Brands in Crisis: Consumer Help For Deserving Victims

Citation:

John Stockmyer (1996) ,"Brands in Crisis: Consumer Help For Deserving Victims", in NA - Advances in Consumer Research Volume 23, eds. Kim P. Corfman and John G. Lynch Jr., Provo, UT : Association for Consumer Research, Pages: 429-435.

Advances in Consumer Research Volume 23, 1996      Pages 429-435

BRANDS IN CRISIS: CONSUMER HELP FOR DESERVING VICTIMS

John Stockmyer, University of Missouri

INTRODUCTION

Product tampering has a devastating impact on consumers and product marketers in the United States. Since the first Tylenol poisoning incident in 1982, reported incidents of tampering have continued; peaking in 1986 when the FDA logged over 1700 complaints about "invasive" tampering (Rosette 1992). According to FDA commissioner Frank Young, more than one billion dollars worth of merchandise was destroyed in 1986 as a result of product tampering emergencies (Stern 1989). Although the number of reported tampering incidents has declined since 1986, tampering complaints have continued at the rate of approximately 500 per year (Rosette 1992). According to FDA forensic research director Fred Fricke, "Every week we get something that's suspected tampering; it never slows down" (Stehlin 1995). The problem is not limited to cases where an actual tampering has occurred. Negative word-of-mouth (Richins 1983) arising from false charges and rumors of tampering are often sufficient to create a market share drop, as demonstrated by the 1993 Pepsi syringe scare (Meyers 1993).

Surprisingly, few marketing studies have addressed consumer response to product tampering. Jackson et al. (1992) examined how demographic factors influence the general level of concern about product tampering. Morgan (1988) reviewed the legal implications of tamper-related issues such as liability and packaging decisions. "Trade" articles focus almost exclusively on the need for better tamper-resistant and tamper-evident packaging. Most of the research on product tampering assumes a crisis management perspective, which focuses on how companies respond to product crises. It is argued here that the crisis management (C/M) approach overlooks consumer reactions, and treats consumer response as a direct function of management response. The objective of this study is to examine consumer response to product crises. Hopefully, this will serve as an impetus for further research on this topic, and provide managers with additional insight into the complicated issues surrounding company crisis response.

The remainder of the paper consists of five sections. The first describes the current crisis management (C/M) approach and its limitations. The second describes helping behavior theory and derives hypotheses regarding consumer response to product tampering. The third presents the experimental methodology used to test the hypotheses. The fourth section describes the results of the experiment. The final section discusses limitations of the study, implications for managerial policy, and directions for future research.

THE CRISIS MANAGEMENT APPROACH AND ITS LIMITATIONS

Interestingly, product tampering has been adopted as a "management" issue. The C/M literature is scattered and largely atheoretical, but it does suggest basic guidelines for handling product crises (Siomkos and Malliaris 1992):

! Get a crisis management team in place before a crisis.

! Quickly (24-72 hours) acknowledge the problem and show concern.

! Voluntarily recall all harmful products if there is a perceived threat.

! Provide direct information regarding harm data to consumers.

The C/M approach is modeled primarily after the successful actions taken by Johnson & Johnson following the Tylenol poisonings in 1982. However, the validity of the C/M approach is increasingly in doubt. The following example demonstrates the poor predictive ability of the C/M approach.

During the Summer 1993 syringe scare, the Pepsi-Cola Co. did not implement a recall, admit there was a problem, or provide timely information to consumers. Their basic strategy was to deny any fault, which led to criticism for reacting inadequately and not being accommodating to the media (Greenberg 1993). However, two weeks after the incident, sales returned to normal levels (Routhier 1993). Pepsi acted inappropriately, yet their market share did not suffer as a result. Clearly, the C/M approach cannot be relied upon as a normative model for company action in a crisis. Siomkos and Kurzbard (1994) express similar concern regarding the acceptance of current practices as laws of general operation. The C/M approach is anecdotal, and it relies on many unstated and untested assumptions.

One unstated assumption is that if the company does not publicly show its concern, consumers will not feel any sympathy for the company, and therefore will not be likely to continue purchasing the company's product. It is quite possible that consumers might feel sympathy for a troubled company even without seeing such a public statement. Also, the link between consumer sympathy and purchase behavior has not been tested.

