A comparative study of generations X, Y, Z in food purchasing behavior: the relationships among customer value, satisfaction, and Ewom

Abstract The COVID-19 pandemic has fundamentally changed many aspects of lives, including business and consumer behaviour. This study investigates Gen-X, Gen-Y, and Gen-Z in their change in food purchasing behaviour during the pandemic. Further, the study examines the relationships among customer value, customer satisfaction, and eWOM across three generations. It is found that (1) Gen-X, Gen-Y, and Gen-Z perceived different value during the pandemic, (2) Gen-X found to be the cohort that is most concerned with safety value in selecting food compared to the other two cohorts, meanwhile Gen-Y is the cohort with the highest concern on information value, (3) customer value has significant impact on satisfaction and eWOM for Gen-X and Gen-Z, however for Gen-Y, customer value impacts satisfaction significantly, surprisingly, it has no impact on eWOM. Managerial implications and future research directions are highlighted in this study.


Introduction
The COVID-19 pandemic has fundamentally changed many aspects of lives, including business (Turner & Aknremi, 2020) and consumer behaviour (Shaw et al., 2020). Among any other industries, hospitality industry has been mostly affected (Nicola et al., 2020) that implicates on the closure of restaurants (Hobbs, 2020). The current changes in public behaviour, include social distancing, large-scale social restrictions, and staying at home (Djalante et al., 2020), impact on attitudes Wahyuningsih ABOUT THE AUTHOR Dr. Wahyuningsih She is an associate professor at the Department of Management, Tadulako University in Indonesia. She finished her PhD degree in marketing from Monash University, Australia. Currently, she is the head of Management and Entrepreneurships Research Center, Faculty of Economics and Business Tadulako University. Her research focuses on customer value, sales management, consumer search behaviour, tourism marketing, and local resource management. and purchasing behaviour (Hasanat, et al). During the outbreak of COVID-19, consumers prefer to buy product online to avoid being affected by the virus (Shaikh, 2020). In the context of restaurant industry, customers might order food using online delivery services (Tomacruz & Flor, 2018). Hence, it can be noticed that COVID-19 has changed purchasing habits of consumers who have to stay at home to buy food using online delivery service.
A business which takes a proactive approach to their changing markets will be able to better serve customers (Guo et al., 2019), especially during COVID-19. Customers will certainly have higher sensitivity to the risks of the virus transmission, more concerned about the nutrition value of the products (Shaikh, 2020); therefore, they demand to seek for hygiene and healthy product (Djalante et al., 2020), including food (Hobbs, 2020). This indicates that COVID-19 has changed what value customers perceive in their decision to purchase products. Understanding what customers will value during COVID-19, delivering superior value to customers (Nasution & Mavondo, 2008), and designing best marketing strategy (Smith & Colgate, 2007) will ensure business' survival and put ahead of major competitors.
Moreover, delivering best value to customers has positive impact on behavioural outcomes: customer satisfaction, repurchase intentions, electronic word-of-mouth (eWOM), and loyalty (Barrutia & Gilsanz, 2012;Martinez et al., 2019;Rousta & Jamshidi, 2020). These previous studies have seen the relationships among the constructs at an aggregate level. Based on the theory of market segmentation, it is suggested that each type of customer needs a specific approach (Wahyuningsih, 2011;Zhou et al., 2020). Each customer segment will perform different attitudes and behaviour; hence, customers will perceive different value in purchasing product (Eggert & Ulaga, 2002). Consumers with different generational cohort tend to perform different values, attitudes, and behaviour (Rudolph & Zacher, 2020).
Consumers who basically prefer to socialize and dine out in restaurants are encouraged to eat at home utilizing online food delivery service due to the practice of social distancing. Hence, COVID-19 has brought social-cultural change (Djalante et al., 2020), and specifically it has created a major impact on the eating and drinking habits of consumers. They have become more health-conscious consumers in an attempt to increase immunity to protect themselves from the spread of COVID-19. According to Zwanka and Buff (2020), COVID-19 has resulted shifts in consumer behaviour; therefore, they suggested researchers to empirically test one or more generational cohorts whether any cohortspecific differences in attitudes and behaviour are apparent. Understanding the shifts of consumer behaviour is critical for business success and theoretical development. Based on these arguments, this research fills the gap as the most recent study which investigates the change of value that customers perceive during the pandemic, and how its impact on customer satisfaction and eWOM across three cohort of generations. Therefore, this study attempts to investigate Gen-X (people who were born between 1965 and 1979), Gen-Y (born 1980-1994), and Gen-Z (born 1995-2000 in their change in food purchasing behaviour during COVID-19. Further, this study sought to examine customer value and its effect on customer satisfaction and eWOM for different generational cohorts. The behaviour of these generations is detailed from their perceived value, level of satisfaction, and activities of doing e-WOM. The relationships among the constructs (value, satisfaction, and e-WOM) will be compared using multi-group analysis for the three generations.

