HÒ CHÍ MINHBÁO CÁO TỐNG KÉT ĐÈ TÀI NGHIÊN cứu KHOA HỌC THAM GIA XÉT GIẢI THƯỞNG ‘’ NHÀ NGHIÊN CỨU TRẺ UEH” NĂM 2024 SHOPPING IN THE ALGORITHMIC AGE: EXPLORING THE IMPACT OF AI-DRIVEN P
INTRODUCTION
Background and statement of the problem
In today's rapidly evolving landscape, data is often referred to as the "new oil," highlighting its crucial role similar to that of oil during the Industrial Revolution The rise of digital marketing enhances the value of this data, making it essential for improving customer experiences while also reducing costs, optimizing storage, and enabling effective data analytics within organizations Therefore, the effective collection, analysis, and application of data are vital for businesses aiming to succeed in the modern digital age.
AI-driven personalization is rapidly becoming a key trend in marketing, leveraging its ability to analyze external data and adapt to achieve specific objectives By understanding user data, AI can identify individual needs and behaviors, providing customized recommendations at the right time This approach benefits both businesses and customers, indicating that AI-driven personalization will continue to evolve, leading to more refined and satisfying user experiences in the future.
In 2023, Statista reported that social networks have become essential to daily life, with over 4.95 billion users, representing about 61.4% of the global population This rise in social media usage is driving the growth of commercial activities on these platforms, creating opportunities for innovative advertising strategies, particularly on mobile devices One such emerging trend is shoppable video, which many businesses are now adopting to enhance their marketing efforts.
Shoppable video represents a succinct advertising format within social networks, designed to exhibit products and streamline direct consumer transactions (Ertekin,
Shoppable video enhances customer shopping experiences by personalizing them to meet individual preferences, enabling quick and convenient connections between users and products or vendors This innovative approach leverages real-time data and emphasizes interactivity and immediacy To fully implement shoppable video, it is essential to utilize customer data for deploying AI-driven personalized algorithms effectively.
AI-driven personalization offers significant benefits for businesses and users, yet it raises serious privacy concerns This creates a paradox where consumers enjoy personalized experiences but worry about the misuse of their personal data Consequently, issues of transparency and privacy governance emerge As applications and websites customize content based on user information, privacy apprehensions can affect consumer purchasing decisions.
Recent studies on personalization in various fields, including advertising, healthcare, and e-commerce, have primarily focused on user privacy concerns, often neglecting the underlying causes of these issues and the user's active role in managing their information (Wang et al., 2019; Dhagarra et al., 2020; Song et al., 2021) Additionally, many of these studies equate personalization with AI-driven approaches, concentrating mainly on measuring outcomes (Aksoy et al., 2023) However, there is a notable gap in research comparing the effectiveness of AI-driven personalization to traditional methods This highlights the urgent need for a comprehensive exploration of personalization, particularly within the context of AI-driven techniques.
This research investigates the impact of AI-driven personalization on consumer purchasing intentions through shoppable videos in Ho Chi Minh City, aiming to enrich existing literature on AI and user data personalization It addresses privacy concerns related to customer interactions in the context of shoppable videos, while also adapting to the changing dynamics of social media platforms The findings are intended to provide digital marketers and business practitioners with valuable insights for developing effective strategies Ultimately, the study seeks to identify better methods for implementing AI-driven personalization to enhance user experiences and boost personalized consumer purchase intentions through shoppable videos.
Research objectives
This study aims to systematically analyze the complex relationships between key variables of AI-driven personalization, including perceived personalization, perceived usefulness, trust, and privacy concerns It also seeks to empirically assess how these factors influence consumer purchasing intentions in the urban context of Ho Chi Minh City Ultimately, the research evaluates the significant impact of AI-driven personalization on consumer intent to make purchases.
This study explores the significant moderating role of information control in online commerce, emphasizing the importance of safeguarding personal information in today's digital landscape.
Based on the research findings, this study aims to provide strategic managerial implications to boost customer preferences for online shopping by effectively utilizing shoppable videos These insights are crucial for developing successful business strategies in the rapidly changing landscape of digital commerce.
Research objects
Research object: The impact of AI-driven personalized Shoppable Video on purchase intentions, considering privacy concerns and the moderating role of Information Control.
Survey participants: Individuals residing, working, and studying in Ho Chi Minh City who have previously engaged in shopping experiences through Shoppable videos on social media platforms.
- Location: This survey was conducted in Ho Chi Minh City, Vietnam.
- Timeframe: Data was collected over a period of 4 days (from 26/1/2024 to 30/1/2024).
Research method
To achieve the research objectives and address the inquiries posed by the research topic, a two-tiered approach was adopted:
The research team undertook a thorough synthesis and analysis of existing literature, enhanced by qualitative research through semi-structured interviews with AI and Marketing experts This preliminary phase focused on refining hypotheses, models, and research scales to prepare for the formal research that follows.
Formal research was conducted using quantitative methods through an online survey featuring a carefully designed questionnaire This survey aimed at a random probability sample of 300 individuals located within Ho Chi Minh City.
Collect survey questionnaire data from the results of online surveys.
- The information garnered from the survey underwent processing via the SmartPLS software.
To assess variable reliability, statistical analysis tools were employed, focusing on the credibility of outcome variables through outer loading coefficients and evaluating causal variables with outer weights Scale reliability was determined using the Cronbach’s Alpha coefficient, while Composite reliability was applied to outcome variables Furthermore, Average Variance Extracted (AVE) and F-square were utilized for the assessment of causal variables.
Research structure
Chapter 01 - Introduction, the current thesis is composed of four themed chapters:
Chapter 02 - Literature review and hypothesis development
Chapter 03 - Research method: This chapter is concerned with the method used for the current thesis, including the research process, measurement scale, questionnaire design, sample and data collection, as well as the sample characteristics.
Chapter 04 - Data analysis and results: This section analyzes the dataset of the research It consists of the following steps: assessment of measurement scales, test for common method bias, assessment of structural model
Chapter 05 - Discussion and conclusion: This final chapter briefs the important results of the current thesis Moreover, the research limitations and recommendations for further research are also mentioned.
LITERATURE REVIEW AND HYPOTHESIS DEVELOPMENT 12
Research theory
Shoppable videos are an innovative form of advertising commonly found on social media, allowing consumers to view and purchase products directly These mobile-friendly videos combine interactivity and immediacy, enabling viewers to engage with the content while seamlessly accessing clickable links and interactive elements This integration facilitates a smooth transition to the shopping platform, enhancing the overall consumer experience.
