Impact of Social Media Influencers on Customer Engagement and Purchase Intention: A Meta-Analysis

Abstract

This research aims at providing a meta-analysis of empirical findings of the literature on the characteristics of social media influencers on customer engagement and purchase intention. For this purpose, researchers derived eight social media influencers’ characteristics, i.e., homophily, expertise, trustworthiness, credibility, congruence with the product, entertainment value, informative value, and attractiveness. The current study synthesizes 176 effect sizes derived from 62 individual studies, and 22,554 individuals act as an aggregate sample. Results revealed that these characteristics have a moderate to high correlation with customer engagement and purchase intention. The entertainment value of social media influencers has the strongest association with customer engagement among all the attributes studied in this analysis. It also concluded that the credibility of influencers impacts purchase intention more than any other attribute. This work provides a novel approach to reducing the heterogeneity in influencer marketing research by empirically specifying the directions of relationships and the extent of the effect of these relationships.

1. Introduction

The popularity of social media among both the young and older generations has led to its widespread usage as a new marketing communication platform [1]. Customers are exposed to a variety of marketing initiatives that businesses promote and carry out quickly, without regard to space or time constraints, in the digital realm of social media networks [2]. Because of social media platform features like live chats and commenting capabilities, individuals may now have direct digital interactions with their favorite brands, famous influencers, and other users [3]. Social media users can communicate their attitudes and ideas about a brand’s products or the actions of other users by simply pressing the “like” or “dislike” button or by posting a favorable or unfavorable comment on a post.

As customers’ reliance on social media apps for decision-making grows (e.g., product reviews), “social media influencers” have been developed and are now being used by businesses as a new channel for promoting a product in the digital era [2]. The term “social media influencer” refers to well-known individuals with expertise in a particular industry, such as music, health, travel, or fashion, who produce and disseminate their information-focused online content to other social media users. Users of social media tend to view social media influencers who are knowledgeable in their specific fields as more convincing and authentic than general celebrity endorsements on social media and in conventional ads [4]. In addition, social media also influences all forms of business. They used to promote their products at low cost, improve communications with customers and other stakeholders, and increase their reach to many customers.

Influencer marketing has grown greatly in terms of numbers. The influencer market increased from $1.7 billion in 2016 to $9.7 billion in 2020 [5]. It rose to $13.8 billion in 2021, showing a consistent increase [6]. The market will have grown to a massive $16.4 billion business by 2022 [7]. An influencer quickly becomes a crucial part of a brand’s marketing strategy [8]. Brands are increasingly working with social media influencers as they become aware of this new chance to connect with their target audiences [9]. It has recently been recorded that 93% of marketers have employed influencer marketing in campaigns, and it is now recognized as a crucial advertising tactic [10].

The number of influencer marketing-related studies has also increased recently [11]. Additionally, an in-depth review of existing literature on influencer marketing reveals that authors have empirically investigated the role of social media influencers’ characteristics in creating purchase intention [12], customer engagement [13], brand equity [14], brand loyalty [15], etc.

However, existing empirical research provides mixed results about the impact of the characteristics of social media influencers on purchase intention and customer engagement. Results are varied in terms of the strength of the relationship between the attributes of social media influencers and customer engagement and purchase intention. For instance, ref. [16] reported a low impact of expertise on customer engagement, while [17] found a high impact of expertise on customer engagement. Similarly, Shan et al. [18] found a negative impact of homophily on purchase intention, while Shen et al. [19] concluded a positive and high correlation between homophily and purchase intention. Thus, despite an enormous number of studies, researchers have reached no consensus on the relationship between the characteristics of social media influencers and customer engagement and purchase intention.

In light of these contradictory empirical results and the growing calls from practitioners who want to apply them, an overall comprehension of the effects of influencer marketing is required. In this regard, professionals might choose their targeting, positioning, advertising, and engagement-related actions more effectively with a streamlined and clear direction. From a theoretical perspective, clarity on this topic would decrease the ambiguity in the influencer marketing literature. Meta-analysis is frequently used to consolidate findings in order to get a more comprehensive picture of a situation and to better comprehend and interpret contradictory findings [20,21].

