Introduction

Driven by the prolonged COVID-19 pandemic and the upgrading of digital technologies, e-commerce has accelerated in recent years. With its continuous development, people’s reliance on e-commerce platforms has gradually changed from a simple consumption model to a complex social model. Social e-commerce has been on the cusp, and the integration of social traffic and e-commerce transaction has been deepening. Firms realized the importance of social attributes, and fully demonstrated the social attributes of social platforms in e-commerce. E-commerce socialization strengthens consumers’ participation in purchase. Consumers make decisions under the guidance of the platform’s own users, which changes consumers’ consumption patterns unconsciously, improves consumers’ trust in the platform, and ultimately affects consumer behavior. It fosters information interaction behaviors such as attention, sharing, discussion and recommendation on online platforms, further promotes consumers to make complex decisions and perform more unpredictable behaviors in online purchase process1. Nowadays, the scale of global social e-commerce industry is growing rapidly, as its user groups are exploding. Not surprisingly, the industry is also rapidly becoming more competitive. Considering the above factors, we propose that it is urgent and necessary to study the psychology and behavior of the users involved in social e-commerce to promote the scientific development of e-commerce.

The social e-commerce conforms to social interaction theory. According to it, human behaviour is influenced by internal psychological intention and external situational factors related to social interaction. Increasing wealth of social factors may promote or inhibit willingness to engage in certain behaviors2. The theory emphasizes that psychological intention is the root cause of individual behavior. It is in line with personality theory, which proposes that individual intention and behavior are governed by personality3. Therefore, in prior literature consumer personality has always been considered an important factor that influences consumer purchasing decision4,5.

Existing studies rarely discussed the long-chain relationship between consumer personality, online social interaction, and deep online consumption behavior, since its exploration is novel but difficult. To analyze the relationship involves multi-disciplinary knowledge such as consumer psychology, sociology and economics. Previous research focused on exploring two types of relationships. First, scholars discussed the relationship between consumer personality and purchase decision, adoption intention and consumption behavior, but lacked the consideration of pre-purchase information acquisition, post-purchase comments, information sharing and other social behaviors in purchasing context4,5. Second, the positive effect of social interaction on consumer behavior was discussed from the perspectives of opinion-based social interaction (e.g., online comment) and behavior-based social interaction (e.g., observational learning), but the antecedent factors (e.g., individual personality) of consumers’ online social interaction were not considered simultaneously6,7.

We contributed to the body of knowledge in two aspects to address the above gaps. First, on the basis of social interaction theory and Big Five personality theory, we integrated consumer personality, online social interaction and deep online consumption behavior into a research framework to systematically explore the internal (personality) and external (socializing) factors that affect consumers’ deep online consumption behavior in the context of social e-commerce. Second, we decomposed the online interaction of consumers into two types. The one is online social interaction that achieves the purpose of interactive experience through sharing, liking, and making friends, etc., which are not associated with a specific purchase action, but daily virtual social life8. The other is consumer engagement such as sharing online information, liking, commenting and recommending (or not recommending) satisfactory (or unsatisfactory) goods and purchase experience after specific purchase actions9. Since the latter is always associated with purchase behavior, we combined it with online purchase decision and defined them as a new construct, i.e., deep online consumption behavior. The adjective “deep” is to emphasize the strength of consumers’ willingness to online consumption and the degree of post-purchase consumer engagement. The newly created construct links online purchase decision and post-purchase consumer engagement together. It reveals that consumers purchasing goods or services would search online information and integrate their needs and preferences into account, and after purchase consumers are willing to express their opinions on the e-commerce platform, comment on manufacturer’s products and services, and even share their suggestions based on their purchase experience online, thus generating in-depth online interaction. We developed the construct to focus our research on consumers’ purchase decision and the positive effect of their post-purchase interactions. Creating a good experience of deep online consumption behavior for consumers is an optimal action to achieve the sustainable development goal of social e-commerce.

Our study therefore aims to answer two research questions. First, does consumer personality significantly influence deep online consumption behavior? Second, how does online social interaction play the mediating and moderating roles? Our results revealed the comprehensive effects of consumer personality and online social interaction on deep online consumption behavior, providing practical guidance for social e-commerce platforms and merchants to improve online marketing strategies and making theoretical contributions to consumer psychology and behavior.

The remainder of our article was structured as follows. "Theory" conducted a brief statement of basic theories to develop our research framework. "Literature review" displayed literature review, following which research hypotheses were developed in "Hypotheses". The methodology including sample and data, measures, reliability and validity, and descriptive analysis, were presented in "Consumer personality and deep online consumption behavior". Moreover, the results were summarized in "The mediating effect of online social interaction". Finally, we provided the conclusion, implications, limitations, and future research in "The moderating effect of online social interaction".

