1. Introduction

As the popularity and use of smartphones has grown, so has mobile shopping. According to the National Statistical Office, the size of the mobile shopping market in Korea was expected to be nearly 50 trillion won by 2015. Korea's online retail sector accounted for 19.6% of total retail sales, which is more than double the rate of major advanced countries such as the USA, 11.0%, Japan 10.8%, and France 6.5%. The Internet retail market in Korea is responsible for 12.7% of retail sales, which is much higher than in the UK (9.6%), Germany (5.0%) and Japan (3.9%) (Euro-monitor International, 2014). The development of the Internet retail market in Korea is due to the world's best IT infrastructure, a well-equipped delivery system, and the rapid supply of smartphones. Recently, online shopping, especially mobile shopping, has become the shopping mainstream due to its time-saving properties. Some think that shopping at a store eventually leads to inefficient time consumption due to frequent long lines at checkout desks; meanwhile, many still value the in-person shopping experience and some companies are expanding their channels through on and offline partnerships (Han, 2017).

Thus, since many consumers still want to directly experience the joys of shopping, the development of complex shopping malls continues. The consumption values of a retail store, regardless of whether it is on- or offline, may affect the attributes selected by the store. The values of shopping can be divided into hedonic and utilitarian. Their effects on the specific attributes of a shopping mall can vary according to specific consumption value. For example, if a consumption mall has easy accessibility and a convenient layout, it will satisfy practical consumption values. On the other hand, the many shops offering a wide variety of merchandise inside a large, uniquely-designed interior serve to satisfy hedonic consumption values. Not many studies have been done on the relative influence of consumption values on functional attributes such as store location, floor layout, and emotional attributes such as types of products and interior atmosphere.

The purpose of this study is to examine the effects of hedonic and utilitarian consumption values on the importance of shopping mall attributes. We will also examine how such consumption values affect shopping mall selection attributes through on- or offline channels as follows. First, we examine the effects of hedonic consumption values on empirical and functional attributes. Second, we explore the effects of practical consumption values on empirical and functional attributes. Third, we analyze the effects of the moderating role of the on- and offline channels when hedonic and utilitarian values affect empirical and functional attributes.

2. Theoretical Background and Research Hypotheses

2.1. Consumer values

2.1.1. Hedonic values

Previous researchers have classified consumer values into two different categories: practical and hedonic (Babin et al., 1994; Chandon et al., 2000; Hirschman & Holbrook, 1982; Stoel et al., 2004). Hirschman and Holbrook (1982) argue that hedonic consumption represents aspects of consumer behavior that are related to the multifaceted, fantastic, and emotional aspects of how consumers experience products. Babin et al. (1994) and Holbrook (1999) found that buyers perceive hedonic value when shopping behavior is assessed independently of planned purchase behavior and that hedonic values are self-deliberated and self-oriented. Chandon et al. (2000) found that hedonistic values such as experience, entertainment, and pleasure were due to fun and enjoyment rather than the fulfillment of tasks.

Bellenger et al. 1976) stated that hedonic consumption value is more subjective and personal than utilitarian value and is perceived through fun and pleasure rather than achievement of purpose. Therefore, hedonic consumption values reflect shopping pleasure or emotional usefulness. MacInnis and Price (1987) suggested that the shopping experience can provide hedonic value by allowing consumers to enjoy the benefits of a product without actually purchasing the product.

Holbrook and Hirschman (1982) argue that hedonistic values are subjective and personal rather than utilitarian values and that fun and playfulness rather than task accomplishment are the results. According to Bellenger et al. (1976), hedonic value reflects shopping's potential entertainment and emotional value.

Cho (2012) studied the effect of service quality on shopping value and use intention was examined by using coffee shop customers and it was verified that interaction and outcome quality had a positive effect on hedonic value.

2.1.2. Utilitarian values

Babin et al. (1994) reported that when consumers reached their desired shopping goals and found the items they were looking for, they received utilitarian value. Chandon et al. (2000) expressed customer value by means of purpose. For example, time savings, convenience, and cost savings can be categorized as utilitarian values. Babin et al. (1994) studied shopping value as a dimension of work, for acquiring products and services, and as a dimension of fun, for enjoyment.

According to Engel et al. (1995), the utilitarian value of shopping arises from gaining goods or services to meet the needs. Babin et al. (1994) suggested that consumers' purchasing behaviors, which emphasize utilitarian shopping values, are logical and reasonable and that they are highly task related. Batra and Ahtola (1991) and Engel et al. (1995) describe utilitarian consumers as being active, task-oriented and rational. Babin et al. (1994) distinguished shopping as task-oriented and empirical shopping. Task-oriented shopping emphasizes utilitarian values by considering shopping a method to acquire the wanted/desired products while empirical shopping emphasizes hedonic value by acting as a means of receiving pleasure through entertainment or emotion.

