INTRODUCTION
Research motivation
In today's digital landscape, brands increasingly leverage social media to connect with local consumers According to Clement (2020), approximately 3.6 billion people were using social media globally in 2020, with projections indicating this figure will rise to nearly 4.41 billion in the near future.
In 2025, the average daily time spent on social media varies significantly by country A special study by GlobalWeb Index, conducted during the Covid-19 lockdown across 17 countries, revealed that nearly half of Internet users increased their social media usage compared to pre-lockdown levels This shift indicates a growing trend of users engaging more with social media platforms, highlighting the potential for companies to leverage social media trading as a crucial factor for success.
Celebrities play a significant role in social media content creation, positively influencing consumers' brand attitudes and purchase intentions (Schlecht, 2003) In Vietnam, influencers, particularly micro-influencers, have become preferred sources for brands seeking to attract customers, thanks to their relatable characteristics and high authenticity, despite having a smaller reach According to Advertising Vietnam (2019), the rise of micro-influencers is notable Furthermore, CafeF (2020) reported a substantial increase in global brands' budgets for branded content, soaring from $3.6 billion in 2016 to $13.7 billion in 2020, with projections to exceed $25 billion by 2025 This trend indicates significant opportunities for influencer marketing, which is also gaining momentum in the Vietnamese market.
2 increasing to 35% compared to 2016 Businesses use many influencers in many different areas, operating on multiple channels such as YouTube, Facebook, Instagram and TikTok
The Vietnamese cosmetics market is rapidly expanding and becoming increasingly competitive, with a notable shift towards online purchasing As of 2020, the market value reached $2.35 billion, with skincare products leading in consumer spending at 8.3 trillion VND Online retail visits surged by 10.63% between Q2 and Q3 of 2020, indicating a strong trend in e-commerce for cosmetics Projections suggest that the market could reach $2.3 billion by the end of 2018, driven by a growing middle class expected to number 33 million by 2020 According to Kantar Worldpanel's Insight Handbook, skincare remains the most popular category, with significant increases in average spending on products such as makeup removers, moisturizers, and sunscreens Given this data, there is a pressing need for brands to enhance their communication strategies to attract customers and gain a competitive edge.
According to the 2020 Vietnam cosmetic market report, the majority of Vietnamese customers rely on online sources for cosmetics information, primarily through platforms like Facebook, websites, and YouTube Digital influencers have emerged as a key channel for effectively communicating with consumers in the cosmetics industry A 2018 report by Celebrity Intelligence indicates that digital influencers have surpassed traditional celebrities, becoming the top choice for brands and significantly impacting consumer opinions and purchasing decisions These influencers play a vital role in shaping beauty trends, models, and product preferences Additionally, leveraging digital influencers has proven successful for cosmetic suppliers, as the same report highlights that investing $1.3 in influencer marketing can yield substantial returns for brands.
The average profit from utilizing digital influencers in Vietnam's cosmetics industry is $11.8, highlighting the effectiveness of celebrities on social media for advertising and public relations As a result, the trend of employing beauty bloggers to boost brand awareness and drive product sales is rapidly gaining popularity in Vietnam.
In Vietnam, the use of digital influencers in marketing is a powerful and evolving method for connecting businesses with consumers; however, there is a lack of analytical studies focused on optimizing their effectiveness This research paper aims to explore the impact of digital influencers on purchase intentions within the cosmetics industry in the context of the technological era.
Recent studies have explored the impact of celebrity characteristics—such as expertise, attractiveness, and trustworthiness—on purchase intentions, as identified by Ohanian (1990) and further examined by researchers like Chaovalit (2014) and Park & Lin (2020) However, there is a notable lack of research focusing on digital influencers, a rapidly emerging and popular form of celebrity in today's market With the rise of the 4.0 technology era, digital influencers are proving to have a more significant impact than traditional celebrities, effectively enhancing company revenues and expanding brand visibility (Le, 2021).
Digital influencers have emerged as a new form of celebrity in the technology 4.0 era, yet research on their impact on purchase intention remains limited A 2019 study by Ambarwati and colleagues focused on the effects of electronic word of mouth (eWOM) specifically on Wardah cosmetics purchase intention, neglecting the characteristics of digital influencers Similarly, Handayani, Evelina, Sylvina, and Lestari (2019) examined the content strategies of digital influencers to engage followers but did not address their influence on purchasing decisions In contrast, Torres, Augusto, and Matos (2019) explored the direct impact of digital influencers on purchase intention, highlighting a gap in the existing literature.
Group authors prioritize the appeal of digital influencers and the alignment between the influencer and the brand, while neglecting to address the compatibility between the celebrity and the product.
