Influencer marketing on Instagram—The optimal disclosure strategy from influencers’ and marketers’ perspectives

Saternus, Zofia; Mihale-Wilson, Cristina; Hinz, Oliver Article — Published Version Influencer marketing on Instagram—The optimal disclosure strategy from influencers’ and marketers’ perspectives Electronic Markets Provided in Cooperation with: Springer Nature Suggested Citation: Saternus, Zofia; Mihale-Wilson, Cristina; Hinz, Oliver (2024) : Influencer marketing on Instagram—The optimal disclosure strategy from influencers’ and marketers’ perspectives, Electronic Markets, ISSN 1422-8890, Springer, Berlin, Heidelberg, Vol. 34, Iss. 1, https://doi.org/10.1007/s12525-024-00743-x This Version is available at: https://hdl.handle.net/10419/315702 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. http://creativecommons.org/licenses/by/4.0/ Vol.:(0123456789) Electronic Markets (2024) 34:60 https://doi.org/10.1007/s12525-024-00743-x RESEARCH PAPER Influencer marketing onInstagram—The optimal disclosure strategy frominfluencers’ andmarketers’ perspectives ZofiaSaternus1 · CristinaMihale‑Wilson2 · OliverHinz2 Received: 11 December 2023 / Accepted: 13 November 2024 / Published online: 10 December 2024 © The Author(s) 2024 Abstract This article explores the impact of different advertising disclosure strategies (i.e., explicit sponsorship disclosure, concealing disclosure, impartiality disclosure, and no disclosure) in influencer marketing on influencer-related outcomes (user engagement, user sentiment, and influencer credibility) and marketer-related outcomes (user attitude towards the brand and users’ intention to purchase). We conducted two field experiments and an online survey with an experimental design in collaboration with an active micro-influencer on Instagram. The results of the studies indicate that from a marketers’ perspective, it is best when influencers promote products as genuine recommendations and use impartiality disclosure. From an influencer’s perspective, the optimal disclosure strategy depends on whether the influencer seeks to improve engagement with their content or their levels of credibility. When influencers’ primary focus is to increase engagement, if they provide information on sponsorship or non-sponsorship, they do not have to worry about decreasing engagement rates due to the employed disclosure strategy. Suppose influencers’ goal is to increase their credibility. In that case, it depends on their content (whether it is rich in genuine recommendations or sponsored content) and the group they want to target—i.e., if they seek to target followers versus non-followers. Keywords Influencer marketing· Disclosure· Instagram· Follower behaviors· Field experiment JEL Classification M37 Introduction Influencer marketing via social media, particularly on Instagram, has established itself as a way to market products and services. Marketers and brands nowadays often pay influencers to endorse their products and brands. With more than 1 billion monthly active users, Instagram is one of the most strategically important platforms for influencer marketing (Mansoor 2021). Influencer marketing is experiencing rapid growth, outpacing traditional paid advertisements on platforms like Instagram, TikTok, and YouTube. A significant trend to note is that U.S. influencer spending1 on Instagram is forecasted to exceed $2 billion for the first time in 2024, reflecting a major shift in how marketers allocate their budgets (Lebow 2023). This trend underscores the increasing importance of influencer collaborations in producing engaging, authentic, and community-driven content. Influencer marketing’s growth is fueled by its ability to connect brands with consumers Responsible Editor: Steven Bellman * Cristina Mihale-Wilson [email protected]t.de Zofia Saternus zofia.sater[email protected] Oliver Hinz [email protected]t.de 1 SAP AG Germany, Dietmar-Hopp-Allee 16, Walldorf69190, Germany 2 Information Systems andInformation Management, Goethe University, Theodor-W.-Adorno-Platz 4, FrankfurtamMain60323, Germany 1 Influencer spending as a key performance indicator (KPI) is increasingly recognized for its role in evaluating the effectiveness of influencer marketing campaigns. This metric provides insights into how much brands invest in influencer collaborations relative to the returns generated. Electronic Markets (2024) 34:6060 Page 2 of 27 through creators who understand and engage deeply with their audiences. This allows for more personalized and relatable content, making it more effective than traditional ads (Brooks etal. 2021). On social media platforms, private and commercial communications frequently go hand in hand. Sponsored Instagram posts imitate and blend with non-sponsored content, making it difficult for users to distinguish advertising from organic, genuine (impartial) recommendations (Campbell & Grimm 2019; De Cicco etal. 2021; De Veirman & Hudders 2020). Hence, various regulatory agencies suggest that