Customer value
Customer value has been widely discussed and has received increasing attention from academics and practitioners (Talwar et al., 2020;Wang & Yu, 2019). This is due to firms which have a strong focus on customer value will form a sustainable competitive advantage (; Woodruff, 1997), specifically in the context of business, management, and marketing (Zauner et al., 2015). If the firms understand what drives value for their customers and provide superior value, they have a greater opportunity to obtain and retain customers which lead to higher profitability (Slater & Narver, 2000). In line with these studies, Woodruff (1997) points out that creating and delivering value to customers will increase the value of the firms. Hence, it can be argued that firms who deliver better value to customers will achieve better performance.
Concept of customer value has been determined in the literature as an assessment of the tradeoff between benefits and sacrifices (Brady & Robertson, 1999;Eggert & Ulaga, 2002;Lapierre, 2000;Ravald & Grönroos, 1996;Ulaga & Chacour, 2001;Walter et al., 2001;Zeithaml, 1988). Benefits defined as a combination of physical and service attributes, as well as technical support available in relation to a particular use situation (Monroe, 1990). Sacrifices are the loss derived from the product or service due to the increment of perceived short-term and long-term costs (Wang et al., 2004). To purchase a product, consumers aggregate the benefit attribute and then compare them to the sacrifices required to purchase the product (Brady & Robertson, 1999). Customer-perceived value captures the extent to which a customer perceives more benefits than sacrifices (Ritter & Walter, 2012). Hence, in delivering value, firms need to maximize benefits to customers and minimize sacrifices that the customers have to give up.
Along with the increased research on customer value, there has been an increase in the diversity of its measurement scale. Sweeney and Soutar (2001) put forward four scale for measuring customer value, quality, price, social value, and emotional value. Other scales examined in prior literature are performance (Bolton & Drew, 1991), utilitarian and hedonic value (Ryu et al., 2010), cost (McDougall & Levesque, 2000, and time/effort/energy (Lapierre, 2000). Having reviewed existing literature, this study adopts the Sheth et al. (1991) scale for measuring customer value. The scale consists of five components, namely, functional, emotional, social, conditional, and epistemic value. There are two main reasons for adopting this scale. First, the five components of value are a well-established theory that has been employed by previous researchers in examining customer value (Biswas, 2017); second, the scale has been well applied in hospitality sector (Talwar et al., 2020), specifically restaurant industry (Thome et al., 2018). Since consumers are facing a specific situation, namely, Covid-19 pandemic, therefore conditional value in this study is labelled as safety value. Safety value is defined as utility derived from the product's ability to maintain and improve immunity and the level of consumer health. In addition, epistemic value is labelled as information value adopted from Talwar et al. (2020), and it is measured in terms of the acquisition of information on different offers, promotions, and the terms and conditions related to the restaurants. Conceptualization of customer value utilized in this study is presented in Table 1.

