Digital innovation has remarkably altered modern societal living patterns, particularly through increased online shopping activities. According to the Ministry of Communication and Digital (Komdigi) statistics, Indonesian internet users will reach 221 million or 79.5% of the population by early 2025. This success has established Indonesia as one of the world's most digitally active nations (Indonesia.go.id, 2025). Furthermore, reports from the Indonesian Internet Service Providers Association (APJII) indicate that Generation Z constitutes the largest proportion of internet users in Indonesia, followed by Millennials and Generation X, while older age groups account for a relatively small share. The high penetration of the Internet, especially among younger age groups, shows that digital activities have become an integral part of the lifestyle of this generation. Such conditions further stimulate shifts in consumer consumption patterns, specifically regarding the growth of online purchasing and engagement on various digital platforms.
Cosmetic products are among the most frequently purchased products after consumers are exposed to promotions or endorsements by influencers on social media. It has been reported that around 68% of consumers in Indonesia frequently acquire products after observing them through diverse social platforms. In addition, the national cosmetics industry has shown consistent growth over the past five years. According to the annual report of the Food and Drug Supervisory Agency (BPOM) as of July 2024, there are 1,178 registered cosmetic industries in Indonesia, with approximately 90% of them being micro, small, and medium-sized enterprises (MSMEs) (Nurhayati-Wolff, 2024; Badan Pengawas Obat dan Makanan, 2025).
One of the platforms contributing to the development of this industry is Sociolla, a beauty e-commerce platform that focuses on cosmetics and skincare products (Sociolla, 2025). Sociolla primarily targets female consumers, predominantly from younger generations, with high levels of digital engagement (Sociolla, 2025). The dominance of skincare products in sales indicates strong emotional involvement of consumers in self-care. Furthermore, the presence of user and expert reviews positions the platform as an influential source of information in the consumer decision-making process (Sociolla, 2025). These conditions suggest that Sociolla provides a relevant context for examining consumer behavior, particularly in relation to impulsive buying.

Table 1. Most Visited Beauty and Cosmetics Website in Indonesia (June 2025)

Based on Table 1, the number of visits to the Sociolla website (214,989 visits) is still lower than that of Sephora Indonesia (229,289 visits). Despite occupying the fourth position on the list of the most visited beauty sites, this achievement shows that Sociolla can still compete and maintain consumer appeal amid competition with global brands such as Sephora. This indicates that Sociolla’s differentiation strategy and omnichannel approach can maintain the existence and interest of consumers in the Indonesian digital beauty market.
Sociolla targets female consumers aged 18–30 years, predominantly from younger generations with high levels of digital engagement (Maulana, 2019). In addition, skincare products contribute the largest share of sales, indicating a consumer tendency toward self-care. This characteristic is further supported by research conducted by Jenifer (2024), which found that Sociolla consumers are largely dominated by younger age groups with a strong interest in skincare and makeup, while also considering the aspect of an enjoyable shopping experience.
In this context, consumer behavior is driven not only by functional needs but also by hedonic motivation. Hedonic shopping motivation reflects consumers’ tendency to shop to obtain pleasure and emotional experiences (Arnold & Reynolds, 2003). Furthermore, electronic word-of-mouth (eWOM) through online reviews and recommendations serves as a source of information that can influence consumer perceptions and emotions during the purchasing process (Pambagyo & Karnawati, 2020). The combination of hedonic motivation and exposure to eWOM has the potential to generate positive emotions, which may ultimately trigger impulse buying behavior, defined as spontaneous purchases made without prior planning (Kotler et al., 2017; Maharanie et al., 2020).
However, despite the extensive examination of factors, including hedonic shopping motivation, eWOM, and positive emotions, in relation to impulse buying, there is still a lack of clarity regarding how these variables interact to influence impulsive purchasing behavior. Several studies have shown that hedonic shopping motivation and eWOM significantly affect impulse buying, both directly and indirectly, through the mediation of positive emotions (Wijoyo & Santoso, 2022; Wulanda et al., 2025). However, these findings remain inconsistent, as other studies have reported that eWOM does not significantly affect impulse buying (Pambagyo & Karnawati, 2020).
This is reinforced by research conducted by Sociolla, which found that consumers are generally dominated by the young age group, namely in the range of 24-25 years, with a strong interest in skincare and makeup products. Sociolla can attract consumers who prioritize product authenticity, completeness of variety, attractive promotions, and a pleasant shopping experience through both offline and online channels (Jenifer, 2024).
These inconsistencies highlight the need for further investigations. Moreover, few studies have comprehensively integrated eWOM into models examining the relationship between hedonic shopping motivation, positive emotions, and impulse buying. Therefore, this study aims to examine these relationships within the context of Sociolla consumers in Jakarta, Indonesia.

