Applying the S-O-R model to explore impulsive buying behavior driven by influencers on social commerce websites

Social commerce (S-commerce)

S-commerce is a business type that combines social media and e-commerce. Consumers use social media and content others post to help themselves make decisions and purchase goods online (Moghddam et al., 2024). Nevertheless, the conceptual separation between s-commerce and traditional e-commerce goes beyond technological integration alone. Traditional e-commerce emphasizes transactional efficiency and product-centered interactions primarily. S-commerce fundamentally reshapes the buying process by integrating social interactions, peer influence, and community-driven decision-making within commercial contexts (Ibrahim, Ariyanti & Iskamto, 2025). This evolution produces what academics describe as “social shopping.” Here, purchasing decisions receive influence from social networks, user-generated content, and collective intelligence instead of solely from product attributes and price comparisons (Moghddam et al., 2024). Social influence theory (Hazari, Talpade & Brown, 2024) provides the theoretical basis for s-commerce. This framework recognizes three influence categories: compliance (behavioral change due to rewards/punishments), identification (behavioral change to maintain relationships), and internalization (behavioral change due to value alignment). Traditional e-commerce operates predominantly through compliance mechanisms like price incentives and promotions. S-commerce employs all three influence categories via social elements such as peer recommendations, influencer endorsements, and community participation (Zhu, Tan & Panwar, 2024). This complex influence framework creates distinctive pathways to impulse buying that are unavailable in standard e-commerce settings.

These platforms based on social media to expand into S-commerce are called S-commerce websites. On the other hand, people increasingly rely on online information to understand products and services when shopping online. However, information such as likes, reviews, and ratings can no longer meet the needs of consumers (Ho & Rajadurai, 2020). Live streaming has been developed and introduced into e-commerce to compensate for the lack of face-to-face and real-time interaction online. Through live streaming, consumers can chat with sellers in real-time, thus stimulating consumers’ interest and purchase intention (Clement Addo et al., 2021; Ho & Rajadurai, 2020).

S-commerce websites such as YouTube, Facebook, Line, and Instagram, which are widely used today, have post and live-streaming functions. These platforms have a significant impact on consumer purchase decisions. Many researchers have discussed online impulse buying behavior. For instance, Xu, Zhang & Zhao (2020) executed their analysis through the S-O-R theoretical framework integrated with dual systems theory to explore how social interaction mechanisms and self-regulatory processes influence consumers’ impulsive purchasing behaviors. This study argues that impulse buying behavior on S-commerce websites should also be explored.

Influencer marketing

Influencer marketing refers to a marketing strategy that uses the influence of key opinion leaders to drive consumers to make purchasing decisions (Choi, Wu & Lee, 2025). Marketing and media are interdependent. Marketing companies use media to push advertisements to attract target audiences. In addition, celebrity endorsement is one of the popular types of marketing (Lou & Yuan, 2019). However, with the changing times, social media has become people’s main information source. This has led to the rise of celebrities, such as influencers, defined as those with a large following on social media (Haenlein et al., 2020). Research indicates that netizens post their creations (e.g., text or video) on social media and have large followings, while consumers believe they are more trustworthy than traditional celebrities (Jin, Muqaddam & Ryu, 2019). For instance, Al-Shehri (2021) examined how social media influencers impact consumer purchasing intentions. The study also explored whether gender differences among consumers or influencers influenced these decision-making processes. Another example is Lou & Yuan’s (2019) study, which constructed a comprehensive theoretical framework called the social media influencer value model. This framework explains the roles of advertising value perceptions and source credibility factors. On the other hand, social media can communicate faster, have a broader target audience, get more feedback, and cost less than traditional media (Lou & Yuan, 2019). Advertisements created and published by influencers can also gain consumers’ trust more than traditional advertisements. Therefore, influencer marketing is considered more cost-effective and effective (Lim et al., 2017; Lou & Yuan, 2019).

Some researchers have pointed out that influencers regularly publish works on social media, which makes fans highly favor the brands or products they introduce, affecting online purchase intention and impulse purchase behavior. For instance, research by Trivedi (2021) compared the relative effectiveness between attractive celebrity influencers and expert influencers in promoting online consumer-brand engagement behaviors. This subsequently resulted in spontaneous online purchasing decisions. Another example is research conducted by Lim et al. (2017) that investigated the effectiveness of social media influencers by analyzing essential factors. These factors included source credibility perceptions, source attractiveness attributes, product congruence alignment, and meaning transfer processes. According to the previous literature, this study focuses on the factors that lead to impulsive buying behavior on S-commerce websites due to influencers.

