1 Introduction

Social media influencers are increasingly popular and affecting consumers’ attitudes, perceptions, preferences, choices, and decisions. Social media influencers are regular everyday people who have created an online presence from the grassroots level through their social media channel or page and, in the process, have created an extensive network of followers (Bastrygina and Lim [10]. In that sense, social media influencers are different than traditional celebrities or public figures, who rely on their existing careers (e.g., actors, singers, politicians) to become popular and exert influence [88].

Influencers first appeared in the early 2000s, and have since progressed from a home-based hobby to a lucrative full-time career. Influencer marketing has become so attractive that with the growing industry, there is an ever-growing set of social media users that aim to become an influencer. Influencers are now capitalizing on their popularity and visibility to further their career in mainstream media such as the film and television industry [1]. The segmentation of influencers is on the number of followers they have, whereby influencers can be classified as micro-, meso- and macro-influencers [44]. According to Lou and Yan [88], posts by influencers have two essential purposes from a marketing perspective: the first purpose is to increase the purchase intention of their followers, and the second purpose is to enhance their followers’ attractiveness and product knowledge. Influencers often curate posts with information and testimonials about the features of the product that they are promoting, which results in increased information value and product knowledge. In the process, they leverage and relay their attractiveness and aesthetic value through the use of sex appeal and posing [104].

Social media influencers have been defined by many scholars in numerous ways. Freberg et al. [44] characterized social media influencers as a new type of independent third-party endorser who shapes audience attitudes through blogs, tweets, and the use of other social media. Abidin [1] construed social media influencers as a form of microcelebrities who document their everyday lives from the trivial and mundane to the exciting snippets of the exclusive opportunities in their line of work, thereby shaping public opinion through the conscientious calibration of persona on social media. De Veirman et al. [28] defined social media influencers as people who built a large network of followers and are regarded as trusted tastemakers in one or several niches. Ge and Gretzel [45] denoted social media influencers as individuals who are in a consumer’s social graph and has a direct impact on the behavior of that consumer. More recently, Dhanesh and Duthler [30] described social media influencers as people who, through personal branding, build and maintain relationships with their followers on social media, and have the ability to inform, entertain, and influence their followers’ thoughts, attitudes, and behaviors. When these definitions are taken collectively and espoused through a marketing lens, social media influencers are essentially people who develop and maintain a personal brand and a following on social media through posts that intertwin their personality and lifestyle with the products (e.g., goods, services, ideas, places, people) that they promote, which can influence the way their followers behave (e.g., attitudes, perceptions, preferences, choices, decisions), positively (e.g., purchase) or negatively (e.g., do not purchase).

Social media influencers, as digital opinion leaders, participate in self-presentation on social media. They form an identity by creating an online image using a rich multimodal narrative of their everyday personal lives and using it to attract a large number of followers [59]. Most critical to their success is the influencer-follower relationship [1], which future follower behavior (e.g., interaction, purchase intention) is dependent upon [13], [37], [126]. Indeed, social media influencers are often perceived to be credible, personal, and easily relatable given their organic rise to fame [28], [31], [104].

In collaborations between brands and social media influencers, the role of a social media influencer is to act as a brand ambassador by designing sponsored content for the brand to convey and enhance its brand image and brand name [104], and to drive brand engagement and brand loyalty [72]. Such content is often curated by social media influencers, as independent third-party endorsers, by sharing their experiences and lives in relation to the brand through pictures, texts, stories, hashtags, and check-ins, among others [28]. Indeed, social media influencers are highly sought after by brands because they have established credibility with their followers as a result of their expertise, which allow them to exert influence on the decision-making of their followers [60]. Moreover, influencer marketing through social media can provide opportunities to influencers and their followers to contribute to the co-creation of the brand’s image on social media [88].

With the growing importance of influencer marketing and the popularity of social media influencers, various brands have started promoting their products with the help of social media influencers in an attempt to influence consumers to behave in desired ways (e.g., forming positive brand attitudes and encourage product brand purchases) [104]. However, consumer behavior is highly complex [81], and increasing inconsistency has been noted in the effectiveness of this medium [124]. Thus, it is essential to understand the factors (i.e., antecedents) underpinning consumer decision making (i.e., consequences or decisions) toward brands promoted by social media influencers, including the factors (i.e., mediators and moderators) responsible for the inconsistency in consumer responses. In this regard, attempts to consolidate extant knowledge in the field is arguably relevant to address the extant gap and needs of marketing scholars and professionals interested in social media influencer marketing.

In recognition of the growing influence of social media influencers and influencer marketing in consumer decision making, this study aims to provide a one-stop, state-of-the-art overview of the articles, journals, methods, theories, themes, and constructs (antecedents, moderators, mediators, and consequences) relating to social media influencer marketing using a systematic review of articles in the area from 2008 to 2021. Though a recent review on social media influencers was conducted by Vrontis et al. [124], the present review remains warranted because the existing review only considered a small sample of 68 articles published in journals indexed in the Chartered Association of Business Schools Academic Journal Guide, and thus, cannot holistically encapsulate the state of the field. Indeed, the insights and the integrative framework resulting from their review was relatively lean, which can be attributed to the sample limitations that the authors had imposed for their review. The same can be said about another recent review by Bastrygina and Lim [10], which considered only 45 articles in Scopus that narrowly focused only on the consumer engagement aspect of social media influencers. To overcome these limitations, the present review will consider a more inclusive search and inclusion criteria while upholding to the highest standards of academic quality by relying on a broader range of indexing sources. The motivation of the present review is also in line with the call by Lim et al. [86] and Paul et al. [98] for new reviews that address the shortcoming of existing reviews in order to redirect research in the area onto a clearer and more refined path for progress. In addition, the present review adopts a bibliometric-content analysis to consolidate current findings, uncover emerging trends and extant gaps, and curate a future agenda for social media influencer marketing. Noteworthily, the rigorous multi-method review technique (i.e., the combination of a bibliometric analysis and a content analysis) adopted for the present review is in line with the recommendation of Lim et al. [86] to facilitate a deeper dive into the literature, and thus, enabling the curation of a richer depiction of the nomological network characterizing the field [94], in this case, the field of social media influencer marketing. In doing so, this study contributes to answering the following research questions (RQs):

  • RQ1. What is the publication trend of social media influencer marketing research, and which are the key articles?

