Keyword Segmentation, Campaign Organization, and Budget Allocation in Sponsored Search Advertising
Sponsored search advertising, where search providers allow advertisers to participate in a real-time auction and compete for ad slots on search engine results pages (SERPs), is currently one of the most popular advertising channels by marketers. Some domains such as Amazon.com allocate in millions of dollars a month to their sponsored search campaigns. Considering the amount of money allocated to sponsored search as well as the dynamic nature of keyword advertising process, the campaign budget planning decision is a non-trivial task for advertisers. Budget constrained advertisers must consider a number of factors when deciding how to organize campaigns, how much budget to allocate to them, and which keywords to bid on. Specifically, they must decide how to spend budget across planning horizons, markets, campaigns, and ad groups. In this thesis, I develop a simulation model that integrates the issues of keyword segmentation, campaign organization, and budget allocation in order to characterize different budget allocation strategies and understand their implications on search advertising performance. Using the buying funnel model as the basis of keyword segmentation and campaign organization, I examine several budget allocation strategies (i.e., search Volume-based, Cost-based, and Clicks-based) and evaluate their performance implications for firms that may pursue different marketing objectives based on industry and or product/service offerings. I evaluate the simulation model using four fortune 500 companies as cases and their keyword advertising data obtained from Spyfu.com. The results and statistical analysis shows significant improvements in budget utilization using the above-mentioned allocation strategies over a Baseline strategy commonly used in practice. The study offers a unique insight into the budget allocation problem in sponsored search advertising by leveraging a theoretical framework for keyword segmentation, campaign management, and performance evaluation. It also provides insights for advertiser on operational issues such as keyword categorization and campaign organization and prioritization for improved performance. The proposed simulation model also contributes a valid experimental environment to test further decision scenarios, theoretical frameworks, and campaign allocation strategies in sponsored search advertising.