Neural network based empirical models are used in a wide range of marketing applications. Despite the growing use of the internet as a direct marketing medium, with every increase in surface postal rates the need to target direct mail offers only to those who are most likely to respond becomes more important. Neural networks are used to learn the critical attributes of a sampling of people in a large database who are amenable to particular kinds of offers; then only people who have those attributes are selected from the entire database to receive the direct mail.

Similarly, as consumer purchasing behavior is increasingly capable of being monitored in real-time using the internet, neural network models can learn not only the preferences of individuals, which then an organization can use to target specific offers; but also the aggregate collection of data across a broad spectrum of customers permits constructing models of aggregate behavior that can suggest future consumption trends. In addition, discovering patterns in purchase behavior can provide the basis for making rational choices for real-time price adjustments as incentives for customers to purchase additional goods or services.