Theory

Social interaction theory emphasizes that psychological and behavioral interactions and effects exist between people, people and group, and different groups in social network10. These effects enhance individuals’ self-role expectation, guiding them to form self-awareness, generate needs, and perform actions. Social communication and interaction between individuals can help to enhance emotion and trust between individuals and acquaintances and even strangers, so that social and environmental factors have the opportunity to influence and change individuals’ attitudes towards certain things or behaviors and then promote some specific behaviors11. The idea of social interaction theory is that individuals’ social actions can affect their decisions and behaviors through the effects of a series of psychological mechanisms. According to the theory, consumers’ online consumption behaviors can be driven by online social interaction through three mechanisms. First, demand inducing mechanism. Through communication, the recommendations and preferences of others induce one’s potential needs10. Second, trust strengthening mechanism. Consumers have higher trust in people in the virtual world after online interaction, so that they may change their consumption behaviors in a compliant manner. Third, emotional compensation mechanism. People who communicate little in real life may get emotional compensation from online interaction and make consumption decisions that deviate from past habits due to the warm recommendations.

In addition to the external social factors, the internal factor that affects individual psychological mechanism is personality. According to Big Five theory, individual personality can be summarized into five dimensions, i.e., openness to experience, conscientiousness, extroversion, agreeableness, and neuroticism12. Openness to experience describes an individual’s cognitive style and represents their tolerance, curiosity and exploration of strange situations. Conscientiousness assesses the organization, persistence and motivation of individuals in goal-oriented behaviors and evaluates the ability of individuals to control themselves and delay the fulfillment of needs. Extroversion indicates an individual’s ability to obtain pleasure through interpersonal interaction. Agreeableness measures an individual’s attitude towards others and reflects the degree of importance an individual attaches to cooperation and interpersonal harmony. Neuroticism reflects an individual’s ability to regulate emotion. It is a sign of an individual’s negative and unstable emotions. The first four dimensions are generally regarded as positive personalities, while the latter often has negative effects on individuals13. In prior literature, the first four personalities were repeatedly considered as the antecedents of positive attitudes, abilities, and behaviors, while the latter’s effects were mostly tested to be insignificant14. Based on the theory, deep online consumption behaviors can be significantly affected by consumer’s personalities through three mechanisms. First, pleasure mechanism. Extroverted and open consumers can get pleasure from online interaction with strangers and providing purchasing information and emotional support to others5. Second, decision responsibility mechanism. Agreeable and responsible consumers make decisions and behaviors in an attitude of being responsible for others and themselves, so they are willing to share consumption experience with others online, and are also willing to obtain information from others to make optimal purchase decisions. Third, emotional catharsis mechanism. Emotionally unstable consumers may be more likely to share their shopping experience in a virtual environment, generate impulsive purchases and make exaggerated comments15.

Integrating the two theories, we found that consumer personality contributes to online social interaction and deep online consumption behavior, and online social interaction plays a mediating role in the relationship between consumer personality and deep online consumption behavior. In addition, consumer personality constitutes the internal cause of deep online consumption behavior, while online social interaction constitutes the external cause. According to the theory of internal and external factors, internal causes are the basis for individual’s intention and behavior, while external causes are the inducement conditions of intention. External causes can only produce effects through the mechanism of internal causes16. Therefore, it is concluded that online social interaction as an external factor plays a moderating role in the relationship between consumer personality and deep online consumption behavior. In conclusion, we proposed that consumer personality can significantly promote deep online consumption behavior, and online social interaction plays both mediating and moderating roles in the affecting path. Hence, the theoretical framework of this study was established, as shown in Fig. 1.

Fig. 1
figure 1

The research framework.

Literature review

Many scholars have studied online consumption behavior. From psychological perspective, Pu et al. (2022) proposed that consumers’ perception of product value and self-cognition of social identity would influence their continuous consumption of online educational products, thus proposing the value-attitude-behavior framework to analyze online consumption17. From the view of informatics, You et al. (2013) considered that consumers’ online social interaction is essentially an information search behavior, which promote their use of information and help them make decisions about online consumption18. From marketing perspective, Yang et al. (2018) pointed out that online consumption behavior may be motivated by a marketing strategy such as China’s “Double Eleven” consumption festival, which could create a specific context and atmosphere19. In addition, scholars have also paid attention to the process of online consumption behavior, focused on exploring online consumption intention20, attitude21, willingness22, and behavior23, as well as the cognition and actions before, during and after purchase24.

Compared with online consumption behavior, deep online consumption behavior pays more attention to the whole process, particularly the continuous behavior after consumption, such as making comments, sharing information and recommending products. Xia et al. (2021) summarized consumers’ deep online consumption behaviors as opening and browsing the shopping platform, buying products, and then collecting, evaluating, and sharing information in it25. They found that these behaviors are closely related to consumers’ personalities. Friedrich (2016) defined the behaviors and intentions of consumers involved in all aspects of social e-commerce as adoption behaviors26. The study focused on consumers’ social and shopping behaviors on social e-commerce platforms, like sharing, forwarding, commenting, and recommending and interacting. The concept resembles the construct deep online consumption behavior that was defined in this study. However, the concept we presented emphasizes consumption behavior over e-commerce adoption behavior, and focuses on a behavioral rather than a technical perspective. A few research have focused on the whole process and the depth of online consumption behavior. Most studies were focused on the mechanism of a specific action and were dedicated to answering what kind of consumers are more willing to post online comments after purchase27. Questions such as what information consumers generally share after online shopping28 and to whom they are usually willing to recommend the goods of which they had a good experience29 were focused and discussed.