Stimulus-organism-response model (S-O-R model)

The S-O-R model was first proposed by Woodworth (1929). It has been applied to environmental psychology to reveal that environmental stimulus (S) affects individuals’ cognition and emotion (O), which in turn affects individual behavioral responses (R) (Mehrabian & Russell, 1974). The S-O-R model demonstrates a notable theoretical enhancement compared to traditional behaviorist stimulus-response models. It distinctly integrates the mediating influence of internal psychological processes. The S-O-R model merges these approaches by identifying the organism component as the vital mediating mechanism. This mechanism facilitates the transformation of external stimuli into behavioral responses.

In recent years, Yang et al. (2021) have used S-O-R models to explore the factors affecting impulse buying behavior. Yang et al. (2021) empirically assessed scientific hypotheses regarding whether Internet utilization elevates life satisfaction levels among elderly populations. The investigation examined whether differences appear in the impact of Internet usage on elderly well-being outcomes. Moreover, it determined through which mechanisms Internet engagement affects elderly individuals’ life satisfaction measures. Moreover, Chan, Cheung & Lee (2017) argue that online environmental stimuli are the key factors that affect impulse buying. Therefore, the present research uses the S-O-R model as a framework to discuss impulse buying behavior. Next, the three concepts of stimulus, organism, and response are discussed further.

Response

The response is defined as outcomes and decisions made by individuals based on cognition and emotion (Mehrabian & Russell, 1974). On the other hand, impulse buying behavior is an important issue in consumer behavior research. Impulse buying behavior is the actual behavior of consumers who experience environmental stimuli and undergo psychological transformation (Chan, Cheung & Lee, 2017). Hence, this study uses impulse buying behavior as the response in the research model. Past research has shown that positive affect (Chen & Yao, 2018), flow state (Ming et al., 2021), and emotional attachment (Chen, Yeh & Lee, 2021) are all related to impulse buying behaviors. As mentioned earlier, this study proposes the following hypotheses.

H8: Positive affect positively impacts impulse buying behavior.

H9: Flow state positively impacts impulse buying behavior.

H10: Emotional attachment positively impacts impulse buying behavior.

This research uses the S-O-R model as a framework to explore the factors that lead to the impulsive buying behavior of influencers on S-commerce websites. This research model was constructed through literature review and hypothesis development, as shown in Fig. 1.

Research subjects and data collection

In this study, influencers are defined as creators and entertainers with over 10,000 followers and who have included advertorials in their creations or posts. This study used an online questionnaire for data collection.

Data was collected using a structured online questionnaire delivered through the Qualtrics platform for 4 weeks (March–April 2023). Participants were obtained through a multi-stage strategy combining convenience and snowball sampling techniques. Initial recruitment employed social media advertisements on Facebook, Instagram, and LINE, targeting users aged 18–45 who followed at least one influencer and had completed purchases through S-commerce platforms within the preceding 6 months. Screening questions confirmed eligibility requirements before granting survey access. The questionnaire took approximately 15–20 min to finish, with progress indicators and attention checks integrated throughout to sustain engagement and ensure data quality.

This study was reviewed by the institutional ethics committee at the National Kaohsiung University of Science and Technology, Taiwan. It was classified as exempt from formal ethical approval, as it involved minimal participant risk and solely relied on previously collected data without engaging directly with individuals. Prior to data collection, individuals were provided with comprehensive written information detailing the study’s aims, procedures, and any potential risks. Written informed consent was obtained, confirming that participation was voluntary and based on understanding the research purpose and participants’ rights.

Comprehensive data cleaning protocols were applied to ensure response validity. Initial screening removed 48 responses (10.7% of total submissions) according to established exclusion criteria: (1) incomplete responses containing more than 20% missing data, (2) responses finished in under 8 min (suggesting inadequate attention), (3) straight-line responses throughout multiple construct scales, (4) failed attention checks incorporated within questionnaire sections, and (5) responses from participants who failed to meet eligibility standards (no influencer following or S-commerce purchase history). Supplementary quality assessments included analyzing response patterns for outliers using Mahalanobis distance (p < 0.001) and detecting multivariate outliers. The final dataset contained 404 valid responses, achieving an 89.3% retention rate following data cleaning protocols, ensuring reliable statistical analysis while preserving sample representativeness.

