The Influence of Self-Service Technology on Customer Word-of-Mouth in Retail

In light of the vital role of word-of-mouth (WOM) in ensuring the success of retail establishments, this study investigates the influence of self-service technology (SST) on customer WOM. Data collected from 361 respondents were used to construct a structural equation model. The results indicate that perceived ease of use, usefulness, and perceived control have varying effects on both SST satisfaction and WOM intention. These findings are of practical significance for store managers seeking to harness SST for enhancing positive word-of-mouth (WOM) and strengthening retail success.


Introduction
Technology has radically changed the business landscape (Gallaugher, 2010;Schön, 2017).Among the range of new technologies developed within the last several decades, self-service technology (SST), or -technological interfaces that enable customers to produce a service independent of direct service employee involvement‖ (Meuter et al., 2000, p. 50), has attracted a great deal of attention from marketing scholars and practitioners (Kelly et al., 2010).Notable examples include automated teller machines (ATMs), pay-at-the-pump machines, and telephone banking or, more recently, automated hotel and grocery store checkouts, airline check-in systems for e-ticket holders, in-store kiosks for product information, web-based purchasing, and Internet transactions (Orel & Kara, 2014;Yang & Klassen, 2008).By replacing human-to-human contact with human-machine interaction (Parasuraman, 2000), SST has effectively reshaped consumers' views of the conception, development, and delivery of services (Huang & Zhou, 2017;Meuter et al., 2005).
If managed properly, SST allows retailers to serve customers in a more efficient way by offering consistent services (Gelderman et al., 2011;Weijters et al., 2007;Yang & Klassen, 2008).SST offers more choices to customers, which can reduce anxiety associated with personnel judgement (Bitner, 2001; Griffy- Brown et al., 2008;Meuter et al., 2000).They also allow retailers to better meet the needs of specific groups of customers, including those who seek more innovative services (Azhar & Lin, 2017;Dean, 2008).Overall, the use of SST makes shopping faster, easier, and more controllable for customers, thus creating an overall more enjoyable retail experience and, by extension, increasing customer satisfaction and loyalty (Bitner et al., 2002;Dabholkar et al., 2003;Meuter et al., 2000, Orel & Kara, 2014).In addition to adding value for customers, SST can empower both employees and customers by offering place and time convenience (Hsieh, 2005;Li & Gregor, 2011;Yang & Klassen, 2008).Further, labor, real estate, and communication costs can be reduced for retailers through the deployment of SST (Canbase, 2009;Orel & Kara, 2014).
To date, the benefits of SST for customers and retailers have been suggested, but limited attention has been paid to the influence of SST deployment on customer word-of-mouth (WOM).The present research sought to address this gap in the literature by using the theory of planned behavior (TPB) and technological acceptance model (TAM) as a framework to predict customer satisfaction and WOM in retail contexts.

Theoretical Framework
The Theory of Planned Behavior (TPB) underscores the significance of both intention and behavioral control in shaping behavior, as evidenced by Ajzen (2002), Armitage and Conner (2001), and Sheeran (2002).TPB finds application in a diverse array of domains, including ethical decision-making (Chang, 1998), health management (McEachan et al., 2011), driving behaviors (Elliott et al., 2007), and technology adoption, such as online grocery shopping (Hansen et al., 2004).In this model, human attitude is considered to be the result of behavioral belief and the evaluation of outcomes, as proposed by Mathieson in 1991.Despite the prevalent use of attitude and subjective norms as predicted outcomes in these models, the present study specifically centers on the outcomes related to the utilization of Self-Service Technologies (SSTs), such as satisfaction and Word of Mouth.In parallel, the Technology Acceptance Model (TAM) is widely employed for forecasting technology acceptance, emphasizing the pivotal roles of perceived ease of use and usefulness (Davis, 1986).TAM's versatility is evident across various contexts, with robust associations established between its constructs and technology adoption (Guriting & Ndubisi, 2006).Moreover, TAM's insights extend to attitudes toward technology, including applications in the realm of retail self-service technology (Childers et al., 2001).This study posits that the foundational concepts shared by TPB and TAM-perceived control, ease of use, and usefulness-exert an influential impact on customer word-of-mouth.This influence is mediated through satisfaction derived from the utilization of self-service technology (SST).

