Introduction

Since Katz and Lazarsfeld’s (1955) report that information flows to opinion leaders before spreading to masses, marketing and branding scholars have recognized their influence on consumers’ brand-related perceptions and behaviors. Considerable theory has emerged to explain how and why opinion leaders and providers of word of mouth (WOM) and eWOM shape consumer–brand relationships (e.g., Fay et al. 2019; Gvili and Levy 2016; King et al. 2014; Risselada et al. 2016; Viswanathan et al. 2018). Scholarly interest in influencers has endured even as consumers shift attention from traditional media to e-commerce and social media platforms (e.g., Jin and Muqaddam 2019; Guttmann 2019; Ki and Kim 2019). Current theory, however, has little to say about the shifting nature of influencers on social media, and shifting nature of influence on brand-perceptions of buyers. For instance, recent reports of Instagram bloggers’ (IBs’) influence on brand-related perceptions among their followers suggests that the theoretical lenses of opinion leadership and eWOM provide inadequate explanations of their nature and effectiveness (Larocca 2018). Unlike opinion leaders and eWOM providers, IBs are emerging as independent entrepreneurs who: (a) build strong emotional connections with their followers, and (b) monetize their influence using existing brands as props in their carefully constructed personal narratives on the medium. Their influence on brand perceptions of Instagram users (IUs) remains unexplained by theory.

This article presents evidence from three studies aiming to address this gap. The first study is exploratory. Based on focus groups of IUs, new constructs relevant to IBs’ influence on brand perceptions are identified and a new conceptual framework, hypotheses and measurement scales are derived. The second study, based on a survey of 494 IUs, helps purify the grounded scales, test the structural coherence of the proposed framework, provide an initial test of hypotheses, and yield a theoretical model. The third study, conducted after a 6-month gap, uses survey data to validate the theoretical model derived from the earlier survey (n = 455). As a result, this article presents new evidence to inform future theory and practice about IUs who: (a) anoint IBs they follow the most as leaders of their tribe, (b) regard tribal leaders as curators of brands, and (c) come to view brands associated with their anointed leaders as tribal artifacts. To aid future theory development efforts, the article presents new constructs, new measurement scales, and new evidence to support grounded hypotheses about IBs’ influence on IUs’ brand perceptions. The article ends with a brief discussion of theoretical and practical implications.

Conceptual background

Current reports versus theories of personal influence

Cursory descriptions of IBs suggest they are similar to opinion leaders and eWOM providers. All commonly promote one set of brands over others and are sought by marketers to reach consumers, and raise the prospects of theory-derived hypotheses for testing in IBs’ contexts (Goldsmith and De Witt 2003; Rogers 1983). However, much of the reported influence of IBs lies beyond the scope of current theories of personal influence and calls for fresh investigation and new theoretical development as an initiating step. First, for instance, opinion leaders and the often anonymous providers of eWOM help brands by providing reviews and making recommendations. In contrast, IBs distinctively leverage their talents with the sociotechnical properties of Instagram, use brands as props to produce and post original content, function as entrepreneurial free agents, attract the bulk of attention toward themselves, and derive economic benefits from brands as a result of the strong following they garner (e.g., Crain 2018; Larocca 2018; Mejia 2018; Pope 2020; Stokel-Walker 2019; Swain 2018). Second, IBs such as Tina Craig and Chiaa Ferrigni are directly motivated by making money, and are uniquely credited with producing sales and revenues for brands in ways opinion leaders and eWOM providers are not (Larocca 2018; Stokel-Walker 2019). Third, opinion leaders are viewed as technically competent early adopters (Rogers 1983; Venkatraman 1989), and highly involved in the consumption of products and brands on which they proffer advice (e.g., Goldsmith and Flynn 1994). IBs are associated with brands about which they do not possess technical competence, nor are they early adopters or consumers (Larocca 2018). Fourth, and most importantly, the difference in their native contexts is striking. The theoretical foundations of opinion leadership in marketing and branding contexts are inextricably linked to a time when brand communications occurred predominantly via print, broadcast, and outdoor media; eWOM providers are uniquely linked to e-commerce sites and blogs. IBs are denizens of the compelling hyper-reality of social media-addicted consumers (e.g., Baudrillard 1996; Blackwell et al. 2017). IBs’ influence is relevant in the context of consumer–brand relationships and brand perceptions shaped more decisively on social than on any other media (Blackwell et al. 2017; Lopez et al. 2017). More than opinion leaders and eWOM providers, IBs share their epistemology with celebrities who trigger reportage in popular media. IBs gain media coverage and trigger gossip and fantasy in the popular press as a result of their popularity (e.g., Rindova et al. 2006). They are enabled, unlike other purveyors of opinion, by a culture that is obsessive about and worshipful of celebrities (McCutcheon et al. 2010).

