The new selling techniques enabled by information technologies in today's marketplaces, such as online sales channels, search portals, and review platforms, changed the consumer-driven demand in many ways. Unlike traditional retail competition, mostly driven by product attributes (e.g., quality, price, etc.), these selling techniques based on information technologies have become more important to consider customer behavior and its resulting effect in shaping demand, in order for firms to better plan their operational strategies. In this dissertation, we investigate different sources of demand uncertainty and obtain insights into operations of the firms competing in the current marketplace. We develop methods for more accurate estimations of demand in the presence of downstream customers' choice behavior or social interactions. We adopt the Markov Chain based model to understand customer demand and validate the model using human-subject experiment and field data. We also conduct empirical research to capture online browsing behavior of consumers and provide implications to operational managers. This dissertation consists of three chapters.

• Chapter 1: The Effect of Social Information on Demand in Quality Competition. This is joint work with Professor Vishal Gaur and Professor Andrew Davis
• Chapters 2: Predicting Order Variability in Inventory Decisions: A Model of Forecast Anchoring. This is joint work with Professor Andrew Davis and Professor Li Chen
• Chapter 3: Predicting Purchase Propensity from Online Browsing Behavior. This is joint work with Professor Vishal Gaur The three chapters are self-contained but are related to one another: Chapter 1 investigates the impact of social information on demand uncertainty using experimental work, Chapter 2 explores the sources of amplified demand uncertainty from the downstream buyers' inventory decisions, and Chapter 3 empirically explores the effect of online browsing behavior on demand prediction and is a work-in-progress. All these chapters commonly focus on the behavioral sources of demand endogeneity. Therefore, this dissertation aims to contribute to improve the accuracy of demand estimation by incorporating those behavioral factors into the models in Operations.