YouTube vloggers' impact on brand equity and purchase intention for low-priced brands

(1)

YouTube vloggers’ impact on brand equity and purchase

intention for low-priced brands

MSc. in Business Administration - Marketing Track

Master Thesis - First Draft

Supervisor: Frank Slisser

Submitted on June 22, 2018

Yinqian Jiang

11770007

(2)

2

Statement of Originality

This document is written by Student Yinqian Jiang who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

3

Abstract

Many researches have conducted focusing on social media platforms and their general influences on brands. However, brands need to understand a platform’s particular attributes in order to better perform influencer marketing, since social media platforms are different on their forms, user profiles, use purposes, etc. This paper focuses on the effects of the most widely-used video-sharing platform, YouTube, on customer perceived brand equity of the advertised brand and the related purchase intention. A survey is conducted and the results indicate that the effect of YouTube videos on brand equity is different based on the forms of advertising information and brand fame. Subjective information has a greater impact on brand equity than objective information, and the brand equity of unfamous brands is more affected by advertising information than famous brands. Also, purchase intention is positively associated with brand equity.

Key words: YouTube, subjective information, brand fame, influencer marketing, brand equity, purchase intention

(4)

4

1. INTRODUCTION

BBC claimed that more than three billion people have registered on at least one social media platforms (Hogenboom, 2018). It also stated that in the United States, the average time adults spend per day on social media platforms is more than two hours (Hogenboom, 2018). Another report even stated that the total Chinese population spend 39.8 billion hours on social media in the half year of 2017, which is enough to build 109 pyramids (China Daily, 2017). Indeed, social media currently is the biggest buzz on the internet (Jin & Phua, 2014), and statistics show that people spend an hour every day just on Facebook (D’Onfro, 2016). At the same time, the extensive usage of social media platforms motivates advertisers to capitalize on it (Jin & Phua, 2014), and it changes the marketing world imperceptibly. Popular social media platforms like Facebook, Instagram, YouTube are becoming the new marketing “battlefields”.

According to the new definition from Felix et al (2017), “social media marketing is an interdisciplinary and cross-functional concept that uses social media to achieve organizational goals by creating value for stakeholders”. Companies open their social media accounts and use for propagating information, holding online campaigns, interacting with followers, etc. They put significant efforts on running these accounts and trying to positively influence their brand image, brand awareness and eventually enhance brand equity and purchase intention. However, even though social media marketing is getting more common, the advertising

saturation in the media causes decreased effectiveness, because consumers are resisting paying attention to them (Martín-Santana & Beerli-Palacio, 2013). Bang and Lee (2016) stated that advertisements sent indirectly by friends are more acceptable compared to those sent directly by brands. Therefore, a new and more effective form of advertising becomes prevalent, it’s

(6)

6 called influencer marketing. For example, Robyn Rihanna Fenty, a famous singer, posts a

fascinating picture of herself on Instagram wearing the latest Adidas shoes, her followers see the post and notice the shoes, at the same time positive attitude and value are added to Adidas’s brand image and brand equity, and their purchase intention of the shoes is motivated to a certain extent.

On the other hand, influencer marketing significantly contributes in marketing campaigns for companies from different industries. The cosmetics industry is one of the representative industries. However, different social media platforms have different features and user profiles, it’s important for brand managers to know how to utilize the effects of influencer marketing on social media platforms that suit their brands best (Erdoğmuş & Çiçek, 2012). Thus, it’s important for researchers to conduct individual research on a specific social media platform to provide insights for marketing and branding (Lee & Watkins, 2016). YouTube, as the second largest search platform after Google since 2008 (Klaassen, 2008), is not

sufficiently studied. Previous studies focus on social networking sites, like Facebook and Twitter, however, YouTube is different since it is a video-sharing site (Kim & Ko, 2012). There is a group of YouTube users that is significant to marketers: the YouTube vloggers. They are the important influencers on this social media platform because they have a large number of subscribers, and they are able to impact mass audiences in terms of what and where to purchase. Furthermore, YouTube vloggers usually add personal opinions and emotions when introducing products. It’s interesting to know whether this subjective information contributes more on changing brand equity and purchase intention than objective information, which means just state the product’s features without adding personal attachments. According to the Forbes (2017) list, the ten

(7)

7 most influential YouTube vloggers on the beauty sector reached over 46.5 million subscribers, which is 2.7 times the current population of Netherlands (Newmark, 2018). They act like opinion leaders and gain a lot of attention and trust. Besides, Lee & Watkins (2016) said it’s not necessary to find a celebrity as the spokesperson for an advertisement, but the similarities between the spokesperson and audiences are important. This statement and the statistics from Forbes indicate an insight that famous YouTube vloggers might have greater influences on consumers compared to celebrities.

Because of the growing importance of influencers from social media platforms, it’s significant for marketers to know their impacts on customers’ perceptions. Most existing studies focus on the impact on consumers’ perceptions towards luxury brands as well as their purchase intentions, but little research has done from the perspective of low-priced brands. However, the target groups for these two types of brands are different, and their consumers tend to have different purchasing processes. Compared to luxury brands, consumers usually have a shorter and different decision making process regarding low-priced brands, because of the price difference. Also, luxury brands put more efforts on establishing a long-term

relationship with consumers in order to increase the likelihood of repeat purchase behavior (Lee & Watkins, 2016). Thus, the conclusions from existing studies on luxury brands are hardly representative for low-priced brands. Additionally, due to the fact that social media marketing is becoming more common, both luxury and low-priced brands are adopting and applying it, it’s important for researchers to find out social media marketing’s different impacts on luxury and low-priced brands. Therefore, the purpose of this thesis is to fill the gap by looking at the impact of influencers from a video-sharing site (YouTube vloggers) on brand equity and

(8)

8 purchase intentions towards low-priced brands. Secondly, from academic aspect, it gives

insights when comparing consumers’ reactions on luxury brands and low-priced brands after exposing to YouTube vloggers’ videos. Last but not least, it is aimed to provide insights for brand managers or marketers about how to utilize the effects of influencer marketing on YouTube. For example, what type of information is more effective on building brand equity and increasing purchase intention? Briefly, an experiment and a survey will be conducted for this thesis. To ensure the consistency of results, this thesis focuses on the cosmetics industry. The research question is formed as below:

What’s the impact of different types of advertising information that YouTube vloggers deliver on the advertised brand’s equity and customers’ purchase intentions for low-priced brands?

