Understanding travel influencers’ video on instagram: A transfer learning approach

Abstract

This study employs a transfer learning technique for video analytics, collecting and analyzing all Instagram video postings (n = 3981) from the top 40 influencer lists and proposing a novel approach for automated video analytics. A structural approach was performed to estimate the effects of 15 video-related features on consumer engagement. We find that video length, speech sentiment, smiling, saturation, and contrast of brightness are negatively associated with consumer engagement, whereas emotions (surprise, happiness, neutral, anger, sadness, fear), product size, and brightness are positively associated with user engagement.