Download as pdf or txt
Download as pdf or txt
You are on page 1of 15

Journal of Services Marketing

Giving back the “self” in self service: customer preferences in self-service failure recovery
Joel E. Collier, Michael Breazeale, Allyn White,
Article information:
To cite this document:
Joel E. Collier, Michael Breazeale, Allyn White, (2017) "Giving back the “self” in self service: customer preferences in self-
service failure recovery", Journal of Services Marketing, https://doi.org/10.1108/JSM-07-2016-0259
Permanent link to this document:
https://doi.org/10.1108/JSM-07-2016-0259
Downloaded on: 17 October 2017, At: 03:01 (PT)
References: this document contains references to 57 other documents.
To copy this document: permissions@emeraldinsight.com
The fulltext of this document has been downloaded 10 times since 2017*
Access to this document was granted through an Emerald subscription provided by emerald-srm:232872 []
Downloaded by UMEA UNIVERSITY At 03:01 17 October 2017 (PT)

For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for Authors service
information about how to choose which publication to write for and submission guidelines are available for all. Please visit
www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company manages a portfolio of
more than 290 journals and over 2,350 books and book series volumes, as well as providing an extensive range of online
products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee on Publication Ethics
(COPE) and also works with Portico and the LOCKSS initiative for digital archive preservation.

*Related content and download information correct at time of download.


Giving back the “self” in self service: customer
preferences in self-service failure recovery
Joel E. Collier
Department of Marketing, Mississippi State University, Starkville, Mississippi, USA
Michael Breazeale
Department of Marketing, Mississippi State University, Mississippi State, USA, and
Allyn White
Department of Marketing, University of Mississippi, University, Mississippi, USA

Abstract
Purpose – When a failure occurs with a self-service technology (SST), do customers want to give back the “self” in self-service? The authors explore
employee’s role in a self-service failure and how the presence of other customers can change that role. Specifically, they examine how the self-
monitoring of customers behavior during a failure can change recovery preferences.
Design/methodology/approach – Data were collected from customers of a movie self-service ticket kiosk and a grocery self-checkout. Three
experiments were conducted.
Downloaded by UMEA UNIVERSITY At 03:01 17 October 2017 (PT)

Findings – Results from these studies find that customers want employees to fully take over a transaction after a failure if it takes place in isolation.
If other patrons are present or waiting in line, then customers prefer the employee to simply correct the problem and let them complete the
transaction. Finally, the servicescape along with the presence of other customers in a self-service area can induce self-monitoring behaviors and
alter optimal recovery strategies.
Research limitations/implications – These findings have implications on the appropriate amount of recovery assistance customers need in a self-
service experience.
Practical implications – This research reveals the social and functional complexities associated with executing a satisfactory SST failure recovery,
particularly with respect to determining the extent to which the employee or customer should control the attempt.
Originality/value – This is the first study to examine the employee’s role in a self-service failure. While other studies have examined customers’
intentions in a self-service failure, authors examine how a service provider can assist in the recovery of a self-service failure.
Keywords Service recovery, Self-service technology, Customer waiting, Self-monitoring
Paper type Research Paper

A Washington D.C. man was arrested at Louis Armstrong International SST implementation, however, broadens the scope for
Airport after he allegedly punched a Continental Airlines check-in kiosk out
of frustration Sunday morning. The blow was enough to shatter the screen potential failures, a situation that can negatively influence
and put the computer terminal out of commission, according to a Jefferson customer evaluations of the service provider. Self-service failure
Parish Sheriff’s Office arrest report. (Economist.com, 2011)
presents a unique retailer dilemma, as customers have assumed
Experts generally contend that self-service technology (SST) a partial employee role to complete a service transaction
represents an important element in today’s retail environment, (Meuter et al., 2005; Zhu et al., 2007; Ding et al., 2011).
offering convenience for many customers, along with long-term Consequently, the risk associated with SST use may be
cost savings sought by providers (Oh et al., 2013; Collier and amplified because of reduced employee support in the event a
Barnes, 2015). These benefits are largely attributable to the problem occurs. Not surprisingly, prior research suggests that
reduced need for employee assistance in carrying out the technical- and process-oriented failures are the primary
customer’s transaction (Zhou et al., 2013). While some SSTs determinants of customer SST abandonment (Forbes, 2008).
still include a degree of employee participation or supervision Though SSTs are becoming ubiquitous in retail environments,
(e.g., airport kiosks, grocery checkout and theater ticketing numerous examples detail how customers’ frustrations because
kiosks), others are completely unattended by employees (e.g. of self-service failures have pushed them to full-service channels
movie rental kiosks and after-hours automated bank tellers). (Forbes, 2008; Zhu et al., 2007, 2013) and even to illegal
The heightened customer responsibility inherent in successful remedies as in the case referenced above (Economist.com,
2011). Another common reaction to SST failures is
embarrassment, as described to Forbes (2008, p. 320):
The current issue and full text archive of this journal is available on I was trying to get money out of a machine in Chicago, late at night. There
Emerald Insight at: www.emeraldinsight.com/0887-6045.htm were many people waiting in line for me. The machine gave me an incorrect
amount of money and then ate my card! I kept trying to fix the problem and

Journal of Services Marketing Received 6 July 2016


© Emerald Publishing Limited [ISSN 0887-6045] Revised 1 February 2017
[DOI 10.1108/JSM-07-2016-0259] Accepted 12 April 2017
Self-service failure recovery Journal of Services Marketing
Joel E. Collier, Michael Breazeale and Allyn White

everyone in line was yelling at me to hurry up. I was so angry and was very encourage self-monitoring, especially when waiting customers
embarrassed.
are funneled into one area, leading to greater visibility of a self-
To compound the problem for retailers, previous research service failure and subsequent recovery.
indicates customers are less likely to notify a company when an From a managerial perspective, this research provides
SST failure occurs, leaving the retailer unable to rectify the guidelines for executing SST failure recovery strategy that align
problem (Robertson and Shaw, 2005). with customer expectations regarding the employee’s role in a
While existing services literature is replete with studies on variety of recovery contexts. Implications from the current
customers’ overall attitude toward SST, current research lacks findings are insightful for the increasing number of retailers
a comprehensive understanding of customer expectations for embracing omni-channel strategies that focus on maximizing
successful self-service recovery and the extent to which a customer relationship management at every provider
provider’s recovery protocol impacts future behavioral touchpoint in a cohesive manner (Friedenberg, 2014). Failures
intentions. Previous research reports that infusing an employee are inevitable in services, but the way management approaches
into the recovery process can be frustrating for customers failure resolution can determine the difference between a
originally seeking to avoid human interaction by choosing this successful SST implementation and customers’ ultimate
channel alternative (Mattila et al., 2011), while other research SST—or provider—abandonment. By understanding the
suggests some customers prefer to have an employee nearby in unique dynamics of self-service failures, management can
the event of a failure (Dixon et al., 2001; Zhou et al., 2013). establish recovery strategies that prevent customer defection in
Moreover, customers have noted benefits in participating in the the event a problem arises. Hence, failure to provide an
recovery when an SST failure occurs (Dong et al., 2008). Thus, appropriate self-service recovery can result in customer
Downloaded by UMEA UNIVERSITY At 03:01 17 October 2017 (PT)

