This dissertation presents innovative demand management techniques for service systems with limited resources. The first study analyzes demand management policies of animal shelters with limited Kennel space as a set of interacting stochastic queueing systems. In practice, there are two main policies being used, which we call "Kill" and "No-Kill" policies. In a "Kill" system, animals may be euthanized if a shelter is full. Many shelters have moved to a "No-Kill" policy, where they avoid killing for space and adopt other approaches to reduce supply and demand mismatch. Our goal is to provide insights on how No-Kill policies, such as coordination, adoption and neutering campaigns, help reduce the animals' killing rate so that the shelter management can choose the way to effectively solve their problems. In the second part, we consider a topic of demand management for the Sports and Entertainment (S&E) industry, called "Scaling the house", i.e., how to divide seats into zones for different prices to maximize revenue across the venue. From the data obtained from several performance venues in the U.S., we find ticket demand is impacted by locations of seats as well as by price. We characterize closed-form solutions for the optimal two-dimensional zoning decision (with row and column cuts) and the one-dimensional decision (with row cuts), and explore when each model should be applied. The third study considers pricing as a tool to manage demand for the S&E tickets. We develop dynamic pricing with demand learning models where demand is also affected by time left until the show dates. Since the show's popularity is usually uncertain to the seller, we propose a method to learn the overall popularity via Bayesian updates. We perform computational experiments to understand properties of the model solutions and identify when demand learning is most beneficial.