How streamer attributes shape buying behaviors through emotional trust and viewer engagement

The rapid rise of livestream commerce has transformed how consumers interact with online shopping. Unlike traditional e-commerce, livestream enables real-time interaction between streamers and viewers, combining entertainment with social connection into a unified experience (Wongkitrungrueng and Assarut, 2020; Xu et al., 2020). In this dynamic environment, consumer decisions are influenced not only by product quality but also by streamer-related source credibility attributes, such as expertise, trustworthiness and attractiveness (Wang et al., 2022; Rungruangjit, 2022). According to source credibility theory (SCT), individuals are more likely to be persuaded by sources perceived as credible, suggesting that a streamer’s characteristics significantly affect consumer attitudes and behaviors (Hovland and Weiss, 1951).

In the marketing literature, trust has traditionally been viewed as a cognitive construct. However, in the livestream context, trust has evolved to include affective dimensions. Emotional trust, in particular, has emerged as a critical factor in shaping consumer responses (Li et al., 2024a; Xu et al., 2020). Furthermore, affective experiences such as brand love can enhance loyalty and willingness to pay a premium price, highlighting the importance of emotional pathways in shaping consumer behavior. Understanding how streamer attributes influence emotional trust and behavioral outcomes is therefore of increasing importance.

According to uses and gratifications theory (UGT), viewers engage in livestream to fulfill their needs for entertainment and social interaction, both of which can evoke emotional and cognitive reactions (Katz et al., 1973). Empirical studies have shown that entertainment and interaction positively affect viewer engagement and purchasing behavior in livestream environments (Tian and Frank, 2024; Zhang and Xu, 2024). These findings align with the affective-cognitive processing (ACP) framework (Edell and Burke, 1987), which posits that both emotional and rational evaluations work together to shape behavioral intentions. Recent studies have also emphasized that emotional trust plays an essential role in maintaining viewer engagement in livestream commerce (Zhou and Baskaran, 2025; Shih et al., 2024).

Although existing research has examined the effects of streamer attributes, such as trustworthiness, expertise, attractiveness, entertainment and interaction, on consumer behavior, most studies have focused on direct effects without exploring the psychological mechanisms that explain these relationships (Zhu et al., 2020; Koay et al., 2021; Sawmong, 2022; Zheng et al., 2023; Liu et al., 2023; Li et al., 2024b). In addition, many studies apply theoretical frameworks in isolation, which limits a holistic understanding of consumer decision-making in livestream commerce. Scholars in marketing have recently called for more integrated approaches and future research agendas that combine different perspectives to better capture complex consumer behavior in digital environments. This suggests a need for an integrated approach that considers both emotional and cognitive pathways in explaining consumer behavior.

Despite the growing body of research on livestream commerce, our understanding of how streamer-related cues translate into actual consumer behavior remains insufficiently understood. In particular, existing studies tend to conceptualize consumer buying outcomes as a single, undifferentiated construct, thereby obscuring important differences in the psychological processes underlying different forms of buying behavior (e.g. Ji et al., 2023; Li et al., 2024b; Li and He, 2025). This study addresses this limitation by explicitly distinguishing between impulsive buying and purchase behavior as two conceptually and temporally distinct outcomes. Impulsive buying refers to spontaneous, unplanned purchase decisions that occur during livestream sessions and are driven by immediate emotional responses, often without prior intention or extensive deliberation (Li et al., 2024b). This construct captures the situational and affective nature of decision-making in livestream environments. In contrast, purchase behavior reflects consumers’ actual purchase actions over time, including purchase frequency, monetary expenditure and continued purchase behavior (Hossain et al., 2023). While impulsive buying represents a momentary decision-making mode, purchase behavior captures the cumulative and sustained outcomes of consumer buying activities. Distinguishing between these two outcomes allows for a more comprehensive understanding of consumer decision-making in livestream commerce, where affective and cognitive processes jointly shape both immediate and longer-term behavioral responses.

To address these gaps, this study proposes and tests a dual mediation model in which streamer attributes, including source credibility characteristics such as expertise, trustworthiness and attractiveness, as well as gratification-related cues such as entertainment and interaction, are associated with both impulsive buying and purchase behavior through the mediating roles of emotional trust and viewer engagement. By integrating SCT, UGT and the ACP framework, this study provides a comprehensive understanding of how streamer characteristics shape consumer behavior in livestream commerce. This integrated approach not only contributes to the literature on digital consumer behavior but also advances the application of classical communication and media theories in a modern retail context.

