Quantifying differences between UGC and DMO’s image content on Instagram using deep learning

Abstract

In the tourism industry, the implementation of effective strategies to promote destinations is considered of utmost importance. Taking advantage of social media, Destination Management Organizations (DMOs) have embraced these platforms as direct channels of communication with potential visitors. However, it remains unclear to what extent these efforts work to effectively construct the desired image and influence visitors’ behavior. In order to explore this phenomenon, this study proposes a comparison of destination images within Instagram, used by both DMOs and visitors (user generated content). Thus, a deep-learning method is presented to automatically compute differences between destination images. Four destinations were selected from Mexico (two urban destinations and two beach destinations). The findings suggest that the images of urban destinations share more significant similarities, particularly in dimensions related to culture, tourist infrastructure, and natural resources when compared to beach destinations. Conversely, the images of beach destinations tend to converge on dimensions such as sun and sand, gastronomy, and entertainment, while differing in aspects related to tourist infrastructure and eco-tourism offerings. It is worth noting that these results underscore the importance of tailoring marketing strategies to the unique characteristics of each destination, taking into account the divergences and similarities in the perceptions of potential visitors.