retailers must understand that optimal SST implementation embarrassment, negative word of mouth (NWOM) and
requires an effective failure–recovery strategy. But what is the defection which ultimately could result in a self-service option
desired role of an employee during a self-service recovery? that customers ignore while increasing the costs of full service
When a failure occurs, do customers want to give back the options.
“self” in the self-service process and let the employee take over?
In other words, to what extent are customers content to take on Theoretical development
the extra effort to perform the service recovery themselves, and
what are the critical determinants to failure resolution? Self-service recovery literature
The current research addresses these questions by exploring Service recovery in general is a topic that has received a great
the optimal strategy for employee intervention during a self- deal of attention from academics as well as practitioners.
service failure. Does the customer prefer to hand off to an Researchers have explored customers’ preferred timing of
employee in the event of a self-service failure, and if so, under recovery and compensation (Kim and Ulgado, 2012),
what circumstances? The theoretical lens guiding this research customers’ attitudes toward excessive recovery efforts (Noone,
is self-monitoring theory (Snyder, 1974), which explains the 2012), the impact of various demographic characteristics on
extent to which individuals attempt to manage public recovery preferences (Chung-Herrera et al., 2010; Mattila,
perceptions of their actions. Using self-monitoring as a 2010) and even the differences between customer involvement
foundation for our studies, we explore whether the receptivity in preferences for service recovery (Bambauer-Sachse and
to recovery options is influenced by the presence of others and Rabeson, 2015). Recovery with a self-service application,
the impression management that takes place in the recovery however, poses a unique challenge to providers seeking to
process. Specifically, we explore how customers’ self- balance customers’ desire for independence with providing the
monitoring behaviors influence self-service recovery right amount of assistance. Recovery in a self-service context
preferences. With a series of experiments, we first explore the has been previously defined as “an immediate problem
extent to which customers want employees to complete a resolution” (Dabholkar and Spaid, 2012, p. 1416) or, in
transaction after a self-service failure or to simply address the essence, the ability to simply resolve a failure to complete a
initial problem and let the customer complete the transaction. transaction. Early qualitative research in self-service recovery
The second research objective relates to how the presence of noted that in 47 per cent of failures, the provider made no
other customers waiting in line impacts a customer’s evaluation recovery attempt, and in 17 per cent of instances, the provider
of the provider’s recovery approach. With waiting customers actually compounded the negative effects of the failure in the
present during a failure, do the self-monitoring actions of the recovery process (Forbes, 2008). Of particular interest in the
customer influence how much aid a customer wants from an findings from the Forbes study, many perceived instances of
employee? In Study 1, we examine how different recovery failure with an SST were a direct result of unprompted or
strategies influence evaluations of satisfaction, intentions to unsolicited actions of an employee. Holloway and Beatty
switch to a full-service option, and other constructs such as (2003) report similar findings with a majority of SST recovery
embarrassment when a technical failure occurs. In Study 2, we attempts still resulting in dissatisfaction.
build on the findings of the Study 1 to explore how recovery These early studies catalyzed recent research on the role of
strategy preferences might change if the customer is the reason employees in the SST recovery process. The results of these
the failure occurred in an SST transaction. Finally, in Study 3, studies put forth equivocal insights regarding customers’
we explore how the environment in a self-service area might desired level of employee assistance during a self-service failure
increase self-monitoring behaviors and, subsequently, the recovery. For example, Mattila et al. (2011) found that
desired role of employees during a recovery. In particular, we customers originally motivated to use SST to avoid human
examine how the servicescape layout of a self-service area can interaction exhibited heightened frustration with employee
Self-service failure recovery Journal of Services Marketing
Joel E. Collier, Michael Breazeale and Allyn White

intervention during recovery attempts. Conversely, Dabholkar presence of others can change the outlook and actions of the
and Spaid (2012) note that customers perceive higher levels of customer hoping to avoid any negative social judgments from
control if employees aid in the SST recovery than if assistance customers waiting to use the technology.
were not provided. Most recently, Zhu et al. (2013) argue that Previous research has noted that self-monitoring individuals
customers’ level of recovery expectancy, defined as anticipated will change their behaviors in a retail environment to account
successful resolution of the problem, is driven by the for the social influence and evaluation of others. In these
interactivity and perceived control over the technology. These contexts, concerns regarding the perceptions of other
expectations in turn largely influence customer responses to the customers have been found to change spending behaviors
failure, such that lower levels of expectancy lead to heightened (Kurt et al., 2011), to modify mood state (Puccinelli et al.,
intentions to seek employee assistance. 2007) and to alter complaint intentions (Bodey and Grace,
While many of these studies examine the effectiveness of SST 2006). Indeed, Ariely and Levav (2000) note that consumer
failure recovery, they largely assume employee assistance is not decision-making is systematically different in a group context
available and/or desirable. However, anecdotal evidence from compared to private consumption because of the impression
practice paints a more complex picture of SST failure management concerns of the customer. In a self-service
experiences, in which employee participation is often necessary in context, Forbes’s (2008) qualitative study found that in an SST
some form (Galer, 2014; Sharp, 2014). Consequently, research failure the social judgment of other customers not only causes
on SST failure recovery must broaden in scope to account for the embarrassment but stimulates changes in customer behavior.
still relevant customer-employee interaction in service provision. With a self-service transaction, failures encourage not only
Are there situations where the customer would rather relinquish employees but also other customers to focus on the actions of
Downloaded by UMEA UNIVERSITY At 03:01 17 October 2017 (PT)

the transaction back to an employee to finish when a failure the customer completing the transaction. Does an audience of
occurs? Additionally, how does the presence of other customers waiting customers in a self-service failure alter how much
during a self-service failure influence customers’ intentions to “help” is desired to rectify the problem while managing the
self-recover from a failure? In considering these issues, self- impression given off to other customers? In most instances,
monitoring theory provides a foundation for explaining customers feel they have the requisite skills to complete a
customers’ self-service recovery preferences. Self-monitoring transaction without employee intervention. When a failure
theory provides insights into how customers’ behaviors and occurs, which can draw the attention of others waiting to use
evaluations of an experience can change depending on the technology, do customers want to prove their self-service
situation and the environment in which a situation occurs. efficacy or do they simply want to hand the transaction off to an
employee? Studies 1 and 2, discussed next, consider the
Self-monitoring theory influence of self-monitoring tendencies on outcomes of varied
Self-monitoring is defined as the extent to which people “can employee-customer participation in SST recovery situations.
and do exercise control over their verbal and nonverbal self- Study 3 then supplements these findings by considering the
presentation” (Snyder, 1979, p. 88). This construct describes extent to which self-monitoring may be influenced by social
the extent to which consumers are concerned with impression influences and servicescape design.
management along with the necessary actions needed to be
portrayed in a positive manner (Hu and Parsa, 2011). For Study 1: employee’s role during an SST failure
example, high self-monitors demonstrate a greater willingness
than low self-monitors to expend effort to ensure that they In Study 1, our goal was to understand if customers want to give
maintain a positive image in front of others. However, existing back the “self” in self-service when a failure occurs. To date, we
literature in consumer neuroscience and cognitive development have very little insight on customers’ preferences for employee
suggests that self-monitoring, like other self-regulatory intervention in a transaction that was initially intended to be
variables, may be situationally influenced or primed by completed solely by the customer. Additionally, SSTs are often
environmental cues (Bengtsson et al., 2011; Schunk and located in public places (or checkout areas) where there are
Zimmerman, 1998). From this perspective, customers may numerous impromptu interactions. Does the presence of other
exhibit a tendency to regulate their behavior in response to customers during an SST failure influence recovery preferences?
situational context cues to present themselves in a favorable Hence, does fear of embarrassment, and ultimately the impression
light to others (Snyder, 1974). Essentially, customers will given off to other waiting customers, impact how much employee
change their behavior in the presence of employees or other interaction is desired when an SST failure occurs?
customers to present a socially desirable image. In a self-service To study the impact of SST failures relevant to managerial
context, self-monitoring presents a unique situation. For strategy, the authors examine five specific outcomes that are
example, the choice of self-service may allow high self-monitors often the constructs of interest in SST failure studies –
to focus on the performance of the task without having to worry satisfaction with the SST transaction, intention to switch from
about the scrutiny of others. Because there is no employee the self-service option to a full-service option on the next visit
involved in the immediate transaction and no other customers (relevant to practitioners as this move would necessitate
are party to a smooth self-service transaction, there is no one to reallocation of retailer resources), propensity to spread
impress. When a self-service failure occurs, this potentially NWOM, susceptibility to embarrassment in the event of a
isolated transaction has now become a public recovery, with an failure, and felt time pressure to complete a transaction. These
employee coming to assist and the possibility of other outcomes represent both an attitudinal and a behavioral
customers that were waiting to use the technology now paying component shown to result from an SST failure (White et al.,
attention to the unfolding process. In this situation, the 2012; Zhu et al., 2013).
Self-service failure recovery Journal of Services Marketing
Joel E. Collier, Michael Breazeale and Allyn White

Satisfaction represents a customer’s overall assessment of the complete the transaction after the failure will not only decrease
total purchase and consumption experience, both actual and perceived time pressure to complete the transaction but also
anticipated (Johnson et al., 1995). The measure of satisfaction lower the chances of embarrassment compared to when the
used in this study is directed specifically toward the SST failure was simply fixed but the customer had to complete the
transaction experience itself. Linked with satisfaction, transaction again without employee involvement. With little
intentions to stay with the provider but switch service delivery perceived social threats in a private setting, customers will
methods should indicate a lesser level of dissatisfaction with the prefer a recovery option that provides more benefits to them in
provider while still representing a degree of dissatisfaction with the service process. Therefore:
the transaction in question. Finally, NWOM, defined as the
interpersonal communication concerning an organization and/ H1. Customers experiencing an SST failure in isolation will
or its products or services that denigrates the object of the be (a) more satisfied with the transaction, (b) less likely
communication (Richins, 1983; Schoefer and Diamantopoulos, to switch to the full-service option on their next visit to
2008), is examined. Existing research often cites NWOM as a the retailer, (c) less likely to spread NWOM about the
behavioral response to dissatisfactory service or service recovery retailer, (d) less impacted by perceived time pressure,
attempts (Blodgett et al., 1993; Ward and Ostrom, 2006). In and (e) less embarrassed when an employee takes over
this study, propensity to spread NWOM is included to further and completes the transaction than when the employee
indicate the degree of dissatisfaction with the service provider. merely fixes the problem and allows the customer to
We have additionally included perceived time pressure and complete the transaction.
embarrassment as constructs of interest because these feelings In a social setting, previous research has found that consumers
Downloaded by UMEA UNIVERSITY At 03:01 17 October 2017 (PT)

