Understanding the relationship between celebrity endorsement on social media and consumer purchasing intention

Article history: Received 21 October 2018 Received in revised form 10 February 2019 Accepted 19 March 2019 The main purpose of this paper is to understand the relationship between the use of celebrity endorsement on social media advertising and consumer purchasing intention. The overflow of advertising on social media is making customers increasingly indifferent to the effects of advertising. There is a research interest to find out how to get consumers attention to social media advertising. Celebrity endorsement has been a strategy long used on traditional media advertising. There is currently, also a need to understand the effect of celebrity endorsement on social media advertising and on online purchasing intention. This paper studies the effect of celebrity endorsement on purchase intention online amongst university students. A conceptual model was constructed with celebrity attractiveness, gender, credibility, celebrity versus non-celebrity, single versus multiple endorsements as independent variables and purchase intention as the dependent variable. A sample of 200 university students was approached to complete a questionnaire designed from the literature as part of a descriptive, crosssectional survey. Statistical reasoning was used to test for correlations and regression. The results supported the notion that celebrity endorsement is a useful method in stimulating consumers’ purchase intention on social media advertisements. However, in comparison with similar research in the field, the notions that the credibility and the attractiveness of the celebrities as being effective on online consumer purchasing intention were not supported. An explanation was offered based on the concept that online trust is mainly derived from peer’s reviews which allows consumer involvement in the effect of social media advertising on purchasing intent unlike like it was in the traditional media.


Introduction
*Companies are constantly striving to gain and retain customers by means of developing their brand image and reputation. Primarily, this calls for better communication strategies with targeted markets. The ever changing competition and the development of the Internet in the new millennium have created all forms of media which are flooded with advertisements (Malik and Qureshi, 2016). Customers have become indifferent to the effects of advertising as they are frequently observed switching channels on TV or skipping when on social Media to avoid advertisements. Companies have long realized that they need a strategy which will make consumers pause and take notice and probably be inspired by advertisements. Celebrity endorsement is presented as one way of doing this in the marketing literature. It involves using celebrities in advertisements. Celebrities are well known personalities to the public whether locally or internationally. Their fame arises from their roles in life, in movies, entertainment, sports, and business. Zipporah and Mberia (2014) considered celebrity endorsement as the hook that will grab the customer's attention on the advertisement. Using celebrities as brand ambassadors is said to facilitate customer engagement as celebrities are perceived to be marketable, inspirational and therefore have influence on people. On social media, people see them as role models. (Anosh and Hamad, 2015). Following celebrity lifestyles affects the purchase intention of consumers (Ahmed et al., 2015). According to Rana and Tuba (2015), celebrity endorsement attracts customers and builds awareness. They also help the customer to remember easily about the advertisement and hopefully the product as well (Shukre and Dugar, 2013). Celebrity endorsement transforms an advertisement and takes it to a new level where visibility of the product is enhanced (Pughazhendi and Ravindran, 2012). Marketers need to find ways to capture the attention of consumers to deliver their message. Celebrity endorsement is a traditional way of attracting consumer attention to advertisements. There has been an escalating interest in research on purchasing intent of consumers on social media sites. We need to understand the impact of celebrity endorsement and how it affects social media advertising and online purchasing intention.
The main purpose of this study is to understand the relationship between celebrity endorsement on social media and consumer purchasing intention. It seeks to explore celebrity endorsement on social media advertising and its impact on the purchasing intent of university students who are from the age group described as the most frequent users of social media.

Literature review
Social media is a phenomenal development that has managed to change how people interact (Ertugan, 2017). It is described as a connection between brands and consumers offering a personal channel for user-centred networking and social interaction (Chi, 2011). Mangold and Faulds (2009) argued that businesses need to learn using social media along with their business plans. Social Media and thus the Social Networking Sides present an opportunity for businesses to reach larger markets. Social networking sites like Facebook, Instagram and Twitter have become outstandingly popular that businesses now make use of them in advertising campaigns.