This research aims to address the existing gap by examining how the characteristics of digital influencers—such as expertise, attractiveness, trustworthiness, and likability—impact consumer perceptions, particularly in the context of cosmetic products in Vietnam Additionally, the study will explore the influence of product fit to offer new insights into this topic.
Research objectives
In this study, I concentrate on the following specific objectives as below:
– To research the effect of digital influencers on the customers’ purchase intention in cosmetics industry
– To propose the solution for firms to attract potential customers by using digital influencers.
Research question
– What are the characteristics of digital influencer influencing to the buying intention of customers?
– Which is the most characteristic effect to purchase intention?
Research scope
This research, conducted in Vietnam from October 2020 to May 2021, examines the purchasing intentions of Vietnamese consumers for cosmetic products, specifically focusing on facial skincare items The study highlights the influence of digital influencers on potential buyers, emphasizing the growing popularity and increasing average spending on skincare products in the country.
Research structure
This chapter introduces the practical and theoretical motivation of research, identifies the research objectives, research questions and research scope
This chapter offers a conceptual definition and foundational theory regarding the terminology of the subject Additionally, it presents the hypothesis and elaborates on the research model, drawing from prior studies and addressing existing research gaps.
This chapter describes the process of research, the way to design sample, questionnaires and method to collect data
This chapter focuses on analyzing the reliability and validity of data collected, regression and identifying the results of tesing hypothesis, providing which hypothesis is supported or not supported
This chapter summarizes the research findings and discusses their implications Additionally, it highlights the limitations of the study, offering recommendations for future research.
LITERATURE REVIEW
Review of related definition
2.1.1 Social media and influencer marketing
Social media is the widely used term refer to the Internet-based networks from which users can communicate and selectively view themselves (Carr & Hayes, 2014) However, Boyd
Social media, as defined by Ellison (2008), consists of websites that allow users to create profiles and connect with each other, while Kietzmann et al (2011) describe it as web-based programs that facilitate design, distribution, communication, and grouping Manning (2014) emphasizes that social media represents new media forms characterized by interactive participation In the context of the technological era 4.0, social media provides an accessible platform to reach potential users According to Tankovska (2021), the three most popular social media platforms globally are Facebook with 2.7 billion active users, YouTube with 2.9 billion, and WhatsApp with 2 billion This vast user base significantly influences consumer behavior, impacting their daily lives.
In 2008, Brown & Hayes defined influencer marketing as a third party that significantly influences customer purchasing decisions without being directly accountable for them They emphasized that this marketing technique utilizes key individuals or opinion leaders to enhance brand recognition and drive consumer purchases, ultimately boosting a company's revenue Carter (2016) further described influencer marketing as a rapidly growing industry focused on promoting products and increasing brand awareness through content shared by influential social media users Today, influencer marketing stands out as one of the most powerful strategies for promoting products and services.
According to Hall (2015), the intrinsic qualities of influencers are essential for attracting brands and advertisers One advantage is that brands can opt for more affordable influencers rather than paying high fees for famous celebrity endorsers Additionally, social media influencers often specialize in specific niches, which enhances their credibility when partnering with brands aligned with their expertise, leading to greater customer trust and support (Hall K., 2016) Thus, while influencer marketing is an effective strategy for boosting brand awareness and customer appeal, its success hinges on the careful selection of influencers, making the choice of an appropriate celebrity crucial for brand and product development.
Digital influencers are increasingly defined by their ability to impact perceptions and behaviors within social media communities According to scholar Morris (2009), digital influence is characterized by the capacity to effect change through social networking These influencers gain trust from their online followers and enhance the reputation of brands or products By leveraging social media communication, digital influencers effectively engage and connect with their audience.
Digital influencers serve as brand communicators for firms, often compensated with free products, exposure, and small cash benefits (Duffy, 2015) These influencers aim to enhance engagement by encouraging their followers to choose products without any cost (Rocamora, 2018).
In 2019, it was noted that digital influencers serve as intermediaries who leverage their extensive and relevant relationships to engage organizational stakeholders Through their interactions, personal appearance, and content shared on social media, these influencers effectively manipulate perceptions and behaviors within their audience.
8 as the person who convey the media message of company, bring the attractiveness of this brand and product to customers through technology platform such as Youtube, Facebook or Instagram
According to Wielki (2020), digital influencers can be categorized into seven distinct types based on their follower count, which significantly impacts their reputation, influence, and potential income from brand collaborations These categories range from social celebrities with over 5 million followers to nano influencers with 1,000 to 10,000 followers The influencer tiers include mega influencers (1 million to 5 million followers), top influencers (over 500,000), macro influencers (100,000 to 500,000), middle-level influencers (20,000 to 100,000), and micro influencers (fewer than 20,000 followers) Each type of influencer adopts specific strategies to align the brand and product messaging with their target audience, tailored to their unique organizational objectives.
