Essays on Entertainment Analytics

This thesis explores live entertainment analytics and revenue management allocation strategies for live entertainment. In Chapter two, we look at empirical factors that effect the success of Broadway shows. How well-known actors (stars) effect film revenues has been a recurring question of entertainment producers and academics. Because a film cannot be disentangled from a star involved, researchers have long struggled to rule out “reverse-causality” that stars have access to higher quality movies. Using a novel data set that includes Broadway show revenues and actor usage, we provide a fixed-effects regression and case studies. We find across multiple specifications that increases in star power in a show improve revenue. Motivated by social grouping and the associated operational challenges, in Chapter three we formulate and study extensions to the Dynamic Stochastic Knapsack Problem (DSKP). We compartmentalize the knapsack according to predefined reward-to-weight ratios, and incorporate a stochastic interaction between the offered set of open compartments and the item placement. Using a specific interaction function inspired by customer choice in the entertainment industry, we provide an algorithm to determine the optimal solution and obtain insights into structural properties. Given the computational complexity of the dynamic program we also propose and analyze via simulation a heuristic algorithm. In Chapter four, in a large sequence of simulations, we propose and study practical heuristic algorithms on which seats should be offered to requests. We propose an algorithm that offers revenue improvements from a “naive” policy on the order of 5-10%. Throughout, we aim for managerial relevance, providing implications to current techniques both in strategy as well as operations.