A case for exploratory research

The epistemic distance between conceptions of opinion leadership and eWOM, and reports of IBs’ personal influence preclude one-shot, theory-derived hypotheses tests. Instead, the gap calls for the derivation of an ontological framework, hypotheses and measurement scales grounded in thick descriptions produced by IUs while discussing their interactions with IBs and brands (e.g., Geertz 1973; Laudan 1977; Li and Du 2011). In other words, this research is rooted in grounded theory and not extant literature on personal influence in brand management. Hence, the discussion that follows is separate from one intending to produce literature-derived hypotheses after exhaustive reviews of literature devoted to branding, consumer–brand relationships or personal influence that have occurred elsewhere. For extensive reviews of consumer–brand relationship literature, see MacInnis and Folkes (2017); for opinion leadership literature, see Goldsmith and De Witt (2003) and Gnambs and Batinic (2013), and for eWOM literature, see Goyette et al. (2010) and Gvili and Levy (2016). The following discussion is devoted to explanation of the process by which grounded insights were drawn from focus group data to produce a new framework, hypotheses, and scales.

Study 1: grounded framework, hypotheses, and scales

Focus groups

Participants for four focus groups were selected from a sample of students enrolled in multiple sections of Marketing Research over three semesters taught by a co-author at a Business School located in Northeastern US. Students were asked to volunteer if they were users of Instagram. All 31 participants (17 males, 14 females) referred to their Instagram feeds at least ten times a day; most said they also visited their Instagram accounts first thing in the morning and the last thing at night, and during the breaks between activities throughout the day. In other words, focus group participants were uniformly classified as heavy users of Instagram. The co-author conducting the focus groups informed all participants that they should answer two questions based on their use of Instagram: (a) which brands do you follow and interact with on Instagram, and how Instagram usage has shaped how you think about the brands you follow, and (b) which IBs do you follow on Instagram, and how do your interactions shape the way you think about the IBs and brands. Most responses to these questions were probed by a co-author, each participant was asked to provide specific examples to illustrate the points they were making. The focus groups, lasting between 55 and 70 min, were recorded and transcribed.

The focus group transcripts were shared among co-authors who independently analyzed the verbal protocols based on the guidelines of Miles et al. (2014). The data analysis process was iterative; each co-author began by independently identifying the major themes in the focus groups. In so doing, the co-authors: (1) drew a list of latent constructs that characterized each of the themes, and supported each theme with key phrases and actual quotes from participant voices, (2) drew box-and-arrows figures that illustrated learning about likely relationships between latent constructs, and (3) made notes about the frequency with which the themes were supported by participant voices. After independent analysis, the co-authors met to reconcile findings. Figure 1 encapsulates the consensus view of co-authors reached after each identified theme was near-fully supported by the data; i.e., there were no instances of dissenting voices. The figure: (1) illustrates the result of structural theorizing and the ontology of ‘IBs’ personal influence on consumer–brand relationships, i.e., anointed tribal leader, leader as curator, and brand as tribal artifact as the three key latent constructs, and (2) makes explicit the data-derived notions of convergence and serves as a basis for the three grounded hypotheses derived from focus groups (see Cavusgil et al. 2008).

Fig. 1
figure 1

A framework of Instagram bloggers’ personal influence on consumer–brand relationships derived from exploration

The key findings that reflect the collection of focus group participants are as follows. If all focus group participants spoke in one voice, they would say the following:

Brands? I don’t follow brands on Instagram, I follow people (IBs). I have anointed the person I follow the most as the leader of my tribe. Brands associated with her/him are artifacts; they are imbued with tribal meaning. S/he is a curator of brands uniquely, especially for me, so I can make a statement about my individuality in my connection with my co-tribalists (if I am not posting on Instagram, I don’t exist). Brand is nothing, acknowledgement from my tribe is everything. What is opinion leader and eWOM? Old people’s words?

