The research aims to: i identify factors influencing the intention to continue using online music platforms, ii measure the extent of influence of these factors on the intention to use m
OVERVIEW OF THE RESEARCH TOPIC
Reason for choosing the topic
Since the dawn of humanity, music has been an integral part of our existence, beginning with the lullabies sung by mothers and grandmothers It transcends individual differences and plays a vital role in our lives, providing strength and fostering connections among people from diverse cultural and linguistic backgrounds (Anthony Storr, 1992) Furthermore, music promotes social integration, enhances individual consciousness, and addresses issues of social exclusion (Welch et al., 2014; Elvers et al., 2017) As Robin Maconie notes, music's influence is profound and far-reaching, underscoring its importance in uniting communities and enriching our lives.
Music transcends language and communicates across cultural and national boundaries It has the power to evoke a wide range of human emotions, although it typically does not elicit the more negative feelings such as terror, fear, or anger, as noted by Charles Darwin in "The Descent of Man."
Music evokes feelings of tenderness and love, fostering a sense of devotion It significantly influences children's language and cognitive skills (Loui et al., 2019) Additionally, music serves as an effective motivational tool during exercise, enhancing enjoyment and extending workout duration A study by Nikol et al (2018) demonstrated that music can increase the time to exhaustion during physical activity.
Listening to synchronized music significantly enhances the experience, with participants reporting a duration of engagement that is two-thirds longer compared to when music is absent As music becomes an essential part of daily life, it is consumed through various mediums such as radio, vinyl records, and digital platforms The rapid advancement of technology, particularly big data and the Internet of Things (IoT), has transformed e-commerce and influenced consumer behavior, leading to a shift in music consumption preferences The International Federation of the Phonographic Industry reported a decline in physical music sales from $23.3 billion in 2001 to $5.0 billion in 2021, while music streaming services surged from $0.1 billion in 2005 to $16.9 billion in 2021, now representing 65% of total music revenue The rise of online music platforms has empowered consumers to access and enjoy music on their terms With intense competition in the streaming market, platforms like Spotify, YouTube Music, and Apple Music have gained popularity, contributing to a global subscriber base of 616.2 million by the second quarter of 2022, up from less than 523 million the previous year.
In 2021, the online music streaming market saw significant growth, with Spotify leading the way at 180 million paid subscribers, followed by Apple Music with 78.6 million and Tencent Music with 76.2 million (Statista) The competition among platforms is intense, as they not only offer music streaming but also value-added services that enhance user experience (Hagen & Luders, 2017) These platforms foster virtual communities where users can share their thoughts and interact with others who have similar musical tastes (Nguyen et al., 2014; Weinberger & Bouhnik, 2021) Additionally, at the end of the year, users receive personalized reports on their listening habits, showcasing the platforms' ability to analyze customer behavior (Kreitz & Niemela, 2010; Weinberger & Bouhnik, 2021).
Personalization, defined as the integration of technology and customer data to customize e-commerce interactions, significantly influences customer attitudes by impacting their information processing and decision-making (Fan & Poole, 2006; Tam & Ho, 2006) This approach enables service providers to offer products and services that align with individual customer preferences (Tam & Ho, 2005; Xu et al., 2014) A 2022 global survey revealed that 62% of consumers would lose loyalty to a brand if their online shopping experience lacked personalization, highlighting the critical role of tailored interactions in maintaining customer loyalty.
According to Statista, the personalization of products significantly influences users' behavioral decisions to continue utilizing online music streaming platforms The authors of the study noted a lack of comprehensive research examining the mediating role of personalization on the relationship between various factors and the intention to persist with these platforms Consequently, they incorporated personalization as an intermediary variable to elucidate its impact on the relationship between other influencing factors and users' intentions to continue using online music streaming services.
Recognizing the significant impact of online music streaming platforms on daily life and the intensifying competition in this sector, the research article titled "Factors Influencing Continuous Intention to Use Online Music Streaming Platforms: The Mediating Role of Personalization" was developed This study aims to explore the relationship between user personalization and the intention to continue using these platforms The authors aspire that the findings will assist businesses in devising effective strategies to attract and retain customers in the rapidly expanding online music streaming market.
Research objectives
This study investigates the main factors affecting young people's intentions to use Online Music Streaming Platforms, emphasizing the mediating role of Personalization The findings aim to assist businesses, developers, and researchers in gaining a deeper understanding of user preferences and needs.
Specifically, the research aims to achieve the following goals:
First, we want to find out the factors that influence the intention to continue using online music platforms.
Second, measure the influence of factors affecting the intention to use music platforms.
This study examines how personalization variables influence users' intentions to continue using a service, highlighting the mediating role of personalization in the relationship between other relevant factors and continued usage intentions.
Fourth, based on the level of influence of the factors, give businesses a way to promptly adapt to the fierce competition in the field of online music platforms.
Scope and objects
The research focuses on the urban area of Ho Chi Minh City, targeting Generation Y (born 1981-1996) and Generation Z (born 1997-2012) The study is conducted from December 2, 2023, to February 17, 2024.
Research methods
The research was conducted in two phases.
The secondary data research team conducted an extensive review of relevant scientific reports to establish a theoretical foundation for developing models and hypotheses This comprehensive analysis provided essential insights that enabled the team to effectively explore and refine their research model.
In the second phase, the team conducted quantitative research through survey questionnaires The implementation process lasts from January 22, 2024, to January 24,
2024, in the form of sending survey forms to subjects of Generation z and Generation Y who are using online music platforms in the region Ho Chi Minh City.
The team successfully identified 329 valid answer sheets from a total of 329 recorded submissions, which provided essential input data for their analysis This data serves as the foundation for the group to investigate and address the target problems outlined at the beginning of the article.
Data analysis methods utilizing SmartPLS 4.0 software encompass various techniques such as descriptive statistics, Cronbach’s Alpha scale testing, and discriminant value testing The analysis also includes SEM model testing, mediation analysis, cross-loading factor evaluation, and the application of the Former-Laker criteria Additionally, it involves calculating the HTMT index, as well as regression and mediation models, along with multivariate regression.
Contributions
The study titled "Factors Influencing Continuous Intention to Use Online Music Streaming Platforms: The Mediating Role of Personalization" seeks to enhance the UTAUT2 model by incorporating personalization as a mediating factor, thereby elucidating the relationship between various influencing factors and the intention to persist in using online music streaming services Additionally, it aims to deepen the understanding of personalization's critical role in fostering continued engagement with these platforms This research also contributes to consumer behavior studies by offering insights into the factors that affect users' intentions to continue utilizing online services, particularly within the music industry.
This research offers valuable insights for online music streaming service providers, enabling them to understand user needs and preferences By leveraging this information, providers can develop effective strategies to attract and retain customers Additionally, the findings enhance the user experience, guiding consumers in selecting the most suitable music streaming platform for their requirements.