Social media channels provide a critical opportunity for sharing electronic word-of-mouth communication, which has been considered a prominent factor in shaping consumer behavior30. In virtual cyberspace, there are interpersonal interactions and interactions between people and organizations (e.g., Internet service providers)31. The consumer psychology for social interaction has been captured by merchants and platforms, resulting in marketing models such as shopping anchors, live sales, and consumer decision portals (e.g., Xiaohongshu) that create an atmosphere of information sharing and integrate social interaction with online shopping32,33. Sestino (2024) found that in luxury shopping context, the high intelligence level of the environment positively affects consumers’ willingness to buy34. Studies have shown that some consumers not only like to receive relevant purchase information shared by others, but also like to share information with others35. Online social interaction between consumers and merchants, platforms, anchors, and other consumers has been shown to effectively stimulate and transmit positive emotions about online consumption intentions, which then induce deep online consumption behaviors36. The higher online social interaction leads to more frequent online consumption and deeper behaviors such as post-purchase comment and recommendation37,38.

Consumers’ positive personalities drive them to carry out online social interaction and deep online consumption behaviors. Liu and Campbell (2017) found that extroversion and openness to experience are significantly and positively correlated with individuals’ participation in various online social interactions, while conscientiousness, neuroticism, and agreeableness can only predict individuals’ participation in a few types of online social interaction39. According to Wang et al. (2012), individuals’ Big Five personalities and their traits such as narcissism, self-esteem, and pursuit of stimulation, are all predictors of their interaction in online social platforms40. Baik et al. (2016) collected data via Facebook and confirmed the significant influence of consumers’ personality traits (including extroversion, self-awareness, desire for uniqueness, and self-esteem) on their online purchase decisions41. Hermes and Riedl (2021) further pointed out that consumers with high scores in openness to experience and agreeableness would prefer to choose online shopping channels42. Finally, through a survey of 559 online brand community users, Marbach et al. (2019) found that individuals’ openness, extroversion, and altruistic personalities are significantly and positively associated with online consumer interaction behaviors (e.g., liking, commenting, and sharing), as consumers with these personality traits are better at discovering social and aesthetic value in online social interactions43.

Hypotheses

Consumer personality and deep online consumption behavior

Scholars have recognized the decisive influence of consumers’ personalities on their preferences, willingness, and behaviors44,45. For example, in terms of the consumption of green products, Duong (2022) proposed that people with agreeableness would care more about others and the environment, so they would be more inclined to consume environmentally friendly green products44. Individuals with conscientious personality would pay more attention to future development, so green consumption is their preferred choice. Furthermore, individuals with extroverted personality are eager to support others, they therefore display pro-environmental characteristics and support green consumption probably. Moreover, individuals with high-scoring open personality would show greater inclusiveness and acceptance of green goods than those with low openness because they are more willing to accept new things. However, fear and anxiety about new things may prevent neurotic individuals from accepting and consuming green products. In terms of automobile consumption decision-making, O’Connor et al. (2022) found that extroverted and open-minded young people would value the brand, style and performance of automobile products, while agreeable and conscientious individuals would care more about the safety and practicality of automobile products45. As can be seen from the two examples above, consumers’ personalities do influence their preferences, and in turn the personalities would affect consumption decisions through concerns about details (e.g., pro-environment and novelty). However, the effects of different personalities on preferences are heterogeneous. That is, people with high extroversion and openness to experience and those with high agreeableness and conscientiousness have large differences in their perception of products (including perceived usefulness, perceived ease of use, perceived enjoyment, and perceived risk) and preferences.

Extroversion traits, such as being talkative, self-assurance, and helpful, tend to make people more prosocial. For one thing, the effect of trust. Extroverts have stronger trust and adaptability in strangers and strange things in the environment, which makes them more willing to accept new things such as online shopping and accept suggestions from strangers on the Internet than introverts46. For another, the effect of emotional disclosure. After online shopping, extroverted consumers hope to share good or bad consumption experience on the Internet, so as to vent their emotions. They may also want to share their information and knowledge with others online, and recommend the goods or services with which they are satisfied to net friends. Sub-hypotheses were then proposed: consumers’ extroversion was positively associated with (H1a) online purchase decision and (H1b) post-purchase consumer engagement.

Individuals with openness to experience trait exhibit higher cognitive abilities, curiosity, and flexibility in action, which make them more willing to perceive and accept new things47. For one thing, te effect of curiosity-hunting. Open consumers are willing to spend time searching for information and products on the Internet, and buy products that are unfamiliar to them but can make them curious at the risk48. McElroy et al. (2007) confirmed that individuals with high-degree openness to experience do tend to experience new patterns of online shopping49. Moslehpour et al. (2018) also stated that the openness to experience is an important driver of consumers’ willingness to consume online50. For another, the effect of desire for exploration. In order to better explore product knowledge, open individuals will actively seek people who can communicate and carry out online communication. When they complete online purchase and consumption, they will verify the product knowledge they have explored with the knowledge shared online. After that, they may initiate a series of actions, such as liking, commenting, etc. Sub-hypotheses were therefore proposed: consumers’ openness to experience was positively associated with (H1c) online purchase decision and (H1d) post-purchase consumer engagement.

