The Super Bowl, to many, is thought to be a marketing holiday—a day when the best creatives in the industry attempt to capture the minds, hearts, and attentions of fans. The hype is not hyperbolic; the Super Bowl is extraordinarily massive in terms of its reach and impact. Super Bowl XLIX (aired in 2015), was the most-watched event in television history, attracting over 100 million viewers. The advertising component was equally as spectacular, with over $385 million dollars spent on ads.
Equally as incredible as game day viewership and ad spend was engagement, specifically on Twitter. The average number of tweets posted by a Twitter user tweeting while watching Super Bowl XLIX was 4.4, compared to an average of 2.3 during regular season games. While a lot of these tweets did focus on the game itself, many focused on non-game components, namely the halftime show and the ads.
While the count of tweets about ads was colossal, these tweets were not evenly distributed across ads. Within the first ten minutes after airing, the top 6 ads were tweeted about a combined 70,415 times, while the bottom 8 ads were tweeted about a mere 3,477 times.
Michael E. Smith and Avgusta Shestyuk of Nielsen, along with their colleagues, were tasked with investigating this huge difference. Using a combination of neuroscience tools, social measures, and TV data the team sought to answer two questions: What made the top ads so great? and How can one evaluate the content of tweets to better understand what viewers think of an ad?
What made the top ads so great?
Using a day-after cued recall, the team tested recall for top ads vs. bottom ads. The top ads scored higher on ad memorability, brand memorability, and brand linkage than bottom ads. In terms of neuroscience, the team made use of EEG and facial coding technologies to test neural difference in viewers while seeing the ads. In terms of EEG, the top ads had better memory scores for branding moments, which was consistent with self-report measures. Looking at facial coding, while there was no significant difference in surprise or negative reactions, viewers smiled three times more in response to top ads than bottom ads.
How can one evaluate the content of tweets to better understand what viewers think of an ad?
The team then moved from understanding how top ads differed from bottom ads to trying to gauge viewer disposition towards the ads via tweet content. The team, acknowledging that not all tweets are created equal, took a look at Super Bowl XLI (aired in 2017) data, and found that sentiment manifested itself in an unexpected place: Emojis.
Around 30% of tweets about Super Bowl XLI ads contained Emojis, with positive Emojis (“heart eyes,” “tears of joy,” etc.) being the most frequently used Emojis across all ads. Importantly, the team found that the use of Emojis was greater in ads with higher Twitter volume.
The Emoji is also unique. Not only is the Emoji used as a succinct expression of sentiment (average tweet length decreases when Emojis are used), but the selection of Emoji can reveal a lot about how viewers understand the narrative of an ad. For example, the most commonly used Emoji in tweets describing a comedic ad by Bai Water was “tears of joy,” signifying a humorous takeaway, while the most commonly used Emoji in tweets describing an Airbnb ad with an accepting message was a heart, signifying a more sentimental takeaway. Additionally, the use of Emojis with skin tone modifiers was 43% for the Airbnb ad (vs. 24% for the Bai Water ad), cluing the research team into a resonance of the ad with a more diverse group of viewers.
Ultimately, through a combination of neural and social measures, both manipulated and observed, Smith, Shestyuk, and their colleagues were able to uncover how and why certain ads resonate with viewers, as well as how moving beyond traditional advertising metrics can provide a fuller, more nuanced view of audience perception.