Customer satisfaction
Customer satisfaction has been considered as a key strategic indicator as a profit generator (Shankar et al., 2003), a company's success, and long-term competitiveness (Woodruff, 1997). A satisfied customer is more likely to stay with the same company (Hultman et al., 2019), and less likely to switch to other companies (Gremler & Brown, 1999). Furthermore, Anderson and Sullivan (1993) pointed out that higher customer satisfaction insulates customers from competitors, enhances a firm's reputation in the marketplace, and lowers the costs of attracting and transacting with new customers. From these advantages, accordingly, customer satisfaction leads to profitability (Athanassopoulos et al., 2001).
Customer satisfaction is a post-choice evaluative judgment concerning a specific purchase selection (Oliver, 1980). Parallel with this argument, Yi (1990) defined customer satisfaction as an attitude like judgment following a purchase act or based on a series of consumer product interactions. There have been various scales to measure customer satisfaction include performance (McQuitty et al., 2000), quality (Ueltschy et al., 2004), expectation (Bendall-Lyon & Powers, 2004), and overall satisfaction (Hultman et al., 2019). Among the scales that have been examined by previous scholars, this study utilizes expectation and overall satisfaction since other scales have been covered in customer value measurement (Barrutia & Gilsanz, 2012).

Electronic word of mouth (eWOM)
The advancements of the internet have radically altered the way people go shopping (Tsao et al., 2015). According to Yen and Tang (2015) information provided, speed, alternative comparison, convenience, and one-to-many reach have facilitated business far beyond the reach of traditional WOM. The role of the internet in disseminating information worldwide such as the current COVID-19 virus spread has created higher awareness and more selective attitudes among global consumers in their shopping behaviours. WOM has been considered as important variable in service industry such as in hospitality sector because of its ability to reduce risk for purchasing intangible offerings (Jeong & Jang, 2011). Similarly, in the restaurant industry, customers will evaluate food quality, services, space, and experiences that determine their satisfaction and positive behavioural intentions, such WOM, as has been demonstrated in some studies (Kim et al., 2009;Ladhari et al., 2008).
Traditional WOM and eWOM are different from which WOM is typically face-to-face communications (King et al., 2014), while eWOM takes place in online platform. eWOM is defined as an exchange of service assessment in electronic commerce (King et al., 2014;Tsao et al., 2015). The popularity of social media sites has taken WOM into eWOM, with the increasing use of Facebook, Instagram, WhatsApp, YouTube, Twitter, and Line to interact among others. eWOM enables the information exchanges on huge number of people, real time/timeless, and no geographical boundary (King et al., 2014). People have now increasingly tended to perceive eWOM as providing reliable information in online activities (Gruen et al., 2006). This study uses eWOM to test the behavioural changes in online food shopping during the pandemic.

Generational cohort theory
Generational cohort theory has been applied in various research fields, such as social science, psychology, politics, education, human resource management, and marketing (Zwanka & Buff, 2020). According to generational cohort theory, a generational cohort is assumed to have distinctive characteristics which are exclusively attached to that specific group (Brosdahl & Carpenter, 2011). In general, people born in the same period have similar values, beliefs, attitudes,  Sheth et al., 1991;Thome et al., 2018 Social The utility derived from the product's ability to enhance social self-concept, augment self-image of the user, being-well respected and have relationships with others. Sheth et al., 1991;Wang et al., 2004 Emotional The economic value or monetary worth of feelings or affective states when customers experience an organization's products and/or services. Sweeney & Soutar, 2001;Sheth et al., 1991 Safety The perceived utility derived from the product's ability to maintain and improve immunity and the level of consumer health. Talwar et al., 2020;Thome et al., 2018 Information A product alternative's ability to incite, curiosity, offer novelty, and/ or satisfy a quest to seek knowledge. Sheth et al., 1991;Talwar et al., 2020;Sweeney & Soutar, 2001 preferences, and behaviour (Parment et al., 2013). Based on these previous studies, it can be noticed that people in the same cohort of generation tend to have the same values, attitude, and behaviour.
Despite the utilization of the theory, however, there have been no consensus on the determination on range of period of each generation. Carpenter et al. (2012) argue that Silent Generation refers to a cohort of people who were born between 1925and 1942, Baby Boomers born between 1943and 1960, Gen-X born between 1961and 1981, and Gen-Y, known as Millennials born between 1982. Meanwhile, according to Twenge (2015, Gen-X born between 1961-1981, Gen-Y born between 1982, and Gen-Z born between 1992-2000. This study utilizes the suggestion of Koksal (2019) and Shams et al. (2020) that determined Gen-X as a group of people who were born between 1965and 1979, Gen-Y born between 1980and 1994, and Gen-Z born between 1995