In 2017, research highlighted the effectiveness of direct marketing on mobile devices, especially through social media platforms, enabling sellers to maximize their marketing reach and capitalize on the "viral value" (Akpinar & Berger, 2017) This approach in social commerce allows consumers to seamlessly "shop where they socialize and socialize where they shop" (Zhou et al., 2013).
Shoppable videos have emerged as a key player in the expanding billion-dollar advertising industry, driven by increasing advertising expenditures (Williams et al., 2020) Unlike regular videos that offer flexibility and convenience for users to pause, search, and watch at their own pace, shoppable videos are designed to create an interactive experience with brands or influencers, aiming primarily to boost sales (Yeon et al., 2019) These videos allow viewers to explore, learn about, and purchase products directly within the content, enhancing the shopping experience by making it more intuitive and convenient Additionally, they serve as a powerful advertising tool for businesses, requiring innovative content that seamlessly integrates product information and facilitates easy cart additions (Taylor, 2019).
Personalization, a concept with deep historical roots and significant academic discussion, originated in the 1980s when Leonard Berry, a marketing professor at Texas A&M University, coined the term "relationship marketing." This foundational idea paved the way for the development of customization in marketing practices.
Personalization, as defined by Cheung et al (2021), refers to the intentional design and production of items tailored to individual specifications In marketing, Chandra et al (2022) describe personalization as the strategic creation of content and products that align with customer preferences, helping to reduce customer fatigue and cognitive load Additionally, Maslowska et al (2016) highlight that personalization extends to communication strategies, involving the customization of messages, services, and products based on individual digital footprints and interests.
The primary aim of personalization, as highlighted by Pappas et al (2014), is to meet customer needs and desires effectively In contrast to earlier research that concentrated solely on the end goals of personalization, Aksoy et al (2023) examine how consumers perceive personalization in a more complex way Their findings reveal a potential disconnect between actual personalization and how it is perceived, particularly regarding its impact on purchase intentions.
AI-driven personalization utilizes artificial intelligence technologies to customize experiences based on individual needs, preferences, and behaviors, making it a rapidly growing area of research Studies indicate that this approach offers significant benefits across various sectors, including marketing, advertising, promotions, and tourism It enhances customer segmentation, allowing for targeted engagement with high-potential user groups and precise advertising strategies The rise of hyper-personalization, which employs advanced AI and machine learning to meet users' real-time needs, further improves customer-centric marketing practices in online businesses.
Consumers are increasingly aware that their personal information is being collected, often prompted by website notices requiring data for access They face a choice: consent to data collection or forgo site access When advertisers obtain explicit or implied consent for data use, consumers tend to worry less about privacy However, many remain unaware that their information has been gathered until they receive marketing communications, which heightens their alertness and concerns As consumers realize that advertisers may acquire their data without consent, privacy worries escalate Altman (1975) emphasized the importance of privacy control, defining it as the selective management of access to oneself Building on this, Shehan and Hoy (2000) identified information control as a crucial factor influencing consumer privacy concerns in the online environment.
In the evolving landscape of digital marketing, consumer privacy has become a critical concern Privacy, as defined by Bart (2005), involves the safeguarding of personal information online and is closely tied to privacy policies, customer consent, and data disclosure practices For consumers, privacy encompasses their ability to manage the collection, usage, and storage of their personal data (Plangger and Montecchi, 2020) This concern reflects individuals' anxieties about potential violations of their privacy (Diev and Hart, 2006a) Furthermore, research by Song (2021) highlights that privacy concerns relate to the extent of consumer fear regarding the collection and misuse of their personal information, emphasizing the importance of transparency in how data is handled.
This study highlights privacy concerns as individuals' worries regarding the collection, use, and security of their personal data These concerns become particularly evident when using Shoppable Video, where the platform gathers users' personal information to recommend videos featuring products tailored to their preferences.
Privacy concerns significantly hinder the adoption of consumer technology Research by Unger (2012) indicates that consumers are less likely to share their information for online personalization due to privacy worries Additionally, Dhagarra (2020) highlights that unresolved privacy concerns can adversely affect consumers' decisions to utilize various services.
Trust is a psychological state characterized by the willingness to accept vulnerability based on positive expectations about another person's intentions or behavior It is established when one party believes in the trustworthiness and integrity of the other during exchanges In the context of e-commerce, trust extends beyond the brand itself to include confidence in the underlying technology.
In alignment with the Technology Acceptance Model (TAM) presented by Davis et al
Perceived usefulness, as defined by Davis (1989), refers to the belief that using a specific system will improve an individual's performance Chiu et al (2009) further elaborate on this concept in the realm of online shopping, describing it as the extent to which consumers feel that online transactions will increase their efficiency Overall, perceived usefulness reflects a user's belief that a system will enhance their work effectiveness.
Customer engagement lacks a universally accepted definition, as scholars approach it from various perspectives to influence desired customer behaviors (Baabdullah et al., 2018) While customer activities may align with social media contexts, they do not inherently add value to the company (De Oliveira Santini et al., 2020) This study defines customer engagement as the degree of interaction and connection between a customer (or potential consumer) and a brand, involving initiatives from both the company and the customer (Vukadin et al., 2018) As this interactive process evolves, the relationship between the client and the business becomes increasingly complex and intertwined.
Consumer purchase intention refers to a clear plan to buy a specific product or service in the future and is shaped by various internal and external factors Internal influences such as necessity, desire, and trust significantly affect this intention, while external factors like economic conditions, marketing promotions, social influences, and brand reputation also play a vital role Understanding these dynamics is essential for businesses aiming to enhance consumer purchase intentions.
Prior relevant studies
Customer experiences in the age of artificial intelligence
Perceived convenience, Personalisation, Al-enabled service quality, Trust,
Relationship commitment, Perceived sacrifice, AI-enabled customer experience
Survey The study emphasizes the importance of considering both the technical aspects of
AI and the human factors of trust, sacrifice, and relationship commitment when designing and implementing Al-powered customer experiences.
Balancing web personalizati on and consumer privacy concerns:
Privacy concerns, Web personalization, Trust,
Consumer privacy concerns significantly weaken the connection between web personalization and website loyalty, with the impact varying based on the degree of privacy apprehensions This relationship is influenced by factors such as consumer trust and reactance.
Exploring the Effects of Personalized Advertising on Social Network Sites
Interactivity, Credibility, Reciprocity, Relevance, Intimacy, Likeability, Consumer Engagement, Perceived Personalization, Purchase
Survey Consumer engagement and purchase intention are increased by perceptions of interactivity, intimacy, relevance, and likeability, thereby enhancing the effectiveness of personalized advertising.
Customer engagement in social media: a framework and meta-analysi s
Emotions, Customer engagement, behavior intention, word of mouth, performance
Customer engagement is a vital area of study, primarily due to its ambiguous definitions and its intersection with traditional marketing concepts This complexity necessitates additional research to clarify its unique role and significance within the marketing landscape.