According to SCT (Hovland and Weiss, 1951; Ohanian, 1990), the persuasiveness of a message is influenced by the perceived credibility of the source delivering it. This credibility is typically defined through three core attributes: expertise, trustworthiness and attractiveness. In the context of livestream commerce, these attributes reflect how knowledgeable, reliable and appealing a streamer appears to viewers. Prior research has shown that these source attributes influence consumers’ trust in the streamer, which subsequently shapes their behavioral responses (Lou and Yuan, 2019).

Unlike many earlier studies that predominantly focused on cognitive trust, the present study adopts the lens of emotional trust, a form of affective-based trust that reflects a viewer’s emotional reliance on and comfort with the streamer (Johnson and Grayson, 2005). This psychological mechanism functions as a mediating pathway through which source credibility influences consumer behavior. For instance, viewers who perceive a streamer as trustworthy or attractive are more likely to develop emotional trust, which in turn increases their likelihood of engaging in behaviors such as interacting during the stream or making a purchase. Thus, emotional trust serves as a key affective mediator that translates source cues into actual behavioral outcomes, aligning with the broader theoretical view that persuasive influence often operates through internal psychological processes rather than through direct effects (MacKenzie and Lutz, 1989). Taken together, SCT provides the theoretical foundation for explaining how streamer expertise, trustworthiness and attractiveness shape viewers’ emotional trust, which subsequently influences consumer behavior in livestream commerce.

UGT posits that individuals actively seek out media to fulfill specific psychological and social needs, such as entertainment, information acquisition, personal identity and social interaction (Katz et al., 1973). In the context of livestream commerce, viewers are not passive recipients but rather active participants who engage with streams to gratify these needs (Hilvert-Bruce et al., 2018). Among the most relevant motivations in this setting are entertainment and social interaction, which act as primary drivers of user engagement with both the streamer and the content.

UGT further suggests that these gratifications do not directly lead to behavioral outcomes such as consumption-related actions. Instead, their influence is channeled through intermediate psychological states, most notably, viewer engagement, which encompasses the cognitive, emotional and behavioral involvement of viewers with the livestream content and the streamer (Calder et al., 2009). For instance, entertaining or interactive livestream may enhance viewer engagement by capturing attention, evoking emotional responses and fostering parasocial interactions (Sjöblom and Hamari, 2017). This heightened engagement, in turn, increases the likelihood of downstream outcomes such as purchase behavior, continued watching, or word-of-mouth communication. Therefore, consistent with UGT, streamer entertainment and interaction influence consumer behavior in livestream through the mediating mechanism of viewer engagement.

The ACP framework provides a theoretical basis for understanding how emotional and cognitive responses interact to shape consumer decision-making (Edell and Burke, 1987). Rather than viewing cognition and affect as independent or sequential processes, this model emphasizes their simultaneous and interdependent influence on how individuals evaluate stimuli and form behavioral intentions. In advertising and persuasive communication contexts, affective reactions often guide attention and judgement before deeper cognitive evaluations occur (Pham, 1998).

In the context of livestream commerce, consumers’ responses to source attributes (expertise, trustworthiness, attractiveness) evoke affective responses (e.g. emotional trust) or stimulate cognitive involvement (e.g. viewer engagement), which then drive actual behavior. Emotional trust reflects a viewer’s emotional reliance and comfort with the streamer (Johnson and Grayson, 2005), while viewer engagement captures the cognitive, emotional, and behavioral involvement of viewers with the livestream content (Calder et al., 2009). Given its multidimensional nature, viewer engagement reflects a cognitive–affective response, where attention, enjoyment and perceived social connection are intertwined, aligning with the ACP perspective of integrated processing. These psychological processes facilitate the translation of perceived source attributes into behavioral outcomes like purchasing or engagement.

According to Shiv and Fedorikhin (1999), affective responses may even dominate consumer choices when cognitive resources are limited or when decisions are made in emotionally charged environments, both common conditions in livestream scenarios. Such affective and cognitive mediators are particularly relevant in livestream, where emotionally immersive and highly interactive environments amplify viewers’ psychological involvement (Hilvert-Bruce et al., 2018; Lim et al., 2020). Therefore, consistent with ACP theory, this study posits a dual mediation mechanism in which emotional trust and viewer engagement act as psychological conduits through which source attributes influence consumer behavior.