sponsored content should be disclosed. In response to various calls for transparency, Instagram implemented a builtin disclosure feature that tags content with “Paid partnership with [Brand].” However, with multiple stakeholders criticizing the position and style of the Instagram-generated tag and research unable to conclusively prove its effectiveness in increasing advertising awareness under real-world conditions (De Cicco etal. 2021), it seems to be influencers’ responsibility to make sponsored content identifiable (Karagür etal. 2022). Although disclosures are essential for ethical and responsible advertising, marketers and influencers might be reluctant to use them. One reason is that marketers and influencers need more certainty regarding the best disclosure strategy for tagging sponsored content (FTC 2016, 2017a). As of 2017, based on a Federal Trade Commission (FTC) mandate, any online sponsored content by a thirdparty source must disclose sponsorship information (FTC 2017b; Stubb etal. 2019). Similarly, in Belgium, the Flemish Regulator for the Media (VRM2) published in 2021 Content Creator Protocol, including rules that influencers must follow when sharing commercial communications, as well as in Germany, according to § 5a UWG 3 (Law against unfair competition), § 6 TMG4 (Telemedia Act), and § 58 RStV5 (Interstate Broadcasting Agreement), the promotional background of a posting must be recognizable to users. However, the guidelines leave room for interpretation regarding their implementation (Krouwer etal. 2017). In Europe, inconsistent court rulings on so-called surreptitious advertising have added to the uncertainty surrounding disclosure, with some influencers having to pay penalties for their content. For instance, influencer Pamela Reif was fined for promoting a brand without a sponsorship contract because her substantial influence made it count as advertising. In contrast, lifestyle influencer Vreni Frost was not fined as she tagged companies for editorial purposes rather than commercial content. Yet another social media influencer, Cathy Hummels, had her case dismissed due to the inability to prove that she was paid for the respective post. These inconsistencies are not limited to sponsored content alone but also extend to situations where influencers promote products they have personally purchased. Consider influencers who, after a long search, find a product that solves a problem they have had for a long time. The influencers might genuinely wish to recommend the product on their social media account if convinced of the product. Here, the question arises: are they supposed to disclose the content as not sponsored? Are they supposed not to disclose at all, since the product was bought by themselves and the persuasion attempt in her media content rests upon a genuine recommendation? Due to the incising discussion about covert marketing in social media and the growing number of legal issues, influencers started to use declarations such as “# non-sponsored” or “this post is not sponsored” to indicate a genuine recommendation for which they were not remunerated in any way (De Veirman etal. 2019; Stubb & Colliander 2019). Another reason for marketers’ and influencers’ reluctance to disclose sponsorships is the ambiguity of whether such disclosures negatively affect brand and product attitudes and influencer evaluations (De Veirman & Hudders 2020; Wojdynski & Evans 2016). Subsequently, by exploiting the room for interpretation in the current regulation, in practice, influencers use a variety of disclosures (De Cicco etal. 2021). Sponsored posts, videos, and other social media content can, for instance, be tagged with explicit sponsorship labels such as “sponsored” content, as “advertising” or “advert”. Simultaneously, content can also be ambiguously marked with concealing tags such as “collaboration,” “in association with,” “thanks to [brand] for making this possible.” In an attempt to further conceal paid advertisement, influencers can also call themselves “ambassador” or try to use inconspicuous and challenging to recognize abbreviations and acronyms such as “ad,”6 “sp,” “spon,” and “collab” (De Cicco etal. 2021). Amidst the current uncertainties around social media sponsorship disclosure and the plethora of possibilities to explicitly disclose and even try to conceal paid advertisements, social media is one of the dominant channels for consumer engagement and purchasing. The rapid expansion in the variety and availability of online products in e-commerce 2 VRM—Vlaamse Regulator voor de Media (Flemish Regulator for the Media) 3 UWG—Gesetz gegen den unlauteren Wettbewerb (Law against unfair competition) 4 TMG—Telemediengesetz (Telemedia Act) 5 RStV—Rundfunkstaatsvertrag (Interstate Broadcasting Agreement) 6 Albeit “ad” is a common label for advertising in English speaking countries such as the USA, Canada, or the UK, in the rest of the world, “ad” is not necessarily well known and a common identifier