should be relevant dimensions of the customer’s perceptions of the will choose the alternatives that supply or may be expected to
overall situation. Perceived time pressure reflects the sense of supply the most social approval, or conversely, those that
urgency that a customer feels as the amount of time required to promise the least disapproval (McDonell et al., 2006).
complete a task approaches or exceeds the amount of time allocated Additionally, Latane (1981) argues that consumers’ behaviors
to that task (Strombeck and Wakefield, 2008). Time pressure has will change based on the real, implied, or imagined presence or
been determined to be a situational component of customer actions of others in a social setting. Hence, the presence of other
attitudes toward SST experiences (Collier et al., 2015). customers waiting to utilize the SST should impact customer
Embarrassment in this context relates to a socially occurring evaluations of recovery attempts. Previous research on service
phenomenon driven by an individual’s concern for others’ opinions recovery has found that group settings can significantly impact
(Miller, 1996). A customer potentially feels embarrassment when different recovery strategies (Zhou et al., 2013). When other
perceiving that a waiting customer judges his inability to complete customers are waiting in line, potentially disapproving of being
an SST transaction unaided. Thus, embarrassment represents a forced to wait even longer, the affected customer will have a
relevant outcome in an SST failure context (Forbes, 2008). heightened level of self-monitoring to save face in front of other
customers. When a customer chooses a self-service option,
Hypotheses there is an expectation that the customer can complete the
When a failure occurs in private, isolated from other customers, transaction apart from employee involvement. When a failure
the customer should evaluate the effort required to fix the occurs, with other patrons waiting to use the SST, recovery
problem as a cost. With most self-service transactions, the takes on an added emphasis for the affected customer to avoid
efficiency of the transaction is a determining factor for choosing appearing incompetent in the self-service process. If an
this channel option (Meuter et al., 2005). When a failure employee not only fixes the problem but completely takes over
occurs, this added cost of time and effort to rectify the problem the transaction with other customers present, this could result in
will change the expectations of recovery. Previous research has a potential social cost of feeling inadequate in the self-service
noted that customers often want a recovery that will process. In essence, the employee is signaling to others waiting
compensate for the cost of a failure (Smith et al., 1999). When that this customer is not capable of completing the self-service
an employee comes over and not only fixes the problem but transaction in a timely manner. This type of recovery could also
takes over the transaction to completion, the additional cost to enhance embarrassment for the customer and increase the
the consumer has now been removed. From a self-monitoring perceived time pressure to complete the transaction with the
perspective, in a private setting, impression management is not employee taking over the transaction process. Additionally, an
a substantial concern. The employee is present to simply aid in employee overtaking a recovery could lead to lower overall
the recovery and is ultimately there to increase the satisfaction satisfaction and greater intentions to switch to a full-service
of the customer in the recovery process. The employee’s job is option. From a self-monitoring perspective, we expect that the
to recover from the perceived “loss/cost” with the self-service presence of other customers actually lowers customer
failure and to compensate for this inconvenience. Hence, receptivity to employee intervention in the service recovery. The
customers will be more than willing to let employees take on customer wants the self-service failure resolved but does not
this added cost to complete the transaction after the initial want to be seen as socially incapable of performing the self-
failure has been resolved. With no other customers present service transaction. Accordingly:
during the failure or recovery, having an employee not only
recover from the failure but complete the transaction for the
customer should increase satisfaction from the experience and H2. Customers experiencing SST failure observed by other
prevent negative future intentions such as switching to a full- waiting customers will be (a) more satisfied with the
service option or NWOM. Additionally, letting the employee transaction, (b) less likely to switch to the full-service
Self-service failure recovery Journal of Services Marketing
Joel E. Collier, Michael Breazeale and Allyn White

option on their next visit to the retailer, (c) less likely to The scenario also denotes whether other customers are present
spread NWOM about the retailer, (d) less impacted by or absent during the failure to examine the impact of an
perceived time pressure and (e) less embarrassed when audience on the outcomes. (See Appendix for the scenarios.)
an employee merely fixes the problem and allows the As a post-survey manipulation check, participants answered
customer to complete the transaction than when the two questions inquiring whether other customers were waiting
employee completely takes over the transaction after and whether the employee/customer completed the checkout
the failure has been resolved. process once the problem was resolved. The sample consisted
of 67 per cent female and the average age was 21.
The source of differentiation between the responses to the two Five dependent variables were measured at the end of the
scenarios should lie in the perceived isolated vs observed nature scenario:
of the customer’s experience with the SST failure. When the 1 satisfaction with the SST transaction;
employee represents the only audience in the service recovery, 2 intentions to switch from a self-service to a full-service
letting the employee take over the transaction process will option on the next visit;
provide more benefits to the customer to compensate for the 3 propensity for NWOM;
costs of the failure with little concerns of social inadequacy in 4 perceived time pressure to complete the transaction; and
terms of self-monitoring. Conversely, when other customers 5 embarrassment.
are present to witness the failure/recovery, the customer will be
more willing to take on the additional effort in the transaction The items were adopted from existing related research (Bougie
process to avoid the social costs of feeling inadequate in front of et al., 2003; Oliver and Swan, 1989; Spears and Singh, 2004;
Collier et al., 2015; Kelly and Jones, 1997) and were measured
Downloaded by UMEA UNIVERSITY At 03:01 17 October 2017 (PT)

others. In this case, the social evaluation of others is a greater


concern than the employee providing additional assistance in on seven-point scales ranging from 1 (strongly disagree) to 7
the recovery process. (strongly agree). We also measured the concept of self-
monitoring or the propensity of customers to alter their
Method behavior in front of others from the research of Lennox and
To test these hypotheses, four scenarios were developed to Wolfe (1984). This variable was used as a covariate in the
manipulate a 2 (employee takes over and finishes transaction vs analysis to control for differing levels of self-monitoring in the
customer finishes transaction)  2 (customers waiting in line vs recovery process.
no customers in line) experimental design. The setting for the The internal consistency for each construct was measured by
study was a self-service movie ticketing kiosk that would allow computing coefficient alpha. Each variable had a coefficient
patrons to purchase movie tickets for upcoming shows. This alpha greater than 0.80 supporting the reliability of scale items.
setting was chosen because it is a relatively new technology To further assess the validity of our dependent variables, a
diffusing in the service landscape of movie theaters. Thus, the confirmatory factor analysis was performed. The results of the
recovery process of this SST may be unfamiliar to some analysis indicate the model’s adequate fit to the data ( x 2 =
customers compared to well established SSTs. Utilizing a 487.33, df = 283, CFI = 0.95, TLI = 0.94, RMSEA = 0.06).
college student panel from a southeastern school, previous All items loaded on their respective constructs with t-values
users of an SST movie ticketing kiosk were recruited. To ranging from 5.30 to 28.24. (See Appendix 2 for the full list of
qualify for the survey, respondents had to denote the last time items) With no concerns over the validity of the measures, a 2 
and location they used an SST movie ticketing kiosk. This was 2 MANCOVA analysis was performed. The analysis revealed a
done to verify that an SST kiosk was available during that time significant difference between recovery type and the presence/
and at that location. absence of other customers with respect to the dependent
A total of 158 previous SST kiosk users were recruited for the variables: Wilks l = 0.68, F(15, 398) = 3.94, p < 0.001 (See
survey. Respondents were directed to an online survey where Table I for the full analysis). The results of the main effects
one of four scenarios were randomly presented (35 1 found that each construct was significantly different across the
respondents were in each scenario). The scenarios describe a manipulations. Further, the univariate analyses provided
situation in which the respondent is going to purchase a ticket interesting insights with respect to the hypothesized
for a movie showing multiple hours in the future but, because of expectations. In an isolated setting when no other customers
the popularity of the film, getting tickets before going to dinner were present during the failure, customer satisfaction was
is a preferred plan. Next, the scenario describes that the significantly higher when the employee took over the service
respondent finds the movie and specific show time and tries to recovery process instead of allowing the customer to finish it
checkout but the SST interface freezes up during the (Memployee = 5.81, SD = 1.11; Mcustomer = 5.08, SD = 1.36, p =
transaction. Thereafter, one of two recoveries options is 0.01). In this instance, customers exhibit heightened
presented to the respondent. In the first recovery, respondents willingness to cede control back to the employee not only to fix
were told that an employee comes over, resets the SST, and the problem but also to aid in completing the transaction.
finds the movie and desired show time again and gets the Interestingly, employees aiding in an SST failure had a
customer back to the final checkout screen. The employee is dramatically different impact when other customers were
present during the remaining transaction until the movie tickets present. In the presence of others, customers rated their
are dispersed. In the second recovery scenario, the employee satisfaction lower when an employee took over the failure
simply resets the SST and lets the customer start over to find recovery and subsequent transaction completion. (Mcrowded =
the movie and show time themselves along with pay for the 4.52, SD = 1.31; Misolated = 5.81, SD = 1.11, p < 0.001).
tickets with no further help or assistance from the employee. Additionally, NWOM and perceived time pressure were all
Self-service failure recovery Journal of Services Marketing
Joel E. Collier, Michael Breazeale and Allyn White

