Service Failure Severity,Customer Satisfaction,and Market Share

Abstract

The generally accepted view among managers and researchers is that the greater the severity of a service failure, the greater the resulting impact on customer satisfaction and business outcomes, such as lost customers and revenue. The research used to defend this viewpoint, however, does not typically address the severity of service failures, like those that result in injury or death (i.e., product-harm crises). This research addresses this issue by examining both minor incidents (i.e., failures that do not result in physical harm) and major incidents (i.e., failures that result in injury or death) in the U.S. airline industry, and the corresponding impact on the customer satisfaction and market share of the firms affected. Our results indicate that minor incidents are more strongly (negatively) related to future market share than are major incidents. Moreover, our findings indicate that only minor incidents are significantly linked to customer satisfaction. We argue that these findings occur for two reasons: First, most customers believe major incidents to be low probability events that are less salient when compared to more probable failures. Second, consumers impacted by major incidents most likely defect and are therefore not captured in future customer satisfaction surveys. Consequently, managers can delude themselves that things have “returned to normal” after a major incident when relying on customer satisfaction scores alone.

Keywords

customer satisfaction service failure product-harm crisis market share airline

Introduction

Managers and researchers are keenly interested in determining the sources and consequences of service failures on customer satisfaction and market share. Without question, negative customer experiences harm firm-customer relationships; a large and growing body of research clearly demonstrates that service failures negatively impact customer satisfaction and future customer purchase behaviors with demonstrable consequences for firm financial performance (Hess, Ganesan, and Klein 2003; McCollough, Berry, and Yadav 2000; Smith and Bolton 1998; Smith, Bolton, and Wagner 1999; Zeithaml, Berry, and Parasuraman 1996).

While most research has focused on positive aspects of the customer experience (Kumar, Batista, and Maul 2011; Luo 2007; Salvador-Ferrer 2010), research suggests that negative customer experiences are more harmful to business outcomes than positive experiences are beneficial to these outcomes (Chevalier and Mayzlin 2006; Luo and Homburg 2008; Mahajan, Muller, and Kerin 1984; Rust and Oliver 2000). Indeed, prospect theory (Kahneman and Tversky 1979) supports the higher relative impact and weighting of negative experiences compared to positive ones. Specifically, the theory assumes a function that is steeper for losses than for gains, where in most instances consumers are risk averse and aggressively avoid potential bad outcomes, even unlikely bad outcomes.

The generally accepted view among managers and researchers is that the greater the severity of a service failure, the greater the resulting impact on customer satisfaction and business outcomes (Smith, Bolton, and Wagner 1999; Weun, Beatty, and Jones 2004). Research in the area, however, has generally not investigated the impact of service failure on satisfaction and financial performance including different categories of failures (i.e., both less and more severe). Rather, the literature on service failures can be divided into two general groups: (1) research investigating noncatastrophic failures, that is, ordinary service failures (e.g., Smith, Bolton, and Wagner 1999; we will refer to these as minor incidents throughout the remainder of this article) or (2) research investigating major incidents, that is, defects or accidents that have the potential to cause serious harm to consumers, including injury or death (referred to in the literature as “product-harm crises”; e.g., Dawar and Pillutla 2000). Yet because little research has combined these two types of service failures into a single empirical examination, our understanding of the relative importance of major versus minor incidents on business outcomes like customer satisfaction and financial performance is limited. While arguments might be made in either direction—with major incidents deemed more impactful because of their more dire consequences or minor incidents more impactful because of their more regular occurrence and the larger cohort of consumers experiencing them—the bottom line is that the existing literature offers little guidance.

The lack of overlap in these research streams most likely results from the differences in the probability of occurrence/recurrence and the different responses required to address minor and major incidents. Clearly, most service failures are not product-harm crises that result in physical injury to the consumer. Yet despite their relative low probability of occurrence relative to minor incidents, major incidents still occur with a degree of regularity across a wide range of product and service categories (e.g., contaminated food products, automobile defects and recalls, airline accidents, pharmaceutical recalls, etc.). Furthermore, these catastrophic failures often receive a high degree of media attention relative to their occurrence (e.g., Block 2009; Hedgpeth 2010; Strait and Jennings 1997), as they represent and tap into some of the public’s worst fears. Therefore, major incidents do not just affect customers who have directly consumed the product/service but also tend to generate negative publicity that impacts the broader market.

Research regarding minor incidents—such as those that lead a consumer to complain to the firm (formally or informally) and/or seek recompense from the firm—consistently finds that they impact customer satisfaction, market share, and other metrics of firm performance (Luo 2007; Smith, Bolton, and Wagner 1999). Similarly, research into the impact of major incidents has found that they too have a measureable impact on satisfaction and market share shortly following the incident, although the longevity of the impact has been found to be relatively short (Cleeren, Dekimpe, and Helsen 2008; Vassilikopouloua et al. 2009). As such, and based on this earlier research, we would expect both major and minor incidents to have a measureable impact on consumer satisfaction and future market share. The question of interest in this investigation, however, is as follows: Does the generally accepted view that the greater the severity of annually aggregated incidents, the greater the resulting impact on annual firm customer satisfaction and business outcomes hold true over time when both major incidents (i.e., product-harm crises) and minor incidents are included in the mix of all service failures? Or, as suggested previously, do minor incidents—because of their more regular occurrence and the larger group of affected consumers—have a greater impact?

This research attempts to provide insight into this question by examining the relationship between annually aggregated major incidents, minor incidents, and future customer satisfaction and market share in the airline industry. Contrary to the accepted wisdom and perhaps intuition, our findings indicate that minor incidents have the greatest impact on consumer satisfaction and market share. In fact, we find that major incidents have no impact on future consumer satisfaction, and suggest that this is likely a result of a form of selection bias, that is, customers influenced by major incidents likely defect, are such a small segment of the population that it is improbable that they would appear in a sample such as ours (or are prevented by legal agreements from discussing their experiences even if contacted), or even possibly die. In any case, they are no longer consumers and thus can have no impact on satisfaction levels gleaned through future consumer surveys. We conclude by outlining the importance of these findings for managers and firms attempting to recover from service failures of both types.

Theoretical Background

Product/Service Failure and Customer Satisfaction

There is a great deal of research examining the relationship between ordinary service failures (minor incidents) and customer satisfaction (e.g., Smith, Bolton, and Wagner 1999). Research into service failures and satisfaction spans a variety of domains, for example, banking (Duffy, Miller, and Bexley 2006), retail (Brown, Cowles, and Tuten 1996), eCommerce (Hsin-Hui, Yi-Shun, and Li-Kuan 2011), mobile telephone (Shapiro and Nieman-Gonder 2006), and health care (Mittal, Huppertz, and Khare 2008). These studies universally find that the effect of a service failure is to lower satisfaction at the time a customer experiences the failure. Likewise, service recovery efforts have frequently been found to mitigate dissatisfaction resulting from these failures, providing motivation for firms to better manage the problems their customers experience.

In the case of airlines—the industry under investigation in this study—researchers have examined the role of service failure on satisfaction (e.g., Anderson, Baggett, and Widener 2009; Bamford and Xystouri 2005; Lapré 2011; Lapré and Tsikriktsis 2006; McCollough, Berry, and Yadav 2000), loyalty (e.g., Zins 2001), and market share (e.g., Rhoades and Waguespack 2005). Of particular relevance to this study, Sajtos, Brodie, and Whittome (2010) examine the interaction between the service brand, trust in the brand, and the customer value-customer loyalty linkage in response to the severity of the service failure for airline passengers. They find that the severity of the service failure significantly impacts this relationship. In the case of their investigation, however, it is important to note that “severe” service failures were defined differently than in the current study, and related to time delays (i.e., waiting time at the airport; Sajtos, Brodie, and Whittome 2010, p. 220).

Clearly, customers’ perceptions of service failures directly and negatively impact their level of satisfaction. Moreover, different service failures vary in terms of their importance to customers. For example, customers will likely perceive a 1-hour flight delay as severe if they miss a special event or an important business meeting. On the other hand, most customers will likely view running out of a desired food item on a flight as an annoyance, but not as a severe failure (assuming that the remaining food items do not violate the customer’s dietary requirements).

When thinking of airline service failure severity, however, it is important to remember that the most basic role of an airline—or for that matter, really any firm in any industry—is to guarantee the safety of the consumer from physical harm while experiencing their product or service. At least from an objective perspective, it is hard to imagine an element of service quality in the airline industry that ranks higher than arrival at one’s destination alive and unharmed. The attention paid to safety procedures prior to take-off on every airline would seem to confirm that this value is indeed paramount. Therefore, the most severe service failures are not time delays, schedule changes, lost baggage, frequent flier miles discrepancies, or being involuntarily denied boarding, although these are typically examined in research on airline service failures and customer satisfaction.¹ The worst, most severe airline service failure is, of course, an accident resulting in injury or death. In fact, regardless of customers’ perceptions of the severity of missing “a special event or an important business meeting” to use the examples mentioned previously, it is difficult to imagine any airline customer who would perceive any noncatastrophic service failure as being more severe than taking a flight that results in injury or death.