Participants explain their attributions toward IBs and brands in the following terms. First, there is a reverence toward the IBs they follow the most; they are anointed as leaders of their tribe and enjoy an exalted social status. Consider a representative voice:

FG1, female: Attitude… is a tribal… tribal branditude… not brandal tribitude; (name of IB). She is the leader… she has (a) tribe on Instagram. She comes first, the brands belong to her. Let’s be clear who comes first (name of IB)… she is (a) tribal leader.

Q: why is (name of IB) your tribal leader? Why are you a tribe versus a group? Why

does she come first?

A:… without attitude, it would be a group. Like my mom baking cupcakes for (a) community…

Second, the IB as tribal leader serves as a curator of brands for IUs’ personal consumption. Instagram is described as the internet of narrowing down. From a bewildering array of available brand options, IBs help IUs focus on specific brands they have carefully curated as part of their compelling, emotion-heavy narrative on the medium. Third, attributions of a tribal leader status and the brand-curation function they serve seem inseparable; together, they trigger attributions of a tribal status to brands associated with the IB. In other words, whether a brand is curated or not, its association with the IB is also sufficient to trigger attribution of an artifact status. Hence, Fig. 1 identifies leader as curator as a partial mediator between the anointed leader and tribal artifact linkage. Some of the voices that lead to the drawing of Fig. 1 are as follows. A participant describes the anointing of the leader in the following terms:

FG3, female: The people who follow her (name of tribal leader)… owe her, pay her respects… she earned it…. This is not a group in that sense, we don’t know each other. This is not some community. She earned the position, there is a lot of edge to her. She earned it, you didn’t….. She may have had a tribe before Instagram, but this is her thing now… her medium. She started the Instagram tribe… I was drawn… not so much to her like we don’t know her, but what she was doing…promoting.

Q: drawn? How?

A:… I found her because of the cool things on her posts… after someone said something. Who else thinks like me like beyond my immediate friends?… But I think she has brought us together because we love her, we owe her. Yeah brands are important, but she is more important that way. Brands won’t matter to me if I don’t see them in her posts first. Otherwise it is just a brand.

A participant, likening IBs’ posts to scrolls of curated catalogs, notes:

FG3, male: You might think, okay someone might think just because I am on Instagram with him I want to be him. I don’t want to be him. Maybe if it came easy, but I don’t have that kind of (credibility). But the guy puts out… a list of things… catalog to choose (from)… if I get into the things, it is meaningful, right?… I have taste too, just like a rich and famous guy.

Once curated, participants note the high likelihood that they would check out the brand online or—if within economic reach—check them out in brick-and-mortar stores. Consider the words of a participant as she describes a low-cost cosmetic brand that was curated by her IB tribal leader:

FG4, female: I googled (name of cosmetic accessory) out. I had to drive to Walgreens. It was that easy to find. Less than ten bucks. I bought it. Put it on my feed. Then waited to see who noticed.

Q: Well? Did they notice?

A: If they didn’t, I’d delete the post (laughs, others join in the laughter)… Of course I treasure it. I am afraid to use it up. I hope it doesn’t dry out. But I want to keep it for long as I can, at least until I get a full time job. Can’t do it on my customer service rep part time job. So I am saving it.

The question whether the IB is regarded as an opinion leader or a word-of-mouth provider is wholly dismissed by focus group participants. A participant notes:

FG2 female: This (Instagram) is not old people media. Nicki Minaj and Serena Williams?… are tribal goddesses, like on Game of Thrones (name of TV show)… Serena’s just (an) opinion leader like Beyonce’s just a singer… she’s more than that. What is opinion leader and word of mouth? Old people’s words?… what you old people use? Young people look for empowerment… (to define) what is nowtoday. They are not talking or using their mouths for me… they are pointing… that like… that brand, and I’ll take it up after that on my own.

Q: please explain that, what does like that brand mean?

A: … It’s attention, a flash going… whom it shines on. Me, or the brand? Definitely me… on me. Definitely on Nicki. Brand is like just something she has.

Q: so how is Nicki’s brand useful to you?