This study integrates SCT, UGT and the ACP framework into a unified model for livestream commerce. This integration is particularly relevant in emotionally immersive, interactive and real-time livestream environments, where viewers’ cognitive and affective responses occur simultaneously. SCT explains how streamer attributes (expertise, trustworthiness and attractiveness) serve as persuasive cues shaping viewers’ initial perceptions. UGT complements this by highlighting viewers’ motivations to engage with livestreams for entertainment and interaction, positioning viewer engagement as a key cognitive mediator. ACP provides a psychological mechanism illustrating how affective responses interact with cognitive involvement to influence behavior.

Overall, the integrated framework shows that streamer attributes (SCT) trigger affective responses (emotional trust), while entertainment and interaction (UGT) drive viewer engagement. Emotional trust also contributes to engagement, and these affective and cognitive processes (ACP) jointly lead to impulsive buying and purchase behavior. This integration demonstrates that these theories complement each other, with SCT providing source cues, UGT explaining engagement motivations and ACP capturing their joint affective–cognitive influence, providing unique insights into consumer behavior in livestream commerce.

In livestream commerce, streamer credibility plays a pivotal role in shaping viewer perceptions and behaviors. According to SCT (Hovland and Weiss, 1951; Ohanian, 1990), three core dimensions (expertise, trustworthiness and attractiveness) determine how persuasive a communicator is perceived to be. Expertise enhances emotional trust by increasing viewers’ confidence in the streamer’s knowledge (Zhou and Baskaran, 2025). Trustworthiness builds emotional trust by fostering a sense of security and transparency, which is crucial in livestream settings where real-time interactions with influencers enhance viewers’ confidence in their purchasing decisions (Ali et al., 2025; Wang et al., 2022). Attractiveness enhances the streamer’s likability and emotional appeal, thereby reinforcing emotional trust in viewers (Ji et al., 2023; Zhou and Baskaran, 2025).

Emotional trust, the belief that the streamer genuinely cares about viewers’ well-being, plays a key role in shaping consumer behavior. Unlike cognitive trust, which is based on competence, emotional trust stems from perceived benevolence and emotional connection (McAllister, 1995). This form of trust is particularly significant in livestream, where parasocial relationships form quickly. Emotional trust drives impulsive buying by reducing hesitation and increasing emotional arousal, leading to spontaneous purchases (Li et al., 2024a). Additionally, emotional trust influences overall purchase behavior, as consumers are more likely to reward emotionally trustworthy streamers with loyalty and purchases, even when other options are available (Xu et al., 2020). Source attributes (expertise, trustworthiness and attractiveness) contribute to these effects, and prior studies have highlighted their role in influencing impulsive buying (Zhu et al., 2020; Koay et al., 2021) and purchase behavior (Sawmong, 2022; Zheng et al., 2023). Accordingly, we propose the following hypotheses:

In livestream commerce, viewer engagement, defined as the cognitive, emotional and behavioral investment in a livestream (Hollebeek et al., 2014; Wongkitrungrueng and Assarut, 2020), plays a central role in shaping consumer responses. Highly engaged viewers are more attentive, emotionally involved, and cognitively focused on the livestream content, which strengthens their connection with both the streamer and featured products. Drawing on UGT (Katz et al., 1973; Hilvert-Bruce et al., 2018), viewers participate in livestreams to satisfy entertainment and social needs. Streamer-provided entertainment, such as humor, creativity, storytelling or visually stimulating content, elicits positive emotions and interactive behaviors, including replying to comments, mentioning viewers’ names or facilitating two-way communication, satisfy social and relational gratifications (Tan, 2008; Tian and Frank, 2024; Liao et al., 2023; Muntinga et al., 2011; Zhang and Xu, 2024). These experiences foster cognitive focus, emotional involvement and a sense of inclusion, encouraging viewers to interact more actively in the livestream.

Prior studies indicate that when viewers experience higher engagement, they are more likely to participate in spontaneous purchases and respond positively to the streamer’s content, affecting both impulsive buying (Dong and Tarofder, 2024; Luo et al., 2024) and overall purchase behavior (Hossain et al., 2023; Liu et al., 2023; Sawmong, 2022; Ma and Mei, 2018; Li et al., 2024b; Yuan et al., 2025). This occurs because engagement enhances attention, emotional involvement and a sense of connection with both the streamer and the featured products, highlighting its role as a psychological mechanism through which entertainment and interaction shape consumer behavior in livestream commerce. Thus, we propose the following hypotheses:

Beyond source credibility attributes, emotional trust also fosters viewer engagement. Emotional trust creates psychological safety and emotional connection, encouraging viewers to pay attention, interact and stay engaged during livestream (Zhou and Baskaran, 2025; Shih et al., 2024). This captures the affective-to-cognitive pathway emphasized by ACP theory, where emotional responses influence cognitive and behavioral involvement. Accordingly, we propose the following hypotheses:

In summary, emotional trust and viewer engagement function as serial mediators through which source attributes influence consumer behavior in livestream commerce. Grounded in SCT, UGT and ACP, and based on the above arguments, we propose the following hypotheses:

The proposed research model is presented in the Figure 1.

