Drawing upon the Stimulus-Organism-Response framework, this study introduces a research model that elucidates the sequential impact of environmental cues within TikTok Live i.e., perceiv
INTRODUCTION
Research background and statement of the problem
In recent decades, the internet has dramatically transformed business and marketing, driving the need for adaptation to digital commerce This evolution leverages the network effect to enhance value and meet consumer demands Technological advancements have reshaped the behavior patterns of individuals and businesses in online commerce, leading to a rapid global increase in online purchasing, especially in countries with advanced information and communication technologies for marketing.
In 2020, social media emerged as the most effective marketing strategy for building customer relationships, thanks to its relatively low marketing costs and the growing number of users that allow brands to reach a broader audience As of January 2023, the user base of social media continues to expand, highlighting its importance in contemporary marketing efforts.
As of 2021, the number of social media users worldwide reached 4.2 billion, reflecting a significant annual growth of 13.2% from the previous year This surge has led to the emergence of online social media marketing, enabling businesses to engage with suppliers and customers, shape perceptions, and gather valuable feedback Additionally, social commerce, a subset of e-commerce, leverages social media to enhance interaction, streamline online transactions, and improve the overall online shopping experience for consumers.
Social commerce, distinct from traditional e-commerce, focuses on promoting products directly within targeted social media platforms It emphasizes value-driven, personalized interactions, integrating social objectives like information sharing on social networks.
TikTok, a short video social media platform created by ByteDance, allows users to capture and share vibrant moments from their lives, appealing to a diverse audience from teenagers to seniors Its unique one-at-a-time content consumption model, driven by an algorithm that tailors videos to individual interests and browsing history, makes it easy for users to discover new information Although TikTok did not initially feature direct shopping, it has evolved into a social commerce platform where marketers can purchase ad space and engage in influencer marketing Influencers can include links to online stores in their bios, enabling products to go viral through their recommendations and directing viewers to merchant websites for easy shopping.
TikTok significantly impacts consumer behavior, with 67% of users finding themselves encouraged to shop impulsively, despite not initially planning to make a purchase Additionally, 74% of viewers express a desire to learn more about brands or products they encounter, and 66% utilize the platform to aid in their buying decisions (TikTok for Business, 2023) This highlights the platform's powerful role in shaping purchasing choices.
The #liklokmademebuyit trend gained significant traction on TikTok, with 7.4 billion views by 2021, highlighting the platform's effectiveness in driving impulse purchases through its hyper-specific algorithm that tailors content to users' interests The introduction of TikTok Shop in August 2021 streamlined the buying process by integrating in-app purchases, eliminating the need for users to navigate away from the platform This shift has led to a remarkable increase in Net Merchandise Value (NMV), which reached approximately VND 93,000 billion in the first half of 2023, marking a 46% year-over-year growth TikTok Shop's revenue has rapidly climbed, surpassing Lazada's in the second quarter of 2023, establishing it as the second-largest e-commerce platform in Vietnam after Shopee Additionally, TikTok's live broadcast feature enhances product promotion for vendors, further boosting sales potential.
TikTok Live facilitates real-time interactions between buyers and sellers, making it a popular platform for brand owners and sellers to promote their products Collaborating with key opinion leaders (KOLs) enhances product visibility, while convenient shopping experiences, enticing promotions, and engaging live-streaming content foster direct engagement between brands and customers Exclusive discounts offered during live streams further incentivize purchases This dynamic has led to a notable increase in product sales, as the TikTok live streaming feature allows customers to ask questions and view products in detail, boosting their confidence in product quality and enabling quicker purchasing decisions Consequently, this ease of transaction can lead to impulsive buying behavior.
Impulsive buying refers to the tendency of consumers to make quick and spontaneous purchases without much deliberation, often driven by factors such as appealing products, discounts, or limited availability (Rook & Fisher, 1995; Rohman, 2009; Edy & Haryanti, 2018) The online shopping environment significantly influences this behavior, as scarcity tactics can prompt rapid decision-making among consumers, enhancing the perceived value of products and boosting consumer satisfaction (Guo et al., 2017; Wang, 2021; Akram et al., 2018) Live streaming services further facilitate impulsive buying by allowing real-time interaction between shoppers and sellers, fostering engagement and increasing purchase intentions Additionally, perceived enjoyment plays a crucial role in enhancing impulse buying behavior This conceptual model categorizes marketing stimuli into perceived product scarcity, streamer trustworthiness, and social presence, while focusing on the effects of perceived enjoyment and involvement on impulse buying behavior.
Research on the impact of TikTok Live on impulse buying behavior remains limited, with few studies conducted in this area Notably, Li et al (2021) found that perceived media richness significantly influenced purchase intentions among Chinese consumers in Changsha While Fadillah and Kusumawati's 2021 study also examined cosmetics in Indonesia, the broader implications of TikTok's growing popularity warrant further exploration As millions engage in live streaming daily, understanding how TikTok Live affects consumer buying decisions is essential The platform's interactive and fast-paced environment may enhance impulsive buying tendencies This study aims to address the existing research gap by investigating the effects of TikTok Live and its in-app shopping feature on impulsive buying behavior, utilizing the Stimulus-Organism-Response (S-O-R) model to elucidate the psychological processes involved Ultimately, this research provides valuable insights for businesses and marketers looking to leverage TikTok's influence in the marketplace.
The S-O-R model is the leading theoretical framework for understanding online impulse purchases, widely adopted by researchers to analyze consumer behavior characterized by sudden emotional responses to stimuli A recent study by Safira Aulia Rahma et al (2023) illustrates the application of this model in examining how marketing strategies influence impulsive buying behavior on TikTok.
This study employs the S-O-R theory to explore marketing stimuli in s-commerce live streaming and their influence on impulse buying behavior It specifically investigates how perceived product scarcity, streamer trustworthiness, and social presence affect perceived enjoyment and involvement, ultimately impacting impulse purchases Given Vietnam's potential as a burgeoning market for live-streaming commerce, the findings aim to enhance existing literature and offer practical insights for TikTok sellers, helping them optimize channel management and improve the overall consumer purchasing experience.