2.1. Consumer Behavior
Individuals or groups shape and reveal their identities by choosing, acquiring, utilizing, or discarding products, services, ideas, and experiences to satisfy their wants and needs. This process—spanning selection, purchase, consumption, and disposal by people, organizations, or collectives—defines consumer behavior (Solomon & Russel, 2024; Kotler, et al., 2022).

2.2. Hedonic Shopping Motivation
Hedonic shopping motivation describes the urge to shop because of wanting to feel pleasure, emotional satisfaction, and a pleasant experience, not just to fulfill a need that is logical. Hedonistic shopping motivation is understood as an individual's drive to obtain personal pleasure that can be realized through the activity of spending time in a shopping center or mall, enjoying the atmosphere or atmosphere it offers, even if it is not making a purchase and simply looking around. Hedonistic shopping motivation is often attributed to consumer behavior oriented towards fulfilling emotional needs, pleasure, and sensory experiences during the shopping process, emphasizing the enjoyment of the shopping activity itself or is often termed as "retail therapy” (Imaliya, 2024; Asj’ari et al., 2021; Solomon & Russel, 2024).

2.3. Electronic Word of Mouth (eWOM)
Online endorsements or critiques, positive or negative, from potential, actual, or past consumers about products represent eWOM, a vital type of Internet-based marketing exchange. eWOM has its roots in the concept of WOM, which was first introduced in the mid-20th century by eWOM originates from Brooks' mid-twentieth-century Word of Mouth theory, defining the phenomenon as interpersonal verbal communication occurring through direct individual talk, in which the recipient views the communicator as having no commercial interest, and the conversation focuses on a brand, product, or service (Slamet & Albab, 2023; Liu et al., 2024)

2.4. Positive Emotion
Positive emotions, such as pleasure, enjoyment, and satisfaction, can arise from consumers' interactions with products, services, or shopping environments. The term 'positive emotions' refers to complex emotional responses that consist of various components, which stem from personal interpretations of existing environmental prospects and incentives, and significantly guide the overall consumer decision-making process and behavior. Positive emotions encompass a spectrum of uplifting feelings, such as joy, gratitude, enthusiasm, happiness and pride (Shiota et al., 2021; Wang et al., 2022).

2.5. Impulse Buying
Impulse buying is the act of buying suddenly without going through a planning process or prior consideration, which generally appears in response to an emotional stimulus or situational condition that is directly encountered. Impulse buying is a form of consumption behavior that appears suddenly and urgently, and has a complex hedonistic dimension where the speed of impulsive decision-making tends to hinder the careful and planned consideration process of various alternatives and long-term consequences. Impulse acquisitions lack prior planning; however, not every unplanned purchase qualifies as impulsive, as some occur simply from forgetting specific needs. Conversely, true impulsive buying is driven by intense emotions and irrational tendencies, distinguishing it from the mere accidental or utilitarian nature of forgotten items (Imaliya, 2024; Masouleh et al., 2012; Santini et al., 2018).

2.6. Research Framework

For more details, refer to Figure 1.

Figure 1. Frame of Mind

2.7. Hypothesis Development

2.7.1. The Effect of Hedonic Shopping Motivation on Positive Emotions
Hedonic shopping motivation is closely associated with individuals’ emotional and imaginative aspects. Based on the study conducted by Mariyana et al. (2023), hedonic shopping motivation has a positive and significant effect on positive emotions. These findings indicate that the higher the level of hedonic shopping motivation experienced by consumers, the greater their likelihood of experiencing positive emotions during shopping. This result is consistent with the studies conducted by Cahyani and Marcelino (2023) and Rabbani et al. (2023), which also reported a positive and significant relationship between hedonic shopping motivation and positive emotions. This indicates that an increase in hedonic shopping motivation directly enhances consumers’ feelings of pleasure, joy, and satisfaction during shopping experiences. Therefore, it can be inferred that hedonic shopping motivation plays a crucial role in shaping consumers’ emotional experiences, particularly within online shopping contexts.
H1: Hedonic Shopping Motivation has a significant positive effect on positive emotions.

2.7.2. The Effect of eWOM on Positive Emotions
Electronic word-of-mouth (eWOM) is an external stimulus that significantly influences consumers’ internal responses, particularly their emotional responses (Wulanda et al., 2025). The dissemination of information through eWOM contributes to the development of positive emotional responses, as consumers tend to form emotional attachments to products and experience feelings such as happiness, satisfaction, and enthusiasm (Wulanda et al., 2025). Empirical evidence supports this relationship in the literature. Putri et al. (2025) found that eWOM positively and significantly affects positive emotions. Miremadi et al. (2021) reported that eWOM communication significantly influences consumers’ emotional responses. Therefore, it can be concluded that eWOM not only functions as a source of information but also serves as a key factor in shaping consumers’ emotional states during the purchasing decision-making process.
H2: eWOM has a significant positive effect on positive emotions.