This research employed a quantitative survey method, using an online questionnaire distributed via Google Forms to collect primary data. The questionnaire was organized into four key sections: an introduction, demographic and psychological questions, screening questions and measurement items. To ensure the relevance and quality of responses, two screening conditions were implemented: participants had to be at least 18 years old and must have previously made at least one purchase through livestream. Those who did not meet these criteria were excluded, as their experiences would not align with the study’s objectives. Before the survey began, respondents were asked to recall and refer to their most recent livestream shopping experience. A brief definition of livestream commerce and an example of a streamer were provided to ensure a common understanding. A pilot test was conducted with 50 individuals who closely resembled the target audience to assess the clarity and reliability of the questionnaire items. All variables were measured using a five-point Likert scale (ranging from 1 = strongly disagree to 5 = strongly agree), and participants were asked to choose the option that best represented their opinions or experiences. To ensure response quality, the Google Form was configured to require participants to log in with a verified email address and restrict each account to a single submission. Data collection took place between September and October 2024, yielding a total of 465 completed questionnaires. After screening for completeness and validity, 386 responses were retained for the final analysis, with incomplete surveys removed from the data set. Sample profile is presented in the Table A1.

The study’s variables were derived from previous studies (see Table A2). Streamer expertise, trustworthiness and attractiveness were adapted from Rungruangjit (2022). Streamer expertise was measured using four items capturing viewers’ perceptions of the streamer’s knowledge, experience and competence related to the product or brand. Streamer trustworthiness was assessed through three items reflecting perceived sincerity, honesty and reliability of the streamer as an information source. Streamer attractiveness was measured with four items focusing on the streamer’s physical and vocal appeal and overall ability to attract attention. Streamer entertainment, interaction, emotional trust, and impulsive buying were adapted from Li et al. (2024a). Streamer entertainment was measured using three items assessing the perceived enjoyment and emotional value of the streamer’s performance. Streamer interaction was captured by four items reflecting viewers’ perceptions of the streamer’s responsiveness and relational closeness during livestream. Emotional trust was measured using four items that reflect the viewer’s belief in the streamer’s emotional care and sincerity. Impulsive buying was assessed using three items focusing on spontaneous, unplanned purchases made during livestream. Viewer engagement was operationalized based on Calder et al. (2009) and Hollebeek et al. (2014), comprising four items to capture cognitive, emotional and behavioral engagement. Purchase behavior was measured through four items adapted from Hossain et al. (2023), assessing both past and future purchasing actions in livestream contexts.

Descriptive statistics of the research sample is reported in the Table A1. The gender distribution was rather balanced, with 59.8% identifying as female and 40.2% as male. The majority of respondents (88.1%) belonged to the 18–27 age group, followed by 9.1% aged 28–43, 2.3% aged 44–59 and only 0.5% aged over 60. In terms of monthly income, 49.5% reported earning less than 3m VND, 31.9% earned between 3m and 7m VND, 12.2% earned between 7m and 15m VND and 6.5% had an income above 15m VND. Regarding their psychographic profile, 67.1% of respondents reported watching livestream for less than one hour per day, 23.6% for 1–2 h, 7.5% for 2–3 h and 1.8% for more than 3 h daily. In terms of livestream experience, 31.8% had started watching livestream within the past six months, 23.6% between six and twelve months ago and 44.5% had been watching for over a year. Table 2 presents the descriptive statistics.

A two-step approach was employed for data analysis using partial least squares structural equation modeling (PLS-SEM) through SmartPLS 4 software. This method is well suited for theory development, predictive analysis and examining complex causal relationships such as mediation effects. It also offers valuable insights into structural relationships in studies with relatively small samples and multiple constructs (Hair et al., 2017). The analysis began with an evaluation of the measurement model to assess the reliability and validity of the constructs, followed by the structural model assessment using SEM and bootstrapping with 10,000 samples.

