Emotional and Cognitive Determinants of Retail Salespersons' Emotional Labor and Adaptive Selling Behavior

Abstract

Purpose: The role of salespersons' emotions in effective selling behavior garners attention among scholars and practitioners. Previous studies have investigated the effects of emotional intelligence and emotional labor on sales success separately. However, to understand the whole process, the relationships among salespersons' cognition, emotions, and behaviors should be considered simultaneously. Accordingly, we uniquely examined how salespersons' emotional intelligence (emotional antecedent) and customer orientation (cognitive antecedent) influence their emotional labor (deep acting vs. surface acting), adaptive selling behavior, and the selling results in the retail environment. Research design, data, and methodology: To improve methodological rigor, we used the dyadic approach. We measured 182 salespersons' emotional intelligence, customer orientation, and emotional labor, and 364 customers assessed the salespersons' adaptive selling behavior and selling results in the insurance and duty-free department retailing sectors. Result: The findings suggest that salespersons' customer orientation and emotional intelligence relate to deep-acting of emotional labor, affecting their adaptive selling behavior and relationship quality with customers. Conclusions: As for managerial implications, sales managers may well consider emotional intelligence levels when selecting salespersons in the retail industry. Additionally, practical training programs are required to cultivate customer orientation, emotional intelligence, and deep acting while performing emotional labor.

1. Introduction

In early studies on marketing, researchers argued that the involvement of emotional factors should be excluded in the process of exchange between firms and customers. This assertion was based on the premise that human beings are purely rational creatures who can make judgments based on reason without being disturbed by irrational emotions. However, more recently, scholars have realized that cognition and emotions are inseparable (e.g., Diaz & Ruiz, 2002; Smith & Bolton, 2002) and in correspondence with this recognition, emotions have become an essential topic in marketing. Especially, the management of salespersons’ emotions has been rising as a critical factor influencing sales performance. As well as the product or service itself, salespersons' relationships with customers are considered as the antecedents of customer satisfaction (CS) and loyalty (Crosby & Stephens, 1987). While salespersons are in contact with their customers, their emotional regulation can critically affect the process of building positive relationships with customers and increasing personal selling that can adapt to specific customer needs.

Nevertheless, most marketing research has paid attention to benefits, compensations, or communication skills as important means to improve salespersons' job satisfaction and selling behavior. Though some research examined the role of emotions, most sales research has not focused on the integrative relationships among cognitive and emotional determinants of a salesperson’s selling behavior and selling results. For example, previous studies have investigated the effects of emotional intelligence (EI), customer orientation (CO), emotional labor (EL), and sales success separately.¹ However, to catch a glimpse of the whole process, the relationships among salespersons’ cognition, emotions, and behaviors should be considered simultaneously with the influences they have on sales performance.

Hence, the main goal of this research is to connect the notions of emotional intelligence and customer orientation with emotional labor to fully understand their effects on selling behavior. More concretely, we posit that salespersons with higher levels of emotional intelligence and customer orientation are more probable to perform deep acting (a specific type of emotional labor), which leads to a demonstration of adaptive selling behavior (ASB) resulting in improvement of relationship quality (RQ) with customers and increase of repurchase intentions (RI).

Though prior psychological studies have discovered that emotional intelligence generally affects performance after regulating the effect of personality traits and cognitive intelligence (Janovics & Christiansen, 2001; Lopes et al., 2006), the limitations of those studies are that they did not sufficiently consider the context of specific occupations. Because salespersons frequently interact with customers in the retail environment and emotional processes play a critical role in influencing customer results in the high-involving retail purchase occasions, the retailing environments such as insurance selling and duty-free department stores were selected as venues of empirical testing in this study.

Moreover, to improve methodological rigor, we used the dyadic approach and collected the data from different sources. Data on the emotional regulation processes (emotional intelligence & emotional labor) and customer orientation were collected from salespersons, whereas data on the salesperson’s adaptive selling behavior and the selling results were collected from their customers. This approach has been rarely taken in research on salespersons’ emotions and cognition (e.g., Chen & Jaramillo, 2014 for exception). The following section delineates the conceptualization and hypotheses on the relationships between emotional intelligence, customer orientation, emotional labor, adaptive selling behavior, and the related consequences. Next, the research method and results of the hypothesis testing are discussed. Finally, implications from practical and theoretical standpoints, limitations of the research, and directions for further studies are presented.

2. Literature Review and Hypotheses Setting

2.1. Emotional Intelligence and Emotional Labor

The notion of emotional intelligence has been introduced by Goleman, in his book, ‘Emotional Intelligence.’ Goleman (1995, 1998) proposed that emotional intelligence can predict performance better than IQ in many life tasks. Although this claim immediately attracted people's attention, it was based on anecdotes and reports rather than rigorous scientific research. Since the introduction of emotional intelligence, an intensive debate has occurred on the definition, instrument, and incremental validity of the emotional intelligence construct. There are researchers who have expressed doubt regarding the construct validity of the concept itself (e.g., Bar-On, 2004; Landy, 2005; Matthews et al., 2004), while others criticized those scholars for being so accurate in different approaches (e.g., Ashkanasy & Daus, 2005; Oatley, 2004).

