Words Meet Photos: How Image-text Thematic Consistency Effect Crowdfunding Success

Narrative text and cover images in crowdfunding appeals have been studied individually, but there is little research on their interaction. This paper investigates the impact of the thematic consistency between cover images and narrative text on crowdfunding success and its boundaries. Using a large-scale field dataset from Kiva, comprising over 2.19 million entries, and applying generative artificial intelligence (GAI) and machine learning algorithms, we quantified the thematic consistency between cover images and narrative text. We found that, overall, the degree of thematic consistency between cover photos and narrative text has a U-shaped effect on the likelihood of crowdfunding success. We also demonstrate that this effect is heterogeneous, varying according to the features and design complexity of the cover image, as well as the length of the narrative text. Our findings offer insights for entrepreneurship research and practice.