Research objectives
This research seeks to enhance the limited literature on customer experiences with TikTok Live by providing valuable insights into the dynamic processes that shape the livestreaming environment and influence impulse buying behavior Specifically, the empirical study aims to examine these interactions and their implications for consumer behavior.
- Examine the effects of the three TikTok Live environmental cues (i.e., perceived product scarcity, streamer trustworthiness and social presence) on customers* experience.
- Explore the mediating role of perceived enjoyment on the relationship between the TikTok Live environmental cues and customers' impulse buying behavior.
- Explore the mediating role of involvement on the relationship between the TikTok Live environmental cues and customers* impulse buying behavior
The findings of this study offer important practical knowledge for TikTok sellers to optimize their channel management that delivers a seamless shopping experience to their customers.
Research scope and objects
The research was conducted over a period of 04 weeks from July 2023 to August
The study focused on customers who engage in live-streaming shopping and have a history of impulsive buying on TikTok Live, examining their behaviors at various times throughout the day on both weekdays and weekends.
The influence of the TikTok environmental cues on customer experience in using TikTok Live, and the relationship between customer experience and their purchase intention towards online beauty products.
Research methodology
This study employed a quantitative approach, beginning with the adoption and translation of measurement scales from previous research into Vietnamese A questionnaire was then developed and pre-tested with 20 customers to ensure clarity before distribution The main survey, conducted over four weeks via a web-based form (Google Form), was followed by data analysis using SmartPLS 4 The analysis involved assessing measurement scales, testing for common method bias (CMB), and evaluating the structural model through hypothesis testing.
Research contributions
This research aims to validate a scale measuring the influence of various factors on impulse buying behavior during TikTok Live sessions By examining the factors that drive impulse purchases, the study enhances our understanding of impulse buying behavior within this platform Utilizing the S-O-R framework, it explores the interactions between stimuli such as product scarcity, live streamer trustworthiness, and social presence, alongside the psychological aspects of perceived enjoyment and involvement, which lead to impulse buying responses The findings provide valuable insights for TikTok sellers, enabling them to optimize their live-streaming strategies, create engaging stimuli, and enhance the overall shopping experience, ultimately boosting impulse buying conversions on TikTok Live.
Research structure
The study has 5 chapters, including:
- Chapter 1: Introduction The author introduces the overview of the research topic
This article explores the factors that influence impulse buying behavior during live-streaming shopping on TikTok Live, framed within the Stimulus-Organism-Response (SOR) model It outlines the research background and problem statement, detailing the objectives, scope, and subject of the study Additionally, the chapter encompasses the research methodology and structure, providing a comprehensive overview of the investigation into consumer behavior in the context of live-streaming commerce.
Chapter 2 of the article presents a comprehensive literature review and hypothesis development focused on consumers' impulse buying behavior in TikTok Live streaming shopping The authors clarify essential concepts and critically analyze various theoretical models, ultimately identifying the stimulus-organism-response model as the most appropriate framework for their research Additionally, the chapter proposes eight hypotheses grounded in prior studies and relevant literature, highlighting the significance of understanding consumer behavior in this emerging shopping context.
In Chapter 3, the author details the research methodology, emphasizing the measurement scale, pilot test, sample design, data collection process, and data analysis methods, which collectively establish the validity of the findings To achieve the study's objectives, the author employs quantitative techniques to manipulate data effectively.
Chapter 4 presents the research results derived from primary data collected via a survey The analysis, conducted using IBM SPSS Statistics 22 and SmartPLS 4, allows the author to identify and elucidate the empirical findings of the study.
In Chapter 5, the author discusses the significant implications derived from key findings, emphasizing their relevance to business decisions The chapter summarizes the results obtained and offers thoughtful insights into potential improvements and changes Additionally, the author explores possible extensions of the findings, highlighting their importance for future business strategies.
LITERATURE REVIEW AND MODEL DEVELOPMENT
Introduction
Chapter 2 focuses on forming the theoretical basis for the research This chapter first provides an overview of relevant research about impulse buying behavior in the context of live streaming commerce Secondly, this chapter highlights the theoretical background of the research by reviewing the literature of the research model, live streaming commerce and online impulse buying behavior From the provided background, key factors influencing the intention to purchase are identified and the need for the study is highlighted as well The subsequent discussion of the S-O-R model explains the theoretical framework of the study, which is established based on the research gap the study sought to fill and hypothesis about the relationship between theories used in the model, will create a foundation for the research model presented at the end of the chapter In conclusion, this chapter consists of two main parts: (1) Theoretical background of the research constructs and (2) Research hypothesis and the research model development.
Related research
(1) Impulse Buying Behavior in Live Streaming Commerce Based on the Stimulus - Organism - Response Framework (Lee and Chen, 2021)
This study focuses on the impulsive buying habits of customers in live streaming commerce, utilizing the stimulus organism response (S-O-R) framework to analyze consumer reactions to various stimuli A total of 433 valid questionnaires regarding the purchasing process on live streaming platforms were collected, and PLS-SEM statistical analysis was employed for evaluation The findings indicate that perceived enjoyment significantly enhances the desire for impulsive purchases However, the study excludes trustworthiness as a stimulus factor due to concerns about live streamers potentially providing misleading information to attract customers Issues such as transaction disputes, fake goods, and poor quality have led to diminished trust in live streaming commerce, prompting a need to reassess the impact of streamer trustworthiness on customer experiences.
Figure 1 Research model of Lee and Chen (2021)
(2) Exploring the TikTok influences on consumer impulsive purchase behavior (Wan etal., 2022)
This study investigates the impact of TikTok on consumer impulse purchasing behavior in Malaysia through the Stimulus-Organism-Response (S-O-R) model It identifies product-related visual appeal and feasibility as key stimuli, while perceived enjoyment and usefulness serve as the organism, ultimately leading to impulsive purchase intentions as the response Utilizing SPSS and Smart PLS, data from 169 online survey responses were analyzed The findings reveal that both visual appeal and product feasibility positively affect subjective enjoyment and perceived usefulness Furthermore, TikTok users' perceptions of the platform's utility and enjoyment significantly enhance their likelihood of making impulsive purchases The paper concludes with theoretical insights and practical recommendations for marketers to harness TikTok's potential in driving impulsive buying behavior.