  • RQ2. Where is research on social media influencer marketing published?

  • RQ3. How has social media influencer marketing research been conducted?

  • RQ4. What are the theories that can be used to inform social media influencer marketing research?

  • RQ5. What are the major themes of social media influencer marketing research?

  • RQ6. What are the constructs (i.e., antecedents, mediators, moderators, and consequences) employed in social media influencer marketing research?

  • RQ7. Where should social media influencer marketing be heading towards in the future?

The rest of the paper is structured as follows. The next section provides an account of the methodology used in the research, followed by the findings and conclusions of the study in subsequent sections.

2 Methodology

This study conducts a multi-method systematic literature review on social media influencer marketing using a bibliometric-content analysis in line with the recommendation of Lim et al. [86] and recent systematic literature reviews (e.g., Kumar et al. [64]. The assembling, arranging, and assessing techniques stipulated in the Scientific Procedures and Rationales for Systematic Literature Reviews (SPAR-4-SLR) protocol by Paul et al. [98] to carry out a systematic literature review are also adopted and explained in the next sections.

2.1 Assembling

Assembling relates to the identification (i.e., review domain, research questions, source type, and source quality) and acquisition (i.e., search mechanism and material acquisition, search period, search keywords) of articles for review. In terms of identification, the review domain relates to social media influencer marketing, but within the subject areas of business management, social sciences, hospitality, tourism, and economics due to their immediate relevance to the review domain, and thus, articles in other subject areas such as computer science, engineering, medical, and mathematics, which are peripheral to the review domain, were not considered. Next, the research questions underpinning the review pertain to the articles, journals, methods, theories, themes, and constructs in the field and were presented in the introduction section. Only journals were considered as part of source type as they are the main sources of academic literature that have been rigorously peer reviewed Nord & Nord, [96]. The source quality was inclusive yet high quality, whereby articles published in journals indexed in the Australian Business Deans Council (ABDC), Chartered Association of Business Schools (CABS), and Web of Science (WOS) were included. In terms of acquisition, the search mechanism and material acquisition relied on the WOS database, which is connected to myriad publishers such as Emerald, Sage, Springer, Taylor and Francis, and Wiley. The search period starts from 2008 and ends in 2021. The year 2008 was selected as the starting year because it was the year that the concept of influencer was first introduced by Kiss and Bichler [63], and thus, a review staring from 2008 can provide a more accurate and relevant account of the extant literature on influencer marketing, particularly from the lenses of consumers and social media influencers. The end year 2021 was selected because it is the most recent full year at the time of search—a practice in line with Lim et al. [83]. The search keywords—i.e., “consumer behavio*” (truncation technique), “social media,” “influencer,” and “marketing”—were curated through brainstorming and endorsed by disciplinary experts in marketing and methodological experts in review studies. In total, 320 articles were returned from the search, but 17 articles were removed as they were related to engineering, mathematics, and medicine, which resulted in only 303 articles that were retrieved for the arranging stage.

2.2 Arranging

Arranging relates to the organization (i.e., organizing codes) and purification (i.e., exclusion and inclusion criteria) of articles returned from the search. In terms of organization, the content of articles was coded based on the key focus of each research question: journal title, method, theory, and construct (antecedent, mediator, moderator, consequence). The bibliometric details of the articles were also retrieved and organized accordingly in this stage. In terms of purification, 89 articles were eliminated as they were not published in journals indexed by ABDC and CABS, with the rest of the 214 articles included for review.

2.3 Assessing

Assessing relates to the evaluation (i.e., analysis method, agenda proposal method) and reporting (i.e., reporting conventions, limitations, and sources of support) of articles under review. In terms of evaluation, a bibliometric analysis and a content analysis were conducted.

For the bibliometric analysis, the Bibliometrix package in R studio software [4] was used to conduct (1) a performance analysis to reveal the publication trend as well as the key articles and journals (RQ1 and RQ2), and (2) a science mapping to uncover the major themes in the field (RQ5) in line with the bibliometric guidelines by Donthu et al. [32]. With regards to science mapping, a triangulation technique was adopted in line with the recommendation of Lim et al. [86] using:

  1. 1.

    co-citation using PageRank, wherein the major themes are revealed through the clustering of articles that are most cited by highly-cited articles,

  2. 2.

    bibliographic coupling, wherein the major themes are revealed through the clustering of articles that cite similar references, and

  3. 3.

    keyword co-occurrence, wherein the major themes are revealed through the clustering of author specified keywords that commonly appear together [32], [64].

For the content analysis, the within-study and between-study literature analysis method by Ngai [95] was adopted (RQ3, RQ4, and RQ6). The within-study literature analysis evaluates the entire content of the article (e.g., theoretical foundation, methodology, constructs), whereas the between-study literature analysis consolidates, compares, and contrasts information between two or more articles. The future research agenda proposal method is predicated on the expert evaluation of a trend analysis by the authors (RQ7). In terms of reporting, the conventions for the outcomes reported include figures, tables, and words, whereas the limitations and sources of support are acknowledged at the end.

3 Findings

The findings of the review are organized based on the research questions (RQs) of the study: articles, journals, methods, theories, themes, and constructs.

3.1 Articles

The first research question (RQ1) deals with the publication trend and key articles of social media influencer marketing research.