In 2023, digital influencers in Vietnam have emerged as powerful entities, rivaling traditional media channels and driving significant success for brands (Brandsvietnam, 2019) By effectively leveraging influencer marketing, companies can successfully engage with young consumers in Vietnam, reaping substantial benefits This research focuses on examining the impact of digital influencers within the cosmetic industry, specifically regarding Vietnamese consumers' purchasing intentions for skincare products.
2.1.3 Definition and segment of cosmetic product
The makeup market offers a diverse range of cosmetic products, featuring various textures, colors, and appearances Defined as items meant to be applied to the human body for cleansing, enhancing attractiveness, beautifying, or altering appearance, cosmetics play a significant role in personal expression and beauty routines.
According to Regulation (EC) No 1223/2009 of the European Parliament and Council, cosmetics serve to cleanse, fragrance, and enhance personal appearance Beauty products not only protect and correct the skin but are also perceived by many as transformative tools that improve features such as the face, eyes, and lips (Cellania, 2014) Furthermore, Benson (2018) suggests that cosmetics play a significant role in boosting individual confidence and allowing self-expression.
The cosmetic industry, as identified by Lopaciuk and Loboda (2013), is categorized into five main types: skincare, haircare, makeup, fragrance, and toiletries, each with distinct features Skincare products, such as cleansers, sunscreens, toners, makeup removers, and moisturizers, are designed to maintain healthy and attractive skin In contrast, haircare products, including shampoos and conditioners, focus on enhancing hair cleanliness and appearance Makeup products, like foundations, lipsticks, and mascaras, aim to improve the consumer's overall look Toiletries, such as toothpaste, care for the body, while fragrances are scented liquids derived from flowers for skin application In Vietnam, Kantar Worldpanel (2019) reports that skincare is the largest segment of the cosmetics market Consequently, this paper focuses on facial skincare products due to their rarity in research and growing popularity in Vietnam.
Before exploring the various types of cosmetics, it's essential to understand the pricing tiers that categorize them within the multi-billion-dollar industry Cosmetics are generally divided into two main segments: drugstore and high-end products (Lopaciuk & Loboda, 2013) These segments can be further classified into three categories: low-priced drugstore items (under 300,000 VND), mid-priced high-end products (ranging from 300,000 to 1,000,000 VND), and luxury high-priced cosmetics (over 1,000,000 VND) (Mai, 2014) Notable brands in the low-priced drugstore category include Maybelline, Loreal, Nivea, and Wet n Wild.
When exploring skincare and beauty brands, notable options include budget-friendly labels like Black Rough, Innisfree, Etude House, and Missha from Korea, as well as Naturie, Shiseido, and Hada Labo from Japan For mid-range products, consider M.A.C, Vichy, Yves Rocher, and Clinique from the US-UK, along with The Body Shop and Laneige from Korea At the high-end luxury tier, brands such as Chanel, Lancome, and Estee Lauder from the US-UK stand out, alongside premium Japanese brands Ski-II and Shu Uemura, as well as Korean favorites Su:m, Ohui, and Sulwhasoo.
The distinction between drug-store and high-end cosmetics is defined not only by price but also by unique features such as formula, pigmentation, and packaging quality While low-end products are affordable and offer sufficient quality, they often rely on cheaper synthetic ingredients, which can diminish the overall application experience In contrast, high-end cosmetics utilize superior quality ingredients, resulting in greater customer satisfaction due to enhanced pigmentation, natural raw materials, and overall performance that surpasses that of drug-store options.
2.1.4 Consumer buying process and Purchase intention
The consumer buying process consists of five stages: need recognition, information search, evaluation of alternatives, purchase decision, and post-purchase behavior (Kotler & Keller, 2008) It begins with need recognition, where consumers identify a problem or desire triggered by internal or external stimuli To effectively address these needs, marketers must gather insights from various customers to understand the conditions that prompt specific needs and create targeted campaigns The next stage, information search, reveals that while manufacturers provide substantial product information, consumers often prefer personal recommendations from family, friends, colleagues, and digital influencers, as these sources are perceived as more trustworthy and effective.
Review of relevant theoretical models
In the era of globalization, digital influencers significantly impact companies aiming to attract potential customers As technology advances, researchers are continually exploring the evolving role of these influencers in marketing strategies.
Numerous theories have been developed to understand and predict consumer behavior, with a focus on effective models that examine consumer habits influenced by public figures, such as the Source Credibility Model.
