1. Introduction
Social commerce is an emerging category of e-commerce based on social network services [
1], and it is a user-centered business model which leverages frequent social interaction and massive user-generated content (UGC) to assist consumers in making purchase decisions. As a quickly emerged area of inquiry for both practitioners and researchers, social commerce is a potential paradigm shift in e-commerce [
2], and its role in promoting the sustainability of business model has been widely recognized [
1,
3]. The increasingly intense commerce competition makes sustaining social commerce become a very challenging task [
3]; thus transaction mechanism analysis in social commerce is of great significance to the sustainable development of such a business model. Among the transaction process, consumer’s purchase intention is the premise of purchase behavior and it is also an important basis for enterprises to formulate market strategies. At present, the study of purchase intention in traditional e-commerce is relatively mature, and perceived risk and trust are considered as two main factors affecting purchase intention and sustainable transaction [
4,
5]. But the “sociality” characteristic of social commerce makes the influencing factors of purchase intention still a topic worth exploring.
Compared with traditional e-commerce, the relationship between consumers in social commerce is closer, and consumers have access to more information via social interaction. For example, users can interact frequently through product experience sharing, product recommendation, and community discussion. Previous studies have shown that such activities as user interaction and word-of-mouth communication in social networks can affect the subsequent purchase intention of users [
6,
7]. Therefore, social commerce is a combination of business activities and social activities. Social attribute is the most essential characteristic that distinguishes social commerce from traditional e-commerce, and it is also the core factor to study the purchase intention of users in social commerce. Meanwhile, it should be noted that the network community will present different forms and characteristics in different cultures [
8], and cultural factors will affect the behavior and attitude of e-commerce users to a large extent. Choi and Geistfeld [
9] found that uncertainty avoidance in culture theory can affect users’ perceived risk and ultimately purchase intention. Doney et al. [
10] also supported that culture will influence the establishment of trust among users. Moreover, due to the prominent social attributes in social commerce, the communication between online users is significantly more frequent. Therefore, the impact of culture on social interaction is more obvious in the context of social commerce, and researches on consumers’ purchase intention also need to consider the antecedent effect of cultural dimensions on social interaction. Based on large-scale research and theoretical comparison, Hofstede [
11] proposed the famous cultural dimensions theory to measure cultural differences, in which “uncertainty avoidance” and “individualism/collectivism” were considered to have a significant impact on social interaction in e-commerce [
12].
In view of the fact that the current researches on the purchase intention of social commerce have not fully revealed the social interaction mechanism and lack the consideration of the antecedent effect of cultural dimension on social interaction, we present the following research questions: (1) How social interaction affects consumers’ purchase intention in social commerce? (2) How uncertainty avoidance and individualism/collectivism in cultural dimensions affect social commerce consumers’ interaction?
In this paper, empirical analysis is conducted to study the influence of social interaction factors (perceived risk, trust, and intimacy) on the purchase intention of Chinese and French consumers, and we also explore the antecedent effect of cultural dimensions (uncertainty avoidance and individualism/collectivism) on social interaction. The results show that the impact of perceived risk on subsequent purchase intention in social commerce will be transferred by trust and intimacy to a certain extent. The intimacy between users contributes to trust-building, and both of their positive impacts on purchase intention would show distinct effects in different cultures. In addition, cultural dimension factors are proved to have a significant effect on users’ social interaction. Although high uncertainty avoidance brings perceived risk, it can promote subsequent trust-building. These findings help provide managerial insights for social commerce community to establish effective trust mechanism in a multicultural context, especially for those cross-border platforms struggling with market entry dilemma.
5. Data Analysis and Results
A total of 558 questionnaires were collected, in which 518 were valid. The following three types were considered as invalid questionnaires: (1) the answer time is less than one minute; (2) respondents who have never experienced social commerce; (3) and extreme answers such as always choose “totally disagree” or “totally agree”.