for advertising (Medienanstalten 2018; Schnoor 2018). Electronic Markets (2024) 34:60 Page 3 of 27 60 has prompted consumers to increasingly rely on intermediaries, such as social media influencers, for product discovery and recommendations (Cao & Belo 2024). Given that influencers play a crucial role in electronic markets, acting as trusted guides who shape consumer preferences and drive purchasing decisions within digital platforms, understanding the effect of various sponsorship disclosures (De Veirman & Hudders 2020; Wojdynski & Evans 2016) is critical—not only for influencers and the brands that use influencer marketing but also for maintaining the integrity of the platforms on which they operate. However, to truly capture the real-world dynamics of sponsorship disclosure, it is crucial that such studies be conducted using authentic influencers on actual social media platforms rather than relying solely on reported behavior or fictitious scenarios (Cao & Belo 2024, Hughes etal. 2019). Research that involves real influencers interacting with their genuine followers provides a more accurate and comprehensive understanding of how different disclosure strategies affect audience engagement and influencer and brand-related perception. Prior work explored the impact of disclosure language in influencer advertising on various outcomes such as ad recognition, brand attitude, and purchase and sharing intention among consumers (Boerman 2020; e.g., De Veirman & Hudders 2020; Evans etal. 2017; Lee & Kim 2020) often by relying on fictitious posts and influencers (De Veirman etal. 2017; De Veirman & Hudders 2020; Evans etal. 2017). Whenever studies used authentic posts, they typically did not measure their impact on a real audience that could be composed of viewers or followers (e.g., Boerman, 2020; Hughes etal., 2019; Stubb & Colliander, 2019). However, distinguishing between viewers and followers is crucial because viewers represent the passive audience who may come across the content incidentally, while followers are an engaged audience who actively subscribe to the content creator’s updates. Understanding the impact on both groups provides a more comprehensive assessment of how authentic posts influence different levels of audience engagement and perception. Against the background that current knowledge on consumers’ perceptions of advertising disclosures might hold only partly in real-world settings, this work investigates the effects of different disclosure strategies based on data from two field experiments and an online survey with an experimental design. In our studies, we partnered with an Instagram microinfluencer (< 100,000 followers (Campbell & Farrell, 2020)). We collaborate with her and her audience to address the following research question: Given four disclosure strategies (i.e., explicit sponsorship disclosure, concealing disclosure, impartiality disclosure, and no disclosure at all) what is the optimal strategy from an influencers’ and marketers’ perspective? In this article, we operationalize optimal disclosure strategy in terms of (i) influencer-relevant outcomes, including changes in their credibility or their ability to engage individuals with their posts and (ii) outcomes relevant to marketers, such as individuals’ attitude towards the brand or their intention to purchase the product showcased in the content. After all, marketers’ motivation to revert to influencer marketing (and in particular micro-influencer marketing) is driven by brand-related and product-sales-related motives (Lou etal. 2019), while in a constantly expanding share of influencers, (micro-) influencers’ goal is to remain relevant and increase the follower base (Wies etal. 2022). In summary, our research offers various contributions to both practical and theoretical domains. On a practical level, it furnishes clear guidance on disclosure strategies for marketers and influencers alike. Theoretically, our work highlights the critical distinction between followers and nonfollowers in influencer marketing. Our findings demonstrate that disclosure strategies impact these two groups in distinct ways, suggesting that future research and theoretical frameworks must account for this differentiation to fully understand the effectiveness of sponsorship disclosures and their influence on consumer behavior. Related work andhypotheses At the core of this research are four disclosure strategies that are recurrently appearing in literature and practice: explicit sponsorship disclosure, concealing disclosure, impartial Table 1 Disclosure strategies Type Description with exemplary disclosure tags Explicit sponsorship disclosure Clear cue about sponsored content through labels such as “advertising,” “sponsored,” “advertisement,” and “advert” Concealing disclosure Hidden cue about sponsored content through the use of inconspicuous and difficult to recognize abbreviations and acronyms such as “ad1,” “collab,” “ambassador,” “sp,” and “spon” Impartiality