Online consumption behavior is driven by hedonic and utilitarian motivators46. For one thing, the effect of hedonic motivator. Individuals with high agreeableness are friendly to others, and have high prosociality and cooperation. Hence, in the context of online consumption, they are willing to share experience and recommend products to help others. These actions in turn make themselves easier to receive emotional support from the environments. Online interaction makes agreeable individuals get happiness from posting online comments and getting feedback46. For another, the effect of patience. Agreeable people are more patient in searching information online and comparing the information from various online channels, thus making better consumption decisions51. That is, agreeableness allows consumers to appreciate their utilitarian motivations for purchasing. Hence, sub-hypotheses were proposed: consumers’ agreeableness was positively associated with (H1e) online purchase decision and (H1f) post-purchase consumer engagement.

Conscientious individuals have strong rationality and sense of responsibility. For one thing, the effect of rationality. Moslehpour et al. (2018) demonstrated that conscientious individuals would spend a great deal of energy collecting information about goods in the hope of making a more accurate judgment about perceived usefulness and perceived ease of use of goods52. In addition, they are more open and tolerant of gathering information from environments. People’s intentions towards and behaviors of online consumption are significantly influenced by their conscientious personality. The relationship is mediated by perceived usefulness, perceived ease of use, and openness to experience. For another, the effect of responsibility. Strong conscientious individuals are more open to commenting on products and sharing purchase experience, especially when they disagree with what the merchants or platforms say online53. Sub-hypotheses were then proposed: consumers’ conscientiousness was positively associated with (H1g) online purchase decision and (H1h) post-purchase consumer engagement.

The relationship between neuroticism and deep online consumption behavior is controversial. For one thing, positive effect. Consumers choose online shopping primarily to meet their happiness or functional needs, and individuals with higher neuroticism have a stronger pursuit of satisfying such needs. Therefore, neuroticism would strengthen consumers’ intention towards online consumption54. Even when external pressure is enhanced, the anxiety of neurotic individuals is strengthened, and then they tend to choose impulsive online shopping for relieving the distress caused by anxiety55. In the impulse buying situation, the neurotic people even prefer to share their anxiety with others, leading others to make similar decisions on irrational online shopping. They could gain peace of mind and relieve anxiety from others’ followings. That is, in this case they will be crazy to get information from the Internet, and also post a lot of comment information. For another, negative effect. People with high neuroticism show more anxiety, unease, and nervousness about environments, which are reinforced by the virtual experience, resulting in a great distrust of online goods and services56. They lack the patience to search for knowledge online, and are friendly to communicate and share knowledge with others. In addition, there also exists a conciliatory view. Simanjuntak et al. (2017) proposed that consumers with neurotic personality rarely purchase online goods based on comments and blog recommendations57. However, high neurotic individuals show strong identification and paranoia, enhancing their willingness to consume online when blog contents are very professional. All in all, we proposed that the effect of neuroticism on online purchase decision and post-purchase consumer engagement is unclear.

In summary, the following hypothesis was proposed. H1: Except for neuroticism, consumers’ Big Five personalities positively affect their deep online consumption behaviors.

The mediating effect of online social interaction

Consumers’ personalities affect their online social interactions, and the differences in personality lead to differences in the pattern and channel of online social interaction. First, Kraut et al. (2002) conducted a research hypothesis and verified that the more extroverted individuals are, the more actively they participate in online social activities58. In turn, the success of online social activities strengthens the development of extroversion. Second, the internet as a more open environment makes open individuals easier to meet the needs of curiosity and risk-motivation. People with high openness to experience are also ready to socialize online59. In addition, high agreeable individuals have a natural sense of trust in others, which makes them more comfortable in online social interaction60. What’s more, conscientious individuals have a greater thirst for information and knowledge. The internet can satisfy the desire more efficiently in current information environments. Therefore, conscientious individuals use social media more frequently for online social interactions61. Finally, those with neurotic personality develop a completely different online self in a virtual environment, bringing them new joy62. High neurotic individuals are also inclined to engage in online social interactions. Several scholars disagreed with this view. They proposed that high neurotic individuals are more anxious, depressed and distrustful in the network environment, making them more resistant to online social platforms63.

Online social interaction positively impacts consumers’ deep online consumption behavior. Xie et al. (2020) developed the concept of “social-consumption conversion” to explore the influencing factors and promoting mechanism of online users’ conversion from social context to consumption context64. They found that the characteristics provided by social e-commerce promote social-consumption conversion. Similarly, based on technology adoption model, Shen and Eder (2011) showed that the interaction in the context of social e-commerce stimulates consumers’ purchase intention, which ultimately changes into actual purchasing behavior65. Phang et al. (2013) revealed that consumers engaging in online social interaction have a significantly higher willingness to consume66. In addition, Hou et al. (2020) claimed that post-purchase consumer engagement positively correlates with online social needs67.

