Manchanda,P . "The effect of banner advertising on internet purchasing" Journal of marketing research [0022-2437] 43.1 (2006). 98-109.

 In this article, internet advertising is examined under the notion that exposure to banner advertising leads to increased advertisement awareness, brand awareness, purchase intention and site visits (as prior research has shown). Specifically, the authors use a “behavioral database that consists of customer purchases at a Web site along with individual advertising exposure,” and see how decisions are influenced by exposure to banner advertising. The main finding is “exposure to banner advertising has a significant effect on Internet purchase behavior, reflected as an increase in purchase probability as a function of banner advertising exposure.” This experiment was able to show the positive effect targeting can have on returns (purchases in this case), and as targeting options increase, the returns should become “relatively higher.” The data also shows that there was a temporal separation between advertising and purchasing behavior, and the authors speculate that the separation is present because the banner advertisements act as a “brandbuilding tool and/or a reminder.” One extremely relevant implication for managers, as indirectly suggested by this data, is “click-through is a relatively poor measure of advertising effectiveness,” because it is a measure of instantaneous behavior.
 While banner advertisements can play a positive role in the world of advertising and marketing, as evidenced by this article, it is important not to generalize this to all banner advertisements. Part of my focus is to provide examples of how banner advertisements can, in fact, be effective. The data from this article certainly suggests that if the creation and promotion of banner advertisements are executed correctly, they can be a useful tool to marketers, providing a crucial starting point for further research of my project.  Additionally, the use of click-through as a measure of advertising effectiveness is poor (to say the least), and I will continue to look for suggestions (in my other sources) to see how to more accurately measure effectiveness.