Celebrity endorsement is a form of advertising that uses famous people. Celebrities come from all walks of life and may be movie stars, prominent business people, sportspeople, artists and the like. According to Pughazhendi and Ravindran (2012), marketers have to put their messages across in such a way that the message is noticeable, memorable and therefore unique. One such way is through celebrity endorsement. Anosh and Hamad (2015) argued that through celebrity endorsement a company can introduce special ways of advertising to their brands. Sabunwala (2013) also agreed with this notion and argues that it is a popular advertising trend and a great way to promote branding and marketing in general. Ligi (2014) explained that celebrities are people in the public eye and appear in magazines and tabloids which are a source of information already. Shukre and Dugar (2013) put forward that celebrity endorsement appeals to affluent people. Those concerned with social status want to be seen wearing the right thing, using the right product and appearances matter to them. Celebrity endorsed advertisements would definitely appeal to them and complement their lifestyle. This is supported by Ahmed et al. (2015) who stated that celebrities are role models and consumers emulate them. They added that sometimes it does not end up in sales but it definitely increases the purchase intention.
The AIDA (Awareness, Interest, Desire, Action) a classic marketing model by Strong (1925) advocates that advertising should be aimed at persuading and creating awareness in such a way that a consumer feels the desire to make a purchase. Celebrity endorsement definitely grabs the consumer's attention in a memorable way. Sheth and Solanki (2015) explained that there are products that consumers purchase frequently which may just look and perform the same as other brands (low involvement products). They said that these are the products where more persuasion is needed and where celebrity endorsement would be ideal. The Elaboration Likelihood model attempts to explain that persuasive messages can change attitudes. Jamil and Hassan (2014) stated that consumers are more likely to be convinced if they believe that the celebrity has a certain attachment to the product and that it means something other than just doing it for the financial benefit.
Gender provides the distinction of being male or female. According to Sawatari (2005), generally there is stereotyping where gender is concerned in advertising. In sports advertisements women get portrayed as weak and focus is on their beauty. They are usually considered in more 'feminine' sports as compared to men. Men portray strength in advertisements whilst women portray beauty and sex appeal. There are different schools of thought where gender is in question. Some studies argue that celebrity's gender has no effect on customer purchase intention (Friedman and Friedman, 1979) whereas others studies argue otherwise (Lafferty et al., 2002). Malik and Qureshi (2016) put forward that celebrity's gender matters depending on the type of product. They depict that different advertisements have different gender orientationsbeauty and food products advertisements are more female, while clothing, banking and financial service advertisements are more male oriented. According to Phang and de Run (2007) old studies viewed women in traditional roles and they were portrayed as nurturing, domestic and caring and men were portrayed as successful, materialistic and with leadership. Society has since moved from these notions and given the promotion of gender equality advertisements have changed in nature as well where gender is concerned. Women still have advertisements portraying beauty but success, leadership and advancement as well. Men are also known for advertising male beauty products and such. Boyd and Shank (2004) suggested that female consumers view female celebrities more trustworthy while male consumers found male celebrities more trustworthy.
Credibility is the extent to which one can be trusted or believed. Khan et al. (2016) suggested that one can easily believe in someone whom they trust.