According to social interaction theory, post-purchase consumer engagement is dominated by online social interaction, and the embedded psychological mechanism of online social interaction is determined by consumers’ personalities. Integrating the above considerations, we proposed a hypothesis.

H2: Consumer online social interaction plays a significant mediating role in the relationship between consumer personalities (extroversion, openness to experience, agreeableness, and conscientiousness) and deep online consumption behavior.

The moderating effect of online social interaction

The theory of internal and external factors is suitable to explain the moderating role of online social interaction as an external factor in the relationship between consumer personality (as an internal factor) and deep online consumption behavior16. The network for online social interaction is not only an element of social e-commerce platform, but also can be separated from the social e-commerce process and become a pure communication carrier. The carrier creates favorable external conditions for stimulating online purchase and consumption.

First, amplification mechanism. As a dialogue field, the carrier enables individuals’ personalities to emerge and magnify the effect, promoting the change from potential purchase impulse to explicit purchase behavior. In the field of online social networks, consumers’ extroversion and openness to experience are released to stimulate the trust, emotion and altruistic motive. It is the mechanism of social e-commerce which creates opportunities to lead to impulse consumption by developing the interactions between different consumers as well as consumers and merchants68. Moreover, the skills and emotions accumulated during online social interaction make extroverted and open consumers more willing and able to share their post-purchase experience and knowledge efficiently. Consumers who prefer to like, comment on and recommend products after purchase are generally active online social media users69.

Second, interaction mechanism. In the context of online interaction, different communication situations, unfamiliar and diverse objects, and uncertainty and high risks thoroughly stimulate different personalities of consumers. These factors interact with each other and jointly influence online purchase decisions and post-purchase behaviors. Jinkyun (2021) proposed that the social presence generated by online socialization would interact with consumers’ conscientiousness, extroversion, agreeableness and other personalities, promoting consumers’ desire to consume in the environment of social e-commerce70. The empirical study of Zhou and Tang (2022) verified that consumers’ extroversion, openness to experience, and agreeableness and conscientiousness have positive effects on their behavior of online comments, and the effects are moderated by consumers’ position in the virtual network71.

We then developed the following hypothesis.

H3: Online social interaction moderates the relationship between consumer personalities (extroversion, openness to experience, agreeableness, and conscientiousness) and deep online consumption behavior.

Methodology

Sample and data

Our study is based on a questionnaire survey of Chinese consumers who have online shopping experiences. The questionnaires were distributed via Sojump (www.sojump.com) in April 2022. It is a professional questionnaire service platform which provides services such as online questionnaire design, retrieval, and data analysis, etc. Until now it has distributed more than 200 million questionnaires and received 16.4 billion responses. Questionnaires distributed through the platform follows the principle of simple random sampling, as each potential respondent has an equal chance of being selected to participate in the survey. Many Chinese scholars have completed data collection for academic research through this platform such as Rui (2017)72 and Li and Yu (2018)73. Now, about 10 million people fill out questionnaires on the platform every day. The sampling frame and procedures for the study are controlled by the platform professional service system. Every Chinese Internet user has the opportunity to be invited to answer our questionnaire. A total of 510 questionnaires were collected in the survey and 313 valid questionnaires were retrieved after manual screening to eliminate invalid ones, with a valid recovery rate of 61.37%. Demographic analysis was carried out on samples of valid questionnaires, and sample distribution is shown in Table 1. As the table shown, most of the respondents are individuals between 18 and 50 years old. It is the main consumer group in China, and the core group of Internet users. Therefore, the sample is representative for the subjects of the study.

Table 1 Distribution of the respondents (N = 313).

Measures

The questionnaire contained 35 items and was divided into two parts. The first part includes 4 items for collecting demographic information such as respondents’ gender, age, education and monthly disposable income. The second part displayed the items of core variables, including the explanatory variable of consumer personality, the mediating and moderating variable of online social interaction, and the explained variable of deep online consumption behavior. The measure of consumer personality followed Big Five theory, including the measures of extroversion, openness to experience, agreeableness, conscientiousness and neuroticism. The measure of deep online consumption behavior was organized from the aspects of online purchase decision and post-purchase consumer engagement. A five-point Likert scale was employed to measure core variables. Levels 1–5 successively indicate the degree of agreement between the items’ description content and the respondents’ actual situation. That is, 1 represents complete disagreement, and 5 represents complete agreement.