Fortunately, relative to the total number of flights the number of airline accidents which involve injury or death is extremely small, and the odds of being injured or killed in an airline accident have dropped substantially over the past few decades (Mayerowitz 2013). In fact, such catastrophic failures occur at a rate far better than the six sigma standard typically associated with service excellence.

But while infrequent, the airlines monitored in our study had over 420 major incidents—accidents that resulted in injury or death—from 1997 through 2009. Furthermore, excluding the attacks of September 11, 2001, there were 13 accidents that resulted in fatalities for the seven carriers under investigation during this same time frame.² And while accidents relative to total flights represent a very small number in total, they often have an oversized impact on the public psyche. For example, for several years after the September 11, 2001, terrorist attacks, airlines experienced a huge decline in passenger traffic on that date, requiring them to discount deeply to lure passengers to flights scheduled on that anniversary (Oremus 2011). Therefore, the potential for airline accidents to have a measureable impact on customers’ perceptions of an airline and on future market share would appear to be obvious. There are, however, several factors that might be expected to impact the effect of major incidents on satisfaction and/or market share.

With regard to the impact of major incidents on customer satisfaction, two issues would appear relevant. The first is the role of uncertainty. Uncertainty has long been known to play a role in consumer choice. This is because uncertainty impacts consumer expectations (Rust, Zahorik, and Keiningham 1996) and expectations strongly impact satisfaction (Oliver 1997, 2010). Therefore, in general, the lower the likelihood that customers will experience an event, the lower the likelihood that such an event should influence customers’ expectations. By extension, if consumers truly view the likelihood that they will personally experience a major incident as extremely small, ceteris paribus its impact on satisfaction would be expected to be much lower than for more probable failures.

Second, in general, people find downside risks to be worse than gains from better than expected outcomes (Rust, Zahorik, and Keiningham 1996). Moreover, fear resulting from high levels of perceived risk often leads potential customers to avoid purchases entirely (McDaniel and Zeithaml 1984; Stuteville 1970). As a result, individuals who believe major incidents to be salient in their decision to fly a major U.S. carrier will be more likely to avoid flying the focal airline altogether. As such, these individuals are not customers of the airline; therefore, their satisfaction will not be gauged in customer satisfaction surveys for the focal airline (assuming those surveys are measuring only actual consumers of the company under investigation, as most do). This form of sample bias has been shown to keep satisfaction levels constant (or even result in raising satisfaction levels) despite significant customer defections (Anderson, Fornell, and Lehmann 1994; Keiningham et al. 2014).

Third, extending the prior argument, while arriving safely to one’s destination is of paramount importance to airline passengers (e.g., Gilbert and Wong 2003), most researchers treat this as a hygiene factor that is not significant in the selection of an airline. For example, Rust, Zahorik, and Keiningham (1996, p. 250) write:

It is hard to imagine a customer who will not report that airline safety is “Very Important.” After all, no one wants to be picked up in charred pieces from the ground and placed in a body bag. At the same time, airline safety is not a determinant of either initial choice or satisfaction. Because just about all airlines have indistinguishable safety records (except for commuter airlines), airline safety will hardly ever lead to any important behaviors. For our purposes, [i.e., understanding what is important to customer satisfaction] airline safety can be ignored.

There is objective evidence to support Rust, Zahorik, and Keinigham’s (1996) assertion of the uniformity of airline safety records. The current safety ratings for the airlines in our investigation—as measured by AirlineRatings.com—are a 7 (out of a possible 7) for all but Southwest Airlines, which received a 5. Moreover, Southwest’s 5-point rating results from its not being certified by International Air Transport Association (IATA), and not from any distinguishable safety deficiencies. AirlineRatings.com awards two points of its total seven for IATA certification. Southwest Airlines, however, is not a member of IATA. Similarly, Rhoades and Waguespack (2005) calculated a safety rating for all airlines in our investigation from 1996 to 2004; the differences between the airlines under investigation were effectively indistinguishable from one another. In fact, even when including low cost carriers (e.g., Airtran, ATA, Spirit, etc.), Rhoades and Waguespack (2005, p. 348) report “that beginning in 2000 the [safety] rates for both groups [i.e. low cost USA carriers and traditional (major) USA carriers] are virtually identical.”

By contrast, Rhoades and Waguespack (2008) also provide anecdotal evidence that drops in service quality levels based upon attributes we would classify as minor incidents have led to correspondingly low industry-level satisfaction ratings. In particular they write (p. 27):

In a 2000 report entitled Why Service Stinks the airline industry had the largest drop in customer satisfaction from 1994 to 2000 of any of the reported industries (Brady, 2000). Consumers were angry about crowded planes, delays, cancellations, complex pricing models, flight restrictions, and employees that seemed to forget that airlines are a “service industry” … The most recent results of the University of Michigan's American Customer Satisfaction Index (ACSI) has airlines ranked below the International Revenue System in terms of customer satisfaction (Yu, 2007).

Furthermore, Anderson, Pearo, and Widener (2008) find that attributes which we would classify as minor incidents significantly impact customers’ overall satisfaction levels at the individual customer level. Similarly, at the aggregate (firm level), Steven, Dong, and Dresner (2012) find that minor incident attributes are linked to customer satisfaction.

It is important to note that unlike major incidents, there is not a uniformity of minor incidents (e.g., airline quality ratings (AQRs), available at http://www.airlinequalityrating.com/) across U.S. airlines. In other words, some airlines perform better/worse than their competitors. As a result, we would expect these incidents to impact choice (e.g., Ostrowski, O Brien, and Geoffrey 1993) and satisfaction (e.g., Anderson, Pearo, and Widener 2008).

Hence we hypothesize that with regard to major and minor incidents on firm-level overall customer satisfaction levels:

Hypothesis 1: Aggregate major incidents (i.e., accidents resulting in injury or death for the airlines under investigation) will not be significantly linked to customer satisfaction.

Hypothesis 2: Aggregate minor incidents (i.e., noncatastrophic service failures such as delays, lost baggage, etc.) will be significantly linked to customer satisfaction.

It is important to note that the hypotheses mentioned previously relate to the relationship between aggregate level (i.e., firm level) major and minor incidents and customer satisfaction, rather than individual consumer perceptions. At the individual level, we would expect all service failures (both major and minor) to negatively impact customer satisfaction. Yet here we are most concerned with the firm effects of these types of incidents on overall consumer satisfaction with the airline, and how these in turn affect the market share of airlines.

Satisfaction and Market Share

Both researchers and managers often assume a positive, even linear relationship between customer satisfaction and market share; as satisfaction improves, the firm ought to continuously win market share from its competitors. Research on the relationship between satisfaction and market share, however, has resulted in mixed findings. Buzzell and Gale (1987), in their seminal study of the Profit Impact of Market Strategy (PIMS) data, found that firms offering superior service had higher-than-normal market share growth. Other studies show positive links between product quality, service quality, and market share (Kordupleski, Rust, and Zahorik 1993; Parasuraman, Zeithaml, and Berry 1985; Reeves and Bednar 1994). Rust and Zahorik (1993) identified a relationship between customer satisfaction and market share. Studies by Anderson, Fornell, and Lehman (1994), Fornell (1992), Griffin and Hauser (1993), and Gronhøldt, Martensen, and Kristensen (2000), however, have all found a negative relationship between customer satisfaction and market share.

Given the conflicting findings regarding satisfaction and market share, Anderson, Fornell, and Lehman (1994, p. 59) noted, “The relationship between customer satisfaction and market share is an emerging issue in need of greater understanding.” Thus far, however, few have taken up this challenge. Two notable exceptions are studies by Gounaris et al. (2001) and Gronhøldt, Martensen, and Kristensen (2000), both of which examine the impact of customer heterogeneity on the relationship between satisfaction and market share for two consumer goods sectors in Greece. Fornell (1992) posits that customer heterogeneity and market (firm/product) heterogeneity both have a moderating effect on the relationship between customer satisfaction and market share. Additionally, Fornell et al. (2006, p. 6) suggest that “fluid markets with low switching barriers and no monopoly power” may negatively impact the relationship between customer satisfaction and market share. This implies a role of competitive intensity on the satisfaction-market share relationship. Most recently, a study by Rego, Morgan, and Fornell (2013) also found a negative relationship between satisfaction and market share across U.S. consumer markets, suggesting that as firms win market share and a larger (and more heterogeneous) group of customers, providing consumer satisfaction may become more difficult, and thus satisfaction may suffer.

In short, the relationship between consumer satisfaction and market share is a complicated one, with evidence pointing in multiple directions, and with a strong possibility that exogenous factors—competition within an industry, the extent of customer heterogeneity, and so forth—make it difficult to confidently hypothesize a relationship one way or the other.