Figure 2 Research model of Chen et al (2020)
(3) The Impact Of Marketing Strategy On Consumer's Impulsive Buying Behavior
On TikTok Live (Safira et al., 2023)
This study explores the impulsive purchasing behaviors of TikTok Live users, proposing a marketing strategy based on the "People-Product-Place" framework within s-commerce live streaming Utilizing the stimulus-organism-response (S-O-R) theory, the research employed a purposive sampling method alongside SEM-AMOS analysis, gathering 170 responses through questionnaires and shopping experiences Findings reveal a significant correlation between users' perceptions of TikTok anchors and their impulsive buying behavior, highlighting the roles of perceived scarcity, involvement, immersion, and participation The concept of rivalry is crucial in understanding how scarcity influences emotional responses to promotions, as it fosters competition and a desire for success among viewers To enhance impulsive purchasing, anchors should prioritize viewer engagement, as higher involvement can lead to more efficient and optimal buying decisions.
Stimulate Organism Response rhree elements of Retail
Figure 3 Research model of Safira et al 2023
(4) How social presence influence impulse buying behavior in live streaming commerce? The role of S-O-R theory (Ming et al., 2021)
This study investigates the impact of social presence, telepresence, and flow state on consumer trust and impulsive buying behaviors in the context of live streaming platforms Utilizing the Stimulus-Organism-Response (S-O-R) framework and structural equation modeling (SEM), data was collected from 405 Chinese consumers who had engaged in live streaming purchases Findings reveal that both telepresence and various aspects of social presence significantly enhance consumer trust and flow state, leading to impulsive buying Furthermore, the relationship between consumer trust, flow state, and impulsive buying is moderated by the consumers' perception of personal power This research provides valuable insights for live streamers and online merchants, aiding in strategies to boost customer purchases and contributing to the growth of live streaming commerce globally.
Theoretical background
The S-O-R (Stimulus-Organism-Response) theory is a cognitive psychology framework that explores human behavior by integrating individual physiological and psychological processes into the traditional stimulus-response model Originally presented by Woodworth in 1929, the theory was later expanded by Mchrabian and Russell in 1974 and refined by Jacoby in 2002, who introduced the concept of the organism as a mediator between stimulus and response This model provides a comprehensive method for understanding how external stimuli influence human cognitive and emotional states (Shah et al., 2020).
The concept of "stimulus" in the model highlights environmental factors that trigger an organism's internal responses (Song et al., 2021) Research indicates that viewers experience a heightened sense of presence when engaging with live streamers, fulfilling their needs and shaping their attitudes and behaviors as potential customers (Gao et al., 2018) This presence in live streaming significantly influences consumer behavior The term "organism" encompasses the emotional and cognitive processes that mediate individual responses to stimuli (Wu and Li, 2018) Affective states reflect emotional reactions to external stimuli, while cognitive states pertain to thought processes in response to such stimuli (Sun and Zhang, 2015; Fu, 2018) This study investigates how viewers' emotional and cognitive states impact their sense of presence in live streaming commerce, focusing on consumer trust and flow state.
In e-commerce, consumer behavior is shaped by emotional and cognitive influences, leading to three main types of purchasing intentions: planned buying, impulsive purchases, and return intentions This study primarily focuses on impulsive buying behavior, reinforcing the validity of the SOR model by highlighting the interconnectedness of these consumer responses.
Live streaming commerce is a dynamic e-commerce model that utilizes live streaming videos on various platforms to facilitate transactions (Xu et al., 2020) This innovative approach employs various stimuli to engage potential customers, as streamers create immersive virtual environments on social media and livestream platforms (Wongkitrungrueng and Assarut, 2020) Through real-time interaction, consumers gain access to accurate information, enjoy entertaining presentations by charismatic streamers, and build social connections with both sellers and fellow viewers, enhancing their overall shopping experience (Sjoblom and Hamari).
2017) Thus, live streaming e-commerce seamlessly combines entertainment, social activities, and product information.
Livestreaming e-commerce relies on three essential components: live streamers, streaming platforms, and logistical support Live streamers come from diverse backgrounds, each bringing unique expertise, reputations, and presentation styles Often categorized as Key Opinion Leaders (KOL) or Key Opinion Customers (KOC), these influencers significantly impact their audience's purchasing decisions through the trustworthiness and authenticity of their content.
Zhao et al (2018) highlight that every live streamer possesses unique streaming preferences, specialties, target audiences, and varying levels of appeal These streamers can effectively communicate business information in an engaging manner, provide immersive product introductions, and cater to client needs within virtual environments.
Live streaming has emerged as a valuable decision-support system for customers during their online shopping experience, offering live product demonstrations and detailed explanations (Cai et al., 2018) This interactive approach enhances participation, visualization, and authenticity in online purchases TikTok Live serves as a prominent platform for social commerce, allowing anchors to share their shopping experiences and recommend products directly to viewers Customers can easily access live product offerings by clicking the "live" icon while browsing, enabling them to engage with anchors in real-time through the comment section Orders can be placed instantly via the yellow cart icon without leaving the live feed, ensuring a seamless shopping experience Additionally, the audio from the live stream remains active on the transaction page, allowing customers to stay informed about product details even after initiating a purchase.
Impulsive purchasing, as defined by West (1951), refers to a spontaneous buying decision made by customers who lack a prior shopping plan Rook (1987) further refines this concept, describing impulse buying as a behavior characterized by a sudden and intense desire to make an immediate purchase This unplanned buying behavior is triggered by stimuli encountered within the store environment, as noted by Applebaum.
In live streaming commerce, impulse buying behavior is driven by consumers making spontaneous and emotional purchase decisions during live sessions Key external factors such as price, visual appeal, social influence, and vendor creativity significantly impact these impulsive purchases Live streamers who showcase products, demonstrate their use, and interact with viewers can heighten shopping awareness, leading to increased impulsive buying tendencies As live streaming commerce gains popularity, businesses must understand and leverage these influencing factors to effectively engage consumers and capitalize on their impulsive buying behaviors.