Literature Review Word-of-Mouth
Positive customer WOM, or -oral, person-to-person communication between a perceived non-commercial communicator and a receiver concerning a brand, a product, or a service offered for sale‖ (Arndt, 1967, p. 190), is a crucial driver of customer loyalty (Garnefeld et al., 2011).It thereby reduces consumer resistance, switching behavior (Wangenheim & Bayón, 2007;Money, 2004), and advertising costs (Trusov et al., 2009) while enhancing new customer acquisition (Trusov et al., 2009), customer purchases, and overall sales (Floyd et al., 2014;Libai et a., 2013;You et al., 2015).The long-term value of a retailer is thus enhanced by positive WOM (Villanueva et al., 2008).Various studies have delved into the effects of negative word of mouth (NWOM).Azemi, Ozuem and Howell (2020) identified three distinct types of customers who engage in online NWOM: tolerable, rigorous, and confrontational online NWOM customers.In a series of experiments, Chan and Cui (2011) discovered that attribute-based NWOM adversely affects dissatisfied consumers, while experience-based NWOM has a positive influence during the post-consumption stage.Moreover, Sandes and Urdan (2013) conducted an exploratory experimental study on electronic word of mouth, revealing that negative comments significantly impact brand image.Additionally, research has shown that the impact of positive word of mouth generally outweighs that of negative word of mouth on brand purchase decisions (East et al., 2008) (Brown et al., 2005), and confusion proneness (Walsh & Mitchell, 2010).Another important antecedent to WOM, however, is customer satisfaction.

Satisfaction with SST and Word-of-Mouth
According to the disconfirmation theory, customer satisfaction is a function of customer expectations regarding products and/or services (Oliver, 1997).When the performance of a product or service meets or exceeds customer expectations, the customer will feel satisfied; if it fails to meet his or her expectations, he or she will feel dissatisfied (Bhattacherjee, 2001;Chen, 2005).The disconfirmation theory posits that consumers' desires, experiences, and expectations interact with product and service performance to affect consumer attitudes (Lee & Joshi, 2006).Satisfied customers exhibit greater loyalty, less tendency to switch (Chen, 2005;Yang & Peterson, 2004), and higher intention to patronize (Fernandes & Pedroso, 2017;Marzocchi & Zammit, 2006).
Given that the use of SST is an aspect of the consumer shopping experience, it was anticipated that satisfaction with SST would affect general customer satisfaction and eventually WOM intention (De Matos & Rossi, 2008;Ha & Im, 2012;Ladhari, 2007, Wangenheim & Bayón, 2007).Thus, it was hypothesized that:

H1: Satisfaction with SST is positively associated with WOM intention.
A diverse range of factors affect customer satisfaction with SST, including but not limited to customer innovativeness and optimism (Abdullah, 2012;Lin & Hsieh, 2006), fun and enjoyment (Fernandes & Pedroso, 2017), product/service price and location (Skoglan & Siguaw, 2004), site design and payment methods (Cho & Park, 2001), waiting times (Marzocchi & Zammit, 2006), convenience (Meuter et al., 2000), human touch (Makarem et al., 2009), process failure, technology failure, and forced use of SST (Jamal, 2004;Liu, 2012).In the present study, it was anticipated that the variables of perceived control, ease of use, and usefulness would affect satisfaction with SST.
Control, Ease of Use, Usefulness, and Satisfaction with SST Perceived control.Perceived control can be used to measure behavioural control in SST contexts.It is a crucial element in the customer's appraisal of the retail experience as it drives intention to use technologies (Collier & Sherrell, 2010;Kuan et al., 2011;Zeithaml et al., 2002).For example, SST with higher perceived control has lower perceived risk and higher perceived value (Zhu et al., 2007).Dabholkar (1996) defined perceived control as the amount of control a customer feels he or she has over a process and its results.SST is perceived as more controllable when customers are able to determine and design the service for themselves and flexibly change their decisions (Lee & Allaway, 2002).In a retail context, perceived control is an intrinsic feeling of independence (Meuter et al., 2000) that can enhance consumers' evaluations of the self-service option and increase their motivation to use SST (Collier & Sherrell, 2010;Dabholkar, 1996).
Ease of use.Customers tend to feel more satisfied when SST is easy to use (Meuter et al., 2000).When an innovation is easy to understand or use, it can be said to possess perceived ease of use (Zeithaml et a., 2002).Perceived ease of use is the degree to which a person believes a particular technology or service can be used effortlessly (Davis, 1986).Perceived ease of use has positive effects on attitudes toward SST (Kim et  Usefulness.When technology is easy to use, customers perceive it as useful because they do not have to spend time and effort figuring out how to operate or employ it and can thus complete their tasks more efficiently (Bruner & Kumar, 2005).Perceived usefulness is defined as the subjective probability of a technology helping a user complete a task ( Within TPB and TAM frameworks, perceived control, ease of use, and usefulness are commonly identified as positive features of SST.These features not only drive the adoption and use of SST, but also satisfaction with it (Bhattacherjee, 2001 H2: Perceived control is positively associated with satisfaction with SST.H3: Ease of use is positively associated with satisfaction with SST.H4: Usefulness is positively associated with satisfaction with SST. If H1 to H4 were established, it was theorized that satisfaction with SST would mediate the relationship between perceived control, ease of use, usefulness, and WOM intention.The definition of mediating effect employed was drawn from the research of Judd and Kenny (2010), MacKinnon and Luecken (2008), and Zhao, Lynch and Chen (2010).It was thus also hypothesized that: H5: Satisfaction with SST significantly mediates the relationship between perceived control and WOM intention.H6: Satisfaction with SST significantly mediates the relationship between ease of use and WOM intention.H7: Satisfaction with SST significantly mediates the relationship between usefulness and WOM intention.
To further delineate the anticipated relationships between variables, the conceptual model is illustrated in Figure 1.