Although researchers like [3] put in the effort to synthesize the literature through a systematic literature review, the scope of a systematic review is limited if a formal meta-analysis is not included. The outputs of the secondary data collected are appropriately summarized by a meta-analysis. In this direction, the current study aims at providing a meta-analysis of the empirical findings of the literature on the characteristics of social media influencers on customer engagement and purchase intention. This research synthesizes 176 effect sizes that are derived from 62 individual studies with a total of 22,554 individuals as an aggregate sample.

There are numerous contributions of this analysis to the existing literature on influencer marketing. Firstly, it provides a statistical compilation of research investigating the relationship between social media influencers and purchase intention and customer engagement. Secondly, by conducting a meta-analysis, it addresses the controversial results in the SMI literature. This study contributes to the literature by minimizing the heterogeneity in influencer marketing research by empirically specifying the directions of relationships and the magnitude of the effect of these relationships. Additionally, it provides guidance to practitioners in deciding which characteristics to seek in an influencer when approaching them to endorse their product.

Further, this article is divided into various sections. The literature review and hypothesis development section sheds light on the existing literature on social media influencers and their characteristics. Hypotheses were formulated in this section based on past literature. The methodology section provides details of the methodology used for data collection, data coding, and data analysis. This is followed by a results and discussion section. Afterward, limitations and future research directions, and practical implications are stated. In the end, the conclusion section is provided to synthesize this research work in a nutshell.

2. Literature Review and Hypothesis Development

2.1. Social Media Influencers

A social media influencer is “someone who has a significant and active following on social media platforms, which one would not know unless one follows them” [22]. Influencers publish material in a certain niche, such as food, travel, fitness, or fashion [22] on social media platforms like Facebook, YouTube, Twitter, and Instagram. To stay up to date on the newest trends, people or customers interested in a specific sector can follow and communicate with influencers. People have the chance to build a fan base and achieve online popularity by uploading original tales and content, which leads to the emergence of social media influencers.

Social media influencers have drawn a lot of interest from academics and business professionals because of their potential as an instrument for brand marketing. SMIs have significantly changed the face of social media marketing [23]. SMIs’ popularity accelerated the collaboration between businesses and social media influencers.

Influencers are distinct from traditional celebrities because they develop their online persona and popularity by creating content for social media platforms. Traditional celebrities, on the other hand, earn recognition through conventional means and use social networking sites as a secondary route for communication with fans. Djafarova and Rushworth [24] found that social media influencers may perform better in terms of an endorsement than traditional celebrities due to the way they communicate and interact with customers, often sharing personal information and having reciprocal interactions. Their frequent sharing of personal information and reciprocal interactions make them more relevant and approachable [25]. Additionally, social media influencers frequently incorporate sponsored posts into their daily stories, creating authentic endorsement content that consumers like [26].

2.2. Social Media Influencers’ Characteristics

2.2.1. Homophily

As in the saying, “birds of a feather flock together”, homophily can be characterized as an individual’s propensity to associate and bond with people having similar traits. Thus, homophily describes the degree to which two people who interact are similar in terms of particular characteristics, such as beliefs, values, education, and social standing. Homophily, according to De Bruyn and Lilien [27], refers to similarities between persons based on their likes, dislikes, values, and experiences. Homophilic individuals usually have traits in common that facilitate easy communication and the development of strong bonds. Numerous studies have investigated homophily in various forms and proven that homophily fosters interaction. Shen et al. [19] concluded that homophily impacts customer engagement and purchase intention. Thus, it can be assumed that:

H1. Homophily with social media influencers impacts the purchase intentions of customers.

H2. Homophily with social media influencers impacts customer engagement.

2.2.2. Expertise

Expertise is a very important factor because it is the outcome of the communicator’s knowledge and professional experience [28] and it has been taken into account in various research on digital influencers [29]. In addition, ref. [30] discovered that influencer expertise was a significant factor in influencing purchase intention. As a result, customers are more inclined to consider the content shared by influencers who are thought of as subject-matter experts [31]. “Expert is often seen as highly knowledgeable and able to provide judgments that are accurate and reliable” [32]. Furthermore, ref. [33] confirmed that expertise impacts purchase intention and customer engagement. Thus, the following hypotheses are made:

H3. The expertise of social media influencers impacts purchase intention.

