This scenario has many segments that illustrate several ways that a profile may be modified by a scenario based upon a customer’s actions. In the Pioneer store, we built some profile attributes to track a customer’s preferences for certain kinds of bikes. The attribute pricePreference tells whether a customer favors economy or expensive products, and weightPreference tells whether a customer prefers sturdy or light products. These are actually implemented as simple integer traits that begin at 0, and then indicate the customer’s bias by being positive or negative.
Each segment catches a certain event, and then increments or decrements these values by an appropriate amount. For example, browsing a cheap product may lower the pricePreference by one. However, adding a very expensive item to the cart increases it by ten. In this way, we can make rough estimates about a customer’s preferences.
The last segment records the types of bikes that a customer purchases. This enables the site to again roughly classify its customers according to the types of bikes that they ride.