Figure 1 indicates that research on social media influencer marketing began to flourish 10 years (i.e., 2018 onwards) after the concept of was introduced in 2008 [63]. This implies that interest in social media influencer marketing is fairly recent (i.e., within the last five years at the time of analysis), wherein its stratospheric growth appears to have coincided with that of highly interactive and visual content-focused social media such as Instagram (e.g., Instagram Stories feature launched in December 2017) [17] and TikTok (e.g., international launch in September 2017) [129]. The growth of triple-digit publications observed in 2021 during the COVID-19 pandemic is especially noteworthy as it signals the importance of social media influencer marketing in the new normal and reaffirms past observations of an acceleration in technology adoption [77], [79].

Fig. 1
figure 1

Publication trend of social media influencer marketing research

Full size image

Table 1 presents the top articles on social media influencer marketing. The most cited article is De Veirman et al.’s [28] (464 citations), which focused on social media influencer marketing using Instagram and revealed the impact of the number of followers and product divergence on brand attitudes among the followers of social media influencers. The burgeoning interest on Instagram as seen through this most cited article despite its recency corroborates the earlier observation on the stratospheric growth in research interest on highly interactive and visual content-focused social media. The top-cited articles in recent years demonstrate increasing research interest in comparative studies (e.g., celebrity versus social media influencer endorsements, [104],Instagram versus YouTube; [108], as well as review studies (e.g., Hudders et al., [48], [124], albeit the latter being limited (e.g., small review corpus, niche review focus) and thus reaffirming the necessity and value of the present review.

Table 1 Top articles on social media influencer marketing
Full size table

3.2 Journals

The second research question (RQ2) deals with the outlets that publish social media influencer marketing research and the source type chosen according to the recommendation of Paul et al. [98] is journals on the basis of academic quality and rigor. In total, the 214 articles in the review corpus were published in 87 journal titles indexed in ABDC, CABS, and WOS. Out of the 87 journal titles, 80 (37.38%) articles are published by the top 10 journals with the most articles on social media influencer marketing, with Journal of Business Research, International Journal of Advertising, and Journal of Retailing and Consumer Services emerging as the top three journals in terms of numbers of articles published in the area (Table 2).

Table 2 Top 10 journals with the most articles on social media influencer marketing
Full size table

3.3 Methods

The third research question (RQ3) focuses on the methods that can inform social media influencer marketing research and were identified and coded manually using the within-study technique and consolidated to portray the outcome of a between-study literature analysis suggested by Ngai [95]. In total, seven categories of methods were employed in 214 articles on social media influencer marketing research (Table 3). As a category, quantitative methods in the form of surveys were most prevalent (n = 64), followed by qualitative methods (n = 52), with individual interviews being the most popular method (n = 19). Experimental (n = 38) and machine learning (n = 33) methods were noteworthy too. Non-empirical methods (n = 19) such as conceptual (n = 9) and review (n = 10) methods were less prominent. Similarly, mix methods (n = 8) were the least popular. As a whole, the review indicates that extant research on social media influencer marketing were mostly empirical in nature albeit in silos (i.e., single rather than mixed methods).

Table 3 Methods employed in social media influencer marketing research
Full size table

3.4 Theories

The fourth research question (RQ4) pertains to the theories that can inform social media influencer marketing research and were identified, coded, and reported using the same Ngai [95] informed within- and between-study literature analysis as reported for the methods in the preceding section. In total, 46 different theories employed in 94 (43.93%) articles on social media influencer marketing research were revealed (Table 4). Persuasion knowledge theory emerged as the most popular theory with eight articles, whereas social learning theory, social comparison theory, social cognitive theory, social exchange theory, social identity theory, social influence theory, source credibility theory, reactance theory, theory of para-social interaction, theory of planned behavior, and uses and gratifications theory were among the other popular theories (n ≥ 3). The broad range of theories indicate that social media influencer marketing is an area of research with multi-faceted aspects worthy of exploration and investigation. The sociological theories manifested in the most ways—namely Bourdieu’s theory, Graph theory, network theory, observational learning theory, optimal distinctiveness theory, social cognitive theory, social comparison theory, social exchange theory, social identity theory, social influence theory, social learning theory, structural hole theory, system justification theory, and theory of para-social interaction—whereas media theories were not far behind—namely advertising literacy theory, media dependency theory, megaphone effect theory, source credibility theory, transfer theory, two-step flow theory, uses and gratifications theory, and visual framing theory. The manifestation of theories that infused “media” and “sociology” together, such as social media influencer value model and social-mediated crisis communication theory, were observed as well. Psychological theories, such as associative learning theory, attachment theory, attribution theory, consistency theory, construal level theory, dissonance theory, dual process theory, elaboration likelihood model, halo effect theory, reactance theory, similarity-attraction model, theory of planned behavior, and theory of reasoned action, and marketing theories, such as Doppelganger effect theory, human brand theory, relationship management theory, and source effect theory, were also noteworthy. Management theories, such as charismatic and transformational leadership theory and resource dependency theory, were also observed. Interestingly, only one economic (i.e., cost-signaling theory) and one technology (i.e., technology acceptance model) theory were observed, which may indicate that the economic and technology aspects are underexplored as compared to the media, psychological, management, marketing, and social aspects of social media influencer marketing.

Table 4 Theories employed in social media influencer marketing research
Full size table

3.5 Themes

The fifth research question (RQ5) involves the mapping of extant research on social media influencer marketing. To do so, three science mapping techniques that rely on different sources of bibliographic data were relied upon—namely (1) a co-citation analysis using PageRank to identify clusters of articles that are most cited by highly-cited articles, (2) a bibliographic coupling to locate clusters of articles that share common references, and (3) a keyword co-occurrence analysis to uncover clusters of author specified keywords that commonly co-appear [32], [65].