H4. The expertise of social media influencers impacts customer engagement.

2.2.3. Trustworthiness

Trust is viewed as a relational quality that develops over time through repeated contact. Based on their trust in their partners, people can predict and evaluate the value of future exchanges. As a result, trust can help keep relationships intact [34]. In a similar vein, trust in the influencer makes followers believe that they will benefit from their relationship with the influencer. As an outcome, customers sought to engage with influencers and intend to purchase the endorsed product. Thus, we assumed the following hypotheses:

H5. The trustworthiness of social media influencers impacts purchase intention.

H6. The trustworthiness of social media influencers impacts customer engagement.

2.2.4. Credibility

In particular, the influencers’ credibility characteristic is the most powerful for influencing consumer behavior. The credibility of the experts gives consumers useful information, improving the effectiveness of the businesses’ promotion. It is important to highlight that celebrity credibility helps businesses exceed customers’ expectations. The customer responds positively to the campaign because of the celebrity’s credibility. The study demonstrates that customers intend to follow the advice of an influencer with a high credibility rating. The credibility of the influencer does appear to be a key factor affecting customer behavior (See Figure 1). In addition, Mainolfi et al. [35] also concluded that the credibility of social media influencers impacts customer engagement and purchase intention. Thus, we make the following hypotheses:

H7. The credibility of social media influencers impacts purchase intention.

H8. The credibility of social media influencers impacts customer engagement.

2.2.5. Congruence with Product

Congruency describes the similarity or consistency between the celebrity and the product [36] According to Lynch and Schüler [37], the transfer of meaning is facilitated and affected by the influencer and brand or product’s congruence. The likelihood of positive responses to the endorsement in terms of customer engagement and even purchase intentions increases with the level of congruence between the influencer and the brand [38]. Thus, we hypothesized:

H9. The congruence of social media influencers with the product impacts purchase intention.

H10. The congruence of social media influencers with the product impacts customer engagement.

2.2.6. Entertainment Value

Another reason why customers utilize social media sites is for entertainment. According to Chen and Lin [39] “entertainment encompasses those emotional aspects such as fun, enjoyment, and pleasure that have a direct impact on the probability that followers will express a more intense attachment to the influencer and therefore it can be considered an antecedent of engagement.” The perception of consumers toward the social media influencer is dependent on their entertainment value and, thus, impacts purchase intentions [40,41]. Thus, it is likely that the perceived entertainment value of influencers may also shape customer engagement and purchase intentions.

H11. The entertainment value of social media influencers impacts purchase intention.

H12. The entertainment value of social media influencers impacts customer engagement.

2.2.7. Informative Value

Another crucial factor that determines whether a target audience sees an individual as an influencer is how informative an appeal is. Customers require clear, easily comprehensible, quick, relevant, and appropriate information regarding products from social media influencers. Peer consumers now regard SMIs as reliable information sources because they offer details about a product or service’s qualities and features in addition to reviews that include details about the users’ actual experiences [42]. The way the influencer’s information worth is perceived by the followers is crucial. Higher perceived information value has an even greater influence on customers’ choice-making [43]. Additionally, Ki and Kim [44] supported the impact of informative influencers on purchase intention. Thus, we propose:

H13. The informative value of social media influencers impacts purchase intention.

H14. The informative value of social media influencers impacts customer engagement.

2.2.8. Attractiveness

According to Erdogan [45], “attractiveness is nothing more than a stereotype of positive connotations attached to a person, and it extends beyond physically attractive attributes to encompass traits like personality and athleticism.” According to Van der Waldt et al. [46], incredibly attractive influencers can have an impact on their followers’ intentions to make purchases of goods or services. Wang and Scheinbaum [47] confirmed and pointed out that attractiveness is a crucial factor in the dissemination of essential messages. SMIs with attractive physiques are more likely to capture and hold their followers’ attention. Thus, we made the following hypotheses:

H15. The attractiveness of social media influencers impacts purchase intention.

H16. The attractiveness of social media influencers impacts customer engagement.