In order to control the total number of items in the questionnaire, consumer personality measurement items were selected from the minimalist version of Chinese Big Five Personality Inventory (CBF-PI) developed by Zhang et al. (2019)74. Three items were reserved for each consumer Big Five personalities. An example item of extroversion is “I am bored at big parties”. An example item of openness to experience is “I am a risk-taker and unconventional person”. An example item of agreeableness is “I think most people have good intentions”. An example item of conscientiousness is “I work or study hard”. Finally, an example item of neuroticism is “I often worry about that something bad is going to happen”. The items of online social interaction were selected and adapted from Cui and Wang (2020), including 6 items, with the example item of “I will communicate with others on the social shopping platform and maintain interaction“75. The measures of deep online consumption behavior were adapted from Hou et al. (2020)67 and Wang (2021)76. Online purchase decisions and post-purchase consumer engagements contained 5 items respectively. For example, an item for online purchase decision is “I consume online almost every week,” and an item for post-purchase consumer engagement is “I will share the shopping experience with my family, friends and net friends after online shopping”.

Reliability, validity and descriptive statistics

First, the exploratory factor analysis of original data showed that the factor loading of some items was below the threshold value of 0.4 and the overall consistency was poor. We therefore removed 5 items based on the analysis of factor loading, including one item of online social interaction, two items of online purchase decision, and two items of post-purchase consumer engagement.

Second, we performed a reliability analysis for the data of reserved items and obtained Cronbach’s α of the whole questionnaire as 0.822, which was between the recommended values of 0.8 and 0.9, showing a good reliability profile of the scale. The KMO value was 0.833, and the Bartlett sphere test p-value was less than 0.001, indicating that the data were suitable for factor analysis.

Finally, we carried out a confirmatory factor analysis and a descriptive statistical analysis. The results were shown in Table 2. In our study, the composite reliability (CR) of all variables was above 0.7, and the average variance extracted (AVE) was above 0.5, proving a good convergent validity. Each square root of AVE was larger than the correlations between the corresponding variable and others, showing a good discriminant validity. All correlations were less than 0.9, showing no serious collinearity problem.

Table 2 Reliability, validity and descriptive statistics.

Data analysis techniques

In our study, structural equation model was used to estimate the coefficients, and Bootstrap process was used to measure the mediating effect of online social interaction. AMOS was used for data analysis, and the maximum likelihood method was used for coefficient fitting. In SPSS software, we introduced regression analysis equations with interaction terms to measure the moderating effect of online social interaction.

Results

Path analysis

Considering the constitutive rather than reflective relationship between the construct consumer personality and its five dimensions as well as between deep online consumption behavior and its two dimensions, we constructed a second-order factor model to measure the path coefficients according to the research framework. However, after the fitting analysis of the initial model, it was found that the parameters of GFI, IFI, CFI and TLI were all less than 0.9 (as shown in Table 3), indicating that the model did not meet the requirement and the model needed to be improved. According to the prompt words of software, we removed two relationships with insignificant effects, i.e., the direct effect of consumer personality on deep online consumption behavior, and the relationship between the construct consumer personality and neuroticism. The variable neuroticism was therefore removed from the model. After the improvement of the model, the fitting parameters were significantly improved, as shown in Table 3. Though the TLI was below the recommended value of 0.9, but it was close to that. Overall, the model has reached an acceptable level with high reliability.

There are two points of concern regarding the above results. First, the effect of consumer personality on deep consumption behavior can only be played through the mediation of online social interaction, which highlights the importance of social interaction and probably explains why social e-commerce flourishes. The answer points to that the online field for social interaction is essential to release the vitality of consumer personalities. Second, the mechanism of neuroticism is different from that of the other four personalities, and it does not have a significant effect on online social interaction and deep consumption behavior. This may not mean that neuroticism does not have an important impact, but only reveals its complex and nonlinear mechanism. That is, the unstable emotions, anxiety and impulsiveness represented by neuroticism probably exert a double-edged sword effect. This reminds us that more careful social occasion and more conservative marketing strategies are needed when motivating highly neurotic consumers in the process of social e-commerce.

Table 3 The fitting parameters of the original and improved models.

The coefficients of the improved model were fitted. The results of path analysis was shown in Fig. 2, and the estimated coefficients were shown in Table 4. According to the figure, the direct effect of the higher-order construct consumer personality on the higher-order construct deep online consumption behavior was insignificant, indicating that hypothesis H1 was rejected. However, the relationships between consumer personality and online social interaction as well as between online social interaction and deep online consumption behavior were significant, indicating that the relationship between consumer personality and deep online consumption behavior could be completely mediated by online social interaction. Hence, hypothesis H2 was supported.

Fig. 2
figure 2

The results of path analysis.

Table 4 Unstandardized and standardized coefficients of the improved model.

* p < 0.100, ** p < 0.010.

According to the standardized coefficients in Table 4, we known that in the context of our study consumer’s openness to experience (estimate = 0.634, p < 0.100), conscientiousness (estimate = 0.547, p < 0.100) and agreeableness (estimate = 0.457, p < 0.100) contribute more to consumer personality than extroversion (estimate = 0.164, p < 0.100), while the effect of neuroticism is insignificant (p > 0.100). The weight of post-purchase consumer engagement (estimate = 0.930, p < 0.010) is greater than that of online purchase decision (estimate = 0.656, p < 0.010) in the composition of deep online consumption behavior. In addition, the relationships between consumer personality and online social interaction (estimate = 0.906, p < 0.100) and between online social interaction and deep online consumption behavior (estimate = 0.983, p < 0.010) are both very strong.