Impact of Trust and Privacy Concerns on Technology Acceptance in Healthcare:
A recent study reveals that patients are more inclined to adopt healthcare technology when they perceive it as beneficial, user-friendly, and trustworthy in safeguarding their privacy This research enhances our comprehension of the impact that trust and privacy concerns have on the acceptance of technology in the healthcare sector.
How trust leads to online purchase intention founded in perceived
Task-Relevant Cues, Aesthetic Relevant Cues, Peer
The perceived usefulness of a brand's social media presence is shaped by atmospheric cues and the level of trust consumers have in the brand and its communication channels Trust not only enhances the perceived usefulness but also significantly influences purchase intentions Additionally, privacy concerns and peer communication play crucial roles in shaping this perception.
Trust in Brand, Trust in
Medium, Privacy Concerns, Purchase Intention limited.
The impact of social network marketing on consumer purchase intention in Pakistan: A study on female apparel
Consumer Purchase Intention, Consumer engagement
Survey The study suggests that social network marketing significantly influences customer purchase intention for women's fashion apparel Brand engagement and customer motivation partially mediate this relationship.
Consumers’ acceptance of domestic Internet-of-T hings: The role of trust and privacy concerns
Perceived usefulness, Perceived ease of use, subjective norm, Intention to use, Privacy concern, Trust
A recent survey revealed that the usefulness and trust in providers play a vital role in the adoption of domestic IoT technology, with user motivations varying significantly between current and potential adopters Interestingly, privacy concerns were found to have no direct effect on the intention to adopt IoT solutions.
When does information transparency reduce the downside of personalizati on? Role of need for cognition and perceived control
Personalization, Perceived control, Privacy concerns, need for cognition, information transparency
A recent survey reveals that the disclosure of data collection methods, whether implicit or explicit, does not affect the effectiveness of personalization on perceived control Interestingly, customers report feeling more in control when companies disclose implicit data collection practices This finding challenges earlier research on targeted advertising.
Lee & Personalisati Personalization, Experi Enhancing privacy assurance
’s (2011) on-privacy paradox: The effects of personalisati on and privacy assurance on customer responses to travel Web sites
Perceived usefulness, Privacy concerns, Privacy assurance, Self-Disclosure Intention,
Adoption Intention ment and study increases perceived usefulness and reduces privacy concerns, but personalization has no effect on privacy concerns or willingness to disclose personal information.
Privacy Concerns in Personalized Advertising Effectivenes s on Social Media
Personalized Advertising, Advertising Value, Purchase Intention,
Survey Personalized ads can be more effective when tailored to users' interests, but privacy concerns can limit their impact.
It is all in the name: A study of consumers' responses to personalized communicati on
Personalized, Perceived personalization, Attentive reading, Positive thoughts, Negative thoughts, Attitude Advertisement, Attitude brand, Intention
Survey When participants felt like a message was tailored to them, they paid more attention and formed both positive and negative thoughts about it
These thoughts ultimately influenced their overall altitude towards the message and their intention to act on it.
Privacy concerns about personalized advertising across multiple social media platforms in Japan:the relationship with information control and
Ad Avoidance, Attitudes towards personalized Ads
Survey privacy concerns play a crucial role in mediating the connection between consumers' control over their information and their tendencies to avoid ads, as well as their perceptions of the intrusiveness of personalized advertising Additionally, the level of information control that consumers possess influences how persuasion knowledge interacts with privacy concerns.
The importance of perceived trust, security and privacy in online trading systems
Perceived security, perceived privacy, perceived usefulness, perceived trust, perceived ease of use, behavioral intention
Users are more likely to trust and utilize online trading platforms that they perceive as secure, user-friendly, and informative Establishing trust through robust security measures is essential for online financial institutions.
A tripartite model of trust in Facebook: acceptance of information personalizati on, privacy concern, and privacy literacy.
Privacy Literacy, Acceptance of Information Personalization, Trust in Data institutions, Trust in
Survey Privacy concerns negatively impact trust, potentially threatening data-driven companies' revenue streams, especially among individuals with good privacy literacy, despite growing calls for privacy literacy education.
“We think you may like this”: An investigation of electronic commerce personalizati on for privacy-con scious consumers
Technology Acceptance Model (TAM), Personalization Features,
Intention to Use, Perceived Usefulness, Privacy
Concerns, General Willingness to Self-disclose
Lab experim ent and study
Consumers’ perceived usefulness of personalization technology positively influences their intention to use an e-commerce mobile app, with privacy concerns and self-disclosure moderating this relationship.
Increased perceived information control mitigates
(2009) online personalizati on: the moderating effects of information control and compensatio n
Control, Privacy Concern, Compensation, Behavioral
Intentions ent privacy concern's negative impact on positive behavioral intentions, while compensation enhances trust's salience to privacy concerns.
The role of live streaming in building consumer trust and engagement with social commerce sellers
Customer trust, Customer engagement, Utilitarian value, Hedonic value
The symbolic value of products significantly influences customer engagement by fostering trust in sellers Additionally, both utilitarian and hedonic values contribute to customer engagement indirectly, enhancing trust in products and subsequently in sellers.
The effect of online privacy policy on consumer privacy concern and trust.
Online privacy, Power distance, Trust, Privacy policy, Willing to provide personal information
Survey The study found a significant correlation between privacy policy content, privacy concern/trust, willingness to provide personal information, and trust, and also found a significant cross-cultural effect.
How customer engagement in the livc-strcamin g affects purchase intention and customer acquisition, E-tailer's perspective
Customer engagement behaviors, Purchase intention, Customer acquisition
Study Customer engagement indicators are not all positively related to purchase intention and acquisition, and the influence of "like" behavior on customer acquisition is not statistically significant.
Research framework and hypothesis development
2.3.1 The relationship between AI-driven Personalization and Perceived Personalization
The assessment of tangible personalization involves measuring objective factors, such as the number of personalization elements used and their ability to distinguish individuals In contrast, perceived personalization is a subjective experience shaped by users' interpretations and cognitive responses (De Keyzer et al., 2022) This perceived personalization is crucial for tailoring advertising messages to match consumer interests, reinforcing the link between perceived personalization and advertising relevance (Back & Morimoto, 2012) Recent studies have demonstrated the effectiveness of personalized advertisements on social media platforms, as highlighted by Tran et al.
