Seema Pai, Boston University
Phone: (617) 353-4423
Fax: (617) 353-4098
S. Siddarth, University of Southern California
Phone: (213) 740-5048
Fax: (213) 740-7828
The Impact of Word-of-Mouth on Purchase Decisions: The Case of Motion Pictures
Word-of-mouth or “Buzz” is a phenomenon that has created several marketing legends. The traffic-stopping retro Beetle, the addictive Pokémon, cuddly Beanie Babies, the hair-raising Blair Witch Project are all examples of blockbuster commercial success driven by customer hype. People like to share their experiences with one another – the restaurant where they ate lunch, the movie they watched over the weekend, the computer they just bought – and when those experiences are favorable, the recommendations can snowball, resulting in runaway success. Word of Mouth (WoM) has been recognized as one of the most influential sources of information transmission since the beginning of society, especially for experience goods (Godes & Mayzlin, 2004) and has been the subject of extensive research in marketing (Anderson, 1998; Bansal & Voyer, 2000; Brown & Reingen, 1987). This prior literature has collectively identified three key dimensions of WoM communication that determine its effectiveness: who (Bone, 1995; Engel, Blackwell & Kegerreis, 1969), how much (Arndt, 1967; Herr, Kardes & Kim, 1991) and what (Crane, 1989; Feldman & Spencer, 1965). The who dimension refers to the source of the WoM information, how much refers to the volume or quantity of WoM and the what refers to the content or the body of meaning conveyed by the symbolic elements, such as the verbal, musical and pictorial cues in the WoM communication. Recent technological advances have enabled consumers to easily access real time information and exchange opinions about companies, products and services on an unprecedented scale. The emergence of online communities, discussion groups, blogs and opinion websites has changed how consumers interact with one another and provided new avenues for word-of-mouth communications. These changes have led to a renewed interest in measuring buzz and understanding its impact on firm performance. Most of the recent literature on the impact of word-of-mouth has focused on the “how much” aspect of WoM, i.e., on establishing that buzz volume significantly impacts product sales (Chatterjee, 2001; Chen et al, 2005; Chevalier & Mayzlin, 2006; Dellarocas et al, 2004; Godes & Mayzlin, 2004). However, this research has either ignored the “what” aspect of WOM or has found that the content or valence of buzz does not impact performance. This is puzzling given the large body of research in consumer behavior and communications showing that content is an important driver of the effectiveness of word of mouth. A shared characteristic of the above studies is that buzz valence is operationalized with easy-to-process summary measures, such as the star ratings that accompany a review, while the actual textual content of the reviews is completely ignored. This approach discards potentially valuable diagnostic information in the actual text and relies on measures that may only be imperfectly correlated with the actual quality, which may explain the null results for content. An important objective of our research is to develop richer measures of buzz valence based on a computerized text-analysis of various sources of WoM and to show how it impacts product sales. Additionally, with regards to the “who” dimension, prior research has generally focused on a single source of buzz, namely, user reviews about a product. However, because all communication sources may not be equally effective (Fishbein & Ajzen, 1975;Sternthal, Dholakia & Leavitt, 1978), we extend the scope of previous work...