Service Failure and Market Share

Yet despite the ambiguous relationship between customer satisfaction and market share, we argue that the relationship between service failures and market share should be expected to be negative. In a specific examination of the market share impact of service failures, Hays and Hill (1999) show that service failures increase customer defection rates resulting in market share changes. Numerous researchers also report similar findings regarding the relationship between service failures and customer defections (e.g., Ahmad 2002; Malhotra and Malhotra 2013; Smith and Bolton 1998). And as mentioned previously, the airline industry itself was punished on one specific date (i.e., September 11) for several years due to a series of particularly catastrophic service failures, resulting in an across-the-board loss of market share, if you will.

It is important to note that despite there being little differentiation among the airlines under investigation regarding major incidents (as noted earlier), making the prediction of a specific airline accident impossible, failures resulting in injury or death do occur. When they do, we would expect them to have at least a short-term impact on potential customers’ decisions to fly the focal airline suffering a major incident.

Our reasoning can be thought of as a continuation of the theme of sample bias resulting from customer defections discussed earlier. In particular, consumers who choose to fly an airline despite prior major incidents have in general, through their very behavior, demonstrated that they do not consider this potentiality relevant. Moreover, those consumers who do believe this relevant will have defected the focal airline to minimize the perceived risk of a major incident. Therefore, as noted earlier, we hypothesize that aggregate major incidents will not have a significant impact on the satisfaction of customers who continue to fly an airline. On the other hand, market share reflects purchase decisions by all customers in the category (whereas consumer satisfaction reflects the perceptions of customers of the focal airline). With regard to major incidents, consumers who believe them relevant to the selection of a particular airline will likely have discontinued the use of that airline. The same would also be expected to hold true for minor incidents that create high levels of consumer dissatisfaction (i.e., high dissatisfaction increases the likelihood of defection).

Because defectors are no longer customers, their satisfaction is not captured in “customer” surveys, as the qualification to participate in such a survey typically requires that a purchase be made within a specified time frame. Because they no longer use the focal airline, however, the airline’s future market share will be impacted.

There is some empirical evidence to support that both major and minor incidents negatively impact market share. Suzuki, Tyworth, and Novack (2001) developed and empirical tested a model of market demand which linked attributes we would classify as both major and minor incidents to market share in the airline industry.

Hence we hypothesize:

Hypothesis 3: Aggregate major incidents (i.e., accidents resulting in injury or death for the airlines under investigation) will be negatively linked to future market share.

Hypothesis 4: Aggregate minor incidents (i.e., noncatastrophic service failures such as delays, lost baggage, etc.) will be negatively linked to future market share.

Conceptual Model

Figure 1 presents our conceptual model of the hypothesized relationships between major incidents, minor incidents, customer satisfaction, and market share. As noted from Hypotheses 3 and 4, we expect both major and minor incidents to be negatively correlated with future market share. Also, we expect minor incidents to be negatively correlated with customer satisfaction (Hypothesis 2). We do not, however, expect a significant relationship between major incidents and customer satisfaction (Hypothesis 1).

Figure 1.

Structure of incidents-customer satisfaction-market share linkage.

At this juncture, the reader should note that there are lead-lag relationships for both satisfaction and market share included in our conceptual model. They occur in our model for theoretical reasons and because of data-alignment issues. These reasons are discussed in detail subsequently (see Theoretical and Empirical Reasons for Lead-Lag Relationships subsection).

The Data

The data examined in this study come from multiple sources. Data regarding the occurrence of major incidents, those incidents resulting in death or physical harm, were compiled from data collected by the National Transportation Safety Board (NTSB) accident database (http://www.ntsb.gov/aviationquery/). Data on minor incidents were collected from the Department of Transportation Air Travel Consumer Reports (http://www.dot.gov/airconsumer) and the AQRs conducted jointly by Dr. Dean Headley, professor at Wichita State University, and Dr. Brent Bowen, dean at the college of aviation at Embry-Riddle Aeronautical University (http://www.airlinequalityrating.com/). This information is collected and reported for the specific calendar month in which the incident/flight occurred. While these data come from multiple sources, for simplicity's sake we will from this point forward refer to the combined major incident and minor incident data as “NTSB data.” All original data were compiled in April 2010 covering available information from 1997 to 2009. Finally, market share data (defined here as the number of enplaned passengers) were also collected from the Department of Transportation Air Travel Consumer Reports.

The data we use as a measure of customer satisfaction with the individual airlines included in the U.S. commercial airlines industry come from the ACSI (Fornell et al. 1996). A standardized, cross-industry, annually updated index of consumer satisfaction with the goods and services consumed across the U.S. economy, the ACSI was launched in the early 1990s and currently measures satisfaction with 45 consumer-focused industries, and more than 235 companies. The ACSI satisfaction scores (or ACSI scores) are derived from survey interviews of a large, randomized sample of actual users/consumers of the companies and industries included for measurement in the study. ACSI data have been used in dozens of academic studies (in marketing and other fields) as an established, reliable indicator of consumer perceptions of their experiences with companies (Aksoy et al. 2008; Anderson, Fornell, and Mazvancheryl 2004; Fornell et al. 1996; Gruca and Rego 2005; Schneider et al. 2009), recommending its use for our study.

The ACSI scores for the airline industry and the companies included therein are constructed, like those for every company and industry measured in ACSI, by estimating partial least squares (PLS) structural equation models of raw survey data relating satisfaction to its determinants (expectations, quality, and value) and its consequences (complaint behavior and loyalty; Fornell et al. 1996). The resulting PLS latent variable index score for satisfaction is a weighted average of three observed measures (i.e., survey items) of satisfaction—overall satisfaction, confirmation/disconfirmation of expectations, and comparison to an ideal—placed on a 0 to 100 scale. The resulting index score represents a cumulative measure of overall satisfaction with a consumption experience and has been validated externally at the firm level through its ability to predict objective measures of financial performance (Aksoy et al. 2008; Fornell et al. 2006; Gruca and Rego 2005; Ittner and Larcker 1998; Schneider et al. 2009). Moreover, in the aggregate, satisfaction has been found to be a significant predictor of the macroeconomic growth within national economies (Fornell, Rust, and Dekimpe 2010).

Descriptive statistics for all of the data and variables described previously are provided in Tables 1 and 2.

Table 1.

Descriptive Statistics for Observed Variables.^a

	N	Min	Max	Mean	SD
Complaints, per 100,000; MI1	91	0.140	5.300	1.393	0.917
Average arrival delay, minutes; MI2	91	1.180	18.480	6.844	3.204
Percent flights cancelled, MI3	91	0.004	0.084	0.019	0.013
Percent flights diverted, MI4	91	0.001	0.004	0.002	0.001
Total number lost luggage, MI5	91	89,886	612,347	298,724	128,589
Number lost luggage per passenger, MI6	91	0.002	0.011	0.004	0.001
Total oversales, voluntary; MI7	91	36,049	261,275	97,013	48,718
Load factor, MI8	89	0.637	0.848	0.750	0.049
Number of accidents, less 9/11; MA1	91	0.000	13.000	4.813	2.940
Number of injuries, MA2	91	0.000	200.000	12.714	27.522
Number of fatalities, less 9/11; MA3	91	0.000	265.000	3.912	28.265
ΔMarket share, as % passenger rev.; MS1	82	−0.095	0.479	0.006	0.072
ΔMarket share, as % total flights; MS2	82	−0.312	0.217	−0.046	0.089
ACSI	91	53.000	81.000	64.626	5.417

Note. MA = major incidents; MI = minor incidents; MS = market share. ^aThe slightly smaller sample for Load Factor (MI8) is explained by two missing observations in the National Transportation Safety Board (NTSB) data set. The smaller samples for the two market share variables is due to data lost through the first-differencing and lagging processes.

Table 2.

Zero-Order Correlations Among the Observed Variables.

	MI1	MI2	MI3	MI4	MI5	MI6	MI7	MI8	MA1	MA2	MA3	MS1	MS2	ACSI
MI1	1.000	0.375**	0.695**	0.390**	0.091	0.276**	0.226*	0.048	0.185	0.093	0.124	−0.141	−0.399**	−0.577**
MI2	0.375**	1.000	0.351**	0.445**	0.382**	0.430**	0.000	0.234*	0.130	0.144	−0.042	0.034	−0.053	−0.160
MI3	0.695**	0.351**	1.000	0.408**	0.236*	0.242*	0.379**	−0.243*	0.182	0.094	0.188	−0.177	−0.437**	−0.339**
MI4	0.390**	0.445**	0.408**	1.000	−0.015	−0.004	0.029	0.172	0.184	0.194	0.095	0.009	−0.106	−0.302**
MI5	0.091	0.382**	0.236*	−0.015	1.000	0.667**	0.381**	−0.090	0.301**	0.137	−0.045	0.114	−0.007	0.151
MI6	0.276**	0.430**	0.242*	−0.004	0.667**	1.000	0.007	0.121	0.067	−0.003	−0.085	0.136	0.026	−0.122
MI7	0.226*	0.000	0.379**	0.029	0.381**	0.007	1.000	−0.427**	0.443**	0.366**	0.056	−0.100	−0.082	0.042
MI8	0.048	0.234*	−0.243*	0.172	−0.090	0.121	−0.427**	1.000	−0.015	−0.157	−0.106	−0.050	0.104	−0.469**
MA1	0.185	0.13	0.182	0.184	0.301**	0.067	0.443**	−0.015	1.000	0.449**	0.154	−0.069	−0.134	−0.114
MA2	0.093	0.144	0.094	0.194	0.137	−0.003	0.366**	−0.157	0.449**	1.000	−0.006	−0.038	0.116	0.079
MA3	0.124	−0.042	0.188	0.095	−0.045	−0.085	0.056	−0.106	0.154	−0.006	1.000	−0.066	−0.243*	−0.039
MS1	−0.141	0.034	−0.177	0.009	0.114	0.136	−0.100	−0.050	−0.069	−0.038	−0.066	1.000	0.336**	0.234*
MS2	−0.399**	−0.053	−0.437**	−0.106	−0.007	0.026	−0.082	0.104	−0.134	0.116	−0.243*	0.336**	1.000	0.339**
ACSI	−0.577**	−0.160	−0.339**	−0.302**	0.151	−0.122	0.042	−0.469**	−0.114	0.079	−0.039	0.234*	0.339**	1.000

Note. ACSI = American Customer Satisfaction Index; MA = major incidents; MI = minor incidents; MS = market share.

