Keyword Advertising and Dynamic Pricing
An Integrated Model and its Application to an Online Market Place
– in englischer Sprache –
Hamburg 2015, 282 Seiten
ISBN 978-3-8300-8000-8 (Print & eBook)
Dynamic Pricing, Dynamic Programming, Keyword Advertising, Marketing, Online-Werbung, Online Marketing, Online Market Place, Revenue Management, Search Engine Marketing, Sponsored Search, Stochastic Simulation, Wirtschaftsinformatik
The emergence of Keyword Advertising as a special form of personalized online advertisement has played an indispensable role in the success of web search engines. The costs for Keyword Advertising are determined by a generalized auction that is conducted for every single search query entered in by a user. So far, current Keyword Advertising models rely on a calculation of revenue that assumes the value per keyword is fixed. This approach, however, does not take into account that many advertisers have the possibility to balance higher costs for better ad positions with increased prices of the advertised products, and vice versa. This book therefore focuses on expanding the pure Keyword Advertising model to additionally optimize it for product prices and to account for capacity constraints. To this end, the author develops the joint multi-keyword and multi-product Keyword Advertising and Dynamic Pricing optimization problem as stochastic program. The major disadvantage of this approach is the intractability of determining the optimal solution. For this reason, the author considers two types of approximation models to the stochastic program: the deterministic approximation model and the probabilistic approximation model. With respect to the deterministic approximation model, the author establishes the general solution of the single product problem under some mild assumptions. In the multidimensional case, he proves asymptotic optimality of the deterministic problem when the budget and inventory are controlled in a first-come-first-serve manner, or via partitioned budget and inventory control. With regards to the probabilistic approximation model, the author presents a novel Poisson-approximation method which allows to numerically calculate the optimal solution in case of Poisson demand. In order to evaluate the revenue performance, a simulation tool was developed which aims to mimic the business and demand process from an online market place’s perspective. The numerous simulation based findings of this work serve as a decision support for practitioners who are considering applying Keyword Advertising and Dynamic Pricing separately or together, using an integrated approach.
Informationen über das Veröffentlichen wissenschaftlicher Arbeiten.