This paper employs social network analysis to explain variation in the pricing of 846 banner advertisements appearing in a community formed by eighty-nine "liberal" and eighty-four "conservative" Weblogs. As predicted, Weblogs that bridge "structural holes" between otherwise disconnected segments of the community command significantly higher prices for their advertisements. Also as predicted, the price of banner ads increases with the number of impressions received, with the size of the ad, when the ad is located higher on the page, and when fewer other ads appear.
Keywords: online advertising, social network analysis, social capital, social networking, Weblogs, social media
Much of the empirical research on the effectiveness of Web banner advertising may be divided into two broad categories-a concern with communication outcomes and a concern with cost effectiveness. In the former group belong studies of the effects of exposure to Web banners on an audience's cognitive, affective, and behavioral responses. Dependent variables in these studies include brand recall and recognition (Briggs and Hollis 1997; Li and Bukovac 1999; Dreze and Hussherr 2003), attitude toward the brand (Dahlen, Rasch, and Rosengren 2003), clickthrough rate (Gatarski 2001; Robinson, Wysocka, and Hand 2007), and purchase intention (Dahlen, Ekborn, and Morner 2000; Gong and Maddox 2003). Independent variables typically include characteristics of the banner ad itself, e.g., the type of appeal (Xie, Donthu, and Lohtia 2004); the information content of the ad copy (Calisir and Karaali 2008), particularly its relevance and degree of personalization (Tam and Ho 2006); the use of animation, sound, or motion (Yoo and Kim 2005; Chen et al. 2009); as well as the banner's size (Sigel, Braun, and Sena 2008; Burns and Lutz 2008), design (Lohtia, Donthu, and Hershberger 2003), location (Ryu et al. 2007), visual complexity (Huhmann 2003), and color scheme (Moore, Stammerjohan, and Coulter 2005).
Empirical studies on the cost effectiveness of Web banner advertising can be further divided into two groups- algorithmic and strategic. Noting that advertisers compete for the premium space on a publisher's Web page, researchers in the former group have treated revenue maximization as an online variant of the well-studied binpacking problem (Dyckhoff 1990). Accordingly, they developed and tested a variety of scheduling algorithms to optimize advertisement inventory (Nakamura 2002), display frequency (Amiri and Menon 2003; Kumar, Jacob, and Sriskandarajah 2006), and budget allocation (Fruchter and Dou 2005).
The "strategic" studies employ a broader frame of reference for the revenue maximization question. Namely, they consider competitive and cooperative relationships that exist among different participants in the online advertising industry (e.g., Sherman and Deighton 2001). And while their designs and results may differ, they do all agree on one point: the motivations and behaviors of several parties can influence the ad pricing decisions. For example, in their examination of the pricing of banner advertisements Li and Jhang-Li (2009) examine the roles of four key players in the online advertising industry-advertisers, visitors, publishers, and channel providers under two market conditions-duopoly, i.e., the presence of two heterogeneous channel providers (e.g., Google for search advertising and Double-Click for display) and monopoly, where the two channels "are merged into a single dominant player with monopolistic power in the market." Kumar, Dawande, and Mookerjee (2007) include advertisers, publishers, and visitors in their model, while Fruchter and Dou (2005) include the role of advertisers, visitors, and two types of publisher-specialized and generic portals.
Unfortunately, the supply of scholarly research on how to "crack the code of social network advertising" falls far below demand from industry professionals (Williamson 2008). …