Results for measurement model are reported in Table 1 Cronbach’s alpha and composite reliability (CR) values exceeded the recommended threshold of 0.70, indicating internal consistency reliability for all constructs (Hair et al., 2017). Additionally, all item outer loadings were greater than 0.70, while average variance extracted (AVE) values above the recommended threshold of 0.50, confirming convergent validity. Accordingly, all constructs demonstrated good reliability and convergent validity.

Discriminant validity was evaluated using both the Fornell–Larcker criterion and the Heterotrait-Monotrait (HTMT) ratio. As shown in Table 2, the square roots of the AVEs for each construct were greater than their corresponding inter-construct correlations, satisfying the Fornell–Larcker criterion. Additionally, all HTMT values in Table 3 were below the threshold of 0.90, further supporting discriminant validity (Henseler et al., 2015).

Before assessing the structural model, multicollinearity should be evaluated by examining the variance inflation factor (VIF) values. As shown in Table 1, all VIF values were below the threshold of 5.0, suggesting that multicollinearity was not serious (Hair et al., 2019). Additionally, Harman’s single-factor test was conducted to assess common method bias (CMB). The analysis revealed that the first factor accounted for 48.67% of the total variance, which is below the 50% threshold (Podsakoff et al., 2003), indicating that CMB is not a serious concern.

Table 4 presents the results for indirect effects. The results indicated that Streamer Trustworthiness and Streamer Attractiveness exert significant indirect associations with Impulsive Buying through Emotional Trust. Specifically, Emotional Trust significantly mediates the relationship between Streamer Trustworthiness and Impulsive Buying (H1b: β = 0.117, t = 3.534), as well as between Streamer Attractiveness and Impulsive Buying (H1c: β = 0.093, t = 2.923). In contrast, the indirect effect of Streamer Expertise on Impulsive Buying via Emotional Trust is not significant (H1a: β = 0.021, t = 0.983), thereby rejecting H1a. Regarding Purchase Behavior, none of the indirect paths through Emotional Trust have statistical significance. The mediation effects of Streamer Expertise (H1d: β = 0.009, t = 0.837), Streamer Trustworthiness (H1e: β = 0.049, t = 1.734) and Streamer Attractiveness (H1f: β = 0.039, t = 1.876) on Purchase Behavior via Emotional Trust are all insignificant. Therefore, H1d, H1e and H1f are not supported.

As presented in Model AEB, Entertainment and Streamer Interaction are significantly associated with both Impulsive Buying and Purchase Behavior through Viewer Engagement Viewer Engagement significantly mediates the effects of Streamer Entertainment on Impulsive buying (H2a: β = 0.093, t = 3.941) and Purchase Behavior (H2c: β = 0.089, t = 3.688). Similarly, Viewer Engagement mediates the relationships between Streamer Interaction and Impulsive Buying (H2b: β = 0.124, t = 3.663), as well as between Streamer Interaction and Purchase Behavior (H2d: β = 0.119, t = 4.092,). Therefore, H2a, H2b, H2c and H2d are supported.

The results further reveal that Emotional Trust is significantly associated with Viewer Engagement, which in turn affects Impulsive Buying and Purchase Behavior. Viewer Engagement significantly mediates the relationship between Emotional Trust and Impulsive Buying (H3a: β = 0.097, t = 3.306), as well as between Emotional Trust and Purchase Behavior (H3b: β = 0.093, t = 3.013). These findings indicate that Viewer Engagement mediates the relationships from Emotional Trust to both Impulsive Buying and Purchase Behavior, supporting H3a and H3b.

Finally, the results confirm significant serial mediation paths from Streamer Trustworthiness to Impulsive Buying (H4b: β = 0.036, t = 3.023) and Purchase Behavior (H4e: β = 0.034, t = 2.939), as well as from Streamer Attractiveness to Impulsive buying (H4c: β = 0.029, t = 2.537) and Purchase Behavior (H4f: β = 0.027, t = 2.243). In contrast, the serial indirect effects of Streamer Expertise on both Impulsive buying (H4a: β = 0.006, t = 0.949) and Purchase Behavior (H4d: β = 0.006, t = 0.938) are not significant, leading to the rejection of H4a and H4d.

