Assortment Optimization in the Presence of Multiple Channels and Consumer Preference Uncertainty

Most models of consumer purchase process that are used in operations focus on purchases from a single channel, typically the offline brick-and-mortar store [3]. However, there has been a proliferation of multiple retail channels, such as online and mobile. This proliferation has created the need for operational decisions to model consumer switching between different channels during the purchase process. Even when shopping for medium-priced items such as backpacks, furniture, or electronics, many consumers typically visit the store’s webpage to learn about product attributes, before potentially visiting the brick and mortar store. The purchase can now happen at any point: either online before offline store visit, in the offline store, or online after the store visit1. Despite the higher variety and convenience that the online channel offers, consumers continue to visit the offline brick-and-mortar stores because they may have a hard time judging the value of certain attributes online. For example, consider a consumer shopping for a backpack. He may be certain about his utility for the color Black, but uncertain about his utility for colors “Pacific Blue” and “Safety Cone Orange”, or the value of a padded hipbelt and an Airmesh back panel. The consumer can learn about the available attributes from the store webpage and then form beliefs about his values for the attributes. However, he may need to physically examine a product with a particular attribute in the brick and mortar store to learn that attribute’s true value.