Recent studies highlight the significance of personalization in enhancing customer perceptions of brands By aligning content with user preferences, cognitive resonance is achieved, which increases the perception of personalization Additionally, emotionally resonant personalization fosters a stronger sense of ownership over the user experience, leading to higher engagement levels This synthesis of findings guides the design of personalized experiences that cater to individual preferences and positively impact perceived personalization Therefore, we propose the hypothesis that
Hl: AI-driven Personalization is positively related to Perceived Personalization
2.3.2 The relationship between Perceived personalization and Perceived usefulness
Personalization plays a crucial role in reducing customers' information overload, as highlighted by Song et al (2021), and recognizes the limited capacity of individuals to process information (Lang, 2000) Liang et al (2012) argue that personalized services enhance customer utility more effectively than traditional services As a result, online retailers can expect positive customer attitudes when services are customized to meet individual preferences This is further supported by Dunne et al (2010), who note that many young people are drawn to social networks due to their personalized alignment with personal interests.
Personalization plays a crucial role in enhancing customer satisfaction, as highlighted by Komiak and Benbasat (2006), who emphasize that tailored products and services cater to individual needs However, Song et al (2021) warn that the effectiveness of personalization relies on customers recognizing the relevance of recommendations to their personal preferences Therefore, the success of personalized marketing strategies is dependent on how well customers perceive these recommendations as aligned with their unique interests Based on this understanding, we propose the following hypothesis.
H2: Perceived personalization leads to greater perceived usefulness.
2.3.3 The relationship between Perceived usefulness and Privacy concern
Roca et al (2009) aimed to validate an extended Technology Acceptance Model (TAM) for online trading, identifying key factors that influence its adoption Perceived usefulness stands out as a crucial determinant in the acceptance of information systems, especially as users increasingly depend on various platforms for daily activities This concept refers to users' beliefs about how effectively a system can improve their performance or provide specific benefits (Dhagarra et al., 2020; Kim et al.).
Research indicates that consumers are more likely to embrace new technologies when they recognize clear benefits (Pizzi & Scarpi, 2020; King & He, 2006) However, as reliance on technology grows, users may prioritize perceived usefulness over privacy concerns, often sharing personal information under the belief that the advantages outweigh potential risks While Baưiopedro et al (2022) highlight the negative impact of privacy issues on perceived usefulness, they do not provide significant evidence for the reverse effect This leads us to propose the following hypothesis.
H3: Perceived usefulness is negatively related to privacy concern.
2.3.4 The relationship between Privacy concern and Trust
Shoppable video represents a significant advancement in advertising, utilizing data-driven personalization to create real-time, interactive shopping experiences tailored to individual consumers (Williams et al., 2020) While this innovative approach enhances user engagement, it also raises concerns about consumer privacy due to the extensive customization based on personal information (Baek and Morimoto, 2012).
As e-commerce evolves and shoppable video becomes a popular purchasing method, privacy concerns pose a significant challenge for businesses striving to build customer trust Research by Hoffman et al highlights the importance of effectively addressing these privacy issues to foster trust among consumers When personal information is misused or harmful content is shared, customers may feel vulnerable, a sentiment that intensifies in systems that rely on personal data The way information is handled greatly impacts customer confidence, and emotional reactions to privacy matters can further heighten these concerns Ultimately, establishing trust depends on the responsible sharing of users' personal information with third parties.
Despite prevailing concerns about privacy, e-commerce website suppliers can potentially assert dominance over these apprehensions (Jaspers and Pearson, 2022).
Conversely, the study conducted by Rosenthal et al (2019) indicates an inverse relationship between privacy concern and trust in Facebook
Consumer perceptions are shaped by their interactions with various messages and the level of trust they have in those sources In light of these findings, this study proposes the following hypothesis.
H4: The level of privacy concern is negatively related to trust.
2.3.5 The role of information control in the relationship between Perceived Usefulness and Privacy Concerns, Privacy Concerns and Trust
This study defines information control as the set of measures that allow individuals to manage access to, and oversee the use and preservation of, information after it has been initially accessed.
Companies must gather extensive data to provide personalized services, but this practice raises significant privacy concerns that can alter user behavior Many users are unaware of the information exchange occurring through personalization algorithms, leading to the covert collection of personal data and jeopardizing user welfare To build consumer trust and alleviate privacy fears, it is crucial for businesses to focus on the responsible use and protection of personal information.
Research consistently highlights the crucial importance of information control in personalization, particularly regarding users' privacy concerns Liu et al (2005) found that comprehensive privacy policies—covering notifications, access, selection, and security—on e-commerce websites can enhance perceived trustworthiness, thereby influencing customer behavior Additionally, Taylor et al (2009) and Morimoto (2020) agree that perceived information control negatively affects users' anxiety, especially related to ad avoidance behavior.
Lambillote (2022) highlights that clear communication, especially for customers with lower awareness, can reduce the negative impacts of personalization on their sense of information control Supporting this, Lee & Cranage (2011) found that privacy measures on websites not only alleviate privacy concerns but also enhance trust and control among users However, the effectiveness of these business practices depends on the sensitivity of the data collected, with Lwin et al (2007) noting that concerns are lessened more significantly when handling low-sensitivity data compared to highly sensitive information.
Given the adverse effects of privacy concerns on the willingness to engage with personalized services (Morimoto, 2020), numerous studies have sought solutions to the personalization-privacy paradox Despite this, research focusing on customer information control as a potential remedy remains scarce This study advocates for empowering customers with control over their information as an innovative strategy to tackle privacy challenges in AI-personalized shoppable videos.
Implementing diverse information control techniques empowers users to manage their personal information effectively, which helps alleviate privacy concerns (Laufer & Wolfe, 1997) Research indicates that enhancing personal information control is crucial for reducing privacy issues and building trust in various settings (Milne & Boza, 1999; Parker & Price, 1994) Personalization relies on users voluntarily sharing their information with companies By ensuring user control over their data, organizations can access and use this information only with explicit consent, fostering trust and reducing privacy fears (Zhang et al., 2014) Therefore, the following hypotheses are proposed:
H5: Information Control moderates the relationship between Perceived Usefulness and Privacy Concerns, such that the relationship is positive when Information Controls are high.
H6: Information Control moderates the relationship between Privacy Concerns and Trust, such that the relationship is negative when Information Controls are high.
2.3.6 rhe relationship between Trust and Customer engagement
Summary
This chapter outlines a research framework examining consumers' purchase intentions through shoppable videos, supported by a literature review that informs the model development It proposes nine hypotheses, seven of which explore the connections between AI-driven personalization factors, perceived personalization, perceived usefulness, privacy concerns, trust, customer engagement, and purchase intention Additionally, two hypotheses investigate the moderating role of information control in the relationships between perceived usefulness and privacy concerns, as well as between privacy concerns and trust The following chapter will detail the methodology employed in this thesis.