disclosure Clear cue about non-sponsored content through tags such as “unpaid, ” “non-sponsored” or “this post is not sponsored.” The main goal of this strategy is to highlight influencers’ altruistic motivation to recommend a product or brand further. No disclosure Ambiguous state in which influencers provide no information/no clues about whether the content is commercial and was sponsored or not. Electronic Markets (2024) 34:6060 Page 4 of 27 disclosure, and no disclosure at all. Table1 provides an overview of the four disclosure strategies and corresponding tagging examples. While explicit sponsorship disclosure and impartial disclosure are transparent ways to inform viewership whether the content is sponsored, concealing disclosure and no disclosure represent attempts to obfuscate information on whether the content is sponsored. Theory and prior research suggest that both influencer and marketer-related outcomes might vary depending on the disclosure strategy employed. Disclosure strategies andinfluencer‑relevant outcomes This article focuses on three influencer-relevant outcomes that directly determine the influencers’ effectiveness as opinion leaders and ambassadors within a marketing campaign. These are influencers’ ability to motivate user engagement (i.e., intense interaction) with their content (Lou etal. 2019; Wies etal. 2022); their ability to induce positive sentiments through their content (Lou etal. 2019); and influencer credibility. The rationale for focusing on these three outcomes stems from the fact that they are the ones of interest to marketers when cooperating with (micro-)influencers (Childers etal. 2019). User engagement User engagement is the social media equivalent of “consumer behavioral engagement” in traditional marketing channels (Tafesse & Wood 2021). The concept of consumer engagement has its roots in relationship marketing (Ashley etal. 2011; Vivek etal. 2012) and has garnered significant attention from both research and practice (Boujena etal. 2021). Vivek etal. (2012) defined engagement as “the intensity of an individual’s participation in and connection with an organization’s offerings or organization activities” (p. 127). Regarding influencers’ performance as opinion leaders, user engagement might be at least if not more important than influencers’ number of followers (Wies etal. 2022). In fact, a high number of followers are relatively meaningless if the audience does not interact with the social media content the influencer provides. As recent literature suggests, a large follower base does not necessarily guarantee high engagement, and a small follower base does not necessarily mean low engagement (De Veirman etal. 2017; Djafarova & Rushworth 2017; Tafesse & Wood 2021). Nanoand microinfluencers, for instance, often generate higher engagement rates than celebrities or mega-influencers because they offer their followers the benefits of personal accessibility and high perceived authenticity (Campbell & Farrell, 2020). Thus, although the number of followers remains important for influencer marketing, user engagement is even more important for measuring online marketing success (Tafesse & Wood 2021; Wissman 2018). Typically, user engagement can be conceptualized in terms of likes, shares, or comments (Tafesse & Wood 2021) and provides essential insights into the success of marketing campaigns by indicating how well the featured content was received (Barger etal. 2016; Gummerus etal. 2012; Lou etal. 2019). Moreover, it ensures that influencers’ content is visible to more people since Instagram’s algorithms tend to highlight posts from accounts that experience a high level of viewership engagement (Saraco 2020). The tenet is that the higher the engagement with influencers’ media content, the better the influencer’s content visibility and greater reach. Thus, it is not surprising that both influencers and marketers seek to optimize user engagement concerning their content (Wies etal. 2022). Against this background, the question arises of whether and how sponsorship disclosure strategies affect users’ engagement with influencers’ content. Prior literature presents ambiguous results on how disclosure could affect social media engagement (Boerman 2020). For instance, Evans etal. (2017) found a negative effect of advertisement recognition on individuals’ engagement intention. In particular, when individuals understand that influencers’ content is a persuasive attempt in the form of advertisement, scholars found that individuals are less willing to share the content with their peers (Evans etal., 2017). In contrast, Johnson etal. (2019) and Lou etal. (2019) found no effect of advertisement recognition on engagement, while Boerman (2020) reports that standardized disclosure versus no disclosure can have a positive effect on individuals’ intention to engage with the post. Similarly, Karagür etal. (2022) found that consumers appreciate transparency, whether the content is sponsored or not, by