KOREASCHOLAR

Introduction
Although hotel employees are trained to deliver the best service, service failures may happen at any time because service is delivered by people to people (Susskind, 2002). Moreover, customers are more impressed by failed services than good services (Titz, 2001). According to the recovery paradox, customers have higher satisfaction level after experiencing a service failure if they receive satisfactory service recovery or compensation (McCollough & Bharadwaj, 1992). With the development of information communication technology and mobile device, customers can receive personalized services in recent days (Migacz, Zou, & Petrick, 2018). They also can easily share their experience on the online review platforms such as TripAdvisor, as well as select hotels based on shared online reviews (Liu & Park, 2015; Nieto-Garaía, Muñoz-Gallego, & González-Benito, 2017). Therefore, it is important for hotel managers to understand the mechanisms for service failure and recovery strategy. Thus, this study aims to examine the relationship between different emotion, customer engagement and brand loyalty under the context from the luxury hotels in China that different service failure compensation strategies are adopted. Particularly, the following two research questions are aimed to be addressed: First, do emotions (anger, regret and helplessness) significantly affect hotel brand loyalty through customer engagement? Second, does compensation type (immediate vs. delayed) significantly affect customer engagement and hotel brand loyalty based on customers’ emotions? The results of this study will benefit industry practitioners for formulating effective service failure recovery strategies.
Theoretical frameworks and hypotheses development
Stimulus-Organism-Response framework
Stimulus-Organism-Response (S-O-R) framework is a commonly used form of behavioral research in which events or occurrences are said to be the result of certain stimulus leading to a certain response, following a set of organism processes (Kim & Lennon, 2013; Mehrabian & Russell, 1974). In behavioral research, the S-O-R theory explains “how” something happens and a variance theory describes “why” (Chiles, 2003). We adopted the S-O-R framework in an attempt to explain the effect of the compensation types (immediate vs. delayed) on hotel brand loyalty. In our research model, customer engagement is used an intervening construct on the causal relationship between emotions of customer (anger and regret as a retrospective emotions, helplessness as a prospective emotion) (Gelbrich, 2010) and hotel brand loyalty. Customer engagement is composed of multidimensional concepts of identification, enthusiasm, attention, absorption, and interaction (So, King, & Spark, 2014). Our model thus explains four basic processes of relationship impact on service failure as “stimulus”, emotions and customer engagement as “organism”, and hotel brand loyalty as “response”. This study also emphasizes compensation type as “moderator”. The model shows how to enhance the understanding of emotions that affect hotel brand loyalty through customer engagement based on the moderating effect of compensations type.
Customer engagement
It is important for a firm to manage customers to improve a firm’s performance. Customer management has transformed from customer transactions, to relationship marketing, and then engaging customers (Pansari and Kumar 2017). There are different definition about customer engagement and most of them define customer engagement as the activity of the customer toward the firm. For example, Pansari and Kumar (2017) define customer engagement as how customer contributes to the firm by “the mechanics of a customer’s value addition to the firm, either through direct or/and indirect contribution.” Vivek et al. (2012) define customer engagement as “the intensity of an individual’s offerings or organizational activities, which either the customer or the organization initiates” (p.127). It has been discussed that customer engagement has been affected by customer emotion and also has significant impact on behaviour intention and brand loyalty. However it has not been discussed under service failure context and when different types of compensation strategies are employed. This study therefore aims to explore this mechanics. Under hospitality context, So, King and Sparks (2014) develop five factors to measure customer engagement: identification, enthusiasm, attention, absorption, and interaction. Since this study also examine hotel guest customers, we adopt the scale of So et al. (2014) due to its comprehensiveness and consistent context.
Service failure and emotion
Customer emotion is an important antecedent of customer engagement. Currently firms have been shifted their focus from selling products to emotional connection with their customers (Pansari and Kumar 2017). Positive emotion may enhance customer engagement and thereby improve customer loyalty. But when service failure occurs, customers have different negative emotions including anger, frustration, helplessness, regret amongst others. These negative emotions of customers disappoint customers themselves and reduce customer loyalty. Different emotions may have different impact on customer engagement. Anger often refers to the attributes of others such as the service providers (Weiner, 1985) whereas regret often refers to the service failure locus of customer himself/herself such as the customer is regret to choose this service provider (Roseman, 1991). Both anger and regret refer to retrospective emotions and when customer would like to solve questions they may also negative emotion of helplessness which is called prospective emotions (Davidow, 2003; Gelbrich, 2010). This study aims to examine and differentiate the impact of two retrospective emotions of anger and regret and one prospective emotions of helplessness. The first hypothesis is therefore proposed:
H1: Anger has negative impact on customer engagement.