The research process involves several key stages, starting with a thorough review of literature and relevant documents to create scales for each construction focus level These scales were then adjusted to fit the specific research context An initial questionnaire was developed in English and later translated into Vietnamese Preliminary research included in-depth interviews with experts to validate and refine the survey, ensuring the model's relevance and the accuracy of its scales.
This ongoing review utilized quantitative techniques to evaluate measurement models and test the underlying framework A two-stage approach was adopted to assess the estimated model, focusing on construct reliability through composite reliability and Cronbach’s alpha, as well as convergent reliability via indicator reliability and average variance extracted (AVE) Additionally, the discriminant validity of the measurement model was examined using the Heterotrait-Monotrait Ratio (HTMT) and cross-loadings.
The study examined the impact of common method bias (CMB) on the results by evaluating the structural model through several criteria, including the R2 of endogenous constructs for predictive ability, the goodness-of-fit of the research model, and VIF values to detect collinearity issues (Henseler et al., 2009) To analyze direct, mediating, and moderating effects as hypothesized, a bootstrapping technique with 5,000 samples was utilized The research process is illustrated in Figure 5.
RESEARCH METHODOLOGY
Qualitative Research
Qualitative research is a crucial first step in the research process, enabling researchers to delve deeply into existing literature related to their field, primarily through secondary sources This approach highlights the importance of acknowledging that initial assumptions and models may be flawed or misleading Therefore, rigorous testing and analysis are essential for refining and developing the researcher's initial ideas into a structured research framework.
Participating in in-depth conversations with industry experts provides essential insights Conducted through live interviews on online meeting platforms, these discussions create an interactive environment that promotes effective knowledge sharing.
Expert participation in discussions is crucial for guiding research teams to reevaluate their models and measurement scales By providing unbiased insights, experts suggest essential modifications that improve the model's relevance to real-world scenarios, thus increasing its applicability and sustainability Furthermore, these discussions play a key role in shaping future quantitative survey questionnaires, ensuring they effectively gather the necessary data.
In this study, we interviewed a research expert and a specialist in Marketing and AI, using predetermined questions designed to gather in-depth insights from both professionals.
The insights garnered from these interviews have enriched our understanding and provided valuable perspectives Subsequent sections will delve into the findings derived from these interviews.
The researcher organized interviews between January 26 and January 30, 2024, taking into account the availability and preferences of participants while accommodating their professional commitments To optimize schedules and reduce travel time, the interviews were held on the Google Meet platform.
An outline featuring open-ended questions was developed to facilitate discussions, with detailed notes recorded during expert interviews, ensuring participants' consent To maintain ethical standards and address privacy concerns in scientific research, all personal information of participants was kept confidential throughout the process.
The results are as follows:
About research concepts and relationships between concepts:
AI-driven personalization in customer experience consists of three key steps Initially, it is essential to collect information from substantial datasets Next, AI analyzes this data to gain insights into customers' demographics and interests, which aids in effective customer segmentation and the creation of tailored experiences.
This data is applied through apps, bots, or software to generate tailored suggestions based on user behavior The degree of personalization depends on the quantity of information gathered.
AI personalization brings products closer to customers, enhances predictability of shopping behavior but raises privacy concerns when personal data is excessively used.
Users' interaction with AI-driven platforms reveals the personalization when suggested content reflects their tasks.
Personalization in business enhances user experience by recommending tailored products While there are valid privacy concerns, experts suggest that these issues do not significantly deter users from engaging with platforms, indicating that privacy is generally maintained at an acceptable level.
Regarding the results of scale development
Experts unanimously agree on the quantity and content of the 45 observed variables derived from the scale's ten variables within the model, though certain considerations warrant attention.
Experts agree on the importance of using two separate scales to assess "AI-driven personalization" and "perceived personalization." However, they emphasize that users often struggle to recognize the significant difference between genuine personalization and what they perceive as personalization.
A thorough evaluation indicates that the observed variables possess content validity, though the range of questions is significant Furthermore, due to the emerging nature of this research area, modifications to the survey questions are essential for improving clarity and achieving the best results.
Appendix 2 details the structure of the in-depth interviews conducted with experts, offering valuable insights that serve as a foundation for identifying three critical issues.
The relationship between Perceived Usefulness and Privacy Concerns is now moderated by the Perceived Privacy variable, replacing the previous moderator of Information Control This change is based on expert insights suggesting that users' attitudes toward privacy and their willingness to share information are more significant factors in addressing privacy concerns than Information Control.
The moderating variable of Willingness to Disclose will be incorporated into the relationship between Perceived Privacy and Privacy Concerns.
The questionnaire covers a wide range of topics, which may make it difficult for participants to understand the specific issues being addressed To enhance clarity, it is essential to refine the questions to better fit the survey's context Additionally, the research team will provide participants with background information on the survey topic and explain any new terms relevant to the study before they begin answering the questions.
3.2.2 Propose an official research theory and model
3.2.2.1 Definition and overview of previous research of Perceived Privacy
Perceived privacy is an individual's subjective assessment of the privacy available in a given context, as defined by Yousafzai et al (2003) as the perception of control over personal information Dinev et al (2013) further describe it as a self-evaluative state where external agents have limited access to personal data Research indicates that privacy awareness is situational, influencing decision-making at specific times (Dinev et al., 2013) Factors such as the sensitivity of shared information, individual control over data, and trust in the managing organization shape this perception (Softer & Cohen, 2014) Moreover, initial perceptions can evolve over time (Chang et al., 2018).
Quantitative methods
The questionnaire underwent meticulous adjustments to ensure clarity before finalization and subsequent distribution The quantitative data were collected through
According to Hair et al (2011), Google Forms should ideally have a sample size of at least 100, with a minimum of five observable variables for each measurement variable The research model detailed in Chapter 2 includes 10 latent variables and 45 measured variables, requiring a minimum sample size of 225 observable variables To achieve reliable results from the PLS-SEM analysis, the survey was distributed to around 315 participants The research team employed various platforms, such as Facebook, Zalo, and Messenger, to connect with friends and acquaintances, ultimately gathering 350 survey responses, of which 315 were deemed valid for analysis.
The study utilizes the PLS-SEM method for data analysis, primarily because it offers significantly higher variance in the dependent variables compared to CB-SEM This makes PLS-SEM a preferred choice for researchers who prioritize the predictive power of their dependent variables (Hair et al.).
2017) Additionally, PLS-SEM offers relative advantages, as it does not necessitate a normally distributed dataset and is immune to multicollinearity issues (Hair et al.,
PLS-SEM is a powerful analytical tool that allows for the simultaneous analysis of models featuring multiple latent variables and parameters, especially those influenced by higher-order variables (Hair et al., 2017) This method facilitates the concurrent estimation of both the measurement and structural models, helping to reduce potential biases and inaccuracies in the estimation process (Hair et al., 2018).