Table I SST recovery with technology causing the failure


Means (SD) Satisfaction Switch NWOM Time pressure Embarrass
Other customers waiting– customer complete transaction 4.93 (1.20) 4.78 (1.22) 3.58 (1.53) 5.58 (1.20) 3.97 (1.89)
Other customers waiting– employee complete transaction 4.52 (1.31) 4.57 (1.66) 3.78 (1.52) 5.11 (1.52) 4.37 (1.78)
Isolated– customer complete transaction 5.08 (1.36) 3.98 (1.39) 2.98 (1.76) 4.24 (1.59) 3.27 (2.05)
Isolated– employee complete transaction 5.81 (1.11) 3.92 (1.81) 2.72 (1.40) 3.64 (1.63) 2.66 (1.80)
Multivariate statistics
Wilks k = 0.68 F-value = 3.94, p-value = < 0.001
Main effects test
Satisfaction: F-value = 8.15, p-value = < 0.001
Time pressure: F-value = 13.17, p-value = < 0.001
Switching: F-value = 2.76, p-value = 0.04
NWOM: F-value = 3.47, p-value = 0.01
Embarrass: F-value = 5.77, p-value = 0.001
Post hoc tests (p value significance)
C-isolated 3 E-isolated 0.01 0.87 0.45 0.08 0.16
E-others waiting 3 E-isolated < 0.001 0.06 0.002 < 0.001 < 0.001
Downloaded by UMEA UNIVERSITY At 03:01 17 October 2017 (PT)

C-others waiting 3 E-others waiting 0.13 0.54 0.56 0.17 0.34


C-others waiting 3 C-isolated 0.57 0.03 0.09 < 0.001 0.11
Note: C-others waiting = customer completes transaction with others customers waiting in line; E-Others Waiting = employee completes transaction with
other customers waiting in line; C-isolated = customer completes transaction in an isolated setting; E-isolated = employee completes transaction in an
isolated setting; Items: Satisfaction: 1 – very dissatisfied; 7 – very satisfied; Switch: 1- not likely; 7 – very likely; NWOM, time pressure, perceived
embarrassment: 1 – strongly disagree; 7 – strongly agree

significantly higher when an employee took over an SST customer having to complete a transaction after the failure has
transaction when others were present compared to an isolated been resolved.
failure. Lastly, customers noted a significantly higher level of As receptive as customers were to employees’ taking over a
embarrassment when the employee took over the transaction in transaction in an isolated environment, the exact opposite
front of an audience (Mcrowded = 4.37, SD = 1.78; Misolated = appears to be the case with a waiting customer audience
2.66, SD = 1.80, p < 0.001). Based on these findings, the present. Customer responses suggest that having an
employee’s role preferred by the customer in an SST failure employee take over the self-service process in front of others
was dependent on the social environment in which the failure created embarrassment and a higher perception that the
took place. In an isolated environment, employees’ taking over transaction needed to be completed quickly. Subsequently,
the transaction ultimately leads to a better perceived service customers had substantially lower evaluations of the service
experience. Conversely, in the presence of other patrons, process when the employee intervened and fully carried out
customers were embarrassed and felt greater time pressure to the recovery, leading to greater intentions to switch to a full-
complete the transaction when an employee took over the service option. With a greater potential for embarrassment
service process. These findings underscore the notion that, in when an employee takes over while others are waiting,
line with self-monitoring theory, customers want to prove their customers will ultimately avoid a transaction method that
ability to complete a self-service transaction in the presence of could cause social harm. Interestingly, although time
other customers. pressure was amplified in the presence of other patrons,
respondents expressed higher satisfaction when allowed to
complete the transaction compared to standing aside while
Discussion the employee finished the interaction. Upon further review,
Do customers want to give up the “self” in self-service when these findings highlight the importance of social costs
there is a problem? The results of Study 1 suggest this is the case compared to functional costs in a SST failure scenario. No
in the context of an isolated setting. Customers welcome this customer wants to feel inadequate in front of others, and an
extra level of help, ultimately enhancing evaluations of the employee taking over a transaction could drive the perception
service experience. In an isolated setting, the results suggest that the customer is not skilled or experienced enough to
that customers see the added value of letting the employee complete the transaction. Hence, in a crowded environment,
complete the transaction after a failure. One explanation for employees are better to resolve the immediate problem (i.e.,
these findings relates to the lack of social pressures or perceived the technical failure) and subsequently let the customer
evaluation from others when completing the transaction. complete the transaction to avoid embarrassment or
Absence of other customers’ scrutiny likely enhanced the perceived pressure to complete the transaction quickly.
receptivity toward the employee aiding in the self-service Even when customers are allowed to complete the
process, which would arguably be more time efficient than the transaction themselves, Study 1 found that the presence of
Self-service failure recovery Journal of Services Marketing
Joel E. Collier, Michael Breazeale and Allyn White

other customers still promotes higher switching intentions than and to whom the initial failure was attributed. The sample was
when a customer completes a failed SST transaction in 64 per cent female and the average age was 25.
isolation. This finding suggests that heavy traffic times during Items measuring the five dependent variables from Study 1
the day will require a more involved service recovery attempt were used again with the exact same scale items. We also
outside of simply taking over the transaction. When a failure included the self-monitoring items from Study 1 to be used as a
occurs at no fault of the customer, having other patrons covariate in Study 2. The validity of the items was assessed
publicly watching the customer recover from the transaction again, and the reliability of the constructs remained at an
could lead to increased anxiety. The employee would be well acceptable level (a > 0.80) with the confirmatory factor
served to apologize for the failure and remain present while the analysis finding that the model fit the data, each item loading
customer (not the employee) completes the transaction, strongly on its respective factor ( x 2 = 496.89, df = 283, CFI =
assuring that no additional failures or delays occur. Employees 0.94, TLI = 0.93, RMSEA = 0.07).
should take on the role of an aid in the process instead of The results of the MANCOVA analysis revealed a significant
completely taking over the transaction. interaction between who recovered from the failure (employee/
customer) and whether other customers were present/absent
Study 2: customer-induced SST failure and the when the customer caused the failure (Wilks l = 0.82, F(15,
employee’s role 360) = 1.72; p < 0.04). In contrast with Study 1, the main effects
test only observed satisfaction (F(3,138) = 2.97; p = 0.04) and
The previous experiment assumes that the failure was a direct
perceived time pressure (F(3,138) = 3.73, p = 0.01) as
result of the technology having a problem. But what if the
significantly different across the groups. Per the univariate
Downloaded by UMEA UNIVERSITY At 03:01 17 October 2017 (PT)

customer is the cause of the self-service failure? Would this


analysis, a significant difference remains between employees
change how receptive customers are to giving the “self” in self-
taking over the service process when other customers are present
service back and allowing the employee to complete the
compared to an isolated setting. Customers were significantly
transaction? Previous research has noted that customers react
more satisfied when an employee took over the transaction when
differently to an SST failure depending on the party at fault
other customers were not present (Mcrowded =5.15, SD = 1.37;
(Snellman and Vihtkari, 2003). Additionally, Choi and Mattila
Misolated = 5.97, SD = 1.12, p = 0.01). Additionally, when an
(2008) found that when the customer feels partially responsible
employee takes over the service failure recovery process with
for an SST failure, service recovery evaluations were higher
other customers present, the perceived time pressure to complete
than when the service provider was at fault. Attribution theory
the transaction is significantly higher compared to the scenario
(Heider, 1958) suggests that customers experiencing a self-
where an employee takes over in an isolated environment
induced failure might be more willing to forgive the failure and
(Mcrowded =4.62, SD = 1.54; Misolated = 3.53, SD = 1.76, p <
the costs attached to resolving it. However, conflicting research
0.001). Consistent with Study 1, having the employee take over
findings suggest that consumers often have a self-serving bias
where a customer attributes successes to himself and failures to in a crowded environment can elicit negative customer responses.
the company (Yen et al., 2004). When a customer is the cause In Study 2, the situational impacts on NWOM, switching
of the self-service failure, do customers want more employee intentions and embarrassment were muted, as compared to
intervention in the recovery process to prevent further delays, Study 1, as these differences did not reach a level of significance.
or are customers still content to process the transaction apart With respect to scenarios in which the customer is left to
from employee intervention once the failure is resolved? Study complete a transaction with no further assistance by the
2 examines our hypotheses in a situation where the customer, employee, minimal differences emerged among any of the
instead of the technology, is the cause of the failure. Does the constructs across experimental cells. Overall, the highest
attribution of fault influence a customer’s response to an satisfaction levels were observed when an employee took over the
employee’s involvement in an SST failure recovery process? transaction in isolation, followed by the customer completing the
Subsequently, how do these two constructs interact to transaction in isolation. Across both studies as self-monitoring
influence the overall evaluation of the experience? theory would predict, having a self-service failure in front of other
customers led to a worse experience regardless of the recovery
Method attempt and who was at fault for the failure.
The same four scenarios were used for this experiment as in To further assess the differences across the customer- and
Study 1 except, instead of the SST causing the problem, the technology-induced failures, we ran simple ANOVAs across
customer causes the failure. The revised scenario reads: the failure types to determine if type of failure significantly
altered the service experience. Regarding overall evaluations,
You scan your credit card to pay for the movie tickets, but before the tickets
are dispensed, you notice that you purchased tickets for the current show the customer-induced failure was significantly higher than the
time instead of the desired future show time. You motion for an employee to technical failure in satisfaction ratings for all types of recovery
come over and help correct the problem. settings. Figure 1 shows a graphical display of the results.
As in Study 1, a college student panel recruited previous SST Similar to previous research, customer inducement of an SST
movie ticket kiosk users. A total of 150 SST movie ticket kiosk failure tends to alter the customer’s attitudes toward various
users were recruited and randomly assigned to one of the four outcomes, perhaps because of a self-serving buffer. Customers
groups via an online survey. After reading the scenario, appear easier to satisfy after the failure and felt less negative
respondents answered three manipulation questions that pressure when the failure was a direct result of the customer’s
assessed whether the SST area was crowded/isolated, whether mistake. Conversely, satisfaction levels were consistently lower
the customer/employee finished the transaction post-failure, when the technology was at fault, supporting the conclusion
Self-service failure recovery Journal of Services Marketing
Joel E. Collier, Michael Breazeale and Allyn White