Mayer and Salovey (1997), who are widely assessed as the creator of the construct, argued that the notion of emotional intelligence should be limited to mental capacities related to the interaction between cognitions and emotions. They argued that emotional intelligence is the capability to evaluate and control one’s own and others’ emotions and the capability to use emotions to plan and achieve fulfillment in one’s life (Bar-On, 2004, Bar-On et al., 2006; Mayer et al., 2008). Taking a similar approach, Kidwell et al. (2011) also defined emotional intelligence as the ability to obtain and utilize information about one’s own or others’ emotions for the purpose of producing positive outcomes in social situations. In sum, emotional intelligence was defined as the ability to control one’s own and others’ emotions (Goleman, 1998). In addition, self-management, relationship management, self-awareness, and social awareness were included as components of emotional intelligence including four constructive abilities. In this study, based on the emotional intelligence model of Mayer and Salovey (1997), the variables developed by Wong and Law (2002) were used, such as regulation of emotion, use of emotion, self-emotion appraisal, and others’ emotion appraisal. In other words, emotional intelligence was categorized into 4 factors and 16 indicators.

Several occupational and social psychological studies discovered a positive relationship between emotional intelligence and personal accomplishment (Côté et al., 2010; Day et al., 2005; Goleman, 1998; Slaski & Cartwright, 2002, 2003; Tsaousis & Nikolaou, 2005). Consistent with these studies, front-line retail service personnel and salespeople who face customers in sales and marketing are better able to recognize and use the positive or negative arise during interactions with customers (Kidwell et al., 2011). Salespeople must be aware of the emotional state of their customers and manage their emotions to improve their performance so that their customers experience high levels of satisfaction (Chen & Jaramillo, 2014; Ladhari, 2007).

On the other side, it is not easy for salespersons to consistently keep positive emotions at work, although they understand the importance of expressing desirable emotions to provide a positive experience for customers. They often find themselves in a situation where they should express emotions that are different from the actual emotions they feel. Acceptable expressions of emotions are determined in the form of display rules, which define the scope, density, and duration of emotions that can be displayed within the work environment (Ashforth & Humphrey, 1993). Rafaeli and Sutton (1991) proposed that display rules are usually social, occupational, and organizational norms. Thus, they are determined by the cultural context of the organization, the industry, and society. In summary, emotional labor can be defined as adjusting one’s emotions or emotional expressions according to display rules (Grandey & Sayre, 2019). The inconsistency between actual emotional states and normative emotional expressions is called ‘emotional dissonance’ (Hochschild, 1983; Middleton, 1989). Emotional dissonance would be recognized as a primary source of stress experienced by service workers (Grandey, 2003).

Salespersons need specific interactions with customers to effectively achieve their sales goals, and for this purpose, there are clearly expected rules of emotion and behavior (Homburg et al., 2011). Even employees who work in a specific environment have different experiences while performing emotional labor because they are less likely to have the same perception of others and work environment (Johnson et al., 2021).

Specifically, Hochschild (1983) introduced the construct of emotional labor in seminal book, ‘The Managed Heart,’ which proposed that there are two approaches to performing emotional labor: surface acting (SA) and deep acting (DA). When employees alter their observable emotional displays (e.g., body language, voice tones, and facial expressions), but do not regulate the emotions themselves, they are conducting surface acting. In contrast, when employees show their genuine emotions after deliberately controlling them, they are performing deep acting (Brotheridge & Grandy, 2002; Hochschild, 1983; Kruml & Geddes, 2000). Service workers or salespersons who have to interact directly with customers are required to express appropriate emotions to customers, whether informal or not, as job requirements, and employees choose one of these two ways to meet the needs of these feeling rules (Klein, 2021).

Since Hochschild (1979) proposed the concept, many researchers have explored qualitative studies to examine the negative effect of emotional labor on employees’ well-being. This issue was investigated in various occupations, such as nurses, hospital workers, debt collectors, waiters, and cashiers (Adelmann, 1995; Smith, 1992; Sutton, 1991; Wharton, 1993). Though Hochschild (1983) focused on service workers who directly interact with customers, previous studies of emotional labor have expanded the concept to include workers in other industries or workers who interact with internal as well as external customers (e.g., Ashforth & Humphrey, 1993; Steinberg & Figart, 1999; Yanay & Shahar, 1998).

Several studies have attempted to measure the dimensions of emotional labor. Morris and Feldman (1996) made the first attempt by operationally defining emotional labor in terms of duration of interaction, frequency of interaction, and emotional dissonance. Brotheridge and Lee (2003) developed the Emotional Labor Scale (ELS) based on a six-factor model of emotional labor: deep acting, surface acting, variety, intensity, frequency of emotional display, and the duration of interaction. Using ELS, they discovered that role identification had a positive effect on deep acting but was negatively related to surface acting. Surface acting also had a positive effect on the emotional exhaustion and depersonalization of the employees. Glomb and Tews (2004) validated the Discrete Emotions Emotional Labor Scale (DEELS), based on a six-factor model of emotional labor. The dimensions of the scale are genuine positive or negative emotions, faking positive or negative emotions, and controlling positive or negative emotions.

While other studies examined the relationship between emotional intelligence and emotional labor generally (Opengart, 2005), this study attempts to investigate the relationship in a more goal-oriented context, the context of selling. How can we, then, conceptualize the relationship between emotional intelligence and emotional labor in the selling context? Salespeople need to recognize acceptable emotions and control their emotions to express them appropriately. In other words, emotional intelligence can help salespersons to manage their emotions more effectively and follow the display rules to achieve favorable results. Thus, emotional intelligence can be regarded as the very ability that is required for salespeople to perform emotional labor (Brotheridge, 2006; Kidwell et al., 2011).