Figure 2 1 Source credibility model (Ohanian, 1990)
According to Ohanian (1990), source credibility theory suggests that individuals are more easily persuaded by reliable figures, encompassing three key components: expertise, trustworthiness, and attractiveness Expertise refers to the knowledge and skills possessed by influencers, which can enhance consumer trust in product quality (Hovland et al., 1953; Amos et al., 2008) Trustworthiness relates to the perceived integrity and reliability of a celebrity, influencing consumers' purchase intentions, especially in the cosmetic industry (Friedman et al., 1976; Hovland & Weiss, 1951) Wei and Lu (2013) further define trustworthiness as the public's confidence in the reliability of information provided Additionally, attractiveness plays a significant role in enhancing the effectiveness of a company's message on social media.
(McGuire, 1969) The scholar Wei & Lu (2013) also found that attractiveness become a means to enhance the consumers by appearance, the similarity and the likeability of someone
Research by Karlins and her colleague (1970) indicates that a source's credibility significantly enhances its persuasive power, as highlighted by cognitive response theory, which focuses on the impact of message ideas They further noted that an individual with an optimistic view of a message, despite lacking credibility, can be more influential than a highly credible source (Aaker & Myers, 1987) Conversely, Hannon and Coney (1982) found that individuals with a negative predisposition may undermine the credibility of a high-status source, as their negative outlook can hinder message acceptance The evaluation of source credibility, particularly in the context of celebrity endorsements, is influenced by attitudes and behavioral intentions, making the credibility model subjective and varied across individuals (Berscheid et al., 1971) Ohanian (1990) identified three key components of the source credibility model: attractiveness, expertise, and trustworthiness, which serve as essential criteria for selecting influencers for brands and products.
The reason choosing model Source credibility model as the theoretical framework
The various consumer behavior models presented illustrate different aspects of purchasing decisions Specifically, the source credibility model is relevant for examining how digital influencers impact customer purchasing intentions for cosmetic products in Vietnam Consumers who engage with technology platforms tend to favor products recommended by influencers, regardless of price range, from affordable to luxury cosmetics The trust followers place in these influencers, based on their reputation, significantly influences their purchasing choices, leading them to select products that resonate with them from credible sources.
This article explores the impact of celebrities on consumer buying intent, building on previous research that assessed the influence of celebrity endorsements on purchase intentions By focusing on the characteristics of influencers, the study proposes the source credibility model as a suitable framework for analyzing how celebrity attributes affect consumer behavior.
Hypothesis development and conceptual model
The research model incorporates three key constructs from the Source Credibility Model: expertise, trustworthiness, and attractiveness Additionally, the study aims to evaluate the impact of two new characteristics of digital influencers—likeability and product fit—alongside the perceived quality by customers, thereby enhancing the overall model.
Figure 2.2 Conceptual model (by authors)
Here is the construct of definition:
1 Expertise The degree of acquiring the perceptions, knowledge, skills
2 Attractiveness A means to enhance the consumers by appearance, the similarity and the likeability of someone (Wei & Lu, 2013)
3 Trustworthiness The dignity, belief, and faith which are belonged to the celebrity
4 Likeability “Affection for the source as a result of the source’s physical appearance and behavior” (Erdogan, 1999)
5 Product fit The congruency between the reviewer and reviewed product
Expertise refers to the level of knowledge, skills, and perceptions acquired (Hovland et al., 1953) In the context of cosmetics, which are closely linked to health care, a digital influencer's expertise significantly impacts consumer purchasing behavior In Vietnam, the prevalence of counterfeit products creates insecurity among consumers (Bao Quang Binh, 2017), leading them to be cautious when considering cosmetic recommendations from peers and celebrities Consequently, a celebrity's expertise becomes a crucial factor in gaining consumer trust for product reviews and influencing their buying decisions.
Previous research highlights the crucial role of expertise in shaping the impact of influencers on purchase intention Gupta et al (2015) established a significant correlation between expertise and consumer purchasing decisions In today's market, celebrity endorsements are vital for business growth, particularly among Indian consumers, where leveraging public figures can enhance brand recognition and solidify its position in customers' minds.
However, Wiedmann & Mettenheim (2019) examined the success of influencers for online influencer events basing on luxury fashion brand Factor expertise as the key role for
17 determining the brand satisfaction, brand image and brand trust and from that, the impact to purchase intention could clarify more by these factors
In a 2008 study published in the Journal of International Consumer Marketing, researcher Pornpitakpan identified the dimension of celebrity credibility, highlighting the impact of expertise on the purchase intentions of Singaporean undergraduate respondents By featuring four Asian celebrities, the research aimed to facilitate viewer engagement and stimulate interest The findings revealed that the expertise of these celebrities significantly influences consumers' intentions to buy, underscoring their effectiveness in promoting products.
Therefore, this research hypothesizes that:
Hypothesis 1: Expertise (EX) of digital influencers positively influences the purchase intention
According to Wei & Lu (2013), attractiveness plays a crucial role in enhancing consumer engagement through appearance, similarity, and likability In the realm of digital influence, physical attractiveness significantly impacts an influencer's ability to reach and captivate an audience This characteristic not only draws in viewers but also encourages them to consider purchasing cosmetic products.