3. Methodology

The meta-analysis method of reviewing the literature is utilized in the current study. It is applied to summarize the findings of several empirical studies by employing an effect size measure, like the correlation coefficient, and then combining these measures [48]. The use of meta-analysis in this study is justified for the following reasons. It is a technique that, in the first place, statistically summarizes the earlier research on various relationships. By collecting and analyzing the quantitative findings of numerous empirical studies, it provides the ability to evaluate the full picture from a research perspective. It not only determines whether a relationship exists but also assesses whether the effect is positive or negative. In addition, it examines potential moderators that evolved through logical thinking and accepted theory and gives stronger statistical power by pooling the results of multiple quantitative research studies and avoiding the statistical limits of a single study. It helps in overcoming the problem of a small sample size and provides greater accuracy because it evaluates numerous research publications in the same domain. Furthermore, it generates new hypotheses which were not posed by individual studies and inspires future studies.

3.1. Data Collection

The design, conduct, and reporting of this systematic review are as per the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines (PRISMA) [49]. For the collection of data, the following keywords were used for different variables (see

Table 1. Keywords.

Variable	Search String
Social media Influencer	“social media influencer*” or “influencer* marketing” or “digital influencer*” or “facebook influencer*” or “instagram influencer*” or “Instafamous” or “snapchat influencer*” or “twitter influencer*” or “vlogger*” or “blogger*” or “influencer* advertising” or “online influencer*”
Customer Engagement	“customer engagement” or “brand engagement” or “consumer engagement”
Purchase Intention	“purchase intention” OR “buy* intention” OR “intention to purchase” OR “intention to buy”

).

A three-step approach was followed in order to collect the relevant literature. Firstly, a search strategy was formulated to search the literature in highly relevant databases such as Scopus and Web of Science. All types of literature, including book chapters, conference proceedings, along with journal articles were synthesized in this analysis. The Scopus database provided 256 research documents and the Web of Science gave a total of 90 research documents. From a total of 346 studies, studies (n = 5) that were not written in the English language and duplicate documents (n = 68) were removed.

In the next step, all of the databases like Emerald, Elsevier, Taylor & Francis, Springer, IEEE, Sage, and JSTOR were explored to double-check the inclusion of all the related research documents. Further, to include theses and dissertations, databases like Proquest and Shodhganga, etc. were tapped. Finally, Google scholar and Google as search engines were searched to gather unpublished work.

In the final step, a backward and forward search was performed to make sure that no relevant study was missed (see Figure 2).

3.2. Inclusion and Exclusion Criteria

The authors have excluded those studies which do not follow inclusion criteria. Research Studies must meet the following conditions in order to get included in this analysis:

• It must investigate at least one hypothesized relationship. Research studies that do not study the attributes of social media influencers or that do not relate to the intention to purchase and to customer engagement are excluded.
• It must be empirical in nature. Research studies that use qualitative methods to study these relationships are also excluded.
• It must state the information required for meta-analysis like correlation value or any other statistics that can be converted into r like t statistics. All those studies that do not provide enough information are excluded.
• Research must be written in the English language. Few researchers report their research results in other languages like German and French. These are excluded on the basis of language criterion.

3.3. Data Coding

Data were extracted using prepared coding tables in Microsoft Excel that specify the data to be collected from each particular study. In a spreadsheet, the authors kept track of the data related to authors, source, title, publication year, sample size, and correlation coefficients from each study. Two authors independently coded the data and reviewed differences in order to ensure the validity and reliability of the data. In the instance of disagreements, specific research was referred back to, and the authors discussed and resolved them together.

3.4. Data Analysis

Systematic reviews are a novel research method that has emerged in the last few years. They feature transparent standards for gathering secondary data, which increases their reproducibility and lessens bias in data collecting and result consolidation. Systematic reviews, field synopses, and scoping reviews can offer a condensed overview of previous research on a particular topic. A researcher may find it useful to carry out an analysis of the overall number of studies that are published each year, their time duration, and other aspects such as their location and context. If a formal meta-analysis is not included, the scope of the systematic review remains restricted. The outputs of the secondary data collected are appropriately summarized by a meta-analysis.