Research design

This study used an online questionnaire. The design of the questions for each of the constructs refers to previous research, including questions about discounted prices and scarcity were taken from Chen & Yao (2018); questions about review quality, observed learning, and positive affect were taken from Xu (2020); questions about social presence were taken from Wang et al. (2021); questions about sense of belonging were referred to Teo et al. (2003), Lin (2008), and Zolkepli & Kamarulzaman (2015); the questions regarding flow state were referred to Ming et al. (2021), and Wang et al. (2021); the questions about entertainment and informativeness were taken from Ki et al. (2020); the questions about emotional attachment were referred to Aw & Labrecque (2020), and Ki et al. (2020); the questions about impulse purchase behavior were referred to Chen & Yao (2018), Xu (2020), and Ming et al. (2021). In addition, this study used the Likert seven-point scale. A total of 7 means strongly agree and 1 means strongly disagree. After the questionnaire was designed, it was translated into Chinese, and experts were invited to assist in revision to improve the quality of the questionnaire. The questionnaire items of this study are indicated in the appendix section of this research.

Descriptive statistical analysis

Reliability, validity, and model fit

Next, this study examined the measurement model using confirmatory factor analysis (CFA) and analyzed the reliability, convergent validity, and discriminant validity according to Anderson & Gerbing (1988). Fornell & Larcker (1981) proposed using composite reliability (CR), Cronbach’s alpha, and factor loading to analyze questionnaire reliability. These values must be greater than 0.7; however, values greater than 0.6 are acceptable for exploratory research (Hair et al., 2019). Table 3 shows that CR, Cronbach’s alpha, and factor loading are greater than 0.6, indicating the questionnaire is reliable.

Convergent validity is judged based on the average variance extracted (AVE), which should be greater than 0.5. Table 3 shows that the AVE ranges from 0.571 to 0.796, indicating the questionnaire has convergent validity. In addition, Fornell-Larcker and heterotrait-monotrait ratio (HTMT) were used to examine the discriminant validity. Fornell & Larcker (1981) proposed that the square root of AVE should be greater than the correlation coefficient between the construct and other constructs in the model. HTMT refers to the estimation of correlations between constructs. An HTMT less than 0.90 has discriminant validity (Henseler, Ringle & Sarstedt, 2015). Tables 4 and 5 show that the values meet the criteria, indicating that the questionnaire has good discriminant validity.

Before examining the structural model, the research model should be examined for multicollinearity. Multicollinearity refers to whether a high correlation between constructs makes the estimation inaccurate. In this study, the variance inflation factor (VIF) was used to determine whether the model had multicollinearity. If the VIF is greater than 5, the research model has multicollinearity. Table 3 shows that the VIF values range from 1.239 to 3.718, indicating no multicollinearity.

Structural equation modeling analysis

Hypothesis testing

After testing the measurement model, it was confirmed that the research questionnaire met the criteria of reliability and validity. Then, the structural model was tested; the purpose was to check the relationship between the constructs, the path coefficient, and $R^2$ are used to detect. The path coefficient is the strength of the relationship between the constructs, and the larger it is, the stronger the relationship is; $R^2$ refers to the degree to which the independent variable explains the dependent variable, ranging from 0 to 1, and the larger it is, the greater the independent variable can explain the dependent variable. In addition, this study uses Bootstrapping to resample 5,000 samples to detect structural models.

As shown in Fig. 2, in this research model, the explanatory power of constructs, including positive affect, was 40.4% ( $R^2=0.404$), flow state was 60.8% ( $R^2=0.608$), the emotional attachment was 54.5% ( $R^2=0.545$), and impulse buying behavior was 4.25% ( $R^2=0.425$). On the other hand, the results of the path coefficient analysis are shown in Table 6. These stimulus (S) factors, including discounted price (β = 0.214, p < 0.001), scarcity (β = 0.192, p < 0.001), review quality (β = 0.253, p < 0.001), and observed learning (β = 0.230, p < 0.001), had positive and significant effects on positive affect belonging to the organism (O), so H1a, H1b, H2, and H3 were supported. These stimulus (S) factors, including social presence (β = 0.447, p < 0.001) and sense of belonging (β = 0.417, p < 0.001), had a positive and significant effect on the flow state belonging to the organism (O), so H4 and H5 were supported. These stimulus (S) factors, including entertainment (β = 0.295, p < 0.001) and informativeness (β = 0.524, p < 0.001), had a positive and significant effect on emotional attachment belonging to the organism (O), so H6 and H7 were supported. Finally, the organism (O) factors, including positive affect (β = 0.299, p < 0.001), flow state (β = 0.179, p < 0.01), and emotional attachment (β = 0.278, p < 0.001), had a positive and significant effect on impulse buying behavior belonging to response (R), so H8, H9, and H10 were supported.

