Google's generalised second price auction on keywords : analysing its profitability

Abstract

Google is known to be the world's biggest search engine which is one of the most extraordinary internet successes of our time. Since its initial public offering at USD$85, Google’s share price has risen by 7 times. This demonstrates the large increase in the company’s market value and investors’ confidence. In the newly published 2009 annual report, Google’s revenue climbed to 23 billion USD, enjoying a 9% increase compared with the previous year. Noticing the company’s motto being "don't be evil", our interest is intrigued by the remarkable commercial success of such an altruistic company. The main source of Google’s revenue comes from the search engine’s online keywords auction. Google uses an auction mechanism termed General Second Price (GSP) auction. Topics that investigate auction equilibrium model and advertisers’ bidding strategy have been intensively studied before. Nevertheless, whether such a mechanism is profitable for Google is a topic of interest. In this paper, we study three different types of keywords auction mechanisms to analyze their corresponding revenue performances. Under certain conditions, when bidders are truthful, we find that GSP mechanism outperforms the others and generates significant revenue for Google. To gain more insight, we also try to examine some theoretical and technical aspects of the GSP model used by Google. Some aspects such as advertisers’ belief on others’ behavior and their respective degree of truthfulness are potential topics for further research.