Firms have increasingly turned to rich digital media, such as videos and photos, to attract attention and boost awareness. Although extant research may help firms promote these media more effectively, the marketing process truly begins with creation of the media. Thus, content creators may benefit from understanding what media is likely to achieve greater popularity, based on its content features. We develop a method to understand the effect of content on the consumption of online videos, and employ our method on a unique dataset including 16,414 videos from 363 YouTube channels. Our approach labels videos as high- or low-performing relative to comparable videos, and leverages random forests to identify content features associated with performance level. We test this method using the personality of speech-driven videos, employing NLP to estimate the extent to which video captions exhibit each of the “big five” personality traits. Our analysis uncovers predictive, economic, and prescriptive insights. We find that using just their personality, we can predict whether videos perform better than expectation with 72% accuracy. Furthermore, videos associated with high-performing personalities can expect a nearly 15% increase in consumption. Finally, we examine which personalities are associated with high consumption, offering prescriptive insights for content engineering.
Krijestorac,, Haris; Garg, Rajiv; and Saar-Tsechansky, Maytal, "Personality-Based Content Engineering for Rich Digital Media" (2019). BLED 2019 Proceedings. 13.