Figure 1 Satisfaction ratings compared across studies Method


For this study, we extend our findings by exploring another
SST Fault Customer Fault self-service context while also further examining how the
7 presence of waiting customers in different servicescapes
impacts the recovery of an SST failure. Not only did we want to
6.5
examine the impact of different types of SST areas but also
wanted to control the specific number of customers waiting in
6
line, to further control for the effect of customers waiting in line
5.5
during a failure. As in the previous two studies, we explored
intentions to switch to a full-service option and perceived time
5 pressure, but we also wanted to explore customers’ intentions
to leave the provider entirely based on how the recovery process
4.5
was handled. Additionally, for this study, we specifically
measured customers’ self-monitoring feelings during the
4
CROWDED ISOLATE EMPLOYEE CROWDED ISOLATE CUSTOMER recovery process instead of using it as a covariate as we did in
EMPLOYEE FINISHES CUSTOMER FINISHES
FINISHES FINISHES the previous studies.
The setting for this study was a grocery store self-checkout
Note: Scale items were on a 1 to 7 scale area. We chose this setting because grocery stores prominently
display both types of SST setups, quad design and multiple
Downloaded by UMEA UNIVERSITY At 03:01 17 October 2017 (PT)

single-line designs. The diversity of grocery store self-service


that customers are less tolerant of technology failures, areas allows for a rich test of how the servicescape and
especially if the failure takes place in front of other customers. specifically waiting customers in the servicescape influence
recovery attitudes.
A 2  2 research design was used with respondents exposed
Study 3: how the servicescape influences to one of the four scenarios, where the customer is either in a
self-service recovery quad SST checkout area or in a multiple single file line SST
In Study 3, we wanted to further explore self-monitoring and area. Sample pictures of a quad and single-line SST design were
the impact of waiting customers on the self-service recovery provided in the survey. Consistent with Study 1, when the SST
preferences of customers. In the previous studies, we controlled has a technical failure, either the employee comes over, simply
for the possible differences of customers’ self-monitoring fixes the problem and allows the customer to complete the
tendencies, but in Study 3 we wanted to explore whether the transaction, or the employee comes over, fixes the problem and
self-service area in which a failure takes place could induce then completes the transaction for the customer. Across all the
higher self-monitoring feelings. Previous research has noted scenarios, respondents are given the information that they are
that publicly consumed goods in a retail environment can make waiting in line to use the self-service checkout area and when it
customers acutely aware of others and, thus, increase self- is their turn at the self-service checkout machine, there are two
customers waiting in line behind them. (See Appendix 1 for
monitoring behaviors (Graeff, 1996). Additionally, Rose and
examples of scenarios).
DeJesus (2007) found that consumers could have an elevated
Respondents for this study were recruited from Amazon
or motivated cognition to self-monitor during a consumer
Mechanical TURK. To qualify for the survey, respondents had
buying process. Thus, we wanted to explore this topic further
to be previous self-service checkout users in a grocery setting.
by considering whether the layout of a self-service area could
Respondents were paid for their participation and were free to
heighten customers’ feelings of self-monitoring because of the
exit the survey at any time. Attention and manipulation check
visibility of customers watching a transaction take place. Self-
measures were placed throughout the survey to assure that the
service areas are usually designed in either a quad set up where
respondents were providing trustworthy responses.
all self-service customers form a single line, and then proceed to Respondents were randomly exposed to one of the four
one of four (or more) self-service machines that are laid out in a scenarios and were asked to answer the following questions, as
box format or in multiple single-file lines designed to be similar if this scenario had actually happened to them. A total of 180
to a traditional full-service checkout. With the quad set up, all respondents filled out the survey but because of attention check
self-service customers are consolidated into one designated failures five respondents were dropped leaving a total of 175
area, thus forcing all customers to view every transaction taking respondents (40 1 respondents for each scenario). The average
place in the quad until a checkout becomes available. In the age of the respondent was 35 with 52 per cent of the sample
alternative setup, multiple single-file lines form with self-service being male.
customers disbursed across multiple lines, and customers The items for intentions to switch to a full-service option and
usually have obstructed views of one another. Hence, the perceived time pressure were the same items used in the
servicescape layout of the self-service area could highlight the previous two studies. For intentions to leave the retailer, we
presence of other customers in a way that enhances self- adapted a three-item scale from Spears and Singh (2004) that
monitoring. This increased self-monitoring should impact asked how likely would the respondent be able to switch to
customers’ preference for handling of the recovery process as another grocery store because of this incident with the anchors
they hope to protect their self-image as a customer who is a of definitely would not/definitely would; Probably not/
capable self-service user. Probably; and Not likely/Very Likely. Self-monitoring was
Self-service failure recovery Journal of Services Marketing
Joel E. Collier, Michael Breazeale and Allyn White

measured using an abbreviated three-item scale from Lennox failure occur, the affected customer would rather complete the
and Wolfe (1984) (Appendix). The reliability of all the items transaction themselves to protect their image in front of other
were assessed again, and all constructs had a Cronbach’s alpha customers. Having an employee complete the transaction for
level greater than .80. them could possibly reflect poorly on the user’s SST ability or
After assessing the validity of the measures, a MANOVA necessary skills to complete the transaction without further
analysis was run and the results found a significant interaction problems or delays.
across the scenarios (Wilks l = 0.19, F(12, 510) = 2.90 and Finally, perceived time pressure was explored to see if
p <0.001). With intentions to switch to a full-service option, the customers feel pressure to complete a transaction after a failure
main effects test found a significant relationship (F(3,171) = occurs. The main effects test was significant across the
2.92 and p = 0.03) across the scenarios. In the univariate test, scenarios (F(3,171) = 4.14 and p < 0.001). In the quad setting,
the results of the analysis found that when an employee tried to when the customer completes the transaction, perceived time
fix and take over the self-service process in a quad setting pressure was significantly higher than when the employee
(Memployee = 3.99, SD = 1.89) this produced a significantly completes the transaction after the failure (Memployee = 3.12
higher intention to switch to a full-service option than when the SD = 1.80; Mcustomer = 3.97, SD = 1.46; p = 0.02).
employee simply fixed the problem and let the customer Furthermore, customers completing the transaction themselves
complete the transaction (Mcustomer = 3.13, SD = 1.74; p = in the individual single file line format had higher feelings of
0.03). Additionally, in the quad setting where more customers perceived time pressure than when the employee took over
are potentially watching a customer experience a failure, an (Memployee = 2.65 SD = 1.86; Mcustomer = 3.25, SD = 1.77; p =
employee taking over and completing the transaction process 0.10). Across the two formats, customers had higher levels of
Downloaded by UMEA UNIVERSITY At 03:01 17 October 2017 (PT)