Table 6:

Direct effect analysis.

Paths		Path coefficient (β)	T-value	P-value	Result
H1a	DP → PA	0.214	4.025	0.000	Supported
H1b	SCA → PA	0.192	3.967	0.000	Supported
H2	RQ → PA	0.253	4.785	0.000	Supported
H3	OL → PA	0.230	4.212	0.000	Supported
H4	SP → FS	0.447	8.524	0.000	Supported
H5	SOB → FS	0.417	8.105	0.000	Supported
H6	EN → EA	0.295	6.422	0.000	Supported
H7	IM → EA	0.524	11.412	0.000	Supported
H8	PA → IBB	0.299	5.309	0.000	Supported
H9	FS → IBB	0.179	2.526	0.012	Supported
H10	EA → IBB	0.278	3.766	0.000	Supported

DOI: 10.7717/peerj-cs.3113/table-6

Notes:

SCA, Scarcity; DP, Discounted Price; RQ, Review Quality; OL, Observational Learning; PA, Positive Affect; SP, Social Presence; SOB, Sense of Belonging; FS, Flow State; EN, Entertainment; IM, Informativeness; EA, Emotional Attachment; IBB, Impulse Buying Behavior.

Comparison of research findings

The research aims to apply the S-O-R model to investigate the factors that influence the impulsive purchase behavior of influencers on S-commerce websites. Furthermore, the sources of the stimulus were categorized into (1) posts, (2) live streaming or videos, and (3) influencers.

First, the factor about posts and message boards/chat rooms.

The present research findings show that discounted prices significantly impact the positive affect. These results are consistent with research by Lee & Chen-Yu (2018), who investigated the mediating role of affective responses to price reductions on consumer perceptions in online retail contexts. Their study utilized a between-subjects experimental design featuring four different discount levels, using denim products as the primary stimulus in a simulated e-commerce environment. By systematically manipulating pricing variables across artificial web-based retail platforms, the researchers established that emotional reactions to discounted pricing were a crucial mediating factor connecting promotional tactics to consumer evaluation processes, including value perception, quality judgments, and savings awareness. Significantly, their findings indicated that when affective responses to price reductions operated as intermediary factors, the emotional stimulation created by discount exposure improved consumers’ quality perceptions of the presented products.

Additionally, scarcity was determined to influence positive affect positively. This result can be contrasted with a previous study by Yu et al. (2024). Their research investigated the associations between perceived scarcity, self-efficacy, self-control, and delayed gratification. They administered a survey to Chinese college students. The findings demonstrated that perceived scarcity was negatively associated with individual self-efficacy, self-control, and delayed gratification. To a certain degree, their findings elucidated how perceived scarcity impedes gratification from motivational and cognitive perspectives and supported additional research on addressing perceived scarcity’s psychological and behavioral effects.

Furthermore, the current research analysis revealed that scarcity perceptions significantly enhance positive affective states among consumers.

Moreover, review quality and observational learning significantly affected positive affect, which is different from the findings of Xu (2020). They found that review quality had no significant effect on positive affect. Because too many reviews lead to information overload, consumers have a negative affect. However, this study suggests that consumers can select reviews more relevant to their needs to read in the message boards of current S-commerce websites, thereby generating a positive affect.

Second, there are factors regarding live streaming and videos. Influencers can use live streaming or video on S-commerce websites to unbox or place advertorials. Social presence was found to influence the flow state significantly. The current study results align with the research by Wei et al. (2024), who investigated the associations between social presence, self-confidence in academic capabilities, and flow on learner engagement in hybrid synchronous learning environments. Using a systematic literature review and thorough examination, the researchers confirmed that social presence, academic confidence, and flow significantly and directly affected learner engagement levels. Furthermore, social presence indirectly affected learner engagement by functioning through academic confidence and flow.