2. LITERATURE REVIEW

Several related sources are reviewed in this chapter in order to create a good

foundation for this research. Firstly, YouTube vloggers are introduced as an important part of influencer marketing. And then four terminologies of brand equity are explained because these are the major aspects this study will focus on. Thirdly, advertising effects will be explained in details since it is the mediator of the relationship between YouTube vloggers and the advertised brand’s equity. This part is consisted of the steps that consumers process advertising

information, different routes they used to process advertising information (Elaboration likelihood Model), and the hierarchy effects of advertising. Besides, brand fame is introduced

(9)

9 because it acts as a moderator on the relationship between YouTube vloggers and the

advertised brand’s equity. Next, the terminology of purchase intention is explained. Finally, previous researches about influencers’ impacts on luxury brands are reviewed, and these helps to define an academic gap and motivate further study. The hypotheses that this research will test are also introduced after each related terminology’s explanation.

2.1 YouTube vloggers

2.1.1 Influencer marketing

YouTube vloggers open their own channels on YouTube, and use the channel as a platform to share opinions, life, products, knowledge, etc., almost everything. But when they got enormous number of subscribers, the situation changed. YouTube vloggers that have

thousands or even millions of subscribers can exert great impacts on spreading information to a mass population base. Thus, brand managers quickly realize this opportunity and collaborate with them for marketing purposes. Normally, consumers tend to put biases and less trust on official advertisements and consider them as exaggerating. According to Bang and Lee (2016), advertisements sent directly by a brand are not preferred by consumers, instead, those sent indirectly by friends are more acceptable. This statement further proves the great impacts of YouTube vloggers because consumers view YouTube vloggers as “friends” once some

similarities are found between them (Lee & Watkins, 2016). Because of this reason, brands frequently collaborate with famous YouTube vloggers for advertisements, and this

phenomenon is very common in the cosmetics industry for both luxury brands (e.g. Yves Saint Laurent, George Armani that sold in department stores) and low-priced brands (e.g. NYX,

(10)

10 Revlon that sold in drug-stores or online only). However, compared to luxury brands, low-priced brands, like NYX and Colourpop, rely heavily on YouTube vloggers to build brand equity (e.g. brand awareness, brand image, brand loyalty, etc.).

Influencer marketing is a new word so there is no academic definition in the existing literature (Johansen & Guldvik, 2017). The relationship between brands and influencers is about business, their collaboration benefits brands by increased brand knowledge, and benefits influencers financially. From this point of view, the definition of influencer marketing by the Word of Mouth Marketing Association (WMMA, an official trade association that focus on word-of-mouth and social media marketing) is not appropriate. WMMA defines it as

“identifying key communities and opinion leaders who are likely to talk about products and have the ability to influence the opinions of others” (Langan, 2016), it states the essence of influencer marketing, but doesn’t point out the mutual beneficial relationship between brands and influencers. However, the definition from Markethub (a leading influencer marketing company) takes this fact into consideration. It defines influencer marketing as “recruiting thought leaders and authorities within your niche to broadcast your message to a wider audience”. In the cosmetics industry, Colourpop (a Los Angeles based company, founded in 2014) is a great example of conducting influencer marketing. It collaborates with famous YouTube vloggers to create and launch new eye shadow palettes. For example,

“KathleenLights”, who has over three million subscribers on her YouTube channel. Colourpop collaborates with her and launched a new eye shadow palette and marked as “KathleenLights X Colourpop”. On the other hand, there are many other vloggers that didn’t collaborate with brands but still introduce their products in their videos. They purchase products on their own,

(11)

11 and eventually find out the products are great and share them to subscribers. In this case, they are also important influencers that have great impacts on brand equity and purchase intention.

2.1.2 Advertising Information

YouTube vloggers’ videos exert an important role in delivering advertising information, and many scholars have already developed several classifications on advertising information when conducting researches. For example, Smith (1991) studied the effects of advertising information on consumers’ inferences on missing product attributes of print advertisement, he classified advertising information into three categories: “verbal information”, “visual

“information, and “the combination of verbal and visual information”. Besides, Harmon et al (1983) mentioned “objective information” and “emotional information” when they were studying comparative advertising (a type of advertising that promotes products by comparing with competitors’) on magazine advertisement. While Royo-Vela (2005) classified TV

commercials content into “informational content” and “emotional content” when he studied the effects of information on audiences’ evaluation. Thus, there are many different ways to classify advertising information. However, taking the features of makeup YouTube vloggers’ videos into consideration, the classification of “verbal information” vs. “visual information” does not fit in this research, because most of times their videos contain both picture and sounds. In addition, the classifications that Harmon et al (1983) and Royo-Vela (2005) used are similar, they distinguished advertising information based on objectivity and subjectivity. Objectivity refers to the extent that advertising information is expressed unbiasedly and not influenced by speakers’ emotions, while subjectivity refers to the extent that advertising information is expressed as affective states. Both of Harmon et al (1983) and Royo-Vela (2005)

(12)

12 refer subjective information by the word “emotional”, and refer objective information by

similar words: “objective” and “informational”. Most importantly, Royo-Vela (2005) studied TV commercials which is close to the advertisement videos of YouTube vloggers. After considering these reasons, this thesis is going to use “objective information” and “subjective information” to classify advertising information delivered by makeup YouTube vloggers. To be more specifically, objective information refers to the advertising information that states a product’s features, for example, the price, weight, packaging description, etc. On the other hand, subjective

information refers to the advertising information that contains the speaker’s emotions, user experiences, and opinions.

2.2 Brand equity/Customer-based brand equity (CBBE)

When people consider about brand equity, it usually indicates two kinds of brand equity, one is from financial perspective and the other is the customer-based brand equity (CBBE). While in this thesis, brand equity only refers to CBBE because the financial brand equity is not involved. CBBE is firstly introduced by Keller (1993) to describe and measure the differential effects of brand knowledge on customer response after exposing to brand’s marketing. Keller believed that brand knowledge is the most valuable asset that can improve marketing

effectiveness. He defined brand knowledge as brand awareness and brand image. Brand awareness refers to the customers’ ability to recognize a brand when perceiving cues, and the ability to retrieve a brand in specific situations (Keller, 1993). On the other hand, brand image indicates the symbolic meanings related to attributes and functional consequences of a product (Padgett & Allen, 1997). With the symbolic meaning, brand image provides information related

(13)

13 to product quality and functionality, and facilitates consumers to differentiate products within the same category (Zembik, 2015). Keller (1993) stated that every time when consumers get in touch with a brand, they store related memory and associate it as their perceptions towards the brand. However, brand image is not only formed by direct experiences, like purchasing experiences, but also can be formed when consumers are exposed to commercials, word-of-mouth descriptions, etc.