led to higher switching intentions than when an employee took time pressure in the quad setting than in the individual single
over and completed the transaction process in an individual line design (p = 0.05). The results of the time pressure analysis
single file line SST design (Mquad = 3.99, SD = 1.89; Mindividual- add an interesting twist to our results in that customers have a
line = 3.06, SD = 1.75; p = 0.02). Consistent with our previous heightened level of self-monitoring during a failure while other
studies, when there is a perceived heightened number of waiting customers are waiting and prefer to handle the transaction
customers watching a failure take place, the affected customer themselves, but, at the same time, they feel a greater sense of
simply wants the employee to fix the problem and let the pressure to complete the transaction quickly even when the
customer complete the transaction without any further failure was not a result of their own actions.
assistance.
Customers’ intentions of leaving the retailer after a self- Discussion & managerial implications
service failure were not as strong as switching to full-service
intentions. In part, these findings offer some encouragement Given the increasing implementation of self-service options in a
for retailers, as customers being dissatisfied with a checkout widening array of service industries (Friedenberg, 2014; Power,
option does not necessarily mean they are willing to leave the 2014), research in this arena must assume a more
store as whole. The main effect test for leaving the retailer was comprehensive approach toward understanding customer
significant (F(3,171) = 2.84, p = 0.03) across the scenarios. In expectations and resulting evaluations of the provider’s SST
the quad setting, having the employee take over the transaction failure–recovery strategy. This series of experiments provides
did not have a significant effect compared to allowing the insight with respect to multiple important elements of the
customer to complete the transaction (p = 0.42). The means recovery. First, in isolation respondents are more satisfied when
were roughly equivalent between these scenarios, but there was the employee takes full responsibility for completing the
a significant difference between an employee taking over the transaction post-failure, regardless of fault attribution. This
transaction in a quad setting (Mquad = 2.37 SD = 1.67) and an finding highlights that customers are happy with the employee’s
individual single file setting (Mindividual-line = 1.52, SD = 1.06; taking on the additional effort to not only rectify the problem
p < 0.001). Again, when an employee took over the transaction but also to complete their transaction. In a private setting where
where it was more visible to other customers, this led to a strong time pressure and embarrassment are less of a concern,
negative reaction by the affected customer. customers want employees to expend that little extra effort to
To further examine whether waiting customers influenced fully recover from the failure. Across all the scenarios,
SST users, we wanted to explore how customers’ self-monitor customers were the most satisfied and more likely to continue
their own behavior when an SST failure occurs with others using an SST if the employee took over the transaction as long
watching. The main effects test found a significant difference as other customers were not present. These results run counter
across the scenarios in customers’ self-monitoring (F(3,171) = to traditional thoughts on self-service that suggest customers
3.79; p = 0.01). When an employee took over the transaction in want as little employee interaction as possible, even when a
the quad set up, self-monitoring was significantly higher than failure occurs. Our findings show that service recovery efforts
when the customer completed the transaction in the same set need to go beyond simply “fixing” the problem to compensate
up (Memployee = 4.85 SD = 1.05; Mcustomer = 4.38, SD = 0.95; for the potential time and effort costs that now need to be made
p = 0.05). Similarly, self-monitoring was significantly higher for to finish a transaction. When the social costs of other customers
consumers in the individual single file line when an employee watching the recovery are removed, customers prefer a higher
took over the transaction than when the customer was allowed level of service recovery in a self-service transaction.
to complete the transaction after the failure (Memployee = 4.84 Second, this research reveals the social and functional
SD = 1.25; Mcustomer = 4.23, SD = 1.05; p < 0.001). In both complexities associated with executing a satisfactory SST failure
SST settings when other customers are waiting and watching a recovery, particularly with respect to determining the extent to
Self-service failure recovery Journal of Services Marketing
Joel E. Collier, Michael Breazeale and Allyn White

which the employee – or customer – should control the attempt. transaction. This subsequently promotes more self-monitoring
Contrary to the results of the isolated setting, customers do not behaviors of customers, which heightens the possible negative
prefer more employee interaction when other customers are feelings that are experienced when a failure occurs. Service
present during the failure. Our results suggest that having an providers need to be acutely aware that the design of a self-
employee take over the transaction fostered higher feelings of service area can directly influence the preferred type of service
embarrassment and time pressure, which led to lower recovery with a self-service transaction. Though pushing all
satisfaction levels. In this situation, a higher level of service self-service users into a quad setting may save retail space, it
recovery (more employee involvement) led to lower satisfaction. also increases the visibility of failures and hence self-
Customers want to maintain their self-image in front of others, monitoring, along with potential feelings of embarrassment
demonstrating that they are a fully capable SST customer who because of employees’ having to come over an aid in the
does not need an employee to complete their transaction like a transaction process. While previous research has examined
novice or a first-time participant. The social costs of appearing service recovery in an environmental bubble, we argue that the
inadequate outweigh the benefits produced by a higher level of servicescape of a self-service area can and does directly
employee involvement in the recovery process. Additionally, in influence self-monitoring behaviors and subsequently impact
situations where the technology is clearly the source of failure, recovery preferences.
the presence of others stimulates higher time pressure, yet less
positive evaluations when the employee completes the Limitations and future research
transaction for the customer. These findings have important With increased SST implementation across service industries,
insights for managers, as the employee’s role is to facilitate and further expansion of this research scope is warranted to
Downloaded by UMEA UNIVERSITY At 03:01 17 October 2017 (PT)

not to commandeer the transaction when other patrons are generalize more fully the findings observed here. Hence, one
viewing the recovery. The amplified time pressure, however, variable of additional importance is the type of service being
would suggest that employees should remain near the SST to provided. The movie ticket purchase services used here
ensure the customer can successfully complete the transaction generally imply customers’ seeking hedonic benefits (a movie)
in a timely manner. associated with relatively low-risk loss in the event of failure. It
Third, when customers are at fault, the nuances of their should be noted that customers’ use of utilitarian-focused SSTs
recovery evaluations are substantially less pronounced, as for completion of more highly involved and potentially
evidenced by the lack of significant differences across Study 2. confidential transactions, such as automated teller machine
One explanation for generally higher reported satisfaction in (ATM) banking and employment applications, could lead to
customer fault situations would be that when the consumer additional considerations for employee involvement during a
assumes responsibility for the problem, he/she takes ownership failure. As well, we did not control for the characteristics of the
of the fault and consequently becomes more flexible with employees or customers during our studies. future research
respect to desired recovery protocol. In other words, the could examine how the psychological tendencies and
priority becomes fixing the problem to alleviate the situation, accompanying behaviors of an employee could influence
and not managing perceptions of other patrons. Further, recovery preferences. Another area of research could further
negative emotions or embarrassment may result from the explore how the length/time of recovery and overall motivations
customer’s recognition of self as the cause of the problem, for customers seeking employee assistance influence recovery
potentially overwhelming the ability to critically evaluate the evaluations. Finally, recruiting procedures for the current
specifics of the recovery. research purposively limited the sample to SST users, to
Fourth, customers tend to feel greater time pressure when control for extraneous effects of nonuser attitudes toward self-
they are responsible for the completion of a self-service service adoption in general. However, the observed trends
transaction that has failed, but even that increased pressure toward encouraging customer adoption of SST for services
does not outweigh the importance of saving face in front of (Galer, 2014) merit understanding nonuser perceptions, as
other customers. When other customers are present and these customers may exhibit differences in expectation that
potentially witnessing the failure and recovery efforts, the would inform recovery strategy upon their expected future
affected customer would rather feel rushed than give the adoption of SST.
impression to fellow customers that he/she does not know what
he/she is doing. Ultimately, this research highlights the
importance of appreciating the social environment in which
References
SST failure occurs and allocating employee resources Ariely, D. and Levav, J. (2000), “Sequential choice in group
accordingly. Situational awareness and the flexibility to select setting: taking the road less traveled and less enjoyed”,
the appropriate recovery path allows customers to protect Journal of Consumer Research, Vol. 27 No. 4, pp. 279-290.
themselves from social embarrassment and ultimately produces Bambauer-Sachse, S. and Rabeson, L.E. (2015), “Service
a better self-service experience customized to the customer’s recovery for moderate and high involvement services”,
current situation. Journal of Services Marketing, Vol. 29 No. 5, pp. 331-343.
Finally, service providers and retailers need to be mindful of Blodgett, J.G., Granbois, D.H. and Walters, R.G. (1993),
how an SST area is designed, specifically as it relates to the “The effects of perceived justice on complainants’ negative
amount of attention that could be directed to a customer word-of-mouth behavior”, Journal of Retailing, Vol. 69 No. 4,
completing a transaction. Our results show that setting up a pp. 399-428.
self-service area where all self-service users are funneled into Bengtsson, S.L., Dolan, R.J. and Passingham, R.E. (2011),
one area draws further attention to customers performing a “Priming for self-esteem influences monitoring of one’s own
Self-service failure recovery Journal of Services Marketing
Joel E. Collier, Michael Breazeale and Allyn White

performance”, Social Cognitive and Affective Neuroscience, Heider, F. (1958), The Psychology of Interpersonal Relations,
Vol. 6 No. 4, pp. 417-425. Wiley, New York, NY.
Bodey, K. and Grace, D. (2006), “Segmenting service Holloway, B.B. and Sharon, E.B. (2003), “Service failure in
complainers and noncomplainers on the basis of consumer online retailing: a recovery opportunity”, Journal of Service
characteristics”, Journal of Services Marketing, Vol. 20 No. 3, Research, Vol. 6 No. 1, pp. 92-105.
pp. 178-187. Hu, H.-H. and Parsa, H.G. (2011), “Self-monitoring, dining
Bougie, R., Pieters, R. and Zeelenberg, M. (2003), “Angry companions and the usage of alternative currencies”,
customers don’t come back, they get back: the experience Journal of Product and Brand Management, Vol. 20 No. 3,
and behavioral intentions of anger and dissatisfaction in pp. 228-237.
services”, Journal of the Academy of Marketing Science, Vol. 31 Johnson, M.D., Anderson, E.W. and Fornell, C. (1995),
No. 4, pp. 377-393. “Rational and adaptive performance expectations in a
Choi, S. and Mattila, A.S. (2008), “Perceived controllability customer satisfaction framework”, Journal of Consumer
and service expectations: influences on customer reactions Research, Vol. 21 No. 4, pp. 128-140.
following service failure”, Journal of Business Research, Kelly, K.M. and Jones, W.H. (1997), “Assessment of
Vol. 61 No. 1, pp. 24-30. dispositional embarrassability”, Anxiety, Stress & Coping: An
Chung-Herrera, B.G., Gonzalez, G.R. and Hoffman, K.D. International Journal, Vol. 10 No. 4, pp. 307-333.
(2010), “When demographic differences exist: an analysis of Kim, N. and Ulgado, F.M. (2012), “The effect of on-the-spot
service failure and recovery among diverse participants”, versus delayed compensation: the moderating role of failure
Journal of Services Marketing, Vol. 24 No. 2, pp. 128-141. severity”, Journal of Services Marketing, Vol. 26 No. 3,
Downloaded by UMEA UNIVERSITY At 03:01 17 October 2017 (PT)