Furthermore, the current study’s results demonstrated that sense of belonging statistically affected flow experiences. These findings align with previous research by Bowles & Scull (2019), who conducted an extensive systematic review of literature from 1990 to 2016 to operationalize school connectedness and identify four component dimensions: attending, belonging, engaging, and flow. The authors then developed a sequential, four-level theoretical framework of school connectedness based on these identified dimensions. The researchers proposed that these four constructs represent the fundamental components required for engagement and theorized that flow experiences emerge due to students’ psychological attachment to their educational institution. Hence, based on the above comparison, the current study indicates that with the help of advanced technology, when consumers watch live streaming or videos, they feel like they interact face-to-face with influencers, creating a social presence. In addition, they can instantly leave messages or interact with others in the chat room to increase their sense of belonging. Consumers have a social presence and a sense of belonging and are easily attracted by the current situation and atmosphere, resulting in impulse buying behaviors.

Finally, the factors regarding influencers. This study found that the entertainment and informativeness of influencers had a significant impact on emotional attachment. These findings are somewhat similar to research conducted by Lee (2025). Their study aimed to examine the impact of augmented reality (AR) technology on consumer perceptions of food-related promotional materials. Participants were presented with AR-enhanced promotional content using an experimental approach and then completed self-assessment questionnaires. The results indicated that informational and entertainment values positively influenced consumer attitudes toward AR-based advertisements. Immersive experiences functioned as a full mediator in the relationship between informational content and attitudinal responses, while entertainment effects were only partially mediated through immersive mechanisms. Besides Lee (2025), previous studies also indicated a positive impact of entertainment and informativeness on flow (Ki et al., 2020; Sánchez-Fernández & Jiménez-Castillo, 2021). Because with the advent of the self-media era, the number of influencers have risen abruptly. Their creations and behaviors will be followed and imitated by fans.

In addition, the research results show that positive affect significantly influences impulse buying behavior. This finding is similar to previous research by Ibrahim, Ariyanti & Iskamto (2025). Their study examined purchasing behaviors within TikTok Shop’s Indonesian marketplace. The research’s main goal was to evaluate the statistical significance of sales promotional activities and hedonic shopping motivations on impulse purchasing behaviors among fashion product consumers in the TikTok Shop digital platform. The empirical results showed that promotional tactics and hedonic motivational elements significantly impacted consumers’ impulse buying behaviors.

Furthermore, the present study showed that flow substantially influenced impulse buying behavior. This finding can be contrasted with previous research by Husada, Prawiyadi & Andreani (2024). Their investigation examined the effect of hedonic browsing behaviors and flow experiences on Instagram concerning food and beverage content on spontaneous online purchasing decisions. The study utilized a survey approach to analyze these relationships. The empirical results demonstrated that flow experiences produced a statistically significant positive impact on online impulsive purchasing behavior. Likewise, hedonic browsing exhibited a statistically significant positive effect on spontaneous online buying decisions.

In addition, this study found that consumers’ emotional attachment to influencers can lead to impulse buying behavior. The present study findings can be contrasted with previous research by Akbar et al. (2020). Their research analyzed the role of impulsive purchasing behavior within consumer research models. The study concentrated on retail consumer groups, and the statistical evaluations revealed that emotional brand attachment negatively impacted impulsive buying behaviors while concurrently showing a statistically significant positive association with post-purchase cognitive dissonance.

Finally, in terms of this study’s mediating results, the current study used positive affect, flow experience, and emotional attachment as mediators. According to the results of the current study, flow experience, and emotional attachment were significant mediators between the indirect impacts of the independent constructs with the impulse buying behavior. The current result of employing positive affect as a mediating variable can be compared to an earlier study by Li & Peng (2021). Their study employed the SOR framework, integrating flow theory and attachment theory, to examine how live streamer characteristics and live streaming environmental factors affect users’ gift-giving intentions through emotional attachment and flow experiences. The empirical findings validated the mediating roles of flow experiences and emotional attachment. Therefore, the results showed that live streamer attributes can generate users’ emotional attachment and flow responses toward the live streamer, thus increasing users’ intentions to give gifts.

The current study’s findings also supported the significant mediating role of flow experience. This finding can be compared to earlier research by Tu, Huang & Tu (2025). Their study employed a cross-sectional survey approach with nonprofit sector employees to examine the impacts of social media engagement and mindfulness practices on psychological distress levels. The research’s main goal was to evaluate whether positive and negative emotional states mediate these relationships. The results verified that positive affect functioned as a significant mediator in the indirect relationships between social media engagement and mindfulness practices and psychological distress outcomes.

