Susana Costa e Silva, a researcher at the Research Centre in Management and Economics (CEGE) at Católica Porto Business School (CPBS), and an associate professor at CPBS, co-authored an article that was published in September in Psychology & Marketing, a scientific publication dedicated to the application of psychological theories and techniques to marketing. In addition to Susana Costa e Silva, the authors of the study are João Alexandre Lobo Marques and Andreia Neto (Saint Joseph University - Macau), and Enrique Bigne (University of Valencia).
Entitled "Predicting consumer ad preferences: Leveraging a machine learning approach for EDA and FEA neurophysiological metric”, the research explores the ability to predict consumer preferences in relation to advertisements by analysing emotions and using artificial intelligence. Among the study's main conclusions, the researchers reveal that these preferences can be predicted using electrodermal activity (EDA) and facial expression analysis (FEA) in advertising videos.
Electrodermal activity (EDA) is a technique often used to assess levels of emotional arousal, stress or reaction to external stimuli. It is a common component of psychophysiological measurement devices, such as polygraphs (lie detectors) and affective neuroscience studies. Facial expression analysis (FEA) is a technology that captures the participant's image and, using artificial intelligence, maps facial muscle movements and associates them with emotions such as happiness, sadness, anger, fear and surprise.
The research was based on the detection of the seven basic emotions according to the psychology standards - Joy, Anger, Fear, Surprise, Sadness, Disgust, and Contemp -, attention and engagement triggered by advertising. The researchers analysed the data gathered from EDA and FEA and, to achieve the results, developed a statistical module to identify the main physiological characteristics, along with an artificial intelligence system based on machine learning techniques. The research proposes an explainable AI module based on feature importance, which discerned Attention, Involvement, Joy and Disgust as the four pivotal features influencing consumer ad preference prediction.
This study shows that smart systems using electrodermal activity (EDA) and facial expression analysis (FEA) can help marketers predict the behaviours consumers have, making their campaigns more targeted and effective. By using these sensors on a participant while they view communication pieces, product demonstrations or adverts, we can directly measure their emotions—like joy, which could mean that the consumer is more satisfied after the stimulus and, probably, their attitude towards the product/brand is better and their intention to buy is higher. This approach helps marketers understand what truly makes consumers happy, or not, instead of relying on their answers, which often don’t match their real buying behavior.
For more details, see the full article here.