2.2. Shopping Mall Selection Properties

2.2.1. Empirical attributes

In the field of marketing, interest in user experience has been increasing (McCarthy & Wright, 2004; Pullman & Gross, 2004; Schmitt, 1999), and various research findings related to user experience have been presented (Hassenzahl & Tractinsky, 2006). Most previous studies related to the user experience have assumed that users' satisfaction with the product is improved when users positively experience the use of a specific product and that such improved user satisfaction can effectively explain and predict user loyalty (Bhattacherjee & Premkumar, 2004; Doll & Torkzadeh, 1988; McKinney et al., 2002).

Hirschman and Holbrook (1982) first proposed an ‘experiential view’ based on the symbolic, hedonic, and aesthetic nature of consumption. They raised the need for an empirical point of view since a method that focuses only on existing cognitive and practical performance cannot properly account for product selection and used by emotional and aesthetic needs. Customer experience is primarily generated when a consumer buys or consumes a product, but it is important for customers to understand advertising and marketing communications, such as color, form, background design, slogan, and masquerade (Brakus et al., 2009). In particular, emotions and senses play an important role in customer experience (Haeckel et al., 2003; Shaw, 2004). Schmitt (1999) argues that the ultimate goal of experiential marketing is to create a holistic experience that integrates the individual experience as a whole. Empirical marketers have seen customers as rational objects and at the same time as emotional objects with the motivation to have pleasant experiences. Berry et al. (2002) and Mascarenhas et al. (2006) found that customer experience was composed of two major components: (1) product and service functions, and (2) emotion and sense. Stores and products are most competitive when providing both cognitive and emotional attributes to the customer. Cognitive attributes refer to the attributes that provide the most intrinsic benefits to product consumption; emotional attributes refer to the feelings consumers have while using the product. Therefore, the overall customer experience should be appropriately combined with tangible and emotional factors throughout the value chain (Mascarenhas et al., 2006).

2.2.2. Functional attributes

Sweeney and Soutar (2001) presented four dimensions of service value perceived by customers at the customer level, taking into consideration customer expectations and performance aspects. Functional value (price/value for the monetary cost paid) is the usefulness of products related to reducing perceived short and long-term costs. The functional value of performance/quality refers to the usefulness of the product from the perspectives of expected performance and perceived quality.

Functional consistency, which mainly deals with a multittribute attitude model, is the degree to which consumers perceive the functional attributes of a particular product or service to a level where they consider it desirable (Kressman et al., 2006; Sirgy et al., 1991). For example, in a coffee shop, the quality of the coffee and other menu items, employee service quality, price, and accessibility are functional attributes. When a consumer using a specific product or store reaches a level at which the functional attributes of the product or store are considered to be desirable, the customer is satisfied. As a result, they show high loyalty through revisits and positive word of mouth (Kressman et al., 2006; Sirgy et al., 1991).

Practical value means an instrumental outcome (Fischer & Arnold, 1990). Therefore, consumption is already understood as a means to achieve the purpose set by the consumer (Rintamäki et al., 2006); functional attributes such as financial savings and convenience increase value (Chandon et al., 2000). Based on previous studies, the following hypotheses were established.

[H1] Hedonic values will have more influence on the empirical attribute importance than practical attribute

[H2] Utilitarian values will have more influence on the practical attribute importance than empirical attribute

2.3. Customer store selection attribute importance by channel

2.3.1. Offline store attributes importance

Store attributes are the characteristics that make up a store. Peter and Olson (1996) defined shop attributes via store evaluation criteria including various dimensions such as product and service dimension. Store attributes can play an important role in store selection and consumer purchase behavior because they can increase the possibility of a store's products being purchased based on the store's favorable image. Store attributes were studied through store evaluation criteria or store image, and the importance of store attributes was considered a very important variable in the store selection behavior of consumers.

Cha and Park (2014) looked at store selection attributes in four dimensions: store accessibility, store size, merchandising, and entertainment facilities. Those stores with a good image will affect the consumer's purchasing behavior for products they sell.

When consumers experience an image of a store that they think is ideal, that image helps the consumer positively perceive the quality of a particular product (Baugh & Davis 1989; Doyle & Fenwick, 1974). In addition, store image formed by various store environment factors has a direct effect on consumers' purchasing behavior (Keller, 1987). According to previous studies, various characteristics of a store, that is, its attributes, are expected to have an important influence on customer purchase intentions. Shen et al. (2016) examined how the attributes of offline channels influence consumer intentions to switch to offline channels and how they also lead to incremental demands on online channels.