However, according to previous studies, the direction of the relationship between emotional intelligence and emotional labor is still unclear. There are some literature works which show that emotional intelligence elucidated a positive and consistent association with deep acting (e.g., Hochschild, 1983; Cheung & Tang, 2009; Lee, 2010; Liu et al., 2008). In contrast, there is still little agreement on the nature of relationship between emotional intelligence and surface acting (Brotheridge & Grandy, 2002; Totterdell & Holman, 2003). Although there are studies that reported a negative relationship between emotional intelligence and surface acting (e.g., Austin et al., 2008), other research did not find a significant relationship (e.g., Johnson & Spector, 2007; Liu et al., 2008). Because salespeople with high emotional intelligence can perceive and regulate their emotions effectively, it is likely that they will control their emotions in accordance with display rules before expressing them to the customers. Therefore, it is likely that salespeople with high emotional intelligence will rely more on deep acting than salespeople with low emotional intelligence. Thus, we can expect that salespersons’ emotional intelligence will be a positive effect on deep acting.

In contrast, what would be the possibility of high emotional intelligence salespeople engaging in surface acting? Though high emotional intelligence salespeople can manage their emotions effectively, psychological resources are still required for them to control their emotions. Therefore, to prevent exhaustion of psychological resources, salespeople with high emotional intelligence may just control emotional expressions if the situation allows them to do so. Thus, they will be more flexible in making choices to control their emotions. This means that high emotional intelligence salespeople may perform both deep acting and surface acting according to the requirements of the situation (Brotheridge & Lee, 2002; Cheung & Tang, 2009).

However, salespeople with low emotional intelligence will engage more in surface acting than in deep acting because they cannot control their emotions effectively. This reasoning may explain why previous research discovered inconsistent results associated with emotional intelligence and surface acting.

In sum, we propose the precise direction of the relationship between emotional intelligence and deep acting. On the other hand, we cannot present a hypothesis indicating a specific direction of the relationship between emotional intelligence and surface acting.

H1: A salesperson’s emotional intelligence is positively related to deep acting.

2.2. Customer Orientation and Emotional Labor

Customer orientation could be defined as marketing-related employees' characteristics to think and act from a more holistic view of the customer. It refers to the tendency to accurately identify customer needs and provide services (Hoffman & Ingram, 1992). Based on Henning-Thurau (2004), customer satisfaction can be understood as satisfying customer needs in the process of salespeople interacting with customers. In addition, the employees with high customer orientation remarked that they enjoyed the act of providing services to customers (Donavan et al., 2004). Customer orientation occurs because one’s role and task in the service or selling process have precise characteristics and suitability. Accordingly, many influencing studies treated customer orientation as a consequence of salespersons’ adaptive selling behavior (Blocker et al., 2011; Franke & Park, 2006; Singh & Das, 2013).

In contrast, many other studies viewed customer orientation as an antecedent such that customer orientation is a cognitive belief affecting adaptive selling behavior or customer satisfaction (Appiah-Adu & Singh, 1998; Brady & Cronin, 2001; Brown et al., 2002; Cross et al., 2007; Donavan et al., 2004; Hamzah et al., 2016; Hennig-Thurau, 2004; Homburg et al., 2011; Jaramillo & Grisaffe, 2009; Macintosh, 2007; Roman & Iacobucci, 2010; Stock & Hoyer, 2005; Zablah et al., 2012). Since we were interested in the role of customer orientation as a cognitive antecedent of emotional labor and adaptive selling behavior, we referred to the previous studies of such an orientation.

Furthermore, many scholars define customer orientation at the firm level. For example, Levitt (1980) defined a firm's customer orientation as fully understanding the target customer to make superior value for the customer consistently. Jaworski and Kohli (1990) included in the concept of customer orientation the concept of accurately making sure of the customer’s current needs and at the same time predicting how the customer’s wants will change in the future. In addition, Narver and Slater (1990) conceptualize customer orientation as developing products by communicating customer needs to companies, avoiding high-pressure sales, and establishing long-term relationships. Henning-Thurau (2004) suggests the definition of customer orientation as the extent that the corporation satisfies customers’ unmet needs in the process of operation. Many studies also applied the spirit of these definitions at the salespeople level.

In communicating with customers, the salesperson's customer-oriented mindset to understand customers' needs and deliver products and services corresponding to them makes similar his or her internal emotions and externally expressed emotions. Therefore, the psychological resources wasted due to emotional inconsistency are relatively small, and the remaining emotional power can be utilized in other areas (Zablah et al., 2012). When an employee engages in deep acting, we can expect that sales activities based on the intention to satisfy customer needs would occur more frequently because salespeople are attempting to control their emotions to emanate positive emotions to customers. Previous studies have also shown that when salespersons engage in deep acting, positive feedback can be obtained from customers while maintaining a lasting relationship with them (Brotheridge & Lee, 2002; Allen et al., 2010).

