Effects of TikTok Influencers on the Purchase Intention of Gen-Z: Cosmetic Industry

. Abstract This study investigates how TikTok influencers affect Generation Z's purchase intention toward the cosmetic industry in HCMC. The research issue is a lack of knowledge on how TikTok influencers affect Generation Z's buying intentions for cosmetics. The research objective is to investigate the relationships among influencers' credibility, physical attractiveness, expertise, social advocacy, customers' trust, gender, and purchase intention. The literature study emphasizes the significance of customer trust and influencer characteristics in influencing consumer intention. A questionnaire survey is used as a data gathering technique in this research employing a quantitative methodology. To analyze the data, SPSS will be used as the statistical program. Descriptive statistics will be computed to summarize the data. Multiple regression analysis will examine the relationship between influencers' characteristics and customer trust toward their purchase intention. According to the findings, all influencers' characteristics, including credibility, physical attractiveness, expertise, and social advocacy, have favorable effects on Gen Z's purchasing intention. Surprisingly, customer trust has the most influence on Gen Z behavior. It indicates that there is a difference in purchase intention for different genders. The outcomes will offer valuable insights to cosmetic industry marketers at HCMC to develop effective influencer marketing strategies for Generation Z consumers.


Introduce the Problem
TikTok has become popular among young individuals aged 18 to 24.TikTok influencers have become powerful drivers of product promotion.This study aims to examine the influence of TikTok influencers on the purchasing decisions of Generation Z in the cosmetics sector of Ho Chi Minh City.The cosmetics industry can effectively engage Generation Z consumers with the insights and recommendations found in this report.

Purchasing Intention
The Theory of Planned Behavior (TPB), developed by Ajzen in 2011, builds on the Theory of Reasoned Action (Ajzen, 1991) and is a significant theory that utilizes individuals' beliefs to predict human behavior (Hegner et al., 2017).TPB suggests that an individual's attitude, subjective norms, and behavioral control influence their intention to perform a specific behavior.Intention acts as a mediator between personal dynamics and behavior, and it is considered a precursor to behavior.Attitudes are individuals' evaluations of behavior, while subjective norms reflect their perception of social pressure.The theory posits that more positive attitudes toward a behavior strengthen the intention to execute it (Armitage & Conner, 2001).Behavioral control factors relate to specific facilitators or obstacles to a behavior.Critics of TPB argue that it is overly rational and neglects affective and cognitive factors that profoundly affect human behavior (Hegner et al., 2017).Therefore, integrating other variables with TPB determinants in one model could be an interesting research avenue for fashion researchers, academics, and advertisers.Purchase intention refers to the willingness and ability of consumers to buy a specific brand in the future, which can be influenced by various factors (Huang et al., 2011).It also includes the extent to which consumers are influenced by recommendations or endorsements while making purchases (Chetioui et al., Published by IDEAS SPREAD 2020).Therefore, purchase decisions can either be driven by emotions or rationality.This study aims to investigate whether individuals are more likely to make a purchase when the advertiser is someone they care about and trust, or if other factors play a role.The research explores the influence of personal relationships and trustworthiness on consumer buying behavior.

Influencers
Traditionally, internet research has predominantly centered on micro-phenomena involving celebrities primarily from the northern hemisphere, thereby shaping users' understanding of typical online practices.In response to this limitation, scholars in the field of Internet research have drawn attention to celebrity cases within their respective study domains (Abidin, 2019;Abidin & Brown, 2019;Semati & Zambon, 2021).Influencers have amassed substantial followings and wield significant influence by actively engaging in content creation on various social media platforms.Renowned for their prowess in producing entertaining and captivating videos, these influencers can attract diverse audiences, transcending age groups, genders, and professional backgrounds (Barta et al., 2023).Particularly, TikTok influencers have achieved acclaim for their expertise, creativity, comedy, or significance in various fields or sectors.The practice of enlisting well-known individuals to promote products or services on social media is known as influencer marketing.Influencers can urge their social media followers to make specific purchases by using their persuasive abilities (Vrontis et al., 2021).In addition, in 2018, ByteDance merged Musical.ly with TikTok, which has since grown to become the seventh-largest app with 689 million users as of 2021 (Mohsin, 2021).