H2: Regret has negative impact on customer engagement.
H3: Helplessness has negative impact on customer engagement.
Service failure compensation
Though service providers aim to deliver zero fault service, it is inevitable service failure may occur that may bring customers anger and dissatisfaction and damage the customer loyalty thereby. It is found that compensation is an effective way to comfort and delight the dissatisfied customers. Therefore, it is important to formulate effective compensation strategy when service failure occurs. Different compensation strategies such as monetary or nonmonetary (Fu et al. 2015), immediate or delayed compensation (Boshoff, 1997; Davidow, 2003), may be suitable to different contexts/situations. According to prospect theory, a customer is risk-reverse in case of gains. A customer may value products available now more than products obtained in the future due to the higher certainty of the former. Similarly, immediate compensation has less uncertainty than delayed compensation, and therefore is supposed to have higher value. Therefore customers with anger are assumed to have higher customer engagement when immediately compensated. On the other hand, regret customers attribute failure to himself/herself and therefore less expect compensation. The immediate compensation may lead to unfair and thereby less effect than delayed compensation. Therefore immediate compensation may not always be superior over the delayed one under different contexts. We therefore propose the second hypothesis:
H1a: Compensation type (immediate vs. delayed) moderates the relationship between anger and customer engagement.
H2a: Compensation type (immediate vs. delayed) moderates the relationship between regret and customer engagement.
H3a: Compensation type (immediate vs. delayed) moderates the relationship between helplessness and customer engagement.
Brand loyalty
Brand loyalty refers to the loyalty of a customer toward the brand both behaviourally and attitudinally (Dick and Basu 1994; Li and Petrick 2008; So, King, Sparks, and Wang 2013). It is a key goal of marketing activities, and its antecedents have been extensively examined such as satisfaction, perceived quality, received value, and brand trust, amongst others. Customer engagement, as the activity of a customer toward to a firm, is naturally viewed to influence brand loyalty. This study therefore adopts brand loyalty as the consequence of customer engagement. Furthermore, we would like to examine if compensation types have moderating effect between customer engagement and brand loyalty. We therefore propose below two hypotheses:
H4: customer engagement has positive impact on brand loyalty.
H4a: Compensation type (immediate vs. delayed) moderates the relationship between customer engagement and brand loyalty.
The research model is shown in Figure 1 where all hypotheses are demonstrated. Our research model is developed based on the S-O-R framework in which emotions are antecedent of customer engagement, and customer engagement impacts hotel brand loyalty. This research model also shows the moderating effects of compensation types has on causal relationships between the aforementioned constructs.
Methodology
Scenario design
Scenario based questionnaire is designed to obtain quantitative data for analysis. Based on the interview with hotel managers/operators, one service failure scenario and two compensation scenarios (immediate and delayed) are designed. In-depth interviews with a couple of hotel managers and guests were conducted to verify the realisation of the scenarios formulated. The questionnaire begins with a screening question: in the previous 12 months have you ever had experience staying in a four- or five-star hotel? The survey would only continue if the answer is “yes”. Then the participant is asked to write down the name of this hotel and read the below service failure scenario thereby. Service failure scenario: Imagine you have checked into this hotel again. During your stay in hotel, you send your coat for laundry. It is a nice coat and you bought it a year ago with the price of 1000RMB. However when you collect the cleaned coat, you notice that there is a damage on your coat which makes you cannot dress this coat anymore. You therefore call the service counter for complain. Immediate and delayed compensation scenarios were designed as follows: Immediate compensation scenario: after 15 minutes, the duty manager of the hotel went to our hotel and expressed his sincere apology. You showed him about the damage and informed him the original price of your coat. The manager offered you the cash compensation with the original price of your coat and you agree with this. After half an hour you received 1000RMB cash as the compensation. Delayed compensation scenario: after 15 minutes, the duty manager of the hotel went to your room and expressed his sincere apology. You showed him about the damage and informed him the original price of your coat. The manager said according to the hotel policy, they need to check how this happened and confirm the price of your coat first before making the compensation for you. After two weeks you left the hotel, you received 1000RMB compensation which is transferred into your bank account directly. Participant emotion is measured after the participants read the service failure scenario and before they read the compensation scenario. Each participant is randomly assigned to be involved in one compensation scenario only. Customer engagement and hotel brand loyalty are measured after the compensation happened.
