What Makes Them Click: Neuroscience at Scale Paves a Path to Image Effectiveness

Visual content dominates the online consumer experience, with images and videos playing a key role in decision-making. While images and videos are everywhere, says SMU Cox Marketing Professor Gijs Overgoor, the tools are lacking to measure their effectiveness. In new research, Overgoor and coauthors explore how AI and neuroscience can be combined to understand consumer responses to advertising images and examine click-through behavior in online marketing contexts. 

Getting Neural

The research challenge was in overcoming the obtuse nature of traditional machine learning models. “The problem with these models is that they're often a black box,” says Overgoor. “They don't really give us insight into the consumer response to images." The authors asked themselves: Can we present a framework or develop a bridge that can take the scale of machine learning with the insights of neuroscience? "That's how we came up with our NeuroAI framework that can give these insights at scale," Overgoor relays.

The authors developed a “NeuroAI” framework which combines the scalability of machine learning with the capabilities of neuroscience. By training models on fMRI brain response data the researchers predicted how images will be processed at scale. fMRI stands for functional MRI scanning techniques which measure blood flow to the brain. This gave them a window into how the consumer’s brain processes these images, the cognitive demand, and whether certain brain regions are activated in response to a specific image. For example, more complex images require greater mental effort, which is relevant to online browsing, where consumers face information overload. Overly complex images that require significant processing effort can cause users to disengage. Activations in other regions of the brain can be positively associated with marketing outcomes.

Real Estate Apps

Specifically, the research uncovered relevant findings for hospitality and real estate marketing. They learned that images that activate certain brain regions responsible for spatial navigation and awareness perform better. This "navigational affordance" allows consumers to mentally visualize themselves moving through a space, like imagining their experience in a hotel room or Airbnb property, which positively influences booking decisions. In simple terms, Overgoor describes the discovery: "You imagine being in a space, and walking around. If the image allows you to do that, then that would be helpful."

The study experimented with both hotel and Airbnb properties, testing over 3,000 New York properties. Findings revealed that optimizing images for lower cognitive demand and higher navigational affordance significantly increases the likelihood of booking. Overgoor noted substantial revenue improvements of approximately $1,000 annually per property in the Airbnb properties.

Overgoor says that their brain-based measurement approach represents a fundamental shift in how to evaluate marketing effectiveness across industries. Beyond travel and hospitality, potential applications include automotive design, food marketing (menu and recipe imagery), retail experience design, and product photography. 

Their scaled and novel approach could extend from static images to videos and even textual content. Beyond visual processing, similar neural measurement approaches could apply to text-based advertising, headlines, and news articles. “We don't have to limit ourselves to just images,” he says. Having begun this line of research as a Ph.D. candidate in 2017, Overgoor describes this work as being ‘in the early days of exploration.’

Professor Gijs Overgoor will present “Creative That Converts: A Neuroscientific Approach to Image Effectiveness” at the Marketing Effectiveness Accelerator in New York City September 12th. It is based on the paper, “The Champion of Images: A NeuroAI Framework for Understanding Image Effects on Consumer Decision-Making” with Gijs Overgoor of Cox School of Business, 91×ÔÅÄ; Hang-Yee Chan of King’s College, London; William Rand of North Carolina State University; with Willemijn van Dolen and H. Steven Scholte of University of Amsterdam.  

Written by Jennifer Warren.