Generation Z
Human resources are said to be the most important value of an organization, so to achieve outstanding performance, investing in human resources is extremely necessary.With the advent of management science, motivation, and reward have been linked to general organizational problems (Hansen et al., 2002).According to Grubb (2017), Generation Z includes people born from 1998 to 2010s, while according to Grace & Seemiller (2019), they were born between 1995 and 2010.This shows that in 2017, they were between the ages of 7 and 21 years old.According to Wood (2013), Generation Z has certain outstanding characteristics, including the ability to adapt to the global world and new technologies.They are also known to be more diverse, socially aware, and globally minded than previous generations (Kahawandala et al., 2020).This may have to do with the fact that Gen Z is by far the most diverse group, growing up in more difficult financial circumstances, making the world wary of anti-discrimination laws, and increasing globalization at breakneck speed (Stukey, 2016).While other generations have had to adapt to social media and technology, Generation Z is the first to be born into it.There are differences in Generation Z return behavior in different cultures, and each country's culture influences how Generation Z returns and accepts different return policies from Asia to Europe (Serravalle et al., 2022).Social media influencers have a positive impact on Gen Z shopping behavior, especially for fashion and beauty products (Pham et al., 2021).TikTok is a rapidly growing social media platform among Gen Z. Therefore, studying Gen Z's buying behavior on TikTok can help us better understand how influencers on TikTok can influence the buying behavior of this generation.This research can help retailers and manufacturers better understand the Gen Z consumer market and develop appropriate marketing strategies.

Cosmetic Industry
The cosmetics industry in Vietnam presents a vast array of products, encompassing color cosmetics, skincare, hair care, and toiletries, thereby satisfying the burgeoning demand for enhanced physical appearance (Nguyen & Homolka, 2021;Milstein et al., 2001).The burgeoning rise in average incomes and the burgeoning expansion of the middle class, particularly among the younger demographic, have engendered substantial growth within the cosmetics market (Virac Research, 2022).Notably, TikTok, as a potent social media platform, has emerged as a preeminent channel for cosmetics brands to engage and interact actively with potential customers, leveraging its ability to disseminate viral and widely shared content (Sheak & Abdulrazak, 2023).Augmenting TikTok's influence by cosmetics influencers who consistently generate succinct videos spotlighting products, sharing experiential insights, and furnishing reviews, thus instilling trust and fostering product adoption (Indrawati et al., 2023).

Ho Chi Minh City
Ho Chi Minh City (HCMC), with its population of over 9 million, represents a significant market for consumer behavior research (Statista, 2023a).TikTok has gained remarkable popularity among Vietnamese youth, particularly Generation Z, making it an ideal platform to study the impact of influencers on their purchasing behavior.Statista (2023b) reports that 67% of Gen Z in Vietnam actively used TikTok in 2022, indicating its widespread adoption.The thriving cosmetic industry in Vietnam has embraced influencer marketing as a Published by IDEAS SPREAD prominent strategy.HCMC, as a major economic and cultural center, has seen a significant presence of cosmetic brands engaging in influencer campaigns (Nguyen et al., 2022).The intense market competition in HCMC due to numerous local and international cosmetic brands makes it an excellent location to study the intersection of TikTok influencers, Generation Z, and cosmetic industry trends.The city's urban lifestyle and vibrant consumer culture provide valuable insights into consumer behavior and preferences.Studying this context will help businesses develop effective differentiation strategies within this highly competitive landscape.

Influencer's Credibility
The purpose of this study is to investigate the impact of influencers' credibility on customers' purchase intentions in the field of marketing communication.Previous research has used factor models to explore potential concerns, but their applicability to capturing the multidimensional nature of source trustworthiness in this context has been questioned.To address this issue, a more generalized approach is adopted, which considers the multifaceted aspects of source credibility.The term "credibility" in this study is defined as the perception of the accuracy of specific information by the recipient (Hovland & Weiss, 1951).The findings of this research highlight the significant influence of influencers on customers' purchasing intentions on social media platforms.Specifically, influencers are shown to have both a direct impact on customers' attitudes and an indirect influence on their purchase intentions through their mindsets.However, it is crucial to note that some experts argue that brand reputation holds greater importance than the trustworthiness of internet influencers in influencing customers' purchase intentions (Hmoud et al., 2022).Conversely, other studies have reported that influencers' credibility traits, such as attractiveness, trustworthiness, and expertise, positively influence purchase intention (Martiningsih & Setyawan, 2022;Li & Peng, 2021).Based on these findings and a comparison with other scientific papers, the following hypothesis is proposed: Hypothesis 1 (H1): Influencers' credibility has a positive influence on purchase intention.