Previous research has explored the influence of attractiveness on purchasing intentions, with Widyanto & Agusti (2020) highlighting the significant role of beauty influencers in shaping Generation Z's purchase decisions, particularly in the cosmetics sector in Indonesia The findings indicate that attractiveness notably impacts Generation Z's intent to buy cosmetic products, underscoring the importance of this demographic as a key market in a developing country like Indonesia This study reveals the potential of Generation Z consumers in the cosmetic industry, emphasizing the need for further research in this area.
Besides, the researcher Pornpitakpan (2008) said that his research results of attractiveness factor of celebrities have significant impact on purchase intention In this technology era
4.0 appearance of Internet entertainers also brings various benefit for them to attract new customers for the brands Thanks to the research, the testing of this factor could be clarified in significant results
Thus, the hypothesis would be:
Hypothesis 2: Attractiveness (AT) of digital influencers positively influences the purchase intention
Trustworthiness, defined as the dignity, belief, and faith attributed to a celebrity (Friedman et al., 1976), plays a crucial role in consumer purchasing decisions For consumers to feel confident in their product choices, they must trust the digital influencer promoting it In Vietnam, where counterfeit products are prevalent and often unreliable (Bao Quang Binh, 2017), digital influencers enhance trustworthiness, reassuring consumers that they are purchasing genuine products.
A study by Park & Lin (2020) revealed the attitudes of Chinese users towards live streaming shopping facilitated by internet celebrities As streaming media platforms gain popularity in the digital era, the research found that the trustworthiness of Wanghong, or internet celebrities, significantly influences consumers' buying intentions in China.
Research by Pornpitakpan (2008) highlights the significant impact of celebrity trustworthiness on purchase intention, particularly regarding Internet public figures Additionally, a study conducted in Thailand revealed that the credibility of beauty bloggers significantly influences Thai consumers' intentions to use cosmetics Chaovalit (2014) further emphasized the role of beauty bloggers in enhancing credibility within the cosmetics industry Consequently, the trustworthiness of influencers is a crucial factor to examine in this research.
Therefore, the hypothesis would be:
Hypothesis 3: Trustworthiness (TR) of digital influencers positively influences the purchase intention
Research on influencer marketing reveals differing perspectives on likeability and attractiveness Some studies, such as those by Kahle & Homer (1985), equate the two, while others, including O’Mahoney & Meenaghan (1998), treat likeability as an independent factor influencing effectiveness Erdogan (1999) defines likeability as the affection for a source stemming from their physical appearance and behavior, highlighting its distinct role in marketing effectiveness.
This research distinguishes likeability from attractiveness, focusing on likeability as a measure of a digital influencer's coverage or virality, while attractiveness pertains solely to their physical appearance Unlike traditional influencers, where measuring a celebrity's fanbase can be challenging, digital influencers' coverage is quantifiable through metrics such as likes, shares, and subscriptions on social media platforms As Patel (2019) notes, the engagement and reach of digital influencers can be analyzed in detail, including data categorized by age, region, and lifestyle This information enables companies to assess the popularity of digital influencers effectively, facilitating the selection of suitable influencers for brand promotions Overall, this variable is crucial for understanding the evolving landscape of digital influence.
In their 2017 study, Gauns, Pillai, and colleagues investigated how celebrity endorsements influence customer purchase intentions, specifically focusing on Goan consumers The research highlighted that likeability is a key attribute of celebrities that significantly impacts purchasing decisions.
The study explores the purchasing intentions of Goan consumers in India, highlighting the influence of endorsement credibility on brand attitude and purchase intention, as demonstrated by Vien et al (2017) in Malaysia Their findings indicated that while the likeability of endorsers did not affect brand attitude, it positively influenced purchasing intentions for badminton players However, there is a lack of research on the relationship between the likeability of digital influencers and purchase intention This study aims to investigate this relationship within the Vietnamese context, assessing the impact of digital influencers on consumer purchasing behavior.
This study proposed the following hypothesis:
Hypothesis 4: Likeability (LI) to digital influencers positively influences the purchase intention
Product fit hypothesis recommend that the review of influencers is more effective when having the link between characteristics of influencers and products they review (Kim & Na,
The alignment between an influencer's image and the properties of the product they promote is crucial for enhancing the credibility of reviews Research by Kamins (1990) highlights that product fit, or the congruence between the reviewer and the product, significantly influences consumer purchase intentions, as noted by Speed and Thompson (2000) This indicates that a strong connection between digital influencers and the products they endorse is vital in guiding consumers' decisions in the cosmetics market Furthermore, leveraging various digital influencers tailored to different product segments can effectively boost product promotion and enhance perceived value.