Theoretical implications

Next, the theoretical implications of this study are explained. First, earlier studies have analyzed the connection between S-commerce websites, live streaming, or influencer marketing and impulse buying behavior. As an illustration, Lee & Chen (2021) investigated the mediating effect of price discount affect, characterized as emotional activation generated by price reductions, within two different relational frameworks: initially, between price discounts and consumer perceptions covering perceived savings, quality, and value; and subsequently, between perceived value and purchase intentions in the online apparel retail setting.

However, these issues are related. Therefore, this study simultaneously examines the impact of S-commerce websites, live streaming, and influencer marketing on impulse buying behavior. This combined methodology tackles a significant theoretical deficiency by showing how numerous digital touchpoints collaborate synergistically to affect consumer behavior. The investigation strengthens the S-O-R theory by supplying empirical validation that stimulus complexity in digital settings demands multi-dimensional interpretation rather than single-factor evaluation. Moreover, this unified framework enriches the theoretical comprehension of omnichannel consumer experiences, demonstrating how diverse digital stimuli operate together to produce more impactful psychological responses than separate channels functioning independently.

Second, Hsu (2020) argues that consumers watching videos from YouTube accounts can generate a sense of belonging and influence impulse buying behavior through a flow state. This study extends Hsu’s (2020) viewpoint to explore live streaming and videos on S-commerce websites. The results show that consumers who watch live streaming or videos on S-commerce websites also develop a sense of belonging and significantly affect the flow state, influencing impulse buying behavior. The theoretical significance resides in confirming the cross-platform applicability of psychological mechanisms underlying consumer engagement. This discovery reinforces the theoretical basis of flow theory in digital commerce environments and offers empirical validation for the universality of belonging-flow pathways across various social media platforms. Furthermore, the research adds to social identity theory by illustrating how digital communities cultivate psychological attachment that surpasses conventional brand-consumer relationships.

Finally, the results show that entertainment and information significantly affect consumers’ emotional attachment to influencers, affecting impulse buying behavior. These results are the same as those found by Ki et al. (2020). This means that although the research area is different, entertainment and informativeness are important factors affecting consumers’ emotional attachment to influencers, affecting impulse buying behavior. This cross-cultural confirmation enhances the theoretical strength of para-social relationship theory in digital marketing environments, offering evidence for the universal relevance of entertainment and informativeness as core drivers of emotional connections between consumers and digital personalities. The uniformity across various geographical settings adds to the generalizability of attachment theory in modern digital commerce contexts.

Furthermore, this investigation provides several additional theoretical contributions. The research expands the conventional S-O-R model by integrating multiple organism variables concurrently, showing that positive affect (Tu, Huang & Tu, 2025), flow state (Husada, Prawiyadi & Andreani, 2024), and emotional attachment (Pang & Zhang, 2024) function as parallel mediating processes rather than competing routes. This multi-mediation methodology enhances our theoretical comprehension of how environmental stimuli convert into behavioral responses through intricate psychological mechanisms. The study also adds to impulse buying theory by recognizing specific digital precursors distinctive to S-commerce settings, such as observational learning and social presence, thus broadening the theoretical scope of impulsive consumer behavior beyond conventional retail environments (Ibrahim, Ariyanti & Iskamto, 2025). Lastly, the research progresses influencer marketing (Choi, Wu & Lee, 2025) theory by supplying empirical validation for the varying effects of different influencer-related stimuli on separate psychological states, providing a more sophisticated theoretical framework for understanding how various aspects of influencer content generate specific consumer reactions.

Practical implications

According to the research results, this study has some practical implications. First, the findings show that observational learning significantly influences positive affect, influencing impulse buying behavior. In addition, Lu et al. (2021) mentioned that observational learning observes the number of likes, shares, and reviews and the sales volume of products. In other words, consumers observe that sales volume also affects their positive affect and leads to impulsive buying behavior. However, most S-commerce websites do not offer sales volume to consumers. It is suggested that influencers can provide relevant data during live streaming. Alternatively, S-commerce websites can design such functions in the future. It is believed that it will be beneficial to product sales.

Beyond basic sales volume displays, platform developers should create advanced social proof systems incorporating real-time purchase alerts, peer comparison features showing “similar customers also purchased,” and dynamic availability indicators that refresh based on current stock levels. These elements should be tactically placed during high engagement periods identified through user behavior analysis.