One also believes in the contributions of someone who has experience. According to Samat et al. (2016) celebrity credibility acts as a stimulus to consumer attitude and leads to purchase intention. They added that marketers especially look at how credible one is in selection of the celebrity to endorse the product. Celebrities, therefore, have to be convincingly honest (Apeyoje, 2013). According to Anosh and Hamad (2015) the source credibility model reflects the extent of the consumer's belief in the endorsement as well as the extent of their belief in the skills and experience of the celebrity. Credibility of the celebrity may attract the consumer without considering too much of the product credentials. According to Sallam and Abdelfattah (2017), source credibility is affected by some characteristics of the consumer comprising initial attitude, level of involvement and their extent of being authoritarian. They explained that those with a negative attitude towards the advertisement need to be persuaded by a more credible source while those with a positive view usually get influenced by a less credible source. Yoon et al. (1998) cited in Sallam and Abdelfattah (2017) suggested that highly authoritative consumers are influenced by a highly credible source as opposed to less authoritative consumers. Samat et al. (2016) suggested that attractiveness has to do with the characteristics of the celebrities that are attractive to the consumer. Jamil and Hassan (2014) added that attractiveness improved social acceptance and is an attention catcher in the media. Similarly, Ohanian (1991) depicted beauty as a greater recommendation than any letter of introduction. McGuire (1985) developed the Source Attractiveness model and explained that the aspects of attractiveness that attracted consumers were the fondness, awareness and extent of similarities to them that they saw in the celebrity. If the consumer sees something of him or her in the celebrity then they pay attention to what is being said. According to Pughazhendi and Ravhindran (2012) consumers were preoccupied with celebrities and this result in them being easily attached to something celebrities may do which may resonate with something in their lives. They added that consumers also become attached to anything that brings good memories and something that makes them feel a connection or mirrors their self-identity. Dzisah and Ocloo (2013) stated that one of the ways human beings acquired knowledge is through imitation. Celebrities who are physically attractive positively influence consumers' beliefs (Anosh and Hamad, 2015). They added that personality traits and lifestyles are also included in the attractiveness, not just the physical beauty. The Source attractiveness model by McGuire (1985) emphasizes this clarity and explained that attractiveness has to do with how likeable one is (likeability), if any similarities are perceived to exist between the endorser and the consumer and familiarity. This is the reason why well-liked public personalities are hired to endorse products (Apeyoje, 2013).
Using or not using celebrities in advertising is also discussed in the literature. Sabunwala (2013) asserted that using celebrities is trendy and results in brand success. According to Khan et al. (2016), the mere presence of a celebrity in an advertisement holds the audience's attention. They assert that the presence of their favourite celebrity is what motivates some consumer to purchase instead of the product itself (Nyarko et al., 2015). According to Erdogan (1999), the use of celebrities can surpass geographical boundaries and enter into new international markets. In this way, international celebrities overcome the challenge of culture that many companies face when breaking new ground in foreign markets. Malik and Qureshi (2016) regarded using celebrities in advertisements as important. They explained that consumers are subjected to too many advertisements and have become indifferent to them and usually do not bother to watch them. In traditional marketing consumers actually take commercial breaks as time to get that snack and do quick tasks. Malik and Qureshi (2016) asserted that marketers need something to stimulate consumers and provide a relief to the advertisements and they state that using celebrities is one way. However, Pughazhendi and Ravindran (2012) urged companies to note that even though using celebrities in advertisements can result in interest and attention, it is not a guarantee for changes in attitude towards the product or brand. Nyarko et al. (2015) noted that celebrities may introduce negative or undesirable traits and meanings towards the brand. They also added that using a celebrity may result in people ignoring the product and aiming their focus on the person.
Single versus multiple endorsements pertain to products and celebrity endorsements. According to Tamizhjyothi and Rajakumar (2011) multiple celebrity endorsement is where a number of celebrities represent one brand and single product endorsement is where only one celebrity represents one brand. If they are representing many brands or products then it would be a multiple product endorsement. Multiple celebrity endorsement may depend on the nature of the product and the image that the company wants to convey. For example when dealing with beauty products and fashion apparel it is ideal to use different celebrities as they are supposed to capture a wider audience and also represent the versatilities in body types, skin types and colour. Jayswal et al. (2013) stated that multiple endorsements are viewed by consumers with match up in mind. Hsu and McDonald (2002) also share this view and advocate for compatibility when engaging in multiple endorsements. Erdogan (1999) pointed out that there is a danger in using multiple products endorsement as that distinctness is lost when the celebrity is involved with many brands. In addition it makes it overly obvious that it is all about financial gain rather than the product itself. Consumers therefore become suspicious of the motive. Um (2008) added that when distinctness is lost negative attitudes towards the advertisement may occur and affect purchase intention. It is vital to note that multiple endorsements are costly as it involves a number of high profile people and it may also result in a case of too many cooks spoiling the meat. He explains that since there are various people with different personalities, there may be many different meanings transferred which may result in incongruence. In addition it may cause confusion (Erdogan, 1999).