being more willing to engage with the post. Due to the mixed results presented by the extant body of the literature, this article seeks to investigate the sub-research question RQ1: whether and how explicit sponsorship disclosure, impartiality disclosure, concealing disclosure, and no disclosure might impact viewers’ engagement with content. Following Evans etal. (2017), one would expect that explicit disclosure would negatively influence user engagement. In contrast, following Karagür etal. (2022), one would expect to see a positive influence of explicit sponsorship disclosure and impartiality disclosure on user engagement with the post. We empirically investigate this sub-research question in Study 1. Electronic Markets (2024) 34:60 Page 5 of 27 60 User sentiment Besides user engagement, this article also seeks to empirically investigate the effect of various disclosure strategies (i.e., explicit sponsorship disclosure, impartiality disclosure, concealing disclosure, and no disclosure) on user sentiment. We empirically investigate this sub-research question in Study 2. Generally, user sentiment reveals individuals’ subjective attitudes toward the content (Lou etal., 2019). Influencers must stay attuned to what users engage with and what consumers prefer or dislike. In today’s competitive influencer landscape, with rapidly changing consumer preferences and the transient popularity of influencers and brands (Wies etal. 2022), staying successful requires recognizing shifts in user preferences through sentiment analysis (Homburg etal. 2015). Viewership sentiment is especially important for microinfluencers with < 100,000 followers, as these influencers cannot afford to miss out on shifts in viewership preferences. Many interactions may indicate engaging, interesting content likely to provoke the user’s positive attitude toward the influencer. However, these interactions (e.g., likes and comments) do not convey the underlying sentiment of these engagements (Gräve 2019). Generally, user sentiment is extracted from users’ comments on the social media platform in question (Lou etal. 2019). It can be an abstract key performance indicator for the quality of content but also a good proxy for a campaign’s success. Thereby, more positive sentiment indicates higher success (Gräve 2019). Despite the importance of consumer sentiment for measuring the success of a campaign, it is surprising that only little literature investigates the effects of sponsorship disclosure on user sentiment. The work of Lou etal. (2019) is a notable exception. In their study, the scholars hypothesize a negative effect between sentiment and sponsorship disclosure by expecting to measure the least positive sentiment for explicit sponsorship disclosures, followed by ambiguous disclosures and no disclosure. However, based on data from influencer marketing campaigns from 41 apparel brands in the U.S., the scholars found no statistically significant effect of the three investigated disclosures on sentiment (Lou etal. 2019). Notably, Lou etal. (2019) use a supervised learning algorithm to conduct sentiment analysis on users’ comments. In this article, we follow a distinct approach and conceptualize sentiment in terms of visual cues represented by an emoji scale (for more details, please see Study 2 in the following sections). Our approach builds on prior literature suggesting that emoji can have similar properties to lexical sentiment indicators, and sentiment expressed in the form of an emoji scale reflects the same findings as content-based analysis (Phan etal. 2019). Formally, we investigate following research question: RQ2: What is the effect of various disclosure strategies (i.e., explicit sponsorship disclosure, impartiality disclosure, concealing disclosure, and no disclosure) on user sentiment. Influencer credibility Credibility is another construct that is particularly important for influencers. Because at its core, influencer marketing rests on the basic principle of peer endorsement and electronic word of mouth [eWOM] where a typical satisfied customer “endorses or demonstrates a product or service and acts as a source of information to influence the acceptability of the message” (Munnukka etal., 2016, p.182), it is essential that influencers come across as everyday individuals that are trustworthy and competent (De Cicco etal. 2021). In line with this notion, influencer credibility, sometimes also referred to as source credibility, is typically conceptualized as a composite construct of influencer trustworthiness, perceived influencer expertise, and perceived similarity (Vrontis etal. 2021). Prior literature has shown that the acceptability of the message conveyed by the influencer hinges on recipients’ perception of the influencer’s credibility (Munnukka etal. 2016). Furthermore, research on interpersonal relationships has repeatedly illustrated the negative attributions resulting from impartial, equivocal, and deceitful disclosures (Carr & Hayes, 