Influencer's Physical Attractiveness
According to Patzer (1985), physical attractiveness plays a significant role in influencing people's perceptions, attitudes, and behaviors.It has been found that individuals who are considered physically attractive are more likely to be perceived positively and receive greater attention and likes from others.This effect is not limited to specific cultures but extends across different societies and contexts.Physical attractiveness plays a crucial role in a person's perception, as it sends a powerful nonverbal message that profoundly affects one's view of another person.Those who are more physically attractive tend to receive more attention and positive evaluations from others compared to those who are less attractive (Kleck & Rubenstein, 1975).In marketing, the use of attractive individuals to endorse products has been shown to generate more positive responses from customers compared to traditional forms of advertising (Till & Busler, 2000).The inclusion of visually appealing individuals in promotional campaigns enhances customer engagement and generates favorable reactions (Ozanne et al., 2019).Additionally, utilizing influencers who promote products related to their lifestyle and image can be more convincing to buyers.Following the research of Kahle & Homer (1985), the match between the celebrity and the product can create a sense of synergy that can be appealing to customers.Therefore, the hypothesis suggests that customers' perceptions of physical attractiveness and the match between the celebrity and the product positively influence their purchase decisions.Furthermore, compared with other studies, including Le (2022) in Vietnam, it is said that KOLs's appearance has an indirect effect on customers' buying behavior.Eroğlu (2019) has provided evidence that physical attractiveness is a key factor that impacts purchasing intentions.This finding suggests that physical attractiveness plays a significant role in influencing customers' decision-making processes.Therefore, the hypothesis is proposed: Hypothesis 2 (H2): The influencer's physical attractiveness has a positive influence on purchase intention.

Influencer's Expertise
The term "expertise" is generally understood to mean a person's high level of knowledge and experience in a specific field.Experts are individuals who have reached elite or peak performance levels in areas such as sports, music, science, or mathematics.When someone reaches this level, they are referred to as experts or other similar terms like masters, prodigies, or geniuses (Bourne et al., 2014).Festinger (1954) argued that social comparison theory suggests that people tend to use others as a source of judgment for their performance and behavior.These comparisons give knowledge about capabilities, social standing, and performance as well as those of others, enabling movement through the social world with ease.This suggests that it is likely that a source with more expertise will be more persuasive and influence consumer perception and behavior.In a study conducted in Thailand, expertise was found to be one of the key factors influencing consumer behavior (Lin & Nuangjamnong, Published by IDEAS SPREAD 2022).Customers, particularly when considering products like cosmetics or chemicals, often choose qualified influencers to ensure their safety.Similarly, Le et al. (2021) hold the view that expertise ranked second in its effect on consumer purchase intention, following trustworthiness, in the study of Vietnam.Vietnamese internet users prioritize trustworthy sources of information due to the presence of incorrect material on social media.On the other hand, the importance of expertise in influencing purchase decisions can vary based on culture and region.Although Widyanto & Agusti (2020) found that expertise may not be crucial for Generation Z consumers in the Jabodetabek area, it still plays a significant role in influencing buying decisions in Vietnam.The skill and experience demonstrated by influencers can easily persuade Vietnamese consumers to make purchases.Therefore, the following hypothesis is formulated: Hypothesis 3 (H3): Influencers' expertise has a positive influence on purchase intention.2.6.4Influencer's Social Advocacy Abidin (2020) highlighted the significant attention social media influencers, especially those on TikTok, receive from younger audiences in the digital world, emphasizing their pivotal role in driving social advocacy on the internet.Social advocacy entails the sharing, discussion, and evaluation of products, services, or brands through diverse social media channels (Lovett et al., 2010).The customer-influencer relationship assumes a crucial role in facilitating social advocacy, as customers actively engaging with influencers contribute to disseminating positive information about products, thereby effectively influencing their purchase decisions (Le et al., 2021).However, diverse perspectives exist regarding the impact of social media influencers on consumer purchase intentions.Some scholars argue that this impact is mediated not solely through social advocacy but also by customer attitudes towards the product or brand endorsement (Lim et al., 2017).Notably, the presence of online social influencers not only influences purchase intention but also fosters interactive engagement between influencers and consumers (Pham et al., 2021).This engagement cultivates trust and connection, ultimately leading to increased social advocacy and consequent influence on customer purchasing decisions.Consequently, the following hypothesis is proposed: Hypothesis 4 (H4): Influencers' social advocacy is positively related to the purchase intention of products.