Structure of the Report
Chapter 1: Overview of the research topic
Chapter 2: Theoretical framework and research model Chapter 3: Research models
LITERATURE REVIEW
Theoretical Framework
The Unified Theory of Acceptance and Use of Technology - UTAUT2, Venkatesh et al., 20J2.
The Unified Theory of Acceptance and Use of Technology (UTAUT) (Venkatesh et al.,
The UTAUT model, established in 2003, has served as a foundational framework for assessing technology acceptance and usage behaviors, primarily from an organizational perspective However, it overlooks the cognitive and psychological factors that impact technology adoption (Chen & Holsapple, 2013) To address these limitations, Venkatesh et al (2012) revised the model, introducing UTAUT2, which incorporates three additional factors: Hedonic Motivation, Price Value, and Habit This updated framework not only emphasizes technology usage but also accounts for the individual needs and motivations of consumers.
Figure Ì.Unified theory of acceptance and use of technology (UTAUT2) Model
The UTAUT2 model has been extensively studied by scholars to analyze usage behavior and intention, demonstrating strong explanatory power (Baabdullah, 2018) Venkatesh et al (2012) noted significant enhancements in the model, with the variance in behavioral intentions rising from 56% to 74%, and technology usage increasing from 40% to 52% This theory effectively clarifies how individuals and organizations accept and utilize technology across various consumption contexts (Escobar-Rodriguez & Carvajal-Trujillo, 2013).
Utilizing online music streaming platforms involves technological aspects and various consumer contexts This research employs the UTAUT2 model as its analytical framework, emphasizing the consumer perspective, which is essential for the success of any platform in the competitive market.
Related research articles
2.2.1 Studied by Cheng et al in 2020: Role of Personalization in Continuous Use
Intention of Mobile News Apps in India: Extending the UTAUT2 Model
The study by Yanxia Cheng et al (2020) empirically tests the extended Unified Theory of Acceptance and Use of Technology 2 (UTAUT2) by incorporating "personalization" as both an antecedent and a moderating factor The research aims to identify the key determinants influencing users' continuous intention to utilize mobile news applications.
This study involved a comprehensive analysis of a sample survey conducted with 309 respondents who had prior experience using a news app Utilizing quantitative methods, including explanatory and confirmatory factor analysis, structural equation modeling, and the Hayes process, the research aimed to identify the moderating effects between various variables.
The study reveals that performance expectancy (PE) is the most significant factor influencing continuous use intention, followed by habit (H), hedonic motivation (HM), and favorable conditions (FC) Additionally, personalization plays a crucial moderating role between the UTAUT2 constructs and continuous use intention, particularly affecting performance expectations and habits This highlights the importance of PE, H, HM, and personalization in driving user engagement.
FC as key factors that trigger users’ intention to continuously use news applications and provides an integrated framework to evaluate the moderating effect of personalization on technology acceptance.
I ont FH1OUX I st mention raci I Hating nruilhtirH
Figure 2 Research model ofCheng et al (2020)
2.2.2 Research by Ifekanandu in 2023: Influence of Artificial Intelligence (AI) on
Customer Experience and Loyalty: Mediating Role of Personalization
This study seeks to evaluate the impact of artificial intelligence on customer experience and customer loyalty; as well as the mediating effect of personalization on this relationship.
A quantitative online survey was conducted and 636 responses were collected The collected data was analyzed using IBM's AMOS for SEM and several findings were made
Artificial intelligence significantly enhances customer experience and loyalty, with personalization serving as a key mediator in this relationship.
2.2.3 Choi et al researched in 2011: The Influence of Social Presence on Customer Intention to Reuse Online Recommender Systems: The Roles of Personalization and Product Type
This study investigates the causal relationship between social presence and reuse intention, examining how trust mediates this connection By offering personalized recommendations tailored to users' preferences, the research assesses the impact of social presence across two distinct product contexts: hedonic and utilitarian.
The experiment involved 368 college students from two private universities in Korea, all participants had experience shopping on the Web The author uses SPSS and Amos to analyze the data.
Research indicates that a stronger social presence enhances users' intention to reuse recommendations and builds trust in the recommendation system Additionally, the impact of social presence on the intention to reuse is more pronounced for hedonic products compared to utilitarian ones.
Figure 4 Research model of Choi et al (2011)
2.2.4 Researched by Leninkumar in 2017: The Relationship between Customer
Satisfaction and Customer Trust on Customer Loyalty
Due to fierce competition among banks, many researchers have focused on the relationship between customer satisfaction and customer loyalty.
A study was launched at Sri Lankan commercial banks, with 300 survey forms issued to customers, of which 210 were collected The researchers used SPSS and SmartPLS to analyze the data.
Research indicates a strong positive relationship among customer trust, loyalty, and satisfaction Specifically, customer satisfaction plays a crucial role in fostering customer loyalty, while also serving as a key factor in building customer trust Additionally, the findings reveal that customer satisfaction indirectly enhances loyalty through the mediation of customer trust.
Figure 5 Research model of Leninkumar (2017)
2.2.5 Al-Hashem Adel Odeh, Abu Orabi Tareq studied in 2021: Electronic customer satisfaction using electronic personalization and social media marketing model
Electronic personalization and social media are revolutionary marketing strategies designed to enhance customer satisfaction by utilizing modern tools for product delivery and services tailored to individual preferences.
The research employed a quantitative methodology by distributing questionnaires to a convenience sample of 622 guests at five-star hotels in Jordan Out of these, 573 questionnaires were deemed valid and subsequently analyzed using AMOS software.
The study reveals that social media marketing significantly enhances customer satisfaction and e-personalization, with e-personalization acting as a mediator in this relationship Therefore, 5-star hotel management should implement electronic marketing strategies and models to boost customer satisfaction amid global competition.
Figure 6 Research model ofAl-Hashem Adel Odeh, Abu Orabi Tareq studied in 202/
2.2.6 Studied by Tarhini et al (2019): An analysis of the factors affecting mobile commerce adoption in developing countries
This study explores the factors influencing the adoption of mobile commerce among consumers in Oman, a developing country It examines both the barriers and facilitators that impact mobile commerce activities in this specific context.
A questionnaire was distributed to a convenience sample of 530 e-commerce users in Oman, yielding 432 responses Following the research process, the final sample size was determined to be 430 after screening The authors utilized AMOS for data analysis and validation.
The structural equation model reveals that consumers' behavioral intention (BI) to adopt e-commerce is significantly influenced by factors such as information quality, habits, performance expectations, beliefs, enjoyment motives, service quality, price value, and favorable conditions However, effort expectancy, social influence, self-efficacy, and system quality were found to have no significant impact on BI This research provides valuable insights for local e-commerce businesses to formulate effective organizational strategies, particularly in marketing and mobile application development, to enhance customer attraction.