From the above results, some knowledge can be obtained. First, the attitude of liking to accept new things influenced by openness to experience is the fundamental basis for consumers to choose online purchase, so its effect is the largest. The willingness to spend time preparing and the willingness to help others dominated by conscientiousness and agreeableness constitute the driving factors of online purchase decision and post-purchase engagement respectively, so they also have strong effects. The fact that extraverts like to interact with new people may simply mean that they are good communicators, but this is less directly related to the willingness of online purchase, so its effect is smaller. As for the insignificant effect of neurotic personality, the reason may be its double-edged sword effect as mentioned above. The knowledge can provide inspiration for us to better classify and manage potential consumers with different personalities.

Mediating effect

We used the Bootstrap process to measure the mediating effect proposed by hypothesis H2. We performed 2000 repeated samplings. The method of maximum likelihood estimation was used. We set the percentile confidence intervals (PC) and bias-corrected confidence intervals (BC) at 95%. The results are shown in Table 5. The table shows that neither interval contains 0, and p < 0.010. Therefore, online social interaction has a significant mediating effect on the relationship between consumer personality and deep online consumption behavior, playing a strong and positive complete mediating role (estimate = 0.890, p = 0.000). Hence, hypothesis H2 was accepted.

Table 5 The mediating effect of online social interaction in the relationship between consumer personality and deep online consumption behavior.

In order to further explore the detail of H2, i.e. measuring the mediating effect of online social interaction in the relationships between the sub-dimensions of consumer personality and the sub-dimensions of deep online consumption behavior, we changed the second-order model into a first-order model. That is, we removed the two second-order latent variables and directly linked the sub-dimensions. In the new model, four consumer personalities were set as independent variables and two types of deep online consumption behaviors were designed as dependent variables. After adjusting the fit coefficient of the model, the Bootstrap process was used for repeated sampling, and the results of the mediating effect test were shown in Table 6. As can be seen from the table, the mediating effect of online social interaction on the relationships between consumer extroversion and online purchase decision as well as post-purchase consumer engagement are not significant ( confidence intervals contain 0 ). However, online social interaction could significantly mediate the relationships between consumer personalities (openness to experience, agreeableness, and conscientiousness) and deep online consumption behaviors (both online purchase decision and post-purchase consumer engagement), since the confidence intervals do not contain 0, and the p-values are less than 0.010. Through the mediating paths, consumer openness personality plays a stronger role than agreeable and conscientious.

Table 6 The mediating effect of online social interaction in the relationships between the sub-dimensions of consumer personality and the sub-dimensions of deep online consumption behavior.

Moderating effect

Several regression models with interaction terms were developed to test the moderating effects. They were fitted in SPSS 26.0 and the results were shown in Table 7. The explained variable for M1 and M2 is online purchase decisions, and the explained variable for M3 and M4 is post-purchase consumer engagement. M1 and M3 are models with control variables and explanatory variables, while M2 and M4 are added moderating variables and the interaction terms. The five panels in the table show the estimation results of moderating effects of online social interaction in the relationships between five consumer personalities and deep online consumption behaviors.

The direct effects of consumer extroversion on online purchase decision (estimate = 0.125, p < 0.100) and post-purchase consumer engagement (estimate = 0.205, p < 0.010) are significantly positive. The effects of agreeableness on online purchase decision (estimate = 0.352, p < 0.100) and post-purchase consumer engagement (estimate = 0.245, p < 0.010) are also significantly positive. Similarly, the effects of openness to experience on them (estimate = 0.239 and 0.402 respectively, p < 0.010) are significant. Moreover, conscientiousness has significant positive effects on online purchase decision (estimate = 0.151, p < 0.050) and post-purchase consumer engagement (estimate = 0.350, p < 0.010). However, the effects of neuroticism on deep online consumption behaviors are not significant. We thus concluded that although in the second-order structural equation model we did not confirm the significant direct effect of consumer personality on deep online consumption behavior, the direct effects of consumer personality sub-dimensions on deep online consumption behavior sub-dimensions are all significant when the constructs are downgraded. The effect of neuroticism was still excluded in estimation. Thus, the sub-hypotheses of H1 were all supported.

In the occurrence mechanism of online purchase decision, consumer extroversion (estimate = 0.103, p < 0.100), agreeableness (estimate = 0.119, p < 0.100), openness to experience (estimate = 0.116, p < 0.100), and conscientiousness (estimate = 0.116, p < 0.100) have significant positive interaction effects with online social interaction. However, the interaction between neuroticism and online social interaction is significantly negative (estimate=-0.138, p < 0.050). For promoting post-purchase consumer engagement, conscientiousness has a significant positive interaction effect with online social interaction (estimate = 0.124, p < 0.050), while other consumer personalities do not. Consequently, hypothesis H3 was partially accepted.

Table 7 The results of moderating effect test.