2014; Toma & Hancock, 2012). Therefore, concealing disclosure and no disclosure that appears to be clouding or misleading the user’s impression of the influencer’s bias may negatively impact influencers’ credibility. Case in point: If users see a post without a sponsorship disclosure, it should be safe to assume the content of the post is unbiased. However, the post’s content might lead users to suspect that the post is sponsored after all, even though not disclosed. To resolve this issue, one can expect users to downgrade the influencer’s credibility (Carr & Hayes 2014; Jeong etal. 2019). This can be explained using Psychological Contract Violation (PCV) theory. The PCV theory of trust has been well established in the context of buyer-seller relationships in an e-marketplace (Pavlou & Gefen 2005). It refers to the perception that there has been a breach of the psychological contract between two parties (i.e., seller and buyer). The psychological contract is an unwritten set of expectations and obligations between two parties (Pavlou & Gefen 2005). Such psychological contracts can refer to rules and expectations that focus on the long-term relationship between parties and include implicit obligations beyond those that can be explicitly described in legal terms. Such relational psychological contracts emphasize trust, loyalty, and mutual commitment Electronic Markets (2024) 34:6060 Page 6 of 27 to each other. A psychological contract breach occurs when one party feels that the other party has violated one or more of the implicitly set unwritten rules and expectations (Pavlou & Gefen 2005). In our research context, psychological contracts entail user beliefs and expectations that influencers should behave in a particular manner—e.g., provide organic recommendations and explicitly disclose sponsorship if their content is sponsored (Wang & Wang 2019). Thus, individuals’ interaction with an influencer’s content can constitute an implicit psychological contract that the influencer is honest and trustworthy. Psychological contract violation in an influencer-viewership relationship occurs when viewers perceive that the influencer has breached their implicit obligations or expectations. Viewers develop certain expectations regarding the influencers’ authenticity, transparency, and honesty in their content. When influencers engage in behaviors that violate these expectations, such as undisclosed sponsorships or misleading endorsements, viewers may feel a sense of betrayal, disappointment, or mistrust. For example, suppose viewers believe an influencer genuinely uses and recommends a product but later discover that the influencer was paid to promote it without disclosure. In that case, viewers may perceive a psychological contract violation. This violation undermines the trust and perceived authenticity of the influencer-viewership relationship. Lack of disclosure or misleading disclosure for content that resembles an advertisement violates the psychological contract between the user and influencer (Wang & Wang 2019), so users decrease their trust in the influencer as a source for credible and unbiased information. In turn, with increasing recent evidence that transparency on sponsorship reduces perceptions of manipulative intent while increasing consumers’ perceived credibility of the marketing agent (Abendroth & Heyman 2013, Cao & Belo 2014, Evans etal. 2019, Wang & Wang 2019), we can expect that: H1: Compared to explicit sponsorship disclosure, concealing disclosure and no disclosure will negatively impact influencer’s credibility. Thereby, prior literature indicates that influencers’ credibility might be higher for impartiality disclosure than explicit sponsorship disclosure (e.g., Carr & Hayes, 2014; Dekker & Van Reijmersdal, 2013; Hwang & Jeong, 2016). For instance, Hwang and Jeong (2016) investigated differences in blogger credibility for blog posts with explicit sponsorship disclosure and blog posts tagged as genuine recommendations—i.e., impartiality disclosure. The scholars report a higher blogger credibility for genuine recommendations. Similarly, Carr & Hayes (2014) and Dekker & Van Reijmersdal (2013) found that consumers evaluate influencers more positively and perceive them as more credible when using extended disclosures. Disclosure strategies andmarketer‑relevant outcomes To be successful, businesses must handle their customers interactively, collaboratively, and in a personalized manner (Wieneke & Lehrer 2016). Given this context, it becomes crucial to anticipate how different disclosure strategies may influence consumers’ reactions to both the brand and the product. Compared to other marketing activities, such as celebrity endorsement or advertising in the traditional media, influencer marketing, and particularly micro-influencer marketing, represents a rather effective yet inexpensive way to reach a broad customer base (Lou etal. 2019). Nevertheless, influencer marketing also induces costs that need to be