2.3.2. Online store attributes importance

In an analysis of pricelevel by store type, Park (2006) showed that the price levels of online stores were lower than those of offline stores. This was also the case when it came to delivery price. Price variance at online stores was lower than that at offline stores. However, in a comparison between pure online and hybrid stores, the price differences were lower.

Szymanski and Hise (2000) studied the factors that affect customer satisfaction with online shopping mall users and found that convenience, product information, financial stability, and site design have a significant effect on customer satisfaction. Chu et al. (2017) showed a model that examines the effect of multichannel retailers broaden the offline store on positive online-attitude and onlinepurchase intention.

Chung (2017) showed that with the increase of smart customers and the popularity of online shopping, the growing number of online stores makes it difficult to compare various prices and delivery options especially in case of multiple purchasing.

[H3] The influence of hedonic consumption values on the empirical and functional attribute importance of a shopping mall will be larger on offline channels than on online channels.

[H4] The effects of practical consumption values on the empirical and functional attribute importance of a shopping mall will be larger for online channels than for offline channels.

[Figure 1] Study Model

3. Research Method

3.1. Sampling and Survey Procedures

A survey was conducted to verify the research hypotheses established. The questionnaire items for the research were revised and modified to fit the findings of previous researchers. In order to select and elaborate on survey items, the literature was reviewed and corrected. From December 1, 2016, to December 14, 2016, 250 questionnaires were directly distributed, and 150 online questionnaires were distributed by SNS and e-mail. From them, 374 questionnaires were collected, with 344 used for empirical analysis after 30 unfaithful or incomplete questionnaires were excluded.

This study was conducted by the following analysis method. First, reliability and exploratory factor analyses were conducted using SPSS 22.0 to verify the reliability and validity of the measured variables. Next, the structural equation model (SEM) was used as a statistical method for testing the hypotheses in this study.

The subjects were composed of 203 males (59%) and 141 females (41%). By age group, 2% were teenagers, 44% were in their 20s, 19% in their 30s and 36% were age 40 and above. Regarding respondent occupation, there were 115 (33%) students, 185 (54%) office workers, 9 (3%) housewives, and 35 (10%) others. Twenty-one percent of respondents had a monthly income of less than 500,000 won, 11% earned between 500,000 and 1 million won a month, 24% between 1 million and 3 million won, 25% between 3 million and 5 million won, 17% between 5 million and 10 million won, and 2% had a monthly income above 10 million won.

[Table 1] Demographic characteristics of survey respondents

4. Empirical Results

4.1. Evaluate metrics

First, reliability and validity analyses were conducted on the constituent concept measurement items measured by multiple items (Churchill, 1979). An exploratory factor analysis was conducted for reliability and validity, and Cronbach's coefficient was examined. Principal component analysis (PCA) was used for factor extraction, and factors were extracted based on Eigen-Value 1. A VARIMAX rotation was applied to the factor rotation method. [Table 2] shows the results of the exploration-based analysis. Cronbach's alpha coefficients were all 0.8 or higher (Nunnally, 1967).

[Table 2] Exploratory factor analysis

Exploratory factor analysis showed that the validity of the discriminant and convergence validities was confirmed to some extent; the validity factor was analyzed by using SPSS Amos 20.0. As shown in [Table 3], the chi-square value of the measurement model is significant (p = 0.00), but it is sensitive to model complexity and sample size. Therefore, it was judged to be appropriate to evaluate the model fit by considering NFI, CFI, and RMSEA. The overall model fit was rated satisfactory since NFI = 0.982, CFI =0.994, and RMSEA = 0.038. The composite reliability (CR) and the average variance extracted (AVE) meet the criteria proposed by Bagozzi and Yi (1988) (CR over 0.6, AVE over 0.5) and factor loadings were statistically significant (p<0.01), indicating convergent validity. [Table 3] shows the results of confirmatory factor analysis.

[Table 3] Confirmatory factor analysis results

X² = 72.237 (p=0.01, df=48), NFI=0.982, CFI=0.994,

RMSEA=0.038 (*p <0.05)

The square root of the average variance extraction (AVE) was used to verify the discriminant validity among each factor with single dimensionality. As a result, the value of the square root of AVE is larger than 0.5 and is larger than the correlation value outside the diagonal line in the related row and column, as shown in [Table 4]. Therefore, it can be said that the validity of the discrimination between the different construction concepts has been verified, in which corresponding differences in the measurement results are to be shown.

[Table 4] Analysis of discriminant validity through correlation analysis

Note: The values presented on the diagonal line are the square root of AVE.

