The Champion of Images: A NeuroAI Framework for Understanding Image Effects on Consumer Decision-Making
Publication Date
5-22-2025
Abstract
In today's e-commerce landscape, consumers' first interactions with products often occur through product thumbnails, called "champion images". Champion images are common in search results and influence click-through behavior. This research investigates what makes these images click-worthy by developing a novel NeuroAI approach-an image mining method that bridges neuroscience and computer vision models. Using a large-scale open-source fMRI dataset, we trained prediction models of neural responses to images and conducted two novel fMRI studies to validate this approach. We applied these models to analyze a global online travel agency platform's hotel search clickstream data. Examining the modeled neural responses to champion images in hotel listings, we found that images with lower neural processing demands were more likely to attract click-throughs, which persisted after controlling for hotel quality, price, and image type. We identified three key neurally-informed features that could influence click-through behavior: retinotopic demand, implied motion, and navigational affordance. Our approach explains additional variance in click-through behavior beyond existing non-neurally informed image metrics. This research establishes a neurobiologically grounded framework for understanding visual marketing, offering theoretical insights into processing fluency in online search environments and practical recommendations for optimizing product images to improve click-through rates.
Document Type
Article
Keywords
E-Commerce, Image Mining, Machine Learning, Computer Vision, Neuroscience, fMRI
Disciplines
Marketing
DOI
10.2139/ssrn.5272560
Source
SMU Cox: Marketing (Topic)
Language
English