According to Khan et al. (2016), purchase intention refers to the plan to purchase something at that moment or a later stage. According to Morwitz (2014) companies used this to forecast sales as well as to find out if what they are doing is going to have any effect on the customers. Younus et al. (2015) provided some factors which may affect purchase intention like packaging, value placed on the product by the consumer, their knowledge as well as celebrity endorsement. They mentioned that use of celebrities enhances product worthiness in the consumer's mind. Maliq and Qureshi 2016 added that the intent to buy can be affected by word of mouth, celebrity presence and that promotional activities act as stimulus. Younus et al. (2015) explained that companies must avail all the information to the consumers as this help the customers in making a decision. Mirabi and Goohari (2015) explained that making a decision is a complicated process and evaluations take place where consumers use their perceptions on things like quality and price. This may change their attitude towards the product. They added that companies need to be aware of what stimulates consumers to make a purchase in order to meet their needs. Nyarko et al. (2015) also emphasized that and added that this information would be useful to managers in coming up with strategies and improving communication. Fig. 1 shows the study model employed for the study with the celebrity gender; credibility attractiveness, celebrity/ non celebrity and the single/multiple endorsements as the independent variables of the study depicting celebrity endorsement; and purchase intention as the dependent variable.

The conceptual model of the study
The model proposes that celebrity endorsement comprises five variables which take part in affecting purchase intention and as with the purpose of this study the purchasing intention online. In parallel to the conceptual model of the study in Fig. 1 the hypotheses below were formulated: H1: The gender of the celebrity has a positive significant effect on consumer purchase intention online. H2: The credibility of the celebrity has a positive significant effect on consumer purchasing intention online.
H3: Attractiveness of celebrity has a positive significant effect on consumer purchasing intention online. H4: Celebrity or non-celebrity endorsed advertisement has a positive significant effect on consumer purchasing intention online. H5: The nature of endorsement whether single or multiple have an effect on consumer purchasing intention online.

Methodology
This study is descriptive in nature and correlational in type as it clearly defines its objectives, variables and makes predictions (Kothari, 2004). It was conducted at one point in time making it cross-sectional. Celebrity endorsement made up of five dimensions was the independent variable with the purchasing intention as the dependent variable. The individual study unit analysis comprised of the students of Near East University. The study took place in a natural environment with minimum interference and is regarded as a field study. A quantitative, deductive logic was used. According to Kumar (2011), this type of approach can be replicated in future for verification.
A survey method was used with questionnaires as the medium to collect data in the measurement of variables and their relationship to each other. The questionnaire was adapted from the works of Carvalho (2012) and Zafoe (2013) who studied the same subject. The questionnaire was designed in two sections. The first section was on the personal details of the respondents while the second section consisted of attitude statements measuring each variable on a 5-point Likert Scale. The dependent variable was purchase intention. The independent variable of celebrity endorsements included the dimensions of gender, credibility, attractiveness, celebrity and non-celebrity endorsements, single and multiple endorsements. A pilot survey was first carried out on an initial sample of 20 respondents. This mini survey helped to refine the attitude statements to be measured to the understanding of the respondents (Kothari, 2004 A non-probability convenience sampling technique was used on a population of university students at the Near East University, northern Cyprus. Malhotra (2009) suggested that the sample size should be above 150 and more ideally around 200 to 300 respondents. The sample size for this study consisted of 200 students from the Near East University. The respondents were approached directly and asked to voluntarily complete the questionnaires.
The study used Cronbach's alpha as the coefficient of internal consistency to estimate the reliability of the responses.