5.1. Demographic Statistics
Among 518 valid questionnaires, there were 291 Chinese users and 227 French users. The number of male and female respondents was approximately the same. Because social commerce is popular among young people, the age of respondents is mainly between 18 and 37 years old, accounting for 74% of the total. Most of the respondents had a bachelor degree or above.
Table 2 gives the detailed demographic information, and descriptive statistics results are shown in
Table 3.
5.2. Reliability Assessment
The reliability of internal consistency was tested using Cronbach’s alpha and composite reliability (CR), which measured the reliability, and they should be higher than the minimum cutoff score of 0.7. The result of reliability analysis shows that the Cronbach’s alpha and composite reliability of the six variables involved in the questionnaire are greater than 0.7, and the F test results are significant (see
Table 4). This means the measurements are reliable, and the factors measured the constructs consistently.
5.3. Validity Assessment
The measurement model’s validity was evaluated by examining content validity and construct validity. Content validity is a subjective evaluation index. Our questionnaire has used the existing literature and theories for reference, and the most items were adjusted according to experts’ advice and the attributes of social commerce.
Construct validity was tested using the subcategories of convergent validity and discriminant validity. Convergent validity was assessed by examining the Cronbach’s alpha (>0.7), composite reliability (>0.7), average extracted variance (AVE) (>0.5), and factor analysis [
46].
Table 4 shows that the values of CA, CR, and AVE for each model construct satisfy the thresholds.
In factor analysis method, KMO test and Bartlett’s test of sphericity were conducted first. Then exploratory analysis was carried out by using Varimax orthogonal rotation to extract factors with eigenvalues greater than 1. Generally, the original variables is suitable for factor analysis when KMO-value are greater than 0.7.
In the validity test of cultural dimension, the KMO-value was 0.781 and the Bartlett’s test of sphericity coefficient was 818.469 at the 0.000 significance level, which meant the measurements of cultural dimensions were reliable and could be used for factor analysis. Then, the varimax method was applied to rotate the scale to extract the factor whose eigenvalue is greater than 1. Two factors were obtained, which were consistent with the theoretical model. The cumulative contribution rate of these two factors was 56.74%, which had a good factor interpretation rate (see
Table 5). After orthogonal rotation, the loading matrix of each factor was obtained (see
Table 6).
The data in
Table 6 shows that all indicators load more strongly on their corresponding constructs than on other constructs in the model, which suggests sufficient convergent validity and discriminant validity of the model constructs. Similarly, for the validity test of social interaction variables and dependent variables (purchase intention), the KMO-value and Bartlett’s test of sphericity coefficient conformed to the requirements.
Table 7 is the factor loading matrix of social interaction variables, and the dependent variable does not rotate because it has only one factor.
Factor loading is an alternative to test convergent validity and discriminant validity [
5]. Besides, discriminant validity was also assessed by comparing the construct AVEs with inter-construct correlations. The results in
Table 8 also show that the square root of AVE is larger than the corresponding inter-construct correlations. Thus, the model constructs demonstrate sufficient discriminant validity. Also, we do not observe high correlations among predictors in the table, indicating the absence of multicollinearity.
5.4. Results of Hypotheses Testing
We tested our hypotheses using Structured Equation Modeling (SEM) technique, because SEM can simultaneously analyze all paths with latent variables in one analysis [
47]. Samples from China and France were analyzed.
Table 9 presents the results of model fit. The results show good fit of the model with data CMIN/DF < 3, Good Fit Index (GFI), Comparative Fit Index (CFI), Normed Fit Index (NFI), Tucker–Lewis Index (TLI), and Adjusted Goodness of Fit Index (AGFI) to be greater than 0.9. In addition, the root mean square error (RMSEA) is less than 0.08.