SmartPLS 4 leverages the PLS algorithm to evaluate the precision of scales, R² values, and f² values, while the bootstrapping method is applied to determine the significance of the path coefficients.
Data analysis process
To ensure accuracy in data analysis, it is crucial to summarize the information collected through Google Forms and perform meticulous checks to eliminate any errors The final dataset consists of 315 samples, which undergo descriptive statistical analysis using SPSS 20.0 to assess the characteristics of the research sample Following this, the author evaluates the research model utilizing two different models: (1) Effect Indicator (Reflective).
Measurement Models) and (2) Composite Indicator (Fonnalive Measurement Models) as proposed by Henseler et al (2009).
Evaluating the measurement model requires a thorough assessment of both reliability and validity Reliability is measured using Cronbach’s Alpha and the Composite Reliability coefficient (CR) Meanwhile, validity, which includes both convergent and discriminant validity, is evaluated through the Cross-Loading coefficient, Average Variance Extracted (AVE), and the correlation matrix of the research variables.
Cronbach (1951) introduced a formula to evaluate internal consistency reliability by examining the correlation among observable variables This method, known as Cronbach's alpha, assumes that all observed variables possess similar reliability levels However, it is important to note that the Cronbach’s alpha coefficient can be sensitive to the number of observable variables within each scale, often leading to an underestimation of internal consistency reliability.
Where: a: Cronbach's Alpha Coefficient k: Number of items in the scale
2 : The variance for all items on the scale ơi : The variance of individual item i2
Cronbach's Alpha coefficient ranges from 0 to 1, with values above 0.70 indicating acceptable reliability; ideally, values closer to 1 reflect greater reliability Understanding these ranges is essential for interpreting reliability analysis effectively.
Cronbach’s Alpha value range Interpretation a >= 0.90 Excellent
Composite reliability (CR) takes into various outer loadings of the observable variables and is determined by the formula (Hair et al., 2018).
In the context of latent variable analysis, li represents the fully standardized loading of the observable variable i, while ei signifies the measurement error associated with that variable The variance of this measurement error, denoted as var(ei), is calculated using the formula 1 - 12, provided that the Composite Reliability (CR) is 0.6 or greater.
The evaluation of average variance extracted (AVE) for each construct involves examining the outer loadings of the indicators, constituting the method for assessing convergent validity.
To achieve adequate convergent validity, the construct score must account for at least 50% of the variable's variance, which is determined by the square of the outer loadings exceeding 0.708 (Henseler et al., 2015) Average Variance Extracted (AVE) serves as a key indicator of convergence by measuring the variance extracted from all items associated with a single construct (Hair et al., 2010) An acceptable level of convergence is indicated by an AVE greater than 0.50, signifying that the construct score captures the majority of the indicator variance (Hair et al., 2017).
Formative measurement models necessitate a unique method for evaluating convergent validity, unlike reflective measurement models that rely on internal consistency reliability To effectively establish convergent validity for constructs measured formatively, it is crucial to include additional well-measured variables within the nomological network of each formative construct in the survey The evaluation of these constructs and their convergent validity should take into account the statistical significance, magnitude, and collinearity of the indicator weights (Hair et al., 2017).
Discriminant validity is crucial for ensuring that distinct concepts are adequately differentiated, indicating that the indicators used are not unidimensional (Henseler et al., 2009) The Fornell-Larcker criterion and cross-loadings are two key measures for assessing discriminant validity in PLS path modeling According to the Fornell-Larcker rule (Fornell and Larcker, 1981), a latent variable must explain more variance than any other latent variable linked to its indicators, with the average variance extracted (AVE) of each latent variable exceeding its highest squared correlation with other latent variables Additionally, the heterotrait-monotrait ratio (HTMT) is evaluated by comparing the average of heterotrait-heteromethod correlations, further supporting the assessment of discriminant validity.
To ensure discriminant validity, it is essential that the Heterotrait-Monotrait ratio (HTMT) remains below the threshold of 0.9, as outlined by Henseler et al (2015) This criterion involves comparing correlations between indicators measuring different constructs with the average of monotrait-heteromethod correlations, which assess relationships within the same construct.
Assessing Structural Model
To evaluate the relationship between research variables, it is essential for the researcher to follow specific steps: first, analyze the multicollinearity issues within the structural model; second, determine the magnitude and significance of the relationships present in the structural model Í2.
(3) assessment of the impact factor; (4) assess the coefficient of determination R2.
Multicollinearity occurs when independent variables in a model exhibit a strong correlation, which can bias the results and alter their relationship with the dependent variable This issue typically arises when linear correlations exist among multiple independent variables, leading to problems such as limiting the R-squared value and distorting the regression coefficients' signs.
To evaluate multicollinearity, the author utilizes the variance inflation factor (VIF), establishing acceptable thresholds of VIF < 5 and tolerance (TOL) > 0.20 (Hair et al., 2017) In PLS-SEM analysis, multicollinearity is indicated by a VIF value greater than 5.00 or a tolerance value less than 0.20 (Hair et al., 2017).
PLS-SEM is advantageous as it does not assume normal data distribution, which means parametric tests for assessing the statistical significance of outer weights, outer loadings, and path coefficients are not applicable Instead, PLS-SEM utilizes bootstrapping to calculate standard errors, enabling the evaluation of significance levels based on statistically significant coefficients.
Hair et al (2018) recommend increasing the sample size to around 5,000 using bootstrapping techniques This approach allows for the calculation of the standard error, which is essential for determining the experimental t-value and p-value across all path structures in the structural model A t-value exceeding 1.96 signifies statistical significance at the 5% level.
3.5.3 Assessing Coefficient of determination (RA2)
The coefficient of determination, R², is the primary metric used to evaluate endogenous latent variables It is calculated as the squared correlation between the predicted and actual values of a specific dependent variable R² values range from 0 to 1, with higher values indicating more accurate forecasts In PLS path models, R² values of 0.75, 0.50, and 0.25 are categorized as significant, moderate, and weak, respectively (Hcnsclcr et al., 2009).
For each impact in the path model, one can evaluate the effect size using Cohen’s
(1988) /2 The increase in R2, indicating the proportion of the endogenous latent variable's variance that remains unexplained, is utilized to calculate the effect size f2
/2 values of 0.02, 0.15, and 0.35 can be interpreted as indicators of whether a predictor latent variable has a weak, medium, or large effect at the structural level (Cohen, 1988).
