Keywords

1 Introduction

With the development of information technology, Internet has taken people’s daily life beyond physical limitations. Internet allows them to have a quick access to information, learn whenever and wherever they want, interact with others, and share experiences online. This new phenomenon where people gain information from online social groups coined a new business model, known as Internet celebrity economy. It is a business chain that includes several elements, from Internet celebrities, social media platforms, e-commerce websites to advertising agencies. Internet celebrities conduct marketing campaign via social media and influence their fans’ shopping decisions. Internet celebrities, known as “Wang Hong” in Chinese society, have become famous on social or live streaming media. Characteristics of Internet celebrities include a high number of followers, and the abilities to advance their Internet traffic to real cash. Most Internet celebrities show their brilliance, professional skills, or unique personality via social media to attract audiences’ eyeballs and keep customers sticky. Throughout this innovative business chain, Internet celebrities are responsible for keeping the platform’s network traffic busy and achieve their commercial goals via social media (Xu 2017).

One can categorize Internet celebrities in three different types, product sellers, content creators, and public figure. The first one finds its origin on Taobao, a leading online shopping site in China. Product sellers are good at maintaining a good interaction with followers and guiding them to online stores from a live streaming event. Content creators work on different channels to deliver visual contents, such as articles or video clips, to their followers. They make money from their fans by stealth advertising because Chinese audience does not favor display advertising. Also, one also sees certain controversial public figures enjoying their popularity on the Internet.

Zhang Dayi and Papi Jiang are two famous Internet celebrities in China. Zhang Dayi sells clothes and cosmetics on Taobao. She makes a considerable amount of revenues by posting pictures of herself wearing the clothes she sells. Since she is her own shop’s only model, some of her followers became her loyal customers. Her annual income was reported as RMB300 million (US$46 m) in 2016 (Tsoi 2016). On the other hand, Papi Jiang, a fashionista, creates original content to attract fans. She posts satirical video clips on social media, such as Weibo and YouTube. In those video clips, Papi Jiang talks about her daily life, show business, hot issues and relationships in a sarcastic manner. She secured a joint investment of RMB12 million (around US$1.84 million). This piece of news surprised the audience and demonstrated that Internet celebrities have the abilities to advance their Internet traffic to real cash.

While China’s Internet celebrity industry started developing rapidly in 2014, some western Internet celebrities have become key opinion leaders (KOL) in America and Europe. Youtube is an important Internet celebrity incubator. Content creators can earn money from advertisements found on their video clips and from YouTube Premium subscribers watching their content. Many famous YouTubers, such as PewDiePie, Germán Garmendia, and Jenna Marbles, have racked up millions of subscribers over the years through a direct relationship with their fans. Following Facebook and Youtube, Instagram (also known as IG) enhances the trend of Internet celebrity. Instagram provides photo and video-sharing social networking service and supports Internet celebrities to attract followers rapidly. For example, Selena Gomez, an actress, and singer, has more than 140 million followers, while Cristiano Ronaldo, one of the world’s leading soccer players, has more than 150 million in January 2019Footnote 1. Some Internet celebrities successfully established their brands on IG, and sell products effectively. Hence, with the developing of social media and mobile devices, social media stars are becoming more influential in the celebrity marketing. Furthermore, Platforms, MCN (multi-channel network) and Internet celebrity are working together to improve the efficiency and enlarge the size of Internet celebrity economy.

While the phenomenon of Internet celebrities continues to grow, what matters to business owners the most though, is whether Internet celebrity economy is taking effect? Past studies on celebrity endorsement mostly focused on experts or celebrities in the entertainment industry via conventional platforms, such as television, newspapers or magazines. Research from Amos et al. (2008) shows that the endorser’s characters, such as his/her professional knowledge or physical attractiveness and so on, help achieve a better advertising result. As using Internet celebrities for product endorsements only started a few years ago, there have been very few analyses of the effect of interactive Internet celebrity’s endorsement via social media platforms such as live-streaming, Facebook or Weibo. The few pieces of research on Internet celebrities focus mainly on how they manage their audience. Cocker and Cronin (2017) point out that Internet celebrities get their audiences’ attention by sharing their daily activities, ideas or shopping experiences online, and interact with their followers. Marwick and Boyd (2011) think that Internet celebrities take contexts and their audiences into account, and maintain a process of managing impressions by presenting tailor-made performances. However, rare studies that explore how Internet celebrities influence consumers and their purchase intention.