Proposed model and hypotheses development
Marketing strategy must be designed relevant to targeted market (Hunt & Arnett, 2004) based on segmentation approach (Brengman et al., 2005). This segmentation significantly contributes to the better understanding of consumer behaviour and demands (Zhou et al., 2020). One of market segmentation can be identified based on generational cohort (Chaney et al., 2017). Mannheim (1952) argue that although people of the same generation have experienced the same events, they may interpret them in different ways, so that generations are not homogenous. Therefore, it can be noticed that people from different generation tend to perform different attitude and behaviour. This study examines these differences, specifically the relationships among customer value, satisfaction, and eWOM (depicted in Figure 1).
Each segment of consumers tends to perceive different values (Bhatnagar & Ghose, 2004). Study by Parment et al. (2013) found that younger generational cohorts more likely select high prices with good quality and value. Parallel with this study, found that price sensitivity differs significantly between generations. In addition, Lo et al. (2020) investigate millennial consumers on their attitudes and behaviour toward restaurant. They suggest the menu information which customers value most, namely, nutrition and sustainability. In addition, Yarimoglu (2017) has identified that Gen-Y and Gen-Z perform different perception of service quality and value. Therefore, it is proposed that: H1: There are significant differences in customer value which consist of (a) functional value, (b) social value, (c) emotional value, (d) safety value, and (e) information value perceived by three generational cohorts Marketing research should aim to improve understanding of characteristic, lifestyles, and attitudes and consumption behaviour of each generation (Williams & Page, 2011). They point out that Gen-Y and Gen-Z determined as generations who need constant change; therefore, they are difficult to retain. In addition, study by Wen-Chuah et al. (2017) suggest that Gen-Y is the most value-conscious, least satisfied, and loyal customers compared with all other generations. Although Gen-Y found to be less satisfied, WOM should be more effective for this generation. Based on these previous studies, each generation perceive different value and might exhibit different satisfaction and eWOM. Accordingly, it is proposed that: H2: There are significant differences in relationships among customer value, satisfaction, and eWOM across three generational cohorts Customer value is identified in the level of buy or not to buy a product (Holbrook, 1994), meanwhile satisfaction is a judgment made by consumers after buying a product (Sweeney & Soutar, 2001). Therefore, the decision to purchase (or not to purchase) the product occurs before consumers demonstrate their feelings. Previous studies found that customer value has a strong, positive influences on customer satisfaction (Martinez et al., 2019). According to Oliver (1993), customer value, theoretically, can be considered as a cognition-based construct capturing any benefit-sacrifice discrepancy, meanwhile customer satisfaction is primarily an affective and evaluative response. Customer satisfaction is the result of a customer's perception of the value received in a transaction or relationship (Guo et al., 2019). This is in agreement with Lapierre (2000) who support this view that providing customers with good value may be the only reliable way to achieve sustained customer satisfaction. Accordingly, it is proposed that:

H3: Customer value has a positive influence on customer satisfaction
Most conceptual and empirical contributions to value literature conceive a direct impact of customer value on behavioural outcomes which consist of repurchase intentions, searching for alternatives and WOM, neglecting the role of satisfaction (Eggert & Ulaga, 2002). According to Sheth et al. (1991), buying decision is determined by cognitive and affective factors in consumer markets. Hence, it can be argued that there is a direct link between customer value and behavioural outcomes since value involve cognitive and affective aspects. This is in line with the study by Wu (2014) who suggest that value lead to the increase WOM intentions. With the rapid development of information and smartphone technology, consumers can share their consumption experiences at any time on Internet including social media, or known as eWOM (Wu & Lin, 2017). Furthermore, Gruen et al. (2006) found that customer value has a positive and significant effect on eWOM and loyalty. Based on this discussion, it is proposed that:

H4: Customer value has a positive influence on eWOM
The effect of customer satisfaction on eWOM, as one of behavioural intentions, has been examined by previous research (Jeong & Jang, 2011;Muskat et al., 2019). As pointed out by Padma and Ahn (2020), guests' satisfaction has a positive impact on customers' word of mouth and revisit intentions. Parallel with this study, Chang et al. (2009) found that customer satisfaction is antecedents of repurchase intentions and WOM. To recommend to other people, consumers need experience to consume the purchase product, so they can evaluate and make a judgment towards the product. Satisfaction is one of the predictors to eWOM intentions (Ryu et al., 2010). Given that the customer is satisfied, the results would be positive evaluation (Soderlund, 2003), including positive eWOM. In addition, Ha and Jang (2013) suggest that satisfying consumption experiences lead to more favorable future behavioural intentions, such as repurchase intention and positive WOM intention. Therefore, it is proposed that: H5: Customer satisfaction has a positive influence on eWOM 3. Methodology

Data collection and analysis
Online survey was used to obtain data from consumers who purchased food from restaurant using online application during COVID-19 (i.e. March to May 2020). Pre-test of the questionnaire was conducted in May 2020; then, the final questionnaire was revised based on the pre-test feedback. The questionnaire was distributed using Google online survey application through social media such as WhatsApp, Facebook page to potential respondents across Indonesia, during the first week of June 2020. The online survey covered most provinces across Indonesia. In total, 865 responded to the survey; nevertheless, only 577 out of the 865 observations (66.7%) met the research criteria; hence, those were retained for analysis. To test the hypotheses, various analyses were used including correlation analysis, Analysis of Variance, and Structural Equation Modelling.

Psychometric properties
The structured questions used in this study were formulated both from existing measurements and information from the literature review. Metrics scale using 1-7 Likert scale was employed for questions related to the main variables of customer value, customer satisfaction, and eWOM. The measurement items for customer value were adopted from Sheth et al. (1991) with some minor refinements adjusted to the context of this study. The original measures consist of five components i.e. functional value, social value, emotional value, conditional value, and epistemic value. To measure functional, social, and emotional value, this study adopts the existing measures from previous studies (; Jeong & Jang, 2011;Sweeney & Soutar, 2001;Talwar et al., 2020). To measure conditional value, this study has made adjustment to the situation or circumstances that the customer faces i.e. pandemic COVID-19. Items were developed using the concept of conditional value suggested by Sheth et al. (1991) consist of health safety, the restaurant concern for public health, and comply with health regulation. Based on these refined items, in this study, conditional value is labelled as safety value. The last component to measure customer value is epistemic value. Measure for epistemic value is based upon the existing literature put forward by Thome et al. (2018) that consist of new information (e.g., new menu items), novel information (e.g., price, healthy food, etc.), and substantial information (e.g., nutritional differences of food). In this study, these scales are defined as information value as suggested by Talwar et al. (2020). Hence, based on these adjustments, five scales of customer value employed in this study, namely, functional value, social value, emotional value, safety value, and information value.
To measure customer satisfaction, this study utilized overall satisfaction and expectation adapted from previous studies (Barrutia & Gilsanz, 2012;Han & Jeong, 2013;Oliver, 1980). The measures for positive eWOM) were adopted from Yen and Tang (2015) which consist of recommended and shared information to other people regarding consumer experiences in purchasing food online from a restaurant. In addition, the determination of generational cohorts was adapted from Koksal (2019) and Shams et al. (2020).
Reliability and validity were examined to establish the psychometric properties of the instrument. The mean, standard deviation, correlations, coefficient alpha, and average variance extracted (AVE) for the constructs were presented in Table 2. Cronbach alphas for each factor were calculated with the ranges from 0.79 to 0.90. The examination of reliabilities reveals that all constructs demonstrate good reliability as indicated by the Cronbach Alpha values which are greater than 0.70 (Nunnally, 1978); thus, providing evidence that the measures were reliable. Discriminant validity was assessed using the Fornell and Larcker (1981) method. In this method, the covariances (between latent factors) were compared to AVE of the indicators of each variable. Discriminant validity is established if performed when the AVE is larger than the square root of the covariance. As presented in the table, discriminant validity was established for all constructs. In addition, average variance extracted of all constructs is greater than the threshold level of 0.50, these demonstrate convergent validity. Overall, psychometric properties were established for all constructs examined in this study.