Table 10 and
Table 11 report the results of proposed hypotheses in China and France, respectively. As hypothesized, H2, H3, H4, H5, and H6 are all supported in both China and France. However, Hypothesis 1 does not pass the significance test. In group China, although the influence in Hypothesis 7 is significant, the path coefficient is positive, that is, individualism is positively related to intimacy, which is contrary to the original hypothesis. Hypothesis 7 is not supported in group France because of low-level significance.
6. Discussion
The purpose of this study is to investigate the influence of social interaction on purchase intention and the antecedent effect of cultural dimensions on social interaction by comparing Chinese and French groups. Furthermore, we discuss the conclusion of every hypothesis one by one in the following.
6.1. The Impact of Social Interaction on Purchase Intention
Hypothesis 1 is not supported both in China and France group, which indicates that there is no significant correlation between perceived risk and purchase intention of users in social commerce. Although some scholars [
48,
49] have found that perceived risk of users in traditional e-commerce can significantly affect purchase intention, shopping environment of social commerce will is much more complex because user interaction and information exchange will affect purchase intention as well. To some extent, the influence of perceived risk on purchase intention may be transferred by trust and intimacy. For example, given that a user is not familiar with a website and is worried about the risks, he may give up buying in traditional e-commerce scenario. But if the user has integrated into the community to create a sense of trust, or interact with other users frequently, the trust and intimacy perceived will weaken his suspicions about the purchase risk, and even generate the purchase intention. In social commerce, perceived risk may have less impact on purchase intention than trust and social relations. Farivar et al. [
50] state that trust toward the site members indirectly increases purchasing intentions, and the trust also reduces perceived commerce risk. Through a rigorous and quantitative meta-analysis, Wang et al. [
51] found that trust had a stronger effect on individual behavior than risk in social media platforms.
The test result of Hypothesis 2 shows that users’ trust in sellers or other users in social commerce can significantly affect their subsequent purchase intention, which is consistent with the findings in traditional e-commerce [
17,
33]. The path coefficient of trust on purchase intention under Chinese culture is twice as that of France, which may be related to the fact that French consumers prefer rational consumption. Comparatively, Chinese consumers’ long-term strengthened trust in the social community will drive their willingness to buy in the community. A reasonable explanation is that China’s social commerce shopping platforms are more diverse, and the management level and regulations are uneven, which may lead to users’ difficulty in making accurate judgments, so users generally have a lower (compared with French) sense of trust in the platform. In the absence of trust, if a social commerce platform can stand out and fully win the trust of users (including friend trust, business trust, and platform trust), the consequent purchase intention increment will be significant.
Hypothesis 3 reveals the importance of social interaction among users in social e-commerce to purchase intention, which is in accordance with [
13,
52]. Word-of-mouth recommendation among friends is one of the core drivers of consumer behavior in this business model. Besides, the purchase intention of French users is more affected by friends’ intimacy than that of Chinese users. Although France is a country of individualism culture and has relatively loose interpersonal relations, this does not mean that French users are less susceptible to the opinions of community friends. As indicated in [
18], the fewer the number of a user’s friends, the more likely he is to adopt their opinions. Therefore, for French consumers with loose social relations, close friends in online communities are valuable resources to assist shopping decisions and avoid perceived risks.
H4 discussed the relationship between social intimacy and trust. The results show that the intimacy between community members in both cultures has a positive impact on trust between them, which is also verified in [
52]. Compared with Chinese culture, the trust between French users is more affected by intimacy, which is similar to Hypothesis 3 and also reflects the prominent influence of intimacy in individualism culture.
6.2. The Impact of Culture Dimension on Social Interaction
According to the result of Hypothesis 5, the higher uncertainty avoidance of social commerce users will lead to the increase of perceived risk in the context of both Chinese and French culture, which is consistent with that of traditional e-commerce. Additionally, Hypothesis 6 verifies that uncertainty avoidance has a positive impact on subsequent trust-building in both cultures. Downey’s research [
10] may be applied to explain this phenomenon: after eliminating the worrying factors, users with higher degree of uncertainty avoidance will build higher trust in the future. The conclusion of Hypothesis 6 further complements the previous debate [
10,
39] about the effect of uncertainty avoidance on trust.