Perceived product scarcity, as defined by Hamilton et al (2019), refers to the lack of available products and services that consumers can access, either immediately due to stock shortages or later due to regulatory issues In the realm of live streaming, scarcity highlights the limited availability of products based on time or quantity, with "time limitation" indicating a fixed sales period and "quantity limitation" capping the number of items available for purchase (Tueanrat et al., 2021) This sense of scarcity can lead to feelings of restriction, prompting consumers to act impulsively to regain a sense of freedom Consequently, the perceived scarcity triggers emotional and cognitive responses that drive impulse buying behavior, making it a crucial stimulus for consumers.
In the realm of live streaming, a streamer's trustworthiness is defined by the perceived honesty, integrity, and credibility they convey to their audience When viewers trust streamers, they are more likely to consider the products being promoted as valuable, leading to increased purchase intentions Therefore, a high level of trustworthiness among streamers not only boosts customer confidence but also significantly enhances their likelihood of making a purchase (Song and Liu, 2021; Lu and Chen).
In 2021, trust in streamers significantly reduces consumer uncertainty regarding product quality and fit, thereby enhancing their desire to purchase and encouraging impulsive buying behaviors Influencers with high credibility positively influence consumers' perceptions of brand reliability and their intention to buy (Chung and Cho, 2017) The situational context of live streaming commerce acts as a catalyst, shaping consumers' cognitive and emotional evaluations Many live broadcasters have established themselves as trustworthy sources, attracting viewers to the products and brands they endorse (Xu et al., 2020) Consequently, credible influencers foster positive attitudes toward brand trustworthiness and purchasing intentions (Chung and Cho).
2017) As a result, the trustworthiness of the streamer in live streaming commerce is a powerful motivator to alter consumer behavior.
The concept of social presence, rooted in the ideas of directness and intimacy from social psychology, plays a crucial role in online live shopping experiences Unlike traditional human-to-human interactions, the "human-to-machine contact experience" in live shopping offers a unique dynamic With advancements in virtual reality, social presence has emerged as a key element in enhancing online shopping During livestream shopping, consumers benefit from real-time visibility of sellers and can engage through text, fostering a strong sense of social presence that enhances their interpersonal communication and positive perceptions of the interaction The live broadcast interface often includes a direct payment link, allowing consumers to make quick purchasing decisions based on real-time updates and direct seller advice.
In the context of live streaming commerce, "stimulus" refers to environmental elements that activate an organism's internal states (Song et al., 2021) Research indicates that viewers experience a heightened sense of presence when engaging with live streamers in real time, which not only fulfills their needs but also shapes their attitudes and behaviors as potential customers (Gao et al., 2018) Consequently, the sense of presence in live streaming commerce serves as a powerful stimulus that significantly influences consumer behavior.
This study explores the influence of hedonism on consumer behavior in live streaming commerce, emphasizing enjoyment as a key factor Defined by Park et al (2012) as the intrinsic joy experienced during interactions, enjoyment becomes crucial when users find live streaming shopping pleasurable and rewarding The technology behind live streaming facilitates real-time audio-visual content delivery, creating an immersive environment that enhances viewer presence Additionally, the participation of influencers sharing personal experiences fosters a co-experience that boosts engagement and enjoyment (Briindl et al., 2017) This heightened emotional response can lead to increased impulse buying and immediate purchasing decisions (Lee and Park, 2014) Therefore, developing e-commerce systems that understand and enhance customer delight is vital for promoting impulse buying behavior (Do et al., 2020).
Hypothesis development
2.4.1 Effects of perceived product scarcity, streamer trustworthiness, and social presence on perceived enjoyment
Research by Shah et al (2015) indicates that consumer perceptions are significantly influenced by product scarcity, leading to an increased desire and perceived value of the product (Akram et al., 2018) This heightened scarcity enhances both perceived enjoyment and overall value in the eyes of consumers Furthermore, Hamilton et al (2019) suggest that promoting products as available for a limited time through live streaming commerce can enhance the customer experience, fostering greater emotional investment due to the product's limited availability Therefore, the study proposes the following hypothesis.
Hl: There is a positive relationship between product scarcity and perceived enjoyment.
Trustworthiness in live streaming refers to a streamer's perceived honesty, integrity, and plausibility, which significantly enhances viewer enjoyment, engagement, and loyalty When viewers trust a streamer, they believe in the quality of service provided and feel assured that their interests will not be exploited (Wongkitrungrueng and Assaru, 2020; Lu et al., 2010) This trust is fostered through a strong emotional connection between the audience and the streamer (Kim and Park, 2013) As viewers recognize the streamer's expertise and ability to meet their needs, they are more likely to continue watching and seek assistance (Zhang et al., 2022) Additionally, interactive features like bullet screens allow viewers to ask questions and share comments, further enhancing their enjoyment of the live stream (Hsu, 2022).
In overall, high streamer trustworthiness encourages customers to buy the products and can raise their enjoyment Thus, the study proposes the following hypothesis:
H2: There is a positive relationship between streamer trustworthiness and perceived enjoyment.
Social presence on live streaming platforms significantly enhances viewer interaction and connection with both the streamer and fellow viewers, leading to a greater sense of community and enjoyment Engaging through live chat, comments, and interactive features boosts overall satisfaction, as highlighted by Lombard and Ditton (1997), who noted that social presence primarily generates pleasure Research indicates that social presence not only increases enjoyment (Choi, 2016; Liu et al., 2019) but also fosters a sense of belonging among viewers (Gao, 2017) In the realm of e-commerce, Hassanein (2007) emphasizes that social presence influences the enjoyment of the shopping experience, while Cyr (2007) confirms a positive correlation between social presence and perceived pleasure Ultimately, heightened social presence leads to a more enjoyable experience and a stronger hedonic mindset (Lombard et al., 1997), supporting the hypothesis that social engagement is crucial for enhancing viewer satisfaction.
H3: There is a positive relationship between social presence and perceived enjoyment.
2.4.2 Effects of perceived product scarcity, streamer trustworthiness, and social presence on involvement
Scarcity drives individuals through competitiveness and a desire to succeed, as noted by Malhotra (2010) It can imply a perceived threat to personal freedom, while time and material constraints positively influence customer engagement (Wang et al., 2022) In live-streaming commerce, anchors create a competitive shopping environment to boost consumer involvement by implementing strategies like time-limited campaigns and rewards for active participation, such as leaving comments This social interaction not only enhances customer engagement but also fosters a sense of community around the live-stream event, leading to the following hypothesis.