The C/M approach does not address brand loyalty or perceived company fault, which are important factors influencing purchase intention. Another problem with the approach is its associated cost if precisely followed. Many smaller firms could probably not afford a recall, testing and/or replacement of the questionable products.

The C/M approach relies on many untested assumptions, ignores potentially important factors, and is extremely costly to implement. The C/M approach has proven to be ineffective in some cases. Even in situations where the approach seems effective, there is no proof that an alternate approach would not have been equally effective. An examination of how consumers respond to product crises is necessary to truly understand the impact of any company strategy on consumer response.

A HELPING BEHAVIOR PERSPECTIVE

This section describes how a reliance on helping behavior literature is applicable in a product crisis context. Some important factors that influence the likelihood of helping are discussed in detail. These factors are then operationalized in a product tampering context, and hypotheses are constructed regarding consumer response to product tampering.

For this study consumer repurchase intention is viewed as a form of helping behavior. The intention to repurchase represents a willingness to assist a victim (the affected company) in distress. Although no prior research could be found in which consumers specifically view companies as victims, this idea has been suggested in previous crisis management research (Kaufmann, Kesner and Hazen 1994). The theory base that guides this study is the norm of justice, which is a rule-based standard that directs people to help those who merit, or deserve assistance (Lerner and Meindl 1981). The critical determination to be made is how one judges when another deserves help. Social psychology literature identifies many factors that influence deservingness and willingness to help. The factors examined in this study are: whether the needy party is trying to help themselves, the relationship between helper and victim, the helper's cost of helping, and the antecedents of helping.

Determination of Self-Help

People feel less empathetic concern for "victims" who are not trying to help themselves, and feel they do not deserve assistance (Betancourt 1990). Self-help is operationalized as the company taking constructive action to "help itself" by recalling products and increasing security following a tampering incident. Company action is viewed as self-help because the primary motivation is to limit its legal liability which would result from additional consumer injuries.

H1: Following a product tampering incident, consumers will rate firms that take constructive action (a) as more deserving of their business, (b) with more sympathy and (c) with higher purchase intentions than firms that do not take constructive action.

The Relationship Between Helper and Victim

Piliavin et al. (1981) reported that people are more likely to be empathetic toward those with whom they have close ties. Clark et al. (1987) and Schoenrade et al. (1986) show that people are more helpful toward those they know and care about than toward superficial acquaintances. The closeness of the relationship is operationalized as the loyalty the consumer has to the brand. This reasoning is consistent with Aaker (1991) who describes brand loyalists as "friends of the brand," because the users generally have an emotional attachment to it. The absence of brand loyalty is viewed as a more superficial relationship.

H2: Following a product tampering incident, consumers who are loyal to the victimized brand will (a) rate the victimized brand as more deserving of their business, (b) feel more sympathy for the victimized brand and (c) report higher continued purchase intentions for the victimized brand than do consumers who are not loyal to the victimized brand.

The Cost of Helping

Another important component of the helping behavior approach is the perceived cost of helping, which is operationalized as the helper's perceived physical risk associated with future consumption of the product. Brand loyalty has proven to be a perceived physical risk reliever for products (Roselius 1971) and services (Mitchell and Greatorex 1993). The following hypothesis extends the concept of risk reduction to a tampering context, where the perceived physical risk can be extreme.

H2d: Following a product tampering incident, consumers who are loyal to the victimized brand will perceive less risk associated with the continued use of the victimized brand than will consumers who are not loyal to the victimized brand.

The Effect of Deservingness, Sympathy and Risk on Purchase Intention

In the tampering context helping is represented by purchase intention. Generally, people are more likely to help others when they feel that the "victim" is deserving of their help (Betancourt 1990; Lerner and Meindl 1981). It has also been shown that helping behavior is directly correlated with empathetic concern (Betancourt 1990). Piliavin et al. (1981) shows that, in general, when the cost of helping increases, the likelihood of helping decreases. A study by Shotland and Straw (1976) found that this relationship holds when the cost of helping is operationalized as the possibility of receiving physical harm.

H3: Following a tampering incident, there is (a) a positive relationship between the perceived deservingness of the victimized brand, (b) a positive relationship between the sympathy for the victimized brand and (c) a negative relationship between the perceived risk of continued use of the victimized brand and purchase intention.