outweighed by the benefits it generates. Business logic mandates that marketers do not invest in influencer marketing without first analyzing the potential return on investment and comparing it to other marketing strategies. Although it remains unclear how influencer marketing’s success can be measured (Childers etal. 2019), marketing campaigns’ end goal is to promote the brand and ultimately drive sales and revenue up (Campbell & Farrell 2020). How customers perceive and experience a brand is reflected in their attitudes. At the same time, consumers’ intention to purchase the advertised products is a good indicator that the marketing campaign induces sales. Accordingly, in this study, we focus on users’ attitudes toward the brand and individuals’ purchase intention as two key performance metrics for influencer marketing success (Campbell & Farrell 2020). Attitude towardsthebrand Faced with highly competitive markets, the fleeting whims of consumers, and the popularity of products and brands, developing a solid consumer-brand relationship is becoming more critical than ever. In theory, brands can engage in consumer-brand relationships and try to strengthen this relationship by pursuing marketing campaigns in social media via (i) their brand accounts and pages or (ii) the influencers (Lou etal. 2019). In recent years, influencers, especially micro-influencers, have become increasingly popular. These influencers are seen as trusted figures by the general public and their followers (Boerman 2020). People tend to find influencers more trustworthy and reliable than traditional advertising sources (Lou etal. 2019). In marketing partnerships, influencers often act as brand ambassadors (Boerman 2020) and create content that seamlessly fits with regular non-commercial posts (De Veirman & Hudders 2020). As a result, studies consistently show that brand-promoted messages are perceived as more biased compared to influencerpromoted brand messages (Lou etal. 2019). Electronic Markets (2024) 34:60 Page 7 of 27 60 In general, branding and creating a solid brand play an essential role in the success of the products and services of a company. For instance, since branding can have strong signaling effects for products that are initially unknown to consumers, it is vital that consumers develop positive attitudes toward the brand. After all, positive attitudes towards the brand (which can culminate into “brand love”) bring a variety of advantages. Such advantages include increased brand loyalty or willingness to pay price premiums (Batra etal. 2012). Due to the importance of attitudes towards the brand for a company’s success, scholars investigated the effects of sponsorship disclosure on brand attitudes. The findings are mixed, suggesting that sponsorship disclosure could have both positive and negative impacts on consumers’ brand attitude. For instance, in a study with television ads, Boerman etal. (2012) found that when individuals recognize the content as advertising, they are more critical of the content message, leading to less favorable attitudes toward the brand. Similarly, a meta-analysis by Krouwer etal. (2017) reported that disclosing sponsored content reduced brand attitudes. In contrast, more recent research argues that transparency on sponsorship reduces perceptions of manipulative intent (Abendroth & Heyman 2013, Cao & Belo 2014, Wang & Wang 2019), which, in turn, translates into a positive effect on brand attitude, by mitigating the negative impact of advertising recognition (Campbell and Evans 2018, Evans etal. 2019). Given that individuals generally respond more negatively when they perceive manipulative intent (Campbell and Evans 2018), we can expect that: H2: Compared to explicit disclosure, all other disclosure strategies will result in less positive brand attitudes. User intention topurchase Besides the advantages of influencer marketing for branding, marketers collaborate with influencers to promote and sell their products and services. To assess the effectiveness of marketing campaigns and estimate the return on investment, marketers frequently rely on users’ intent to make a purchase as a practical proxy for their actual purchasing behavior. Consequently, it is understandable that there are existing studies investigating the link between sponsorship disclosures and individuals’ purchase intentions. Again, prior literature presents ambiguous results. While various studies report that disclosure that helps recognize the content as sponsored (i.e., explicit sponsorship disclosure) and thus harms consumers’ intentions to purchase (Boerman etal. 2015; Van Reijmersdal etal. 2016; Wojdynski & Evans 2016), other studies found no significant direct effect of disclosure on the intention to purchase. Yet, recently, research increasingly reports a positive link between transparent sponsorship disclosure and consumers’ purchase intentions (Wang & Wang 2019, Woodroof etal. 2020). Drawing on these recent insights paired with the