A study of the effects of reconstructive memory points the way to dealing with the damage to a business’s reputation that follows an instance of negative publicity. The study contradicts the commonly held myth that it is best to avoid communicating for a time and let consumers “forget” an unfortunate incident. Instead, given what is now known about reconstructive memory processes, the crisis situation can be used as a means to reestablish a relationship with consumers. This research investigation proposes that postcrisis communication efforts should be focused on emotionally connecting with consumers via autobiographical-referencing advertising. Moreover, although the study focuses on crisis management, the lessons of reconstructive memory can be applied to all phases of brand management.
For example, see Kate MacArthur, “Chili-Finger Fiasco Continues to Poke at Wendy’s,”Advertising Age76 (April 25, 2005): 20-20.
2.
Allan J. Liddle, “Tipster Fingers Hoax Suspect as Wendy’s Awaits Sales Comeback,”Nation’s Restaurant News39 (May 23, 2005): 3-5.
3.
Numerous Web sites criticized Wendy’s “Free Frosty” promotional efforts; here is one that reported on the “frosty” reception: Rick Aristotle Munarriz, “Wendy’s Frosty Reception,” TheMotleyFool.com, published on May 16, 2005, at www.fool.com/News/mft/2005/mft05051603.htm.
4.
Roland T. Rust, Valarie A. Zeithaml, and Katherine N. Lemon, “Customer Centered Brand Management,”Harvard Business Review82, no. 9 (September 1, 2004): 110-118.
5.
for applications to marketing and advertising, see Kathryn A. Braun, “Post-experience Effects on Consumer Memory,”Journal of Consumer Research25 (March 1999): 319-334.
6.
for a discussion of the importance of loyalty when something goes wrong, see Anna S. Mattila, “The Impact of Relationship Type on Consumer Loyalty in a Context of Service Failure,”Journal of Service Research2 (No-vember 2001): 91-101.
7.
See Harvard Business Essentials: Crisis Management (Cambridge, MA: Harvard Business School Press, 2004).
8.
(While this article deals primarily with Northwest’s advertising after a pilot strike, it mentions the snowstorm as well.) For a detailed discussion of the Detroit snowstorm, see Roger Hallowell, “Northwest and the Detroit Snowstorm,” Harvard Business School Case no. 800053 (Harvard Business School, Cambridge, MA, July 1998).
9.
Liddle, “Tipster Fingers Hoax Suspect as Wendy’s Awaits Sales Comeback.”
10.
Phillip Kotler, John Bowen, and James Makens, Marketing for Hospitality and Tourism, 3rd ed. (Upper Saddle River, NJ: Prentice Hall, 2003).
11.
See Hans Baumgartner, Mita Sujan, and James R. Bettman, “Autobiographical Memories, Affect and Consumer Information Processing,”Journal of Consumer Psychology1, no. 1 (1992): 53-82.
12.
See Suzanne Vranica, “Focus Groups Favor Nostalgic Messages,”Wall Street Journal, October 10, 2001, eastern edition, p. B6-B6.
13.
Vanessa Patrick and Valerie Folkes, “Whodunnit? Accessibility of Blameworthiness in the Firestone Tire Recall,”American Marketing Association Conference Proceedings 13 (2002): 8-8.
14.
Daniel D. Muehling and David E. Sprott, “The Power of Reflection: An Empirical Examination of Nostalgia Advertising Effects,”Journal of Advertising33, no. 3 (2004): 25-36.
15.
For a description of what makes a brand worthy of a relationship, see Charles L. Martin, “Engineering High-Meaning, High-Involvement Brands as a Relationship Marketing Strategy,”Journal of Product Management7, no. 1 (1998): 6-26.
16.
see also Louise Kramer, “Jack in the Box Uses Ads, Menu to Fuel Comeback,”Advertising Age68 (December 1, 1997): 50-50.
17.
Bill McDowell, “Jack in the Box Battles Back with Some Help from a Clown,”Advertising Age67 (November 25, 1996): 11-11.
18.
G. W. Prince, “Yesterday, Today and Tomorrow,”Beverage World, March 15, 2000, pp. 52-54.
19.
See also Marcia K. Johnson, Mary A. Foley, Aurora F. Suengas, and Carol L. Raye, “Phenomenal Characteristics of Memories for Perceived and Imagined Autobiographical Events,”Journal of Experimental Psychology: General117 (December 1988): 371-376, noting that consumers are more likely to have source confusion errors when there are similarities between messages.
20.
Kathryn A. Braun-LaTour and Michael S. LaTour, “Assessing the Long-Term Impact of a Consistent Advertising Campaign on Consumer Memory,”Journal of Advertising33, no. 2 (2004): 49-61.
21.
We hypothesize that the structural relations proposed hold in the case of the autobiographical ad context and not the control ad. Therefore, the hypothetical model references the experimental group’s analysis. The same formulation is tested with the control group to verify the lack of viable structural relations there.
22.
Note that we do not endorse the intentional falsifying of documents, as per John E.H. Sherry, “False and Misleading Promotional Material: A Word of Caution,”Cornell Hotel and Restaurant Administration Quarterly34, no. 4 (August 1993): 14-15.
23.
See Kevin Lane Keller, “Memory and Evaluation Effects in Competitive Advertising Environments,”Journal of Consumer Research17 (March 1991): 463-476.
24.
That research was based on the following earlier investigated idea that imagining an event from childhood makes it seem more likely to have happened: MaryAnne Garry, Charles G. Manning, Elizabeth F. Loftus, and Stephen J. Sherman, “Imagination Inflation: Imagining a Childhood Event Inflates Confidence that It Occurred,”Psychonomic Bulletin & Review3, no. 2 (1996): 208-214.
25.