The purpose of this study is to investigate the effect of the Internet celebrity’s endorsement. As this newly-emerged advertising method has not been verified and campaign field not explored, whether the result will be the same as those from the past studies seems unknown at the moment. Against the backdrop as mentioned above, this study proposes three main objectives:

  1. 1.

    Explore the influence of the endorser’s characters on the endorsement effect.

  2. 2.

    Analyze the influence of the consumer’s characteristics on the endorsement effect of Internet celebrities.

  3. 3.

    Explore whether the interactive atmosphere created by Internet celebrities for the endorsed products influence the endorsement effect.

This paper consists of five sections. The first explains the motivation and objectives of the study. The second describes the theoretical foundation and the development of hypotheses. The third explains the design of the empirical study and the survey method by using questionnaires. The fourth tells the results of the analysis. The fifth are the conclusion and suggestions for future studies.

2 Theoretical Foundation and Hypotheses Development

2.1 Celebrity Endorsement in Social Media Marketing

Internet celebrities are individuals with a pronounced characters and charisma. Through multiple ways of presenting themselves, they disseminate messages quickly and gain attention from a large number of users. Their followers are netizens and fans of a certain type. Internet celebrities exercise a high capacity in disseminating information and influencing people in a specific domain or group. The author considers that Internet celebrities have the following pronounced characteristics: 1. They draw support from Internet platforms to become famous. 2. Each has a distinctive personality with features based on his/her physical appearance, expertise, such as creative writing, or knowledge in a particular area, such as fitness or beauty. 3. They continuously and routinely create content online. 4. They build a strong bonding with their fans and possess a strong interactive ability. 5. They have a large number of fans and fans tagging. 6. They exercise a high capacity in disseminating information and influencing people.

To generate revenues, Internet celebrities will become the endorser for a company or a particular product. An endorser is any public figure, such as an actor, an actress, an athlete, an entertainer or a political figure, who uses his/her popularity or personal achievement to recommend a product or a service to the public through any kind of format for a company or an organization (McCracken, 1989). As Internet celebrities have many fans, they have the advantage of becoming an endorser to promote a product or a service. Freiden (1984) categorizes endorsers as four types, the celebrity, the expert, CEO, and the typical consumer. According to an empirical study by Friedman and Friedman (1979), when a product involves in high financial, performance, and physical risks, using an expert will reach a better endorsement effect. When a product involves high psychological and social risks, using a celebrity will achieve a better endorsement effect. There is no evident difference between using a celebrity or an expert as an endorser when it comes to general consumption goods. From the statements above, one learns that different types of endorsers create different kinds of endorsement effect. Wang (2005) points out further that the appropriate use of endorsers will enhance the consumer’s behavioral intention. It is evident that professional opinions provided by experts intensify the attitude toward an endorsed product. However, for a product that requires user’s experience, viewers of the advertisement are prone to develop a sense of trust and better behavioral intention because they can identify themselves with typical consumers.

Different types of endorsers may have different endorsement effects. With the newly emerged Internet celebrity economy, their way of endorsing a product may influence the endorsement effect differently. However, in today’s society, the roles of different types of endorsers tend to overlap, so their boundaries become vaguer and vaguer. As far as how Internet celebrities advance Internet traffic to real cash in China, there are several ways. 1. Product selling. Internet celebrities use Weibo windows to display products to their fans and direct them to e-commerce platforms, such as Taobao. 2. Marketing endorsement. Internet celebrities advertise for major brands on social platforms, such as Weibo or WeChat by placing advertising messages cleverly via different forms, such as in a blog article or a video clip. By doing so, they earn endorsement fees from sponsors. 3. Being a content producer. They make money by selling their content IP (intellectual property). An example is Zhang Jiajia. She initially became famous by telling bedtime stories on Weibo. Those stories later became a book under the title “I Belonged to You,” which was later adapted to a script and then a film. Thus, Zhang also became a scriptwriter and a director. 4. Receiving tips from fans. Through interacting with fans on live streaming platforms, such as Douyu, Huajiao, Panda or Miaopai, Internet celebrities receive virtual gifts as tips from fans. 5. Becoming entertainers and participating in productions of online dramas, TV series or movies.

Through observations on various Internet celebrities’ modes of endorsement, this study concluded that they are as the following. 1. Via social platforms such as online live streaming, Weibo or Facebook. 2. Independent creations of content, including online live streaming, video clips, images, and texts, to present products. 3. Stimulating fans’ interests for the endorsed product by interacting with them. 4. Encouraging fans to forward so the content can go viral. 5. Redirecting to online stores by displaying a shopping link, or to a physical store. Table 1 summarizes comparisons of different endorsement methods between an Internet celebrity and a non-Internet celebrity.

