Unveiling the power of video sentiment analysis for predicting advertising effectiveness: exploratory research on femvertising

Abstract

This research challenges the conventional method of textual sentiment analysis in evaluating advertising effectiveness. We examined two distinct video ads for a cosmetics brand, employing video sentiment analysis to gauge their emotional impact on 35 female participants. This assessment was conducted using a self-reported questionnaire and a biometric platform integrating facial detection analysis (using Affdex software by Affectiva) and Galvanic Skin Response (GSR). Both video and text sentiment analyses were performed on two 30-s television ads—one promoting women’s empowerment and the other emphasizing anti-aging properties. While the textual analysis suggested a generally positive tone for both ads, video sentiment analysis, incorporating facial expressions via the Face API, revealed a significant contrast. The ad highlighting youthful appearance displayed higher positive emotional content, particularly “surprise,” compared to the “femvertising” ad promoting women’s empowerment. This heightened emotional response, as identified by video sentiment analysis, had a notable impact on participants’ emotional reactions to the ad, correlating with increased purchase intention and ad enjoyment. Conversely, no significant correlations were observed between emotions, purchase intent, and ad enjoyment for the femvertising ad. This outcome underscores the superiority of video sentiment analysis over traditional textual analysis, which fails to capture the emotional nuances conveyed through facial expressions. In summary, this study demonstrates the potential of video sentiment analysis in predicting purchase intent and assessing advertising appeal.