This study offers important insights into the role of emotional trust and viewer engagement in shaping impulsive buying and purchase behavior in livestream commerce. The findings suggest that when emotional trust is present alone, trustworthiness and attractiveness are insufficient for deliberate purchase decisions, becoming relevant only when viewers are actively engaged in the livestream. In contrast, for impulsive buying, trustworthiness and attractiveness foster emotional trust regardless of engagement, making viewers more prone to spontaneous purchases. These patterns align with studies highlighting that affective cues such as warmth, sincerity, and likability effectively encourage immediate, emotion-driven responses rather than deliberate decisions (Ali et al., 2025; Wang et al., 2022; Ji et al., 2023; Zhou and Baskaran, 2025). Interestingly, expertise does not appear associated with either impulsive or deliberate purchase behavior, even when emotional trust and engagement coexist. This suggests that emotional trust in livestreams is rooted in perceived benevolence, warmth and interpersonal closeness rather than competence. While expertise, product knowledge and presentation skills provide rational cues, they lack emotional resonance, making them less salient in motivating action, contrasting with traditional e-commerce contexts where competence-based trust plays a central role (McAllister, 1995; Zhou and Baskaran, 2025).

Specifically, trustworthiness and attractiveness are sufficient to encourage impulsive buying via emotional trust, even without active engagement, whereas their influence on deliberate purchases emerges only with sustained viewer involvement. Features such as appealing appearance, pleasant voice, natural expressions, friendly gestures and warm communication help viewers feel emotionally connected to streamers, promoting longer viewing, more interaction and openness to spontaneous purchases. Deliberate decisions, however, require deeper involvement, including attention to product details, interaction with the streamer and cognitive processing. This distinction supports prior research indicating that trustworthiness and attractiveness more effectively drive impulsive responses, while deliberate decisions rely on sustained engagement and information processing (Zhou and Baskaran, 2025; Ali et al., 2025; Li et al., 2024a; Xu et al., 2020).

By contrast, expertise mainly provides informational value and signals competence, but it does not appear to foster emotional closeness or psychological safety. Even when viewers are actively engaged, expertise alone may not be sufficient to prompt purchasing behavior, as it lacks the affective and relational qualities that characterize effective persuasion in livestream commerce. While prior studies have documented positive effects of expertise on purchase-related outcomes in more information-driven or utilitarian online contexts (e.g. Zhu et al., 2020; Koay et al., 2021), the present findings suggest that such effects may not readily translate into livestream environments where emotional resonance and social connection play a more central role. This observation underscores the limited role of purely cognitive cues in highly social and interactive commerce settings.

In addition, entertainment and interactivity play an important role in enhancing viewer engagement, which is closely associated with both impulsive buying and purchase behavior. Entertainment elements, such as humorous performances and enjoyable content, help sustain attention and create positive emotional states, while interaction fosters social closeness through real-time responses and reciprocal communication. These interpretations are consistent with prior studies, suggesting that such features make viewers feel valued and psychologically closer to the streamer, thereby increasing their willingness to remain engaged and engage in purchasing behaviors within the livestream commerce context (Zhang and Xu, 2024; Dong and Tarofder, 2024; Hossain et al., 2023).

The results further suggest that emotional trust is closely associated with both impulsive buying and purchase behavior when accompanied by viewer engagement, a pattern that aligns with prior studies in livestream commerce (Zhou and Baskaran, 2025; Shih et al., 2024). Emotional trust fosters psychological safety and emotional connection, encouraging viewers to engage more actively in the livestream. Increased participation exposes viewers to product-related cues and social influence, making purchases more likely. This aligns with findings that emotional trust enhances attentiveness, interaction and sustained involvement, facilitating consumer responses. The mechanism highlights that affective and participatory processes are interconnected, reflecting the experiential and relational nature of livestream commerce.

Overall, this study contributes to livestream commerce research in three important ways. First, it distinguishes between two distinct psychological pathways: an emotional route, where trustworthiness and attractiveness foster emotional trust that translates into impulsive buying, and a rational route, where active engagement is necessary for purchase behavior to emerge. Second, the findings highlight the dual role of entertainment and interaction, which influence consumer behavior through engagement, underscoring their unique ability to simultaneously stimulate positive emotions and strengthen social bonds. Third, the results demonstrate the distinctive nature of livestream commerce compared to traditional e-commerce, where emotional resonance, real-time interaction and communal experiences often outweigh purely cognitive evaluations, explaining why expertise plays a limited role in shaping both impulsive buying and purchase behavior in this context. Research summary summarizes the main findings and implications of this study.