Table 1. A comparison chart of endorsement methods between an Internet celebrity and a non-Internet celebrity.
Full size table

2.2 Research Hypotheses

The process of Internet celebrities’ endorsement is via interaction with netizens (or fans) on social media or other online platforms. Since digital media’s content is presented as multi-media, the content by an Internet celebrity’s endorsement can be considered as a message, Internet celebrities themselves as message transmitters, and the party that interacts with them as message receivers. One can infer how the message of endorsement influences the message receiver’s attitude, and furthermore his/her purchase intention by applying the Elaboration Likelihood Model (ELM; Petty and Cacioppo 1981). According to ELM, during the process of attitude change or formation, the message is more persuasive when the receiver holds a higher value toward its source. When the message’s source credibility is low, consumers will have doubts about the message (Eagly and Chaiken 1993; Zhang and Buda 1999). When an endorser has higher credibility, the message receiver has a more positive understanding of messages in the advertisement, and attitudes toward the advertisement and brand.

While a message receiver deliberates the content, it is possible that receiver will go through in-depth thinking, analyze and evaluate the message content. This is called the central route. On the contrary, it will be called a peripheral route when a message receiver makes his/her judgment based simply on a positive or negative element of the message without in-depth analysis. Via which path a message receiver deliberates a message depends on his/her ability and motivation. When message receiver has better expertise or a higher motivation deciphering the content of a message, the central route will be taken. Otherwise, the message is deliberated via a peripheral route (Petty and Cacioppo 1981). As a result, the author proposes the following hypophyses:

H1::

The message’s source credibility will positively affect the consumer’s brand attitude.

H1a::

The consumer’s expertise level moderates the association between source credibility and brand attitude.

H1b::

The consumer’s degree of product involvement moderates the association between source credibility and brand attitude.

H2::

The message’s source credibility will positively affect the consumer’s attitude toward advertising.

H2a::

The consumer’s expertise level moderates the association between source credibility and attitude toward advertising.

H2b::

The consumer’s degree of product involvement moderates the association between source credibility and attitude toward advertising.

H3::

The content quality will positively affect the consumer’s brand attitude.

H3a::

The consumer’s expertise level moderates the association between content quality and brand attitude.

H3b::

The consumer’s degree of product involvement moderates the association between content quality and brand attitude.

H4::

The content quality will positively influence the consumer’s attitude toward advertising.

H4a::

The consumer’s expertise level moderates the association between content quality and attitude toward advertising.

H4b::

The consumer’s degree of product involvement moderates the association between content quality and attitude toward advertising.

Social media created an online virtual community in which social awareness is present, similar to the one found in a community’s behaviors and process (Blanchard and Markus 2004; Forster 2004). To put this a more concrete way, this sense of virtual community (SOVC) is defined as feelings, identity, a sense of belongings or association, and attachment and association to a group of which members share the same interest or goal (Koh and Kim 2003), which involve four aspects, its members, influence, fulfillment of needs and emotional connection. Past studies that explored SOVC reveal a new discovery in the purchase intentions of live streaming room users. The identity, emotional connection and influence of the members during live streaming are all associated positively with the users’ purchase intention. As a result, users are more willing to purchase virtual gifts on real-time media. This is because members have a sense of belongings, and consider themselves as part of the live streaming room. This recognition propels them to contribute, and take part in their “community” (Kim et al. 2004). Kim et al. (2009) also discovered that a higher SOVC might increase the members’ purchase intention. As a result, we argue that a virtual community helps enhance its members’ attitudes toward the brand and the advertising, and therefore has an influence on the purchase intention, and proposes the following hypotheses:

H5::

SOVC will positively affect the brand attitude.

H6::

SOVC will positively affect the attitude toward advertising.

According to Bauer (1960), consumers select products based on the brand’s popularity or reputation. These two elements also serve as an assurance to reduce risks of the product they purchase. In consumers’ decision-making behaviors, the purchase intention has a significant influence on the purchasing policy. The factors that lead consumers to be willing to purchase a product depend on consumers’ belief, attitude, quality assessment and perceived value toward the product or the brand. Zeithaml (1988) proposes using the perceived value as a way to measure the purchase intention and uses maybe buying, desire to buy and considering to pay as variables for the item variables. Purchase intention arises from consumers’ evaluation of the product or brand attitude, as well as action triggers that serve as extrinsic factors, such as advertising message.

