Many of us remember the 1993 Jack-in-the-Box scandal involving tainted hamburger meat: An outbreak of E. coli food poisoning left more than 700 people ill and four dead (New York Times, 1993). In the wake of the scandal, sales at Jack-in-the-Box plummeted. Although this outcome is not surprising, it is less obvious how the scandal affected competitors such as McDonald’s and Burger King. Did these firms benefit from Jack-in-the-Box’s scandal as consumers looked elsewhere for their hamburgers or did the scandal raise concerns about the safety of hamburgers served by fast food chains and thereby depress sales for their brands as well? More generally, how can a competitor of a scandalized brand assess the likelihood that a scandal will spill over? What strategies can be used to reduce the likelihood of spillover or to respond to spillover when it does occur? These are the questions that Alice Tybout, Harold T. Martin Professor of Marketing at the Kellogg School, and Michelle Roehm (Wake Forest University; PhD from the Kellogg School’s Department of Marketing) sought to answer in their research.
Conditions Which Make Brand Spillover Likely
Drawing upon theories of how consumers organize information about brands and product categories and how they use this information when making judgments, Tybout and Roehm outline conditions under which a brand scandal is likely to spill over. They propose that a brand scandal is more likely to spill over to beliefs about the general product category when: (1) the brand is a typical rather than an atypical member of the product category and (2) the scandal pertains to an attribute that is strongly associated with the product category. Thus, a scandal involving hamburgers at Burger King should be more likely to spill over to beliefs about traditional fast food restaurants than a scandal involving either ice cream at Burger King or hamburgers at Dairy Queen. This is because Burger King is more typical of traditional fast food restaurants than Dairy Queen and hamburgers are more strongly associated with traditional fast food chains than ice cream.
Further, they predict that a brand scandal is more likely to spill over to a particular competitor is perceived to be highly similar on the attribute involved in the scandal. For example, a A scandalized brand may benefit from drawing attention to its similarities to competitors.scandal involving hamburgers but not a scandal involving ice cream might be expected to spill over from Burger King to Hardee’s because Hardee’s is perceived as similar to Burger King in terms of its hamburgers but not its ice cream offerings.
These brand scandal spillover effects are especially likely when contextual factors, such as advertising seen in close proximity to learning about the scandal, highlight attributes that are shared by members of the category. By contrast, if contextual information highlights attributes that are unique to brands in the category, the likelihood of a spillover effect may be mitigated.
Finally, Tybout and Roehm predict that the effectiveness of a competitor’s denial of the scandal behavior will depend upon whether or not spillover has occurred. When spillover has occurred, a formal denial should prompt people to correct their mistaken inference. However, if spillover has not occurred, a denial is expected to boomerang, creating the very scandal effect it was intended to correct. The experiments were conducted to test these predictions.
Scandal in the Fast Food Category
The first experiment focused on brands in the fast food category. The scandalized firm was either Burger King or Dairy Queen and the company was purported to have either misled customers about the nutritional content of its burgers or served tainted ice cream. The scandal was introduced by asking MBA students to read and evaluate a fictional newspaper article that described one of the four scandals (Burger King-hamburger, Burger King-ice cream, Dairy Queen-hamburger, Dairy Queen-ice cream). Measure were taken both before and after reading about the scandal to assess its affect on attitudes toward the traditional fast food category to which Burger King belongs, and attitudes toward the dessert fast food category to which Dairy Queen belongs. Measures of the effect on a competitor, Hardee’s, were also taken.
The findings reveal that the scandal spilled over and affected attitudes toward the category when the scandalized brand was typical of that category and it involved an attribute strongly associated with the category. Thus, the Burger King-hamburger scandal affected attitudes toward the traditional fast food category but not the dessert fast food category, whereas the Dairy Queen ice cream scandal affected attitudes toward the dessert fast food category but not the traditional fast food category. Neither the Burger King-ice cream scandal nor the Dairy Queen-hamburger scandal spilled over to category attitudes.
Spillover to the specific competitor, Hardee’s, was also found, but only when the scandal pertained to an attribute on which the brands were similar. Specifically, the Burger King-hamburger scandal and the Dairy Queen-ice cream scandal spilled over to Hardee’s but the Burger King-ice cream scandal and the Dairy Queen-hamburger scandal did not. These findings are consistent with Tybout and Roehm’s theorizing regarding the role played by brand typicality and the scandal attribute in creating spillover.