2.2.7 Barata and Coelho researched in 2021: Music streaming services: understanding the drivers of customer purchase and intention to recommend
This study investigates the key factors affecting music consumption on streaming platforms, with a particular focus on the intention to adopt premium (paid) versions of these services It also explores the likelihood of recommending music online to others.
This study utilized structural equation modeling (SEM) to analyze data from 324 music streaming service users, complemented by in-depth semi-structured interviews These interviews provided valuable insights into the profiles, behaviors, and motivations of new music consumers.
Research concepts
The rapid advancement of modern Information Technology, particularly the Internet of Things (IoT), has led to a significant surge in digital platforms Social networking sites and streaming services are thriving, solidifying their role in the global digital economy These platforms are reshaping various markets by introducing innovative methods of working, creating, interacting, and capturing value.
Streaming platforms like Netflix, Audible, Spotify, and Amazon Prime have revolutionized online interaction by providing essential technical infrastructure and user-friendly interfaces These platforms grant users access to extensive catalogs of music, movies, television shows, audiobooks, and podcasts, fostering opportunities for creators and businesses to distribute their intellectual property rights effectively.
This study focuses on the role of online music streaming platforms, which enable music owners to stream their songs and earn revenue based on listener engagement These platforms offer consumers a vast digital music archive accessible via the Internet, with features such as offline downloads and the ability to upload personal music collections to the cloud, often referred to as a "music store." Additionally, users can create and share playlists, discover new music, and curate personal collections Notable platforms in this space include Spotify, Apple Music, Amazon Music, Google Play Music, and Pandora The predominant business model consists of two service types: a free, ad-supported version and a premium subscription that provides unlimited access, offline listening, and mobile app benefits.
2011 calls this business model the 'two-tier freemium model'.
Music recordings play a vital role in daily life, enhancing activities such as commuting, work, and leisure (Fuentes et al., 2019) The rise of online music streaming platforms has made them an integral part of people's routines, leading to intense competition in the digital music industry To secure a strong position, these platforms must continuously adapt to the diverse needs of users, striving to become the preferred choice among numerous available options.
Proposed research model and hypothesis
The authors took the basis from a number of related research articles in section 2.2 along with some of the following research models:
Research on technology acceptance and use models - UTAUT
From the above research models, the authors propose the following research model:
Artificial Intelligence (AI) refers to machines that can mimic human intelligence to perform tasks typically requiring human cognition (Patrick, 2020) These machines are capable of understanding their environment and optimizing their actions to achieve specific goals (Wirth, 2018) Research by Prentice et al (2020) indicates that AI significantly impacts Employee Service Quality, which in turn affects Customer Satisfaction and Loyalty Additionally, Chen et al (2023) found that AI chatbots enhance Service Quality and Customer Loyalty by fostering Cognitive Trust, Perceived Value, Satisfaction, and Emotional Trust Thus, AI is crucial in delivering value to both users and businesses, significantly affecting the likelihood of continued product usage.
Artificial Intelligence (AI) plays a crucial role in understanding customer preferences and behaviors, including their desires, preferred platforms, and satisfaction levels (Davenport & Kirby, 2016; Boyd & Holton, 2018; Duan et al., 2019) By analyzing emotions and locations, AI can fulfill customer wishes that are difficult for humans to achieve (Inversini & Masiero, 2014; Davenport et al., 2020) The personalization of products through AI not only enhances customer experiences but also opens new opportunities for businesses (Kumar et al., 2019) Key applications of AI in personalization include automated ad copy generation, recommender systems, and targeted marketing strategies aimed at improving customer satisfaction and generating leads (Deng et al., 2019; Zanker et al., 2019; Syam & Sharma, 2018) Among these, recommendation systems stand out as a prominent example of AI-driven personalization (Zanker et al., 2019) Furthermore, the impact of AI on customer loyalty is significantly influenced by its ability to personalize offerings (Ifekanandu et al., 2023), suggesting that personalized products may enhance customers' intentions to continue using them.
Hypothesis 1 (Hla) Artificial Intelligence significantly influences the Continuous Use Intention on online music streaming platforms.
Hypothesis 1 (Hlb) Personalization plays a mediating role between Artificial Intelligence and Continuous Use Intention on online music streaming platforms.
Online trust refers to the confidence that individuals have in risky online situations, where they expect their vulnerabilities will not be exploited (Corritore, Kracher, & Wiedenbeck, 2003) It involves a willingness to accept risk based on positive expectations about others' intentions (Park, 2005) Research by Garbarino and Johnson (1999) shows that trust is a crucial predictor of customers' future purchasing intentions Additionally, Lu et al (2021) found that trust in a platform significantly influences the intention to continue using it Consequently, trust is a key factor affecting users' intentions to persist with online music platforms.
Research indicates that consumers tend to undervalue their freedom of choice when receiving guidance from trusted experts, often leading them to accept such advice more readily (Clee & Wicklund, 1980) In the context of online retail, trustworthiness is heightened (McKnight et al., 2002) Personalized recommendations on online music platforms foster trust among users, encouraging their continued engagement However, personalization can also raise privacy concerns, as it exposes users to the risks associated with personal data collection Nonetheless, studies suggest that trust can enhance a person's willingness to embrace vulnerability, driven by positive expectations of another party's intentions or actions (Rousseau et al., 1998) Additionally, factors such as familiarity and positive past interactions with a company help reassure consumers that their personal information will be collected and utilized fairly (Chellappa).
When consumers trust a retailer, they believe their personal data is secure and used ethically This trust encourages users to share their information with online music platforms, leading to personalized content that enhances user control and reduces mistrust Consequently, it is anticipated that personalization strengthens the relationship between online trust and continued usage Based on this premise, the authors proposed several hypotheses.
Hypothesis 2 (H2a) Online-Trust significantly influences the Continuous Use Intention on online music streaming platforms.
Hypothesis 2 (H2b) Personalization plays a mediating role between Online-Trust and Continuous Use Intention on online music streaming platforms.
Social Influence is defined as the extent to which individuals perceive that important people in their lives, such as colleagues, friends, and family, believe they should adopt a specific technology (Venkatesh, Thong, & Xu, 2012) Users are more likely to form intentions to use technology when they receive encouragement and motivation from those whose opinions they value (Alalwan et al., 2016) Research has shown that Social Influence significantly predicts the intention to use various applications, including mobile commerce (Verkijika, 2018), mobile payments (Khalilzadeh et al., 2017), and diet apps (Okumuặ et al., 2018) Furthermore, studies by Shen et al (2009) and Zhou and Li (2014) indicate that Social Influence also plays a crucial role in Continuous Use Intention.
Straub (2004) argued that having a sense of Social Presence among users has a positive effect on Reuse Intention.