Though experiments have not often been used in hospitality research, they offer an excellent way to theoretically test different strategies; here, an experiment allowed us to compare the status quo promotional ad against the autobiographical ad. See Ann Lynn and Michael Lynn, “Experiments and Quasi-Experiments: Methods for Evaluating Marketing Options,”Cornell Hotel and Restaurant Administration Quarterly44, no. 2 (April 2003): 75-84. Data were collected in late May and early June 2005, several weeks after Ayala was arrested. Note that the autobiographical ad contained more words than the control ad did. Future research might consider looking at whether the number of words makes any difference.
26.
Wendy’s spokesman Denny Lynch said the sales slump was greatest in San Jose (where the allegation was made) and in Las Vegas (where the incident generated the most publicity). “The sales in Las Vegas have been hurt as if the incident occurred there, which isn’t fair,” Lynch said, as quoted in Brian Haynes, “LV Woman Arrested in Finger Case,”Las Vegas Review Journal, April 22, 2005, www.reviewjournal.com/lvrj_home/2005/Apr-22-Fri-2005/news/26353584.html.
27.
Attitudes were measured on four Likert-type 10-point scales anchored by unfavorable-favorable, bad-good, pleasant-unpleasant, and negative-positive. Higher values indicate more positive attitudes. These scales were highly correlated (Cronbach’s alpha = .89 for attitude toward the ad, Cronbach’s alpha = .95 for Wendy’s, and Cronbach’s alpha = .95 for Mc-Donald’s), so they were combined to form an attitude index. Note that Burger King was meant to be filler so as not to draw the participants’ attention directly to the Wendy’s and McDonald’s comparison (though the attitude toward Burger King scales were also correlated).
28.
These four affect adjectives, which were thought to represent positive, nostalgic feelings, were adapted from Marian Chapman Burke and Julie A. Edell, “The Impact of Feelings on Ad-Based Affect and Cognition,” Journal of Marketing Research 16 (February 1989): 69-83. Participants rated how much each word related to the restaurants. The scales were anchored by 1 = not at all associated and 10 = highly associated. The four affect scales were highly correlated, Cronbach’s alpha = .93 for Wendy’s and Cronbach’s alpha = .89 for McDonald’s.
29.
Specifically, they rated on a 1 to 10 point scale, anchored by 1 = not at all associated and10= highly associated, the question, “How much do you associate _____ with your childhood?” They answered the question, “How much did you enjoy going to _____ when you were a child?” anchored by 1 = not at all and10= extremely well. They were also asked to give an estimate of how often they went to each fast-food restaurant as a child, anchored by 1 = never and10= every day. These three measures were correlated, Cronbach’s alpha = .87 for Wendy’s and Cronbach’s alpha = .83 for McDonald’s.
30.
Two filler events were also included: “Finding an extra toy in a kid’s meal” and “Going through the drive-through after winning a game.” Participants rated how confident they were that these four events happened to them under the age of ten at Wendy’s, where 1 = not at all confident and10= extremely confident. They then rated their confidence these same events had occurred to them at McDonald’s.
31.
The fault measure was anchored by 1 = completely Wendy’s fault and10= not at all Wendy’s fault. The feeling toward Wendy’s after-the-incident measure was anchored by 1 = not very favorable and10= very favorable. How likely they are to visit Wendy’s in the future, how likely they are to order Wendy’s chili in the future, and how likely they are to take their own children (once they have them) to Wendy’s were all anchored by 1 = not very likely and10= very likely. These five measures were highly correlated, Cronbach’s alpha = .88, and were combined for the purpose of our causal model.
32.
Maximum likelihood analysis was used for model testing. For the autobiographical group, c2 = 11.29, p = .08, 6 df, goodness of fit index (GFI) = .923. A nonsignificant chi-square is highly desirable as an indicator of fit. The common benchmark for GFI is .9. See James L. Arbuckle and Werner Wothke, Amos 4.0 Users Guide (Chicago: Small Waters, 1999).
33.
For the control group, ϰ2 = 26.77, p < .001, 6 df, GFI = .847.
34.
The t-test statistics for these measures are affect toward Wendy’s, t(98) = 3.26, p = .001; attitude toward Wendy’s, t(98) =4.7, p < .0001; fault for the situation, t(98) = 4.76, p < .0001; feelings after the situation, t(98) = 6.02, p < .0001; reward with visit, t(98) =5.7, p < .0001; buy chili, t(98) = 3.31, p < .0001; and take children, t(98) =4.2, p <.0001.
35.
The t-test statistics for these measures are associate Wendy’s with their childhood, t(98) = 4.37, p < .0001; enjoyed Wendy’s during their childhood, t(98) = 3.78, p = .0003; and how often frequented Wendy’s, t(98) = 3.28, p = .001.
36.
The t-test statistics for these measures are find-inganextratoy, t(98) =2.4, p = .01; drive-through, t(98) = 3.18, p = .002; family trip, t(98) = 4.8, p < .0001; and Wendy’s Playland, t(98) = 3.94, p < .0001.
37.
All of these measures were on a 10-point scale with higher values indicating more positive feelings and attitudes (or higher confidence). None of these differences was significantly different. In addition, the correlations between feelings toward Wendy’s and feelings toward McDonald’s were not significant; neither were attitudes toward Wendy’s and attitude toward McDonald’s.
38.
The number of people who mentioned the finger incident upon viewing the ad was significantly larger (22 percent) in the control (Frosty) group than in the autobiographical group (10 percent), ϰ2(N = 100) = 2.67, p = .10.
39.
MacArthur, “Chili-Finger Fiasco Continues to Poke at Wendy’s.”
40.
see also Hy Mariampolski, Qualitative Market Research: A Comprehensive Guide (Thousand Oaks, CA: Sage, 2001), for a discussion of early memory probing.