Measurement Scale
The research papers were evaluated using a five-point Likert scale, as established in prior literature, and translated into Vietnamese to align with the current research context Participants assessed their level of agreement with various factors, providing ratings on a scale from 1 (Strongly disagree) to 5 (Strongly agree).
Factor Items Items (Vietnamese) Sources
All: The Al-driven personalized shoppable videos on this platform were recommended based on my past digital search behaviors.
All: Nội dung cá nhân hóa dựa trên AI trôn nền tảng có shoppable video này được dựa trên hành vi tim kiêm trong quá khử của tôi.
AI2: The Al-driven personalized shoppable videos on this platform were recommended based on my past purchasing behaviors in digital channels.
AI2 là nền tảng video có thể mua sắm, cung cấp nội dung cá nhân hóa dựa trên trí tuệ nhân tạo, được xây dựng dựa trên hành vi mua hàng từ các kênh thương mại điện tử.
AI3: The AI-driven personalized shoppable videos on this platform with were recommended based on my past behaviors on this platform
AI3 là nền tảng shoppable video sử dụng trí tuệ nhân tạo để cá nhân hóa nội dung, dựa trên hành vi trước đây của người dùng.
AI4: The Al-driven personalized shoppable videos on this platform with were recommended based on my past behaviors on other digital platforms
AI4 sử dụng công nghệ AI để cá nhân hóa nội dung, cho phép tạo ra video có thể mua sắm dựa trên hành vi trước đây của người dùng trên các nền tảng kỹ thuật số khác.
Personalization pp 1: This Al-driven personalized shoppable
PPI: Shoppable video được cá nhân hóa này đưa
Tran et al.(2020) video makes purchase recommendations that match my needs. ra những đề xuất mua sắm phù họp với nhu cầu của tôi.
PP2: 1 think that this AI-driven personalized shoppable video enables me to order products that are tailor-made for me.
PP2: loi nghĩ rằng shoppable video được cá nhân hóa này giúp tôi đặt hàng những sán phâm dành riêng cho minh.
PP3: Overall, this Al-driven personalized shoppable video is tailored to my situation.
PP3: Nhìn chung, shoppable video được cá nhân hóa này phù hợp với hoàn cánh của tôi.
PP4: This Al-driven personalized shoppable video makes me feel that
PP4: Shoppable video được cá nhân hóa này khiến tôi cảm thấy mình là khách hàng đặc biệt.
PP5: I believe that this AI-drivcn personalized shoppable video is customized to my needs.
PP5: Tôi tin răng shoppablc video được cá nhân hóa này được tùy chinh phù hợp với nhu cầu cúa tôi.
PU1: The platform with AI-driven personalized shoppable video improves my performance when searching for and purchasing goods.
PU1: Nền táng có shoppable video cải thiện hiệu suất cúa tôi khi tìm kiếm và mua hàng hóa.
PU2: The platform with AI-driven personalized shoppable video increases my productivity when searching for and purchasing goods.
PU2: Nền táng có shoppable video giúp tăng năng suất cúa tôi khi tìm kiểm và mua hàng hóa.
PU3: The platform with Al-driven personalized shoppable video makes it easier to search for and purchase goods.
PƯ3: Nên tảng cỏ shoppable video giúp tìm kiếm và mua hàng dề dàng hơn.
PU4: The platform with Al-driven personalized shoppable video enhances my effectiveness in goods searching and purchasing
PU4: Nền táng có shoppable video giúp nâng cao hiệu quả của tôi trong việc tìm kiếm và mua hàng.
PU5: The platform with AI-driven personalized shoppable video is useful for searching for and buying goods.
PU5: Nen táng có shoppable video rất hừu ích cho việc tỉm kiếm và mua hàng hóa.
PCI: 1 feel uncomfortable when information is shared without my permission.
PCI: Khi tôi nhận được nội dung được cá nhân hóa trên nền tảng có shoppable video, tôi cảm thây không thoải mái khi
Baek and Morimoto (2012) thông tin bị chia sẻ mà không có sự cho phép của tôi.
PC2: I am concerned about misuse of my personal information.
Khi nhận được nội dung cá nhân hóa trên nền tảng video có thể mua sắm, tôi cảm thấy lo ngại về việc thông tin cá nhân của mình có thể bị lạm dụng.
PC3: It bothers me to receive too much personalized material.
Khi tôi nhận được nội dung cá nhân hóa trên nền tảng có video có thể mua sắm, tôi cảm thấy phiền khi phải tiếp nhận quá nhiều nội dung được tùy chỉnh.
PC4: I believe that my personal information is often misused.
Khi nhận được nội dung cá nhân hóa qua video có thể mua sắm, tôi cảm thấy lo ngại về việc thông tin cá nhân của mình có thể bị lạm dụng.
PC5: I think companies share my information without permission.
Khi nhận nội dung cá nhân hóa từ video có thể mua sắm, tôi cảm thấy doanh nghiệp đang chia sẻ thông tin của tôi mà không có sự cho phép.
PR 1:1 think I would have enough privacy when the privacy information is collected and used.
PR1: Tôi nghi tôi SC có đù quyền riêng tư khi thông tin về quyền riêng tư được thu thập và sử dụng.
PR2: I think I would be satisfied with the privacy
1 have when the privacy information is collected and used.
PR2: Tôi nghi tôi sẽ hài lòng với quyền riêng tư mà tôi có khi thông tin quyền riêng tư được thu thập và sử dụng.
PR3: I think my privacy would be protected when the privacy information is collected and used.
PR3: Tôi nghĩ quyền riêng tư của tôi sẽ được bảo vệ khi thông tin về quyền riêng tư được thu thập và sử dụng.
WT1: 1 am likely to disclose my personal information.
WT1: Tôi có khả năng tiết lộ thông tin cá nhân cùa mình.
WT2: 1 am willing to disclose my personal information.
WT2: Tôi sằn sàng tiết lộ thông tin cá nhân cùa mình.
WT3: Disclosing my personal information on AI-drivcn personalized platforms is unlikely for me (reverse coded).
WT3: Tôi không chắc chăn trong việc tiêt lộ thông tin cá nhân trôn nền táng được cá nhân hoá bời AI.
Trust TRI: The platforms with
AI-driven personalized shoppable video would be trustworthy in handling my information
TRI: Các nền tảng có shoppable video sè đáng tin cậy trong việc xứ lý thông tin của lôi.
TR2: The platforms with AI-driven personalized shoppable video would tell the truth and fulfill promises related to the information provided by me.
TR2: Nen tảng có shoppable video sè nói sự thật và thực hiện các cam kết liên quan đến thông tin mà tôi cung cấp.
TR3: I trust that the platform with AI- driven personalized shoppable video would keep my best interest in mind when dealing with my personal information.