Collier, J.E. and Barnes, D.C. (2015), “Self-service delight: pp. 158-167.
exploring the hedonic aspects of self-service”, Journal of Kurt, D., Inman, J.J. and Argo, J.J. (2011), “The influence of
Business Research, Vol. 68 No. 5, pp. 986-993. friends on consumer spending: the role of agency-
Collier, J.E., Moore, R.S., Horky, A. and Moore, M.L. (2015), communion orientation and self-monitoring”, Journal of
“Why the little things matter: exploring situational influences Marketing Research, Vol. 48 No. 4, pp. 741-754.
on customers’ self-service technology decisions”, Journal of Latane, B. (1981), “The psychology of social impact”,
Business Research, Vol. 68 No. 3, pp. 703-710. American Psychologist, Vol. 36 No. 4, pp. 343-356.
Dabholkar, P.A. and Spaid, B.I. (2012), “Service failure and Lennox, R.D. and Wolfe, R.N. (1984), “Revision of the self-
recovery in using technology-based self-service: effects on monitoring scale”, Journal of Personality and Social
user attributions and satisfaction”, The Service Industries Psychology, Vol. 4 No. 6, pp. 1349-1364.
Journal, Vol. 32 No. 9, pp. 1415-1432. McDonell, J., Strom-Gottfried, K.J., Burton, D.L. and Yaffe, J.
Ding, D.X., Hu, J.P. and Sheng, O.R.L. (2011), (2006), “Behaviorism, social learning, and exchange
“E-SELFQUAL: a scale for measuring online self-service theory”, Contemporary Human Behavior Theory: A Critical
quality”, Journal of Business Research, Vol. 64 No. 5, Perspective for Social Work, Allyn & Bacon, Chicago, IL, Vol.
pp. 508-515. 349 No. 85.
Dixon, A.L., Spiro, R.L. and Jamil, M. (2001), “Successful Mattila, A. (2010), “Do women like options more than men?
and unsuccessful sales calls: measuring salespeople An examination in the context of service recovery”, Journal of
attributions and behavioral intentions”, Journal of Marketing, Services Marketing, Vol. 24 No. 7, pp. 499-508.
Vol. 65 No. 3, pp. 64-78. Mattila, A., Cho, W. and Ro, H. (2011), “The role of self-
Dong, B., Evans, K.E. and Zou, S. (2008), “The effects of service technologies in restoring justice”, Journal of Business
customer participation in co-created service recovery”, Research, Vol. 64 No. 4, pp. 348-355.
Journal of the Academy of Marketing Science, Vol. 36 No. 1, Meuter, M.L., Bitner, M.J., Ostrom, A.L. and Brown, S.W.
pp. 123-137. (2005), “Choosing among alternative service delivery
Economist.com. (2011), “Sometimes you just can’t take it modes: an investigation of customer trial of self-service
anymore”, available at: www.economist.com/blogs/gulliver/ technologies”, Journal of Marketing, Vol. 6 No. 2, pp. 61-83.
2011/11/airport-kiosks (accessed 23 January 2017). Miller, R.S. (1996), Embarrassment: Poise and Peril in Everyday
Forbes, L.P. (2008), “When something goes wrong and no one Life, Guilford Press, New York, NY.
is around: non-internet self-service technology failure and Noone, B.M. (2012), “Overcompensating for severe service
recovery”, Journal of Services Marketing, Vol. 22 Nos 4/5, failure: perceived fairness and effect on negative word-of-
pp. 316-327. mouth intent”, Journal of Services Marketing, Vol. 26 No. 5,
Friedenberg, M. (2014), “Retail must reinvent itself for the pp. 342-351.
omnichannel future”, available at: www.cio.com/article/ Oh, H., Miyoung, J. and Seyhmus, B. (2013), “Tourists’
749879/Retail_Must_Reinvent_Itself_for_the_Omnichannel_ adoption of self-service technologies at resort hotels”, Journal
Future (accessed 16 April 2014). of Business Research, Vol. 66 No. 6, pp. 692-699.
Galer, S. (2014), “Three steps to boost employee engagement”, Oliver, R.L. and Swan, J.E. (1989), “Consumer perceptions of
available at: www.forbes.com/sites/sap/2014/04/16/three-steps- interpersonal equity and satisfaction in transactions: a field
to-boost-employee-engagement/ (accessed 16 April 2014). survey approach”, Journal of Marketing, Vol. 53 No. 2, pp. 21-35.
Graeff, T. (1996), “Image congruence effects on product Power, B. (2014), “Redbox breathes life into movie rental
evaluations: the role of self-monitoring and public/private scene”, available at: http://thechronicleherald.ca/business/
consumption”, Psychology and Marketing, Vol. 13 No. 5, 1200471-redbox-chain-breathes-new-life-into-movie-rental-
pp. 481-499. scene) (accessed 15 April 2014).
Self-service failure recovery Journal of Services Marketing
Joel E. Collier, Michael Breazeale and Allyn White

Puccinelli, N.M., Deshpande, R. and Isen, A.M. (2007), Zhou, Y., Huang, M., Tsang, A. and Zhou, N. (2013),
“Should I stay or should I go? Mood congruity, self- “Recovery strategy for group service failures”, European
monitoring, and retail context preference”, Journal of Journal of Marketing, Vol. 47 No. 8, pp. 1133-1156.
Business Research, Vol. 60 No. 6, pp. 640-648. Zhu, Z., Nakata, C., Sivakumar, K. and Grewal, D. (2007),
Richins, M.L. (1983), “Negative word-of-mouth by “Self-service technology effectiveness: the role of design
dissatisfied customers: a pilot study”, Journal of Marketing, features and individual traits”, Journal of the Academy of
Vol. 47 No. 1, pp. 68-78. Marketing Science, Vol. 35 No. 4, pp. 492-506.
Robertson, N. and Shaw, R.N. (2005), “Conceptualizing the Zhu, Z., Nakata, C., Sivakumar, K. and Grewal, D. (2013), “Fix
influence of the self-service technology context on it or leave it? Customer recovery from self-service technology
consumer voice”, Services Marketing Quarterly, Vol. 27 failures”, Journal of Retailing, Vol. 89 No. 1, pp. 15-29.
No. 2, pp. 33-50.
Rose, P. and DeJesus, S.P. (2007), “A model of motivated
cognition to account for the link between self-monitoring
Further reading
and materialism”, Psychology and Marketing, Vol. 24 No. 2, Collier, J.E. and Kimes, S.E. (2013), “Only if it is convenient:
pp. 93-115. understanding how convenience influences self-service
Schoefer, K. and Diamantopoulos, A. (2008), “The role of technology evaluation”, Journal of Service Research, Vol. 16
emotions in translating perceptions of (in)justice into No. 1, pp. 39-51.
postcomplaint behavioral responses”, Journal of Service
Research, Vol. 11 No. 1, pp. 91-103. Appendix 1
Downloaded by UMEA UNIVERSITY At 03:01 17 October 2017 (PT)