METHODOLOGY

Research Design

The hypotheses were tested in a 2 x 2 between-subjects experiment. The two independent variables were brand loyalty (high/low), and company action (yes/no). The dependent measures assessed four main constructs: deservingness, sympathy, perceived risk and purchase intention (Refer to Appendix 1). Perceived fault was also assessed, because Folkes et al. (1987) and Jorgensen (1994) showed that perceived controllability was inversely related to repurchase intent. Also, Schmidt and Weiner (1988) found that when people attribute difficulties to controllable factors they are usually less willing to help than when they attribute the difficulties to factors beyond the individual's control.

Subjects

Participants in the study were students enrolled in upper-level business courses at a large Midwestern University. Eighty-four subjects participated in pretests. A total of 145 participated in the final experiment. One subject's data was deleted from the analysis due to missing information, leaving a final n of 144 (36 per cell). Although the use of a student sample is not ideal, Dipboye and Flanagan (1979) suggest that the use of a homogeneous sample in the early stages of exploratory research may be useful in enhancing the internal validity of the experimental treatments. Also, the tasks required of the subjects were within the domain of their normal experience.

Pretests

Extensive testing was undertaken to make sure the instructions were clear, to check the reliability of the measures, and to gain additional insight into the tampering phenomenon.

Procedures / Stimuli

Students were randomly assigned to the four treatment conditions. A one page (double-sided) "survey" was distributed. Subjects were informed that this was a study about consumer attitudes toward product marketing practices. Participation was voluntary, and subjects were informed that their identities would remain anonymous. Subjects responded to the brand loyalty manipulation, brand loyalty check, and the product tampering scenario on page one. The second page included the dependent measures and company action manipulation check. The entire procedure took less than 15 minutes. The dependent items were randomized to minimize order bias. Half of the subjects received randomized version A; the other half received its mirror-image, version B.

The initial section included the brand loyalty manipulation. Half of the subjects listed a brand of food or beverage that was one of their favorite products that they purchased on a regular basis (High brand loyalty condition). The other half was instructed to imagine a hypothetical brand of food or beverage that they had previously found satisfactory, but was not one of their favorite brands (Low brand loyalty condition). The subjects then read the following scenario:

"Now, imagine that your (favorite/hypothetical) brand has been "Tampered" with, and the consumption of this product has resulted in serious injuries to several people. Soon after the injuries were first reported, the "Tamperer" was caught, convicted, and put in jail. As the crisis unfolded, the company decided that:

1) "The tampering incident was beyond their control (Not their fault)."

2) "It was an isolated incident."

The third company statement served as the company action manipulation. The constructive action treatment read as follows:

3) "In the best interest of public safety, the company voluntarily removed all of the product from store shelves nation-wide, and made company-wide changes specifically designed to minimize any security threat in the future."

The no constructive action treatment read as follows:

3) "Initiating a recall, or changing company procedures was not necessary."

RESULTS AND ANALYSIS

Manipulation Checks

Brand Loyalty. This scale consisted of three items (a=.83) which assessed product liking (really like/really dislike), product seeking (avoid it intentionally/seek it intentionally), and substitutability (anything would be a better substitute/there are NO acceptable substitutes). The results of an ANOVA test indicated that the brand loyalty manipulation was perceived as intended (F_1,143=256.29, p<.0001), indicating that subjects in the brand loyal group exhibited higher brand loyalty than did those in the non-loyal group (means were 4.34 versus 3.00).

Company Action. The results of an ANOVA test indicated that the company action manipulation was perceived as intended (F_1,143=38.87, p<.0001), indicating that subjects in the constructive action group perceived greater company action than did those in the non-action group (means were 3.87 versus 2.78).

Fault. Analysis indicated that subjects did not perceive a significant difference in company fault between the brand loyalty conditions (Brand Loyalty F_1,143=2.14, p=.15). However, subjects did perceive a significant difference (at a=.10) between company action conditions (Action F_1,143=3.87, p=.051), suggesting that company action reduces company fault in the eyes of consumers (means were 2.72 versus 2.45). Apparently, subjects modified their perceptions of the past event with new information regarding present company behavior. Although unexpected, this result had little influence on the results of the study. The partial correlation between fault and purchase intention (controlling for deservingness, sympathy, and risk) was insignificant (partial r=-.04, p=.309).