Another important framework about CBBE is introduced by Aaker (1996a), he looked at the product/service value provided to customers, CBBE consists of any assets or liabilities that linked to the brand name and symbol, at the same time, add to or subtract from the

product/service value. There are many dimensions that can affect CBBE, but the core

dimensions are: perceived quality, perceived value, brand uniqueness and willingness to pay a price premium (Aaker, 1996a). Each dimension will be introduced in details as following:

2.2.1 Perceived quality

According to Aaker (1996a), perceived quality is the core dimension because it is highly associated with other key measurements of CBBE, such as willingness to pay a price premium, brand usage and purchase intention. The most acceptable definition of perceived quality

viewing it as the customers’ judgement of a brand (relative to alternative brands) in terms of its overall excellence, esteem and superiority (Netemeyer et al, 2009). Thus, perceived quality is different from objective quality since it measures at a very abstractive level of customers’ assessment of a brand (Aaker, 1996a). As a result, perceived quality is applicable across product classes as it can be measured in different scales, for example, “brand A has high quality

(14)

14

2.2.2 Perceived value

Perceived value is more about the trade-off between what customers give out and what they receive in return (Kirmani & Zeithaml, 1993). It is also abstractive because customers make an overall assessment on the brand utility based on the trade-off (Netemeyer et al, 2009). Aaker (1996a) stated that perceived value can be viewed as an indicator of brand success at creating value proposition, and value proposition usually refers to functional benefits. People often argue that perceived quality and perceived value are similar measurements, and

customers do not have the intention to distinguish them (Aaker, 1996a). While on the other hand, some theorists suggested that they are different because perceived value relates more about a brand’s functional utility, and perceived quality associates a higher prestige aspect of a brand (Netemeyer et al, 2009). While in this thesis, perceived quality and perceived value are considered as two separate dimensions of CBBE, and will be measured separately in the following research.

2.2.3 Brand uniqueness

Brand uniqueness is also a core dimension of CBBE as it is able to indicate price level to a certain extent (Aaker, 1996b). It refers to the level of uniqueness that customers perceive a brand when comparing to competitors, if the brand is highly unique in its product class, then it has the ability to set higher price than competitors (Aaker, 1996b). Thus, this shows the

relationship between brand uniqueness and willingness to pay a price premium. Netemeyer et al (2009) stated that customers’ judgement on the brand uniqueness can be affected by both direct experience and advertising claims. The higher uniqueness level that customers perceive, the higher price premium they are willing to pay in the marketplace (Aaker, 1996b).

(15)

15

2.2.4 Willingness to pay a price premium

The willingness to pay a price premium is considered as a loyalty indicator because it shows to what extent do customers willing to pay more for the brand’s product in comparison to another brand that offers similar products (Aaker, 1996a). For example, a customer is willing to pay extra €400 for a MacBook Pro rather than a Toshiba laptop. Aaker (1996a) argued that the willingness to pay a price premium is the strongest single measurement of CBBE because any drivers of CBBE should eventually affect customers’ willingness to pay a price premium. In other words, if a variable does not have any impacts on customers’ willingness to pay a price premium, then it contributes little value on indicating CBBE. In addition, Blackston (1995) said that the willingness to pay a price premium is a consequence of managing other dimensions of CBBE well, especially the perceived quality, perceived value and brand uniqueness.

The main relationship that this thesis focuses on is between the types of advertising information that YouTube vloggers deliver and brand equity. There are several existing sources that have already studied related areas. For example, Chan (2015) analyzed Chinese’s

perceptions of informative and emotional advertising. The author’s finding suggested

marketers to provide emotional commercials because customers are more likely to view this kind of commercials as interesting and appealing, and eventually leads to higher score on “liking” the commercial and perceived brand image, compared to “dull” and “uninteresting” informative commercials. In addition, Lee & Burns (2014) studied the effect of informational and emotional advertisement strategy on brand attitude and brand recall. They also found that emotional ad strategies led to greater change in brand attitude compared to informational ad strategies. As discussed before in 2.1.2 Advertising information, objective information refers to

(16)

16 the product’s features, and subjective information is more related with the speaker’s emotions and opinions about the product. This thesis will use “objective information” and “subjective information” to classify advertising information, so combining with the two findings from Chan (2015) and Lee& Burns (2014), the following hypothesis is proposed:

H1. Subjective information delivered by YouTube vloggers has greater impact on advertised brand’s equity than objective information.

2.3 Advertising effects

Many researchers have already studied different factors that can impact the advertised brand’s equity. For example, Jung et al (2016) found that consumers react differently towards the advertised brand, depends on how they perceive online advertising. When they perceive advertising as valuable information, they generate favourable attitude or behavior intention towards the brand or product. In addition, Bang & Lee (2016) studied online advertising on social media in specific, they stated that an ad placed inside users’ timeline and sent by

unfamiliar advertiser is more favorable than the ad placed outside timeline and sent by known others. Nevertheless, consumers are exposed to hundreds of ads per day (Edwards & Lee, 2002), it raises the issue of effectiveness, consumers are overload and believe advertising is deceptive and misleading (Rojas-Mendez et al., 2009). Friestad and Wright (1994) used the persuasion knowledge model to describe how consumers’ attitude change and how they cope with

advertising. According to the model, if consumers recognize advertisement as a purposive tactic that aims to achieve managers’ purpose, like selling more products, then consumers would

(17)

17 generate a defensive coping behavior automatically, they would choose to reduce attention from advertising.

In late 20th_{century, researchers began to conceptualize the perceptual process when}

customers change their attitudes and behaviors towards advertising. McGuire (1972) used six stages to identify advertising effects, which is known as the Information-Processing Model (IPM). More specifically, customers are firstly presented to the advertising communication and attention to it, then they comprehend arguments and yield conclusions. After that, the new attitude formed on customers’ mind is retained and eventually impacts their behaviors (McGuire, 1972). In 1996, Scholten modified McGuire’s model into five stages, named as

exposure, reception, persuasion, retention and behavior. Even though different researchers use different numbers of stages to summarize the advertising effects on customers, it’s roughly the same. Besides, the elaboration likelihood model has great influence on this perceptual process, it indicates the way that consumers process advertising. Cacioppo and Petty (1980) first came up with this model and stated that there are two ways of processing advertising information: central route processing and peripheral route processing. In the context of watching lipsticks videos of a YouTube vlogger, customers that use the former way usually pay more attention to product information, like the price, colors, texture etc. While other customers that use the peripheral route processing are easily influenced by information that are not important. They are more likely to purchase the lipstick just because the YouTube vlogger recommends it. Overall, advertising effects act like a mediator on the relationship between advertising information and the advertised brand’s equity. Furthermore, Olney et al (1991) conducted a study on TV commercials, and developed a flow chart to explain the hierarchical effects of

(18)

18 advertising, from ad content through emotions and attitudinal responses to viewing behavior. To be more specifically, the advertising effects flow when customers are analyzing ad content in terms of the advertising appeals and uniqueness (Olney et al, 1991). Thus, customers’

knowledge about the brand is renewed by adding new knowledge and updating associations. They also stated that based on the content analysis, customers generate subjective emotions, for example, they feel pleased about the ad, or not happy with it because of unpleased content. Further, advertising starts to affect the aspect of customers’ attitude because the viewing experience would be evaluated automatically on their mind (Olney et al, 1991). The authors measured this aspect by using terms of hedonism, utilitarianism and interestingness. Eventually they founded out that advertising effects flow to the behavior aspect as customers would increase or decrease the ad viewing time. Therefore, the following hypothesis is proposed:

H2a. Advertising effects positively mediate the causal relationship between advertising information and brand equity.