Schunk, D.H. and Zimmerman, B.J. (1998), Self-Regulated


Learning: From Teaching to Self-Reflective Practice, Guilford
Press, New York, NY. Machine at fault – employee recovers – isolated
Sharp, B. (2014), “Are self-service machines in our On Friday night, you and a friend decide to go to dinner and
supermarkets really the way forward? The Independent”, see a highly anticipated movie that has just been released.
available at: www.independent.co.uk/news/business/analysis- You decide to go and get your movie tickets before getting
and-features/are-selfservice-machines-in-our-supermarkets- dinner to make sure you get a seat. At the movie theater, you
really-the-way-forward-9198140.html) (accessed 16 April approach a self-service ticket kiosk to purchase your tickets
2014). for a show later in the night. After finding your movie and
Smith, A.K., Bolton, R.N. and Wagner, J. (1999), “A model of the desired time of the show, you move to the checkout
customer satisfaction with service encounters involving screen. You scan your credit card to pay for the movie tick-
failure and recovery”, Journal of Marketing Research, Vol. 36 ets but before the tickets are dispensed the kiosk interface
No. 3, pp. 356-372. screen freezes. Noticing no other customers around you,
Snellman, K. and Vihtkari, T. (2003), “Customer complaining you motion for an employee to come over and help correct
behaviour in technology-based service encounters”, the problem. The employee apologizes and informs you that
International Journal of Service Industry Management, Vol. 14 sometimes the kiosk locks up when there is heavy use and
No. 2, pp. 217-231. simply needs to be reset. After cancelling out your previous
Snyder, M. (1974), “Self-monitoring of expressive behavior”, attempt to purchase the tickets, the employee resets the
Journal of Personality and Social Psychology, Vol. 30 No. 4, kiosks and proceeds to the main menu. With no other cus-
pp. 526-537. tomers waiting to use the kiosk, the employee finds your
Snyder, M. (1979), “Self-monitoring process”, in Berkowitz, desired movie and show time and asks you to swipe your
L. (Ed.), Advances in Experimental Social Psychology, Vol. 12, credit card again. This time the kiosk presents a confirma-
Academic Press, New York, NY, pp. 85-128. tion screen and the tickets for the future show are dispensed.
Spears, N. and Singh, S.N. (2004), “Measuring attitude toward You take your tickets and proceed to dinner before your
the brand and purchase intentions”, Journal of Current Issues movie.
and Research in Advertising, Vol. 26 No. 2, pp. 53-66.
Strombeck, S.D. and Wakefield, K.L. (2008), “Situational
Machine at fault – customer recovers – others
influences on service quality evaluations”, Journal of Services
present
Marketing, Vol. 22 Nos 4/5, pp. 409-419.
Ward, J.C. and Ostrom, A.L. (2006), “Complaining to the On Friday night, you and a friend decide to go to dinner and
masses: the role of protest-framing in customer-created see a highly anticipated movie that has just been released. You
complaint web sites”, Journal of Consumer Research, Vol. 33 decide to go and get your movie tickets before getting dinner
No. 2, pp. 220-230. to make sure you get a seat. At the movie theater, you
White, A., Breazeale, M. and Collier, J.E. (2012), “The effects approach a self-service ticket kiosk to purchase your tickets for
of perceived fairness on customer responses to forced SST a show later in the night. After finding your movie and the
migration”, Journal of Retailing, Vol. 88 No. 2, pp. 250-261. desired time of the show, you move to the checkout screen.
Yen, H.R., Gwinner, K.P. and Su, W. (2004), “The impact of You scan your credit card to pay for the movie tickets but
customer participation and service expecations on locus before the tickets are dispensed the kiosk interface screen
attributions following service failure”, International Journal of freezes. Noticing the growing line of customers behind you,
Service Industry Management, Vol. 15 No. 1, pp. 7-26. you motion for an employee to come over and help correct the
Self-service failure recovery Journal of Services Marketing
Joel E. Collier, Michael Breazeale and Allyn White

problem. The employee apologizes and informs you that Quad SST set up – employee recovers
sometimes the kiosk locks up when there is heavy use and sim-
You arrive at the grocery store, complete your shopping and
ply needs to be reset. After cancelling out your previous
decide to use the self-service checkout. You proceed to a single
attempt to purchase the tickets, the employee resets the kiosks.
line of waiting customers that feeds four self-scanning stations
With a line of customers now waiting to use the kiosk, you find
placed in a quad design. You wait in line and watch the other
the desired movie and show time again and swipe your credit
customers in the quad scanning their groceries until a self-scan
card. This time the kiosk presents a confirmation screen and
machine becomes available. You then approach the available
the tickets for the future show are dispensed. You take your
self-scan machine and start the check-out process. You finish
tickets and proceed to dinner before your movie. scanning most of your items when the self-service screen
freezes and does not allow you to scan the next item. Noticing
Customer at fault – employee recovers – others two customers waiting in line to use the self-scan machines,
present you motion for an attendant to come over and help correct the
problem. The attendant apologizes and informs you that
On Friday night, you and a friend decide to go to dinner and
sometimes the kiosk locks up and simply needs to be reset.
see a highly anticipated movie that has just been released.
After clearing out your scanned order, the attendant resets the
You decide to go and get your movie tickets before getting
kiosk and then proceeds to rescan all of your items for you and
dinner to make sure you get a seat. At the movie theater, you
completes the checkout process all the way to the payment
approach a self-service ticket kiosk to purchase your tickets
screen. You pay, take your groceries and leave the store.
for a show later in the night. After finding your movie, you
Downloaded by UMEA UNIVERSITY At 03:01 17 October 2017 (PT)

move to the checkout screen. You scan your credit card to


Single line set up – customer recovers
pay for the movie tickets, but before the tickets are
dispensed you notice that you purchased tickets for the You arrive at the grocery store, complete your shopping and
current show time instead of the desired future show time. decide to use the self-service checkout. You proceed to the
Noticing the growing line of customers behind you, you self-scanning station. The self-service area has four self-
motion for an employee to come over and help correct the scanning machines with a line that forms behind each
problem. The employee apologizes that you had a problem machine. You select a line and watch the customer in front
and states that he will need to reset your transaction. After of you scan their groceries. After that customer finishes, you
cancelling out your previous attempt to purchase the tickets, approach the self-service technology and start the check-out
the employee proceeds to the main menu. With a line of process. You finish scanning most of your items when the
customers waiting to use the kiosk behind you, the employee self-service screen freezes and does not allow you to scan the
finds your desired movie and show time and asks you to next item. Noticing two customers waiting to use the self-
swipe your credit card again. This time the kiosk presents a scan machine you are using, you motion for an attendant to
confirmation screen and the tickets for the future show are come over and help correct the problem. The attendant
dispensed. You take your tickets and proceed to dinner apologizes and informs you that sometimes the kiosk locks
before your movie. up and simply needs to be reset. After clearing out your
scanned order, the attendant resets the kiosk and proceeds
to the main menu where you can rescan your groceries.
Customer at fault – customer recovers – isolated
Once the menu is available, the attendant returns to the
On Friday night, you and a friend decide to go to dinner and employee station and allows you to complete the remaining
see a highly anticipated movie that has just been released. You transaction. After rescanning your entire order, you finish
decide to go and get your movie tickets before getting dinner the checkout process, pay, take your groceries and leave the
to make sure you get a seat. At the movie theater, you store.
approach a self-service ticket kiosk to purchase your tickets
for a show later in the night. After finding your movie, you Appendix 2
move to the checkout screen. You scan your credit card to
pay for the movie tickets, but before the tickets are
Scale items
dispensed you notice that you purchased tickets for the
current show time instead of the desired future show time. Satisfaction
Noticing no other customers around you, you motion for an How satisfied would you be with the way in which the problem
employee to come over and help correct the problem. The with the self-service kiosk was handled:
employee apologizes that you had a problem and states that  very dissatisfied-very satisfied;
he will need to reset your transaction. After cancelling out  unhappy-happy;
your previous attempt to purchase the tickets, the employee  frustrated-pleased; and
proceeds to the main menu. With no other customers  very unfavorable-very favorable.
waiting to use the kiosk, you find the desired movie and
future show time and swipe your credit card. This time the
kiosk presents a confirmation screen and the tickets for the Switching to a full service option
future show are dispensed. You take your tickets and How likely would you be to switch from this self-service
proceed to dinner before your movie. method of buying tickets to a full-service option:
Self-service failure recovery Journal of Services Marketing
Joel E. Collier, Michael Breazeale and Allyn White

 definitely would not-definitely would; Susceptibility to embarrassment


 probably not- probably; and  I feel uncomfortable in a group of people.
 not likely- very likely.  I probably care too much about how I come across to
others.
 I feel humiliated if I make a mistake in front of a group.
Negative word of mouth  I am very much afraid of making a mistake in public.
Based on the way in which the self-service problem was  I often feel unsure of myself.
handled in the scenario:  It is unsettling to be the center of attention.
 I would likely say negative things about this experience.  I often worry about looking stupid.
 I would not recommend using this self-service option to  I am concerned about what others think of me.
my friends and relatives.  I worry about making a fool of myself.
 I would discourage friends and relatives from using this  What other people think of me is very important.
self-service option at this movie theater.

Self-monitoring
Time pressure  In social situations, I have the ability to alter my behavior
 I would feel pressed for time because others were waiting. if I feel something else is called for.
 I am often able to read other peoples’ true emotions from
Downloaded by UMEA UNIVERSITY At 03:01 17 October 2017 (PT)

 I would feel rushed because others wanted to use the


checkout option. their eyes.
 I would feel the need to hurry because I would not want to
 I have found that I can adjust my behavior to meet the
keep other customers waiting. requirements of any situation I find myself in.
 I would feel that the problem I experienced negatively
impacted other customers. Corresponding author
 I would feel that the problem I experienced caused a Joel E. Collier can be contacted at: jcollier@business.
problem for other customers. msstate.edu

For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: permissions@emeraldinsight.com

You might also like