The Relative Effectiveness of Celebrity Endorsement for Print Advertisement

Abstract

The paper reports an experiment that investigates the effect of celebrity endorsers with good or bad fit, an unknown endorser, and compares this to an endorser free control condition. The relative effectiveness of celebrity endorsement is investigated on the basis of Spears and Singh (2004) and the meaning-transfer-model of McCracken (1989). In a pre-test study the fit between products and different celebrities was investigated. On the basis of these results different combinations of advertisements, high and low involvement products and ‘high’ fit score and ‘low’ fit score for celebrity were compared with the combinations without an endorser, and with an unknown (no celebrity) person. The experiment shows that celebrity endorsement is not always effective. This result was also found for the advertisements with the endorsement of celebrities who were found to match best with the products at hand. The results of this experiment therefore suggest that the considerable amounts invested in celebrity endorsement deserve serious consideration.