The first step in the study is to determine the effect size, which is a measurement of the magnitude of the treatment effect and the degree of the correlation between two variables in a meta-analysis. The effect size was calculated for each study, and a summary effect was produced. The correlation between two continuous variables was presented in the data set research, serving as the effect size index. The real effects were presumed to be regularly distributed since a random-effects model for meta-analysis was presumed. The fixed within-study (sampling) error and an additional source were considered by the random model (between-studies variance). A variety of metrics were used to assess the true variance and dispersion patterns. One of them was the Q statistic, which is a calculation of weighted squared deviations and the proportion of actual heterogeneity to all observed variance, indicated by the symbol I square. The same was calculated as follows [50].

$$Q={\displaystyle \sum }_{i=I}^k{\left(\frac{Yi-M}{Si}\right)}^2$$

$$I^2=\left(\frac{Q-df}{Q}\right)\times 100\%$$

For all studies undertaken (k), the effect size values (R+) and the upper and lower limits of the effect size (UL and LL) have been determined.

4. Results and Discussion

Meta-analysis is a research technique that has an advantage over conventional literature reviews in that it measures the strength of a relationship between two variables. Meta-analysis examines pertinent studies to identify significant trends. It formulates a more significant causal effect of independent variables. The results are then combined to determine the overall effectiveness and generalizability of their causal connections.

Table 2. Meta-analysis of the relationships under study.

Hypothesis	SMI Characteristics	Dependent Variable	K	Total Sample	Effect Size	LL	UL	Q	I2	p	Cited Source
H1	Homophily	PI	19	7372	0.45	0.34	0.55	130.02	95.49	0.000	[1,2,18,28,32,51,52,53,54,55,56,57,58,59,60,61,62]
H2	Homophily	CE	9	3031	0.53	0.09	0.80	122.44	93.47	0.000	[16,18,35,61,63,64,65,66,67,68]
H3	Expertise	PI	33	12149	0.50	0.44	0.62	844.30	96.80	0.000	[1,2,17,28,32,33,38,57,62,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87]
H4	Expertise	CE	6	1723	0.46	0.25	0.62	86.65	94.23	0.000	[16,17,33,66,69,88]
H5	Trustworthiness	PI	28	9697	0.55	0.46	0.62	844.30	96.80	0.000	[1,2,17,28,33,38,38,51,53,55,56,57,59,62,71,72,73,75,76,78,79,80,82,83,84,86,87,89,90]
H6	Trustworthiness	CE	7	1927	0.45	0.24	0.62	129.94	95.38	0.000	[16,17,33,66,68,88,91]
H7	Credibility	PI	14	5270	0.57	0.52	0.62	71.21	81.74	0.000	[11,32,33,58,59,76,81,82,92,93,94,95,96,97]
H8	Credibility	CE	5	1493	0.50	0.22	0.71	59.43	93.27	0.000	[33,65,67,97]
H9	Congruence	PI	7	1779	0.56	0.48	0.63	15.86	62.18	0.015	[28,33,60,73,79,98]
H10	Congruence	CE	4	1178	0.46	0.26	0.62	18.80	84.12	0.000	[16,33,99,100]
H11	Entertainment Value	PI	5	1971	0.48	0.24	0.66	75.10	94.67	0.000	[1,56,76,81,85]
H12	Entertainment Value	CE	4	1473	0.62	−0.01	0.80	49.07	93.89	0.000	[65,67,101]
H13	Informative Value	PI	5	2248	0.39	0.23	0.53	28.43	85.93	0.000	[1,44,56,72,90]
H14	Informative Value	CE	2	652	0.54	−0.93	0.97	7.68	86.97	0.006	[65,101]
H15	Attractiveness	PI	26	11527	0.47	0.40	0.55	554.14	95.49	0.000	[1,2,12,17,33,44,51,52,54,56,70,71,73,74,75,77,78,80,82,83,84,85,86,87,92,98]
H16	Attractiveness	CE	5	1355	0.51	−0.45	0.55	130.02	96.92	0.000	[16,17,33,66,88]

provides the summary of the meta-analysis of each attribute of SMI with customer engagement as well as purchase intention.