Chaudhuri (1999) points out that the brand attitude will influence the decision maker’s purchasing behavior. It may affect his/her loyalty to the brand. A positive brand attitude will lead the consumer to trust the brand even more and to think that there is a lower possibility of uncertainty when purchasing the product, thereby, increase the consumer’s purchase intention. Good brand and attitude toward advertising make it easy to resonate in consumers and persuade them. Consumers, therefore, will have interests in the product and purchase it. As a result, this study proposes:

H7::

Brand attitude will positively affect the purchase intention.

H8::

Attitude toward advertising will positively affect the purchase intention.

Based on the process of a message communication and the consumer’s experience in the virtual community, this study establishes a research framework (as shown in Fig. 1) from the influence created during the Internet celebrities’ endorsement process, the content of the endorsement, and the SOVC created while interacting with fans, to how the process influences brand attitude and attitude toward advertising, and furthermore the consumer’s purchase intention.

Fig. 1.
figure 1

Research framework

Full size image

3 Research Method

3.1 Developing a Questionnaire

A questionnaire was developed for the survey used in this study. It is divided into four sections. The first section is a photo of the endorser to examine to what extent is the endorser familiar to the consumer, and his/her popularity among the fans. The second section is the demonstration of an actual case where an Internet celebrity endorsed for a product. The demonstration is presented via a video clip in which a simulated situation of the respondent experiences the whole process of a commercial clip on Weibo. The third section is to examine every question of the study’s constructs. The fourth section is relevant questions of variables demographic statistics.

To control the effects of product types, we chose two types of products for the survey. One product is clothing (a consumer product), and the other is gaming (an entertainment product). In terms of selection of Internet celebrities, an e-commerce type and a gaming type were chosen respectively. The former mainly advertises and endorses her brand, and the latter for different game companies and other advertisements. Thus, there are two types of Internet celebrities and products respectively. Besides, the author also used a control group by adding a movie star as an endorser. The combination became four different kinds, an Internet celebrity and a movie star with a clothing product, and an Internet celebrity and a movie star with a gaming product respectively. By having four different combinations, one could avoid potential influence on the statistic results as a single endorser, or a single product would have.

In the construct measurement, this study used the three dimensions of an endorser’s source credibility proposed by Ohanian (1991). They are attractiveness, expertise, and source credibility. 15 items are developed based on the message’s source credibility, i.e. each dimension consists of five items. The quality of the content consists of five items (Zhang 1996). The consumer’s expertise level is measured by three items (Sussman and Siegal 2003). The degree of the consumer’s involvement is a revised version based on a measure designed by Laurent and Kapferer (1985), and five items were created for the study. In the construct for SOVC, we adapted the questions proposed by McMillan and Chavis (1986), Blanchard (2007) and Hsieh et al. (2018) to design items, including members, emotional connection, influence and fulfillment of needs. There are four second-order constructs, and each construct has three items for the evaluation. Furthermore, there are five, four and three items for measurements corresponding to brand attitude, attitude toward advertising, and purchase intention respectively. All items are based on Likert’s five-point scale.

The survey was conducted online. The fill-in process is as the following. When a respondent opens the questionnaire, the respondent will first see a photo of an Internet celebrity endorser. The respondent gives a truthful answer as to whether he/she recognizes or knows about the endorser. Four questions were added and still used the Likert’s five-point scale. The questions include:

  1. 1.

    I think the person in the photo is (endorser’s name).

  2. 2.

    I think (endorser’s name) is a famous Internet celebrity.

  3. 3.

    I am a fan of (endorser’s name).

  4. 4.

    I usually visit the social media’s account (Weibo or live streaming) of (endorser’s name) to learn about her latest development.

After answering the above questions, a video clip will be shown to the respondent, who will see a simulated situation of he/she watches a presentation of an endorsement campaign, from an advertising copy, images or animation, netizens’ comments to the content of discussions. After viewing the video clip, the respondent will carry on answering items in each construct and of variables concerning demographic statistics.

3.2 Survey Administration

As the development of Internet celebrities in China is phenomenal, and the scale of Internet celebrity economy considerable, this study chose China as the research scope to collect samples, and Internet celebrities on Sina Weibo as the objects of the study. There are three reasons for choosing Internet celebrities on Sina Weibo. Firstly, Sina Weibo had 392 million active users as of Feb 2017. It has vast and wide-range users. Secondly, most of Internet celebrities in China choose to set up a Sina Weibo account and run a social group from there. Thirdly, Sina Weibo works closely with major mobile phone makers and top mobile phone apps to build a complete ecosystem for its platform. As its sales campaigns and online content become more and more complementary to each other, Sina Weibo provides a robust platform for the development of Internet celebrity economy. An online questionnaire was used to collect data. The pilot run’s sampling period was between March 15 and 24, 2018. 60 copies of valid pilot-run questionnaires were collected. After verifying the reliability and validity of all constructs, the result indicated that the items in the model constructed in this study proved to have good reliability and validity. A formal survey could be conducted.