TR3: Tôi tin tưÓTig rằng nền tảng có shoppable video sẽ luôn đặt lợi ích cùa tôi lên hàng đâu khi xú lý thông tin cá nhân.
TR4: The platforms with AI-driven personalized shoppable video are in general predictable and consistent regarding the usage of information.
TR4: Việc sữ dụng thông tin cá nhân của các nên tảng có shoppable video thường dễ dự đoán và nhất quán.
TR5: The platforms with AI-driven personalized shoppable video are in general predictable and
TR5: Các nen táng có shoppable video luôn trung thực với khách hàng khi sủ dụng thông tin mà consistent regarding the usage of information. họ cung cấp.
IC1: I was informed about the personal information that the platform with Al-driven personalizes shoppable video would collect about me.
IC1: Tôi đà được thông báo về nhừng thông tin cá nhân nền táng có shoppable video sè thu thập.
IC2: The platform with AI-driven personalized shoppablc video explained why personal information was being collected.
IC2: Nen táng có shoppable video giải thích lại sao thông tin cá nhân được thu thập.
IC3: The platform with AI-driven personalized shoppable video explained how personal information collected about me would be used.
IC3: Nền tảng có shoppable video giải thích cách thông tin cá nhân của tôi sè được sử dụng.
IC4: The platform with AI-drivcn personalized shoppable video gave me a clear choice before using personal information about me.
IC4: Nen tảng có shoppablc video cho lôi lựa chọn rõ ràng trước khi sử dụng thông tin của tôi.
CE1: 1 spend more time on platforms with
CE1: Tôi dành nhiều thời gian hơn cho các nền táng
Al-driven personalized shoppable video.
(2020) CE2: I would become a follower of a page with
AI-driven personalized shoppable video.
CE2: Tôi sè theo dõi và trờ thành người theo dõi của một trang có shoppable video.
CE3: I would be likely to try and keep track of the activities of a seller that has a page with
AI-driven personalized shoppablc video.
CE3: Tôi sẽ cố gắng theo dõi các hoạt động cúa người bán cỏ một trang với shoppable video.
CE4: 1 am likely to revisit the seller's page to watch their AI-driven personalized shoppable video in the near future.
CE4: Tôi sẽ xem lại trang của người bán đẻ xem shoppable video của họ trong tương lai gần.
CE5: I am likely to recommend sellers that have a page with
AI-driven personalized shoppable video to my friends.
CE5: Tôi se giới thiệu những người bán có một trang với shoppable video cho bạn bè cúa lôi.
CE6: I encourage friends and relatives to do business with a seller that has a page with
AI-driven personalized shoppable video.
CE6: Tôi khuyến khích bạn bè và người thân làm kinh doanh với người bán có một trang có shoppable video.
CE7: In the near future, I will definitely buy products from a seller that has a page with AI-drivcn personalized shoppable video.
CE7: Trong tương lai gần, tôi chac chan sẽ mua sán phâm từ một người bán có một trang với shoppable video.
CE8: I consider a seller who has a page with Al-driven personalized shoppable video to be my first choice when buying a product.
CE8: Tôi coi một người bán có một trang với shoppable video là lựa chọn đầu tiên của lôi khi mua một sản phâm nào đó.
PI1: In the future, I intend to continue shopping online based on AI-driven personalization shoppable video.
PI1: Trong tương lai, tôi dự định tiếp tục mua sắm thông qua shoppable video đà được AI cá nhân hóa.
PI2: My general intention to buy online on AI-driven personalization shoppable video is very high.
PI2: Ỳ định khi mua sắm thông qua shoppablc video đã được AI cá nhân hóa cùa tôi rất cao.
PI3: 1 will shop online in the future based on AI-driven personalization shoppable video.
PI3: Tôi sè tiếp tục mưa sắm trực tuyến thông qua shoppable video được AI cá nhân hóa trong tương lai
Sample characteristics
After an exhaustive survey, 301 legitimate responses were utilized for additional investigation Table 4 provides additional information about the profiles and purchasing habits of respondents.
Name of variables Frequency Percentage
Regarding gender distribution, women constituted the majority of survey respondents, accounting for 59.1%, surpassing their male counterparts at 40.9%.
The survey revealed a diverse age demographic among participants, with a significant majority of 85% (n = 256) aged between 18 and 24 Participants aged 25 to 34 represented 12%, while those in the 35 to 44 age group accounted for 3% Importantly, there were no participants older than 44 years in the survey.
Educational backgrounds varied, with the majority being high school graduates, making up 83.4% of respondents The remaining participants held Associate's degrees (3.7%), Bachelor's degrees (8%), and Master's degrees or higher (5%).
Students formed the largest segment among participants, representing 83.4% (n 251) Officials/employees constituted 10.6% (n = 32), and traders comprised 6% (n 18) of the respondents.
Over half of the participants (n = 171), representing 56.8%, reported a monthly income of less than 5 million, highlighting the student-centric demographic Additionally, 25.9% (n = 78) fell within the 5-10 million income bracket, while 6.3% (n = 19) earned between 10-15 million Those earning 15-20 million comprised 5% (n = 15), and another 5% (n = 15) reported incomes exceeding 20 million.
TikTok emerged as the predominant platform for shoppable videos, with 68.8% (n 207) of respondents utilizing it, significantly surpassing other platforms like Facebook (15%), Instagram (12%), and YouTube (4.3%).
A recent study on platform engagement with shoppable videos revealed that 33.2% of participants had over one year of experience, while 20.3% had 9-12 months of experience Additionally, 27.9% reported having 3-6 months of experience, and 18.6% indicated they had less than three months of exposure to shoppable videos.
Chapter 3 offers an extensive examination of the research process, encompassing critical components such as defining the research problem and objectives, and structuring a systematic theoretical framework to underpin a robust research model This model serves as the cornerstone for refining the initial scale, engaging in group discussions, and conducting preliminary research, including in-depth interviews with experts, to refine the research model and adjust the scale accordingly Subsequent steps involve refining the scale and research model, and constructing a theoretical framework to accommodate these modifications The process continues with the selection of the survey sample, data collection and processing through descriptive statistical analysis, hypothesis testing, scale evaluation, and culminates in the formulation of conclusions and proposal of implications.
The scale, derived from a synthesis of prior studies and complemented by qualitative surveys, effectively addresses the research questions The resultant model comprises
10 research concepts, encompassing a total of 45 inquiries.
The author utilizes SPSS 20.0 software for data processing and statistical sample representation, while the measurement and PLS-SEM structural models are thoroughly assessed using SmartPLS 4 software This methodology guarantees a comprehensive analysis of the research data, enhancing the reliability and validity of the study's results.