Customer Trust in Influencer Marketing
Customer trust is defined as confidence in the supplied content of products, people, or organizations that increase the purchase likelihood of a customer (Morgan & Hunt, 1994).Influencer marketing is a method of marketing that entails collaborating with people with a large online following to promote a product, brand, or service (Saima & Khan, 2020).According to Araujo et al. (2022), Generation Z consumers manifest their strong engagement with social media platforms like TikTok, which prompts TikTok influencers' potential to influence their purchase decisions through large followings and persuasive content.In the context of influencer marketing, trust plays a crucial role as consumers rely on influencers' recommendations and perceive them as credible sources of information (Lou & Yuan, 2019).Customers want to be confident in the safety and effectiveness of cosmetics applied to their bodies.As a result, influencers who promote cosmetic products must be transparent about their relationships with brands and provide their followers with accurate and unbiased information about the products they endorse.Several prior pieces of research investigated the link between customer confidence in influencer marketing and purchase intent.According to Le (2022), trust in KOLs has a proportionate influence on purchase intent.Similarly, Dao (2022) conducted a survey that discovered that 80% of Gen Z respondents indicated that influencers' recommendations on cosmetic goods convinced them.However, the effect of consumer trust on influencer marketing is not always beneficial.If a TikTok influencer is viewed as deceitful, it might lead to a decline in trust and purchase intention among Generation Z customers and vice versa.
Hypothesis 5 (H5): Customer trust in TikTok influencer marketing positively impacts their purchasing intention on cosmetic products.

Generation Z's Gender
Generation Z, known as the iGeneration, Post-Millennials, or Digital Natives, comprises individuals born between 1997 and 2012 (Dimock, 2019).This generation, which is heavily reliant on technology and social media, has been significantly influenced by online content creators, including social media influencers (Chiu & Ho, 2023).Research consistently indicates that gender can impact attitudes and behaviors toward cosmetics.Phua et al. (2017) found that female consumers generally exhibit a higher interest in cosmetics and are more likely to engage in cosmetic-related activities than males.A business review from Harvard conducted by Silverstein & Sayre (2009) revealed that females tend to have a more positive perception of cosmetics and display a higher beauty involvement and propensity to purchase cosmetics than males.However, Hoang et al. (2022) discovered that gender did not Published by IDEAS SPREAD influence purchase intention for cosmetic items.It is crucial to critically evaluate these studies, as the influence of gender may vary depending on the type of cosmetic product marketed or specific cultural and regional contexts, which may limit the generalizability of their findings.Therefore, there is a need for further research specifically examining the effects of gender on Generation Z's purchase intention towards cosmetic products in HCM City.
Hypothesis 6 (H6): There is a relationship between Generation Z gender and their cosmetic purchase intention.

Research Methods and Materials
The present study adopts a pragmatic research philosophy, which underscores the significance of practical, applied research and accommodates diverse perspectives.Pragmatism proves particularly suitable for this investigation, given its alignment with the study's objectives and intricacies, allowing for the integration of both quantitative and qualitative data collection techniques.Within this pragmatic framework, a deductive research approach is chosen, where specific hypotheses derived from existing theories and literature will be tested, facilitating a systematic investigation to either validate or refine these hypotheses through empirical data analysis.By exploring preexisting ideas instead of formulating new theories, the deductive approach provides a well-structured framework for this study.To implement the chosen deductive research approach, the survey method is employed as the research strategy, with surveys designed using Google Forms and disseminated to the targeted sample through popular social media platforms in Vietnam, such as Facebook, Instagram, Zalo, and TikTok.Given the research objectives and the complexity of assessing Generation Z's purchasing intentions influenced by TikTok influencers, a mixed-methods approach is adopted.By combining quantitative and qualitative methods, this approach ensures a comprehensive and multidimensional understanding of the research phenomenon, facilitates the gathering of both subjective and objective data, and thereby contributes to comprehensive and reliable insights.