H4: There is a positive relationship between perceived product scarcity and involvement.
Trust in a streamer is the belief that they provide quality service without exploiting or harming their audience, fostered by the emotional bond between the streamer and viewers When audiences perceive streamers as experts who can fulfill their needs, they are more inclined to watch their content and seek their guidance This trust enhances viewer engagement during live streaming, as it creates a sense of reliability and authenticity Trustworthy streamers are recognized as credible sources, motivating viewers to actively participate in discussions and interactions Therefore, the study posits a significant positive relationship between streamer trustworthiness and viewer involvement.
H5: There is a positive relationship between streamer trustworthiness and involvement.
Social presence plays a crucial role in live streaming, significantly influencing consumer experiences and impulse buying behavior Research by Lin and Lee (2019) indicates a positive correlation between social presence and viewer involvement When viewers feel a heightened sense of social presence during a live stream, they are more likely to engage actively with the content, interact with the streamer, and connect with other viewers This engagement can take various forms, including commenting, participating in live chats, sharing the stream, or becoming loyal subscribers to the streamer's channel Therefore, the study proposes a hypothesis linking social presence to increased viewer involvement.
H6: There is a positive relationship between social presence and involvement.
2.4.3 Effects of perceived enjoyment and involvement on impulse buying behavior
In the realm of live streaming, impulse buying is significantly influenced by the excitement and enjoyment viewers experience while engaging with live content This positive emotional response serves as a key driver of impulse buying behavior, as highlighted by Verhagen and Dolen.
Engaging live streams create an emotional connection between viewers and streamers or showcased products, fostering trust and familiarity that can trigger impulse purchases Research by Shen and Khalifa (2012) highlights that arousal and pleasure significantly enhance the likelihood of impulsive buying Furthermore, customers who enjoy live streaming shopping are often influenced by marketing promotions, which can further stimulate impulsive purchasing behaviors (Xu et al., 2020) This enjoyment also encourages exploratory behavior, such as increased browsing, ultimately leading to more impulsive decisions (Guo and Poole).
2009) According to Xiang et al (2016) customers' perceived enjoyment of a social commerce platform has a beneficial impact on their propensity to make impulsive purchases Thus, the study proposes the following hypothesis:
H7: There is a positive relationship between perceived enjoyment and impulse buying behavior.
In live streaming commerce, consumer involvement enhances product visualization and helps predict preferences (Dimoka et al., 2012; Hong and Pavlou, 2014) High-quality product descriptions improve consumer perception and increase emotional engagement (Noriega, 2008) Engaged customers are better at managing risk, potentially leading to impulsive purchases (Venkatraman, 2006) Overall, customer involvement is a critical factor influencing online purchasing decisions (Prentice et al.).
Participation in various product categories significantly influences impulse purchasing behaviors (Jones et al., 2003) Impulsive purchases often stem from strong emotions triggered by the proximity to a desirable item (Rook and Gardner, 1993) Additionally, a higher level of consumer involvement with a product correlates with increased impulse buying tendencies (Liang, 2012) Therefore, this study proposes a hypothesis based on these findings.
H8: There is a positive relationship between involvement and impulse buying behavior for the current study.
Summary
This chapter outlines the research framework based on the SOR model and a comprehensive literature review of its constructs, alongside four pertinent studies Additionally, it introduces eight proposed hypotheses The subsequent chapter will focus on the methodology employed in this study.
RESEARCH METHODOLOGY
Introduction
Chapter 2 establishes the theoretical foundation necessary for the advancements discussed in Chapter 3, which focuses on three key areas: the design of the research procedure, the implementation of quantitative research, and the development of measurement scales.
Research procedure design
• Reliability (Cronbach's Alpha & • The collinearity issues (VIF value) composite reliability) • The predictive power (R2)
• Convergent validity (AVE) •Hypotheses testing (bootstrapping 5,000):
• Discriminant validity (HTMT) direct effects, mediating effects
Measurement scale
Following a comprehensive literature review and detailed interviews, the author developed and refined 25 research-related questions The study employed a 5-point Likert scale questionnaire to assess various variables, allowing respondents to express their level of agreement ranging from "1" (strongly disagree) to "5" (strongly agree).
The TikTok Live environment is characterized by three key elements: perceived product scarcity, streamer trustworthiness, and social presence Perceived product scarcity is assessed using four observed variables derived from Akram (2018)’s single-item constructs Meanwhile, streamer trustworthiness is evaluated through four items that have been adapted based on the research of Park et al (2020), Xiao et al (2018), and Ohanian (1990).
Four single-items of social presence were adopted from the study of Barfield et al (1995); Hassanein and Head (2007); Wang et al (2009).
The customer experience on TikTok Live is influenced by perceived enjoyment, which is measured using five items derived from established scales by Parboteeah et al (2009), Xiang et al (2016), Lee et al (2021), and Parboteeah et al (2016) Additionally, involvement is assessed through a single-item construct based on the same authors' research Furthermore, impulse buying behavior among customers is evaluated using four items adapted from the findings of Jones et al (2003).
PPS1 When I shop on TikTok Live, I consider when the item will go off the shelf.
PPS2 When 1 shop on TikTok Live, I worry about the limited time it is available to purchase.
PPS3 When I shop on TikTok Live, I consider the limited quantity of that item.
PPS4 When I shop on TikTok Live, I worry that the item is sold out.
STL I feel the live streamer on TikTok
Live is dependable Park et al (2020);
Xiao et al (2018) and Ohanian (1990) ST2 I feel the live streamer on TikTok
ST3 I feel the live streamer on TikTok Live is trustworthy.
ST4 Ĩ feel the live streamer on TikTok Live is sincere.
SP1 In the scenario of livestreaming on TikTok, I feel as if the products are in front of me.
Barfield el al (1995); Hassanein and Head (2007); Wang et al
SP2 In the scenario of livestrcaming on TikTok, I feel like I am shopping in a real shopping mall.
SP3 In the scenario of livestreaming on TikTok, I feel like I'm dealing with people face to face.
SP4 In the scenario of livestreaming on TikTok, I feel like the streamers, and I know each other's thoughts.