3.5.1 Foundational themes (or foundational knowledge)

The foundational themes and the top articles for each foundational theme in social media influencer marketing research are depicted in Table 5. In essence, foundational themes exemplify the perspectives that a field’s research relies upon, and thus, these themes may encompass articles inside and outside that field [32]. In the case of social media influencer marketing, four foundational themes were revealed by the co-citation analysis using PageRank. Noteworthily, the PageRank scores indicate article prestige, wherein a higher score indicates that the article is cited more by highly-cited articles in the field, whereas the betweenness and closeness centrality scores reflect the article’s relevance across and within themes, wherein a higher score indicates greater relevance across and within themes, respectively [32].

Table 5 Foundational themes in social media influencer marketing based on co-citation analysis using PageRank
Full size table

The first foundational theme depicts the foundations and models for social media influencer marketing. The articles in this foundational theme signify the key characteristics of concepts associated to social media influencer marketing, such as the concept of engagement [49], “Instafamous” [55], influencer marketing [88], and social media influencers [44], including the difference between traditional celebrities and contemporary social media influencers [104].

The second foundational theme denotes the influence and impact perspectives for social media influencer marketing. The articles in this foundational theme represent a collection of insights in relation to influence and impact. For example, the most prestigious article under this theme examines the impact of the number of followers of Instagram influencers and the divergence of the products promoted by these influencers on the brand attitudes of their followers [28]. Other examples of influence and impact outcomes include attitudes and behavioral intentions [37], engagement [120], perceptions Lee & Watkins, [67], and purchase decisions [31].

The third foundational theme highlights the importance of endorsement and resonance perspectives for social media influencer marketing. The articles in this theme, which are widely cited by highly cited articles on social media influencer marketing, emphasize the importance of endorsement and resonance literature in grounding the reasons for and outcome of social media influencer marketing. This can be seen by the prominence of celebrity endorsement (e.g., [34], Mccracken, [93], Silvera & Austad, [107]) and congruence (e.g., Till & Busler, [116]; [[128]] literature that make up the most prestigious articles under this theme.

The fourth foundational theme relates to the profiling and measurement perspectives for social media influencer marketing research. This theme signifies and reaffirms the value of personal characteristics (e.g., personalities, profiles; [31], Ferchaud et al., [40]), measurement scales (e.g., expertise trustworthiness and attractiveness; Ohanian, [97]), and evaluation methods (e.g., structural models; Fornell & Larcker, [43]) in guiding and informing social media influencer marketing research, and thus, they form a considerable part of the knowledge relied upon by research in the field.

3.5.2 Major themes (or major research streams)

The major themes build upon the foundational themes to curate new knowledge and understanding on social media influencer marketing [32]. To uncover the major themes, a keyword co-occurrence analysis was initially conducted to gain a sense of the nomological network for the major themes [94], followed by a bibliographic coupling to gain an in-depth understanding of the content under each major theme in the field [32].

The keyword co-occurrence analysis indicates that four major themes characterize the knowledge curated by extant research focusing specifically on social media influencer marketing (Fig. 2 and Table 6), which is triangulated by the six major themes revealed through bibliographic coupling, in which four bibliographic coupling clusters corresponds to two keyword clusters (Table 7). The key peculiarities of these themes are presented as follows.

Fig. 2
figure 2

Nomological network of research streams in social media influencer marketing research

Full size image
Table 6 Major themes (clusters) and topics (keywords) in social media influencer marketing research based on keyword co-occurrence analysis
Full size table
Table 7 Major themes (clusters) in social media influencer marketing research based on bibliographic coupling
Full size table

Parasocial interactions and relationships in social media influencer marketing. This major theme is most prominent (eight keywords) and relatively recent (2020.1429–2020.7499). This theme highlights the importance of the “credibility” (n = 6), “persuasion knowledge” (n = 7), and “source credibility” (n = 7) of social media influencers as essential “persuasion” (n = 5) factors that influence the “parasocial interactions” (n = 8) and “parasocial relationships” (n = 12) in social media influencer marketing. Most research in this area is conducted in the context of “Instagram” (n = 27), wherein “purchase intention” (n = 13) is a common outcome expected and examined. Noteworthily, extant research concentrating on influencing parasocial interactions have highlighted the importance of self-influencer congruence (Shan et al., [105]; [128] and the value of message value [88] and credibility [108], including the moderating role of audience comments [102], in fostering consumer trust and purchase intention toward branded content [88], [102], Shan et al., [105], [108], [128], whereas those focusing on developing and managing parasocial relationships emphasized the importance of being entrepreneurial (Fink et al., [41]) and personal branding (Ki et al., [61]) in the pursuit of becoming famous and garnering brand equity and loyalty among followers [18], [55], [57].

Sponsorship in social media influencer marketing. This major theme is fairly prominent (six keywords) and recent (2019.8–2021). This theme highlights the importance of “sponsorship disclosure” (n = 6) in “celebrity endorsement” (n = 5) and among “social media influencers” (n = 60) engaged for “native advertising” (n = 7) in “influencer marketing” (n = 63), with “YouTube” (n = 9) featuring prominently in this space. Noteworthily, extant research on this theme is divided into two notable streams, wherein the first stream sheds light on the commercialization and value of social media influencer marketing (Britt et al., [16]; Harrigan et al., [47]; Hudders et al., [48]; [124],), which highlights the importance of the second stream pertaining to the impact of disclosure (i.e., macro, micro—e.g., declaring sponsorship to establish and reaffirm the credibility of social media influencers and the brands they represent) on the behavioral responses of social media followers [13], [30], [58], [104], [110].

Authenticity of marketing and public relations in social media influencer marketing. This major theme is fairly prominent (five keywords) but with a longer history (2017.4286–2021) than the other major themes. This theme highlights the continuing importance of “authenticity” (n = 7) in the “marketing” (n = 5) and “public relations” (n = 7) endeavors of “influencers” (n = 29) on “social media” (n = 56). Thus, it is no surprise that extant research in this theme have focused on traditional marketing concepts such as advertorial campaigns [1], personal branding [59], rhetoric [45], strategic communication [33], and self-presentation [6].