2.7.3. The Effect of Hedonic Shopping Motivation on Impulse Buying
Hedonic Shopping Motivation reflects consumers’ drive to engage in shopping activities to obtain pleasure, entertainment, and emotional experiences (Arnold & Reynolds, 2003). The findings of Ramdani et al. (2024) demonstrate a positive and significant effect of Hedonic Shopping Motivation on Impulse Buying. This finding is consistent with the studies conducted by Pambagyo and Karnawati (2020) and Widagdo and Roz (2020), which also indicate that Hedonic Shopping Motivation has a positive and significant influence on Impulse Buying. This relationship is based on the notion that hedonic shopping tends to trigger consumers to make impulsive purchases (Widagdo & Roz, 2020). Therefore, it can be concluded that the higher the level of Hedonic Shopping Motivation, the greater the likelihood that consumers will engage in impulse buying behavior.
H3: Hedonic Shopping Motivation has a significant positive effect on Impulse Buying

2.7.4. The Influence of eWOM on Impulse Buying Behavior.
eWOM, which encompasses online reviews, ratings, and consumer recommendations, facilitates consumers’ access to and assessment of product-related information generated by other users (Hennig-Thurau et al., 2004). This information serves as a persuasive stimulus that influences consumers’ evaluations, attitudes, and purchase-related judgments (Cheung & Thadani, 2012). Furthermore, exposure to positive eWOM is likely to enhance consumers’ desire and confidence in a product, thereby increasing the likelihood of spontaneous and unplanned purchasing behavior. The results of the study conducted by Shiu et al. (2023) indicate that positive eWOM can lead to impulse purchases directly, meaning that exposure to positive information from reviews or recommendations of other consumers can elicit an emotional response, which then strengthens the consumer’s tendency to make spontaneous purchases. Similar studies have demonstrated that eWOM significantly impacts impulse purchases, as constant digital exposure to product discussions heightens consumer desire (Kurniawan et al., 2023). Consequently, individuals frequently acquire items they previously had no intention of buying, driven by the persuasive power of online social interactions (Shiu et al., 2023; Kurniawan et al., 2023).
H4: eWOM has a positive and significant effect on Impulse Buying

2.7.5. The Influence of Positive Emotion on Impulse Buying
Positive emotions, including feelings of pleasure, excitement, and satisfaction, have been shown to diminish cognitive control while enhancing consumers’ propensity for spontaneous behaviors. In this context, affective states serve as a critical driver that shifts decision-making processes from rational evaluation to emotion-driven responses, thereby increasing the likelihood of unplanned purchasing behavior (Beatty & Ferrel, 1998; Rook & Fisher, 1995). Research results demonstrate that positive emotions significantly and positively influence impulse purchases. These findings align with prior empirical evidence, confirming that heightened emotional states effectively stimulate spontaneous consumer acquisitions and unplanned buying decisions (Listriyani & Wahyono, 2019; Wisesa & Ardani, 2022).
H5: Positive Emotion has a positive and significant effect on Impulse Buying

2.7.6. The Role of Positive Emotion Mediation on Hedonic Shopping Motivation and Impulse Buying
Hedonic Shopping Motivation reflects consumers’ desire to seek pleasure, enjoyment, and experiential value during the shopping process (Arnold & Reynolds, 2003). When consumers are driven by hedonic motives, they are more likely to experience heightened positive emotions, which reduce cognitive control and increase their tendency to act spontaneously. These affective responses facilitate impulse buying by shifting decision-making from rational evaluation to emotion-driven behavior (Beatty & Ferrel, 1998; Rook & Fisher, 1995). Research results demonstrate that Positive Emotion significantly mediates the relationship between Hedonic Shopping Motivation and Impulse Buying, aligning with evidence from Dewi and Adi (2023). These findings correspond with those of Renaldi and Nurlinda (2023), who established that hedonistic triggers influence impulsive actions through emotional states. Furthermore, the study identified complementary partial mediation, suggesting that Positive Emotion partially explains this interaction, while other factors also contribute (Kholis et al., 2023). Therefore, Positive Emotion can be regarded as a key mediating variable that explains how hedonic shopping motivation leads to impulse buying behavior.
H6: Hedonic Shopping Motivation has a positive and significant effect on Impulse Buying through the mediation of positive emotions.