According to H2a, advertising effects is assumed as the mediator of the relationship between advertising information and brand equity. Olney et al (1991) mentioned four

hierarchical steps of advertising effects: knowledge, emotion, attitude and behavior. Through researches on TV commercials, they found that ad contents exerted great influence on respondents’ viewing time, and the relationship is mediated by their emotions and attitudes towards TV commercials. In other words, behavior is influenced by knowledge, emotion and attitude. They also found that respondents’ emotions after watching TV commercials are

(19)

19 positively related with their viewing time. Therefore, to further study the mediating effect of advertising effects, the following hypothesis is proposed based on Olney et al (1991)’s findings:

H2b. Among the three steps of advertising effects, emotion is the most affected step by advertising information, and it has the greatest influence on brand equity.

2.4 Brand fame

Brand fame is supposed to have great impact on the relationship between advertising information and brand equity. It indicates to what extent do customers know about a particular brand, in other words, how famous it is. As mentioned before, Loftus and Loftus (1980) stated that memory is durable, and that’s why once information is stored on customers’ mind, it then becomes hardly to decay. To be more specifically, if customers have already stored much and strong brand knowledge in mind, then it’s harder to attach different information, compared to other brands that have weak brand fame. For example, Mercedes-Benz is a famous high-end automobile brand which is mainly associated with older population, then it is difficult for Mercedes-Benz to target younger customers because the association in mind is hard to change. Based on Loftus and Loftus (1980)’s statement, it’s reasonable to propose the following

hypothesis:

H3. Advertising information exerts more positive impact on the brand equity of unfamous brands than famous brands.

(20)

20

2.5 Purchase intention

Purchase intention refers to the purchase willingness towards a certain product. There are numerous factors that may have impacts on customers’ purchase intention, and

researchers are trying to discover the rules among all the possible factors. Johansen and Guldvik (2017) stated that the relationship between influencer marketing and consumers purchase intention is indirect and not obvious. This result is unexpected because it contradicts other previous researches. They also argued that influencer marketing doesn’t perform more efficient than regular online advertisement as a marketing tactic. However, Lee and Watkins (2016)’s work indicated different results, they measured the attractiveness of influencers from YouTube, in relationship with customers’ perceptions towards luxury brands. They proved that higher attractiveness positively impacts the advertised brand’s image stored in customers’ mind, and motivates purchase intention. On the other hand, it’s interesting to point out that celebrity endorser is not a key indicator of customers when choosing beauty products (Ngnoubamdjum & Zahn, 2016). Nevertheless, non-celebrity endorsers can actually draw customers’ attention to the product itself, instead of the person that endorsing the product, and customers actually are more in favor of this advertisement form (Ngnoubamdjum & Zahn, 2016). So, this argument gives an insight for this thesis to focus on the impact of YouTube vloggers (a type of non-celebrity endorsers) on consumers’ perceptions on advertised brands.

Besides the main focus (the relationship between advertising information and brand equity), this research also aims to find out the relationship between brand equity and purchase intention. Previous researches have already studied similar topics. For example, Lee & Watkins (2016) focused on the impact of YouTube vloggers’ personal attractiveness on customers’

(21)

21 perceptions towards luxury products and purchase intention. They measured luxury brand value and brand-user-imagery-fit first, then linked to customers’ purchase intentions. Their findings proved that luxury brand value, brand-user-imagery-fit and purchase intention increased significantly after exposing to YouTube vloggers’ videos. Due to the fact that brand value, brand-user-imagery-fit and brand image are all parts of brand equity, the following hypothesis is proposed:

H4. Customers’ purchase intention is positively associated with brand equity.

2.6 Influencers’ impacts on luxury brands

Existing literature puts more emphasize on discovering the impacts of social media influencers on customers’ perceptions towards brands or luxury brands in specific. Godey et al (2016) conducted a research on the impact of social media marketing efforts (SMMEs) on consumers’ perceptions on brand equity and their response towards luxury brands. They classified SMMEs into entertainment, interaction, trendiness, customization and word-of-mouth. Their results proved that SMMEs have a great positive effect on rising brand awareness and brand image, which eventually leads to increasing preference, greater loyalty and higher price premium towards the brand. On the other hand, Lee & Watkins (2016) looked at YouTube in details. They studied the relationship between influencers on YouTube and consumers’ perceptions towards luxury brands as well as purchase intentions. They measured the attractiveness of YouTube vloggers by three factors: social attractiveness, physical

(22)

22 beliefs, social status, etc. as vloggers). The three factors are the antecedents of para-social interaction (PSI), which is a term to measure audiences’ likelihood of frequent interaction with a YouTube vlogger. They proved that the higher level of each factor leads to higher PSI, so, higher likelihood for audience to view vloggers as “friends” and more frequent interaction with them, which eventually results in more positive perceptions and higher purchase intentions towards luxury brands.

2.7 Conceptual framework and hypotheses

(23)

23 Independent variable: By reviewing existing literature and taking the features of makeup YouTube videos into consideration, this thesis is going to classify the advertising information delivered by makeup YouTube vloggers into two categories: [1] Objective information, [2] Subjective information. The classification is based on the different subjects of makeup YouTube vloggers’ videos. For example, “Monthly Favorite”, “Unboxing and First Impression”, etc. Videos from different subjects usually reveal different information (e.g. packaging, design, use

experience), which can be classified into objective information and subjective information. To be more specifically, [1] Objective information is about the product itself. It can be described or presented, and the aspects include packaging, design, texture, weight, size, smell, price, etc. Usually, videos like “Unboxing and First Impression” mainly contain this type of information. Secondly, the [2] Subjective information is more about makeup YouTube vloggers’ personal opinions. How they feel about the product? what’s the use experience? Do they like it or not? Videos like “Monthly Favourite” and “Product Review” usually contain this type of information.

Mediator: according to Olney et al (1991), advertising effects can be measured

hierarchically in four steps. The information delivered by makeup YouTube vloggers influences the advertised brand’s equity by the way as Olney et al (1991) stated, impacting customers’ knowledge, then emotions, attitude and eventually change behaviors. Thus, advertising effects act as the mediator on this relationship between YouTube vloggers and the advertised brand’s equity.