Hypotheses Tests

Company Action. The "action" hypotheses were tested using ANOVA. The cell means and critical F-values are reported in Table 1. The results indicate that after a tampering incident, constructive company action is:

! positively associated with deservingness (means: 3.57 vs. 3.84, p=.015)

! positively associated with sympathy (means: 3.40 vs. 3.67, p=.045)

! not significantly associated with risk (means: 2.86 vs. 2.65, p=.13)

! not significantly associated with purchase intention (means: 3.37 vs. 3.56, p=.16)

Hypotheses 1a and 1b are supported; hypothesis 1c was not supported.

Brand Loyalty. The "loyalty" hypotheses were tested using ANOVA. The results indicate that after a tampering incident, brand loyalty is:

! positively associated with deservingness (means: 3.46 vs. 3.95, p<.0001)

! positively associated with sympathy (means: 3.39 vs. 3.67, p=.038)

! negatively associated with risk (means: 2.98 vs. 2.52, p =.001)

! positively associated with purchase intent (means: 3.01 vs. 3.92, p<.0001)

Hypotheses 2a, 2b, 2c and 2d are supported. No significant A X L interactions were found.

Purchase Intention (PI). The following hypotheses were tested using partial correlations (See Table 2). The results indicate that after a tampering incident, there is a:

! positive relationship between deservingness and PI (r=.35, p<.0001).

! non-significant relationship between sympathy and PI (r=.08, p=.167).

! negative relationship between perceived risk and PI (r=-.57, p<.0001).

Hypotheses 3a and 3c are supported; hypothesis 3b is not supported.

DISCUSSION

Company Action

There was no significant link between company action and purchase intention. It appears that the subjects felt more sympathy towards their favorite brands and firms that took action, but that sympathy did not translate into an increased willingness to purchase from (help) the affected company. The fact that sympathy was not significantly related to purchase intent is surprising, because sympathy is thought to be a key factor in regaining market share, and is a major element of the C/M approach. Given the magnitude of the action manipulation, this finding suggests that company action may not be a critical factor in an attempt to regain market share. This could be welcome news for affected companies, because it suggests that extensive (costly) action may not be necessary in the wake of a product crisis incident, as was demonstrated in the Pepsi case. However, it would seem unwise to rely upon the results of this single study for directing managerial policy. The subjects' age, and their potentially different risk perceptions (from the average consumer) may have biased the results. It is also possible that consumer purchase intentions may be influenced indirectly through sympathy and deservingness.

TABLE 1

PANEL A: CELL MEANS FOR DEPENDENT MEASURES

PANEL B: ANOVA F-VALUES FOR DEPENDENT MEASURES

TABLE 2

PARTIAL CORRELATIONS BETWEEN PURCHAE INTENTION AND BRAND LOYALTY, DESERVINGNESS, SYMPATHY, RISK

Brand Loyalty

The most interesting finding is the relationship between brand loyalty and deservingness. Brand equity, or brand loyalty confers many benefits to the firm that owns the brand, such as enhanced profit margins, trade leverage, and higher perceived quality to name but a few (Aaker 1991). However, deservingness of help/patronage has not previously been associated with brand equity. It is known that consumers exhibit strong feelings of commitment toward highly visible products such as cars, clothing, etc., but the results of this study suggest that consumers may feel commitment toward a broader scope of products than was realized. In this study subjects reported strong brand loyalty toward 45 different products ranging from candy bars to fish sticks. Subjects provided both product and brand name specific information.

CONCERNS / LIMITATIONS

Brand loyalty is moderately correlated with deservingness and risk (Loyalty-Risk, r=-.34, p<.001; Loyalty-Deservingness, r=.44, p<.001). This might lead one to question whether these constructs are truly independent. It is possible that an underlying factor (such as liking the product) is solely responsible for driving the purchase intention result. This possibility was investigated in two ways. First, a principal components factor analysis with oblique rotation was conducted to assess discriminant validity between the constructs. A five factor solution was chosen on a priori grounds, and the scree test also suggested retention of five factors. The five factors accounted for 79.3% of the variance. The results indicate that, with the exception of purchase intention item 1 double-loading on the deservingness factor, these five constructs appear to demonstrate discriminant validity (Refer to Table 3). Second, a partial correlation analysis was conducted to assess the impact of each variable on purchase intent (Refer to Table 2). The results suggest that deservingness and perceived risk influence purchase intention independently of brand loyalty.