2.7.7. The Role of Positive Emotion Mediation on eWOM and Impulse Buying
The mediating role of Positive Emotion in the relationship between Electronic Word-of-Mouth (eWOM) and impulse buying behavior reflects the underlying affective mechanism through which external information influences consumer actions. As a source of social and informational influence, eWOM can shape consumers’ perceptions and behavioral responses (Cheung & Thadani, 2012). In addition, eWOM evokes emotional reactions that influence consumer decision-making processes (Verhagen & van Dolen, 2011). In this context, Positive Emotion functions as an internal psychological pathway through which eWOM translates into impulse-buying behavior, as affective states play a crucial role in driving spontaneous purchasing decisions (Beatty & Ferrel, 1998; Rook & Fisher, 1995).
Research conducted by Panasea et al. (2025) indicate that the influence of E-WOM on Impulse Buying is fully mediated by positive emotions. These findings indicate that eWOM cannot directly encourage impulse buying but can do so indirectly through a positive emotional response in consumers. Moreover, the results demonstrate that Positive Emotion fails to provide significant mediation between eWOM and consumers' Impulse Buying. While eWOM stimulates favorable consumer emotions, this emotional impact lacks sufficient statistical strength to drive spontaneous purchasing decisions through specific internal psychological pathways (Wulanda et al., 2025). These inconsistent findings highlight that the mediating role of Positive Emotion remains context-dependent, thereby justifying further investigation into its role within the relationship between eWOM and impulse buying.
H7: eWOM has a positive and significant effect on Impulse Buying through Positive Emotion mediation.

This quantitative study utilized an explanatory research design to analyze the causal relationships among variables using survey methods. Causal links represent the effects of independent factors on the dependent variables. Specifically, this study examines how hedonic shopping motivation and eWOM influence impulse buying, while exploring positive emotions as a mediator among Sociolla customers located throughout the Greater Jakarta region (Sugiyono, 2013). The target population included Sociolla consumers across Greater Jakarta who purchased products through digital platforms or physical stores. This study employed non-probability sampling, specifically purposive sampling, in which respondents were selected based on predetermined criteria relevant to the research objectives (Sugiyono, 2013).
The respondent criteria used in this study were Sociolla consumers who are domiciled in the Greater Jakarta area and have made at least one transaction through the Sociolla digital platform (website or application) or shopped offline at Sociolla in the last six months. Following Hair et al. (2019), the optimal sample size ranges from five to ten times the total number of indicators. Since this study utilizes 17 indicators, the minimum required sample size is 125 respondents (Hair et al., 2019). A total of 125 valid responses were collected and used for further analysis.
This study incorporates two primary independent variables suspected of influencing consumer impulse buying behavior: Hedonic Shopping Motivation (X1) and electronic Word of Mouth (eWOM) (X2). Additionally, this study utilized Positive Emotion (Z) as the mediating variable within the framework. Finally, the dependent variable investigated in this analysis is represented by the Impulse Buying (Y) construct.
Primary data were collected through a structured questionnaire, utilizing measurement instruments developed from specific research indicators. All items were measured using a five-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). The questionnaire was distributed online via Google Forms over the period of [December 2025–February,2026], and disseminated through social media platforms, including WhatsApp, Instagram, and TikTok, to reach the target respondents.
These indicators represent hedonic shopping motivation, eWOM, positive emotions as a mediator, and impulse buying as the dependent variable. Respondent feedback provided quantitative data necessary for statistical analysis. The data were analyzed using Structural Equation Modeling (SEM) with the Partial Least Squares (PLS) approach, processed using SmartPLS software. The analysis included tests of convergent validity and discriminant validity, reliability testing using Cronbach’s Alpha and Composite Reliability, and hypothesis testing through bootstrapping procedures. All respondents participated voluntarily, and their responses were kept anonymous to ensure confidentiality and adhere to ethical research standards.

3.1. Data Analysis
Primary data were collected through structured questionnaires using measurement instruments derived from each of the research variables. These include hedonic shopping motivation, eWOM, mediating positive emotions, and impulsive buying. The variables were categorized as exogenous or endogenous, following the non-circular structural requirements of PLS-SEM. The assessment of the outer model, also known as the measurement model, examines the reliability and validity of the indicators that represent the underlying latent variables. Essential indicators employed in this analysis included internal consistency reliability, convergent validity, and discriminant validity, confirming that the measurement tools precisely captured the conceptual framework (Hair et al, 2017).
The structural model was tested to determine the strength and significance of the relationships between the constructs. First, the R² value measures the variance of endogenous constructs explained by exogenous variables within the structural model. Second, the effect size (f²) evaluates how each exogenous construct specifically contributes to R² fluctuations in the endogenous variable, reflecting its structural model impact. F-square values of 0.02, 0.15, and 0.35 categorize these influences as weak, moderate, or strong, respectively (Hair et al., 2017). Third, Q² evaluates the model’s predictive relevance, where a value exceeding zero indicates sufficient predictive capacity, confirming that the structural model can accurately forecast endogenous constructs (Hair et al, 2017). Fourth, effect size (q²) assesses the unique contribution of each exogenous construct toward the predictive relevance and overall forecasting power of the structural model’s various endogenous latent variables (Hair et al., 2017).
The construct of positive emotion as a mediating variable was tested using indirect effect analysis with a bootstrapping method. The significance of mediation was analyzed based on the confidence interval of the indirect effects. If the interval does not include zero, the mediation effect is considered to be significant (Hair et al., 2014). This study uses a significance level (α) of 5% (α = 0.05) in accordance with the general standard in quantitative research in marketing science. To determine the validity of each hypothesis, the researcher evaluates specific conditions that dictate whether to accept or reject the proposed relationships between the variables: (1) if the p-value ≤ 0.05 and the T-statistic ≥ 1.967, then H0 is rejected, Ha is accepted; and (2) if the p-value > 0.05 or the T-statistic < 1.967, then H0 is accepted, Ha is rejected (Hair et al., 2019).