Data Collection Methods
This study employs non-probability sampling, specifically convenience sampling, due to the absence of a comprehensive population list.Convenience sampling is chosen as a practical means to gather data from readily available individuals, focusing on Gen Z users of TikTok in Ho Chi Minh City.Conducting a complete population survey is infeasible due to time and resource limitations, making non-probability and convenience sampling suitable methods for this research.
The estimated population residing in Ho Chi Minh City hovers around 9.2-9.3 million (Statista, 2023a;World Population Review, 2023), constituting over 9% of the entire country's population, which stands at 99.5 million (Trading Economics, 2023).As indicated in the 2021 Statistical Yearbook of the General Statistics Office (GSO), the age group spanning 15-19 years comprises approximately 1.3 million individuals, while the age group of 20-24 accounts for approximately 3.3 million individuals, and the age group of 25-29 includes approximately 5.5 million individuals.Given the reasonable assumption of a relatively uniform distribution within each age group, it can be estimated that the population of individuals within the Gen Z age group (16-26 years) resides in Ho Chi Minh City.Furthermore, laborers aged 15 and above who are actively working in Ho Chi Minh City represent 47.3% of the total population in this locality (GSO, 2021).Based on data collected from Decision Lab, the proportion of TikTok users belonging to Gen Z in Vietnam is estimated to range from 55% to more than 60 percent (Aspin, 2022).Consequently, considering the amalgamation of the above-mentioned data, the estimated total number of individuals belonging to the Gen Z age group in the Ho Chi Minh City area falls within the range of 700,000 to more than 800,000 people.
The study used the Central Limit Theorem (CLT) assumption that was first developed in 1733 by Abraham de Moivre to identify the minimum sample size (Dodge, 2008).
Where n is the number of participants required, p is the proportion matching the selected category, q is the proportion not matching the selected category, Z is 1.96 of the value from the normal distribution at a 95% confidence level, and e is the margin of error required equal to 5% (Saunders et al., 2009).
Following the same techniques applied in the primary survey, an initial pilot sample comprising 53 observations was conducted.The outcome of this pilot survey revealed that approximately 80 percent of the respondents made cosmetic purchases influenced by TikTok influencers.In light of potential concerns regarding data reliability, a second pilot sample involving 30 observations was conducted.In this subsequent pilot, only 18 out of the 30 individuals purchased cosmetics under the influence of TikTok influencers, indicating that 60 percent of the participants fell into this specific category.Using the formula specifications, the sampling for this study was obtained as follows: n = 1.96² × (18/30) × (12/30)/(0.05² ) = 370.This number indicates that the sample size required is at least 370.
All pilot surveys are to be prepared for the official survey with the design of three questionnaires (screening questionnaires, demographic questionnaires, and measurement questionnaires) for variables.
After conducting two pilot surveys to refine the official survey instrument, the data collection process commenced.The target sample size for the official survey was set at a minimum of 370 respondents.Despite encountering challenges in achieving this exact target, the survey still managed to yield a substantial number of participants, albeit slightly below the desired proportion.Although the initial proportion of respondents was less than expected, it was not statistically significant.Consequently, to meet the minimum sample size requirement, the research team increased the number of individuals to be surveyed while ensuring it remained within the team's operational control.Thus, the data collection process effectively balanced the need for an adequate sample size with practical constraints.