Choi and Lee (2014) emphasize that recommendation systems benefit significantly from identifying friends or neighbors with similar interests, fostering a sense of social presence among users Personalization can effectively pinpoint individuals with shared tastes and preferences within the same social network, enhancing the influence of recommendations Li and Karahanna (2012) further assert that social network-based personalization promotes ongoing usage intentions by delivering suggestions from peers within the same network Based on these insights, the authors propose several hypotheses regarding the effectiveness of social connections in personalized recommendation systems.
Hypothesis 3 (H3a) Social Influence significantly influences the Continuous Use Intention on Online Music Streaming Platforms.
Hypothesis 3 (H3b) Personalization plays a mediating role between Social Influence and Continuous Use Intention on Online Music Streaming Platforms.
Habit is defined as the automatic performance of behaviors learned over time, reflecting prior experiences When users are satisfied with an application's performance, they are likely to develop habitual behaviors that encourage future use Research indicates that habit is a significant predictor of technology usage across various domains Therefore, fostering users' habits is crucial for their continued engagement with online music platforms.
Online music platforms leverage user data to enhance personalization, allowing for precise targeting of content based on listening history, preferences, and even emotions (Prey, 2018) This tailored approach not only increases user engagement but also fosters a habit of regular use, as personalized suggestions resonate with individual experiences (Tam & Ho, 2006) When users perceive that the service aligns with their needs, their intention to continue using the platform grows stronger, creating a positive feedback loop between personalization, habitual use, and sustained engagement Consequently, users' satisfaction with personalized experiences reinforces their commitment to the service, leading to the following hypotheses.
Hypothesis 4 (H4a) Habit significantly influences the Continuous Use Intention on online music streaming platforms.
Hypothesis 4 (H4b) Personalization plays a mediating role between Habit and Continuous Use Intention on online music streaming platforms.
Performance Expectancy (PE) reflects how effectively new technologies enhance user efficiency and convenience (Venkatesh, Thong, & Xu, 2012) In the realm of communication technology, users appreciate mobile applications for their ability to streamline goal-oriented tasks (Chopdar et al., 2018) Consumers are more likely to adopt a new system when they believe it saves time and effort compared to traditional methods If online music streaming platforms offer more convenient access to songs and information, users are likely to respond positively and continue using them Previous studies have demonstrated a significant link between PE and the intention to continue using applications, particularly in sectors like food ordering (Alalwan, 2020) and news (Ye et al., 2019).
Research by Cheng et al (2020) indicates that personalization negatively moderates the relationship between performance expectancy and the intention to continue using a mobile app Specifically, when personalization is low in a new mobile app, it enhances performance expectancy, which in turn encourages users to keep using the app.
A study by Ifekanandu et al (2023) highlights the significant role of personalization in enhancing user satisfaction in fashion apps, demonstrating its positive mediation between Information Quality, System Quality, and Service Quality, ultimately influencing Purchase Intention This research builds on the UTAUT2 Model, emphasizing the need for further exploration of personalization's multifaceted roles in user experience Based on these findings, we propose two hypotheses to extend the understanding of personalization in the context of fashion apps.
Hypothesis 5 (H5a): Performance Expectancy significantly influences the Continuous Use Intention on online music streaming platforms
Hypothesis 5 (H5b): Personalization plays a mediating role between Performance
Expectancy and Continuous Use Intention on online music streaming platforms
Effort Expectancy refers to a consumer's comfort level with technology use, significantly impacting user engagement and adoption rates Research indicates that perceived ease of use enhances user interaction with technology, while complexity can deter application adoption and return intentions In the context of mobile applications, effort expectancy is a crucial factor influencing user behavior Additionally, comfort in operating related tasks is shown to affect users' intentions to utilize news applications Consequently, when users find online music streaming platforms easy and convenient to navigate, they are more likely to develop positive intentions towards using them.
Research by Krishnaraju and Mathew (2016) indicates that personalization enhances self-reference, content relevance, and navigational ease, ultimately reducing cognitive load and encouraging technology adoption Additionally, Koufaris (2002) found that when personalization offers users choices, it fosters a sense of control and enhances the overall user experience, leading to greater comfort and enjoyment on the platform Online music streaming services leverage personalization features, enabling users to customize their homepage and interface according to their preferences, such as music genres and artists This tailored approach allows users to prioritize their favorite songs and stay updated on relevant topics with minimal effort Consequently, when users perceive personalized content as a means to simplify and enhance their experience, they are more likely to save time and increase their intention to reuse the platform.
Hypothesis 6 (H6a) Effort Expectancy significantly influences the Continuous Use Intention on online music streaming platforms.
Hypothesis 6 (H6b): Personalization plays a mediating role between Effort Expectancy and Continuous Use Intention on online music streaming platforms
RESEARCH MODELS
Research process
Firstly, through real-life observations, we selected the topic and identified issues related to the topic, including objectives and relevant issues.
Secondary data analysis involves the examination, synthesis, and comparison of existing data to develop models and formulate hypotheses This process includes reviewing reference materials, particularly prior research by relevant international authors, which serves as a foundation for creating questionnaires and selecting suitable scales for the questions.
In our research, we conducted quantitative analysis to assess the validity of our research model and measurement scales We employed Likert scales to gather data from a sample of 329 participants This comprehensive evaluation involved rigorous testing to ensure the accuracy of the scales and the overall appropriateness of the research model in relation to real-world data.
Processing and analyzing data using statistical methods entails the collection and examination of data to uncover relationships between variables and test research hypotheses This approach enables researchers to derive significant conclusions that contribute to both the academic field and practical applications.
Secondary data research methods
This research establishes a clear direction and theoretical foundation for the proposed model by utilizing a variety of reference materials and studies from international authors focused on the intention to continue usage As a result, the gathered information provides a robust and precise basis for the proposed model and hypotheses within the research process.
3.2.2 Results of secondary data research
Through an extensive review of models and hypotheses from esteemed international scientific journals like ScienceDirect and MDP1, our research has identified key variables and developed an appropriate research model The selection of independent variables is grounded in their frequent use by various authors and their significant impact on the dependent variables relevant to our study topic.
After thorough research and analysis of academic sources, we developed a proposed research model featuring six key factors that influence users' intention to continue using online music platforms These factors include Artificial Intelligence (AI), Online Trust (OT), Social Influence (SI), and Habit (H).
This study examines the influence of five key variables: Performance Expectancy (PE), Effort Expectancy (EE), and the additional factor of Personalization (P), totaling seven variables We will employ quantitative analysis methods to assess their impact on the dependent variable, Continuous Intention (CI), and to validate the hypotheses outlined in Chapter 2.
Quantitative Research Methods
To ensure reliable research findings, it is essential to estimate a minimum sample size of at least N5, where N represents the number of questions (Hair et al., 1998) For this study, the required sample size is 415, calculated as 205 The research team collected actual data from 329 respondents, which exceeds the minimum requirement of 205 This larger sample size enhances the reliability of the regression model results.