4.1. Respondent Characteristics
This research was conducted in the Greater Jakarta area with respondents who had shopped at Sociolla for the past six months. Data were collected by administering online and paper-based surveys to participants who fulfilled the study's eligibility requirements. "The researcher successfully gathered data from 125 participants. The characteristics of the respondents analyzed included domicile, knowledge of Sociolla, shopping intensity, gender, and age.

Table 2. Respondent Characteristics

Based on Table 2, most participants hailed from Jakarta (54.4%), followed by Tangerang (20 %) and Depok (9.6 %). Most respondents were familiar with the Sociolla platform (90.4%), while 4.8% said they did not know it. Regarding the frequency of shopping, the majority of respondents shopped 2–3 times in a period (47.2%), followed by 1 time (16.8%), 4–5 times (12.0%), and more than 5 times (6.4%). The survey results show a gender imbalance, with women comprising 64.8% of the respondents and men 17.6 %. The age groups were led by 18–23-year-olds (36.0%) and 24–29-year-olds (29.6%).
The distribution of respondents, which is dominated by consumers from urban areas such as Jakarta, indicates high exposure to digital platforms and online shopping environments, potentially increasing susceptibility to impulse buying behavior due to frequent exposure to marketing stimuli. The high level of familiarity with Sociolla suggests that respondents have prior experience with the platform, which may reduce perceived risk and enhance trust, thereby strengthening the influence of eWOM on spontaneous purchasing decisions.
The predominance of consumers who shop 2–3 times reflects a moderate purchasing frequency, which may indicate repeated exposure to shopping activities and a reduced level of cognitive deliberation, thus increasing the likelihood of impulse buying.
The dominance of female and younger respondents (18–29 years old) further reinforces the relevance of the sample, as these groups are generally more responsive to hedonic motivations, emotional stimuli and social influence, particularly in digital environments. Overall, these characteristics indicate that the sample aligns with consumer segments that are more susceptible to hedonic motivation, eWOM influence, and positive emotional responses, thereby supporting the relevance of this study in explaining impulse-buying behavior.

4.2. Measurement Model (Outer Model)
The measurement model evaluation utilized Partial Least Squares Structural Equation Modeling. Convergent validity, discriminant validity, and construct reliability were assessed before testing the structural model for significance.

4.2.1. Validity and Reliability

4.2.1.1. Convergent Validity
Higher loading factor values indicate that the indicator has a stronger contribution to reflecting the latent construct being measured. The specific loading factor values for every research indicator are detailed comprehensively in Table 3, presented in the section below:

Table 3. Convergent Validity Test

Source: SmartPLS Output Results (v.4.2.9)

For more details, refer to Figure 2.

Figure 2. Structural Models

Source: SmartPLS Output Results (v.4.2.9)

As shown in Table 3, all indicators exhibit loading factor values above 0.70. Consequently, every indicator satisfied the requirements for convergent validity and remained appropriate for representing the latent constructs analyzed throughout this structural equation model. These findings suggest that each indicator is strongly associated with its corresponding construct, indicating that the measurement items effectively represent the intended variables of this study. The consistently high loading values also imply that the constructs are clearly defined and measured consistently across indicators.
Moreover, the strength of these loadings reflects that the measurement model meets the criteria for convergent validity, as the indicators within each construct share a substantial amount of variance. This condition enhanced the overall reliability of the model and supported its appropriateness for subsequent structural model evaluation.

4.2.1.2. Discriminatory Validity
A construct achieves sufficient discriminant validity when the square root of the AVE exceeds its correlation with any other latent construct within the structural model.

Table 4. Fornell-Larcker Criterion Value

Source: SmartPLS Output Results (v.4.2.9)

Table 5. Cross Loading Value

Source: SmartPLS Output Results (v.4.2.9)

The findings reveal that every measurement item exhibits cross-loading scores above 0.70, with their highest values occurring specifically within their respective constructs. Consequently, these findings confirm that each indicator accurately represents its intended latent construct, successfully satisfying all established discriminant validity criteria. Furthermore, referring to the Fornell–Larcker criterion in Table 4, the square root of the AVE for each construct exceeds its correlation with the other constructs. This indicates that each construct possesses adequate distinctiveness and represents a unique aspect that is not captured by other variables in the model. The cross-loading results in Table 5 further reinforce this finding, as each indicator shows the highest loading on its respective construct compared to the others. This pattern suggests that the indicators are properly aligned with their intended constructs and that there is minimal overlap between the variables. Taken together, these findings demonstrate that the measurement model meets the requirements for discriminant validity, confirming that each construct is sufficiently distinct and appropriate for subsequent structural analyses.