PEI Shopping with live streaming commerce was exciting.
PE2 Shopping with live streaming commerce was enjoyable.
PE3 Shopping with live streaming commerce was interesting.
PE4 I found my visit to live streaming commerce was fun.
PE5 Shopping with live streaming commerce was fun for its own sake.
Pilot test
INI I am very interested in the products and services offered by TikTok Live.
IN2 The products promoted in the TikTok Live are important enough for me.
IN3 My level of involvement with the products and services offered by TikTok Live is high.
IN4 I am particularly engaged with TikTok Live’s buying and selling environment.
IB1 I purchased a product I did not originally intend to buy or purchased a virtual gift to reward the streamer.
IB2 I have noticed a lot of products I have recently purchased via TikTok Live are rarely used.
IB3.1 did not think deeply when I bought these products when watching TikTok Live, or when 1 bought virtual gifts to reward the streamers.
Watching TikTok Live often compels me to purchase products or virtual gifts for streamers, as the experience creates an irresistible urge to engage and reward.
A pilot test involving 20 participants was conducted to assess the clarity of the final questionnaire and the interview duration before employing a convenience sampling method To ensure representative feedback, participants from diverse demographics, including gender, age, monthly income, and weekly TikTok Live usage, were invited The author provided a brief overview of the test's objectives, encouraging participants to ask questions about any confusing aspects and to share their comments.
Regarding survey feedback, all twenty participants agreed that five minutes was an appropriate amount of time Additionally, they said that each question was simple to understand.
Sample design
A recent report by Singapore-based DataReportal highlights Vietnam's significant presence on TikTok, ranking 6th globally with approximately 49.9 million users Livestream shopping has emerged as the most popular activity, with 62% of respondents indicating that shopping is their primary reason for watching live streams The average viewing time for livestreams is 1-3 hours per week, according to 38% of participants With around 77.93 million internet users in Vietnam as of February 2023, this means that 64% are active on TikTok Notably, a TikTok study reveals that half of its users have made direct purchases while engaging with TikTok Live In November 2022, TikTok Shop generated impressive sales of VND 1,698 billion, selling 13 million products and involving 32,000 sellers, positioning it just behind Lazada and Shopee in the Vietnamese e-commerce landscape This research focuses on individuals participating in livestream shopping via TikTok Live.
According to Suhr (2006), the sample-to-item ratio should ideally be at least 5-to-1, meaning that a study with 25 items would necessitate a minimum of 125 respondents Additionally, Suhr suggests that a 10-to-1 ratio could be a more realistic goal, which indicates that the 276 respondents in the current study are acceptable.
Data collection
In July and August 2023, a comprehensive survey was conducted in Vietnam's urban areas over two weeks to capture diverse consumer behaviors Data collection occurred at various times, including weekdays and weekends, to ensure a representative sample Questionnaires were distributed through face-to-face interactions and social media platforms like Facebook, utilizing a convenient sampling method due to limited resources The research aimed to include a wide range of respondents, varying in gender, age, income, and TikTok Live engagement frequency, resulting in a nuanced dataset A total of 335 questionnaires were distributed, with 276 valid responses analyzed, ensuring confidentiality for all participants.
Data analysis method
Descriptive statistics were conducted to provide a general picture of respondents' backgrounds and choices Numerical information is summarized and presented in a manner that is illuminating and useful.
This study utilized single respondents to assess both independent and dependent variables, necessitating the examination of common method variance prior to implementing the measurement model To address this concern, Harman's one-factor test was performed to detect any presence of common method variance, which poses a risk to the validity of survey data and can lead to erroneous interpretations.
A confirmatory factor analysis was conducted to assess Harman's one-factor test, following the guideline that common method variance is indicated when a single component accounts for over 50% of the covariance in both dependent and independent variables (Podsakoff et al., 2003).
3.7.3 Partial least square - structural equation modeling
This study utilized Partial Least Squares Structural Equation Modeling (PLS-SEM) to evaluate the research model, as it effectively assesses the psychometric properties of constructs, including reliability and validity, while also examining the relationships between exogenous and endogenous variables (Chin et al., 1998) PLS-SEM's variance-based approach is particularly advantageous for predictive applications, emphasizing explained variance (R²) (Hair et al., 2014) Consequently, this method is well-suited for analyzing relationships and predicting outcomes within the context of our research.
In this study, we employed the SEM-PLS method using SmartPLS version 4.0.9.5 for data analysis, following a structured approach Initially, we assessed the measurement model to ensure the reliability and validity of the constructs, utilizing the Fornell-Larcker criterion and the heterotrait-monotrait ratio (HTMT) to determine discriminant validity Subsequently, we analyzed the structural model to explore the direct relationships between exogenous and endogenous variables, implementing a bootstrapping procedure with 5000 bootstrap resampling to enhance the robustness of our findings.
The initial phase of evaluating the measurement model involves assessing the convergence and discriminant validity of indicators and latent variables, as outlined by Fornell and Larcker (1981) This step aims to analyze the relationship between latent variables and their indicators The criteria for measuring reliability and convergent validity differ depending on whether the variables are reflective, formative, or a mix of both (Dijkstra, 2010) In this research, all latent variables are reflective, thus focusing solely on relevant criteria such as indicator loadings, Cronbach's alpha, composite reliability, and average variance extracted (AVE).
To evaluate a reflective measurement model, the first step is to analyze the indicator loadings, which reflect the relationship between indicators and their constructs (Henseler et al., 2009) Hair et al (2018) state that loadings above 0.708 are deemed acceptable, as they indicate that the construct accounts for over half of the variance in the indicator, thus ensuring sufficient item reliability Although loadings between 0.4 and 0.7 can be utilized in initial scale development, they should be considered for removal only if their exclusion enhances composite reliability (Chin, 1998b; Henseler et al., 2009).
Composite reliability, or construct reliability, assesses the internal consistency of scale items Higher composite reliability values typically signify greater reliability, as noted by Hair et al (2018) This measure is essential for ensuring the dependability of research instruments.
According to 1998 research, reliability values of 0.6 are deemed acceptable for exploratory studies, while Nunnally (1994) asserts that composite reliability should reach at least 0.7 in early research and 0.8 or 0.9 in more advanced studies However, values exceeding 0.95 may indicate redundancy among items, which can compromise construct validity (Diamantopoulos et al., 2012; Drolet and Morrison, 2001).
