This paper addresses the determination and evaluation of optimal Internet marketing strategies when a firm is advertising on multiple websites. An optimization model is constructed for the determination of the optimal amount of click-throughs subject to a budget constraint. The underlying network structure of the problem is then revealed and exploited to obtain both qualitative properties of the solution pattern as well as computational procedures. In addition, three different pricing schemes used in Internet marketing are quantitatively compared and indices that can guide marketers to shift from one scheme to another are proposed. Finally, two numerical examples are constructed that demonstrate two paradoxes: (1) that advertising on more websites may reduce the total responses and (2) that advertising on more websites may reduce the click-through rate. Through the analysis of the network model, such puzzling phenomena are then quantitatively explained.