EXPRESS: Predicting the Memorability of Brand Slogans

One component of a brand's identity is the brand slogan, and an important desideratum for the slogan is that it is memorable. We develop predictive models of how well a slogan is remembered and to what extent it cues its brand. To do so, we identify proposed memorability factors through a systematic literature review and operationalize them using a combination of natural language processing, human ratings, and constructs derived from text embeddings. We evaluate the predictive accuracy of models that use these factors and text embeddings on recognition and cued brand recall judgments for existing slogans. We use our models to assess the extent to which the literature is missing textual characteristics important for predicting memorability. Our models provide a tool to help marketers select between slogans, which we demonstrate by predicting the memorability of candidate slogans for new brands.