The first step in the study is to determine the effect size, which is a measurement of the magnitude of the treatment effect and the degree of the correlation between two variables in a meta-analysis. The effect size was calculated for each study, and a summary effect was produced. The correlation between two continuous variables was presented in the data set research, serving as the effect size index. The real effects were presumed to be regularly distributed since a random-effects model for meta-analysis was presumed. The fixed within-study (sampling) error and an additional source were considered by the random model (between-studies variance). A variety of metrics were used to assess the true variance and dispersion patterns. One of these was the Q statistic, which is a calculation of weighted squared deviations and the proportion of actual heterogeneity to all observed variance, indicated by the symbol I square. For all studies undertaken (k), the effect size values (R+) and the upper and lower limits of the effect size (UL and LL) have been determined. Moreover, the value of I² statistics was found to lie in the range of 75–100%, which further suggested a high level of heterogeneity among the studies [45]. Heterogeneity within the studies backed our decision to use the random-effects model.

Table 2 summarizes the random-effects average correlations between SMI characteristics and purchase intention and customer engagement across studies. SMI attributes were found to have moderate to high associations with both purchase intention and customer engagement.

The effect size computed with the correlation value between homophily and purchase intention from 19 studies was 0.45. Thus, the relationship between homophily and purchase intention as indicated by H1 was positive and significant. It supports existing literature that shows a positive association between homophily and purchase intention [54,92] and contradicts the studies that suggest a negative relationship between homophily and purchase intentions [61].

H2 shows a significant and positive relationship between homophily and customer engagement with an effect size of 0.53. It was calculated on the basis of correlation values of 9 studies. It supports existing literature that shows a significant association between homophily and customer engagement [65] but contradicts the studies that depict a weak association between homophily and customer engagement [61].

The effect size between the relationship shown in the third hypothesis was positive and significant (R+ = 0.50). It was calculated from correlation values derived from 33 studies related to expertise and purchase intention. It supports past literature that shows a significant association between expertise and purchase intention [81] but contradicts the studies that depict a weak association between expertise and purchase intention [2].

In the fourth hypothesis, the effect size computed from correlation values between expertise and customer engagement was 0.46. It was positive and significant, thus, H4 was accepted. This hypothesis was based on past literature [88]. It is contradictory to studies that show a weak association between expertise and customer engagement [16].

The fifth hypothesis reported a positive and significant impact of trustworthiness on purchase intention with an effect size of 0.55. It was calculated on the basis of correlation values derived from 28 studies. This supports past literature that shows a positive association between trustworthiness and purchase intention [18] but contradicts [2] which shows a weak association between these variables.

The effect size computed with the correlation value between trustworthiness and customer engagement from 7 studies was 0.45. Thus, the relationship between trustworthiness and customer engagement as indicated by H6 was positive and significant. It supports existing literature that shows a positive and significant association between trustworthiness and customer engagement [17].

H7 shows a significant and positive relationship between credibility and purchase intention with an effect size of 0.57. It was calculated on the basis of correlation values of 14 studies. It supports existing literature that shows a significant association between credibility and purchase intention [58].

The effect size between the relationship shown in the eighth hypothesis was positive and significant (R+ = 0.50). It was calculated from correlation values derived from 5 studies related to credibility and customer engagement. It supports past literature that shows a significant association between credibility and customer engagement [67] but contradicts past literature that depicts a weak association between these variables [97].

In the ninth hypothesis, the effect size computed from correlation values between congruence and purchase intention was 0.56. It was positive and significant, thus, H9 was accepted. This hypothesis was formulated on the basis of past literature that shows a positive and significant association between the congruence of influencers with the product and purchase intention [33].

The tenth hypothesis reported a positive and significant impact of congruence on customer engagement with an effect size of 0.46. It was calculated on the basis of correlation values derived from 4 studies. This supports past literature that shows a positive association between congruence and customer engagement [33].

The effect size computed with the correlation value between entertainment value and purchase intention from 5 studies was 0.48. Thus, the relationship between entertainment value and purchase intention as indicated by H11 was positive and significant. It supports existing literature that shows a positive and significant association between entertainment value and purchase intention [85] but contradicts past literature that depicts a weak association between these variables [56].

H12 shows a significant and positive relationship between entertainment value and customer engagement with an effect size of 0.62. It was calculated on the basis of correlation values of 4 studies. It supports existing literature that shows a significant association between entertainment value and customer engagement [101].

The effect size between the relationship shown in the thirteenth hypothesis was positive (R+ = 0.39). It was calculated from correlation values derived from 5 studies related to informative value and purchase intention. It supports past literature that shows a positive association between informative value and purchase intention [2].