A: It’s about reputation. If I am not on Instagram, my friends will forget… it never happened… people will forget what I was doing… they won’t know who I am, I will not exist… I want them to notice… show my uniqueness. I am unique. Different. Nicki makes it easy, she says, that brand. Now I’ve got something to say to others… Nicki is making up my museum for me which makes me interesting… (so that) people appreciate me on Instagram. Then others notice…. I am always checking… did others notice? If others don’t say they notice, am I doing justice to myself?

The significant difference between focus group findings and current notions in the literature are worthy of note. While the literature offers rich insights into tribes (Maffesoli 1996), and into brands that command a tribe-like following (e.g., Taute and Sierra 2014); the literature is silent on IBs as tribal leaders or curators, or endowers of a tribal artifact status on brands. Similarly, while current writings call firms and managers to curate the content of the online communication to manage relationships with key stakeholders and customers (see Rosenthal et al. 2017; Kilgour et al. 2015), the literature is silent on the curatorial function of opinion leaders or eWOM providers. The new testable hypotheses derived from focus group voices that add value to the state of the art and serve as the basis for the hypotheses-testing studies that follow are:

H1

The greater the extent to which a person is anointed the leader of the user’s Instagram tribe, the greater the likelihood that the tribal leader serves the function of a curator of brands.

H2

The greater the extent to which a person is anointed the leader of the user’s Instagram tribe, the greater the likelihood that the brand associated with the tribal leader is regarded as a curated artifact of the user’s tribe.

H3

The greater the extent to which the tribal leader is regarded as a curator of brands, the greater the likelihood that the brand curated by the tribal leader is regarded as an artifact of the user’s tribe.

Derivation of grounded measurement scales

The newness of the exploratory-data-derived latent constructs precluded the use of theory-derived measurement scales for the studies that followed. Hence, grounded measurement scales were derived from focus group transcripts based on the guidelines of Churchill (1979) and Hinkin (1995). Briefly, the focus group excerpts that led to derivation of each of the themes were compiled, and key words used in the description were identified. Based on the key words, and the contexts in which they were spoken, the co-authors framed the Likert scales for each of the indicator or measured variables. This process was iterative, and the wordings of the scales were refined for clarity. Table 1 shows the result of this iterative process; it includes the constitutive definition of the three latent constructs, and the indicator or measured variables—expressed as Likert scale items—derived from participant voices.

Table 1 Definitions and scales derived from focus group data

Study 2

For the second study, a Qualtrics questionnaire that included scales shown in Table 1 was circulated to a nationwide sample of Instagram users registered with Amazon Mechanical Turk (only people who indicated that they used Instagram at least once a day were permitted to complete the questionnaire). The study and questionnaire were approved by the Institutional Review Board (IRB) of one co-author’s college; all participants signed an informed consent form before completing the questionnaire. The purpose of the study was academic; there is no conflict of interest. M-Turk samples were used because they have stood up to repeated tests of reliability (e.g., Litman et al. 2017). The survey data included 494 completed surveys (see Table 2 for a brief description of sample).

Table 2 Description of samples

As Table 1 shows, the questionnaire included an eight-item Likert scale for measuring anointed tribal leader, a ten-item Likert scale for measuring tribal leader as curator, and a seven-item Likert scale for brand as tribal artifact. Table 1 also shows the result of the first step of scale-purification based on Churchill (1979) and Hurley et al. (1997). The table shows the factor loadings for each indicator variable from the rotated component matrix (obtained from Varimax, orthogonal rotations; extraction method: principal component analysis). We also tested the indicator variables with oblique rotations (Promax) with both principal component analysis and maximum likelihood as extraction methods; the results are comparable with the ones obtained by Varimax rotation shown in Table 1. Hence, a five-item scale for anointed tribal leader, a four-item scale for tribal leader as curator, and a four-item scale for brand as tribal artifact served as a basis for fitting a structural equation model and serve as a basis for testing three hypotheses simultaneously (all indicator variables with factor loadings of .725 or higher).

Based on Anderson and Gerbing’s (1988) guidelines, a two-step process of structural equation model construction and hypotheses testing was employed (using EQS.2 software). The first step examined whether any structural model existed, and whether such a model had acceptable goodness of fit. Robust estimation procedure was employed to preclude problems caused by non-normality in the data. Each CFA iterations relied on Lagrange multiplier test to: (a) indicate the cross loading of measured variables on latent factors, and (b) help remove the measured variables one by one over the three iterations which indicate the cross loading of measured variables on latent factors; i.e., we removed two measured variables over two iterations based on the strength of the cross-loading variables corresponding to the CFA iterations (see Table 3 for results of CFA iterations). The iterations stopped when the measurement model yielded an RMSEA of .046.