4.2.1.3. Reliability
In PLS-SEM, reliability is assessed using Cronbach's alpha and composite reliability. Constructs are considered reliable when the composite reliability exceeds 0.7, and Cronbach's alpha is also advised to exceed 0.7. The resulting Cronbach's alpha and Composite Reliability scores are presented in Table 6 below:

Table 6. Reliability Test

Source: SmartPLS Output Results (v.4.2.9)

The findings in Table 6 demonstrate that all constructs achieved composite reliability or Cronbach’s alpha above 0.70, with the AVE exceeding 0.50. This confirms high internal consistency, satisfying all the criteria for reliability and convergent validity. Therefore, having established a robust measurement model, this study assessed the structural relationships within the inner model. High Cronbach’s alpha values indicate that the indicators within each construct are internally consistent, suggesting that they measure the same underlying concept in a stable manner. In addition, the composite reliability values further support the reliability of the constructs, indicating that the measurement model demonstrates a satisfactory level of consistency with minimal measurement errors. Furthermore, the AVE values, which exceed the recommended threshold of 0.50, imply that each construct accounts for a substantial proportion of the variance in its indicators. This provides additional support for the adequacy of convergent validity in the measurement model. Taken together, these findings suggest that the measurement instruments employed in this study are both reliable and valid, thereby providing a sufficient basis for proceeding to structural model analysis and hypothesis testing.

4.2.1.4. VIVID

Table 7. VIF Value

Source: SmartPLS Output Results (v.4.2.9)

Referring to the Variance Inflation Factor (VIF) values displayed in Table 7, the entire construct shows a value below the threshold limit of 5. This demonstrates that no multicollinearity problems exist within the analyzed research model. Thus, the variables used do not cause distortions in the significance test or the estimation of the coefficient (weight estimation). These results suggest that each indicator contributes distinctly to its respective construct, with no substantial overlap among the indicators. The low level of multicollinearity indicates that the relationships between variables can be estimated more precisely without bias arising from redundant information among the predictors. In addition, the acceptable VIF values indicate that the model satisfies the assumptions necessary for accurate path coefficient estimation in PLS-SEM, thereby supporting the credibility of the hypothesis-testing results. Overall, the VIF findings enhance the robustness of the model, allowing the structural relationships identified in this study to be interpreted with greater confidence.

4.2.1.5. HTMT

Table 8. HTMT Test

Source: SmartPLS Output Results (v.4.2.9)

Referring to the Heterotrait–Monotrait Ratio (HTMT) results in Table 8, all values remain below the 0.90 threshold. These findings confirm that the discriminant validity requirements are satisfied, demonstrating that the research constructs possess an adequate degree of statistical separation. Consequently, the measurement model was considered valid and distinct for further analysis. These findings suggest that each construct is distinct from the others, indicating that each variable represents a different concept within the model. The HTMT values also confirm minimal overlap among constructs, reducing the likelihood of multicollinearity at the latent level. In addition, the alignment between the HTMT results and earlier discriminant validity assessments (Fornell–Larcker and cross-loadings) further supports the overall soundness of the measurement model. This consistency provides strong evidence that the constructs are well defined and appropriate for further structural model analysis.

4.3. Structural Model (Inner Model)
After evaluating the outer model, the subsequent phase involves testing the inner model was tested.

4.3.1. R Square

Table 9. R Square Test Results

Source: SmartPLS Output Results (v.4.2.9)

As shown in Table 9, the coefficient of determination (R²) for the impulse-buying construct is 0.277. This value indicates that the independent variables within this structural model account for 27.7% of the variance in impulse buying behavior. Thus, the remaining 72.3% of the variability is accounted for by other external influences omitted from this study. This result suggests that the model has moderate explanatory power, indicating that while hedonic shopping motivation and eWOM contribute to explaining impulse buying behavior, a substantial proportion is still influenced by other factors not included in the model, such as situational variables, personality traits, or promotional stimuli.
Additionally, the R² value of positive emotions was 0.271, indicating that the studied independent variables explained 27.1% of its variance. Consequently, 72.9% of the observed fluctuations were attributed to additional factors that were not incorporated into this structural model. This finding indicates that positive emotions are partially explained by the proposed stimulus variables; however, other psychological or contextual factors may also play a significant role in shaping consumers’ emotional responses.
Overall, these R² values indicate that the model demonstrates an acceptable level of explanatory capability in behavioral research, particularly in the context of consumer behavior, where multiple external and internal factors simultaneously influence decision-making processes.