For influencer marketing strategy enhancement, organizations should establish multi-level partnership structures that utilize the varying impact of stimulus channels (Choi, Wu & Lee, 2025). High-entertainment influencers should be assigned to emotional connection campaigns targeting lifestyle and aspirational merchandise, while high-informativeness influencers should concentrate on practical products requiring logical validation. Marketing professionals should execute flexible content approaches where promotional activities (reduced prices, limited availability) are coordinated with influencer content schedules to maximize positive emotion enhancement during live-streaming events.

Platform designers should construct engaging social presence (Mahmood, Jusoh & Nor, 2024) capabilities, including virtual collaborative shopping experiences, instant peer communication during live broadcasts, and competitive community activities that strengthen a sense of community. These capabilities should integrate artificial intelligence (AI)-powered customization that adjusts social interaction levels based on individual user flow state activators, establishing optimized psychological routes to purchasing decisions.

As mentioned earlier, entertainment and informativeness significantly affect emotional attachment and lead to impulse buying behaviors. Additionally, Casaló, Flavián & Ibáñez-Sánchez (2020) and (Martínez-López et al., 2020) found that the originality and uniqueness of influencers’ creations improve the willingness of fans to interact with them, follow their suggestions, and recommend their work to others. Companies that place product or brand information into influencers’ creations will be better than traditional advertising. Organizations should develop thorough influencer assessment systems that evaluate audience size and emotional connection creation potential through indicators such as comment sentiment evaluation, recurring engagement frequencies, and para-social bond strength measurements. Strategic collaborations should emphasize influencers who show reliable capability to produce flow states through content engagement and social presence stimulation. Influencer marketing will be one of the most important marketing strategies in the future.

Research limitations and future research directions

This study tried its best, but there are still some imperfections. For example, the respondents were mostly teenagers, students, and women, which may have biased the results. Impulse buying behavior should vary by gender or age. Also, there are many categories of influencers, such as by the number of followers or the nature of their creations (e.g., cuisine and cooking). However, this study did not explore the impact of different types of influencers in detail. Finally, impulse buying behaviors may lead to returns, but little research has focused on their relationship. This demographic bias potentially restricts the generalizability of results across various age cohorts, educational levels, and cultural settings where impulsive purchasing behaviors may differ considerably. Future investigations should utilize stratified sampling across varied demographic categories and incorporate multi-group examinations to explore limiting conditions.

Second, the research fails to control for essential confounding factors, including influencer classification (micro vs. macro influencers), product types (hedonic vs. utilitarian), and participants’ previous experience with S-commerce platforms. These uncontrolled elements may substantially moderate the connections between stimuli and psychological reactions, potentially exaggerating or concealing actual effects. Therefore, future researchers are advised to implement experimental frameworks controlling for influencer attributes, product categories, and user proficiency levels to strengthen causal conclusions and improve theoretical comprehension of when and how S-commerce stimuli most effectively prompt impulsive purchasing behavior.

Although this study utilizes methodologically sound cross-sectional survey approaches with suitable statistical examinations, the empirical framework lacks the complexity and refinement necessary for strong scientific inference and causal comprehension. Future research should embrace a longitudinal methodology to monitor the temporal progression of psychological state modifications leading to impulsive purchases.

Future studies should broaden this S-O-R framework to include emerging technological developments transforming S-commerce environments (Dabija & Frau, 2024). Incorporating generative artificial intelligence in influencer marketing creates compelling research possibilities, particularly investigating how AI-created content, virtual influencers, and customized recommendation systems alter traditional S-O-R mechanisms. The emergence of multisensory extended reality metaverse settings provides exceptional opportunities to examine S-O-R processes in immersive environments (Lăzăroiu & Rogalsk, 2024). Future investigations should explore how sensory enhancement in virtual shopping experiences intensifies or reduces traditional psychological mediators and whether virtual presence creates unique organism states beyond current conceptual frameworks. Furthermore, algorithmic predictive modeling and sophisticated customer behavior analytics facilitate real-time stimulus enhancement based on individual psychological characteristics. Research should investigate how dynamic customization of promotional strategies, social proof systems, and influencer alignment affects impulse buying patterns.

Finally, incorporating online trust, perceived risk, and purchase intention elements within the S-O-R framework would offer a more thorough understanding of consumer decision-making mechanisms, particularly addressing how trust mediates the connection between environmental stimuli and psychological reactions in S-commerce environments (Lăzăroiu et al., 2020).