This study provides several important theoretical contributions to the livestream commerce literature. First, this study advances SCT by disentangling the differential roles of expertise, trustworthiness and attractiveness in livestream commerce that is highly interactive, spontaneous, and emotion-driven context. Prior studies have often conceptualized source credibility as a composite construct or focused on general streamer or content-related characteristics (e.g. Zhu et al., 2020; Koay et al., 2021; Zheng et al., 2023). By separating these dimensions and examining their influence through emotional trust, and grounded in SCT theory (Hovland and Weiss, 1951; Ohanian, 1990), this study shows that trustworthiness and attractiveness, rather than expertise, are key factors that facilitate impulsive buying behavior. Notably, emotional trust alone appears insufficient to translate streamer attributes into purchase behavior, pointing to a fundamental distinction between impulsive buying and more deliberative purchase behavior. This finding nuances SCT by emphasizing that competence is less relevant in environments where relational and affective cues dominate, and that trust is built more on perceived sincerity and relational appeal. Thus, the study contextualizes SCT to emotionally immersive commerce platforms and highlights trustworthiness as the most central attribute for triggering impulsive buying through emotional trust, followed by attractiveness, while expertise plays a limited role in both buying outcomes.

Second, this study contributes to UGT by introducing emotional trust as a psychological gratification that arises from the viewer–streamer relationship. While previous UGT studies have focused on hedonic, informational or social gratification (Wongkitrungrueng and Assarut, 2020; Zheng et al., 2023), our findings, grounded in foundational UGT literature (Katz et al., 1973; Hilvert-Bruce et al., 2018), indicate that emotional trust itself is a sought-after outcome that motivates viewer engagement and buying behaviors. Emotional trust acts as a non-functional yet affective gratification, deepening our understanding of why viewers remain engaged in livestream even without strong informational utility. This expands the UGT framework to include emotion-based gratifications as key drivers of behavior in real-time interaction contexts.

Third, by integrating emotional trust and viewer engagement as serial mediators, this study advances the ACP model by demonstrating a dual-pathway mechanism from streamer attributes to consumer behavior. The affective (emotional trust) and cognitive-affective (viewer engagement) processes are not independent but serially connected, reflecting a layered psychological response in livestream environments. While prior livestream commerce studies have examined affective and cognitive processes, these mechanisms have not been fully explored in the context of how specific streamer attributes translate into different forms of buying behavior. Extending this literature and drawing on ACP theory (Edell and Burke, 1987; Pham, 1998; Shiv and Fedorikhin, 1999; Lim et al., 2020), our findings clarify how an ordered affective-to-cognitive process operates in livestream commerce, with emotional trust forming the psychological foundation that supports subsequent viewer engagement, jointly shaping buying decisions. This layered mechanism suggests that affective responses precede and enable deeper cognitive-affective engagement, ultimately translating into consumption outcomes. By explicitly modeling this ordered relationship, this study refines the ACP framework in the context of livestream commerce.

Moreover, the findings reveal differentiated mediation structures across buying behaviors. Specifically, trustworthiness and attractiveness influence purchase behavior only when affective and cognitive mechanisms operate jointly, whereas impulsive buying can be triggered at the affective level alone or through the full serial pathway. In contrast, streamer expertise plays a limited role in shaping either form of buying behavior within this affective–cognitive process.

Overall, by integrating SCT, UGT and the ACP model, this study clarifies how emotional trust and viewer engagement function as core psychological mechanisms linking streamer attributes to consumer behavior, highlighting the emotionally immersive and participatory nature of livestream commerce.

This study offers practical implications for livestream commerce businesses, platforms, marketers and streamers aiming to enhance viewer engagement and strengthen buying behaviors. First, the findings reveal that streamer trustworthiness and attractiveness significantly foster emotional trust, which in turn is associated with higher levels of engagement and impulsive buying responses. To strengthen perceived trustworthiness, streamers should present detailed and verifiable product information, such as specifications, warranty terms or third-party certifications (e.g. eco-labels, safety approvals); conduct live demonstrations or unboxings to transparently showcase product quality; and share authentic personal experiences or customer testimonials during the livestream to enhance credibility. Moreover, streamers are encouraged to tailor these strategies to the type of product being promoted: for technical or high-involvement products (e.g. electronics, appliances), providing detailed demonstrations and certifications is particularly important; for lifestyle or experience-oriented products (e.g. cosmetics, food, travel services), emphasizing personal experiences and customer testimonials is likely to be more effective. To increase attractiveness, streamers should ensure high-quality visual and audio production using professional lighting, HD video and clear sound, while adopting an expressive and friendly on-camera style that includes positive facial expressions, warm tone and confident delivery. In addition, streamers should adjust their on-camera style according to the target audience: for younger or entertainment-focused viewers, more expressive gestures, humor and lively storytelling may enhance engagement, whereas for professional or older audiences, a calm, confident and clear communication style may be more effective. These adjustments are expected to optimize viewer engagement for different audience segments. For platforms and businesses, selecting streamers who demonstrate a strong combination of trustworthiness, and attractiveness is essential. In line with previous research on virtual environments (e.g. Ramlatchan and Watson, 2020), platforms should offer features that enhance visual and vocal presentation, such as HD video and voice clarity. Additional platform tools can be designed to suit specific products and audiences, such as visual product info cards or certification badges for technical products, and gamified engagement features or interactive polls for lifestyle products. Structured Q&A functions and segment-specific streamer training programs can further reinforce these qualities. Marketers should align brand collaborations with streamers who reflect the brand’s image in terms of reliability, sincerity and appeal.