Respondent profile
Majority of respondents were Gen-Z (50%), most respondents (60.1%) purchased food via online for 1-5 times (as the least frequency), and 55% of respondents have reduced their online shopping due to the pandemic. In terms of restaurant types, 42% of respondents (the highest percentage) purchased food from fast casual restaurants. Type of food they purchased mainly were fast food (38%) and main dishes (36%). In addition, most of respondents (53%) used "GoFood" application in purchasing food via online.

Different perceptions of customer value components across three generational cohorts
Analysis of variance (ANOVA) was conducted to examine differences on perceptions of customer value components across Gen-X, Gen-Y, and Gen-Z. The results of hypotheses testing as shown in Table 3 indicate that there are no significant differences in the mean of functional value, social value, and emotional value as perceived by the generational cohorts; thus, H1a, H1b, and H1c were not supported. The results further indicate that there are significant differences in the mean of safety value and information value among Gen-X, Gen-Y, and Gen-Z; hence, the results support H1d and H1e. Specifically, Gen-X exhibits the highest mean of safety value (6.14, p < 0.01) compared to Gen-Y (6.11, p < 0.01) and Gen-Z (5.81, p < 0.01). Similarly, Gen-Y were shown to have the highest mean of information value (5.86, p < 0.01) compared to Gen-X (5.82, p < 0.01) and Gen-Z (5.53, p < 0.01).
The finding of this empirical research has shown that Gen-X, Gen-Y, and Gen-Z perceived different value during COVID-19 (see , Table 3 and Figure 2). Each generation will respond differently to COVID-19 based on the experiences that generation has had (Rudolph & Zacher, 2020). Such experiences are assumed to have an impact on different attitudes, values, beliefs, motives, and behaviour (Chaney et al., 2017). Gen-X, the oldest generation investigated in this study, found to be the cohort that is most concerned with safety value compared to the other two cohorts. In purchasing food online, Gen-X will select restaurants that implement health safety standards, care for public health, and comply with health regulations. Rising health consciousness caused by COVID-19 will lead to increasing demands for food safety and balanced diets (Zwanka & Buff, 2020). Safety was ascribed by previous researchers as most important aspect perceived by consumers in their shopping behaviour (Josiam et al., 2005). Other reasons underlying this is 7.60** Gen-Y> Gen-X> Gen-Z * = p < 0.05, ** = p < 0.01, *** = p < 0.001. ns = not significant because Gen-X is committed to family values, responsible for running households, taking care of children and caring for elderly parents (Taylor & Gao, 2014). They not only have to increase their personal health but also responsible for the lives of other family members. Our research suggests that COVID-19 has conveyed Gen-X perceive safety value as the priority in their decision making to purchase products in restaurant industry.
Gen-Z, the youngest generation examined in this study, is the cohort that has the lowest level of concern for safety value. This is in line with Perna's (2020) argument that Gen-Z was found to have least fear and low susceptibility to COVID-19. He points out that Gen-Z expresses the lowest concerns about corona virus, on the other hand, Gen-X is the cohort who are more anxiety about the virus. The value most sought by Gen-Z in purchasing food online from a restaurant is emotional value, which consists of enjoyment, convenience in putting order, payment, no need to cook, no need to dress up, and transaction security. Gen-Z is found to be pragmatic persons and fast-changing consumer technologies (Schwieger & Ladwig, 2018), wherein this type of persons will perceive for home-delivered food due to convenience reasons (Tomacruz & Flor, 2018).
In addition to safety value, the results also indicate that there are significant differences in information value perceived by three generational cohort (see, Figure 3). Gen-Y is the cohort with the highest concern on information value. In purchasing food, Gen-Y will consider whether the restaurants provide new, novel, and substantial information about their offerings. This result contradicts to the study of Lo et al. (2020) who suggest that Gen-Y is the generation that is most conscious of sustainability and personal health issues in selecting restaurants. The different finding may be due to different conditions and focus of both researches. The study of Lo et al. (2020) was conducted in normal conditions and focus on Gen-Y segment, whereas our study was carried out during COVID-19 and expanded investigation on other generational cohorts.