Method Sample and Procedure
A descriptive cross-sectional research design was implemented in the current study as its primary aim was to discover which factors affect consumer SST use in retail contexts (Burns & Bush, 2006).A questionnaire was administered via Qualtrics using an online panel.Australian shoppers over 18 years old who had used supermarket self-checkout systems in the previous 12 months were invited to fill out the questionnaire.Prior to the study, the questionnaire was evaluated by 10 Australian shoppers, three SST scholars from Australian universities, and one Australian supermarket manager.A total of 361 participants (N = 361) completed the questionnaire with no missing values.
Perceived control.Conceptualizing perceived control as a uni-dimensional construct, Dabholkar (1996), Yen and Gwinner (2003), and Zhu (2002) defined it as an individual's desire to exhibit mastery over his or her environment.In the current study, perceived control was measured using three items.
Ease of use.In an SST context, ease of use can be defined as the degree of complication and confusion that users experienced during technology use (Dabholkar & Bagozzi, 2002;Zhao et al., 2008).The construct of ease of use in the current study was adapted from Dabholkar and Bagozzi (2002) and Zhao, Mattila and Tao (2008).Three items were used to measure ease of use.
Usefulness.The usefulness of SST in a retail context can be defined as the efficiency and speed of shopping performed via the technology in question (e.g., self-scanning checkout) (Weijterset et al., 2007).The measure of perceived usefulness, as conceived by Weijters, Rangarajan, Falk and Schillewaert (2007), was implemented in the present study.Three measurement items were used to measure usefulness.
Satisfaction with SST.To assess overall satisfaction with SST in a retail context, three items were adopted from Zhao et al., (2008)'s study of satisfaction with library self-checkout systems.
WOM intention.Ha and Im (2012)'s three-item measurement of WOM intention was adopted in this study to measure the likelihood of customers recommending stores equipped with SST to others.

Results
The constructs of perceived control, ease of use, usefulness; satisfaction with SST; and WOM intention, each containing three items, were first purified using confirmatory factor analysis (CFA).Perceived ease of use had factor loadings of .80-.95, usefulness had factor loadings of .76-.94, perceived control had factor loadings of .78-.88, satisfaction with SST had factor loadings of .92-.94, and WOM intention had factor loadings of .91-.98.The composite reliability of constructs ranged from .85 to .96(Appendix A), which exceeded the acceptable level of .70 (Fornell & Larcker, 1981).Thus, the reliabilities were in the acceptable range and all measures had sufficient consistency.
Hypotheses 1 to 4 were tested using linear regression analysis.Variance inflation factor (VIF) analysis was conducted to ensure the assumption of multicollinearity.VIF values (perceived control = 1.56, ease of use = 1.67, usefulness = 1.90, satisfaction with SST = 2.36) were below the cut off value of 10 (Aiken et al., 1991); therefore, the assumption of multicollinearity was confirmed.Perceived control, ease of use, usefulness, and satisfaction with SST were entered into a linear regression equation (Model 1) and accounted for a significant amount of variance in WOM intention: R ² = .59,F (4,356) = 128.21(p < .001)(Table 1).