Table 3 Sample one (n = 494)

After completing the CFA, the hypothesized paths were specified to run the SEM procedure on EQS. As Fig. 2 shows, the three hypotheses were supported by the survey data; i.e., all paths are significant, and the model has excellent fit. The measurement and structural parameters for the revised theoretical model and the standardized solutions for the hypothesized model are shown in Tables 4 and 5.

Fig. 2
figure 2

Nationwide sample 1 (n = 494, for purification of scales, estimation of reliability and validity, derivation of a theoretical model)

Table 4 Measurement and structural parameters from the revised theoretical model
Table 5 Standardized solution for the hypothesized model

Reliability and validity

Table 6 reports key statistics attesting to reliability and construct validity of scales. The Cronbach’s alphas for the three latent constructs ranged from .887 to .912; the construct reliability ranged from .89 to .95—both indicating acceptable reliability of scales based on Churchill (1979) and Hair et al. (1998). The significant path parameters point to the convergent validity of scales; i.e., the latent factors are related in ways they were hypothesized. The average variance extracted (AVEs) calculated for each of the three latent constructs using factor loadings produced by the standardized solution are greater than .5 (ranging from .68 to .8), exceed the squared correlations among all latent constructs attest to discriminant validity of scales (e.g., Hair et al. 1998). In other words, the shared covariance between any two latent variables is exceeded by the variance captured by the latent constructs based on the measured variables (e.g., Fornell and Larcker 1981).

Table 6 Key statistics and correlations among factors (S1 = sample one, n = 494; S2 = sample 2, n = 455)

Study 3

To further test the robustness of the theoretical model yielded by Study 2, a second survey was conducted after a period of 6 months (see Table 2 for brief description of sample taken for Survey 2). The same survey instrument was re-circulated via Amazon’s Mechanical Turk service to a nationwide sample of Instagram users. To prevent overlap, the M-Turk service was used to ensure that the questionnaire was not sent to IP addresses that had participated in the first survey. As Fig. 3 shows, the second survey validates the theoretical model produced by the first survey; all hypothesized paths are supported (please see Tables 5 and 6 for measurement and structural parameters for the revised theoretical model, and for the standardized solutions for the hypothesized model based on the second survey).

Fig. 3
figure 3

Nationwide sample 2 (n = 455 for validation of the theoretical model)

Implications for future theories of personal influence

The purpose of this research is to address the gap between current personal influence theories and reported influence of IBs on brand-perceptions. In this regard, findings suggest that future theories of personal influence are more likely to hold real world analogs if they can explain how and why a new generation of IBs serve buyers’ needs beyond brand-related evaluations, critique, review and recommendations that eWOM producers and opinion leaders provide. The present study highlights the unique needs of IUs served by IBs, i.e.: (1) socioemotional needs for venerating tribal leaders and connecting with other tribe members in ways that brands and traditional opinion leaders and eWOM providers do not, and (2) needs for curated catalogs of brands they deem as artifacts and collectors’ items imbued with sacred and spiritual meaning. Definitions of personal influence-related constructs of opinion leaders and eWOM providers deserve reexamination and re-construction based on emerging realities of prevalent social media, strong social media engagement of buyers, and the emergence of IBs as social media celebrities.

This research also makes a theoretical contribution by yielding new, grounded measurement scales, tested for their reliability and validity by two samples separated by 6 months. They make a contribution because they assess attitudes of those influenced, when current theory of personal influence is almost entirely informed by measures based on influencers’ self-reports. For instance, people are attributed higher level of influence when they deem themselves as knowledgeable, central to interpersonal networks, influential in terms of recommending products and brands (e.g., Goldsmith and Flynn 1994). Flynn et al. (1996) scale relies on self-assessment, with Likert items worded as: Other people rarely ask me about rock cd’s before they choose one for themselves, and People that I know pick the rock music [clothing, “green” products] based on what I have told them. Similarly, Goyette et al. (2010) scale for eWOM intensity is worded as: ‘I spoke of this company much more frequently than about…’, and scale for positive valence WOM as: “I am proud to say to others that I am this company’s customer.” This article presents new scales that channel the voices of those influenced; they hold the potential to shape future theories of personal influence in ways that speak to the practical realities of social media users and not just influencers.