Data Analysis Methods
This study involves data analysis using descriptive statistics, including measures such as mean, median, standard deviation, min, max, range, and variance.The reliability assessment includes Cronbach's Alpha, and exploratory factor analysis (EFA) involves indicators like the Kaiser-Meyer-Olkin (KMO) test, Bartlett's test, Eigenvalues, and Total Variance Explained.SPSS software is utilized for Pearson's correlation analysis, regression analysis, and independent sample T-test to examine relationships and predict outcomes.
Descriptive statistics play a pivotal role in research, offering insights into data analysis.They include Mean, Median, Mode, Standard deviation, Min, Max, Range, and Variance (Hair et al., 2009).Mode frequently represents value for categorical data, while numerical data includes remaining factors.
Cronbach's alpha is a commonly used reliability metric that ranges from 0 to 1, with higher values indicating greater reliability.An alpha value greater than 0.7 is normally appropriate (Hair et al., 2009).A good item-total score correlation is indicated by a corrected item-total correlation greater than 0.3.An "alpha if item deleted" analysis is conducted to evaluate scale reliability while eliminating a particular item.Hair et al. (2009) argue that a considerable drop in alpha indicates how crucial the item is to maintaining scale dependability.
Exploratory Factor Analysis (EFA) is a statistical method used to condense a huge collection of measures into a smaller set of components by discovering similar patterns.This technique makes it easier to identify hidden connections and interconnections in the data, which improves comprehension of the data's structure.During the conduct of EFA, several key factors warrant consideration.Firstly, the Kaiser-Meyer-Olkin (KMO) test is employed to evaluate whether each factor adequately predicts a sufficient number of items, with KMO values ranging from 0 to 1. Values above 0.9 indicate superb suitability, 0.8 to 0.9 indicate great suitability, 0.7 to 0.8 Published by IDEAS SPREAD indicate good suitability, 0.6 to 0.7 indicate mediocre suitability and values below 0.5 are considered unsuitable (Hutcheson & Sofroniou, 1999).Secondly, the Bartlett test of Sphericity is used to assess meaningful connections between variables and determine the presence of significant relationships among them (Hair et al., 2009).A statistically significant result (p < 0.05) indicates sufficient correlations among variables, thereby validating the continuation of the analysis.Conversely, non-significant results suggest the potential removal of factors.Thirdly, Eigenvalues in EFA represent the amount of variation in the data explained by each factor, with higher values signifying a greater portion of the overall variation in the data.Factors with eigenvalues greater than 1.0 are deemed reasonable for further analysis, whereas values less than 1.0 indicate a potential lack of relevance or informativeness for the respective factor (Hair et al., 2009).Fourthly, the Total Variance Explained index provides a measure of how well a factor solution accounts for the representation of all variables together.An index greater than 50% indicates that the extracted factors explain over half of the total variability in the dataset.Finally, Factor Loadings indicate the degree of correlation between each variable and the identified factors.Higher loadings signify a stronger alignment between the variable and the factor, indicating the variable's representative nature of the respective factor (Hair et al., 2009).
Regression is a widely used statistical approach that investigates the relationship between a dependent variable and independent variables (Kumari & Yadav, 2018).Within linear regression analysis, it is important to consider two types of coefficients: unstandardized and standardized coefficients.Unstandardized coefficients reflect relationships in their original scales, while standardized coefficients enable a comparison of effects, irrespective of the units of measurement (Kim & Ferree, 1981).Researchers often prefer adjusted R-square over R-square to estimate the model's predictive ability more realistically, as adjusted R-square accounts for predictors and addresses potential overfitting (Johnson & Wichern, 2007;Hair et al., 2009).Another valuable tool in regression analysis is the Durbin-Watson statistic, which detects autocorrelation in residuals, ensuring the independence assumption for reliable regression results (Wooldridge, 2020).A Durbin-Watson value near 2 suggests no autocorrelation, close to 0 indicates positive autocorrelation, and near 4 implies negative autocorrelation.In the context of hypothesis testing, ANOVA is employed to compare the means of groups and assess the significance of independent variables concerning the dependent variable.Hypothesis testing involves setting up null and alternative hypotheses, with a common significance level of 0.05 and p-values below 0.05 considered statistically significant (Tabachnick & Fidell, 2019).To evaluate multicollinearity in regression models, researchers rely on the Variance Inflation Factor (VIF).VIF values exceeding 10 indicate strong multicollinearity, and researchers aim to keep VIF values as low as possible to ensure the reliability of regression estimates (Hair et al., 2009).

Descriptive Statistics
After running SPSS, the findings show that TikTok influencers have a considerable influence on both men's and women's purchasing behavior, with over 80% of users choosing to buy cosmetics under their influence but the demand for buying women is still higher than that of men.Gen Z, aged 19 to 23, is the most affected demographic, with more than 90% confessing to purchasing cosmetics from TikTok influencers.The 24-26 age range is closely followed by the 15-18 age group, which is less affected.In addition, the group with an income of about VND 1.5 million -VND 4.7 million is the group that is willing to pay more for cosmetics.TikTok, launched in Vietnam in 2016, has only gained significant traction in Gen Z since 2021 when survey data shows that the number of Gen Z users have used TikTok for 1-2 years with the most votes and most of these spend 2-3 hours surfing TikTok.There are various purposes for using TikTok including watching funny videos that received the highest number of responses, following influencers on TikTok, and new trend updates.Overall, the descriptive statistics from this study shed light on the multifaceted nature of TikTok's appeal, emphasizing its role in entertainment, influencer engagement, and trend awareness among its user base.Additionally, skincare was the most popular among Generation Z with more than 70%, followed by makeup and hair care items.For products that need more careful consideration when buying related to facial skin, Gen Z will be influenced by TikTok influencers more than lowend products.The trend of foreign beauty is also a factor, with imported cosmetic products being preferred over domestic products and price is a significant factor for Gen Z.When intending to buy a cosmetic product, Gen Z will actively search for information and consider product reviews from previous buyers rather than asking friends or relatives for advice.Based on Gen Z gender people who have not purchased cosmetic products before, the purchasing intention of males for cosmetic products in the future is less than that of females but is insignificant.Of those people who do not intend to buy cosmetics, males account for 100%.It is found that Gen Z males are more consistent in choosing not to purchase, while females are more consistent with purchase intentions.Published by IDEAS SPREAD Cronbach's Alpha Coefficient was examined with six variables: influencer's credibility (CR), influencer's physical attractiveness (PA), influencer's expertise (EX), influencer's social advocacy (SA), customer trust in influencer marketing (CT), and purchasing intention (PI).