Coding of the satisfaction scale
All Artificial intelligence (Al) used on online music platforms helps me discover many good music
AI2 Artificial intelligence (AI) used in online music platforms satisfies my consumption needs
AI3 Random playlists on online music platforms renew my music experience
AI4 New artists I've never heard of are recommended by artificial intelligence (AI) that match my interests
AI5 Automatic AI-recommended playlists that match my music taste
AI6 Podcasts recommended by artificial intelligence (AI) fit my needs
OT1-OT4: Eid (2011), Merrilees and Fry (2003)
OT5-OT7: Morgan and Hunt (1994)
OT1 I believe that online music platforms are trustworthy and honest
OT2 I believe that online music platforms build trust among their customers
OT3 I believe that online music platforms deliver on the promises and commitments they make
OT4 I believe that online music platforms will keep my personal information secure
OT5 I believe that music streaming platforms have my best interests al heart
OT6 I trust the information that online music platforms provide
OT7 I believe that music streaming platforms really care about their customers
Sil Friends and colleagues around me use music streaming platforms, and their use influences my use of those platforms.
SI2 My boss or people I admire use music streaming platforms and their use influences my use of the platforms
SI3 On social networks, people use music platforms and share trends about it, creating joy and motivation for me to continue committing to those platforms.
SI4 In general, people around me support and share my joy in using online music platforms
Venkatesh, Thong, & Xu (2012), Verplanken & Orbell (2003)
Hl Using online music platforms has become a habit of mine
H2 I’m addicted to using online music platforms
H3 I have to use online music platforms
H4 Using online music platforms is something I do without thinking
PEI I find online music platforms very useful in my daily life
PE2 Using online music platforms helps me get things done faster
PE3 Using an online music platform increases my work productivity/performance
PE4 Online music platforms allow me to listen to music with good sound quality
EEl Learning how to use online music platforms was easy for me
EE2 My interactions with music platforms are clear and understandable
EE3 I find online music platforms very easy to use
EE4 I can easily become proficient in using online music platforms
Pl Online music platforms understand my needs
P2 Online music platforms can provide me with personalized songs that suit my daily activities
P3 It's important for me to be able to customize my account on online music platforms
I highly value the ability of online music platforms to recommend songs, artists, and podcasts that align with my personal preferences, enhancing my listening experience.
P5 It’s important for me to be able to create custom music playlists
P6 It's important for me to get information about the bands/ musicians/ singers 1 follow
P7 It's important for me to be able to share music, playlists or podcasts on social media
CIl I am willing to use paid online music platforms, use them and share them with others through social networking sites on a regular basis in the future.
CI2 I plan to use online music platforms for a long time
CI3 I will continue to use online music platforms
This study employed a convenience random sampling method, utilizing social networks like Facebook and Zalo to distribute survey questions By posting survey information in Facebook groups with a high participation of UEH students, the research team benefited from a convenient approach that facilitated quick information collection while saving time.
To identify the appropriate respondents for the study, the research team incorporated a screening question in the survey questionnaire This approach led to the selection of 329 qualified samples, which were subsequently coded and analyzed using SmartPLS software to evaluate the research hypotheses.
The following are the steps of data processing and analysis performed by the group:
Step 1 Reliability Analysis: Using SmartPLS 4.0 software to determine the reliability of the survey variables using Cronbach's Alpha coefficient Eliminating inappropriate observed variables.
Step 2 Discriminant Value Analysis: Firstly, we analyse cross-loading factors and apply the Former-Laker criteria, calculating HTMT index and therefore, eliminate inappropriate variables.
Step 3 Regression Analysis and Hypothesis Testing
Using regression analysis to test the research model.
Applying mediation analysis to test research hypotheses.
Step 4 Analysis and Interpretation of Results
Analysing the results obtained from the regression and mediation models.
Interpreting the significance of the results obtained and drawing conclusions for the study.
RESEARCH RESULTS
Descriptive statistics
We collected a total of 329 responses through a Google Form survey, all of which met our criteria Following the data collection and categorization process, our team encoded the information and utilized Smart PLS 4 software for in-depth analysis.
In a survey of 329 responses, female participants comprised 69.3% (228 individuals), while male participants accounted for 30.4% (100 individuals) Additionally, 1 participant identified as another gender, representing 0.3% of the total This data highlights a significant gender discrepancy among the respondents.
The survey revealed that the majority of participants, 78.7%, were aged between 18 and 25, with 259 responses In contrast, the age group of 26 to 40 years had the lowest participation, contributing only 5.8% with 19 responses Additionally, participants under 18 accounted for 15.5% of the total, with 51 responses Overall, the 18 to 25 age group comprised nearly four-fifths of all survey participants.
The majority of survey participants reported a monthly income below 5 million VND, with
The survey gathered a total of 207 responses, representing 62.92% of participants, with 105 individuals earning between 5 million VND and 15 million VND per month, making up 31.92% Additionally, 11 respondents reported monthly earnings between 15 million VND and 30 million VND, which accounted for 3.34% Only six participants indicated a monthly income exceeding 30 million VND, representing 1.82% of the total responses This significant disparity highlights the differing usage needs for online music platforms across various income levels, particularly showing a scarcity of data from higher-income individuals.
Table 4 Monthly income/ allowance structure
The survey results indicate that a significant majority of respondents, comprising 245 individuals or 74.5%, are students In contrast, 37 participants from the business sector represent 11.2% of the total responses Other fields include Administration with 13 responses (4.0%), Arts with 8 (2.4%), Politics with 2 (0.6%), Homemaking with 3 (0.9%), Engineering with 7 (2.1%), and Other fields with 14 responses (4.3%) Overall, the data on age, income/allowance, and occupation show consistent trends across the different categories.
Reliability and validity testing of the scale
4.2.1 Internal consistency reliability (Cronbach’s alpha)
Cronbach’s alpha is a key measure of internal consistency reliability, with acceptable values ranging from 0.7 to 0.8, and values from 0.8 to nearly 1 indicating good scale reliability (Hair et al., 2021) The reliability analysis results presented in Table 6 demonstrate that all scales achieved internal consistency, with Cronbach’s alpha values between 0.784 for the CI scale and 0.897 for the OT scale These findings confirm that the measured variables possess significant reliability and are valuable for the research process.
Table 6 Reliability analysis and average variance extracted (AVE) of the scales
All scales demonstrated internal consistency reliability, as indicated by Cronbach's alpha However, as noted by Hair et al (2021), a reliable scale also requires evaluation through the composite reliability coefficient (CR), which is deemed more dependable The CR assesses the internal consistency of indicators within a scale and serves as an alternative to Cronbach's alpha (Netemeyer et al., 2003) Acceptable CR values range from 0.6 to 0.7, while optimal reliability is found between 0.7 and 0.9 According to Table 6, the CR values for the CI and EE scales range from 0.875 to 0.921, indicating high reliability across all tested scales, including AI.
CI, EE, H, OT, p, PE, and SI are acceptable and suitable for further evaluation in the study.