The relationship among customer value, customer satisfaction, and eWOM across three generational cohorts
The results of hypotheses testing are presented in Table 4 which indicate that overall, there is a significantly different relationship among customer value, customer satisfaction, and eWOM between Gen-X, Gen-Y, and Gen-Z; hence, the result supports H2. Specifically, for Gen-X, customer value was confirmed to have a significant impact on customer satisfaction (β = 0.37, p < 0.001) and eWOM (β = 0.35,

All Generation
Gen-X Gen-Y Gen-Z  p < 0.001), but customer satisfaction was not found to have a significant impact on eWOM. For Gen-Y, the results indicate a significant impact of customer value on customer satisfaction (β = 0.52, p < 0.001).
Whereas customer value and customer satisfaction were not found to have a significant impact on eWOM. Furthermore, the results support the view of the impact of customer value on customer satisfaction for all respondents (β = 0.61, p < 0.001), Gen-X (β = 0.37, p < 0.001), Gen-Y (β = 0.52, p < 0.001), and Gen-Z (β = 0.72, p < 0.001) led to acceptance of H3. In addition, customer value was found to have a significant impact on eWOM (β = 0.29, p < 0.001), and customer satisfaction also found to have a significant impact on eWOM (β = 0.17, p < 0.01); therefore, H4 and H5 were accepted.
This study has empirically demonstrated positive and significant relationships among customer value, satisfaction, and eWOM (refer overall model depicted in Figure 4). However, this relationship does not apply the same for three generational cohorts. Our finding identified that only Gen-Z exhibits a common behaviour (the same as overall model), meanwhile Gen-X and Gen-Y perform differently. The results suggest that the higher the value perceived by Gen-Z, the higher their level of satisfaction; therefore, they are more likely to recommend and share to other people. This finding corresponds with those of prior studies (Chang et al., 2009;Martinez et al., 2019) which examined positive and significant relationship between perceived value, satisfaction, and WOM. Gen-Z is a unique and truly digital native generation, as they are always following the development of information and communications technology (Ghani et al., 2018). Although Gen-Y and Gen-Z like to work with the latest technology, the younger one is more technology addicted and technology savvy (Kolnhofer et al., 2017). This is the underlying reasons why Gen-Z is likely to recommend and share their experience electronically to other people (eWOM). This is in agreement with the study of Sladek and Grabinger (2014), when Gen-Z has a good or bad experience with a company, they are not going to keep it to themselves, but share their evaluation or judgment via smartphone, social media, and the Internet.
Different from Gen-Z, the results for Gen-X indicate that there is positive and significant relationship between perceived value and satisfaction. Customer value also performs significant effect on eWOM, but the effect of satisfaction on eWOM found to be not significant. It means that Gen-X has willingness to recommend and share their perceived value to other people. Gen-X does not need satisfaction as an evaluative judgment to proceed their intentions to do eWOM. This finding is in agreement with previous research who argue that customer value has a direct impact on behavioural intentions, including eWOM (Eggert & Ulaga, 2002;Gruen et al., 2006). These prior studies agree that there is a direct link between customer value and behavioural outcomes because value involves cognitive and affective aspects. This finding is in agreement with Rousta and Jamshidi (2020) study who point out the positive effect of tourist's local food consumption value on intention to recommend. Hence, it can be argued that the studies which examined customer value utilizing emotional scale in its measurement, it will have a direct effect on behavioural intentions.
Unlike the other two generations, when Gen-Y perceives high value, they will perform high level of satisfaction, but do not have intentions to recommend and share their perceived value and satisfaction to other people. This finding indicates that Gen-Y is the most inexplicable cohort in their behaviour in terms of relationships among value, satisfaction, and intentions to do eWOM. The rationale behind this behaviour might be explained as follows. According to Kolnhofer et al. (2017), millennials are more pragmatic, embrace environmental protection, and they are self-

Safety Value Information value
Gen-X Gen-Y Gen-Z Total 6.14 6.11 5.81

Figure 3. Safety value and information value perceived by three generational cohorts.
Notes: significant ***p < 0.001, **p < 0.01, *p < 0.05. not significant assured and achievement focused. Besides, study of Parment et al. (2013) suggest that Gen-Y has no emotional connection or loyalty. This finding is predictable, since "millennials believe rules are made to be broken and tend to be here today and gone tomorrow" (Gursoy et al., 2008).