Variable measurement
Customer engagement is measured using 25-item scale developed by So et al. (2014) in which five factors are involved: identification, enthusiasm, attention, absorption, and interaction. Particularly, identification is measured by four attributes: “When someone criticizes this brand, it feels like a personal insult”, “When I talk about this brand, I usually say we rather than they”. “This brand’s successes are my successes”. “When someone praises this brand, it feels like a personal compliment”. Enthusiasm is measured by five attributes: “I am heavily into this brand”. “I am passionate about this brand” “I am enthusiastic about this brand” “I feel excited about this brand” “I love this brand”. Attention is measured by five attributes: “I like to learn more about this brand” “I pay a lot of attention to anything about this brand” “Anything related to this brand grabs my attention” “I concentrate a lot on this brand” “I like learning more about this brand” . Absorption is measured by five attributes: “When I am interacting with the brand, I forget everything else around me” “Time flies when I am interacting with the brand” “When I am interacting with brand, I get carried away” “When interacting with the brand, it is difficult to detach myself” “In my interaction with the brand, I am immersed” “When interacting with the brand intensely, I feel happy”. Interaction is measure by five attributes: “In general, I like to get involved in brand community discussions” “I am someone who enjoys interacting with likeminded others in the brand community” “I am someone who likes actively participating in brand community discussions” “In general, I thoroughly enjoy exchanging ideas with other people in the brand community” “I often participate in activities of the brand community”. Three emotion of anger, regret and helplessness are included as the measurement of emotion. Particularly, according to Gelbrich (2010), three attributes are adopted to measure anger “I would feel angry with the hotel/hotel employees”, “I would feel mad with the hotel/hotel employees”, and “I would feel furious about the hotel/hotel employees”. Three statements are employed to measure regret (Tsiros & Mittal 2000): “I would feel sorry for choosing this hotel”, “I regretted choosing this hotel”, and “I should have chosen another hotel”. Four statements are used to measure helplessness (Gelbrich 2010): “I would feel helpless”, “I would feel lost”, “I would feel defenceless”, and “I would feel stranded.” Five statements are used to measure brand loyalty (So, King, Sparks, & Wang 2013): “I would say positive things about this brand to other people.” “I would recommend this brand to someone who seeks my advice.” “I would encourage friends and relatives to do business with this brand.” “I would consider this brand my first choice to buy services.” “I would do more business with this brand in the next few years.” A seven-point Likert scale ranging from 1 (=disagree strongly) to 7 (=agree strongly) is adopted for all measurement.
Data collection and analysis method
In-depth interview with managers from upscale hotels and customers will be used to finalize scenarios. Opinions of academic experts will be used to revise variable measurements and questionnaires. Convenience sampling method will be adopted to obtain about 400 respondents who has experience of staying at four- or five-stars hotels in China in the previous year. Regarding with data analysis, Partial least square structural equation modelling (PLS-SEM) is used to test the hypotheses proposed.
Expected results
The manipulation check has been conducted to verify the scenarios designed. The negative relationship between emotions and customer engagement are expected and compensation timing (delayed or immediate) may moderate this relationship. Most importantly, it is expected that this moderating effect varies when different emotions and customer engagement are examined.
Contributions
The theoretical contributions have three folders. Firstly, this study first considers compensation timing into the examination of relationship between different negative emotions and customer engagement, after service failure occurs. Secondly, this study adopts stimulus-organism-response theory to explore the mechanism how service failure could be well recovered by relationships of different negative emotions, effective compensation type, customer engagement, and brand loyalty. Thirdly, this study applies second order factor for the measurement of customer engagement and also divides negative emotions into retrospective and prospective ones to shed light on customer engagement in the context of service failure and compensation. The practical implication of this study will benefit industry practitioners for their formulation of compensation strategies. Especially as the development of big data, hotel industry is able to adopt different strategies for individuals to maximize customer experience. The findings of this study could propose different strategies for different situations/individuals thereby.

• Doris Chenguang Wu(Sun Yat-sen University, China)
• Namho Chung(Kyung Hee University, Republic of Korea)
• Zhaohan Hua(Sun Yat-sen University, China)
• Hee Chung Chung(Kyung Hee University, Republic of Korea)