Second, the study highlights the importance of entertainment and interactivity in building engagement. Streamers are encouraged to incorporate humor, storytelling or interactive games to make sessions more enjoyable, while also addressing viewers by name, responding quickly to comments and acknowledging participation to reduce psychological distance. Streamers should also consider differences between returning and new viewers, using interactive strategies to sustain engagement among highly involved or loyal audiences. These strategies are expected to maintain or increase engagement and encourage repeated impulsive buying. Platforms can facilitate these efforts with low-latency chat, emoji reactions, real-time polls and gamified features like loyalty points, ranking systems or achievement badges, which help encourage continued participation and repeated buying activities. These findings align with prior research (Shin et al., 2024; Liu and Tanaka, 2020), emphasizing the value of real-time interaction and tailored gamification in enhancing user experience and encouraging purchase behavior.

This study reveals that emotional trust is associated with buying behaviors through viewer engagement. Meanwhile, streamer expertise shows no significant indirect effects on buying behaviors. Trustworthiness and attractiveness are associated with impulsive buying through emotional trust alone and through a serial mediation pathway involving emotional trust and viewer engagement, whereas their effects on purchase behavior emerge only via serial mediation. Entertainment and interaction are also associated with buying behaviors indirectly through viewer engagement.

This study has limitations that should be addressed in future research. First, this study focuses on five streamer attributes, including expertise, trustworthiness, attractiveness, entertainment and interaction. Future research could explore additional factors such as humor (Hou et al., 2020), empathy (Li and He, 2025), celebrity–product congruence (Rungruangjit, 2022), and influencer–follower congruence (Maghraoui and Khrouf, 2025), among others, to deepen the understanding of consumer behavior in livestream commerce. Beyond streamer attributes, future studies could examine platform, product, environmental and social factors, as these may also impact viewers’ emotional and behavioral responses. Second, future research could explore other aspects of trust, such as cognitive trust, and examine various dimensions of viewer engagement, such as involvement and commitment, to understand their impact on buying behaviors in livestream settings. Third, the cross-sectional design limits conclusions about causal relationships. Longitudinal studies could offer deeper insights into how streamer characteristics influence consumer engagement and buying behaviors over time. Finally, the use of convenience sampling raised several statistical concerns. In particular, the sample was heavily skewed toward younger (Gen Z) and low-income groups, which constrains the generalizability of the results to other demographic segments and economic contexts. Future research should employ more rigorous sampling strategies, such as probability-based or stratified sampling, to improve the representativeness of the findings.

Ethical approval was waived as the research did not involve vulnerable populations and posed minimal risk to participants. All participants provided informed consent prior to completing the questionnaire.

Research summary

Main findings

• Streamer trustworthiness and attractiveness increase emotional trust, thereby shaping impulsive buying via, but not purchase behavior.

• Streamer entertainment, interaction, and emotional trust enhance both impulsive buying and purchase behavior through viewer engagement.

• Serial mediation through emotional trust and viewer engagement is observed for trustworthiness and attractiveness, but not for expertise.

Theoretical implications

• Trustworthiness and attractiveness, rather than expertise, influence impulsive buying via emotional trust, refining SCT.

• Emotional trust, as an affective gratification that motivates engagement and buying, extends UGT and refines Affective–Cognitive Processing (ACP), highlighting an affective-to-cognitive pathway in livestream commerce.

• Integrating SCT, UGT and ACP explains the serial mediation through which streamer attributes influence emotional trust, thereby enhancing viewer engagement and ultimately shaping buying behavior.

Managerial implications

• Streamers should enhance trustworthiness and attractiveness via clear communication, demonstrations and expressive presentation.

• Entertainment and interactivity increase engagement, fostering buying.

• Platforms should provide high-quality video, interactive features and gamification to strengthen engagement.