Moderator: this research takes the potential consequences of brand fame into

consideration by adding it as a moderator on the relationship between YouTube vloggers and the advertised brand’s equity. It simply includes two levels, [1] Famous brand and [2] Unfamous

(24)

24 brand. By using these two extreme classifications, the potential influences could be revealed more obviously.

Dependent variables: there are two dependent variables in this research, brand equity and purchase intention. Firstly, brand equity (CBBE) refers to assets or liabilities associated with brand name or symbol that add to or subtract from the product/service value (Aaker, 1996a). It has many different dimensions, but based on Aaker (1996a)’s theory, there are four core dimensions and this research is going to measure brand equity based on them: [1] perceived quality, [2] perceived value, [3] brand uniqueness, [4] willingness to pay a price premium. In addition, purchase intention will be measured by the likelihood that customers are willing to purchase the advertised product.

3. METHODOLOGY

This chapter relates to the empirical part of this thesis. Firstly, the collected sample will be described and explained in details, and followed by the research procedure that is presented in details. Lastly, the measurement of variables will be presented and explained.

A quantitative study will be conducted to find out the impact of advertising information on brand equity and purchase intention towards low-priced brands. To make the results more representative and precise, the research is limited to the cosmetics industry. A 2x2 between-subject experimental design will be used in this research. The first factor is about the types of advertising information that YouTube vloggers deliver, and there are two levels: objective information and subjective information. The other factor is brand fame, it consists of two levels:

(25)

25 famous brand and unfamous brand. This research involves six groups in total, two of them are control groups. Online surveys via Qualtrics will be used to measure responses, and

respondents from different groups will answer same questions based on different videos. The surveys consist of fixed-response questions, so respondents are required to select answers from Likert-scaled choices. To eliminate the interference of different effects from different YouTube vloggers, all the videos will be chosen from one makeup YouTube vlogger,

“KathleenLights”. Furthermore, to include the moderator (brand fame) in the research,

Maybelline is selected as the famous brand and Colourpop is selected as the unfamous brand. Both of them are brands based in the United States. Maybelline is a very well-known cosmetics brand, founded in 1915 and owned by L’Oréal. On the other hand, Colourpop is founded in 2014, it’s a young brand with less reputation and fame compared to Maybelline, but Colourpop is rising quickly in recent years.

3.1 Sample

The respondent population will be limited to females because they are the major customers for cosmetic products, so the non-probability sampling is used for this research. Due to the fact of having six survey groups, the sample size is expected to be at least 180

respondents (anonymously). In addition, the convenience sampling technique will be adopted by posting the survey link on social media platforms (e.g. Facebook).

(26)

26

3.2 Research procedure

First, the respondents will be grouped randomly into six groups (four groups + two control groups), then, each respondent (except respondents from control groups) will watch a (different) video based on the group she is assigned to. There are four different videos relate to four groups based on the different types of advertising information that video contains. A survey will be given for each respondent after exposing to the video to record responses. Respondents from the two control groups will be asked to fill out the survey without watching any videos, so the control groups’ responses are treated as the responses before exposing to videos. By comparing responses from the first four groups and the two control groups, the differences of customers’ attitude towards brand equity and purchase intention between before and after exposing to advertising information delivered by YouTube vloggers shall be disclosed. The following tables show detailed information of grouping and the survey procedure:

Table 1: Videos Used for Surveys and the Difference

Famous brand Unfamous brand

Objective information Group A, Video 1 Group B, Video 2 Subjective information Group C, Video 3 Group D, Video 4

No video Control Group 1 Control Group 2

(27)

27 Pre-Tests 1. Each respondent will answer two questions (details explained in 3.3.1

Pre-Test 1). It’s aimed to ensure if respondents perceive

Maybelline/Colourpop as a famous or unfamous low-priced cosmetics brand;

2. Each respondent will be randomly assigned into four groups.

Respondents from the same group will watch a same video* and answer one question (details explained in 3.3.2 Pre-Test 2). It’s aimed to ensure if respondents perceive the information in Video 1 & 3 as objective, and perceive the information in Video 2 & 4 as subjective.

* The videos refer to the four videos that are planned to be used in the main research.

Watch a video Each respondent (from Group A, B, C, D) will watch one video**.

**The video is different based on the assigned group. Different videos contain different advertising information (details presented on Table 1), but comes from the same YouTube vlogger.

Survey Each respondent will be asked to fill out a survey to measure and record advertising effects (for Group A, B, C, D only), respondents’ perception towards brand equity, and their purchase intention towards the advertised product.

(28)

28

3.3 Pre-Tests

Before conducting the main research, it’s important to take a pre-test to minimize the likelihood that respondents might misunderstand questions, in order to improve reliability and validity (Saunders et al, 2012). Therefore, two pre-tests are conducted, and each pre-test will be explained in details as below:

3.3.1 Pre-Test 1

The first pre-test is to make sure the two brands selected (Maybelline and Colourpop) for the moderator (brand fame) are perceived accurately by respondents. From the perspective of this research, Maybelline is selected because it’s one of the most famous low-priced

cosmetics brands, it has a strong brand fame, while Colourpop is the opposite and it’s

considered as unfamous with a weak brand fame. To make sure this study can be successfully conducted, respondents should perceive these two brands’ fame in the way as the researcher’s. The pre-test was distributed to 10 respondents from the sample population, and each of them answered the two questions (adopted from Yoo et al, 2001) in Table 3 which indicate their perceptions of how famous these two brands are. The collected results (see Appendix 1) show that 100% respondents think Maybelline is more famous than Colourpop as a low-priced cosmetics brand. Even though the pre-test result is consistent with the expectation, it can’t be guaranteed that everyone perceives the two brands in this way. Therefore, the beginning of the main survey is modified by adding questions from the pre-test 1. Respondents assigned to Group A, C, and Control group 1 are going to respond questions regarding to Maybelline, and respondents assigned to Group B, D, and Control group 2 are going to respond questions regarding to Colourpop.

(29)

29 Question 1 I have heard about and I know Maybelline/Colourpop.

Strongly Somewhat Neither agree Somewhat Strongly disagree disagree nor disagree agree agree 1 2 3 4 5 Question 2 I am able to recognize Maybelline/Colourpop easily from among other

competitive brands.

