Design Research Associate
Key stakeholders have a seemingly endless supply of modern technologies and methodologies available to answer vital business questions. But even bleeding-edge tools may not capture the complete, nuanced picture that makes the difference between a rousing success or an expensive mistake.
A major movie studio asked us to predict the success and target audience of an upcoming movie based on emotional reactions and attitudes towards the trailer.
100 participants watched the movie trailer as Affectiva’s artificial intelligence (AI) technology recorded and coded their facial expressions in real time. Participants then answered 15 diary-style video questions using Vidlet’s mobile platform, describing their affective responses and attitudes towards the trailer in their own words.
The AI data uncovered several patterns of emotional response, including one key moment where the slow build of the plotline is interrupted by a sudden burst of colorful and fast-paced visuals, loud EDM music, mysterious drug use, and sexual scenes involving rebellious teenagers. The AI revealed participants reacted exactly how the trailer editors intended: their eyes widened and jaws dropped with surprise.
But, so what? Unlike happiness and anger, surprise can be positive or negative; was this jaw-dropping moment a good or bad thing? The AI data could not uncover this nuance on its own.
The Vidlet platform provided the answers. Mobile diaries told us the tonal shift was genre-defying and edge-of-their-seat exciting to some viewers, while others found the shock-value content to be trashy, shallow, and disappointing. By uncovering subtleties lost on the AI, we could segment viewers three ways: Excited; Intrigued; and Lost. These groups allowed us to assess demographic patterns, learn improvements for the trailer, and most importantly, gauge how likely participants were to see the movie and why.
Valid and unbiased AI can reveal the how of human behavior, but not the motivation. Qualitative data captured with the Vidlet platform unlocked the emotional complexity behind facial expressions – and gave a major movie studio a better understanding of the target audience for their multimillion-dollar film.