A potential minor problem with the study is the use of the "hypothetical brand." It is not clear whether subjects envisioned these as national brands, store brands, or generic "brands." However, because the subjects were instructed to imagine a product at the "brand" level, there is some confidence that they did not view the hypothetical product as being unbranded, or generic.

TABLE 3

FACTOR SOLUTION OF KEY MEASURES: PRINCIPAL COMPONENTS ANALYSIS WITH OBLIMIN ROTATION

A serious threat to the study is the possibility that hypotheses 2 and 3 are not specific to the tampering context. One might suspect that loyal brands elicit greater sympathy, deservingness, purchase intent, and less physical risk in any context. In order to show that brand loyalty leads to a differential change in any factor because of the tampering, one must compare the end results to "baseline" measures of these factors before the tampering incident. Although this possibility cannot be completely eliminated, there are many reasons why this issue was not considered to be a major concern in this study. First, existing research does not show that brand loyalty ever elicits sympathy or deservingness from consumers. Second, because the study required subjects to select a brand of food or beverage, which must pass strict FDA quality and safety guidelines, it is almost certain that the (normal) level of perceived risk associated with these products approaches zero. It is also assumed that consumers do not normally attribute feelings of sympathy or deservingness to such products. Therefore, because deservingness, sympathy and perceived risk in all likelihood normally approach a similar minimal level across brands, a pre-crisis baseline is "built-in" to the study. Therefore, any significant increase in deservingness, sympathy, or perceived risk associated with brand loyalty must have been elicited by the tampering incident. Third, all hypotheses are directly related to the tampering context, and the items used to test the hypotheses clearly indicate a connection to the tampering incident. Many of the dependent measure items listed in Appendix 1 directly, or indirectly mention the tampering incident. However, because not every item specifically mentions the word tampering, the ANOVAS and correlations were conducted using each individual item of each scale. The results of the individual item runs were essentially identical to the run made with the summated scales. Therefore it can be concluded that all of the items and the hypotheses are specifically tied to a tampering incident.

Crisis Management Implications

Many questions regarding the efficacy of the C/M approach are raised. The results of the study suggest a possible alternative strategy for handling a no-fault crisis situation. The current C/M approach suggests placing a heavy emphasis on risk reduction and extensive use of the media to elicit consumers' sympathy. Perhaps a more effective strategy would be to allocate a larger portion of the public relations message towards increasing consumers' awareness of the company's use of manufacturing and distribution processes designed to provide customers with products of the highest possible quality, reliability, and safety. Thus, consumers may be more likely to view the company as one with high integrity, and it therefore does not deserve to be "harmed" by a tamperer. Extreme caution must be taken when developing a crisis management strategy. Reliance on any one approach is probably unwise because of the wide range of factors that influence consumers. While this study does not capture all of the important factors; it does identify a previously unstudied factor (deservingness) and suggests that sympathy may not be a critical determinant of market share rebound. Perceived risk and deservingness appear to be the factors that have significant impact on purchase intent following a product tampering incident. A logical next step is to test the effectiveness of the previously mentioned "integrity appeal" versus the current C/M approach.

Future Research Directions

The deservingness factor needs further research attention. Currently, deservingness is not a recognized component of brand loyalty or brand equity, yet this study suggests that it should be in the future. Additional research is needed to explore other factors that might be influenced by deservingness. Also, research into the origins of company/brand deservingness seems warranted. Is deservingness the result of frequency of use, attributions toward the company/brand, truly liking the product, or something else? Perhaps the helping behavior framework could be used to model the relationship between consumers and retailers. A key question might be to determine what attributes a retailer must have to deserve consumers' patronage. Additional research is needed to investigate this issue. This study suggests that deservingness and sympathy toward brands are elicited by crisis conditions. The possibility that consumers possess substantial initial feelings of sympathy and deservingness toward brands should also be investigated.

APPENDIX 1

SELECTED MEASURES

REFERENCES

Aaker, David A. (1991), Managing Brand Equity, The Free Press, New York.

Betancourt, Hector (1990), "An Attribution-Empathy Model of Helping Behavior: Behavioral Intentions and Judgements of Help-Giving," Personality and Social Psychology Bulletin, 16 (September), 573-591.