Strongly Somewhat Neither agree Somewhat Strongly disagree disagree nor disagree agree agree 1 2 3 4 5

Table 3. Pre-Test 1 Questions

3.3.2 Pre-Test 2

The second pre-test is about the videos that this research is going to use. The four videos are classified by the types of advertising information it contains. Because it’s difficult to completely eliminate objective information in subjective videos, vice versa. In this research, if a video is considered as containing mostly objective information, then it’s perceived as an

objective video, vice versa. Therefore, it’s important to make sure that respondents perceive these videos as they supposed to. For example, information in video 1 and 2 are designed to be objective (See Table 1 for details), while the pre-test is able to indicate whether respondents consider the information in video 1 and 2 as mainly objective. The question (see Table 4) is adapted from Leclerc et al (1994)’s research which was originally aimed to distinguish utilitarian products and hedonic products. Answers that scores 1, 2 or 3 indicate that the video mainly contains objective information, 5, 6 and 7 scores indicate that the video mainly contains

subjective information. Scores 4 means no obvious differences so it’s hard to conclude whether the video mainly contains objective or subjective information. This pre-test was distributed to

(30)

30 20 respondents, they were grouped randomly into Group 1, 2, 3, and 4. Each video was

watched and evaluated by 5 respondents. Respondents from group 1 watched and evaluated video 1 (the objective video about Maybelline), respondents from group 2 watched and evaluated video 2 (the objective video about Colourpop), respondents from group 3 watched and evaluated video 3 (the subjective video about Maybelline), and respondents from group 4 watched and evaluated video 4 (the subjective video about Colourpop). The results (see Appendix 2) show that all the four videos are qualified because the two objective videos are perceived containing more objective information than subjective information, and the other two subjective video are perceived containing more subjective information than objective information. However, it also cannot be guaranteed that everyone perceives these four videos in the same way, thus, the question from pre-test 2 is also added to the end of the main survey, but before demographics.

After watching the video, please indicate how objective or subjective the information is:

Definitely Somewhat Neither subjective Somewhat Definitely objective Objective objective nor objective subjective Subjective subjective

1 2 3 4 5 6 7 Table 4. Pre-Test 2 Question

3.4 Measurement of variables

Besides demographic questions, questions that measure advertising effects, brand equity and purchase intention are adopted from previous studies. The advertising effects are measured in terms of knowledge, emotion, and attitude, while the measurement scales are adopted from Olney et al (1991)’s work. Brand equity is measured in four aspects from

(31)

31 customers’ perspective: perceived quality, perceived value, brand uniqueness and willingness to pay a price premium (Netemeyer et al, 2004). On the other hand, purchase intention is measured by the likelihood of purchasing. The measurements of brand equity, purchase intention and advertising effects (see Appendix 3) are explained in details as following: Brand equity

Perceived quality: aimed to find out respondents’ attitudinal assessment of the two brands. For example, “Compared to other brands of, I think Maybelline is of very high quality” (adopted from Netemeyer et al, 2004). Each question is measured by a 5-point Likert scale, 1 stands for “Strongly Disagree”, and 5 stands for “Strongly Agree”.

Perceived value: aimed to find out respondents’ objective assessment of the utility of the two brands. For example, “What I get from Maybelline is worth the cost” (adopted from Netemeyer et al, 2004). Each question is measured by a 5-point Likert scale, 1 stands for “Strongly Disagree”, and 5 stands for “Strongly Agree”.

Brand uniqueness: aimed to find out respondents’ overall attitude of the two brands compared to competitors. For example, “Maybelline is ‘distinct’ from other brands of the cosmetics product class” (adopted from Netemeyer et al, 2004). Each question is measured by a 5-point Likert scale, 1 stands for “Strongly Disagree”, and 5 stands for “Strongly Agree”.

Willingness to pay a price premium: aimed to find out respondents’ loyalty level of the two brands. For example, “The price of Maybelline would have to go up quite a bit before I would switch to another brand” (adopted from Netemeyer et al, 2004). Each question is measured by a 5-point Likert scale, 1 stands for “Strongly Disagree”, and 5 stands for “Strongly Agree”.

(32)

32 Purchase intention

It’s aimed to find out respondents’ willingness to and likelihood of actually purchase products from the advertised brands. For example, “I would purchase this product over other brands” (adopted from Allen et al, 2008). Each question is measured by a 7-point Likert scale, 1 stands for “Improbable”, and 7 stands for “Probable”.

Advertising effects

Knowledge: aimed to find out the extent that respondents perceive the video contains product-related information. For example, “This video contains sufficient information about the features of the brand’s products” (adopted from Olney et al, 1991), 1 stands for “Strongly disagree”, 7 stands for “Strongly agree”. The scales are different among questions but all measured by a 7-point Likert scale.

Emotions: aimed to find out the respondents’ emotional feelings about the video. For example, “I think this video is annoying/pleasing” (reversed scale, adopted from Olney et al, 1991). The scales are differently measured by emotional responses, but all measured by a 7-point Likert scale.

Attitudes: aimed to find out respondents’ attitudes about this video. For example, “I think this video doesn’t keep my attention/keep my attention” (reversed scale, adopted from Olney et al, 1991). The scales are also different among questions but all measured by a 7-point Likert scale.

(33)

33

4. RESULTS

The research results are recorded in this chapter. Firstly, the result of data collection is simply introduced together with the steps executed in the data preparation stage. And then, hypotheses testing is explained in details including choosing analysis tools, presenting important data and accepting or rejecting hypotheses.

4.1 Data collection and data preparation

This research involves six groups in total, so at least 180 respondents are required. Also, due to the fact that this research has to be distributed online only because respondents need to watch videos. Therefore, the survey was distributed in several online platforms in order to get sufficient participants, including Facebook, UvA student emails and Amazon Mechanical Turk. There are 511 received responses in total, 143 (28%) of them are males, and 368 (72%) are females. So firstly, the data is sorted by the gender, and all the responses from males are deleted because this study focuses on females only. Secondly, because the surveys contain videos, all the answers that took less than the video’s duration time are considered as invalid and deleted. Thus, there are 220 responses are considered as valid and used for the following analysis.

This research is a between-subject experiment and involves six groups, so the original data is distributed into six blocks that actually use same questions. To group data in orders and to combine data into one block, a new column called “Group” is added. This variable is valued from “1” to “6”, and each value indicates the group that each respondent is assigned to. Then, the data is sorted by “Group”, and the data from different blocks are moved to the first block.

(34)

34 Any other meaningless variables are deleted, e.g. columns of the other five blocks, survey start date and end date, etc.

Thirdly, there are two kinds of data that needed to be treated as missing data. The respondents that are assigned to the two control groups are not required to watch a video, therefore, they didn’t answer any video-related questions. Secondly, in case they don’t know the two brands (Maybelline and Colourpop) that this research focuses, the question scale of control groups’ surveys is modified. The scale used in the other four groups’ surveys is “1 Strongly disagree”, “2 Somewhat disagree”, “3 Neither agree nor disagree”, “4 Somewhat agree”, “5 Strongly agree”. And the scale for control groups is “1 Strongly disagree”, “2 Somewhat disagree”, “3 I don’t know”, “4 Somewhat agree”, “5 Strongly agree”. Therefore, value “3” in control groups’ answers should also be treated as missing data. Further, missing data is represented by value “999”.