DATA AND METHODOLOGY
Research process
Here is the process of this research:
Reviewing the related research and conceptual theory
Identifying the objectives of research, research methodology, research questions, research scope
Depending on the previous study to build the hypothesis and finalize conceptual model
Designing the survey questionnaire and sample of survey
Conducting pilot test and adjusting questionnaire after receiving feedback of pilot test
Analyzing data, check the supported or not supported hypothesis and conceptual model
Conclusion the findings and propose some implication and limitations
Sample design
According to Hair et al (2009), the optimal sample size for a survey is determined by multiplying the number of questions by five In this study, which involved measuring seven variables through 23 questions, the minimum required sample size is calculated to be 115 (23 questions x 5) Increasing the number of responses will enhance the accuracy of the scale items.
This study utilized non-probability sampling, as defined by Leary (2004), which is suitable for hypothesis testing in research The surveys were conducted primarily in Hanoi and other major provinces and cities, targeting diverse consumer groups who follow digital influencers for skincare cosmetic purchases Due to the impact of Covid-19, the survey was predominantly administered online via Google Forms, resulting in a total of 235 responses.
The target respondents for this survey include individuals who have purchased or intend to buy facial skincare products based on reviews from digital influencers The selection process was random and not restricted by demographic criteria After filtering the data, responses from individuals who have never purchased facial skincare products were excluded, resulting in a total of 187 valid survey responses.
Measurement scale
This research adapts the measurement framework established by Hair et al (2009) to focus on digital influencers, utilizing the source credibility model proposed by Ohanian (1990), which encompasses three key variables: Expertise (4 items), Attractiveness (5 items), and Trustworthiness (3 items) A 5-point Likert scale will be employed for the survey questionnaire, ranging from strongly disagree to strongly agree Additionally, the study introduces the variable "Likeability," based on Tripp et al (1994), which includes 4 items and also utilizes the 5-point Likert scale Finally, the research will examine the independent variable of product fit.
The study utilizes 24 variables based on Kamins (1990), measured using a 5-point Likert scale with three items Additionally, the purchase intention variable, as defined by Tingchi-Liu & Brock (2011), consists of four items Detailed measurement scale items for these variables are provided in the table below.
Strongly disagree Disagree Neutral Agree Strongly agree
No Variable Coding Question Source
EX1 The digital influencers are knowledgeable about the product
The digital influencers are qualified with the certificate about cosmetics
The digital influencers are experienced in using a variety of facial care products
The digital influencers are skilled at using cosmetics properly and well Example: Makeup skills, using cosmetics for facials - skincare
I like seeking about the product through the recommendation of digital influencers with beautiful face, smooth skin Ohanian
I like seeking about products through the recommendation of attractive digital influencers
No Variable Coding Question Source
I like seeking about products through the recommendation of elegant digital influencer
I like seeking about the product through the recommendation of digital influencer with classy style
I like seeking about the product through the recommendation of digital influencer with sexy style
I think digital influencer is a trustworthy person when recommending a product (according to your personal sensing)
I think digital influencer is a honest person when recommending a product
I think digital influencer is a dependable person when recommending a product
I like exploring about the product through the introduction of likable digital influencer
I like exploring about the product through the introduction of pleasant digital influencer
I like exploring about the product through the introduction of popular digital influencer
I like exploring about the product through the introduction of agreeable digital influencer
Digital influencers' image is suitable for the product they recommend
No Variable Coding Question Source
The link between digital influencers' career field and the product they recommend
19 PF3 Digital influencers support to increase the value of the product
PI1 Products recommended by digital influencers caught my attention
Products recommended by digital influencers effect my buying intention
I buy this product because my favorite digital influencers recommending the product
I bought this product because I trust the recommendation of the product review of digital influencers.
Data analysis methodology
The descriptive analysis categorizes demographic information, including gender, age, occupation, monthly income, type of product, product segment, and frequency of use This data will be visually represented through text and charts for clarity and ease of understanding.
The data collected would be analyzed by SPSS software
Cronbach's alpha is a commonly used metric for assessing reliability in research (Hair et al., 2009) This study employs Cronbach's Alpha to evaluate the reliability of measurement items for each variable The alpha value indicates the strength of the relationship between individual items and the overall score, with values ranging from 0 to 1 According to researchers George & Mallery (2003), this value serves as a critical indicator of consistency in measurements.
A reliability score under 0.5 indicates low reliability, while a score above 0.9 signifies strong reliability Nunnally & Bernstein (1994) emphasize the significance of the correlation between individual items within a construct and the total score They also note that item-total correlations should ideally be greater than 0.3; items with correlations below this threshold should be excluded from the analysis.