Clark, Margaret S., Robert Oullette, Martha C. Powell and Sandra Milberg (1987), "Recipient's Mood, Relationship Type, and Helping," Journal of Personality and Social Psychology, 53 (1), 94-103.

Dipboye, Robert and Michael F. Flanagan (1979), "Are Findings in the Field More Generalizable than in the Laboratory?" American Psychologist, 34 (February), 141-150.

Folkes, Valerie, Susan Koletsky and John L. Graham (1987), "A Field Study of Causal Inferences and Consumer Reaction: The View from the Airport," Journal of Consumer Research, 13 (March), 534-539.

Greenberg, Keith Elliot (1993), "Pepsi's Big Scare," Public Relations Journal, 49 (August), 6-7.

Jackson, Gary B., Ralph W. Jackson and Clyde E. Newmiller, Jr. (1992), "Consumer Demographics and Reaction to Product Tampering," Psychology and Marketing, 9 (January), 45-57.

Jorgensen, Brian K. (1994), "Consumer Reaction to Company-Related Disasters: The Effect of Multiple Versus Single Explanations," in Advances in Consumer Research, Vol. 21, eds. Chris T. Allen, and Deborah Roedder John, Nashville, TN: Association for Consumer Research, 348-352.

Kaufmann, Jeffrey B., Idalene F. Kesner and Thomas Lee Hazen (1994), "The Myth of Full Disclosure: A Look at Organizational Communications During Crises," Business Horizons, (July-August), 29-39.

Lerner, Melvin H. and Hames R. Meindl (1981), "Justice and Altruism," in Altruism and Helping Behavior: Social, Personality and Development Perspectives," eds. J. Philippe Rushton and Richard M. Sorrentino, Hillsdale, NJ: Lawrence Erlbaum, 137-165.

Morgan, Fred W. (1988), "Tampered Goods: Legal Developments and Marketing Guidelines," Journal of Marketing, 52 (April), 86-96.

Meyers, Gerald C. (1993), "Product Tampering and Public Outcry," Industry Week, August 2, 41.

Mitchell, V.W. and M. Greatorex (1993), "Risk Perception and Reduction in the Purchase of Consumer Services," The Service Industries Journal, 13 (October), 179-200.

Piliavin, Jane A., John F. Dovido, Samuel L. Gaertner and Russell D. Clark, III. (1981), Emergency Intervention, New York: Academic Press.

Richins, Marsha L. (1983), "Negative Word-of-Mouth by Dissatisfied Consumers: A Pilot Study," Journal of Marketing, 47 (Winter), 68-78.

Roselius, Ted (1971), "Consumer Rankings of Risk Reduction Methods," Journal of Marketing, 35 (January), 55-61.

Rosette, Jack L. (1992), "Tamper Evident Packaging: Law Enforcement and the Consumer," FBI Law Enforcement Bulletin, (September), 16-19.

Routhier, Rick (1993), "Punching Out A Hoax," Sales & Marketing Management, October, 12.

Schmidt, Greg and Bernard Weiner (1988), "An Attribution-Affect-Action Theory of Behavior: Replications of Judgements of Help-Giving," Personality and Social Psychology Bulletin,14 (September), 610-621.

Schoenrade, Patricia A., Daniel C. Batson, Randall J. Brandt and Robert E. Loud, Jr. (1986), "Attachment, Accountability, and Motivation to Benefit Another Not in Distress," Journal of Personality and Social Psychology, 51 (3), 557-563.

Shotland, R. Lance and Margaret K. Straw (1976), "Bystander Response to an Assault: When a Man Attacks a Woman," Journal of Personality and Social Psychology, 34, 990-999.

Siomkos, George J. and Peter G. Malliaris (1992), "Consumer Response to Company Communications During a Product Harm Crisis," Journal of Applied Business Research, 8 (Fall), 59-65.

Siomkos, George J. and Gary Kurzbard (1994), "The Hidden Crisis in Product-Harm Crisis Management," European Journal of Marketing, 28(2), 30-41.

Stehlin, Isadora B. (1995), "FDA's Forensic Center: Speedy, Sophisticated Sleuthing," FDA Consumer, 29(6), 5-9.

Stern, Walter (1989), "A Common-Sense View," Packaging, May, 37-41.

----------------------------------------