Scale value
The outer loading coefficient is a key metric for evaluating the contribution of variables in a model Hair et al (2021) suggest retaining variables with outer loading coefficients exceeding 0.7, while those ranging from 0.4 to 0.7 may be considered for retention or elimination based on their significance in impacting the research model.
The analysis reveals that the variable "All" has the lowest coefficient at 0.735, indicating it remains within a safe range Consequently, this suggests that all variables in the scale, including AI, are performing adequately.
CT, EE, H, OT, p, PE, and SI, meet the criteria and are retained for further study.
Table 7 Outer loading coefficients ofobserved variables
Al CI EE H OT p PE SI
Al CI EE H OT p PE SI
The Average Variance Extracted (AVE) is a crucial metric for evaluating the convergent validity of a scale, representing the total average of the squared factor loading coefficients of observed variables An AVE value of 0.5 or above signifies that the research concept accounts for more than half of the variance in its observed variables, as outlined in the PLS-SEM book by Hair et al (2021) According to the analysis presented in Table 6, all scales satisfy this criterion, with the p scale recording the lowest AVE value of 0.595 and the EE scale achieving the highest at 0.745.
Discriminant validity will be assessed using three criteria: (1) the Cross Loading coefficient, (2) the Fornell-Larcker criterion as established by Fornell and Larcker (1981), and (3) the Heterotrait-Monotrait Ratio of Correlations (HTMT) based on the research by Henseler and Ringle (2015) While it is not mandatory to apply all three criteria for evaluating the discrimination of the research model, Hair and Hull (2017) strongly recommend utilizing a combination of all three for a comprehensive assessment.
To assess discriminant validity, as suggested by Fornell and Larcker (1981), the outer loading of latent factors must exceed the cross-loading coefficients of other factors in the model Table 8 demonstrates that the outer loading coefficients for the observed variables in the parent factor, such as All-AI, AI2-AI, AI3-AI, AI4-AI, AI5-AI, and AI6-AI, are consistently higher than the cross-loading coefficients of the same observed variables in the other factors This indicates that all observed variables serve as strong measures of their respective parent variables, thereby confirming the discriminant validity of the research model.
AI CI EE H OT p PE SI
AI CI EE H OT p PE SI
AI CI EE H OT p PE SI P2 0.577 0.567 0.631 0.491 0.543 0.803 0.613 0.610
To evaluate discriminant validity according to Fornell and Larcker's (1981) criterion, the square root of the average variance extracted (AVE) for each variable must exceed the correlation with other variables Table 9 demonstrates that the square root of the AVE for variables AI, CI, EE, II, OT, p, PE, and SI is greater than their respective correlation coefficients with other variables This confirms that the scales for AI, CI, EE, H, OT, p, PE, and SI maintain discriminant validity, with no violations present.
AI CI EE H OT p PE SI
4.3.2.3 Heterotrait-Monotrait Ratio of Correlations
Henseler and Ringle (2015) identified the limitations of the Fornell-Larcker criterion and introduced the Heterotrait-Monotrait Ratio of Correlations (HTMT) as a more effective method for assessing discriminant validity in models According to the HTMT analysis, if any value exceeds the threshold of 0.9, discriminant validity cannot be assured However, as shown in Table 10, all HTMT values are below 0.9, confirming that the model maintains discriminant validity.
Al CI EE H OT p PE SI
Al CI EE H OT p PE SI
The model's discriminant validity is confirmed through three measurement criteria, indicating that the latent concepts are measured independently and accurately Each observed variable is associated with only one latent concept, enhancing the research results' accuracy This validation of discriminant validity demonstrates the model's strong capability to explain the data effectively.
Model Structure Evaluation
Multicollinearity is a significant issue in research, as highlighted by Daoud (2017), because it can distort research outcomes This phenomenon arises when independent variables in a linear regression model exhibit high correlation with one another, allowing one variable to be accurately predicted by the others To assess multicollinearity, Daoud (2017) recommends using the Variance Inflation Factor (VIF) and has developed a scale for interpreting VIF values in regression models.
VIF = 1: No correlation between independent variables
1 < VIF < 5: Weak correlation between independent variables
VIF > 5: High correlation between independent variables
Table 11 reveals that all VIF values fall below 5, ranging from 1.403 to 2.567, which demonstrates weak correlations among the independent variables and confirms the absence of multicollinearity Consequently, this linear regression model is expected to exhibit enhanced accuracy, reliability, and predictive capability.
Table ỉJ VỉF - Variance Inflation Factor
Chicco et al (2021) highlight the widespread use of the R-squared metric for evaluating regression analyses across various scientific disciplines A recent study by Ozili (2023) categorizes the R-squared index into three distinct levels, particularly emphasizing its application in social science research.
0 < R-squared < 0.09: The R-squared index is too low, indicating that the model does not fit the data set.
0.10 < R-squared < 0.50: An R-squared index within this range is acceptable, provided that many variables in the model have statistical significance.
0.51 < R-squared < 0.99: An R-squared index within this range is completely acceptable Table E shows two different R-squared values:
The R-squared value for the CI variable indicates a significant relationship between seven variables—six independent variables (AI, OT, SI, H, PE, EE) and one mediating variable (P)—and the dependent variable (CI) With an R-squared value of 59.7% and an adjusted R-squared of 59.4%, the model demonstrates a strong fit to the dataset, explaining 59.4% of the variation in the CI variable.
The R-squared value for the p variable indicates the influence of six independent variables—AI, OT, SI, H, PE, and EE—on the mediating variable (P) With an adjusted R-squared value of 67.9%, the model demonstrates a strong fit for the dataset However, this value slightly decreases to 67.7%, reflecting the variation among the independent variables.
SI, H, PE, EE) will lead to 67.7% of the variation of the dependent variable (CI).
The adjusted R-squared value for the Personalization (P) variable is 67.9%, surpassing the 59.4% adjusted R-squared value for Continuous Use Intention (CI) This indicates that the independent variables have a more significant influence on Personalization than on Continuous Use Intention.
The f-square test, as described by Umaroh and Barmawi (2020), is a valuable tool for evaluating the relative influence of an exogenous construct on an endogenous construct This test measures how much an independent variable affects a dependent variable in comparison to other independent variables within the model A high f-square value signifies that the exogenous variable exerts a greater impact on the endogenous variable than its counterparts, thereby helping to identify whether an independent variable significantly influences the dependent variable.
Umaroh and Barmawi (2020) mention four levels of f-square test:
0 < f-square < 0.019: the exogenous construct has a very small impact on the endogenous construct
0.020 < f-square < 0.15: the exogenous construct has a small impact on the endogenous construct
In the context of evaluating exogenous constructs, an f-square value between 0.16 and 0.35 indicates a medium impact on the endogenous construct, while an f-square value greater than 0.35 signifies a large impact This assessment can be further supported by referencing the F table for a comprehensive analysis of the results.
