Abstract
Our research is inspired by a leading online retailer using clickstream big data to estimate customer demand and then ship items to customers or hubs near customers by a mode of “delivery before order making” (DBOM) mode. Using clickstream data to obtain advance demand information in order quantities, we integrate the forecasting with a singleitem uncapacitated dynamic lot sizing problem in a rolling-horizon environment. Using the simulated clickstream data, we evaluate the performance of DBOM mode.