*p < .05; **p < .01.

Theoretical and Empirical Reasons for Lead-Lag Relationships

The literature argues for the following chain of effects, as they relate to consumer experiences and business outcomes: incidents (major and minor) impact the consumer, which in turn impact the consumer's satisfaction, which, finally, impact business outcomes (e.g., market share). Not all of these events, however, occur simultaneously. Clearly, an incident must first occur before it can impact consumer satisfaction at some point in the future, though in the abstract a consumer may begin to become dissatisfied very shortly after a service failure occurs. Thus, while there is some lead-lag in this relationship, the time between these events may be quite small. On the other hand, satisfaction is a leading indicator of financial and accounting-based measures of performance, such as productivity, sales, and net income (Anderson and Fornell 2000). Anderson, Fornell, and Lehmann (1994, p. 61) observe “a change in customer satisfaction is not reflected all at once in returns. Rather, a percentage point change in customer satisfaction in one period carries over to future periods, consistent with the cumulative nature of customer satisfaction.” We would likewise expect a similar outcome from service quality failure incidents and business outcomes (Kordupleski, Rust, and Zahorik 1993).

The lead-lag relationships between satisfaction and/or service quality and business outcomes have been demonstrated by many researchers (e.g., Aksoy et al. 2008; Anderson, Fornell, and Mazvancheryl 2004; Fornell et al. 2006; Gruca and Rego 2005; Kordupleski, Rust, and Zahorik 1993). The appropriate length of the time lag, however, is less certain. The relationship is in large part dependent upon the frequency of purchase/usage, as the impact of satisfaction on business outcomes is felt through the consumer’s future purchase decisions. The length of the lead-lag relationship therefore relies on the purchase cycle, which not only varies across product and service categories, but in the airline industry varies greatly across individual travelers. For instance, ABC News reported that “more than 85 percent of the public flies infrequently if at all. More than seven in ten can be described as infrequent fliers, flying once or twice a year or less; and 14 percent have never flown” (Sussman 2001). Gallup (2013) reports similar findings regarding the infrequency of air travel for most Americans; specifically, they find that in 2012, 48% of Americans did not fly at all, 27% flew 1 or 2 times, and 25% flew 3 or more times. As a result, we would anticipate that significant time might pass before airlines experience the maximum effect of both major and minor incidents and customer satisfaction on market share.

Taking all of this lack of theoretical or empirical certainty into account, the choice of the appropriate market share lag to use when analyzing our statistical model was done empirically. We tested lags at t + 1 and t + 2 (i.e., t + 1 = focal year plus one year, and t + 2 = focal year plus two years). Our results indicate that t + 2 performs best in our models. Given that most Americans fly infrequently, this lag appears both reasonable and defensible. It is important to note, however, that this finding may in part be a function of the data alignment issues regarding the different metrics. In particular, the ACSI _t ₊₁ data overlaps the market share _t ₊₁ data which could have impacted the choice of market share lags. We discuss these issues in more detail next.

Alignment of Data

One complication with our use of ACSI scores as a proxy of aggregate consumer satisfaction in the airline industry is the timing of the data collection and resulting data, and aligning this data with the data on aggregate major incidents, minor incidents, and market share (i.e., the NTSB data). That is, while the NTSB data are collected on a precisely defined annual basis, with the “reported incidents” variables discussed previously referring exclusively to a single calendar year, the ACSI data is less timing precise. The data for the ACSI's airline industry study are collected during the first quarter of each year (with interviewing completed from January through March), and potential respondents are screened prior to interviewing and asked if they have flown on a commercial airline “in the past twelve months.” During any given year, the final sample of respondents might, therefore, have flown on an airline at any time between early January of the previous year through late March of the year of interviewing, for example, for the 1997 ACSI airlines study, respondents might have flown as early as January 1, 1996, and as late as March 31, 1997, depending on the specific opening and closing dates of interviewing.

Figure 2 shows the potential overlaps in time for the ACSI in times t and t + 1 and the NTSB at time t. The reader will see that the potential overlap in terms of months is greater for ACSI _t ₊₁ and NTSB _t . Only the first 3 months of the ACSI _t and NTSB _t actually overlap. We also see that the data collection period for ACSI _t ₊₁ overlaps the data collection period for market share _t ₊₁. As we would expect the impact of customer satisfaction on market share to be visible as a lead-lag relationship (i.e., one that occurs in the future), this overlap in data may impact the strength of various market share lags.

Figure 2.

Data alignment between NTSB data (i.e., major and minor incidents), American Customer Satisfaction Index (ACSI) data (i.e., customer satisfaction), and market share data.

While little can be done to perfectly align these two distinct samples of data retroactively, this complexity does require consideration during modeling. Given that the time frame for ACSI _t ₊₁ most closely corresponds to the NTSB _t data, we use these time frames in our model. Additionally, the imperfect data alignment would be expected to in part impact the selection of the appropriate market share lag. As noted earlier, this was done empirically, with market share _t ₊₂ found to perform best in our analyses.

Statistical Analysis and Results

Statistical Methods

To analyze the data and operationalize the conceptual model presented in Figure 1, linking aggregate major and minor incidents in the airline industry, customer satisfaction, and market share, we estimate a partial least squares path model (PLS-PM).³ The PLS-PM approach was chosen for several reasons, as PLS-PM is recommended for addressing several issues present in our data, allowing us to model the multiple and dynamic structural relationships between the variables (themselves multivariate, latent components) while addressing these issues. First, PLS-PM is sometimes called a “soft modeling” approach because of its greater flexibility relative to competing techniques, and particularly when the theory guiding the analysis is less well defined, as is the case here, and the distribution of the variables under investigation is uncertain (Chin 1998; Wold 1980). Second, PLS-PM is particularly useful when the assumption of a multivariate normal distribution is violated, sample sizes are relatively small, and there is an emphasis on prediction of outcome variables rather than comparison of observed and best-fitting models, all features indicative of our study (Chin 2010; Manuel, Almeida, and Coelho 2010). Finally, PLS-PM has been used extensively in marketing studies of this kind and is particularly predominant in the customer satisfaction literature (Fornell 1992; Fornell et al. 1996; Henseler, Ringle, and Sinkovics 2009; Hulland, Ryan, and Rayner 2010).

In brief, and like most SEM techniques, PLS-PM is an iterative approach for estimating complex measurement (outer) and structural (inner) latent variable models. The PLS-PM algorithm does not attempt to maximize the covariances across an entire data matrix with the goal of identifying structurally invariant parameters, but rather estimates (and iteratively reestimates) weights and loadings within the specified latent variables until an optimal prediction of the outcome variable is achieved; PLS-PM is aimed at best-fitting path coefficients, rather than a best-fitting covariance matrix (Vilares et al 2010; Vinzi, Trinchera, and Amato 2010). In addition, as we will discuss subsequently, PLS-PM permits the specification of formative latent variables, where the manifest variables are thought to be causes of the latent variable rather than the other way around. Thus, we propose to use these PLS-PM methods to test the conceptual model linking the ACSI, NTSB, and market share data discussed previously.

Furthermore, it is important to note that the PLS-PM we propose to test includes two types of latent variables, both formative and reflective latent constructs.⁴ The distinction between these two types of indicators is critical. Reflective indicators (such as the market share [MS] latent variable) are related (in a theoretical sense) as unobservable predictors of the underlying observed variables that comprise them, and are expected to be empirically unidimensional. Because there are strong reasons (both conceptual and empirical) to suggest that the two measures of market share examined in this study—change in market share as percentage of total flights, and change in market share as a percentage of passengers—are strongly and positively correlated, and are in fact predicted by the latent construct of market share, modeling these variables reflectively is appropriate.

