journey for that product or service.

SS: In your opinion, Matt, what types of clients should lead campaign targeting with emotions?

MK: I think any product that has an emotional decision behind it, which could be an expensive product such as a new car or jewelry. The beauty category is another good candidate because those purchases also have strong emotions that drive the buying process. Aligning attention to positive attention would be beneficial to the brand.

It’s also important to test the thesis behind an ad’s emotional content to see if it drives the response you want. An insurance ad may opt for a fear-based creative, but does that actually work with the content? I think all brands will appreciate testing this and leveraging that insight.

PC: In addition to what Matt just said, we found that additional emotions performed really well that weren’t in the initial analysis. In the campaign Matt mentioned, we targeted millennial moms and anticipated that a high joy emotion would motivate them to take a specific action. What we found, however, is that anger also drives behavior. 

I can imagine sometime in the future having a matrix that says, “here are the emotions” and “here are the KPIs” that those emotions deliver.

MK: The reason why this is so important to us at Media.Monks are that the shift from linear to CTV is accelerating, but much of the inventory isn’t up for sponsorship. That means we have limited opportunities in CTV, and we need to ensure that we make the most of our opportunities. Being able to have emotional resonance and drive positive attention helps us to maximize the impact of those opportunities.

SS: Final question, what advice would you give advertisers interested in this approach?

MK: It’s important to set aside some budget to test new options. Things like BrandDiscovery and even Disney’s and NBCUniversal’s approach to emotion-based targeting are disruptive. So setting some budget aside to test better and understand the emotions that drive their KPIs is a new way to consider the challenges of engaging an audience and increasing positive attention.