Hypothesis 7 concludes that the degree of individualism has a significant impact on the user intimacy in Chinese social commerce, but not in France. In Chinese group, individualism, contrary to expectation, leads to a higher degree of intimacy among consumers. One possibility is that intimacy is more important in individualistic than in collectivistic crowds [
53]. The scarcity of intimacy for individualists makes the individualistic users treasure their social relations in online community. This finding reflects the complexity and contradiction of the characteristics of social commerce users. Nowadays, consumers are getting rid of their traditional persona settings, and there is a trend of mutual influence of consumption concepts. A joint study by Ali Research Institute and Boston Consulting finds that the sales growth of male cosmetic products is 1.5 times that of overall cosmetics, and consumers over 40 years of age are buying more high-end outdoor goods than the average of other age groups. In modern urban society, there is a group of people who have many social activities because of their occupation and other reasons, but they are eager to be alone in their hearts. The majority of people with such characteristics are younger generation, and they are the main users of social commerce. These social phenomena provide a realistic explanation for Hypothesis 7.
6.3. Key Findings
In summary, the key results of this research are that the impact of perceived risk on subsequent purchase intention in social commerce will be transferred by trust and intimacy to a certain extent. That is to say, perceived risk may have less effect on purchase intention than trust and social relations in social commerce. Besides, the intimacy between users contributes to trust-building, and both of their positive impacts on purchase intention would show distinct effects in different cultures. The influence of trust on Chinese users’ purchase intention is more obvious, while intimacy can stimulate French users’ purchase intention more. This may be related to the general lack of trust in the platform among Chinese users and the relatively loose interpersonal relationships among French consumers. This shows that scarce resources (trust for Chinese users, and intimacy for French users) play a more important role in decision-making. Moreover, cultural dimensions are proved to have a significant effect on users’ social interaction. Although high uncertainty avoidance brings perceived risk, it can promote subsequent trust-building.
7. Conclusions
Based on the social interaction theory and Hofstede’s cultural dimensions theory, this paper studies how social interaction affects consumers’ purchase intention, and explores the antecedent impact of cultural dimensions on social interaction.
7.1. Theoretical Contributions
The contribution of this paper is twofold. First, we incorporate risk perception, trust, and intimacy into the research model, and find that social interaction has different effects on users’ purchase intention in different cultures, which further complements the research on purchase intention in social commerce. Second, based on the cultural dimensions of uncertainty avoidance and individualism/collectivism, we explore the influence of cultural dimension on social interaction, which helps to explain the current debate in cultural dimension research, such as whether high index of uncertainty avoidance will hinder the trust-building.
7.2. Practical Implications
For one thing, in view of the core role of social interaction in social commerce, managers should direct users’ attention to social functions, such as creating common topics, creating interest-based communities, and creating more interesting ways of content sharing, and finally to enhance consumer stickiness. Meanwhile, the involvement of social relationship helps to solve the difficult problem of trust-building between buyers and sellers in e-commerce. Behind social interaction is the trust value of the whole platform, and trust is the catalyst of platform transactions. In [
3], the constant comparison analysis also suggests that trust is the essence for the continuity and sustainability of social commerce. Therefore, it is necessary to establish an effective trust mechanism in social commerce community. For another, in the context of the rising cross-border e-commerce, a single management strategy does not apply to users in different cultural backgrounds. In order to enter a new market with different culture, website functions and social promotions should be designed accordingly.
7.3. Limitations and Future Research
The samples of this research have limitations. More data from different cultural backgrounds could be collected in the future to enhance the generalization of the research findings. With the deepening of globalization, multicultural integration and collision (such as in North America) will also have a certain impact on social commerce. Therefore, other kind of user behavior in multicultural context is still a topic worth exploring.