Conclusion

Findings and discussion

First, the direct effect of consumer personality on deep online consumption behavior is insignificant. Online social interaction completely mediates the effect. The finding has never been demonstrated in existing literature, deepening our understanding of the complex generating mechanism of deep online consumption behavior.

Second, consumer personalities including extroversion, openness to experience, agreeableness, conscientiousness, and neuroticism have significant direct effects on online purchase decision. However, the direct effects of the first four are positive, while that of the last is negative. The finding is consistent with existing studies77,78.

Third, consumer extroversion, openness to experience, agreeableness, and conscientiousness have significant positive effects on post-purchase consumer engagement, while the effect of neuroticism is not significant. The finding is partially consistent with Ul Islam et al. (2017)79, which found that five personalities have significant effects on post-purchase consumer engagement, but the effect of conscientiousness is negative. In addition, our results differ from Picazo-Vela et al. (2010)80, which confirmed the positive effects of conscientiousness and neuroticism but did not support the significant effect of extroversion, openness to experience and agreeableness on post-purchase consumer engagement. Our study offers a new perspective on the controversial issue.

Fourth, online social interaction can play a significant positive mediating role in the relationships between consumer personalities and deep online consumption behaviors. It confirms the value of creating a social e-commerce environment81. Different from the previous scholars which focused on fragmented studies39,82, our study helps uncover the complex affecting mechanism of consumer personality on deep online consumption behavior by way of integration.

Finally, consumers’ extroversion, openness to experience, agreeableness, conscientiousness and neuroticism can interact with online social interaction, and in turn significantly affect their online purchase decisions. However, only the effects of the first four personalities are positively reinforced by online social interaction, and the negative effect of neuroticism is negatively reinforced. In addition, the positive interaction of conscientiousness and online social interaction promotes post-purchase consumer engagement, but the interactions of other personalities and online social interaction cannot make significant effect. Since only a few of existing literature focused on the interaction effect between online social interaction and consumer personality in the context of online consumption68,70, we successfully built a new theoretical framework.

Implications

The theoretical contribution of our study to the body of knowledge is threefold. First, it introduced the theory of social interaction into online consumption research, emphasizes the important role of online social interaction in the generation mechanism of deep online consumption behavior, forms a new research framework of consumer personality - online social interaction - deep online consumption behavior, and demonstrates the mediating and moderating effects of online social interaction. It improves our understanding of the mechanism of online social interaction in the context of online consumption. Second, it integrates online purchase decision with post-purchase consumer engagement, expanding and defining the new concept of deep online consumption behavior. The proposal of this concept helps us understand the goals and functions of social e-commerce more systematically, which is conducive to guiding our future research direction in social e-commerce. Third, it provides some new insights and evidence into the Big Five’s decisive role in online consumption, providing a deeper understanding of controversial issues, such as which personality traits are the main drivers of online consumption behavior.

Our study inspires the development of social e-commerce and the improvement of online marketing. First, the goal of the development of social e-commerce and online marketing should not be limited to the promotion of online purchase decision, but should also pay attention to the promotion of post-purchase consumer engagement. Providing consumers with more satisfactory products, services and purchasing experiences, and allowing consumers to take the initiative to start likes, positive comments and recommending products are conducive to the sustainable development of social e-commerce. Second, social e-commerce should attach great importance to guiding potential consumers’ online social interaction. Daily online social interaction before purchase plays an important role in information sharing, emotional transmission and drainage in communication, enhances the stickiness between potential consumers and platforms and merchants, as well as gives full play to the positive effects of online social interaction on online consumption. Third, social e-commerce companies can consider the application of machine learning, social text mining and other technologies to the exploration and construction of a database of potential consumers’ personalities, and make precise improvements in online marketing strategies to enhance marketing performance based on an optimal understanding of consumer personalities.

Limitations and future research

Some limitations do exist in our study. First, all of the samples of our study were from China, and most of the respondents were young people. The samples with the potential problem of selection bias needed to fully comply with the requirements of random or stratified sampling. People with different cultures may have differences in consumption views and factors affecting consumption behaviors, which affect the universality of our results. Second, despite 313 valid questionnaires being collected, there is still a gap between the sample size and the large population of Chinese consumers. Third, the questionnaire for our study was collected at once, and so the data belong to horizontal data. A lack of time series data may affect the accurate inference of causality between constructs. Finally, in order to avoid potential covariance problem caused by excessive interaction terms and meet the need of measuring higher-order models, we did not synchronously consider the mediating and moderating effects in the structural equation model for parameter estimation. The development of the models, bringing about local optimization, may affect the accuracy of the overall model estimation, resulting in an overestimation of the mediating and moderating effects. In the future, we will improve sampling method, sample size, data collection procedure, and estimation method to optimize data analysis results and better guarantee the reliability of our findings. In particular, we plan to conduct confirmatory and comparative studies by collecting data from multicultural individuals in order to generalize the findings to non-Chinese populations. In addition, we are considering to enrich and extend the research framework, for example, replacing the Big Five theory with other personality theories such as HEXACO, or introducing more situational factors, e.g. culture and knowledge, as moderator variables.