This study used online questionnaires for data collection. The distribution period was between April 15 and May 20, 2018. 512 respondents were collected at the end of the survey. 466 copies were valid after removing questionnaires with foolproof questions being answered or careless responses. The valid answer rate is 91%. The gender ratio of valid samples is 36.3% male and 63.7% female. The main age distribution is between 26 to 30 years old (47%), and then 21 to 25 years old (37%). In terms of occupation, office workers are the majority (55.4%), and full-time students (37.1%) come in second. Most of the respondents (42.5%) spend daily an average of one to three hours online. When asked how frequently they took the initiative to follow an Internet celebrity, most of them (46.1%) answered they rarely followed.

4 Research Results

To test the study’s framework and different hypotheses, this study used PLS method for analysis, and used the Multi-Group Analysis (MGA) from SmartPLS 3.0 to analyze the discriminant effect of endorsements by an Internet celebrity and a movie star.

4.1 Reliability and Validity

Firstly, a test was conducted on the reliability and validity of each item’s construct. Reliability was tested by Cronbach’s Alpha value and composite reliability, of which each construct is above 0.86, demonstrating each measuring item is consistent and stable. Validity was tested by the average variance extracted (AVE), and each construct is above 0.5. While applying factor loading to verify the construct validity of each sub-item, each construct reached the threshold of 0.65 (Hair et al. 2006). Also, the discriminant validity analysis was examined. Based on the suggestions by Hair et al. (2006), the discriminant validity of each construct in this survey is defined as the following, the correlation of two different constructs should be smaller than the square root of each construct’s AVE, as the values presented diagonally in the Table 2. As shown in Table 2, the square roots of the AVE are all greater than all other cross-correlations. This proves that each construct has a certain degree of discriminant validity (Hair et al. 2006).

Table 2. Discriminant validity and correlation matrix
Full size table

4.2 Result of PLS Analysis

We tested the Hypotheses H1 to H8 with the overall sample. The PLS results of the structural model show that the path coefficient of the influence of the endorser’s source credibility on brand attitude and attitude toward advertising are significant (β = 0.26, p < 0.001; β = 0.287, p < 0.001). Therefore hypotheses H1 and H2 are supported. However, the path coefficient of the moderating effect of the consumer’s expertise level and degree of product involvement did not reach the significance level. Therefore H1a, H1b, H2a and H2b are not supported. The path coefficient of the influence of the content quality on the brand attitude did not reach the significance level, but showed an effect of the marginal support (β = 0.153, p < 0.1). Therefore H3 is marginally supported. The path coefficient of the influence of the content quality on the attitude toward advertising is not significant (β = 0.126, p > 0.05). Hence, H4 is not supported.

Moreover, the path coefficient of the moderating effect of the consumer’s expertise level and degree of product involvement did not reach the significance level. Therefore, H3a, H3b, H4a and H4b are not supported. The path coefficients of the influence of the SOVC on the brand attitude and attitude toward advertising are significant (β = 0.368, p < 0.001; β = 0.288, p < 0.001). Therefore H5 and H6 are supported. The path coefficient of the influence of the brand attitude and attitude toward advertising on the purchase intention reached the significance level (β = 0.619, p < 0.001; β = 0.216, p < 0.001). Therefore, H7 and H8 are supported.

This study compared the effects of endorsements between an Internet celebrity and a movie star. The result shows that from examining (without moderating effect) the outcome of the structural model of a movie star’s endorsement, the path coefficient for the influence as an endorser’s source credibility on the attitude toward advertising did not reach significant level (different from an Internet celebrity’s endorsement), but reached the marginal support one (β = 0.186, p < 0.1). Furthermore, the path coefficients for the influence of the content quality on both brand attitude and attitude toward advertising both reached significant levels (β = 0.19, p < 0.05; β = 0.373, p < 0.001), which is different from the outcome of the Internet celebrity’s endorsement. However, the moderating effects of the consumer’s expertise level and degree of product involvement are not significant. No significant difference was observed between endorsements by Internet celebrities and movie stars in other areas.

