Evaluation and associations: A neural-network model of advertising and consumer choice

We propose a neurally motivated network model of advertising where advertisements create complementarities among the directly observable attributes of a product, allowing a brand to obtain social value. The formalization of this process allows us to examine the indirect effects of advertising in addition to how ads might affect the decision utility of a product. We focus mainly on the externalities, both positive (spillovers) and negative (dilution), predicted by this model.

► We expand the hedonic pricing model by adding a neural-network structure on product attributes.
► Advertisements act by changing the connection strengths in this network.
► Changing these connections change consumer perceptions of products.
► The network structure predicts that advertisements will tend to produce externalities.
► The predicted externalities match with findings from the marketing literature.