Formative indicators, on the other hand, impose fewer restrictions and expectations on the observed variables, and do not assume that the block of manifest variables included in the latent variable to be unidimensional or homogenous. Contrary to the market share latent variable, there is less reason to believe that the several variables included in either the major or minor incident latent variables below to be unidimensional, and because we have “no specific expectations about patterns or magnitude of intercorrelations between the indicators,” we model both as formative indicators (Diamantopoulos, Riefler, and Roth 2008, p. 1205). In a practical sense, the most important interpretational consideration in differentiating these two types of latent variables is that with our reflective indicator, examination of the factor loadings derived from the measurement model is most appropriate, while examination of the weights of the two formative latent variables is most relevant (Chin 2010; Harris 1989).

Variables Defined

As noted earlier, our analysis creates latent variables to determine the relationship between major incidents (MA), minor incidents (MI), MS, and customer satisfaction (CS). For clarity, they can be defined as follows:

Major incidents:

Accidents resulting in injury or death. In the case of this analysis, a latent variable was constructed to represent MA based upon three variables: (1) number of accidents, (2) number of injuries, and (3) number of fatalities.

Minor incidents:

Noncatastrophic service failures (e.g., delays, lost baggage, etc.). In the case of this analysis, a latent variable was constructed to represent MI based upon eight variables: (1) complaints, (2) average arrival delay in minutes, (3) percent flights cancelled, (4) percent flights diverted, (5) total number of lost luggage, (6) number of lost luggage per passenger, (7) total oversales (voluntary), and (8) load factor.

It is important to note that the above-mentioned classification of Major Incidents and Minor Incidents reflects our distinction of severe service failures as those incidents that result in injury or death. As such, it is an objective criterion, not a customer-perceived measure of severity as is typically used in most of the literature on service failure and recovery. Indeed, some failures which we would classify as Minor Incidents would likely be perceived by customers as terrible experiences (e.g., a flight delay that resulted in missing an important event). Nonetheless, we are confident that the overwhelming majority of airline customers would consider a flight resulting in injury or death to be more severe than any noncatastrophic service failure; this also holds true for airline managers.

Market Share:

The portion of the market accounted for by a specific airline. In the case of this analysis, a latent variable was constructed to represent MS based upon two variables: (1) percent of passenger revenue and (2) percent of total flights.

Customer satisfaction:

The overall (cumulative) judgment that the focal airline provided a “pleasurable level of consumption related fulfillment, including levels of under- or overfulfillment” (Oliver 2010, p. 8). In the case of this analysis, customer satisfaction is represented by the ACSI.

Measurement Model Results

Based on the techniques described previously, we turn now to a discussion of the measurement model results for the specified PLS-PM outlined in Figure 3. These results are presented in Tables 3 and 4.

Figure 3.

Partial least squares path model coefficients.

Table 3.

Measurement Model Statistics.

LV	MV	LV TYPE	LOAD	WT
MI	MI1	F	0.749	0.228
	MI2		0.398	−0.021
	MI3		0.669	0.670
	MI4		0.402	−0.121
	MI5		0.031	−0.354
	MI6		0.329	0.209
	MI7		0.094	0.224
	MI8		0.511	0.703
MA	MA1	F	0.727	0.908
	MA2		−0.140	−0.541
	MA3		0.593	0.444
MS	MS1	R	0.641	0.367
	MS2		0.938	0.815
ACSI _t	ACSI _t		1.000	1.000
ACSI _t ₊ ₁	ACSI _t ₊ ₁		1.000	1.000

Note. ACSI = American Customer Satisfaction Index; LV = latent variables; MV = manifest variable; LV type: F = formative, R = reflective; LOAD = PLS-PM loadings; MI = minor incidents; MA = major incidents; MS = market share; WT = standardized weights.

Beginning with a discussion of the one reflective latent variable in our model, the ultimate dependent latent variable MS (market share), we observe a loading for market share as percentage change in passenger revenue (MS1) of λ = .641, and a loading for market share as percentage change in total flights (MS2) of λ = .938. While the loading for MS1 is slightly below the often-prescribed threshold of 0.707 for acceptable standard loadings in PLS-PM (Chin 2010), for theoretical reasons, given that smaller loadings of λ = .50 have been argued to be acceptable in some instances (Barclay, Higgins, and Thompson 1995; Chin 1998), and because the average variance extracted (AVE) statistic indicates an acceptably-fitting component (AVE = .65), we will maintain the model in its current form (Fornell and Larcker 1981). Regardless, these results indicate that the market share latent component is more strongly predictive of market share as percentage change in total flights, with a stronger relationship between this variable and the component than between it and market share as percentage change in passenger revenue.

Turning to the two formative constructs specified in the model, the two primary independent exogenous variables in our model, and beginning with the minor incidents formative construct (MI), the results for this component (and again, focusing now on standardized weights rather than loadings) show that two variables, MI3 (percentage of flights cancelled) and MI8 (load factor), are most strongly predictive of the extracted latent variable.⁵ With the largest weight among the variables of w = .703, MI8 is the strongest predictor of the minor incidents latent component, followed closely by MI3 (w = .670). The remaining variables are less strongly related, with a few exhibiting small negative coefficients.⁶ But as suggested by the simple correlations in Table 2, and confirmed by examining variance inflation factors for the MI latent variable (not reported here, but none greater than VIF = 3.00), strong collinearity that might make the estimates within the formative construct unstable is not a significant problem in this instance (Hair, Anderson, and Tatham 1992).

The second formative construct, major incidents (MA), shows that one variable (MA1, number of accidents) dominates as a predictor of major incidents (w = .908), while a second variable (MA3, number of fatalities) also has a strong positive weight (w = .444). Interestingly, MA2 (number of injuries) also shows a reasonably strong weight as a predictor of the construct, but the weight is negative (w = −0.541), suggesting that number of injuries has a negative effect on the extracted latent variable. Nevertheless, the strong relationship between number of accidents and major incidents (and, by extension, between major incidents, customer satisfaction, and market share) is the most noteworthy result.

Finally, because the lagged customer satisfaction (ACSI) included in the model reflects a single observed variable, analysis of measurement model statistics does not apply in this instance.

Structural Model Results

Focusing now on interpretation of the structural model, the path coefficients represented graphically in Figure 3 confirm most of our expectations and hypotheses. (Table 4 provides interconstruct correlations among the latent variables.) Beginning with the major incidents (MA) latent variable, and the hypothesized nonsignificant relationship between MA and future customer satisfaction (ACSI _t ₊₁), the coefficient is small (−0.04; p > .05) and as expected insignificant.⁷ Major incidents—and most significantly, given the results of the measurement model for this formative construct, number of accidents, with a lesser effect from number of fatalities—do not significantly impact future customer satisfaction with an airline. MA does, on the other hand, have a significant negative effect on future market share (MS), and in this case the coefficient is both significant and in the expected direction (−0.22; p < .05). Therefore, and again reflecting back on the measurement model results, major incidents as number of accidents and fatalities (primarily) can be seen to have a significant negative effect on future market share for an airline.

Table 4.

Inter-Construct Correlations.^a

	MI	MA	ACSI _t	ACSI _t ₊ ₁	MS_t+2
MI	0.462	0.209	−0.761	−0.831	−0.332
MA	0.209	0.547	−0.270	−0.211	−0.264
ACSI _t	−0.761	−0.270	1.000	0.753	0.362
ACSI _t ₊ ₁	−0.831	−0.211	0.753	1.000	0.192
MS_t+2	−0.332	−0.264	0.362	0.192	0.802

Note. ACSI = American Customer Satisfaction Index; MI = minor incidents; MA = major incidents; MS = market share. ^aThe diagonal values represent the square roots of average variance extracted (AVE) values, with the exception of the two observed ACSI variables. The off-diagonal values represent interconstruct correlations.

Regarding the minor incidents (MI) construct, we observe significant effects between MI and both future customer satisfaction (−0.82, p < .001) and future market share (−0.54, p < .05). MI shows a substantially stronger effect on satisfaction than major incidents, underscoring the importance of this predictor in influencing consumer's satisfaction in the future. In fact, MI explains almost all of the 69% explained variance in ACSI (about 99% of the explained variance is explained by MI). MI also has the strongest effect on future market share (−0.54), and this coefficient is also significant (p < .05) and in the expected direction. Together, the three predictors of market share (major and minor incidents, and customer satisfaction) explain about 18% of the variance in market share (R ² = .18).⁸

Finally, while reasonably large (−.30), the coefficient from customer satisfaction (ACSI) to future market share is insignificant, with a p value (p > .05) too large to declare significance under the traditional thresholds. Given the strong observed relationship between the MI and ACSI variables, the strong relationship between MI and MS, and the fact that the ACSI variable has a significant simple correlation with both observed changes in market share variables (see Table 2), it would appear that the effect of ACSI on MS is being suppressed by the MI latent variable.⁹

General Discussion and Conclusions

The analysis reported here advances the theoretical and empirical research regarding the relationship between services failures, customer satisfaction, and market share in several ways. First, our findings indicate that at the firm level, service failures do not appear to follow a simple pattern with regard to their impact on consumer satisfaction and market share. Despite the generally accepted view among managers and researchers that the greater the severity of the service failure, the greater the resulting impact on customer satisfaction and business outcomes (e.g., Smith, Bolton, and Wagner 1999; Weun, Beatty, and Jones 2004), our findings indicate that product-harm crises do not have the same impact on consumers’ perceptions or behavior in the airline industry. Interestingly and importantly, major incidents for airlines—accidents, injuries and fatalities—demonstrate a weaker relationship to future market share than do minor incidents, like flight cancellations and airline load factor. Moreover, major incidents showed no significant relationship with consumer satisfaction, while minor incidents were strongly and negatively related to future consumer satisfaction, as anticipated. Finally, as in some previous research, we find a negative (but insignificant) relationship between customer satisfaction and market share.

