Abstract
Keyword advertising is a thriving industry in recent years. Advertisers always have difficulties to select lots amount of keywords, and have no idea how to evaluate the value of each keyword effectively. In this research, we employ social network analysis methodology in keyword advertising. First of all, we provide clear image of current market situation among 6 industries. Then, we did experiments to seek the relations among impressions, CPC and CTR. Moreover, we examine these variables again by adding social network indicator, degree centrality. Finally, we use regression model to compare the prediction capability of these variables to CTR. As we found in the experiments, the more impression the keyword has, the higher CPC it will be. However it will have lower CTR, vice versa. From the point view of social network, higher impressions will attract more bidders, which is also the definition of degree centrality. High degree will positively related to CPC but negatively to CTR. Furthermore, it has been proven that impression, CPC and degree centrality can predicted CTR in simple regression model.