Engagement and influence in social media influencer marketing. This major theme is fairly prominent (five keywords) and recent (2019.4–2020.6). This theme encapsulates “social media marketing” (n = 16) research that concentrates on the “social influence” (n = 5) of “opinion leadership” (n = 5) and the equivalent outcome of “brand engagement” (n = 5), with “Twitter” (n = 7) featuring prominently in this space. Noteworthily, the prominent studies under this theme concentrate on the power of social networks of social media influencers, including examining the influence of the number of followers [28], measuring the influence of customer networks [63] and social media influencers [5], and the value of opinion leaders [87] and sponsored campaigns [49] across these networks.

Taken collectively, these themes, which were triangulated across two bibliographic sources of data (i.e., keywords and references) and analytical techniques (i.e., keyword co-occurrence analysis and bibliographic coupling), suggests that social media influencer marketing has tremendous commercial value, which justify the sponsorship that brands are willing to provide to social media influencers in return for marketing and public relation campaigns for their brands and products. Nevertheless, it is important to note that the power of social media influencers resides in their authenticity, which is a crucial reason as to why social media influencers are followed and relied upon by their followers. The management of parasocial interactions and relationships are also highly important as they are essential to foster desired engagement among followers and influence their behaviors in ways desired by social media influencers and the brands that they represent. The next section provides a deeper dive into the mechanisms (constructs) that transpire in social media influencer marketing.

3.6 Constructs

The sixth research question (RQ6) involves the unpacking of constructs that relevantly explain consumer behavior toward social media influencer marketing, which were revealed through the same within- and between-study literature analysis as reported in the methods and theories sections previously [95]. The constructs (Fig. 3) were arranged according to testable categories in the form of antecedents (Table 8), mediators (Table 9), moderators (Table 10), and consequences (Table 11), with each category having sub-categories that encapsulate relevant constructs that fall under the theme of that sub-category. The thematic naming of sub-categories are mostly self-explanatory (i.e., audience-, brand-, content-, influencer-, social-, and technology-related), with only one sub-category being uncommon yet sensible due to the unique nature of the context under study—that is, the comparative-related sub-category, which captures the essence of constructs where comparison exist between two or more sub-categories (e.g., influencer-follower relationship is a construct that accounts for the comparison transcending the audience- and influencer-related sub-categories, whereas product-endorser fit is a construct that reflects the comparison between the brand- and influencer-related sub-categories).

Fig. 3
figure 3

Consumer behavior toward social media influencer marketing

Full size image
Table 8 Antecedents in social media influencer marketing
Full size table
Table 9 Mediators in social media influencer marketing
Full size table
Table 10 Moderators in social media influencer marketing
Full size table
Table 11 Consequences in social media influencer marketing
Full size table

In terms of antecedents, four sub-categories emerged, namely comparative-, content-, influencer-, and social-related antecedents (Table 8). The comparative-related antecedents (six counts) comprise of influencer-follower relationship (two counts) and perceived similarity (four counts). The content-related antecedents (36 counts) consist of authenticity (four counts), disclosure (14 counts), informativeness (nine counts), message construal (one count), perceived quality (two counts), perceived quantity (two counts), perceived originality (one count), and post credibility (three counts). The influencer-related antecedents (34 counts) consist of engagement and interaction (two counts), influencer attractiveness (10 counts), influencer credibility (six counts), influencer expertise (nine counts), influencer likeability (one count), perceived trustworthiness (five counts), and perceived uniqueness (one count). The social-related antecedent (four count) contains parasocial relationship (four count) only. In total, 18 antecedents emerged across four sub-categories. Content-related antecedents appear to be the most researched (36 counts), followed by influencer-related antecedents (34 counts), with few studies examining comparative- (six counts) and social- (four count) related antecedents. Disclosure (14 counts) is the antecedent that has been studied the most, followed by influencer attractiveness with 10 counts. As a whole, there is good breadth and depth for antecedents as a category, but is mixed for its sub-categories.

In terms of mediators, seven sub-categories were revealed, namely audience-, brand-, comparative-, content-, influencer-, social-, and technology-related mediators (Table 9). The audience-related mediators (13 counts) comprise of attachment (one count), attitude (five counts), interest (one count), psychological ownership (one count), and trust (five counts). The brand-related mediators (eight counts) consist of brand recognition (five counts), product attractiveness (one count), and sponsorship transparency (two counts). The comparative-related mediator (four counts) contains self-influencer connection (four counts) only. The content-related mediators (seven counts) encapsulate disclosure (two counts), message appeal (one count), message credibility (one count), message process involvement (one count), and source credibility (two counts). The influencer-related mediators (15 counts) encompass engagement and interaction (two counts), expertise (two counts), influencer credibility (five counts), opinion knowledge leadership (five counts), and perceived popularity (one count). The social-related mediators (three counts) include electronic word of mouth (one count) and parasocial interaction (two counts). The technology-related mediators (two counts) incorporate perceived ease of use (one count) and perceived usefulness (one count). In total, 22 mediators were revealed across seven sub-categories. Influencer- and audience-related mediators appear to be the most researched with 15 and 13 counts respectively, followed by brand- (eight counts) and content- (seven counts) related mediators. Attitude, brand recognition, influencer credibility, opinion leadership knowledge, and trust are the mediators studied the most with five counts each. Overall, there is reasonable breadth and depth for mediators as a category, but is mixed for its sub-categories.