Cronbach’s alpha is a key reliability measure in test development and usage, recognized as one of the most significant statistics in research (Cortina, 1993) It has become a standard in multiple-item measurement studies (Schmitt, 1996) According to Cronbach (1951), a value above 0.7 is deemed acceptable for evaluating the reliability of measurement instruments.
Convergent validity is a crucial aspect to be examined in the measurement model
Convergent validity measures how closely related items within a scale correlate, with the Average Variance Extracted (AVE) serving as the key metric for assessment To calculate AVE, the squared loading of each indicator on the construct is averaged According to Fornell and Larcker (1981), an AVE value of 5 or higher is ideal, indicating that the latent variable effectively represents over half of the variation in its observed variables Conversely, an AVE below 5 suggests that the observed variables contain more error than the variance explained by the latent variable, reflecting poor convergence.
Following the establishment of convergent validity, the next crucial step is to assess discriminant validity within the measurement model Discriminant validity is essential as it verifies that a measurement effectively differentiates and accurately represents distinct phenomena of interest, setting them apart from other measures in a structural equation model (Hair et al.).
Henseler et al (2015) introduced the HTMT (Heterotrait-Monotrait) ratio of correlations as a novel method for evaluating discriminant validity According to the HTMT rule of thumb, problematic discriminant validity is indicated by values approaching 1 Researchers have suggested varying threshold values for HTMT; Clark and Watson (1995) and Kline (2010) recommend a threshold of 0.85, while Gold et al (2001) advocate for 0.90 Furthermore, Henseler et al (2015) assert that a threshold of 0.90 should be applied to structural models involving constructs that are conceptually similar, such as cognitive satisfaction, affective satisfaction, and loyalty (Hair et al., 2018).
When assessing a structural model, it is vital to first investigate the presence of multicollinearity, as this can significantly impact regression results Ensuring that collinearity does not affect the relationships within the model is crucial for accurate evaluation.
To identify the existence of multicollinearity in the research model, the Variance Inflation Factor (VIF) index is often used According to Hair, Ringle, and Sarsledl
In 2011, it was established that a model demonstrates significant multicollinearity when the Variance Inflation Factor (VIF) exceeds 5 Nevertheless, research by Mason and Perreault (1991) indicates that multicollinearity concerns can still arise even when the VIF is between 3 and 5.
In structural model evaluation, once collinearity is confirmed not to be an issue, the next focus is on the coefficient of determination (R²) This essential metric, as highlighted by Chin (1998a), assesses the quality of the model by quantifying the variance explained by each endogenous construct R² serves as a key indicator of the model's explanatory capacity, as noted by Shmueli and Koppius (2011).
Summary
Chapter 3 discussed the research methods and the process of how this research was implemented As a result, we presented the foundation to build measurement scales and generalize the equation of choosing samples.
RESEARCH RESULTS
Introduction
This chapter outlines the findings from the data collection process, which includes descriptive statistics of demographic profiles, an evaluation for Common Method Variance, an assessment of the measurement model, an analysis of the structural model, the results of direct and indirect effects, and a predictive assessment.
Descriptive statistics
Times spent on TikTok Live per week
Table 2 has shown the characteristics of the respondents involved in the survey
The survey collected 276 valid samples, revealing that 34.80% of respondents are male and 65.20% are female, indicating a stronger interest in online shopping via TikTok Live among women The predominant age group is 18 to 25 years old, followed by 26 to 35 years old, while those aged 36 to 45 represent the smallest percentage This age distribution aligns with Generation Z's familiarity with technology and diverse shopping methods In terms of income, 52.20% of respondents earn between 5-10 million VND monthly, and 27.20% earn below 5 million VND, suggesting a prevalence of low to medium income levels Conversely, respondents with monthly incomes above 20 million VND tend to avoid impulsive purchases on TikTok Live Additionally, 61.60% of respondents spend 2-7 hours per month on TikTok Live, with only 12.3% spending 7 to 21 hours, indicating frequent engagement that may encourage impulse buying behavior.
Common method variance (Harman single factor test)
Common method variance occurs when a single factor accounts for a significant portion, usually over 50%, of the covariance between dependent and independent variables, as indicated by Harman's single-factor test results.
3) reveal 25 factors with eigenvalues greater than 1 emerging from the analysis Approximately 47.234% of the variance is accounted for by the first factor, which is below the 50% limit tolerable to consider that variations in responses are not caused by the instrument These findings propose that this study does not suffer from the CMV issue.
Table 3 Testing results of common method variance
Initial Eigenvalues Extraction Sums of Squared
Measurement model assessment
In the initial phase of our study, we assessed the measurement model to confirm the convergent and discriminant validity of the constructs This evaluation involved a detailed analysis of indicator loadings, Cronbach's alpha, composite reliability, and Average Variance Extracted (AVE), with the findings presented in Table 4.
The analysis of Table 4 reveals that the indicators' loadings exceed the threshold of 0.707, indicating that over 50% of the variance in each indicator is accounted for by its corresponding latent construct, thus confirming indicator reliability (Benitez et al., 2020; Hair et al., 2017) Additionally, the study meets the Cronbach's alpha and composite reliability criteria, with values ranging from 0.867 to 0.951 for composite reliability and 0.795 to 0.931 for Cronbach's alpha, highlighting a strong correlation among indicators Furthermore, the Average Variance Extracted (AVE) values, ranging from 0.620 to 0.828, exceed the acceptable threshold of 0.5, indicating that the constructs effectively capture the variance of their respective items (Hair et al., 2017) Together, these metrics provide robust evidence for the convergent validity of the study's constructs.
Table 4 Testing results of reliability and convergent validity
Construct Items Loadings Cronbach’s alpha
Discriminant validity refers to the distinctiveness of a construct, ensuring that it represents a unique phenomenon not captured by other constructs in the model (Hair et al., 2013) As demonstrated in Table 5, the HTMT values for all constructs are below the acceptable threshold of 0.85, as recommended by Clark and Watson (1995) and Kline (2010) This indicates that the measurement model exhibits sufficient validity and discriminant validity.