4.2. Research Results

The research hypotheses were tested by using SPSS Amos 20.0. The fit index was satisfactory for the study model with X² = 73.219 (df = 49, p <0.05), NFI = 0.981, CFI = 0.994, RFI = 0.975, IFI = 0.994, TLI = 0.992 and RMSEA = 0.038. Table 5 shows the results for [hypotheses 1] and [hypotheses 2]. To verify [hypotheses 1] and [hypotheses 2], the procedure was as follows. We estimate the equally constrained model so that the hedonic value is equal to the value of the path coefficient for the empirical attribute importance and the value of the path coefficient for the functional attribute importance. We then estimate an unconstrained model that does not place constraints on the two coefficients. The chi-square increment values obtained in each model are compared, and the hypothesis is adopted if the difference is significant. For one degree of freedom, there is a significant difference (α = 0.05 level) when the difference between the two models is over 3.84. The same applies to the case where utility value is related to functional attribute importance. The results show that the difference between the paths of [hypotheses 1] and [hypotheses 2] is statistically significant in hedonic value (∆ chi-square (1) = 12.623) and practical value (∆chi-square (1) = 34.953).

[Table 5] Results of research hypotheses

* p<0.05, ** p<0.01 *** p<0.001

The hypotheses test found that the hedonic value was found to affect both empirical attribute and functional attribute importance, but the analysis showed it had more influence on the empirical attribute importance. On the other hand, practical value was found to affect only functional attribute importance. This implies that consumers are more likely to respond to empirical factors than to functional factors when they value hedonic value while practical value consumers tend to consider functional factors more important than empirical factors. The results of the above verification are shown in [Figure 2].

[Figure 2] Results of research hypotheses

4.3. Verification of the control effects of consumers’ consumption values

To verify the effects of online and offline channel control on a complex shopping mall, we compared the total sample (n = 344) by on- and offline customer groups. [Table 6] compares the path coefficients of the on- and offline consumer groups. The results of the comparison show that both hedonic and practical values exert more influence on empirical properties in offline channels. On the other hand, both groups, hedonic and practical, demonstrated more influence on functional attributes when they were part of online channels. However, an analysis of the incremental value of chi square and the p value found statistical significance only when the hedonic value affected empirical properties.

[Table 6] Comparison of On- and Off-line Consumers

* p<0.05

The effect of an on/off-line channel adjustment that has more influence on empirical properties in off-line purchasing has been proven empirically. As a result, the hypothesis that the effect of hedonic value on empirical attributes is greater for off-line consumers is adopted. On the other hand, the effect of hedonic value on functional attributes and the effects of utilitarian values on functional and empirical attributes was not statistically significant in either on/off line group.

5. Conclusions

The purpose of this study is to determine whether consumption values affect shopping malls’ empirical and functional attributes and how on- and off-line channels affect those effects. The result showed that hedonic value exerted more influence on the empirical attribute importance than on functional attribute importance of shopping malls; these results were statistically significant (p<0.001). In addition, utilitarian value was found to have more influence on the functional attributes than on the empirical properties of shopping malls; these findings were also statistically significant (p<0.001). However, these relationships differed by channel. Consumers' consumption values were more influential on empirical attributes in off-line channels but were more influential on functional attributes in online channels. However, this was statistically significant (p<0.05) only when the hedonic value influenced empirical attributes. These results provide the following implications for companies operating complex shopping malls.

6. Implications

First, since hedonic value is more important for its experiential property, experiential marketing should be strengthened for pleasure value-oriented customers. Consumers care more about meaningful interactions with shopping environment than product feature itself. They need to feel inspired, empowered and heard. Especially, new generation of consumers is all about the emotional connection they can form with brands and products. Getting consumers to have feelings towards shopping mall brand and products is the first step towards creating a lifelong customer. The next step is getting those same consumers to share those feelings with others. Practical value, though, is evaluated as more important for its functional property. Therefore, for practical value-oriented customers, it is necessary to better promote quick product selection and convenient shopping such as swift payment and proper refund and exchange system and so on. And, while closing transactions at online shopping site is one important goal, companies should not lose importance and power of their online site as an information and communications vehicle.

Second, the way most companies operating a shopping mall market is based only on customers’ past purchasing behavior, regardless of whether the channel is on- or off-line. However, companies must recognize that their customers' consumption values differ by channel and thus any marketing should take that into account. For off-line channels, companies need to introduce more enjoyable marketing techniques that can entice consumers to buy products through pleasing experiences, while on-line channels need to improve their systems so that customers can quickly process payments, refunds and exchanges. This study has several limitations which suggest future directions for research. This study divided consumers' consumption values into only hedonic and utilitarian groups and limited the choice attributes of complex shopping malls to experiential and functional attributes. Future research on consumption values will need to take into account other attributes.