We conjecture that these findings result from several causes. First, consumers who believe major incidents to be relevant in their decision to fly an airline would be prone to defect from that airline; assuming competition exists, few rational consumers would continue to give their business to an airline they believed would cause them harm. As a result, these consumers' perceptions would not be gauged in customer satisfaction surveys, as they no longer fly the airline. Second, customers rightly believe major incidents to be highly unlikely occurrences, and therefore they do not appear as salient issues to them. Without question, commercial air travel is very safe relative to other modes of transportation (Mouawad and Drew 2013), and therefore customers are more likely to focus on those issues that they may experience in their routine travel.

It is important to note here that we believe the weaker impact of major incidents when compared to minor incidents vis-à-vis service failures in the airline industry in large part reflects the fact that major incidents are extremely unlikely events. Indeed, according to one estimate, “the annual risk of being killed in a plane crash for the average American is about one in 11 million,” making it about 16 times less likely than being struck by lightning in any given year (Ropeik 2011). Were major incidents higher probability events, it seems likely that they would have a much stronger impact on market share and customer satisfaction. Similarly, were the probability of experiencing a major incident and a minor incident even equal, it is difficult to imagine a scenario where major incidents would not have a stronger impact on future market share. The reality, however, is that virtually any firm with a high probability of product-harm crises would cease to exist.¹⁰ Therefore, by their very nature, they must be very low likelihood occurrences. Instead, we are suggesting that in an environment where consumers view major incidents as being extraordinarily unlikely, their rare occurrence (if managed properly) will not hold the same sway as minor incidents with relatively high probabilities of occurrence.

This does not mean that major incidents have no impact on consumers. One of the most important findings of our study is that major incidents negatively impact future market share but not consumer satisfaction. We believe this discrepancy to be very telling. If market share declines but consumer satisfaction remains the same (or increases), this may signal that dissatisfied consumers are defecting due to their concerns about major incidents, and as a result their voices are not being heard in customer satisfaction surveys. This phenomenon has been observed in the past, with firms bleeding customers, rapidly losing revenue, and even nearing bankruptcy, but actually experiencing an increase in customer satisfaction, precisely because the (much) fewer cohort of remaining customers are the most satisfied and most staunchly loyal (Fornell et al. 2005). As a result, managers can easily delude themselves into believing that things have returned to normal after a major incident because satisfaction levels are at or near pre-major incident levels. Therefore, managers must not rely exclusively on consumer satisfaction levels post any major incident; they also need to determine the impact on customer defection levels, and where possible identify lost customers in an attempt to understand what can be done to win them back.

Our findings also point to the difficulty in linking consumer satisfaction to market share, reinforcing the mixed results from earlier studies. Our results show no significant impact of satisfaction on future market share. This may indicate that customer satisfaction takes more time to impact airlines in terms of future consumer behaviors (and market share). It may also speak to the somewhat restricted ability of consumers to switch suppliers in the airline industry, where the major carriers control hubs and can effectively limit competition from smaller airlines. Finally, this may also indicate a problem of self-selection; specifically, customers who choose to defect are not measured in customer satisfaction surveys, therefore there is the potential for satisfaction score inflation as only the most satisfied customers remain.

This research, and a growing body of additional works, has shown that the commonly held belief by managers that higher consumer satisfaction levels lead linearly to higher market share is a more complicated issue (Anderson, Fornell, and Lehman 1994; Fornell 1992; Griffin and Hauser 1993; Gronhøldt, Martensen, and Kristensen 2000; Rego, Morgan, and Fornell 2013; Rust and Zahorik 1993). One strong driver of both consumer satisfaction and market share, however, may play a critical role in keeping both aligned, specifically preventing major and minor incidents whenever and wherever possible. While this may seem obvious, the need for vigilance in operational excellence cannot be overstated.

Limitations and Future Research

As with all scientific research, there are limitations to this study that need to be noted. In particular, this investigation has examined the service failure (major and minor)-customer satisfaction-market share relationship for only a single industry, and an industry (commercial airlines) where the results may lack generalizability to other industries. Because of both limitations in free market competition in this industry and the somewhat unique nature of the major incidents that plague the industry (i.e., low-probability, but very high profile), it is not unreasonable to suspect that consumers may behave differently under these circumstances toward this industry than toward others. Therefore, future research should aim to replicate these findings within other consumer industries. In our estimation, applying this type of analysis to an industry with both greater competition and a more regular occurrence of major incidents—such as the automobile industry—would be a useful exercise in helping to clarify the independent effect of these service failures on satisfaction, and likewise on future market share.

Furthermore, because of limitations to the existing theory defining the impact of major and minor incidents on customer satisfaction and market share, a situation discussed at the beginning of this study, future research building upon this nexus and refining the conceptual model we have proposed would be helpful. Combined with the previous recommendation, therefore, tests of industries in addition to airlines to refine the model we have tested in this study would, in our opinion, be of greatest value.

Finally, research expanding on these findings by extending them to consumers in other nations and cultures is needed. Substantial prior research in marketing and consumer behavior has regularly discovered that consumers from distinct cultures display behaviors substantially different from one another. Indeed, a variety of recent studies have found that consumers across different nations are “satisfied differently” and that this is due in large part to culture (e.g., Morgeson et al. 2011). Indeed, one prominent cross-cultural difference lies in risk averseness (sometimes called “uncertainty avoidance;” Hofstede 2001). As such, examining how consumers who are culturally predisposed to address risk differently respond to service failures of various kinds, and how this impacts future satisfaction and business outcomes, is an important next step, particularly as corporations continue to seek revenue growth by expanding into new national markets.

These limitations notwithstanding, the results in this study present important new insights into the nature of services failures as both major and minor incidents, and their resulting impact on consumer satisfaction and market share.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

References

Aksoy

Lerzan

Cooil

Bruce

Groening

Christopher

Keiningham

Timothy L.

Yalcin

Atakan

(2008), “Long Term Stock market Valuation of Customer Satisfaction,” Journal of Marketing, 72 (4), 105–122.

Ahmad

Sohel

(2002), “Service Failures and Customer Defection: A Closer Look at Online Shopping Experiences,” Managing Service Quality, 12 (1), 19–29.

Anderson

Eugene W.

Fornell

Claes

(2000), “The Customer Satisfaction Index as a Leading Indicator,” in Handbook of Service Marketing & Management, Swartz

Teresa

Iacobucci

Dawn

, eds. Thousand Oaks, CA: Sage, 255–268.

Anderson

Eugene W.

Fornell

Claes

Lehmann

Donald R.

(1994), “Customer Satisfaction, Market Share, and Profitability: Findings from Sweden,” Journal of Marketing, 58 (3), 53–66.

Anderson

Eugene W.

Fornell

Claes

Mazvancheryl

Sanal K.

(2004), “Customer Satisfaction and Shareholder Value,” Journal of Marketing, 68 (4), 172–185.

Anderson

Shannon W.

Scott Baggett

Widener

Sally K.

(2009), “The Impact of Service Operations Failures on Customer Satisfaction: Evidence on how Failures and Their Source Affect what Matters to Customers,” Manufacturing & Service Operations Management, 11 (1), 52–69.

Anderson

Shannon

Pearo

Lisa Klein

Widener

Sally K.

(2008), “Drivers of Service Satisfaction Linking Customer Satisfaction to the Service Concept and Customer Characteristics,” Journal of Service Research, 10 (4), 365–381.

Bamford

David

Xystouri

Tatiana

(2005), “A Case Study of Service Failure and Recovery within an International Airline,” Managing Service Quality, 15 (3), 306–322.

Barclay

Donald

Higgins

Christopher

Thompson

Ronald

(1995), “The Partial Least Squares (PLS) Approach to Causal Modeling: Personal Computer Adoption and Use as an Illustration,” Technology Studies, 2 (2), 285–309.

10.

Block

Melissa

(2009), “Cause of Buffalo, N.Y., Crash Being Probed,” NPR All Things Considered [Radio Broadcast]. National Public Radio. February 13, 2009. (accessed March 06, 2011), [available at HighBeam Research: http://www.highbeam.com/doc/1P1-160981244.html].

11.

Brady

Diane

(2000), “Why Service Stinks,” Businessweek. (October 23), 118–128.

12.

Brown

Stephen W.

Cowles

Deborah L.

Tuten

Tracy L.

(1996), “Service Recovery: Its Value and Limitations as a Retail Strategy,” International Journal of Service Industry Management, 7 (5), 32–46.

13.

Buzzell

Robert D.

Gale

Bradley T.

