NEURO-ALGORITHMIC CONSUMER INSIGHTS: INTEGRATING ARTIFICIAL INTELLIGENCE, NEUROMARKETING, AND PERSONALIZED MARKETING - A CONCEPTUAL FRAMEWORK AND SIMULATION-BASED PROOF OF CONCEPT

There is an increasing tendency in marketing practices of using artificial intelligence (AI) for message personalization while neuromarketing can provide data on consumer behavior that are impossible to obtain from self-reports. Despite similarities, the fields have been developing separately. This paper presents the Neuro-Algorithmic Consumer Insights (NACI) model where AI-based personalization leads to physiological responses (arousal, motivation for an approach, attention), which influence the results from self-reports of emotional response to purchase execution. The results can be further used as predictors of purchase behavior. For the verification of the proposed model, an experiment was conducted with 600 subjects (289 personalized messages and 311 generic). The analysis included Welch tests with Holm correction, hierarchical regression, bootstrapped mediation analysis, and multiple logistic regression. It was demonstrated that personalization leads to arousal (d = 0.45), greater left frontal alpha asymmetry (d = 0.431), attention (d = 0.24), engagement (d = 0.40), and intention to purchase (d = 0.37). It was found that physiological variables can account for 24.1% of variability in intention to purchase over demographic and personalization variables (F(4, 590) = 52.11, p < .001). It is estimated that attention and engagement accounted for 59 percent of the overall effect. Incorporation of physiological factors resulted in the increase in cross-validated AUC for purchase from .61 to .69 according to logistic regression. The results indicate that the designed pipeline reveals previously confirmed effects but does not yield novel insights. The article also reviews limitations of the research and deals with ethical issues raised by it.