Cronbach's Alpha
After running Cronbach's Alpha, the results obtained of Cronbach's Alpha of all independent variables and a dependent variable have values greater than 0.7 (CR=0.819,PA=0.856,EX=0.851,SA=0.887,CT=0.859, and PI=0.880) which are regarded as acceptable.Furthermore, all its observed variables have corrected items and the total correlation is greater than 0.3, which is considered satisfactory.

Factor Analysis
An exploratory factor analysis (EFA) was performed on 20 variables with 5 independent variables.The KMO coefficient was 0.905 (>0.5), indicating adequate factor analysis.Bartlett's test result was 5025.181,rejecting Published by IDEAS SPREAD hypothesis H0 and accepting hypothesis H1.The total variance explained was 73.105%, with eigenvalues of 1.432.Factor loading coefficients were greater than 0.5, with CR4 having smaller loadings.After removing the above CR4 variables, the second exploratory factor analysis yields a KMO coefficient = 0.949>0.5,indicating appropriateness for the research data.Bartlett's test yields 4784.586 with Sig.= 0.000<0.05,which means rejects hypothesis H0 and accepts hypothesis H1.However, CR1 converges in two factors, requiring elimination from the model.For the third time, tested KMO index = 0.949 > 0.5 and Bartlett's test result = 4436.320with Sig.= 0.000<0.05.Finding EX2 had a factor loading coefficient in two factors, excluding it from the study.
Following an exploratory factor analysis, there were 5 items eliminated including CR4, CR1, EX2, EX1, and CT1 respectively.The remaining variables' factor loading coefficients meet the value requirements.

Correlation Analysis
The sig. value between the dependent variable (PI) and five dependent variables (SA, PA, CT, EX, and CR) is 0.000 less than 0.05 indicating that there is a linear relationship between them.PI has a significant positive link with SA and CT and has an average level of linear correlation with them (r=0.374and r=0.490 respectively), while PI has weak links with PA,EX,and CR (r=0.294,r=0.280,and r=0.272 respectively).Despite a weak linear correlation, their association is positive.

Regression Analysis
The findings of the regression analysis present that R-square = 0.619 and an adjusted R-squared = 0.613, indicating that the independent factors included in the regression study affect 61.3% of the variation of the dependent variable.
The value DW = 2.034 is near 2, the result did not contradict the autocorrelation assumption.Based on the result obtained from ANOVA, the sig value of 0.000 was much smaller than 0.05, thus rejecting the H0 hypothesis and accepting the H1 hypothesis.Therefore, the regression model was suitable.
According to the findings of the coefficient analysis, it indicates 5 independent variables (SA, PA, CT, EX, and CR) had a T-test sig value of 0.000<0.05so they are significant in the regression model and have an impact on the dependent variable PI.Additionally, their regression coefficient has a positive sign, therefore, they have a positive Published by IDEAS SPREAD effect on the variable PI.Because the value of VIF was equal to 1, multicollinearity does not occur.Hypothetical conclusion:

Independent Sample T-Test
The Sig. F-test is equivalent to 0.001<0.05,indicating that there is a difference in variance between the two groups of men and women using the T-test results in the equal variances not assumed row to analyze.Sig.T-test equals 0.022<0.05implying that null hypothesis H0 is rejected and hypothesis H1 is accepted.As a result, there is a disparity in purchasing intention between males and females.Therefore, it is concluded: H6: There is a relationship between Generation Z's gender and their cosmetic purchase intention (Accepted).