According to Hair et al (2009), Exploratory Factor Analysis (EFA) is a technique that condenses a large set of variables into a smaller, more manageable group while preserving the essential characteristics of the original variables To effectively conduct EFA, certain prerequisites must be adhered to.
• Factor loading is accepted when > 0.5
• Kaiser – Meyer – Olkin (KMO) of Sampling Adequacy > 0.5 In case the KMO value is under 0.5, that means no distinction and validity between these variables
The independent sample T-test, also known as Student’s unpaired t-test, is an inferential statistical method used to compare observations from two unmatched study groups (Hill, 2006) In contrast, Student’s paired t-test is applicable when two groups of the same size are matched in ways other than the experimental intervention (Hill, 2006) This research focuses on the gender groups "male" and "female," which are not matched, making the independent sample T-test the appropriate choice for analysis.
Mishra et al (2019) explain that this test involves a continuous normally distributed variable, referred to as the test variable, alongside a categorical variable that has two categories, known as the grouping variable The analysis includes calculating the mean, standard deviation, and the number of observations for each group.
In this study, the significance degree will be calculated using a sample group of 28 participants Levene’s test will first determine the significance level, with the P-value being utilized for the Student’s unpaired t-test A P-value greater than 0.05 indicates that the results are insignificant, suggesting equal variances between the two groups Conversely, a P-value less than 0.05 signifies unequal variances.
3.4.2.4 One-way Analysis of variance (ANOVA test)
The Independent sample T-test is used to compare observations from two research groups, whereas Analysis of Variance (ANOVA) is employed for comparing observations from three or more groups (Mishra et al., 2019) According to Larson (2008), the ANOVA test is defined as a statistical method for this purpose.
“ a kind of statisticsal technique for analyzing variation in a response variable (continuous random variable) measured under conditions defined by discrete factors (classification variables, often with nominal levels)”
One-way ANOVA is utilized when analyzing a single independent variable, while two-way ANOVA involves two independent variables (Mishra et al., 2019) This statistical test evaluates one continuous dependent variable in relation to at least three categorical independent groups, aiming to determine the significance of the P-value Similar to the T-test, a P-value greater than 0.05 indicates no significant difference and equal variance, whereas a P-value less than 0.05 signifies a significant difference and unequal variance.
Simple linear regression is a statistical method used to predict the value of a response variable based on an explanatory variable (Hill, 2006) It effectively measures the relationship between independent and dependent variables, allowing for insightful forecasting The regression function illustrates this relationship clearly.
In that case, the abbreviations of each part will be shown:
𝛽 1 : Coefficient between dependent variable and independent variable
In this model, it is essential for the author to assess the P-value to determine the support for the hypothesis; a P-value greater than 0.05 indicates no support for the hypothesis, while a P-value less than 0.05 signifies support Following the P-value analysis, the researcher should examine R-squared (R2) and Adjusted R-squared (Adjusted R2) According to Hill (2006), R-squared measures how much variance in the independent variable accounts for the variance in the dependent variable, with higher R2 values indicating a stronger explanatory power A model is considered a good fit when R2 exceeds 50% Adjusted R2 also assesses model fit, adjusting for the number of independent variables; an increase in Adjusted R2 upon adding a new independent variable suggests that the variable effectively contributes to explaining the dependent variable.
In regression statistics, the Variance Inflation Factor (VIF) is essential for assessing multicollinearity among variables A VIF value below 2 indicates that multicollinearity is not a concern, allowing the variable to remain in the model, while a value of 2 or higher suggests the need for further evaluation or removal of the variable.
ANALYSIS RESULTS
Descriptive statistics
This section presents the analysis and findings from the survey data collected through both online (via Google Forms) and offline methods A total of 235 participants expressed their willingness to participate, but only 187 responses met the criteria for qualification The qualified data indicates that these participants regularly use facial skincare products and have purchased or intend to purchase cosmetics based on recommendations from digital influencers The details of the participants involved in the survey are outlined below.
Gender: Among 187 respondents, there are 89.3% of respondents related to 167 persons were women while 10.7% of ones related to 20 persons were man
From 16 to 22 years old From 23 to 30 years old From 31 to 40 years old Over 40 years old
Age: In a group 187 respondents, the largest percentage of age accounts for 48.1% (90 responses), from 23 to 30 years old Meanwhile, the second rank belongs to group age from
16 to 22 years old, took 47.1% (88 responses) Two lowest group of age is from 31 to 40 years old and over 40 years old, accounted for 4.3% and 0.5%, respectively, at 8 and 1 responses respectively
The survey on marital status revealed three categories: dating, married, and single The majority of respondents, 112 individuals or 59.9%, identified as dating, while 58 respondents (31%) reported being single, and 17 respondents (9.1%) were married.