Discussion and Implications
The current study implemented TPB and TAM as frameworks to analyze the relationships between perceived control, ease of use, and usefulness; customer satisfaction with SST; and customer WOM intention.As hypothesized, customers who exhibited higher satisfaction with SST had higher intention to recommend stores to others.This finding is in accordance with the results of research conducted by De Matos and Rossi (2008), Ha and Im (2012), Ladhari (2007), and Wangenheim and Bayón (2007).
Additionally, customers who perceived SST as more controllable, easier to use, and more useful were more likely to feel satisfied with SST, which is consistent with hypotheses formulated based on evidence provided by Bhattacherjee (2001), Chen and Chen (2009), Dabholkar and Bogozzi (2002), Fernandes and Pedroso (2017), Wang (2012), and Yen and Gwinner (2003).As predicted, perceived control, ease of use, and usefulness affected customers' intention to recommend stores to others.This relationship was mediated by satisfaction with SST, which is consistent with predictions regarding the mediation process as defined by Judd and Kenny (2010), MacKinnon and Luecken (2008), and Zhao, Lynch and Chen (2010).As a result, the current study confirms that SST deployment influences customer WOM through satisfaction with SST, which contributes new knowledge to the breadth of SST literature and suggests managerial implications.
In previous studies, TPB and TAM have been predominantly used to predict satisfaction with, attitudes toward, adoption and use of technology.The current study confirms that TPB and TAM can also predict customer WOM intention.However, an additional theoretical component must be added to fully explain the phenomenon or phenomena under observation; in the present study, disconfirmation theory was employed.
The results indicated that perceived control, ease of use, and usefulness influence WOM.This suggests SST deployment has an impact on consumer behavior that extends beyond satisfaction with, attitudes toward, and the adoption and use of the technologies in question.
The current study thus provides insights for store managers or other actors making decisions regarding SST deployment, offering suggestions to maximize the benefits of SST in retail environments.To enhance positive customer WOM, store managers can prioritize the usefulness of self-checkout systems as usefulness was found to be the most important factor in generating customer WOM.An important priority in the design of self-checkouts is the speed of transactions: redundant instructions and complicated procedures should be avoided.In addition to design, stores can limit self-service transactions to customers purchasing fewer items to facilitate speedy transactions and to shorten lines.
To further foster customer WOM, perceived control and ease of use must also be considered.Managers should prioritize user friendliness, focusing on the option and customization functions included in self-checkout system design.If design alterations are not feasible, managers can consider alternative ways to enhance customer satisfaction with SST.Reducing failure incidents and providing customers with autonomy in their decision to use or not use SST are two such ways to enhance customer WOM (Jamal, 2004;Liu, 2012).

Limitations and Future Research
While the present study undeniably offers valuable insights for retail managers navigating the implementation of Self-Service Technology (SST), it is crucial to recognize its inherent limitations.The primary constraint lies in its exclusive focus on Australian supermarket customers, which naturally restricts the generalizability of its findings to a broader international or multi-sector context.Nonetheless, the research methodology employed in this project displays flexibility, allowing its application in various settings, including small to medium-sized retailers and diverse geographical locations, facilitating exploration into nuanced variations in consumer behaviour when engaging with SSTs (Arnould & Thompson, 2005;De Mooij, 2010).The study's research design, using a cross-sectional approach, offers valuable snapshots of consumer behaviour but comes with limitations in terms of internal validity (Howitt & Cramer, 2010).Therefore, there is a compelling case for conducting longitudinal and experimental studies in future research endeavours to better understand causal relationships between variables and enhance depth of insight (Goodwin, 2009;Howitt & Cramer, 2010).Additionally, the study didn't investigate the potential influence of situational factors like queue length and waiting time, nor did it explore moderating effects related to variables such as age and gender.These unexplored dimensions could significantly contribute to a more comprehensive understanding of how situational elements interact with and moderate SST usage.Furthermore, it's essential to note that the data collected primarily represents experiences of customers frequenting large retailers, particularly in the expansive Australian supermarket landscape.As a result, this data may not fully represent the behaviours and preferences of customers in small or medium-sized retailers.
To gain a holistic understanding of consumer SST usage, future research should explicitly address dynamics within smaller retailers, recognizing potential differences in SST adoption and use in such contexts.

Conclusion
Self-checkouts not only offer convenience to customers and reduce costs for retailers, but they also enhance customer satisfaction and increase the likelihood of positive word-of-mouth when managed properly by retailers.However, the current research is subject to various limitations.