In terms of moderators, six sub-categories were unpacked, namely audience-, brand-, comparative-, content-, influencer-, and social-related moderators (Table 10). The audience-related moderators (10 counts) comprise of advertisement literacy (one count), audience engagement (two counts), domains of interest (one count), envy identification (one count), interaction propensity (one count), purchase intention (one count), self-discrepancy (two counts), and social identification with social commerce (one count). The brand-related moderator (one count) consists of brand attitude (one count) only. The comparative-related moderators (three counts) contain perceived closeness (one count), perceived fit (one count), and product-endorser fit (one count). The content-related moderators (nine counts) encapsulate audience comments (one count), disclosure (one count), download volume (one count), message process involvement (one count), message valence (one count), number of hashtags (one count), online ratings (one count), structural assurance (one count), and visionary insights (one count). The influencer-related moderators (four counts) encompass influencer socio-economic status (one count), number of followers (one count), perceived self-serving motive (one count), and type of influencer (one count). The social-related moderators (two counts) include parasocial relationship (one count) and parental mediation (one count). In total, 27 moderators were unpacked across six sub-categories. Audience-related moderators (10 counts) appear to be the most researched, followed by content-related moderators (nine counts). All moderators had only one count except audience engagement and self-discrepancy, which have two counts, and thus indicating its breadth but not depth.

In terms of consequences, three sub-categories were unveiled, namely brand-, influencer-, and social-related consequences (Table 11). The brand-related consequences (73 counts) comprise of brand attitude (17 counts), brand awareness (one count), brand involvement (two counts), brand purchase or patronage (46 counts), brand recall (two counts), and brand trust (five counts). The influencer-related consequences (19 counts) consist of engagement and interaction (11 counts), following influencer (five counts), and influence (three counts). The social-related consequences (12 counts) contain recommendation and referral propensity (nine counts) and social sharing (three counts). In total, 11 consequences were unveiled across three sub-categories. Brand-related consequences (73 counts) appear to be the most researched, followed by influencer- (19 counts) and social- (12 counts) related consequences. Brand purchase or patronage (46 counts) represent the most studied consequence, followed by brand attitude (17 counts) and engagement and interaction (11 counts). Taken collectively, the consequences unveiled indicate its depth but not breadth.

4 Trend analysis and future research directions

Agendas for future research are a hallmark of systematic literature reviews [84]. While there are many approaches to develop future research agendas, the present study adopts an approach that the authors found to be most objective and pragmatic—that is, a trend analysis from thematic and topical perspectives. The suggestions for future research based on the analysis from these perspectives are presented in the next sections.

4.1 Thematic perspective

The thematic perspective comprises a trend analysis of bibliographic clusters representing the major themes of social media influencer marketing research. The choice of focusing on bibliographic clusters as opposed to keyword clusters was a deliberate decision taken in light of the finer-grained research streams in the former (six clusters) over the latter (four clusters), as well as the availability of the alternative perspective (i.e., the topical perspective) that will use keywords to shed light on the topical trend in the field.

The productivity of the six major themes (research streams) in social media influencer marketing research has generally improved in recent years, particularly in 2021, with the exception of research on parasocial relationships in social media influencer marketing (Cluster 6), which experience a slight decline (i.e., seven in 2020 to six in 2021). Though closely-related research on parasocial interactions has proliferated (Cluster 5), the difference between the two research streams and their relatively lower number of studies as compared to other research streams suggest that new research in both streams is very much required. Similarly, the research stream on disclosures (Cluster 4) is highly important, yet it remains relatively low as compared to its more popular counterpart, that is, the research stream on commercialization and value of social media influencer marketing (Cluster 3), both of which are important research streams to the larger umbrella research stream on sponsorship revealed by the keyword co-occurrence analysis. While the research streams on authenticity (Cluster 2) and engagement and influence (Cluster 1) in social media influencer marketing are highly popular, further research remains necessary in light of the evolving changes in the social media landscape. Notwithstanding the productivity of the research streams, several promising avenues avail for advancing knowledge across all research streams.

In terms of engagement and influence in social media influencer marketing (Cluster 1), the emergence of augmented, virtual, and mixed realities, including the metaverse, signals the need for new research that unpacks the opportunities for engagement in these new social avenues along with the effectiveness of these avenues as compared to existing avenues for social media influencer marketing. In addition, the nature of engagement will benefit from finer-grained examination to account for the differences between its varied cognitive, affective, and behavioral manifestations [80], [85], which remains underexplored in social media influencer marketing.

In terms of authenticity in social media influencer marketing (Cluster 2), the key markers of authenticity and the strategies to communicate and strengthen a sense of authenticity are potential avenues to enrich understanding of this area. Noteworthily, future research on authenticity will need to go beyond traditional measures (e.g., scales; Ohanian, [97]) and engage in purposeful exploration to uncover the attributes and actions that if available and taken will enhance followers’ perceptions of the authenticity of social media influencers. In this regard, future qualitative and experimental research in this research stream is encouraged, wherein the former will lead to the discovery of new authenticity markers that the latter can test for cause and effect. Such research should lead to meaningful extensions on the understanding of authenticity that goes beyond treating the concept as a singular construct in the field.

In terms of commercialization and value of social media influencer marketing (Cluster 3), the potential of non-economic returns of social media influencer marketing could be explored in future research. With the advent of corporate social responsibility and environmental social governance (Lim et al., [83], it is imperative that the expectations and evaluations of returns goes beyond those that are economic in nature (e.g., sales) [78]. The advocacy and support of socio-environmental causes (e.g., hashtags of actions and statements) could be explored, which can be subsequently useful to develop sustainability ratings beneficial for illustrating the impact of both social media influencers and the brands that they represent.

In terms of disclosure in social media influencer marketing (Cluster 4), future research could explore the different ways in which explicit and implicit disclosures could be curated and signaled by social media influencers to their followers. Such research should be potentially useful as not all social media platforms provide options of explicit labels (e.g., sponsor ad) to social media users, especially when such social media posts are not paid to extend its reach and thus relies on social media users themselves to self-disclose. Moreover, the effectiveness of these forms of disclosure, including their combination, have not been adequately studied and thus should be worthwhile exploring. The negative connotation that may be attached to such disclosures should also be addressed in ways that make such disclosures an asset rather than a liability.