H —> p, OT —* CI, PE —► CI, SI —ằ Cl, SI —► p have almost very little impact on the dependent variable.
In particular, the relationship AI —> CI has almost no impact.
The rest of the relationships have a small impact on the dependent variable.
The f-square value may be small due to the presence of multiple independent variables in the model Nevertheless, it serves as one of several metrics for researchers to assess the model's effectiveness, although it should not be viewed as a definitive measure.
Table J3 /-square value f-square AI ( 1 0
Regression Result - Mediation Analysis
Figure Ỉ0 Path analysis ofevery' research construct Artificial Intelligence (AỈ), Online Trust (OT), Social Influence (SI), Habit (H), Performance Expectancy (P), Effort
Expectancy (EE), Personalization (P), Continuous Use Intention (CI)
Following the analysis of the R-square and f-square values, we proceed to evaluate each path within our research model The Structural Equation Model results, illustrated in the figure above, include the Original sample (O), Sample Mean (M), Standard deviation, T Statistics, and p-values, which are compiled in Table G below.
Hla is rejected as it is statistically insignificant (p=0.844>0.05, p=-0.015) Therefore,
The study reveals that Artificial Intelligence (AI) and Online Trust do not significantly influence Continuous Use Intention, as indicated by the rejection of H2a (p=0.404) and H3a (p=0.332) In contrast, Habit is found to positively impact Continuous Use Intention with a significant coefficient of 0.322 (H4a supported, p=0.000) Additionally, Performance Expectancy is shown to have no effect on Continuous Use Intention, leading to the rejection of H5a (p=0.118) However, Effort Expectancy is confirmed to positively influence Continuous Use Intention with a coefficient of 0.190 (H6a confirmed, p=0.009).
Hla, H2a, H3a, H5a are rejected and H4a, H6a are statistically confirmed.
This research will conduct a Mediation Analysis using the methodology outlined by Taylor, MacKinnon et al (2008), employing SmartPLS to generate a bias-corrected percentile bootstrapping table with a 95% confidence interval based on 5,000 bootstrap samples Following the guidelines of Preacher and Hayes (2008), the presence of a zero within the confidence intervals indicates an insignificant mediating effect, suggesting that no mediation is present.
(Hair Jr, Hult et al 2021) in his book: “Partial least squares structural equation modeling (PLS-SEM) using R: A workbook'* has characterized the outcomes of Mediation Analysis into 5 categories:
Complementary mediation, also referred to as partial mediation, occurs when both direct and indirect effects are statistically significant and align in the same direction, whether positive or negative.
Competitive mediation, also referred to as partial mediation, reveals that both direct and indirect effects are statistically significant, yet they operate in opposing directions—one effect is negative while the other is positive.
Indirect-only Mediation (known as full Mediation): Only the indirect effect is statistically significant
Direct-only Non-mediation (no Mediation): Only the direct effect is statistically significant
No-effect Non-mediation (no Mediation): neither the indirect nor the direct effect is statistically significant
The book of (Hair Jr, Hull et al 2021) is the development of (Zhao, Lynch Jr et al 2010)’s academic work in the field.
Figure JI Mediation analysis procedure (Zhao, Lynch Jr el al 20JO)
Table 15 illustrates the direct effects of Independent Variables on Dependent Variables when a mediator is present, as well as the indirect effects mediated by the mediator Additionally, it highlights the total effects of Independent Variables on Dependent Variables without the mediator's involvement These total effects are essential for estimating the percentage of influence that the indirect effects, facilitated by the mediator, contribute to the overall relationship (Hair Jr, Hult et al., 2021).
Table H indicates that Hlb is rejected, as personalization does not mediate the relationship between artificial intelligence (AI) and continuous use intention (CI) The indirect effect of AI on CI through personalization is statistically insignificant (p=0.061 > 0.05, p=0.047), and similarly, the direct effect of AI on CI is also found to be insignificant.
(p=0.844>0.05, p= -0.0Ỉ5) Therefore Personalization is not qualified to be the mediator of the relationship between Artificial Intelligence and Continuous Use Intention.
H2b is accepted because Personalization is qualified to have a mediating effect upon the relationship between Online Trust and Continuous Use Intention The indirect effect (OT
The study reveals a statistically significant relationship between Online Trust and Continuous Use Intention (p=0.042), while the direct effect is insignificant (p=0.404) This indicates that Personalization acts as a mediator, categorizing the relationship as Indirect-only Mediation or Full Mediation.
The indirect effect (SI —> p —► CI) of Social Influence and Continuous Use Intention with the involvement of Personalization as a mediator is statistically insignificant
(p=0.096>0.05, p=0.032) The direct effect (SI —> CT) of the two is also insignificant
(p=0.332>0.05, p=0.062) Therefore, there is a No-effect Non - mediation relationship between Social Influence and Continuous Use Intention H3b is rejected.
H4b is dismissed as personalization does not serve as a mediator between habit and continuous use intention This indicates a direct-only relationship, where the indirect effect of habit on continuous use intention through personalization is statistically insignificant.
(p=0.654>0.05, p=-0.007) but the direct effect (H —> P) is significant (p=0.000 Continuous Use Intention Complementary Partial Mediation
Discuss research results
4.6.1 Impact on Continuous Use Intention
4.6.1.1 Artificial Intelligence to Intent to Continue Using
A recent study indicated that Artificial Intelligence (AI) has no significant positive impact on users' intent to continue using music platforms, with a weighting of -0.015 and a p-value greater than 0.05, aligning with findings from Prentice et al (2020) While AI analyzes customer data to tailor music products and services, this deep analysis can lead to user concerns about data privacy and security Users may fear personal information leaks and question the ethical implications of AI, including potential bias and privacy issues, which can ultimately affect their willingness to engage with these platforms in the future.
AI for malicious purposes However, this result contrasts with results from research by Bhagat el al (2022), Chen el al (2023), and Ifekanandu el al (2023), which all show that
AI significantly influences user intent and enhances customer loyalty, highlighting the necessity for more research into its effects on consumer behavior Understanding both the positive and negative aspects of AI's impact is crucial for clarifying its role in shaping human interactions with brands.
4.6.1.2 Online-Trust and Continuous Use Intention
A recent study indicates that Online-Trust does not significantly influence Continuous Use Intention on online music streaming platforms, with findings mirroring those of Khatib et al (2019), which also reported no significant correlation between trust and loyalty in this context Users are likely to continue utilizing familiar and convenient platforms primarily due to their user experience rather than trust Factors such as sound quality, user interface, and search capabilities play a crucial role in their decision-making, as users prioritize features that align with their preferences over trust considerations Additionally, the low subscription prices and the absence of mandatory upgrade packages on many platforms reduce users' concerns regarding potential losses from unsafe transactions or risks to personal data.