Implications for future theories of branding on social media

Tribes have brands

The notions that tribes or consumer tribes exist (O’Reilly 2012), or that brands have tribes (Ruane and Wallace 2015), are not new to the literature. Maffesoli (1996) defined tribes as ephemeral gatherings relevant to the post-modern world of fragmented individuals over 2 decades ago. Some brands are known to produce tribal following; please see Badrinarayan et al. (2014) and Taute and Sierra (2014) for detailed deconstruction of ‘brands have tribes’ construct. This research reports evidence about the Instagram context in which the opposite is true; i.e., tribes have brands. Tribal leader, tribal affiliation, and curation are central in the consciousness of IUs; brands are secondary considerations. The findings also contrast with current notions that use the terms ‘tribal leader’ and ‘opinion leader’ interchangeably (Cova and Cova 2002); the reported studies find them epistemologically distinct.

Brands as secondary triggers of emotions

The studies suggest that IBs leverage their talents in ways that serve as primary triggers of emotionality; Instagram users are emphatic that they follow other people on the medium and that brands are afterthoughts. This is a significant finding unique to this study; it contrasts sharply with currently popular notions of brands and brand messages as the primary triggers of emotions (Thomson et al. 2005). A brand’s power, the literature notes, is reflected in its image and equity, and by its cognitive, emotional and sensory associations (Cho et al. 2015). A brand’s image is more positive when buyers say that it is good value for money, provides a good reason for purchase, is interesting and possesses a personality, different from other brands, and suggests who is consuming the brand (Martinez and de Chernatony 2004). If a brand is preferred over another with the same features, same services, and same prices, it is regarded as indicator of brand equity (e.g., Yoo and Donthu 2001). Current theories speak of a brand’s power to shape users’ perceptions, particularly when they represent knowledge (Keller and Lehmann 2003), possess a personality (Aaker 1997), or seem humanlike (Aggarwal and McGill 2007). It is likely that the present study produces findings contradictory to these notions because they emerge from exploration of branding in social media contexts whereas extant thinking about branding is deeply rooted in traditional print, broadcast, and outdoor media. In the IU’s context, the strong tether between a brand’s power on Instagram and IBs’ power as brand-curating, artifact-producing tribal leaders is evident from the reported studies. In this context: (1) the brand’s image is indistinguishable from IBs’ image and attractiveness to IUs, and (2) the brand’s equity is enmeshed with the IBs’ equity; i.e., in the number of their followers, the number of posts and reposts they garner, and the monetary value of their personal brand.

Brands as curated artifacts

The notion of brands as curated artifacts deserves focused attention from scholars aiming to explain user–brand relationships that emerge as a result of social media usage, and represents a significant contribution of this research. Currently, most discussions about curation have occurred outside of branding contexts, i.e., in discussions of digital assets produced by information sharing on social media (Yakel 2007; Tous et al. 2018), and digital curation by institutions and libraries for preserving knowledge (e.g., Dallas 2016). Marketing literature is largely concerned about managers either as curators of brand messages (Kilgour et al. 2015) or as co-creators of brand meanings (e.g., Rosenthal et al. 2017). The findings point to IBs and not managers as the chief curators of brands and endowers of the ‘tribal artifact’ status to brands. IBs serve as tribal leaders who narrow down choices and curate brands; e.g., a focus group participant explains: “(IBs) They are not talking or using their mouths for me… they are pointing… “that” like… “that brand,” and I’ll take it up after that on my own.”

Findings resonate, however, with current discussions of artifacts as anthropological constructs. Curated tribal brands, as do artifacts in common parlance: (1) reflect the sense-making and learning that results from IU–IB interaction (e.g., Kleinsmann et al. 2013; Singh et al. 2009), (2) serve as receptacle of an IB’s and their tribe’s shared understanding and knowledge (e.g., Kreiner 2002), and (3) hold symbolic meaning for the Instagram tribe’s identity (e.g., Schultz et al. 2006; Vilnai-Yavetz and Rafaeli 2006). Findings also resonate with recent cultural trends that favor transformation of crowded living spaces into living museums, with fewer and carefully curated artifacts that hold meaning (e.g., Kondo 2015). The trend is significant; thrift stores reportedly refuse to take in objects people want rid from their living spaces (NPR 2019). Our findings suggest that IUs seek curation and a careful, tasteful narrowing down of brand choices made by their tribal leaders. The notion of brands as curated artifacts deserves accommodation by theories of branding in light of intensifying user–social media interactions.

