Seeing is believing? Data mining to create a choice-based conjoint approach for restaurant mobile marketing

Abstract

We propose an online impression-management conceptual model and present a two-step approach to unfold the importance of multiple attributes on a restaurant marketplace mobile app. First, our study uses OpenTable platform data to provide insights into the visual attractiveness of restaurant images. In total, 92,934 images were scraped and further labeled through the Tencent Multi-Label API, which identified six categories of image cues. Second, an exploratory Choice-Based Conjoint (CBC) analysis was performed to assess the importance of attributes at different levels from 746 valid survey responses. Furthermore, we performed Latent Class Analysis (LCA) to categorize the consumers by their choice features in which we identified two groups: “visual-image-cues responsive” and “price-rating-cue responsive.” These categories advance the proposed online impression-management conceptual model and indicate that visual information is the most important attribute in decision making. Our study provides empirically based suggestions to enable practitioners to strengthen their m-commerce marketing strategies.