5 Conclusion

This study aims at exploring the influence of an Internet celebrity’s endorsement on the consumer’s purchase intention by focusing on three areas for investigation. They include the Internet celebrities’ characters (i.e. source credibility), the published content (i.e. its quality) of the endorsed products, and the overall atmosphere created (i.e. SOVC) by the Internet celebrities. Based on the study’s framework, the endorser’s credibility and SOVC are both positively associated with the consumer’s brand attitude and attitude toward advertising and increases their purchase intention. However, the influence of the content quality on the brand attitude and attitude toward advertising did not reach the significant level. We argue that in the context of Internet celebrities’ endorsement, consumers are prone to be attracted by the Internet celebrities’ characters and immerse themselves in the influence of SOVC without paying too much attention to the content of the endorsed product published by the endorser. Instead, a peripheral route is taken upon receipt of the product’s information. The inference and association made by the consumer are simple without deliberating about the content.

Also, the moderating effects of the consumer’s expertise level and degree of product involvement are not significant. One potential reason is that this is because the mean of these two factors based on the chosen product categories, clothing and gaming, is not high, and even toward a lower end. Furthermore, these two product categories do not require a high purchase or usage cost. Consumers for these two categories do not have enough motivation and the ability to learn about the products. As a result, there is no significant influence on brand attitude and attitude toward advertising.

The difference between a product endorsement by movie stars and by Internet celebrities is that the former’s source credibility on the influence of attitude toward advertising did not reach a significance level. We infer that when the image of a movie star is positive, it will lead the consumer to associate it with the brand image. Therefore, the brand image is enhanced positively without much influence on the consumer’s attitude toward advertising. For the endorsement by movie stars, the influence of the content quality on brand attitude and attitude toward advertising both reached a significance level, but the same situation was not observed in the endorsement by Internet celebrities. We infer that most of the information published in a movie star’s endorsement campaign is advertising copies by the product sponsor. When a movie star has a relatively positive image, the more outstanding an advertising copy (i.e. a higher content quality) is, the more significant influence there will be on brand attitude and attitude toward advertising because consumers equate the image of the movie star to the one of the sponsors. In the context of an Internet celebrity’s endorsement campaign, consumers pay more attention to the Internet celebrity him/herself.

According to the result of this study, the characters of an Internet celebrity is a paramount factor to an endorsement campaign. This is because source credibility enhances significant brand attitude and attitude toward advertising, and therefore the purchase intention. Thus, in a business process, Internet celebrities need to continually maintain and improve their attractiveness, expertise, and source credibility, as well as add a personal flavor to attract the consumer’s attention. They also have to pay attention to their own life because many times consumers judge whether an endorser is trustworthy or not based on some of his/her everyday behaviors. At the same time, SOVC plays a vital role. In the process of the interaction between the consumer and Internet celebrities, a new social circle is formed, and a feeling of belonging to the same community emerged. Internet celebrities are therefore encouraged to develop their communities, and exert their formidable abilities to interact and to run social groups to interact with fans and share experiences. All of these create a sense of membership, influence, emotional connection and fulfillment of needs among fans in a social group. When a member considers him/herself as part of a community, a belief is formed to want to contribute and grow with it. This will help with the brand attitude and attitude toward advertising for the product endorsed, and will even promote a purchase intention so that the member shares the same thoughts as other members and share the same shopping experiences.

While there is no salience effect on the content quality in an Internet celebrity’s endorsement campaign, the effect is evident in a movie star’s one. Nowadays as the role of an endorser is getting vaguer and vaguer, a movie star can also be an Internet celebrity or vice versa. Paying attention to the content quality of an endorsed product still merits great attention. Furthermore, as there is no significant difference between a product endorsed by an Internet celebrity or by a movie star, sponsors with budget concerns are advised to choose the former as the costs may be lower and conveys distinctive characters. If the Internet celebrity’s personality happens to be identical to the brand image, targeted marketing’s purpose will be fulfilled.

In terms of future studies, this study explored the influence of an Internet celebrity’s endorsement on the consumer’s purchase intention. However only China was selected as the research area. Studies in the future may choose multiple locations. Different cultural backgrounds may lead to different research results. Moreover, this study only chose one platform, Weibo, and investigated how Internet celebrities endorsed a product on it. Studies in the future can extend the research spectrum to other platforms, such as live streaming, explore the characteristics of endorsement campaigns by an Internet celebrity on such platform, or compare the differences of endorsement campaigns on various platforms.