Managerial implications

This research finds that IBs’ influence on brand perceptions is more expansive than currently reported. For instance, while popular reports of IBs refer almost entirely to luxury fashion brands (Carbone 2019), focus group participants mentioned no luxury brands; instead they referred to IBs’ influence on their relationships with every-day use brands associated with personal grooming and lifestyles and available in drugstores, big-box stores, and shopping malls. In other words, Instagram represents a viable social media platform for a wide variety of consumer brands interested in reaching IUs—and relevant to a broader spectrum of managers than current writings in popular literature would suggest.

Similarly, the context of IBs’ influence on user–brand relationships seems expansive and relevant to managers. American buyers are spending more time on social media than watching TV, and spending on social media advertising is projected to exceed $37 billion in 2020 (Guttmann 2019). Roughly 370 K IBs command over 100,000 followers each on the medium and jostle with brands for attention (Mention 2018). If personal influence on brand choices was once the realm of paid spokespersons who might have earned their celebrity status elsewhere, or unpaid but identified opinion leaders and unpaid, anonymous providers of WOM and eWOM, some influence is shifting in favor of IBs willing to produce compelling, entertaining content on photo and video sharing social media in ways that gain them—and not the brand—a tribal following. The hyper reality of tribal leaders, curation, and tribal artifacts triggered by user–IB interaction seems compelling enough to compete with messages from brand sponsors alone.

The findings echo the concerns raised by other scholars, i.e., a brand’s strategy for building a relationship with their users are resisted by newly empowered and connected consumers on social media (Leitch and Merlot 2018). Moreover, the studies support the notion that IBs as tribal leaders are usurping managerial power to define what brands mean. In the context of a hundred million users of Instagram, brand power is inseparable from the power of IBs as tribal leaders. IBs, not brand managers, are: (1) producing original content that IUs consume, (2) addressing IUs unmet socioemotional needs for tribal affiliations and tribal leaders who can produce curated catalogs for consumption, and (3) providing the fodder and social cache to IUs by producing re-postable content.

Even as IBs encroach on the influence of brand sponsors, some brands are coopting IBs and paying them for creating content. In this regard, the studies raise caution about brand managers ceding or losing messaging and curatorial power to independent IBs. As Lieber (2014) reports, Tina Craig first translated her ability to use GoogleAds to channel revenues from other brands to her, and eventually used her curatorial powers as a blogger to channel web users to brands and directly produce sales for handbag manufacturers (estimated $20,000). Brand sponsors are ceding not just the sense-making process, they are letting independent IBs dip into their revenue streams at best, or plain delegating tangible parts of their revenue generation function to effective IBs. Brayanboy, for instance, who is contracted to post 5–7 Instagram brand-related posts for Gucci, emerged from obscure origins in the Philippines and built a following of 650,000 with clever use of the medium (Larocca 2018). These developments raise caution; even though hiring IBs to promote brands is an easy option, it is unlikely to serve as a substitute for connection with the albeit rapidly changing reality of user–brand interactions occurring on social media in general, and IUs in particular.

Limitations and conclusion

The findings and implications are consistent with the intents of the studies; i.e., the generation and testing of grounded evidence to stimulate new thinking and research. No single study can capture the entire complexity of IBs’ influence on user–brand relationships; new exploration and hypotheses testing are clearly necessary before widely generalizable findings are produced. In particular, further testing of our hypotheses and scales across multiple samples and longitudinal designs can shed more light on personal influence of IBs. The directions of relationships tested are based in grounded theory; exploration of alternative linkages is similarly left to future research. The SEM procedure was used to test hypotheses simultaneously, inferences of causality are not implied (e.g., Fornell and Larcker 1981). The assessment of dependent measure (brand as tribal artifact) is subject to inflation because of common methods bias, and reliance on self-reports (Podsakoff et al. 2003). This issue is addressed by following Conway and Lance’s (2010) guidelines; i.e., with special emphasis on and reports of the construct reliability and discriminant validity of scales. Future research based on independently derived ways of assessing the artifact-status of brands is likely to shed more light.