In terms of parasocial interactions in social media influencer marketing (Cluster 5), the multitude ways in which parasocial interactions could be curated represent a potentially fruitful avenue for future exploration. At present, the general focus has been on the influence of social media influencer credibility and the congruence of such interactions to follower expectations and perceptions [108]. In this regard, future research is encouraged to explore the different ways in which parasocial interactions could be curated, and in the midst of doing so, theorizing the entry points and sustaining factors that make such interactions parasocial between social media influencers and their followers. Given the complex nature of parasocial interactions, future research in this space could benefit from employing neuroscientific tools (e.g., eye tracker, wearable biosensors, [73], [74] to gain nuanced insights into biological responses that can be used to supplement self-reported responses in order to better ascertain the parasocial nature of interactions among social media influencers and their followers.

In terms of parasocial relationships in social media influencer marketing (Cluster 6), deeper insights on what makes parasocial relationships gratifying and lasting should be developed in future research. Such research should provide a better understanding on the constitution of parasocial relationships and how social media influencers can foster and maintain them over time. Nevertheless, errors or mistakes are bound to happen (e.g., slip of inappropriate word, unintentional non-disclosure of sponsorship). Thus, the repair and recovery of negatively-affected parasocial relationships among social media influencers and their followers could also be given scholarly attention in future research.

Taken collectively, these suggestions for future research should enrich research across all research streams in social media influencer marketing. The next section builds on the insights from this section and takes a closer look on topical trends in the field (Fig. 4).

Fig. 4
figure 4

Productivity trend of major themes in social media influencer marketing research. Note: Cluster 1 = Engagement and influence in social media influencer marketing. Cluster 2 = Authenticity in social media influencer marketing. Cluster 3 = Commercialization and value of social media influencer marketing. Cluster 4 = Disclosure in social media influencer marketing. Cluster 5 = Parasocial interactions in social media influencer marketing. Cluster 6 = Parasocial relationships in social media influencer marketing

Full size image

4.2 Topical perspective

The productivity of topical research in social media influencer marketing has evolved over the years (Fig. 5). Noteworthily, the extant literature on social media influencer marketing has been largely predicated on “communication management”, “centrality”, and “viral marketing” up to 2018. Newer research has nonetheless made a stronger and more explicit connection to “influencer marketing” and “social media”, with “Instagram” emerging as the most prominent social media in the field. The transmission of “eWOM” or “electronic word-of-mouth” and how this translates into “parasocial interaction” or “immersion” between “social media influencers” and “followers” has taken center stage alongside “online marketing” and “social media marketing” considerations such as “advertising”, “brands”, “brand awareness”, and “purchase intention” from a “neoliberalism” perspective.

Fig. 5
figure 5

Productivity trend of major topics in social media influencer marketing research

Full size image

Notwithstanding the trending topics in social media influencer marketing revealed by the trend analysis, it is clear that new research focusing on new phenomena is very much required. For example, new social media platforms such as Clubhouse and TikTok have been extremely popular platforms for social media influencers in recent years, and thus, future research should also consider exploring social platforms other than Instagram. Furthermore, the proliferation of augmented and virtual realities remains underexplored for social media influencer marketing. The rebranding of Facebook to Meta is a signal of the future rise of the metaverse. New research in this direction focusing on new-age technologies for social media influencer marketing should provide new knowledge-advancing and practice-relevant insights into contemporary trends and realities that remain underrepresented in the literature. Similarly, the diversity and evolution of social media followers also deserve further attention in light of accelerated technology adoption by societies worldwide in response to the COVID-19 pandemic and the new normal [77], as well as the changing nature of generational cohorts in the society [79].

5 Conclusion

The importance of consumerism for business survival and growth albeit in a more authentic, meaningful, and sustainable way [76] along with the increasing use of digital media such as social media [82] have led to the proliferation of social media influencer marketing and its burgeoning interest among academics and professionals [10], [124]. This was evident in the present study, wherein the consumer behavior perspective of social media influencer marketing took center stage. Using the SPAR-4-SLR protocol as a guide, a bibliometric-content analysis as a multi-method review technique, and a collection of 214 articles published in 87 journals indexed in ABDC, CABS, and WOS as relevant documents for review, this study provides, to date, the most comprehensive one-stop state-of-the-art overview of social media influencer marketing. Through this review, this study provides several key takeaways for theory and practice and additional noteworthy suggestions for future research.

5.1 Theoretical contributions and implications

From a theoretical perspective, this study provides two major takeaways for academics.

First, the review indicates that most articles on social media influencer marketing published in journals indexed in ABDC, CABS, and WOS were not guided by an established theory, as only 94 (43.93%) out of the 214 articles reviewed were informed by theories (e.g., persuasion knowledge theory, social learning theory, source credibility theory, theory of planned behavior). This implies that most articles relied on prior literature only to explain their study’s theoretical foundation, which may be attributed to a lack of awareness on the possible theories that may be relevant to their study. In fact, a similar review on the topic albeit with a relatively smaller sample of articles (i.e., 68 articles only) due to protocol limitations (i.e., CABS-indexed journals only) had acknowledged the issue but unfortunately failed to deliver a collection of theories informed by prior research [124]. In this vein, this study hopes to address this issue as it has revealed 46 different theories that were employed in prior social media influencer marketing research, which can be used to ground future research in the area. Furthermore, the list of theories can be used to justify the novelty of future research where a new theory is applied. In addition, future studies can take inspiration from the manifestation of theories emerging from multiple theoretical perspectives, such as the social influencer value model and the social-mediated crisis communication theory informed by the media and sociological theoretical perspectives, to develop new theories in the field, which may be challenging but certainly possible [81]. Alternatively, future studies can consider theoretical integration by using two or more theories in a single investigation, which can reveal richer insights on the phenomenon (e.g., which theoretical perspective is more prominent or which factors from which theoretical perspective yield strong impacts and therefore warrant investment prioritization).