In contrast, some salespeople may sense the customer's internal desires with only emotional expressions and with not much of customer orientation. Often they involve surface acting, which uses only typical emotional expressions rather than various emotional display rules and pretends to conform to the expression standards. As a result, the size of the emotional discrepancy between the inner and the expressed emotions and thus the work stress will be relatively large (Zablah et al., 2012). Then, salespersons with a high level of customer orientation tend to satisfy customers’ needs and know that customer satisfaction and loyalty increase when they express sincere emotions (Stock & Hoyer, 2005).

Therefore, it is possible to express the same emotions as the salespersons genuinely feel by controlling their emotions before or during interactions and face-to-face communication with customers (Brady & Cronin, 2001). On the other hand, employees with low customer orientation have a weak tendency to increase satisfaction or loyalty by satisfying customer needs. In this case, salespeople are highly likely to control only the expression of apparent emotions without making any effort to control the emotions themselves when confronting customers. Therefore, we describe the research hypotheses regarding the relationship between customer orientation and emotional labor as follows.

H2.1: A salesperson’s customer orientation is positively related to deep acting.

H2.2: A salesperson’s customer orientation is negatively related to surface acting.

2.3. Emotional Labor and Adaptive Selling Behavior

Adaptive selling behavior was defined as changing one's own sales behavior based on perceived information about product sales obtained by salespeople in interactions with customers (Sujan, 1986). Previous quantitative studies on adaptive selling behavior focused on the impact of cognitive activities, such as salespeople's information processing and collecting ability and changes in communication strategies (e.g., McIntyre et al., 2000; Robinson et al., 2005; Weitz, 1981; Weitz et al., 1986).

Therefore, the capability to acquire necessary information during interactions and face-to-face communication with customers and the effects of a knowledge structure for effective intelligence processing are proposed as essential requirements for a salesperson to perform adaptive selling behavior (Spiro & Weitz, 1990). Empirical research on adaptive selling behavior has focused on fixed properties such as similarity and relative expertise in the customer-salesperson dyad (Woodside & Davenport, 1974; Riordan et al., 1977).

Salespeople use cognitive processes to collect and use emotional information during their interactions with customers (Kidwell et al., 2011). Salespeople with higher levels of emotional intelligence and customer orientation are expected to show different behavioral strategies when they perform emotional labor because they possess the mental ability and mindset that can be used to collect and use emotional information effectively while communicating with customers. This advantage results in higher levels of psychological resources and lower levels of emotional dissonance while performing emotional labor, which leads to more attention to customers’ individual needs and more effective use of communication strategies to satisfy them. Hence, the salesperson’s choice of behavioral strategy for emotional labor will have a different influence on adaptive selling behavior demonstrated by him or her.

Salespeople selecting surface acting tend to show superficial and ‘fake’ desirable emotions and experience a higher level of emotional dissonance, which makes them focus on their own emotional state rather than reacting to customers’ emotions. Therefore, they are likely to provide similar sales services that do not adequately reflect customers’ individual needs because they hardly have the necessary resources to obtain information on customers' emotions (King & Emmons, 1990). Hence, it would be also difficult for the salesperson to form a rapport with them (Hennig-Thurau et al., 2006; Grandey, 2003).

Salespeople conducting deep acting, on the other hand, will have lower emotional stress due to the consistency of the feelings expressed and the feelings they feel because they will strive to elicit the desired emotions from deep within (Grandey, 2003). Thus, they don't need a lot of emotional resources to regulate his or her emotional stress, and their resources can be used to understand the customer's situation. In the end, it was confirmed that the emotional exhaustion felt by the salesperson prevented the salesperson from performing adaptive selling behavior. Hence, they are highly likely to perform active adaptive selling behavior by forming deep relationships with customers, acquiring highquality emotional information from customers, and responding appropriately to customers’ emotional changes. They will be emotionally stable themselves and have great respect for others in personal relationships (Denham, 1993; Zhou & George, 2003). In addition, they will be able to firmly manage tasks, fulfill their responsibility, and consider viewpoints of other people. In this aspect, deep acting is a good candidate for an antecedent of adaptive selling behavior.

Based on these assumptions, we propose the following research hypotheses.

H3.1: Deep acting is positively related to adaptive selling behavior.

H3.2: Surface acting is negatively related to adaptive selling behavior.

2.4. Adaptive Selling Behavior and Customer Responses

Salespeople need to perform the sales activities to provide the customer with proper product/service recommendations and fit with the customers' requirements to satisfy them (Alavi et al., 2019; Furrer & Sollberger, 2007). In other words, the salespersons should be able to carry out adaptive selling to satisfy customer's basic needs and adjust to the individual customer's emotional state. And hence, the salesperson’s adaptive selling positively affects customer satisfaction. For example, Predmore and Bonnice (1994) showed that the more the telemarketers did adaptive selling, the higher success in the selling situation. In addition, some previous studies proved that the more adaptive selling behavior was likely to make more profits (Boorom et al., 1998; Spiro & Weitz, 1990). Adaptive selling behavior affects the sales outcome as well as customer satisfaction. Kennedy et al. (2001) showed that a car salesman who adopted adaptive selling put more faith in their customers' expectations. Moreover, the customers appeared to be more satisfied with the seller's service (Hyun et al., 2021; Oh & Kim, 2022). Many studies indicate that salespeople enhance customer satisfaction by grasping the individual customer’s characteristics, adjusting the proper strategies for the customer, and providing individualized sales services.