Conclusions and implications
This study has compared the relationships among customer value, satisfaction, and e-WOM for generations X, Y, and Z in purchasing food using online delivery service. Three main findings can be concluded; first, responding to government regulation preventing COVID-19, the food online shopping pattern and frequencies are all declining and fast food are preferable than other food types. Second, further insights of this study indicate that there are differences in safety value and information value across Gen-X, Gen-Y, and Gen-Z. Gen-X shows the highest concern on safety value, and Gen-Y more concern on information value. However, there are no differences across those three generations in functional value, social value, and emotional value. Finally, the results indicate that there are relationships among customer value and customer satisfaction for all respondents, and across three generational cohorts. All groups show that there is impact of customer value on eWOM, except for Gen-Y. The impact of customer satisfaction on eWOM was not significantly performed by Gen-X and Gen-Y. In addition, the results of the proposed relationships among the three key constructs show similar findings between all respondents and Gen-Z. Due to the partial supports on the findings across three generational cohorts, careful interpretation must be taken.

Theoretical implications
Main contributions of this study are related to domain of market segmentation and online shopping behaviour. In responding to COVID-19 situation, this study contributes theoretically as follows: first, different value perceptions across three generational cohorts in food online shopping enriche the marketing literature on the different value appreciation among Gen-X, Gen-Y, and Gen-Z. In this sense, Gen-X is more concerned on safety value, since they are responsible for family, community, and workplace. While Gen-Y emphasized on informational value, Gen-Z did not specifically show any tendency to a particular value. Second, the results from the relationships among customer value, satisfaction, and eWOM offer contributions wherein the three cohorts show different responses to perceived value,

Overall Model
Gen-X Gen-Y Gen-Z Notes: significant ***p<0.001, **p<0.01, *p<0.05 not significant  satisfaction, and eWOM. These different responses of each cohort in the structural model were empirically tested and worth the marketing knowledge, particularly contribute to market segmentation literature and consumer behaviour. In terms of segmentation, marketers are suggested to divide market segment not only based on demographics aspect, namely, age, but in more depth on generational cohorts. Therefore, they will be able to identify the behaviour of consumers including their motives and attitudes in purchasing a product.

Managerial implications
The results are likely to be most useful for restaurant managers/owners or managers from a similar context. First, restaurant managers and website providers should understand the target market fit well with their core products (food types, etc.). The COVID-19 has changed online food shopping behaviour; thus, managers need to proactively respond to different customers preferences. Second, as there are differences in the relationships among three main constructs, managers should not therefore treat customers in the same manner, and should properly consider customer value components (functional, emotional, social, safety, and information value) to better satisfy, and subsequently to positive eWOM. Overall, managers are suggested to implement better marketing strategy, maintain customer relationships, and make adjustments to the changes of market during pandemic, and the new normal situation.

Limitations and future research
Several limitations are acknowledged in this study. The limitations suggest that caution is needed in interpreting findings from this study, but they also indicate future research directions. First, this study adopts a descriptive research design using online survey to determine the degree to which the variables under investigation are associated. It would be beneficial also to conduct a qualitative research and offline survey. A qualitative research would deepen the understanding of customer perceived value across generational cohort. Second, the present study focused on customer value, customer satisfaction, and eWOM as the key constructs; surprisingly, some findings provide different directions. Hence, future studies are recommended to include other constructs as the predictors of eWOM behaviour; this would give a more holistic interpretation of the online shopping behaviour, particularly in restaurant industry. Third, three generational cohorts were focused on this study; future research should explore baby boomers in addition to the three generational cohorts. This additional segment would provide better explanations for relationships among the variables across different segments. Finally, the context of this study was the restaurant industry in Indonesia and during COVID-19. To establish robustness of the findings of this research, future research might explore the framework of the present study to be extended to different industry (i.e. hotel industry, tourism industry, etc.), in cross-cultural background (e.g., other eastern countries, western countries, other ethnicity background, etc.), and in a "normal" situation.

Funding
The authors received no direct funding for this research.