4.3.2. F Square (f²)

Table 10. F Square Test Results

Source: SmartPLS Output Results (v.4.2.9)

In Table 10, the effect size (f²) test results show that all relationships between variables in the model are in the small effect category. The greatest influence was shown by the relationship between EWOM to Positive Emotion, with an f² value of 0.141, which is close to the moderate category and shows a relatively stronger contribution than other relationships. This finding indicates that among the predictors, eWOM plays a more prominent role in influencing consumers’ emotional responses than other variables, highlighting the importance of social and informational cues in shaping positive emotions.
Specifically, the impacts of eWOM (0.055), hedonic motivation (0.035, 0.117), and positive emotion (0.078) on their respective constructs were categorized as small effects, demonstrating that each predictor exerted a limited but statistically relevant influence. Although the effect sizes are relatively small, these results remain meaningful in the context of consumer behavior research, where behavioral outcomes are typically influenced by multiple factors simultaneously rather than a single dominant predictor.
These results demonstrate that although exogenous variables contribute to the model, their influence remains limited, suggesting that alternative significant factors outside this research likely explain the observed variation in positive emotions and impulse buying. Overall, the f² results complement the R² findings by indicating that while the model includes relevant predictors, the magnitude of their individual contributions is modest, reinforcing the need for future research to incorporate additional variables to enhance the model’s explanatory power.

4.3.3. Predictive Relevance (Q²)

Figure 3. Predictive Relevance

Source: SmartPLS Output Results (v.4.2.9)

Based on Figure 3, the results are summarized in Table 11 below:

Table 11. Predictive Relevance

The predictive relevance (Q²) test results revealed values of 0.260 for impulse buying and 0.327 for positive emotion. As both coefficients exceed zero (Q² > 0), the structural model demonstrates strong predictive power. Consequently, this research framework possesses adequate predictive relevance in explaining endogenous variable variations, confirming its suitability for further statistical analysis. These results suggest that the model can predict the observed data with a reasonable degree of accuracy, particularly for positive emotions, which demonstrates a higher Q² value than impulse buying. This implies that the model is more effective in explaining consumers’ emotional responses than their purchasing behavior.
In addition, the positive Q² values indicate that the model has sufficient predictive capability, as the exogenous variables meaningfully contribute to explaining endogenous constructs. Overall, the Q² results support the previous findings from the R² and f² analyses, indicating that although the explanatory and effect sizes are moderate to small, the model still maintains adequate predictive relevance and can be considered reliable for interpreting consumer behavior patterns.

4.4. Hypothesis Test
For more details, refer to Figure 4.

Figure 4. Hypothesis Testing

This study analyzes Hedonic Shopping Motivation (X1) and eWOM (X2) against Impulse Buying (Y), alongside testing Positive Emotion (Z) as a mediating variable. The standard applied for direct hypothesis testing was a p-value of < 0.05 (significance level = 5%) or a statistical T-value of > 1.960. This confirms the statistical significance of these results (see Table 12).

Table 12. Hypothesis Testing

Within the framework of PLS-SEM, the validity of every hypothesized path is evaluated via computational simulation. Specifically, this process employs a bootstrapping technique on the collected data to generate the necessary test statistics. The empirical outcomes of the resampling procedure for the current investigation are as follows.

4.4.1. The Effect of Hedonic Shopping Motivation on Positive Emotions
Hypothesis testing for the initial proposition on the link between Hedonic Shopping Motivation and Positive Emotion yields a path coefficient of 0.307, p-value of 0.001 (< 0.05), and t-value of 3.369 (> 1.960). This empirical evidence confirms that the psychological urge to seek pleasure during shopping significantly shapes a consumer's mood; therefore, Hypothesis 1, which states that "Hedonic Shopping Motivation has a Positive and Significant Influence on Positive Emotion," is formally validated. This result indicates that hedonic motivation serves as an important internal driver that enhances consumers’ emotional experiences during shopping. In line with consumer behavior theory, individuals who seek enjoyment and experiential value are more likely to develop positive emotional states, which may further influence their subsequent purchase behavior.

4.4.2. The Effect of EWOM on Positive Emotion
The evaluation process for the second hypothesis regarding the effect of digital messaging or EWOM on Positive Emotion yielded a coefficient of 0.337, a p-value of 0.000 < 0.05, and a t-statistic of 4.075 > 1.960. These findings illustrate that online recommendations from fellow consumers exert a significant and positive impact on enhancing a customer's sense of joy. Consequently, Hypothesis 2, which asserts that "EWOM has a Positive and Significant Effect on Positive Emotion," is confirmed within the framework of this study. This finding highlights the role of eWOM as a social influence mechanism that provides information and triggers emotional responses. Positive reviews and recommendations can create feelings of excitement and confidence, thereby strengthening consumers’ positive emotions.