In recent years, adaptive selling has been recognized as one of the crucial traits for the salesman to maintain a connection with customers (Park & Holloway, 2003; Park & Deitz, 2016; Robinson et al., 2002). Adaptive selling behavior, secured by a series of efforts based on investigating the current situation and customers' characteristics, ultimately improves the ability of the salesperson and makes him or her adequately respond to appeals for help. Therefore, adaptive selling behavior takes charge of a critical role to create the enhanced quality of relationship with the customers and the great outcome of the result (Porter et al., 2003; Sujan et al., 1994; Verbeke et al., 2011). Due to the characteristics of adaptive selling behavior, salespeople build long-term relationship correlated with their customers, and salespeople and customers get all the positive benefits (Dunlop et al., 1988). Also, it was found that female salespeople and new employees in the position have a stronger adaptive selling behavior (Park & Deitz, 2016).

Relationship quality was a comprehensive concept that reflects the overall strength of the relationship between a company and its customers. It generally consists of three variables: overall satisfaction, trust, and commitment to a company or employee (Gremler & Gwinner, 2015). Several previous studies regard customer satisfaction as an antecedent variable of the relationship quality (Crosby & Stephens, 1987; Teas, 1993). In other words, customer satisfaction positively impacts the customer service commitment and trust in the organization. Customer commitment will increase along with overall customer satisfaction such that higher the customer assesses the level of satisfaction, the higher the level of attachment. In addition, customer satisfaction improves customers’ confidence in the organization (Anderson & Narus, 1990). According to Ganesan (1994), an ongoing overall satisfaction provides a sense of trust put together by mutual interest. Garbarino and Johnson (1999) explain that the factors of service outcome positively affect customer satisfaction, and satisfaction also has positive impacts on trust, commitment, and thus relationship quality.

Customer satisfaction affects repurchase intentions (Oliver, 1980). Jones and Sasser (1995) found strong relationships between customer satisfaction and repurchase intentions in auto repair service and pharmaceutical industries. Likewise, the effects of overall satisfaction on repurchase intentions were generally positive in many studies (Anderson et al., 1994; Halstead & Page, 1992; Labarbera & Mazursky, 1983; McDougall & Levesque, 2000; Oliver, 1980; Szymanski & Henard, 2001; Parasuraman et al., 1996; Reichheld & Aspinall, 1993). In addition, Szymanski and Henard (2001) present that there was a positively significant relationship with the effect of overall satisfaction on repurchase in their meta-analysis of previous studies.

Trust and relationship quality were a customer's belief in the expertise and promises of salespeople and were possible only when an interdependent relationship is premised. It was also the customer's confidence which makes salespeople fulfill their obligations to customers (Agnihotria & Krush, 2015; Vanneste et al., 2014). Due to this interdependent relationship, trust and long-term relationships are important factors influencing the customer (Sirdeshmukh et al., 2002). Doney and Cannon (1997) showed that trust in and relationship quality with the supplier would positively impact repurchase intentions. A study of sales in online shopping behavior, Gefen (2000) and Lynch and Ariely (2000) found that customer trust positively affected repurchase intentions. Based on these previous findings, we were following hypotheses of replication.

H4.1: Adaptive selling behavior is positively related to customer satisfaction.

H4.2: Adaptive selling behavior is positively related to relationship quality.

H4.3: Customer satisfaction is positively related to relationship quality.

H4.4: Customer satisfaction is positively related to repurchase intentions.

H4.5: Relationship quality is positively related to repurchase intentions.

3. Research Design and Methods

3.1. Dyadic Samples and Data Collection Procedure

In this study, the unit of analysis is a dyad relationship between salespersons and customers. The participants were salespersons from two retail service industries (five insurance companies and two duty-free department stores). We selected the organizations because the salespeople were expected to have display rules with different characteristics. Convenience sampling was used for this study. In total, 570 respondents (380 customers and 190 salespersons) completed the questionnaires out of the 400 (customers) and 200 (salespersons) dyadic samples. Of these, we used a total of 182 data salespersons’ self-responses and thus 182 individually matched independent pairs of salesperson-customer dyadic data for the final hypotheses test. In other words, the responses of two customers were matched with that of the corresponding salesperson. The summed average of two customers’ responses was used.

We asked the manager for cooperation in conducting the survey, distributed and collected the questionnaire to the salesperson, and received responses from two customers who were serviced by the salespersons. Insurance salespeople accounted for 65.4%, while duty-free retail shop salespeople constituted 34.6%. The majority of participants were male (51.1%) and between 20 and 29 years old (43.4%). As for the type of employment, 54.9% of the responses were from regular employees and 45.1% from non-regular employees. Moreover, most of them had worked more than two years in their current organizations (47.8%). The gender of the customer was male (47.3%) and female (52.7%). By age, 29.1% were in their 20s (29.1%), 40.7% were in their 30s, and 19.5% were in their 40s. In addition, most of them had a bachelor’s degree (43.4%), marital status married (54.9%) and single (45.1%) participated. Finally, the average monthly income was less than 2 million won (16.2%), 2 to 3 million won (32.4%), 3 to 4 million won (26.9%), and 4 million won or more (24.5%).

