Is a picture worth a thousand views? Measuring the effects of travel photos on user engagement using deep learning algorithms

Abstract

Travel photos inform and inspire consumers by conveying a first-hand destination experience. Despite the proliferation of consumer-generated travel photos in online travel review sites, deconstructing the effects of photos on consumer engagement remains a challenge to the tourism industry. We provide a framework to process and interpret various photographic elements on user engagement using deep learning algorithms. We posit that a photo can invoke consumers’ subjective interpretations of photos representing authentic, creative, or emotional dimensions of the destination experience. A structured crowdsourced categorization process was deployed to measure the interpretive dimensions of the photos. The objects in photographs are identified using a novel deep learning algorithm for controls. We use narrative framing concepts to theorize their influence on user engagement in an online travel review site setting. Relevant three sets of hypotheses are tested using a large dataset of photo-based travel reviews sampled between 2012 and 2014. A negative zero-inflated binomial regression is used to estimate the effect of subjective interpretation of photos on user engagement, accounting for overdispersed excess zeros associated with count outcomes. Results support the hypotheses. The additional analyses explore other plausible influential attributes to user engagements to complement our main findings. We discuss the theoretical contributions to the online-image-persuasion stream of research and practical implications for online tourist review sites.