Table 5 Testing results of discriminant validity
IB IN PE PPS SP ST
Structural model assessment
Before evaluating the relationships in the structural model, it is essential to confirm that collinearity does not impact the regression outcomes The VIF values presented in Table 6 are below the standard threshold of 5, indicating that the models are free from multicollinearity issues (Hair et al., 2011).
Table 6 Testing results of VIF
IB IN PE PPS SP ST
The significance of the structural parameters was assessed using a bootstrapping procedure with 5,000 samples The results, illustrated in Figure 7, include path coefficients, significance levels, and variance explained, all of which aligned with the anticipated algebraic signs and demonstrated statistical significance, indicated by t-values exceeding 1.96 at the 95% confidence level Consequently, all hypotheses were confirmed as valid based on these criteria.
Figure 7 Structural model of the study
According to Hair et al (2011), the coefficient of determination, or R2, measures the significance of path coefficients The R2 value for perceived enjoyment is 0.545, indicating that 54.5% of the variance is attributed to augmented reality (AR) attributes, while 55.5% is explained by other independent variables and random error For involvement, the R2 value is 0.498, suggesting that TikTok environment cues, such as perceived product scarcity, streamer trustworthiness, and social presence, account for 49.8% of the variance Additionally, the R2 for purchase intention is 0.609, meaning that 60.9% of the variation in impulse buying behavior is explained by the mediating variables of perceived enjoyment and involvement.
Results of direct and indirect effect
Table 7 Results of direct effect
Hl: There is a positive relationship between perceived product scarcity and perceived enjoyment 0.282 4.532 Supported
H2: There is a positive relationship between streamer trustworthiness and perceived enjoyment 0.240 2.809 Supported
H3: There is a positive relationship between social presence and perceived enjoyment 0.318 4.091 Supported
H4: There is a positive relationship between perceived product scarcity and involvement 0.288 4.434 Supported
H5: There is a positive relationship between streamer trustworthiness and involvement 0.187 2.593 Supported
H6: There is a positive relationship between social presence and involvement 0.327 4.223 Supported
H7: There is a positive relationship between perceived enjoyment and impulse buying behavior 0.435 6.036 Supported
H8: There is a positive relationship between involvement and impulse buying behavior 0.377 5.327 Supported
Hypothesis 1 posits that perceived product scarcity positively affects customers' perceived enjoyment during TikTok Live sessions The findings reveal a significant and strong correlation, with standardized coefficients of 0.282, a significance level of p < 0.05, and a t-value of 4.532, exceeding the critical threshold of 1.96 Therefore, the results support Hypothesis 1.
Hypothesis 2 assumes that streamer trustworthiness positively influences customer's perceived enjoyment during TikTok Live sessions The result indicates a strong impact of compatibility on customers' perceived enjoyment with large magnitude and positive significance The effect is consistent with standardized coefficients at 0.240, significant level at p < 0.05 and t-value at 2.809 which is higher than the threshold of 1.96 Thus, hypothesis 2 is supported.
Hypothesis 3 assumes that social presence positively influences the customer’s perceived enjoyment during TikTok Live sessions Indeed, the results of the linear structural model SEM show the significant value of the factor al p < 0.05, standardized coefficients at 0.318 and t-value at 4.091 which is higher than the threshold of 1.96 Thus, hypothesis 3 is supported.
Hypothesis 4 assumes that perceived product scarcity positively influences customer's involvement during TikTok Live sessions The effect is consistent with standardized coefficients at 0.288, significant level at p < 0.05 and t-value at 4.434 which is higher than the threshold of 1.96 Thus, hypothesis 4 is supported.
Hypothesis 5 assumes that streamer trustworthiness positively influences customer's involvement during TikTok Live sessions The result indicates a strong impact of streamer trustworthiness on customers' involvement with large magnitude and positive significance The effect is consistent with standardized coefficients at 0.187, significant level at p < 0.05 and t-value at 2.593 which is higher than the threshold of 1.96 Thus, hypothesis 5 is supported.
Hypothesis 6 assumes that social presence positively influences customer’s involvement during TikTok Live sessions The effect is consistent with standardized coefficients atO.327, significant level at p < 0.05 and t-value at 4.223 which is higher than the threshold of 1.96 Thus, hypothesis 6 is supported.
Hypothesis 7 assumes that perceived enjoyment positively influences the customers’ impulse buying behavior during TikTok Live sessions As can be seen from Table 7, the results of the linear structural model SEM show the direct effect of perceived enjoyment on customers' impulse buying behavior Specifically, the value of standardized coefficient is at 0.435; t-value is higher than 1.96 and significant level is at p > 0.05 which is not statistically significant As a result, hypothesis H7 is not supported.
Hypothesis 8 assumes that the customer's involvement positively influences their impulse buying behavior during TikTok Live sessions The effect is consistent with standardized coefficients at 0.377, significant level at p < 0.05 and t-value at 5.3 27 which is higher than the threshold of 1.96 Thus, hypothesis 8 is supported.
Our analysis examined the indirect effects of TikTok environment cues—such as perceived product scarcity, streamer trustworthiness, and social presence—on customers' impulse buying behavior using a multiple mediator model Utilizing PLS bootstrapping, we identified significant mediation effects: perceived product scarcity influences impulse buying through perceived enjoyment and involvement; streamer trustworthiness affects impulse buying via perceived enjoyment and involvement; and social presence impacts impulse buying through perceived enjoyment and involvement All relationships were statistically significant, with confidence intervals not straddling zero.
Table 8 Results of indirect effect
Standardized beta coefficients p- value Lower limit Upper limit
Summary
This chapter presents findings from a quantitative study examining impulsive buying behavior among TikTok Live customers The research tested proposed hypotheses using a structural model, validated through Cronbach's alpha, composite reliability, and average variance extraction Results confirmed that the measurement model met the standards for convergent and discriminant validity Additionally, the structural model demonstrated strong predictive capability, indicated by the R2 value reflecting explained variation Notably, standardized p coefficients highlighted the significant impact of perceived enjoyment and involvement on customers' impulse buying behaviors on TikTok Live.
Overall, the findings of the quantitative research provide an understanding of the research model The next chapter presents the discussion and recommendations of the study.