The Brain as Predictor

Consumer neuroscience – applying neuroscience methods and insights to marketing issues – has gained considerable popularity in recent years amongst scholars and practitioners alike (Smidts et al., 2014). As noted by Ariely & Berns (2010), there appears to be good reason for this enthusiasm: brain data is considered less noisy than data obtained through conventional marketing methods. It is thought that data from smaller samples can generate more accurate predictions, making neuroscience methods cheaper and faster than traditional methods.

Although much progress has been made relating brain activity to choice behaviour, evidence that neural measures could actually be useful for predicting market-level responses remains limited. To be of added value, neural measures should significantly increase the accuracy of predicting consumer choices, above and beyond conventional measures.

In this line of research, we set out to investigate this possibility. We obtain both stated preference measures and/or measures of actual purchase behaviour from consumers in combination with neural measures (electroencephalography; EEG or functional magnetic resonance imaging; fMRI) in response to advertisements for commercially released products, to probe their potential to predict individual preferences and sales in the population at large.

Examples of applications concerned the prediction of advertising recall (Chan et al., 2019), commercial success of movies (Boksem & Smidts, 2015), video engagement (Tong et al, 2020), and microlending (Genevsky & Knutson, 2015).

Recently the potential for brain activity to improve predictions of aggregate level behaviour has motivated a new stream of research on neural forecasting. For example, in a study by Genevsky et al. (2017) neural activity collected via fMRI is used to forecast real-world crowdfunding outcomes. They find that neural activity can forecast market success months later and, in fact surpasses predictions made using traditional behavioural methods, such as self-report ratings. In addition to demonstrating the plausibility of neural forecasting, these findings suggest a new perspective on how individual choice might scale to the aggregate level (Knutson & Genevsky, 2018).