Prevention and Response
A second experiment explored strategies for preventing spillover and responding to spillover when it occurs. This study focused on a single scandal, the fictitious Burger King-hamburger scandal described earlier, and its spillover to a new competitor, Wendy’s. Wendy’s belongs to the same traditional fast food restaurant category as Burger King and the two chains are viewed as similar in their hamburger offerings, making it likely that spillover would occur in the absence of a strategy for preventing it. An advertisement for McDonald’s shown prior to introducing the scandal was used to either encourage or reduce the likelihood that spillover would occur. When the ad highlighted attributes that McDonald’s has in common with other fast food chains, this priming of similarities was expected to promote spillover. However, when the ad highlighted attributes unique to McDonald’s, this focus on how brands differ was expected to reduce the likelihood that spillover would occur.
After reading the newspaper article that introduced the Burger King scandal, participants read a second article describing a new menu offering at Wendy’s that either included or did not include a denial that Wendy’s had ever or would ever engage in the scandal behavior (i.e., mislead customers about the nutritional content of its hamburgers). Inclusion of the denial was expected to be beneficial when spillover had occurred because it would motivate consumers to distinguish between Burger King and Wendy’s behavior. By contrast, inclusion of the denial was expected to be harmful when no spillover had occurred because it would cause consumers to wonder why Wendy’s “doth protest too much.”
The results supported the predictions. When the McDonald’s ad encouraged spillover by emphasizing similarities between fast food chains, attitudes toward Wendy’s were more favorable if the company denied the scandal behavior rather than remaining silent. However, when the McDonald’s ad reduced the likelihood of spillover by emphasizing unique attributes of the chain, attitudes toward Wendy’s were more favorable in the absence versus the presence of a denial. In this case, Wendy’s drew unwanted and unnecessary attention to the possibility it engaged in the scandal behavior, resulting in something the researchers dub a “boomerang effect.”
The final experiment examined a different product category, athletic shoes, and a different type of scandal, one related to corporate actions rather than to attributes of products. The fictitious scandal stated that Nike caused significant water pollution as a byproduct of its manufacturing processes. The effect of this scandal on attitudes toward two competitors, Reebok and Asics, who either issued or did not issue a denial, was assessed. The Nike scandal was expected to spill over to Reebok because Nike and Reebok are perceived as highly similar. As a result, attitudes toward Reebok should be more favorable when it issued a denial than when it did not. However, the scandal was not expected to spill over to Asics because Asics is not viewed as highly similar to Nike. Therefore, attitudes toward Asics should be more favorable when no denial is offered, compared to when the scandal behavior is denied. The results of the experiment supported these predictions, thereby generalizing the findings in the first two experiments.
An intriguing and unanticipated result also emerged. When spillover occurred, the pollution scandal was less detrimental to Nike itself. Apparently, if consumers infer the scandal might be a common practice within the industry, they are less likely to think ill of the scandalized company. The scandalized company “rebounds” while its unwitting competitors suffer.
Roehm and Tybout’s findings show that a scandalized brand may benefit from drawing attention to its similarities to competitors. Conversely, a competitor that is perceived as similar to the scandalized brand would do well to differentiate itself as much as possible from the scandalized brand—issuing a post-scandal denial only if it has already failed to differentiate itself.
Editor’s Note: Tybout and Roehm’s experiments involved fictional scandals about real brands. Immediately following each experiment, all research participants were fully debriefed to insure that they understood that the scandals were not true. Attitudinal measures taken following the debriefing revealed that the debriefing was successful in returning participants’ attitudes toward the companies to their pre-experimental levels.
Roehm, Michelle and Alice M. Tybout (2008). “Managing the Unthinkable: What to Do When a Scandal Hits Your Brand,” in Bobby J. Calder, editor, Kellogg on Advertising and Media. Hoboken, NJ: Wiley, pp. 159-177.
New York Times (1993). “Jack in the Box’s Worst Nightmare,” February 6 (Accessed: December 6, 2007).