During the survey, each variable (independent variable, moderator, mediator, dependent variables) is measured by several questions, thus, a mean of these questions is conducted for each variable. The means of questions related to perceived quality, perceived value, brand uniqueness, willingness to pay a price premium are recorded as “PQTOT”, “PVTOT”, “BUTOT”, “WTPTOT” respectively. And a mean of these four value is also computed and

recorded as “Brand equity”. The same step is also conducted for another dependent variable, purchase intention, and the mediator, advertising effects, and they are recorded as “PITOT”, “AdEffectsTOT” respectively. In addition, advertising information (X) and brand fame (M) are measured in scales, while they are considered as categorical variables in this research. The reason of scaling these two variables in surveys is to improve data accuracy, so based on the

(35)

35 results, these two variables are computed as category variables to meet the research need. Thus, two dummy variables are added and are given values based on the related scaled answers. The first dummy variable measures the subjective level of advertising information, “1” refers to subjective, and “0” refers to objective. The second dummy variable measures the level of perceived brand fame, “1” refers to famous, and “0” refers to unfamous.

Next, the reliability test is conducted to test if the scale of each variable is reliable. The results presented in Table 5 indicate that the Cronbach’s alpha for advertising effects, brand equity and purchase intention are 0.84, 0.86 and 0.91 respectively. Each corrected item – total correlation is higher than 0.3 which means each item is well correlated with its scale. In

addition, none will affect the Cronbach’s alpha significantly if deleted.

Additionally, the statistics in Table 5 show that the correlation between

(36)

36 not significantly correlated with advertising information. Thus, the mediating effect of

advertising effects may not exist, but it will be tested in details on section 4.2.

4.2 Hypotheses testing

In the following sections, hypotheses will be tested one by one using SPSS, and results will be summarized in Table 11.

4.2.1 Hypothesis 1

H1. Subjective information delivered by YouTube vloggers has greater impact on advertised brand’s equity than objective information.

This hypothesis assumes a causal relationship between advertising information (X) and brand equity (Y), and ANOVA analysis is conducted to validate it. Based on the results (see Table 6), this model is statistically significant, p<0.05, and explained 14.2% of variance in brand equity. In addition, the means of brand equity on conditions of subjective information and objective information are plotted in Figure 2. The plot indicates that compared to objective information, subjective information has a greater effect on brand equity. Thus, H1 is accepted.

(37)

37

4.2.2 Hypotheses 2 a + b

H2a. Advertising effects positively mediate the causal relationship between advertising information and brand equity;

H2b. Among the three steps of advertising effects, emotion is the most affected step by advertising information, and it has the greatest influence on brand equity.

H2a is about the mediating effect of advertising effects on the relationship between advertising information (X) and brand equity (Y), and H2b is proposed based on H2a. Therefore, it’s essential to firstly test whether advertising effects function as a mediator, and then H2b can be further tested.

(38)

38 It’s mentioned in section 4.1 that the correlation between advertising information and advertising effects is not significant. The results showed in Table 7 is in compliance with the previous correlation result as the p value for a1 is greater than 0.05, which means the effect of

advertising information (X) on advertising effects (M) is not significant. However, the b1 effect is

significant (p < 0.001), and the positive sign of coefficient (b1 = 0.329) means advertising effects

(M) is positively correlated with brand equity (Y). In addition, the indirect effect (a1b1) of

advertising information (X) on brand equity (Y) is -0.024, and the related confidence interval includes zero (-0.165 to 0.116) which further proves that the mediating effect of advertising effects does not exist. Therefore, H2a is rejected.

H2b further assumes that among the three steps of advertising effects, emotion is the most important one because it is assumed to be greatest impacted by advertising information, it also has the greatest impact on brand equity. Even though H2a is not proved, H2b is tested

(39)

39 here. The first half of H2b is about the relationship between advertising information (X) and advertising effects (M). According to Table 8, the coefficients of a1.1, a1.2, a1.3 indicate that

emotion is not the mostly affected aspect, while attitude is. However, the p values of a1.1, a1.2,

a1.3 are not significant, which indicate the absence of mediating effect. Secondly, the other part

of H2b is related to the relationship between advertising effects (M) and brand equity (Y). Table 8 shows that b1.1 is significant, ad appeal (M1) is positively correlated with brand equity (Y) as

b1.1 = 0.224, p < 0.001. However, the statistics of emotion (M2) and attitude (M3) indicate

opposite results to ad appeal because both b1.2 and b1.3 are not significant. Additionally, the

coefficient of b1.1 is the largest, compared to b1.2 and b1.3. So ad appeal (M1) has the greatest

impact on brand equity (Y). Also, zero is included in all the confidence intervals of a1.1b1.1,

a1.2b1.2, and a1.3b1.3. This further proves that the mediating effect of advertising effects doesn’t

exist, even though advertising information (X) seems to have the greatest impact on attitude, and ad appeal seems to have the greatest impact on brand equity (Y). Because the causal relationship does not exist, these values are meaningless. Therefore, H2b is rejected.

(40)

40

H3. Advertising information exerts more positive impact on the brand equity of unfamous brands than famous brands.

This hypothesis assumes a moderating effect of brand fame on the causal relationship between advertising information (X) and brand equity (Y). Specifically, compared to famous brand, the brand equity of unfamous brand is more positively affected by advertising information. Table 9 shows that c3 = -0.48, and p < 0.05, which means that the interaction

between advertising information (X) and brand fame (M) on brand equity (Y) is significant. Therefore, the moderating effect does exist. Furthermore, the results of conditional effect of advertising information (X) on brand equity (Y) indicate that the causal relationship between X and Y is significant for both famous and unfamous brands (p < 0.05). The effect values (0.923 for unfamous brand, 0.444 for famous brand) indicate that unfamous brand is more impacted on brand equity, compared to famous brand, and this trend is also indicated in the interaction plot in Figure 3. In addition, the R square indicates that this model represents 43.3% of total variance of brand equity, and it’s statistically significant as p < 0.001. Moreover, the brand equity of unfamous brand is more affected compared to the brand equity of famous brand (c2 =

(41)

41 -0.63, p < 0.05) when the advertising information are perceived at the same level of subjectivity. The negative sign of c2 indicates that the level of brand fame is negatively associated with brand

equity, so the brand equity of unfamous brands is more affected after exposing to advertising information, compared to famous brands. This result doesn’t indicate that famous brands have lower brand equity than unfamous brands, because the brand equity here refers to customer perceived brand equity after exposing to advertising information. Overall, H3 is accepted.