The statistical software package SPSS version 21 was used to analyse the data collected from the survey. First, a descriptive analysis was carried out on the respondents then a correlation analysis to depict and the influence and direction of the relationship between the variables. Finally, a regression analysis was employed to determine the relative importance of the celebrity endorsement variables on attitudes towards purchasing intention of the respondents.

Results and discussion
All the 200 questionnaires were regarded as valid. Apart from the screening questions the Section I of the questionnaire included the personal details of the respondents. There were a total of 131 male (65.5 %) and 69 (34.5 %) female students responding. Almost half of the respondents were from the 21-23 age groups. A cumulative percent of 86.5% represented students between the ages of eighteen to 27. This may indicate that younger people shop online more than the people at older ages.
The section II of the questionnaire contained 34 attitude statements of which 28 represented celebrity endorsement and 6 represented purchasing intentions. Celebrity endorsement in return had 5 dimensions which contained attitude statements of 2 for celebrity credibility, 5 for celebrity attractiveness, 5 for celebrity gender, 8 for celebrity versus non celebrity and single versus multiple endorsements.
Cronbach's alphas used to test the consistency of the measuring instrument for the variables included in the model of the study were 0.846 (Celebrity credibility), 0.843 (Celebrity attractiveness), 0.841 (Celebrity gender), 0.844 (Celebrity/ non-celebrity endorsement, and .751 Single/ multiple endorsement respectively. According to DeVellis (2003) an ideal values of alpha should range from .70 to .95. Therefore the research measuring instrument was supported as a consistent in measuring the effects of celebrity endorsement on purchase intention.
The average responses to the attitude statements revealed that celebrity credibility is an important dimension of celebrity endorsement when considering purchasing and trust is slightly more important to them than expertise. Respondents agreed that celebrity attractiveness influences their purchase intention and respect is the most consider dimension under celebrity attractiveness. Another finding was that female celebrities command more attention than male celebrities. Respondents were almost neutral on their responses regarding celebrities using the products they endorse in real life. Respondents slightly agreed that they lose their trust in the celebrity when he or she endorses multiple brands in a short time and multiple celebrities make an advertisement dynamic. On average, the respondents moderately agreed that celebrity endorsement influences their purchase intention and that credibility is the most important attribute of celebrity endorsement that motivates them.
A 2-tailed Pearson Bivariate Correlation analysis carried on the data yielded positive correlations were: celebrity credibility and purchase intention were moderately correlated, r=529, p<0.01; celebrity attractiveness and purchase intention were moderately correlated, r=555, p<0.01; celebrity gender and purchase intention were moderately correlated, r=622, p<0.01; celebrity credibility/ non celebrity and purchase intention were moderately correlated, r=663, p<0.01; celebrity/ non celebrity/ non celebrity and purchase intention were moderately correlated, r=630, p<0.01.
Linear regression analysis carried out to determine if the five dimensions of celebrity endorsement significantly predicted attitudes towards purchasing intention. The results of the regression indicated five predictors explained 55.8% of the variance (R 2 =0.558, F(5,194)=49.014, p=0.000). Table 1 shows each dimension's contribution to the dependent variable, purchasing intention represented by ß values and p-value to show at what level each dimension is significant. Table 1 also listed the research hypotheses and depicted whether they were supported or not supported. Table 1 depicted that the gender of celebrity, Celebrity/ non celebrity endorsement and single/ multiple celebrity endorsement significantly predicted purchase intention. The credibility and the attractiveness of the celebrity is not supported by the model, displaying a low variance of 0.080 and 0.161 respectively and p values > 0.5 indicating their insignificance in explain the relationship.
The results of the analysis from the descriptive statistics, correlation and regression all indicate that all the five dimensions of celebrity endorsement are positively correlated on purchase intention. However, the hypotheses results showed that he only three of the dimensions (Gender, Celebrity/ Non-celebrity, Single/ multiple) positively affect purchase intention. Gender creates an interest in the advertisement and female celebrities grab attention more than the males. However, respondents are more likely to purchase when there is a celebrity of a gender they prefer.