4.6.1.3 Social Influence and Continuous Use Intention
Research indicates that social influence has a minimal effect on users' intentions to continue using online music platforms, with a significance level of 0.062 (p > 0.05) This aligns with findings from Cheng et al (2020), which suggest that personal preferences and needs outweigh social pressures in decision-making The study highlights that customers' knowledge and experience with music services are more critical than external opinions, particularly among younger users aged 18-25, who are less influenced by others While social interactions, such as sharing music on platforms like Facebook, may introduce users to new songs, they do not significantly affect ongoing listening habits, supporting the principles of Social Exchange Theory.
The survey revealed that a significant portion of participants belonged to Generation Z, a demographic known for its proactive approach to information gathering and product comparison This generation increasingly values independent purchasing decisions and is less influenced by the opinions and evaluations of others compared to previous generations.
4.6.1.4 Habit and Continuous Use Intention
Research indicates that habit significantly influences the intention to continue using online music streaming platforms, with a raw weight coefficient of 0.322 and a p-value of less than 0.05 This suggests that once users develop a habit, they are likely to automatically persist in using these streaming apps.
Habits related to technology app usage are developed through repeated behaviors, leading to stronger intentions that influence user actions As these habits solidify, users may find themselves responding automatically in similar situations, bypassing conscious thought processes.
Habit plays a crucial role in consumer behavior, significantly impacting loyalty and customer retention As users grow accustomed to technology, digitization facilitates their ability to access music anytime and anywhere, thereby shaping their habits and influencing their intentions and behaviors.
4.6.1.5 Performance Expectancy and Continuous Use Intention
This research indicates that Performance Expectancy does not significantly influence Continuous Use Intention on online music streaming platforms, particularly when Personalization is considered as a mediator When Personalization is excluded, the total effect of Performance Expectancy on Continuous Use Intention becomes significant, yet it remains the lowest among the variables studied, such as Habit and Effort Expectancy Additionally, previous studies, including Lee, Kim et al (2022), have shown that Performance Expectancy also lacks a direct impact on Customer Satisfaction, which is crucial for Subscription Intention Instead, factors like Price Value, Curation System, and Habit are more likely to enhance Customer Satisfaction Given the limited research on Performance Expectancy in the context of online music streaming, future studies could explore this variable further to better understand customer behavior.
4.6.I.6 Effort Expectancy and Continuous Use Intention
Effort Expectancy significantly influences users' Continuous Use Intention on online music streaming platforms, with a notable weight of 0.190 (p < 0.05), aligning with earlier studies (Gao et al., 2015; Hew et al., 2015; Teo et al., 2015) These platforms offer users convenient access to music anytime and anywhere, emphasizing the importance of an intuitive and user-friendly interface that enhances their appeal Research indicates that lower effort in understanding and utilizing technology correlates with a higher likelihood of continued use Users prefer technologies that prioritize simplicity and efficiency (Agarwal and Prasad, 1999; Davis et al., 1989) Consequently, when users find these platforms easy to navigate and straightforward, they are more likely to return for future use.
4 6.2.I Personalization as a mediator of the relationship between Artificial Intelligent and Continuous Use Intention
Research indicates that personalization does not mediate the relationship between artificial intelligence (AI) and continuous use intention on online music streaming platforms, with a statistically negligible indirect impact (p = 0.061) Consequently, the hypothesis that personalization serves as a mediator in this context is rejected This finding contrasts with Ifekanandu's 2023 study, which demonstrated that personalization significantly enhanced customer experience and loyalty through AI While personalization can positively influence customer engagement, it may also lead to adverse effects, such as diminished user comfort and attention due to competing app features Additionally, reliance on personalized AI systems can cause users to feel a lack of control, resulting in frustration and lower performance expectations Overall, the findings suggest that high levels of personalization may not enhance user retention in streaming music applications.
& Hitt, 2018), and concerns about user privacy from personalizing content will negatively affect users' habitual behavior when using personalized services.
4 Ó.2.2 Personalization as a mediator of the relationship between Online-Trust and Continuous Use Intention
The study reveals that Personalization serves as a critical mediator between Online Trust and Continuous Use Intention in Online Music Streaming Platforms, indicating an indirect-only mediation effect Specifically, Online Trust influences Continuous Use Intention solely through the mediating role of Personalization This aligns with findings from Zhen Shao et al (2019), who emphasized that customer awareness of mobility and personalization enhances trust in mobile payment platforms, thereby encouraging continued usage Similarly, Siau et al (2003) and Huang et al (2014) highlighted the positive relationship between personalization and trust in e-commerce Personalization prompts users to assess their online trust in platforms, which non-personalized information fails to achieve It encompasses the belief that music streaming services provide tailored attention and understand user needs, as noted by Swaid and Wigand (2007, 2009) Furthermore, the transparent collection of personal data to create customized content fosters trust in the security and proper use of user information, emphasizing the importance of clear and intuitive data practices to enhance user trust and encourage ongoing engagement.
4 6.23 Personalization as a mediator of the relationship between Social Influence and Continuous Use Intention
Our study identifies the relationship between Social Influence and Continuous Use Intention as a No-effect non-mediation, indicating that Social Influence does not significantly impact Continuous Use Intention, which aligns with Cheng's (2020) findings Furthermore, the mediating effect of Personalization on this relationship is negligible for online music platforms In contrast, Eun-Jung Lee & Jung Kun Park's research suggests that Subjective Norm positively influences attitudes toward service personalization, asserting that positive opinions from peers enhance individual attitudes toward personalized service features However, our findings diverge from this perspective, as Personalization does not affect the relationship between Social Influence and Continuous Use Intention Cheney (2011) noted that online music platforms utilize big data to tailor music recommendations to individual preferences and lifestyles Given that today's youth share many similarities in preferences and values, Social Influence appears irrelevant to music resources, as music serves primarily as a reflection of individual tastes and styles The rise of online music platforms further consolidates these preferences, highlighting popular songs while diminishing the relevance of specific genres to distinct teenage groups (Anthony Y H Fung, 2014).
4 6.2.4 Personalization as a mediator of the relationship between Habit and
The study reveals that personalization does not mediate the relationship between habit and continuous use intention, indicating a direct-only effect Specifically, while habit significantly influences the intention to continue using, the pathway involving personalization is insignificant These findings align with previous research by Orbell, Blair, Sherlock, and Conner, reinforcing the notion that habit plays a crucial role in determining continuous use intention without mediation.
The findings indicate that personalization does not mediate the relationship between habit and the intention to continue using online music platforms This suggests that as habits develop, behaviors become automatic, diminishing the influence of personalized factors Additionally, habits can be general and may change over time, failing to accurately represent individual preferences In some instances, habits form when individuals are required to engage in specific behaviors for an extended period, ultimately leading to their establishment.
4 Ó.2.5 Personalization as a mediator of the relationship between Performance
Expectancy and Continuous Use Intention