(1987), The PIMS Principles: Linking Strategy to Performance. New York, NY: The Free Press.

14.

Chevalier

Judith A.

Mayzlin

Dina

(2006), “The Effect of Word of Mouth on Sales: Online Book Reviews,” Journal of Marketing Research, 43 (3), 345–359.

15.

Chin

Wynne W.

(2010), “How to Write Up and Report PLS Analyses,” in Handbook of Partial Least Squares, Esposito Vinzi

Chin

Wynne W.

Henseler

Jörg

Wang

Huiwen

, eds. Heidelberg, Germany: Springer.

16.

Chin

Wynne W.

(1998), “The Partial Least Squares Approach for Structural Equation Modeling,” in Modern Methods for Business Research, Marcoulides

George A.

ed. Mahwah, NJ: Lawrence Erlbaum.

17.

Cleeren

Kathleen

Dekimpe

Marnik G.

Helsen

Kristiaan

(2008), “Weathering Product-Harm Crises,” Journal of the Academy of Marketing Science, 36 (2), 262–270.

18.

Dawar

Niraj

Pillutla

Madan M.

(2000), “Impact of Product-Harm Crises on Brand Equity: The Moderating Role of Consumer Expectations,” Journal of Marketing Research, 37 (2), 215–226.

19.

Diamantopoulos

Adamantios

Riefler

Petra

Roth

Katharina P.

(2008), “Advancing Formative Measurement Models,” Journal of Business Research, 61, (12), 1203–1218.

20.

Duffy

Jo Ann M.

Miller

John M.

Bexley

James B.

(2006), “Banking Customers’ Varied Reactions to Service Recovery Strategies,” The International Journal of Bank Marketing, 24 (2), 112–132.

21.

Fornell

Claes

(1992), “A National customer satisfaction Barometer: The Swedish Experience,” Journal of Marketing, 56 (1), 6–22.

22.

Fornell

Claes

Larcker

David F.

(1981), “Evaluating structural equation models with unobservable variables and measurement error,” Journal of Marketing Research. 18 (1), 39–50.

23.

Fornell

Claes

Rust

Roland T.

Dekimpe

Marnik G.

(2010), “The Effect of Customer Satisfaction on Consumer Spending Growth,” Journal of Marketing Research, 47 (1), 28–35.

24.

Fornell

Claes

Mithas

Sunil

Morgenson III

Forest V.

Krishan

M. S.

(2006), “Customer Satisfaction and Stock Prices: High Returns, Low Risk,” Journal of Marketing, 70 (1), 1–14.

25.

Fornell

Claes

Amburg

David Van

Morgeson III

Forrest V.

Bryant

Barbara

Anderson

Eugene

Johnson

Michael

(2005), The American Customer Satisfaction Index at 10 Years. Ann Arbor, MI: Stephen M. Ross School of Business.

26.

Fornell

Claes

Johnson

Michael D.

Anderson

Eugene W.

Cha

Jaesung

Bryant

Barbara Everitt

(1996), “The American Customer Satisfaction Index: Nature, Purpose and Findings,” Journal of Marketing, 60, (4), 7–18.

27.

Gallup (2013), “Poll: Airlines.” (accessed March 17, 2013), [available at http://www.gallup.com/poll/1579/Airlines.aspx].

28.

Gilbert

David

Wong

Robin K. C.

(2003), “Passenger Expectations and Airline Services: A Hong Kong Based Study,” Tourism Management, 24 (5), 519–532.

29.

Götz

Oliver

Liehr-Gobbers

Kerstin

Krafft

Manfred

(2010), “Evaluation of Structural Equation Models Using the Partial Least Squares (PLS) Approach,” in Handbook of Partial Least Squares, Esposito Vinzi

Chin

Wynne W.

Henseler

Jörg

Wang

Huiwen

, eds.Heidelberg, Germany: Springer.

30.

Gounaris

Spiros P.

Avlonitis

George J.

Kouremenos

Athanassios

Papavassiliou

Nikolaos

Papastathopoulou

Paullina

(2001), “Customer Satisfaction and Market Share: What Is the Missing Link,” Journal of Euromarketing, 10 (4), 61–81.

31.

Griffin

Abbie

Hauser

John R.

(1993), “The Voice of the Customer,” Marketing Science, 12 (1), 1–27.

32.

Gronhøldt

Lars

Martensen

Anne

Kristensen

Kai

(2000), “The Relationship between Customer Satisfaction and Loyalty: Cross-Industry Differences,” Total Quality Management, 11 (4/5 & 6), S509–S514.

33.

Gruca

Thomas S.

Rego

Lopo L.

(2005). “Customer Satisfaction, Cash Flow, and Shareholder Value,” Journal of Marketing, 69 (3), 115–130.

34.

Hair

Joseph F.

Jr Anderson

Rolph E.

Tatham

Ronald L.

(1992), Multivariate Data Analysis. New York, NY: MacMillan Publishing Company.

35.

Hauser

Robert M.

Goldberger

Arthur S.

(1971), “The Treatment of Unobservable Variables in Path Analysis,” in Sociological Methodology, Costner

Herbert L.

ed. London, UK: Jossey-Bass.

36.

Hays

Julie M.

Hill

Arthur V.

(1999), “The Market Share Impact of Service Failures,” Production and Operations Management, 8 (3), 208–220.

37.

Harris

Richard J.

(1989), “A Canonical Cautionary,” Multivariate Behavioral Research, 24 (1), 17–39.

38.

Hedgpeth

Dana

(2010), “Safety Agency Seeks Toyota documents; Recalls’ Speed in Question, Move Comes in Advance of Hearings in Congress,” The Washington Post. (February 17), p. A08, (accessed March 06, 2011), [available at http://www.washingtonpost.com/wp-dyn/content/article/2010/02/16/AR2010021605572_pf.html].

39.

Henseler

Jörg

Ringle

Christian

Sinkovics

Rudolf

(2009). “The Use of Partial Least Squares Path Modeling in International Marketing,” Advances in International Marketing, 20 277–320.

40.

Hess

Ronald L.

Ganesan

Shankar

Klein

Noreen M.

(2003), “Service Failure and Recovery: The Impact of Relationship Factors on Customer Satisfaction,” Journal of the Academy of Marketing Science, 31 (2), 127–145.

41.

Hofstede

Geert H.

(2001). Culture’s Consequences: Comparing Values, Behaviors, Institutions and Organizations Across Nations. Thousand Oaks, CA: Sage.

42.

Hsin-Hui

Lin

Yi-Shun

Wang

Li-Kuan

Chang

(2011), “Consumer Responses to Online Retailer's Service Recovery After a Service Failure,” Managing Service Quality, 21 (5), 511–534.

43.

Hulland

Ryan

M. J.

Rayner

R. K.

(2010), “Modeling Customer Satisfaction: A Comparative Performance Evaluation of Covariance Structure Analysis versus Partial Least Squares,” in Handbook of Partial Least Squares, Esposito Vinzi

Chin

Wynne W.

Henseler

Jörg

Wang

Huiwen

, eds. Heidelberg, Germany: Springer.

44.

Huey

Raymond B.

Salisbury

Richard

Wang

Jane-Ling

Mao

Meng

(2007), “Effects of Age and Gender on Success and Death of Mountaineers on Mount Everest,” Biology Letters, 3 (5), 498–500.

45.

Ittner

Christopher D.

Larcker

David

(1998), “Are Nonfinancial Measures Leading Indicators of Financial Performance? An Analysis of Customer Satisfaction,” Journal of Accounting Research, 36 (Supplement: Studies on Enhancing the Financial Reporting Model), 1–35.

46.

Kahneman

Daniel

Tversky

Amos

(1979) “Prospect Theory: An Analysis of Decisions Under Risk,” Econometrica, 47 (2), 263–291.

47.

Keiningham

Timothy L.

Gupta

Sunil

Aksoy

Lerzan

Buoye

Alexander

(2014), “The High Price of Satisfaction,” MIT Sloan Management Review, 55 (3), 37–46.

48.

Kordupleski

Ray

Rust

Roland T.

Zahorik

Anthony J.

(1993), “Why Improving Quality Doesn’t Improve Quality,” California Management Review, 35 (3), 82–95.

49.

Kumar

Vikas

Luciano

Batista,

Maull

Roger

(2011), “The Impact of Operations Performance on Customer Loyalty,” Service Science, 3 (2), 158–171.

50.

Lapré

Michael A.

(2011), “Reducing Customer Dissatisfaction: How Important is Learning to Reduce Service Failure?” Production and Operations Management, 20 (4), 491–507.

51.

Luo

Xueming

Homburg

Christian

(2008), “Satisfaction, Complaint, and the Stock Value Gap,” Journal of Marketing, 72 (4), 29–43.

52.

Luo

Xueming

(2007), “Consumer Negative Voice and Firm-Idiosyncratic Stock Returns,” Journal of Marketing, 71 (3), 75–88.

53.

Lapré

M. A.

Tsikriktsis

(2006), “Organizational Learning Curves for Customer Dissatisfaction: Heterogeneity across Airlines,” Management Science, 52 (3), 352–366.