3.2. Measures

All measures were operationalized as multi-item constructs from prior studies and adapted to the context of this research. With a few exceptions, questions used 5-point Likert-type scales ranging from 1 (very low) to 5 (very high). To assess emotional intelligence, 16 items (four items per stimuli) of the Wong and Law Intelligence Scale (WLEIS) developed by Wong and Law (2002) were measured, conducted a confirmatory factor analysis (CFA) and found that the instrumentations had four branches. These dimensions were consistent with the four components based on Salovey and Mayer's (1990) model. Reliability (internal consistency) estimates (Cronbach’s alphas coefficient) for the 4 measurements of regulation of emotion, use of emotion, self-emotion appraisal, others’ emotion appraisal and were .76, .88, .89, and .85, respectively.

Salespersons need to know about self-emotions and handle them to understand customers' emotions. In addition, moderating the emotions to adjust to the selling situation is required in customer-seller interaction (Spiro & Weitz, 1990). The research has been limited despite the great importance of employees’ customer orientation in applying the marketing concept in a market-driven company. Saxe and Weitz (1982) measured customer orientation at the individual level for the first time. Customer orientation was used to measure variables related to 12 items from the customer dimensions of Saxe and Weitz’s (1982) SOCO scale (Cronbach’s alphas coefficient alpha = .883).

Emotional labor was defined as the act in which salespeople regulate their emotions and expressions in order to achieve the organizational goals in the mutual relationship between salespeople and customers (Grandey, 2003). We used the measure items of Kim and Han (2008) to measure emotional labor. They extracted the items from Brotheridge and Lee's (2003) emotional labor scale, forming two subscales: deep acting and surface acting. The Cronbach alpha for each subscale was .83 (deep acting) and .78 (surface acting).

Adaptive selling behavior was defined conceptually as developing the necessary sales knowledge for each sales situation (Spiro & Weitz, 1990). We followed Comer et al. (1996) and Park and Holloway (2003) to measure adaptive selling behavior, using the 7 items from the ADAPTS scale about adaptive behaviors. This scale proved to be appropriately reliable (Cronbach’s alpha coefficient =.82).

Relationship quality was comprised of two dimensions: trust and commitment. We measured the trust and commitment of a salesperson by using 8 items from Crosby et al. (1990), Morgan and Hunt (1994), Lassar et al. (1995), and Garbarion and Johnson (1999). The items of each variable were revised to fit with the salesperson's service selling context. Two items measurement is commonly used for repurchase intentions (Oliver, 1997). Accordingly, we measured repurchase intentions by the degree of customers’ revisiting the seller (Oliver, 1997) and the degree of recommending to others (McDougall & Levesque, 2000). Cronbach alphas for all measurement items were acceptable, ranging from .759 to .899. Appendix 1 presents all measurement items used in the analyses.

3.3. Validity of the Measurement

As an analysis method of this study, structural equation modeling (SEM) was conducted to test the research hypotheses. Anderson and Gerbing’s (1988) two-step approach was used, whereby a confirmatory factor analysis (CFA) examined the relations of the observed constructs before standardized factor loadings by statistically evaluating the fit among the assessment of the measurement and structural models (Hair et al., 2005; Yang, 2005). We used the matched dyadic data (n=182 salespersons, 182 responses of 364 customers). Also, the method used in each analysis applied the maximum likelihood extraction based on the common variance of the matrix as implemented in SPSS Statistics 24.0 and AMOS 24.0. Specifically, measurement models for six latent variables with a total of fifty measured variables were investigated. With this process, we re-specified measurement models to improve discriminant and convergent validities. We also investigated the overall measurement model, which combined all the measurement models. As presented in Table 1, the model fit estimates indicated that all measurements had appropriate construct validity. In addition, descriptive statistics and correlations of ten constructs involved in this study are presented in Appendix 2. The overall fit of the hypothesized structural model was acceptable and other goodness of fit statistics meet the commonly accepted standards (χ² = 905.762, degrees of freedom = 668, RMR = .035, GFI = .807, IFI = .921, TLI = .910, CFI = .919, RMSEA = .044; Hair et al., 2005).

Table 1: The Results for Confirmatory Factor Analysis

Note: n = salespersons =182 / customers = 364; EI= emotional intelligence; OEA = others’ emotion appraisal; SEA = self-emotion appraisal; UOE = uses of emotion; ROE = regulation of emotion; CO= customer orientation; DA= deep acting; SA= surface acting; ASB= adaptive selling behavior; RQ= relationship quality; CS= customer satisfaction; RI= repurchase intentions; AVE= average variance extracted

4. Results

4.1. The Results of Hypotheses Testing

Figure 1 illustrates the relationships among the constructs, showing standard estimates of each hypothesized path. The hypothesized structural model indicates a good fit in all indices (χ² = 923.195, degrees of freedom = 685, IFI = .921, TLI = .912, CFI = .919, RMR =.037, GFI = .803, RMSEA = .044).

Figure 1: The Results of Model Test

H1, which predicted a positively associated with emotional intelligence and deep acting, was supported (b = .339; t = 2.931). Next, the salesperson’s customer orientation was positively related to deep acting (b = .583; t = 4.168) but did not have a significant relationship with surface acting (b = -.006; t = -.037). Thus, H2.1 was supported, while H2.2 was rejected.