Conclusion
This study investigated the effects of the use of celebrities on online advertising on purchase intention. The study revealed that the level of consumer purchase intention varies with celebrity credibility; celebrity gender; type of endorsement; the single or multiple endorsements and celebrity attractiveness. However, only the Celebrity Gender, Celebrity/ non-celebrity, Single/ multiple dimensions have a significant effect on purchase intention. The study revealed the importance of the use of celebrities on online advertising. Even though the respondents revealed scepticism when faced with whether the celebrities used the products they endorse in real life; the respondents mostly agreed that the celebrity endorsed advertisements grab their attention and facilitate prompt recall of products. It is therefore imperative that companies consider celebrity endorsement when they are looking for ways of differentiating their products and increasing awareness in consumers.
Celebrity endorsed advertisements showed a great impact on the respondents. Respondents showed that they prefer advertisements with celebrities and they are more likely to be motivated to purchase after being subjected to celebrityendorsed advertisements than those without. The researcher would like to point out as a contribution to science, that, despite the positive attitude that respondents have towards celebrity endorsed advertisements the results revealed an almost indifferent response to whether the celebrities use products they endorse in real life. A growing awareness in respondents was noted that celebrity endorsements are about business and not everything celebrities do apply in their real life. The respondents are aware that celebrities transfer meanings to products and they consider the match between celebrity and product to be important. This should be taken seriously when evaluating celebrities to be used in advertisements so that the brand will not suffer the consequences of negative meanings.
The idea that multiple celebrity and single celebrity endorsements both positively impact the purchase intention of consumers was supported. The respondents revealed that this affects their recall process and only slightly agreed that it makes them lose credibility in the celebrity and change their attitude towards the advertisement. Multiple celebrity endorsements were found to be dynamic and appealing. It is therefore vital that a company engages many celebrities to create a dynamic impact and appeal to the masses.
In conclusion, celebrity endorsements are a powerful advertising tool and marketers should take advantage of them as well as the growing popularity in social media platforms to appeal to wider audiences. They appeal more to the masses and reduce the numbing effect of the numerous advertisements that consumers are subjected to on a daily basis. However, marketers should not overlook the drawbacks associated with celebrity endorsements. They therefore need to really do thorough evaluations of celebrities, keep abreast of developments as well as arm themselves with sufficient information to avoid possible damage to brands and organization's reputation as a result of engaging celebrities. Celebrity endorsements should be embraced because they have a positive effect on the purchase intention of consumers.
In comparison with similar papers in the field (Khan et al., 2016;Samat et al., 2016), this paper has reached dissimilar conclusions on the notions that the credibility and the attractiveness of the celebrity were effective on online consumer purchasing intention. This dissimilarity cannot be explained by the sample size and characteristics as similar sample design was also employed by this research. The explanation is considered to be the effect of the concept of "Trust" on online consumer behaviour. Customers are increasingly observed to trust more to peers' reviews on online advertising. Past research on celebrity endorsement is carried out on the traditional media. The research carried out on social media is relatively recent and further research is needed to understand its effects on purchasing intent and other consumer behaviour.

Limitations
Due to time constraints this study was limited to the students at Near East University only. It also used a non-probability sampling technique of convenience sampling which does not really allow generalization of results over a whole population. This study was also based on a quantitative approach and hence only employed quantitative methods. Future studies can be conducted taking these limitations into consideration.

Recommendations for future research
Future research is recommended to focus on a wider population with a study model that includes moderators such as age and gender. Organization's perspective should also be taken into consideration to assess the effects of celebrity endorsement on the goals of the organization. The cultural effect in celebrity endorsement amongst culturally different societies should also be considered for investigation. Finally, future research needs to be carried out on the possible negative effects of celebrity endorsement and how to avoid them.