In the fourteenth hypothesis, the effect size computed from correlation values between informative value and customer engagement was 0.54. It was positive and significant, thus, H14 was accepted. This hypothesis was based on past literature [101].

The fifteenth hypothesis reported a positive and significant impact of attractiveness on purchase intention with an effect size of 0.47. It was calculated on the basis of correlation values derived from 26 studies. This supports past literature that shows a positive association between the attractiveness of social media influencers and purchase intention [68] but contradicts the literature that depicts a weak association between attractiveness and purchase intention [71].

In the sixteenth hypothesis, the effect size computed from correlation values between attractiveness and customer engagement was 0.51. It was positive and significant, thus, H16 was accepted. This hypothesis was based on past literature [88]. It contradicts the existing literature that depicts a weak association between attractiveness and customer engagement [16].

4.1. Limitations and Future Research Directions

There are some limitations in the present research. The meta-analysis is exclusively based on empirical research. Therefore, qualitative investigations can be included in the reporting of future researchers. Through the use of moderators, future research could be advanced even further. Potential moderators could be social media platforms used by influencers, products endorsed by influencers, demographics of the respondents, sample size, etc. In addition, only eight characteristics were considered in the analysis. It was observed that some attributes like authenticity, likability or prestige, etc., are not yet empirically investigated much in the existing literature. These characteristics will be improved with further research, and scholars may use the same ones in meta-analytical studies. Future studies can look into how these characteristics affect other variables like brand equity and brand loyalty. Despite these limitations, the current study is novel and offers academia and researchers a strong platform to develop the discipline and offer new implications.

4.2. Implications

Marketing through influencers gains considerable interest from scholars. Synthesis of existing literature provides practical as well as theoretical contributions. There are numerous contributions of this analysis to the existing literature on influencer marketing. Firstly, it provides a statistical compilation of research investigating the relationship between social media influencers and purchase intention and customer engagement. Secondly, by conducting a meta-analysis, it addresses the controversial results in the SMI literature. This study advances the knowledge base by minimizing the heterogeneity in influencer marketing research by empirically specifying the directions of relationships and the magnitude of the effect of these relationships. Additionally, it provides future research directions and opens up a platform to develop the discipline.

This meta-synthesis also put forth practical implications. Professionals might choose their targeting, positioning, and engagement-related actions more effectively with a streamlined and clear direction. While marketing managers lack an understanding of targeting apt influencers and their impact on the purchase intention of customers, they are therefore hesitant to employ influencers as a strategic tool for marketing. This study demonstrates that the influencers’ characteristics significantly influence purchase intention. It is evident by the results that managers, in order to create intentions to purchase, must target those social media influencers who are perceived as credible by their target customers.

Next, we discovered that attributes of social media influencers significantly impact customer engagement. For enhancing the engagement level of customers, entertaining influencers must be selected by marketing professionals. Entertainment is the most common reason behind customers following social media influencers.

As customers are highly active on social media platforms and follow digital influencers, practitioners are advised to tap these influencers to engage customers and create intentions to purchase their products. This meta-analysis provides guidance to professionals in deciding which characteristics to seek in an influencer while approaching them to endorse their product.

5. Conclusions

Influencer marketing is gaining popularity in the era of social media. It is being widely used by marketers to influence their potential customers. This research meta-synthesizes the existing literature on influencer marketing. It highlights eight significant attributes of social media influencers, i.e., homophily, expertise, trustworthiness, attractiveness, credibility, informative value, entertainment value, and congruence with the product. This analysis is based on 176 effect sizes that are derived from 62 individual studies and a total of 22,554 individuals act as the aggregate sample. Results revealed that SMI attributes have moderate to high associations with both purchase intention and customer engagement. As customers are highly active on social media platforms and follow digital influencers, practitioners are advised to tap these influencers to engage customers and create intentions to purchase their products. This study contributes to the literature by minimizing the heterogeneity in influencer marketing research by empirically specifying the directions of relationships under consideration and the extent of the effect of these relationships. Synthesis of existing literature not only provides practical contributions but also offers academia and researchers a strong platform to develop the discipline and offer new implications.