The survey reveals that students constitute the largest occupational group, comprising 60.4% with 113 respondents Officers follow as the second largest category, representing 27.8% with 52 individuals Additionally, freelancers account for 6.4% with 12 participants, while the sales, customer service, and receptionist roles collectively make up the remaining 5 respondents.
Business owner/Self-employed Freelancer Officer Sales/Customer service/Receptionist
32 people (2.7%), and the last is the group of Teacher and Business owner / Self-employed, respectively 3 people (1.6%) and 2 people (1.1%)
Due to COVID-19 and limited access to online forms, the majority of responses regarding living areas came from Northern Vietnam, comprising 92% (172 responses) In contrast, only 8 participants from Southern Vietnam and 7 from Central Vietnam contributed, representing 4.3% and 3.7% of the total responses, respectively.
The income distribution among the surveyed individuals is categorized into four distinct groups A total of 102 participants were analyzed, revealing that 54.5% earn less than 5 million VND The second income group, earning between 5-10 million VND, comprises 24.1% of respondents, totaling 45 individuals The third group, with incomes ranging from 11-20 million VND, accounts for 16.6% of the participants Lastly, the smallest segment, those earning over 20 million VND, represents only 4.8% of the total.
11-20 million VND 5-10 million VNDOver 20 million VND Under 5 million VND
Product segment: The product segment is divided into 3 main groups: Drug-store product
Products are categorized into three price ranges: budget products under 300,000 VND, mid-end products priced between 300,000 VND and 1,000,000 VND, and high-end products exceeding 1,000,000 VND Among these categories, drug-store products are the most favored, receiving 97 responses, which represents 51.9% of the total Mid-end products follow with 76 participants, accounting for 40.6%, while high-end products attract the least interest.
14 high-end users, accounting for 7.5%
Drug-store product (Price below 300,000 VND/product)
High-end product (Price above 1,000,000 VND/product)
Mid-end product (Price from 300,000 VND - 1,000,000 VND/product)
Product type: Regarding the product type, the author lists 10 product types, namely
In a recent survey on skincare product usage, respondents selected from a range of items, with the top five most popular being cleanser, makeup remover, sunscreen, moisturizer, and peeling The cleanser emerged as the most widely used product, with 185 participants (98.93%) reporting its use Following closely were makeup remover with 164 responses (87.7%), sunscreen at 154 (82.35%), moisturizer at 147 (78.61%), and peeling at 144 (77.01%) Other products saw significantly lower usage, including toner used by 137 participants (73.26%), facial masks by 134 (71.66%), and facial oils or serums by 113 (60.43%) Spot corrector was utilized by only 70 users (37.43%), while anti-aging cream was the least favored, indicating a clear preference for basic skincare essentials among participants.
Scale measurement of constructs
This study aims to assess the reliability of the scale and eliminate any unreliable measuring items by utilizing the Cronbach’s Alpha test This method is employed to evaluate the reliability of both dependent and independent variables.
If the Cronbach alpha value ≥ 0.3, the items will reach the standard (Nunnally & Bernstein,
1994) The results of Cronbach alpha will be presented as below:
Scale Mean if Item Deleted
Scale Variance if Item Deleted
Cronbach's Alpha if Item Deleted
Scale Mean if Item Deleted
Scale Variance if Item Deleted
Cronbach's Alpha if Item Deleted
The above results illustrate that by the Corrected Item-total correlation >0.3, the results of Corrected Item-total correlation is satisfied, Cronbach alpha of each group items mostly
>0.8, this means that the measurement scale is very good
To enhance the reliability of the scales, two items will be removed based on Cronbach's Alpha analysis The item EX2 will be deleted as its removal increases the Cronbach's Alpha for the Expertise group to 0.803, exceeding the acceptable threshold of 0.755 Similarly, item LI1 will be removed to boost the Likeability group's Cronbach's Alpha to 0.865, surpassing the benchmark of 0.825 Consequently, the revised scales will consist of 3 items for Expertise, 5 items for Attractiveness, 3 items for Trustworthiness, 3 items for Likeability, 3 items for Product Fit, and 4 items for Purchase Intention.
Table 4 2 Exploratory factor analysis results of independent variables
Kaiser-Meyer-Olkin Measure of
The study involved 17 independent variables assessed for reliability and validated through exploratory factor analysis (EFA) The Kaiser-Meyer-Olkin (KMO) Measure of Sampling Adequacy was found to be 0.859, indicating that the data is suitable for factor analysis, with a significance value of 0.000, which is less than the threshold of 0.05 (Hair et al., 2009) The analysis revealed that five factors accounted for 73.6% of the total data variation Additionally, the varimax rotation of the rotated component matrix confirmed that the 17 observation items were effectively grouped into these five variables.
Table 4.4 Exploratory factor analysis results of dependent variables
Kaiser-Meyer-Olkin Measure of Sampling Adequacy 0.777 0.5