A Feature-Based Approach to Assessing Advertisement Similarity

This research presents a feature-based statistical model and subsequently explores the degree to which similarity perceptions between two advertisements can be decomposed and explained by a “weighted-and-summed” distance measure, computed on the advertisements' executional elements, after controlling for familiarity and viewers' attitudinal responses toward the advertisements. Furthermore, the authors obtain empirical findings in two major areas: First, variation in similarity ratings can be explained by the advertisements' features, a finding of potential importance for advertisement construction. Second, some, but not all, executional elements that have been shown (in the literature) to drive recall and persuasion are effective at driving perceptions of similarity. This is of practical importance because managers want their advertisements not only to be liked and remembered but also (possibly) to be perceived as similar (or dissimilar) to those for other products. In particular, an understanding of which items drive which constructs (recall and persuasion, or similarity) can contribute to a more effective overall marketing strategy.