54.

Mahajan

Vijay

Muller

Eitan

Kerin

Roger

(1984), “Introduction Strategy for New Products with Positive and Negative Word-of-Mouth,” Management Science, 30 (12), 1389–1415.

55.

Malhotra

Arvind

Malhotra

Claudia Kubowicz

(2013), “Exploring Switching Behavior of US Mobile Service Customers,” Journal of Services Marketing, 27 (1), 13–24.

56.

Mayerowitz

Scott

(2013), “Plane Crash Survival Odds Boosted By Safety Advances,” Huffingtonpost.com (July 7). (accessd August 22, 2013), [available at http://www.huffingtonpost.com/2013/07/07/plane-crash-survival_n_3558538.html].

57.

McCollough

Michael A

Berry

Leonard L.

Yadav

Manjit S.

(2000), “An Empirical Investigation of Customer Satisfaction after Service Failure and Recovery,” Journal of Service Research, 3 (2), 121–137.

58.

McDaniel

Stephen W.

Zeithaml

Valarie A.

(1984), “The effect of fear on purchase intentions,” Psychology & Marketing, 1 (3-4), 73–82.

59.

Mittal

Vikas

Huppertz

John W.

Khare

Adwait

(2008), “Customer complaining: the role of tie strength and information control,” Journal of Retailing, 84 (2), 195–204.

60.

Morgeson III

Forrest V.

Mithas

Sunil

Keiningham

Timothy L.

Aksoy

Lerzan

(2011), “An Investigation of the Cross-National Determinants of Customer Satisfaction,” Journal of the Academy of Marketing Science, 39 (2), 198–215.

61.

Mouawad

Jad

Drew

Christopher

(2013), “Airline Industry at Its Safest Since the Dawn of the Jet Age,” New York Times. (February 12), A1.

62.

Oliver

Richard L.

(1997), Satisfaction: A Behavioral Perspective on the Consumer. New York, NY: McGraw-Hill Irwin.

63.

Oliver

Richard L.

(2010), Satisfaction: A Behavioral Perspective on the Consumer, 2nd ed. Armonk, NY: M.E. Sharpe.

64.

Oremus

Will

(2011), “Are People Still Afraid To Fly on 9/11?” Slate.com (September 9). (accessed August 22, 2013), [available at http://www.slate.com/articles/news_and_politics/explainer/2011/09/are_people_still_afraid_to_fly_on_911.html]

65.

Ostrowski

Peter L.

Brien V

Terrence O

Gordon

Geoffrey L.

(1993), “Service Quality and Customer Loyalty in the Commercial Airline Industry,” Journal of Travel Research, 32 (2), 16–24.

66.

Parasuraman

Zeithaml

Valarie A.

Berry

Leonard L.

(1985), “A Conceptual Model of Service Quality and Its Implications for Future Research,” Journal of Marketing, 43 (4), 41–50.

67.

Reeves

Carol A.

Bednar

David A.

(1994), “Defining Quality: Alternatives and Implications,” Academy of Management Review, 19 (3), 419–445.

68.

Rego

Lopo L.

Morgan

Neil A.

Fornell

Claes

(2013), “Reexamining the Market Share-Customer Satisfaction Link,” Journal of Marketing, 77 (5), 1–20.

69.

Rhoades

Dawna L.,

Waguespack

Blaise

(2005), “Strategic imperatives and the pursuit of quality in the US airline industry,” Managing Service Quality, 15 (4), 344–356.

70.

Rhoades

Dawna L.,

Waguespack

Blaise

Jr (2008), “Twenty years of service quality performance in the US airline industry,” Managing Service Quality, 18 (1), 20–33.

71.

Ropeik

David

(2011), “Risk Reporting 101,” Columbia Journalism Review (March 11). (accessed August 28, 2013), [available at http://www.cjr.org/the_observatory/risk_reporting_101.php?page=all]

72.

Rust

Roland T.

Oliver

Richard L.

(2000), “Should We Delight the Customer?” Journal of the Academy of Marketing Science, 28 (1), 86–94.

73.

Rust

Roland T.

Zahorik

Anthony J.

Keiningham

Timothy L.

(1996), Service Marketing. New York, NY: HarperCollins.

74.

Rust

Roland T.

Zahorik

Anthony J.

(1993), “Customer Satisfaction, Customer Retention, and Market Share,” Journal of Retailing, 69 (2), 193–215.

75.

Sajtos

Laszlo

Brodie

Roderick J.

Whittome

James

(2010), “Impact of Service Failure: The Protective Layer of Customer Relationships,” Journal of Service Research, 13 (2), 216–229.

76.

Salvador-Ferrer

Carmen Maria

(2010), “Quality of University Services: Dimensional Structure of SERVQUAL vs. ESQS,” Service Science, 2 (3), 167–176.

77.

Schneider

Benjamin

Macey

William H.

Lee

Wayne C.

Young

Scott A.

(2009), “Organizational Service Climate Drivers of the American Customer Satisfaction Index (ACSI) and Financial and Market Performance,” Journal of Service Research, 12 (1), 3–14.

78.

Shapiro

Terri

Nieman-Gonder

Jennifer

(2006), “Effect of Communication Mode in Justice-based Service Recovery,” Managing Service Quality, 16 (2), 124–144.

79.

Smith

Amy K.

Bolton

Ruth N.

Wagner

Janet

(1999), “A Model of Customer Satisfaction with Service Encounters Involving Failure and Recovery,” Journal of Marketing Research, 36 (3), 356–372.

80.

Smith

Amy K.

Bolton

Ruth N.

(1998), “An Experimental Investigation of Customer Reactions to Service Failure and Recovery Encounters Paradox or Peril?” Journal of Service Research, 1 (1), 65–81.

81.

Steven

Adams B.

Dong

Yan

Dresner

Martin

(2012), “Linkages between customer service, customer satisfaction and performance in the airline industry: Investigation of non-linearities and moderating effects,” Transportation Research Part E, 48 (4), 743–754.

82.

Stuteville

John R.

(1970), “Psychic defenses against high fear appeals: A key marketing variable,” Journal of Marketing, 34 (2), 39–45.

83.

Sussman

Dalia

(2001), “Poll: Fear of Flying Has Eased,” ABC News. (October 9). (accessed August 28, 2013), [available at http://abcnews.go.com/Politics/PollVault/story?id=39335]

84.

Suzuki

Yoshinori

Tyworth

John E.

Novack

Robert A.

(2001), “Airline market share and customer service quality: a reference-dependent model,” Transportation Research Part A: Policy and Practice, 35 (9), 773–788.

85.

Strait

George

Jennings

Peter

(1997), “Recalled Fen-Phen, Redux Worse than Thought,” World News Tonight with Peter Jennings [Television Broadcast], September 24, 1997. (accessed March 06, 2011), [available at http://www.highbeam.com/doc/1P1-28123415.html]

86.

Vassilikopouloua

Aikaterini

Siomkosa

George

Chatzipanagiotoua

Kalliopi

Pantouvakis

Angelos

(2009), “Product-Harm Crisis Management: Time Heals All Wounds?” Journal of Retailing and Consumer Services, 16 (3), 174–180.

87.

Vilares

Manuel J.

Almeida

Maria H.

Coelho

Pedro S.

(2010), “Comparison of Likelihood and PLS Estimators for Structural Equation Modeling: A Simulation with Customer Satisfaction Data,” in Handbook of Partial Least Squares, Esposito Vinzi

Chin

Wynne W.

Henseler

Jörg

Wang

Huiwen

, eds. Heidelberg, Germany: Springer.

88.

Vinzi

Vincenzo Esposito

Trinchera

Laura

Amato

Silvano

(2010), “PLS Path Modeling: From Foundations to Recent Developments and Open Issues for Model Assessment and Improvement,” in Handbook of Partial Least Squares, Esposito Vinzi

Chin

Wynne W.

Henseler

Jörg

Wang

Huiwen

, eds. Heidelberg, Germany: Springer.

89.

Weun

Seungoog

Beatty

Sharon E.

Jones

Michael A.

(2004), “The Impact of Service Failure Severity on Service Recovery Evaluation and Post-Recovery Relationships,” Journal of Services Marketing, 18 (2), 133–146.

90.

Wold

Herman

(1980), “Soft Modeling: Intermediate between Traditional Model Building and Data Analysis,” Mathematical Statistics, 6 (1), 333–346.

91.

Roger

(2007), “Airlines score lower than IRS in customer satisfaction,” USA Today, (May 15), 5B. (May 22, 2014), [available at http://usatoday30.usatoday.com/travel/news/2007-05-15-airline-survey-usat_N.htm]

92.

Zeithaml

Valarie A.

Berry

Leonard L.

Parasuraman

(1996), “The Behavioral Consequences of Service Quality,” The Journal of Marketing, 60 (2), 31–46.

93.

Zins

Andreas H.

(2001), “Relative Attitudes and Commitment in Customer Loyalty Models: Some Experiences in the Commercial Airline Industry,” Journal of Service Management, 12 (3/4), 269–294.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.16 MB