(42)

42

H4. Customers’ purchase intention is positively associated with brand equity.

This hypothesis assumes a causal relationship between two dependent variables: brand equity and purchase intention. In order to test H4, linear regression is conducted and brand equity is considered as the predictor. The result (see Table 10) indicates that brand equity has a significant influence on purchase intention as p < 0.05. The R square value indicates that this model represents 65% of variance of purchase intention. Therefore, H4 is accepted.

(43)

43 Table 11 Hypotheses Summaries

Nr Hypothesis Accepted/

Rejected H1 Subjective information delivered by YouTube vloggers has greater impact

on advertised brand’s equity than objective information;

Accepted

H2a Advertising effects positively mediate the causal relationship between advertising information and brand equity.

Rejected

H2b Among the three steps of advertising effects, emotion is the most

affected step by advertising information, and it has the greatest influence on brand equity.

Rejected

H3 Advertising information exerts more positive impact on the brand equity of unfamous brands than famous brands;

Accepted

(44)

44

5. DISCUSSION

The research results will be further discussed in the section of general conclusion, following by its theoretical implications and managerial implications. In addition, limitations that this research could not solve will also be explained, and suggestions about future research will be given.

5.1 General conclusion

The result of hypothesis 1 shows that brand equity of the advertised brand is affected to different extents by different forms of advertising information. Compared to objective

information, subjective information has greater impacts on customers perceived brand equity. YouTube vloggers often add personal opinions, feelings and emotions about products in their videos, and this subjective information is more reliable and persuadable than objective information. Thus, the effect of subjective information is larger, and this result is in line with Lee & Burns (2014)’s study, which concluded that emotional ad strategies lead to greater changes in brand attitude compared to informational ad strategies.

So, what if two brands both are advertised in the same form of advertising information? Then the result of hypothesis 3 explains this question by proving that different levels of brand fame impact brand equity differently. Specifically, famous brands, like Maybelline, are less capable to modify their perceived brand equity through advertisements. Loftus and Loftus (1980) stated the logic by indicating the durable feature of memory. Because customers have already known famous brands and stored brand image in their memory, it’s hard to change by

(45)

45 just exposing them to single advertisement. In addition, the result of hypothesis 4 proves that the higher brand equity, the higher purchase intention.

However, there are two rejected hypotheses, and these are related to advertising

effects. Even though many previous researches could be used to support Hypotheses 2a and 2b, the positive mediating effect of advertising effects on brand equity doesn’t exist in this research. Holbrook & Batra (1987) have a research on the mediating effect of customers’ emotions on their responses after watching TV commercials, and they pointed out three limitations of studies on consumer behavior: respondents, instruments and stimuli. These three limitations have great influences on study results because they’re hard to control. The four videos used in this research are not perfectly comparable due to the limitation of video resources, so different products are involved. Since these videos are not prepared for this research, instead, several videos are gathered and edited to fit in with this research. Also, the survey is distributed online through different platforms, respondents are selected by gender instead of distinguishing respondents that actively engaged in cosmetics consumption. These could be the reasons of why advertising effects are found to be negatively associated with advertising information. On the other hand, the relationship between advertising information and advertising effects is not proved in this research, but the causal relationship between advertising effects and brand equity is proved as positive and significant. It’s still important to notice this relationship, because customers perceived brand equity does get affected by the knowledge, emotions and attitude towards the ad after exposing to advertisement.

(46)

46

5.2 Theoretical implications

The result provides some insights on existing literature about influencer marketing and brand equity. Firstly, this research fills the gap of studying influencer marketing’s influence on brand equity by looking at video-sharing platforms, specifically YouTube. The effect of YouTube is measured in types of advertising information. Specifically, subjective information is better than objective information on influencer marketing, because it helps to improve customer perceived brand equity more effectively than objective information. Also, this research uses Loftus and Loftus (1980)’s statement of the durability of memory as a support, proves that unfamous brands are more capable to change customer perceived brand equity through influencer marketing, compared to famous brands. In addition, this research fills the gap of having insufficient studies on low-priced brands. It’s proved that the types of advertising information and the levels of brand fame are significantly correlated with low-priced brands’ brand equity, as well as purchase intention. Last but not least, even though two hypotheses regarding to mediating effect are rejected, there are still theoretical implications. When studying video-sharing platforms, it’s essential to prepare designed videos for researches. In order to achieve accurate results, it’s better to record videos intentionally instead of finding secondary resources.

5.3 Managerial implications

This research also has several managerial implications. For unfamous brands, using influencer marketing to advertise brands and products could be a good choice to increase brand awareness, brand fame, brand equity and purchase intention. When collaborating with

(47)

47 influencers, it’s important for managers to suggest them to add personal opinions, feelings and emotions about products to make it more persuadable to customers, instead of only giving objective information by just introducing products. Also, customers’ knowledge, emotions and attitudes towards advertisements are important to brand equity. It’s wise to set the goal of influencing these three aspects when planning commercials. This suggestion is more useful to famous brands, because it’s more difficult to change customers’ perceptions about the brand by simple influencer marketing. Managers of famous brands need to add new things to

influencer marketing to increase customers’ knowledge about the brand, or attaching positively emotions or attitudes to the brand. In addition, since influencer marketing has such a great impact on brand equity, it’s important for brand managers to set orientation and goals for each case of influencer marketing. For example, indicating the types of subjective information that influencers need to include, e.g. user experience; adding fun staff to videos to strengthen impression. Last but not least, for brands that want to increase sales, it’s wise to start with increasing brand equity.

5.4 Limitations and future research

There are some limitations in this research. Firstly, the survey is lengthy, respondents need to watch a 2-minute video and then answer more than 30 questions. This may hurt data accuracy and negatively impact respondents’ emotions towards the survey, which may cause the rejection of hypothesis 2. Secondly, this research uses non-representative convenience sample because the survey is distributed on platforms like Facebook and university emails. Thus, the external validity of the results cannot be guaranteed. Also, the videos used in survey are not

(48)

48 perfectly comparable due to resource limitation. Future research should consider to record videos intentionally instead of gathering videos online. In addition, there is a common method bias that respondents are asked to participate surveys online, so the data quality is not

guaranteed. Future research should consider to use group interviews to address this problem. Furthermore, this research focuses on the cosmetics industry, so the generalization of results to other industries is also not guaranteed, future research could focus on other industries. Since this research only considers low-priced brands, future research could compare it with luxury brands to see if there are any differences.

YouTube vloggers' impact on brand equity and purchase intention for low-priced brands