Discussion
The study provides evidence that factors such as influencer credibility, physical attractiveness, expertise, social advocacy, and customer trust positively affect purchasing intention.Customer trust and influencer&#39;s social advocacy were found to be the most influential factors in customer buying behavior.Customer trust and influencer's social advocacy were found to be the most influential factors in customer buying behavior.In comparison to the previous study of Lin & Nuangjamnong (2022) customer trust is also the most significant factor that influences customer engagement toward purchase intention the study of Lin & Nuangjamnong (2022) in Thailand delved into the components of trust, such as trust in community members, trust in broadcast, and trust in the product.This provides a more nuanced understanding of how different aspects of trust can affect customer engagement and subsequently influence purchase intentions.These findings are somewhat surprising given the fact that both countries are located in Southeast Asia but with distinct cultural backgrounds, the similarity in findings about the influence of trust on customer purchase intentions suggests that the relationship between trust and consumer behavior may transcend cultural boundaries.Taken together, these results suggest that businesses and marketers should prioritize building trust with their customers, as well as fostering overall trustworthiness in their brand, as indicated in this study.In addition, it is interesting to compare social advocacy in this study with the study of Lin & Nuangjamnong (2022) considered social advocacy, it is proved indirectly with purchase intention that social advocacy and interactivity contribute to perceived information credibility, which subsequently influences brand credibility and attitude, leading to increased purchase intention.Although our research did not explicitly mention perceived information credibility, and the positive relationship between social advocacy and purchase intention, these results provide important insights into whether consumers perceive influencers's advocacy as credible, leading to higher purchase intentions.
The findings of this research align with previous papers, supporting the notion that physical attractiveness and influencer's credibility significantly impact purchasing intention following the research from Le (2022) in the study "How Do the Influencing Factors of Key Opinion Leaders (KOLs) on Social Networks Affect Vietnamese Consumers' Purchase Intention?".Additionally, the study highlights the importance of influencer expertise in influencing customer purchase intention, as demonstrated in previous research.The study also explores the Published by IDEAS SPREAD relationship between gender and purchase intention among Generation Z consumers, The more surprising correlation is a significant association.It showed that women consider buying cosmetics higher than men in this context.This gender effect on purchase intention highlights the need for marketers to consider gender-specific approaches when targeting the Generation Z population, particularly in Ho Chi Minh City.
Finally, these findings present new and timely insights for businesses seeking to optimize their marketing strategies, linking these findings factors.Businesses can increase their credibility among consumers by using influencerdriven social advocacy and interactive content in their campaigns.When influencers genuinely advocate for a brand and actively interact with their audience, it creates a sense of trust and authenticity.This can ultimately elevate the brand's credibility in the market.The study contributes to the existing literature on the impact of TikTok influencers on consumer behavior, specifically in the cosmetic industry and among Generation Z consumers.The findings provide valuable insights for marketers and practitioners seeking to leverage influencers on TikTok, particularly in the context of Ho Chi Minh City.Published by IDEAS SPREAD CR2 CR3 CR4 endorse.
I trust the information shared by social media influencers about the endorsed products.
The credibility of social media influencers plays a significant role in shaping my purchase intention.Influencers' promotion of social causes and initiatives influences my perception of their credibility.

Figure
Figure 1.Conceptual Framework Influencers on TikTok can possess a stronger capacity to persuade consumers when they endorse products that align with their personal lifestyle and projected image Celebrities' physical attractiveness is congruent with the product endorsed, consumers develop a positive attitude towards the product and the advertisement evaluationsThe physical attractiveness of influencers creates the illusion and drives the motivation inside the potential customers to picture what the outcome of the product is if they agree to buy the product Beauty influencers have become a powerful marketing tool for companies looking to attract younger Influencers in the cosmetic field have extensive knowledge.I believe that cosmetic Influencers provide references to customers based on their experience.I believe that users agree with the opinions of the influencer's expertise when purchasing cosmetics.I believe that users accept the cosmetic products recommended by efforts of influencers on social media platforms resonate with my values and beliefs I trust influencers who actively engage in social advocacy and use their platform to promote positive social change.

Table 1 .
Result Cronbach's Alpha of all variables Therefore, in the final results, there were still 5 factors (SA, PA, CT, EX, and CR) with 15 observed variables.The EFA analysis of Purchasing Intention (PI) yielded satisfactory, with a KMO index = 0.829>0.5, Bartlett's test yielded 773.004 with Sig.=0.000<0.05;Eigenvalues=2.951 and total variance explained of 73.780% (satisfactory).As a result, the PI factor remained unchanged for 4 observed variables.

Table 3 .
Result of regression analysis

Table 4 .
Hypotheses testing resultBased on the above two equations, it can be seen that the CT variable with the highest Beta infers that it has the highest influence on PI, followed by the variables SA, PA, EX, and CR.It can be concluded that customer trust has a significant influence on Gen Z's purchasing decisions under the influence of influencers on TikTok.