Neuroforecasting of Consumer Choice

Series of studies have shown that it is possible to predict subject’s value-based choice using functional magnetic resonance imaging (fMRI) data. Several studies were conducted with aim to predict real consumer behavior, not in the laboratory settings, but in the real life. This area of the research was named neuroforecasting. It aims to use group neural activity to predict aggregate choice of a separate independent group. Neuroforecasting is interesting and relevant both for marketers (to increase the accuracy of models of consumer behavior) and for neuroscientists (to develop and validate current models of decision-making). However, this field is relatively young and the number of neuroforecasting studies is relatively small, as the number of studies which have been properly validated using marketing research. Interestingly, neuroforecasting of consumer behavior becomes more and more popular, but there is no single study that investigates neuroforecasting of meals or food items. Therefore, we suggested to (neuro)forecast real life consumer behavior in a restaurant chain using fMRI data that was collected for a sample of participants. During fMRI study thirty participants were exposed to photos of meals from the real menu of the restaurant. In addition, self-reported preferences were collected. fMRI data was be extracted from a priori selected brain regions: the ventral striatum, amygdala, hippocampus, prefrontal cortex, cingulate cortex and insula. We analyzed valuation neural signal encoding preferences for different meals and hypothesized that this signal should correlate with total (real) one-year sales measured by the restaurant chain. Overall, our results demonstrated the feasibility of using neuroimaging data of a relatively small number of participants to forecast real sales.