4.4.3. The Effect of Hedonic Shopping Motivation on Impulse Buying
Referring to the assessment of the third hypothesis concerning the connection between Hedonic Shopping Motivation and Impulse Buying, a coefficient of 0.177 was obtained, with a p-value of 0.029 < 0.05 and a t-statistic of 2.181 > 1.960. These findings suggest that a shopping orientation centered on internal satisfaction significantly and positively triggers unplanned transactions. Hence, Hypothesis 3, which states that "Hedonic Shopping Motivation has a Positive and Significant Effect on Impulse Buying," is fully accepted. This suggests that consumers driven by pleasure-oriented motives are more likely to engage in spontaneous purchases because their decisions are guided more by affective impulses than by rational evaluation.

4.4.4. The Influence of EWOM on Impulse Buying
Data processing for the fourth hypothesis regarding the impact of EWOM on Impulse Buying recorded a coefficient value of 0.224, a p-value of 0.003 < 0.05, and a t-statistic of 3.005 > 1.960. These results prove that public opinion in digital media significantly and positively leads to spontaneous purchasing behavior among consumers. Therefore, Hypothesis 4, which states that "EWOM has a Positive and Significant Effect on Impulse Buying," is declared empirically proven based on the model testing. This finding implies that exposure to online reviews and recommendations can directly influence consumers’ purchase decisions, reducing the need for extensive evaluation and increasing the likelihood of impulsive action.

4.4.5. The Effect of Positive Emotions on Impulse Buying
Analysis of the fifth hypothesis regarding the influence of psychological conditions or positive emotions on impulse buying produced a coefficient value of 0.278, a p-value of 0.001 < 0.05, and a t-statistic of 3.212 > 1.960. These findings demonstrate that feelings of pleasure or happiness experienced by consumers significantly and positively stimulate shopping without prior preparation. Thus, Hypothesis 5, stating that "Positive Emotion has a Positive and Significant Effect on Impulse Buying," is accepted without hesitation. This result reinforces the role of emotional states as a key mechanism in driving impulse buying, where positive feelings can lower cognitive control and encourage immediate purchase behavior.

4.4.6. The Effect of Hedonic Shopping Motivation on Impulse Buying Mediated by Positive Emotions
The sixth hypothesis on the effect of Hedonic Shopping Motivation on Impulse Buying, mediated by Positive Emotion, produced a coefficient of 0.085, p-value of 0.031 (< 0.05), and t-value of 2.153 (> 1.960). This data proves that the desire to shop for pleasure can influence spontaneous buying actions because of the prior emergence of emotional arousal. Consequently, Hypothesis 6, which states that "Positive Emotion is Able to Mediate the Influence of Hedonic Shopping Motivation on Impulse Buying," is accepted. This indicates that positive emotions function as an underlying psychological mechanism through which hedonic motivation translates into impulse buying behavior, supporting the mediating role of affect in consumer decision-making.

4.4.7. The Influence of EWOM on Impulse Buying Mediated by Positive Emotions
Analysis of the seventh hypothesis regarding EWOM's effect of EWOM on Impulse Buying, with Positive Emotion as the mediator, yielded a coefficient of 0.094, p-value of 0.009 (< 0.05), and t-value of 2.619 (> 1.960). These outcomes suggest that online information sparks spontaneous buying by evoking positive consumer feelings. Therefore, Hypothesis 7, which states that "Positive Emotion is Able to Mediate the Influence of EWOM on Impulse Buying," is accepted in this final analysis. This finding confirms that the effect of eWOM on impulse buying is not only direct but also indirect through emotional responses, highlighting the importance of affective processes in transforming external information into actual purchasing behavior.

The results of this study indicate that Hedonic Shopping Motivation and Electronic Word of Mouth (EWOM) have a positive and significant effect on positive emotions and impulse buying. This suggests that the more intense the gratification-oriented buying drive and the more powerful the interaction with external consumer insights and ratings, the more likely consumers are to experience positive emotions that encourage impulse purchases. Furthermore, positive emotions enhance impulse-buying tendencies, making the affective dimension a vital component in understanding consumer purchase patterns. These results verify that unplanned buying is not merely driven by external cues such as EWOM but also by internal drivers and the emotional states experienced by consumers throughout the shopping journey.
From an applied perspective, this study’s results imply that Sociolla's leadership must prioritize digital communication strategies and the development of an enjoyable retail environment. Since favorable EWOM is shown to cultivate consumer emotions, the management of online reviews, testimonials, and social media engagement should be consistent to preserve brand reputation and consumer confidence. Furthermore, the impact of Hedonic Shopping Motivation on impulsive purchases highlights that shoppers are frequently motivated by retail experiences that are both pleasurable and engaging. Therefore, Sociolla can develop a platform display, promotional programs, and marketing strategies that present a more attractive and memorable shopping atmosphere. However, because impulse buying is closely related to emotional impulses, business actors still need to consider ethical responsibility so that the marketing strategies applied do not encourage excessive consumptive behavior.