Deep acting had a positive relationship with adaptive selling behavior (b = .214; t = 3.144), so H3.1 was supported. In contrast, surface acting was significantly related to adaptive selling behavior (b = -.020; t = -.317). H3.2 was not supported, but the path showed negative directions.

Adaptive selling behavior had a positive effect on customer satisfaction (b = .489; t = 3.614) and relationship quality (b = .373; t = 2.029), which confirmed H4.1 and H4.2. Moreover, customer satisfaction was positively associated with repurchase intentions (b = .807; t = 7.405) and relationship quality (b = .692; t = 4.650), which supported H4.3 and H4.4.

Finally, we confirmed H4.5’s prediction that relationship quality was positively related to repurchase intentions, which was supported (b = .364; t = 6.512).

4.2. Additional Analyses for the Mediating Roles of Emotional Labor and Adaptive Selling Behavior

In addition to the hypotheses testing, we did additional model comparison tests to check the soundness of our hypotheses-related reasoning. First, considering the reasoning that emotional intelligence affects adaptive selling behavior through emotional labor, we analyzed the alternative partial mediation model in which emotional intelligence is directly related to adaptive selling behavior, finding that the path coefficient of emotional intelligence → adaptive selling behavior was insignificant (t = .089; p = .912). Further, the full mediation model was no significant effects differently from the partial mediation model. The hypothesized model (emotional intelligence → deep acting & surface acting → adaptive selling behavior) indicated fit indices as follows: Chi-square = 205.830, degrees of freedom = 164, TLI = .958, CFI = .971, RMR = .036, GFI = .902, RMSEA = .038. The alternative partial mediation model fit (emotional intelligence → deep acting & surface acting → adaptive selling behavior and emotional intelligence → adaptive selling behavior) intentions. The bootstrapping analysis was done. The bootstrapping analysis appeared to be better than the alternative methods of Baron and Kenny (1986) or Sobel test since it could be applied even if the normal distribution assumptions of mediating effects did not hold (Preacher & Hayes, 2004). Using AMOS 24, 5,000 samples were extracted through the bootstrapping method. As a result, the mediating effects' standard estimate and standard error were respectively .094 and .041. Nevertheless, the percentile and the bias-corrected measure did not include 0 in the confidence interval of 95% (the interval of percentile = .031 ~ .193; the interval of bias-corrected measure = .033 ~ .199). These results indicate that the mediating effects of emotional labor and adaptive selling behavior are significant. In sum, these additional analyses confirm the critical parts of emotional labor and adaptive selling behavior in the effect of emotional intelligence on customer assessments ofsalespersons' selling behavior, such as customer satisfaction, relationship quality, and repurchase intentions.

5. Discussion and Conclusion

5.1. Implications

Our findings show that salespersons' emotional intelligence relates to emotional intelligence, which affects adaptive selling behavior, driving customer responses such as customer satisfaction, relationship quality, and repurchase intentions in the retail environment. It would then be essential to not only manage salespersons' emotional labor but also consider the emotional intelligence of salespersons. We present more specific interpretations of the study results in the following.

First, if the high emotional intelligence salesperson performs emotional labor, it turns out that he or she appropriately engages in deep acting. As the high emotional intelligence salesperson can recognize and control feelings effectively, they appear to genuinely feel positive emotions that they should present to their customers not just simply by their facial expressions but by adjusting the expression of emotion.

Second, we found that customer orientation was positively related to deep acting. The finding is new and significant that deep acting requires more of the cognitive antecedent (customer orientation) as well as the emotional antecedent (emotional intelligence) (b of EI = .339; b of CO = .583). In contrast, customer orientation was not significantly related to surface acting.

The fact that a salesperson has a high frequency of deep acting does not necessarily mean that the frequency of surface acting is low. In deep acting, psychological resources are consumed a lot to control the emotion itself. Then the resource is limited. Thus, deep acting cannot be used continuously. In this case, to express positive emotions to customers while consuming relatively little psychological resources, even if customer orientation is high, it is possible to use surface acting that only controls the expression of emotions. Then, even if the on his or her psychological resources, some salespersons use surface acting at a high level, and others use surface acting at a low level. Therefore, we may have found that the relationship between customer orientation and surface acting was insignificant. In other words, the emotions expressed as a result of controlling emotions must match the actual feelings. However, since psychological resources can be consumed in controlling emotions, it implies that not only the results of emotional labor but also the process are important.

Third, the notions of deep acting and surface acting require care from salespersons because these are two different strategies of emotional labor that can have diverse performance. Therefore, it is possible to develop the work performance and job satisfaction ability of salesperson through training on the use of deep acting strategies. Also, providing salespeople with opportunities to learn specific skills (i.e., attentional deploy, cognitive change, etc.) of deep acting may contribute to their adaptive selling behavior.

A critical conceptual contribution in terms of academic significance is that this study is almost the start in the sales literature to investigate the impacts of emotional intelligence and customer orientation on emotional labor, effective sales behavior, and outcomes assessed by customers. In particularly, conceptualizing and empirically validating the cognitive (customer orientation) and emotional (emotional intelligence) antecedent roles of salespeople's emotional labor, adaptive selling behavior and their consequences was a significant both a theoretical and practical implication in the retail sales literature.