Second, the review shows that social media influencer marketing research does not have to be limited to a simple direct antecedent-consequence relationship or the multiply of such relationships. Instead, research in the area can benefit from testing the mediating and moderating effects of various factors to enrich the insights derived from their study. Interestingly, the review reveals that antecedents can also play the role of mediators (e.g., engagement and interaction) and moderators (e.g., parasocial relationship) and vice versa, which implies that the conditions in research design setup are fundamental to the conclusions made about the consequences of social media influencer marketing [75], which can take the form of consumer responses to the brand (e.g., brand purchase or patronage), the influencer (e.g., following influencer), and the community (e.g., recommendation, social sharing). In total, seven categories in the form of audience-, brand-, comparative-, content-, influencer-, social-, and technology-related factors that could manifest as antecedents, mediators, and moderators were revealed. Noteworthily, the comparative-related factors such as perceived closeness, perceived fit, perceived similarity, self-influencer connection, and product-endorser fit transcended across multiple categories (e.g., audience and influencer, brand and influencer), which indicate the promise of social media influencer marketing as a research context suitable for the development of new factors to describe consumer behavior of a comparative nature. Indeed, comparative-related factors is, to the best knowledge of the authors, a new categorization that has not been revealed by prior systematic literature reviews, and thus, represent a key contribution to the literature that should be noted in future research and reviews. Moreover, the mapping of constructs in Fig. 3 and their counts in Tables 8, 9, 10, and 11 provide useful starting points to identify the extant gaps in prior research (e.g., brand-related factors remain underexplored as moderators, comparative-related factors remain underexplored as mediators) and to inform the direction of future research accordingly. Finally, the constructs and their associated categories revealed can also be compared and contrasted in future investigations to delineate the difference in impact between constructs of different categories, and when paired with appropriate theories, can provide stronger grounds for managerial recommendations to brands and influencers interested to leverage off the benefits of social media influencer marketing to attract and persuade desired consumer behavior.

5.2 Managerial contributions and implications

From a managerial perspective, this study provides two major takeaways for brands and influencers.

First, the review indicates that brands indirectly influence consumers through influencers—that is to say, the strategy of brands engaging in influencer marketing on social media places influencers at the forefront, with brands taking a backseat in that strategy. This was evident from the literature review, where brand-related antecedents were absent; instead, the influence of brands manifests in the form of mediators (e.g., brand recognition, product attractiveness, sponsorship transparency) and moderators (e.g., brand attitude). In that sense, it is important that brands identify and engage with influencers strategically, particularly those who are perceived to be attractive, credible, engaging and interactive, experts, a good fit for their products, likeable, opinion leaders, popular, trustworthy, unique, and without overly self-serving motives in order to encourage desired consumer behavior toward their brands (e.g., brand purchase and patronage, brand trust), as revealed by the review herein.

Second, the review reveals that social media influencers directly influence consumer behavior toward the brands they promote (e.g., brand attitude, brand awareness, brand involvement, brand recall, brand trust), the influencers themselves (e.g., follower, influence), and the social media community at large (e.g., recommendation, social sharing). In particular, the content that influencers curate on social media can affect how consumers respond to these stakeholders. The review indicates that such content should be authentic, credible, informative, original, and transparent (disclosure). The message appeal and message process involvement are also important mediators to strengthen the influencer’s ability to encourage desired consumer behavior among their followers (e.g., positive audience, brand, influencer, and social behavior), whereas audience comments, assurance, hashtags, insights, and volume of posts can moderate or nullify the potential desired impact that influencers could elicit from their followers on social media. Indeed, the importance of electronic word of mouth, parasocial interaction, and perceptions of closeness and fit have also been highlighted through the review. Importantly, when promoting to kids and youth, it is essential that influencers consider what parents would think about their posts, as parental mediation was observed to occur in the review.

5.3 Review limitations and future review directions

From a review perspective, this study acknowledges three major limitations that can inform the curation of future reviews.

First, the systematic literature review herein does not capture article performance (i.e., citations) because it was mainly interested in unpacking the articles, journals, theories, methods, and content (themes, constructs) underpinning existing research on social media influencer marketing, and it kept in mind the space limitation of the journal. Notwithstanding the comprehensive and rigorous insights revealed using the SPAR-4-SLR protocol, future reviews may wish to pursue an impact analysis, which can lead to rich insights pertaining to article performance (e.g., difference in citations [e.g., total citations, average citations per year, h-index, g-index] between papers with and without theory, using empirical and non-empirical methods, or across different methods and thematic categories).

Second, the systematic literature review herein encapsulates only a qualitative evaluation of the constructs in existing social media influencer marketing research. To build on the insights herein, future reviews may wish to pursue a meta-analytical review, where a meta-analysis involving the antecedents, mediators, moderators, and consequences revealed in Tables 8, 9, 10, and 11 in this review (in the short run) or unveiled in future reviews (in the long run) is performed. Such an endeavor should also provide finer-grained insights on conflicting findings and provide a resolution to such findings in the same study.

Third, the systematic literature review herein focuses only on the consumer behavior perspective of social media influencer marketing, which is mainly due to the maturity of research from this perspective [98], as seen through the number of articles available for review (i.e., 214 articles) under a rigorous protocol (i.e., the SPAR-4-SLR protocol). Moving forward, future reviews may wish to pursue a systematic review of social media influencer marketing from the business and industrial perspective, wherein the impact of influencer marketing on social media for business and industrial brands in general and across different industries are reviewed and reported.