Resumo do trabalho

Mobile technologies are increasingly present in retail, yet most purchases still occur in physical stores. This creates a challenge: designing mobile ads that effectively drive store visits. While geolocation and geobehavior are key, little is known about which visual and textual elements influence this behavior. Therefore, this study aims to analyze the features of push mobile messages that positively impact offline store visits, linking online efforts to offline outcomes in a cross-channel context.

This study intends to identify visual and textual elements of mobile ads using computer vision, providing further comprehension on mobile promotional messages content, and its effect on performance results. Despite advances, little research has addressed how mobile content affects cross-channel behavior. The study fills this gap by analyzing mobile message elements—temporal, spatial, and semantic—and their role in driving in-store traffic.

This study examines mobile marketing’s impact on offline store visits, focusing on visual and textual features of push messages. It draws on contextual marketing and the O2O model, emphasizing geolocation, geobehavior, and mobile ubiquity. Danaher et al. (2015) looked into mobile coupon redemption. Li et al. (2017) then looked at another mobile promotion contextual element: the weather. Ghose et al. (2019) looked at trajectory-based targeting and Högberg et al.(2020), at customer labeling on mobile messages. Lastly, Cliquet (2021) with new methods for enlarging geomarketing.

The method consists of a secondary data study that analyzed which visual and textual features are predictors of higher visit through rates (VTR) in mobile campaigns, using computer vision and machine learning in a sample of 640 creatives from different location-based mobile campaigns. To access the add performance, the study focused on the relationship between visual and textual elements of the creative and the dependent variable, the store visit.

This study found that branding appeal and the presence of people in mobile ads significantly increase store visits. Textual features like “discount,” and “participate” also positively affect visit-through rates (VTR). Using computer vision and machine learning on 640 mobile ad creatives, the study highlights the importance of emotional, contextual, and semantic content in driving offline behavior, offering strategic insights for enhancing mobile campaign performance.

The findings show that mobile adds with branding appeal tend to drive more visits to the offline point of sale than those with purchase appeal. A mobile add displaying people tends to drive more visits to the offline point of sale than one with no person. Regarding wording, “discount” and “participate” displayed a positive effect.

The study aims to contribute by providing clearer directions to advertisers on improving mobile add effectiveness. As theoretical contribution, it provided new perspectives on mobile marketing, location-based communication and push-notification effects on customers’ attitudes and behavior, bringing further insights into the “brand in the hand” marketing era.

Brei, V. A., et al. (2020). Machine learning in marketing: Overview, learning strategies, applications, and future developments. Foundations and Trends in Marketing, 14(3), 173–236.
Nanne, A. J., Wiesel, T., & Pauwels, K. (2020). The use of computer vision to analyze brand-related user generated image content. Journal of Interactive Marketing, 50, 156–167.
Tong, S., Luo, X., & Xu, B. (2020). Personalized mobile marketing strategies. Journal of the Academy of Marketing Science, 48(1), 64–78.