The additional contribution lies in uncovering the relationship between emotional intelligence and emotional labor. Emotional intelligence and emotional labor have been studied independently in sociology and psychology (Opengart, 2005). Furthermore, this study adopts the empirical method based on recent studies of emotional intelligence and emotional labor measurements. Few studies have empirically conducted the impacts of emotional intelligence and emotional labor on work performance in the retail environment. Specifically, many studies of emotional intelligence and performance relationships were based on diverse personality traits in defining emotional intelligence (e.g., Bar-On, 2004, 2006; Goleman, 1998; Slaski & Cartwright, 2002). Further, many emotional labor studies have used qualitative methods (e.g., Rafaeli & Sutton, 1991; Sutton, 1991; Tolich, 1993) to test the relationship between emotional labor and performance. Especially, our analyses and results that emotional intelligence influences customer assessments of selling behavior through emotional labor and adaptive selling behavior add a new conceptualization and its empirical verification to the sales literature. Moreover, our study findings are methodologically meaningful since we take the dyadic approach free from the common method biases usually accompanied by single informant surveys.

From the managerial perspective, this study contributes a deeper comprehension of the nature of the work carried out retail salespeople and some directions for managing them. First, given that emotional intelligence can drive emotional labor, sales organizations can consciously evaluate emotional intelligence competencies when selecting and nurturing salespeople. Often, sales organizations focus primarily on employee incentives, salaries, and well-being to improve performance. However, our findings suggest that sales managers’ understanding of the emotional aspects of sales and service processes is critical.

In addition, the concept of emotional labor can contribute to a basis to perceive positive emotional experience of salespeople deeply. Primarily, sales organizations can develop their sales reps' deep acting capabilities, given the facilitating effects and consequences of deep acting on adaptive selling behavior. In this direction, training on deep acting strategy will become an essential role for improving the image and employees’ work performance of salespeople. This kind of training process could some more integrate the customer's perspective and story. Customer orientation is almost always crucial in driving deep acting with emotional intelligence. In addition, we must acknowledge that large number of customers can perceive whether salespersons are doing either surface acting or deep acting. Moreover, the training process should include the acquisition of strategies to suppress negative emotional expression such as emotional dissonance, job burnout and stress. Examples of the emotion-controlling strategy include recalling situations that evoke specific emotions and resolving negative emotions by focusing on the positive aspects of the situation (Grandey, 2003).

In sum, this study can contribute to marketing knowledge by adding the substantive knowledge of emotions and cognition in the retail workplace of selling. The results of this study may have significations for various domains, such as improving communication, designing customer orientation, emotional intelligence, and emotional labor training programs, and creating more productive work environments.

5.2. Limitations and Future Research Directions

This study, of course, has some limitations despite the interesting findings. First, future research needs to precisely assess why hypothesis 2.2 (customer orientation & surface acting) was not supported. Then, the need for future exploration involves the relationship between customer orientation and surface acting (hypothesis 2.2) and surface acting and adaptive selling behavior (hypothesis 3.2). These relationships can be quite complicated as the adaptable task of emotional intelligence and consumption of psychological resources and require further study.

Elaborating in a related vein, we indeed explored the possibility of the emotional intelligence’s flexible roles. A high emotional intelligence can be linked to DA (deep acting) and SA (surface acting) depending on situational requirements in precise goal-directed settings such as selling. We did a few analyses to check this flexible relationship of emotional intelligence and surface acting. However, we found that the relatively high-level emotional intelligence and surface acting association was positive but not significant: the correlation between emotional intelligence and surface acting was .110 (p= .273) in the split sample of the high emotional intelligence (n= 101; emotional intelligence median = 3.625). The low-level emotional intelligence and surface acting correlation was negative but not significant in the split sample of the low emotional intelligence (n= 81): correlation= -.13; p= .690. This is the canceling-out effect, making the relationship insignificant. Interestingly, our debriefing interviews with about ten selfsufficient salespeople confirmed our expectations to some extent. After discerning the importance of the prospect, they were able to prioritize their conversations and sales efforts (i.e., regulating emotional resources as our terms). We reason that presumably survey methodology may not have been sensitive enough to capture the flexible roles of emotional intelligence and surface acting from the salespersons. We must remember that there is little agreement in sociology and psychology literature on the associated between emotional intelligence and surface acting (e.g., Brotheridge, 2006). This warrants future research. Hence, the reasoning that emotional intelligence would relate to emotional labor (deep acting & surface acting) in a versatile way awaits the elaborated conceptualization and empirical testing.

Second, we need to conceptualize and empirically examine differences in distinct types of retail industry environments. Specifically, we need to consider that selling help and activities may play a more critical part in service sectors than tangible product domains. Hence, the future study may focus on finding the different impacts of retail salespersons' customer orientation and emotional regulation processes on the selling results for services vis-à-vis products.

Third, we had better examine the related effects of more concrete alternative deep acting strategies on effective selling behavior and its consequences. This study showed that the salespeople’s use of deep acting positively influenced on customer satisfaction, relationship quality, and repurchase intentions outcomes through adaptive selling behavior. However, further studies may be valuable to clarify the relative effectiveness of various strategies to promote deep acting.

Finally, this study applied existing tools from occupational and social psychology to measure emotional intelligence and emotional labor. However, new measures and methods will be needed to evaluate more precisely variables for sales situations in the distribution industry. This methodological interest may facilitate the research of cognitive and emotional phenomena of the salesperson in aspects of a sales encounter.

Appendix 1

Appendix 2