well-established relationship between attitudes towards a product and brand and purchase intentions, we posit that: H3: Compared to explicit disclosure, all other disclosure strategies will result in a lower intention to purchase. Potential mediating andmoderating factors Influencer credibility mediate marketers‑relevant outcomes Besides the direct effect of sponsorship disclosures on users’ attitudes toward the brand and intention to purchase, we can also expect indirect effects through influencers’ credibility. As mentioned previously, influencers’ credibility is important not only for the influencer themselves but also for marketers, as it has downstream effects on brand attitude (Hwang & Jeong, 2016) and intention to purchase (Ohanian 1990; Rifon etal. 2004). Our expectation that influencer credibility mediates the effect of various disclosure strategies aligns with theoretical perspectives on the processing of persuasive communications. The Persuasion Knowledge Model (PKM) introduced by Friestad and Wright, (1994) theorises how individuals evaluate and respond to influence attempts from marketers. The PKM model consists of “a target” (i.e., the party who is attempted to be persuaded), “an agent” (i.e., the party who tries to persuade the target), and three types of knowledge on both sides: topic knowledge, agent knowledge, and persuasion knowledge. From the targets’ (i.e., consumers’) perspective, topic knowledge refers to beliefs about the topic or advertisement subject. Persuasion knowledge adheres to beliefs about the marketers’ motives and strategies. Finally, agent knowledge refers to beliefs and perception of the agent’s (i.e., influencer’s) traits, abilities, and goals. Transferring the PKM to the influencer marketing context, the influencer is acting as an agent, and the target is the social media user exposed to the agent’s advertising post. The paid posting on social media can be considered as persuasion attempt (Kim and Song 2018). The response to a persuasion attempt is based on these three types of knowledge and results in the target’s personal persuasion coping behavior (Kirmani and Campbell 2009). Although marketing literature acknowledges several different strategies that consumers can use to respond to persuasion attempts (Kirmani and Campbell 2009), in this study— for the sake of simplicity—we classify consumers’ coping behavior into resistance or compliance with the agents’ request. Several scholars concluded in their studies that persuasion knowledge is a central factor for consumers’ coping behavior Electronic Markets (2024) 34:6060 Page 8 of 27 (Boerman etal., 2015, 2017). Individuals use their persuasion knowledge to evaluate and process advertising situations on a daily basis. If individuals infer that a salesperson has ulterior motives, they activate their persuasion knowledge and become suspicious about the company, resulting in a negative evaluation (Campbell & Kirmani, 2000). Besides persuasion knowledge, in this study we also consider the link between agent knowledge and consumers’ coping behavior. Because in influencer marketing the agent and its characteristics are important factors in conveying the advertisement message to the targets, agent knowledge is most likely at least as important as the persuasion knowledge itself. At its core, influencer marketing rests on the basic principle of peer endorsement and (e)WOM where a typical satisfied customer “endorses or demonstrates a product or service and acts as a source of information to influence the acceptability of the message” (Munnukka etal., 2016, p.182). The targeted individuals’ acceptability of the message conveyed by the agent hinges on targets’ perception of the agents’ credibility (composite of trustworthiness, expertise, and similarity (Munnukka etal. 2016). This is especially the case when the endorsed product or brand is not known to the targeted individuals. Altogether, according to the PKM model, the persuasiveness of a message (in this case, the disclosure) can alter the receiver’s attitude towards the communicator (the influencer), which subsequently affects responses to the endorsed product. As prior research showed, the effect of influencer credibility can be, in some cases, strong and long-lasting. Fink etal. (2020) for instance report that in a Facebook community, celebrity-endorsed marketing influencer credibility has a multi-year lasting positive effect on the intention to purchase. Additionally, conceptualizing influencers’ credibility as a construct composed of trustworthiness, perceived influencer expertise, and perceived similarity, prior research indicates that credibility influences attitudes and purchase behaviors mostly through the trustworthiness component because the influencers try out the products themselves (Uzunoğlu & Kip 2014). Against this background, a decrease in influencers’ credibility is expected to negatively impact consumers’ attitudes toward the brand and their purchasing intentions. This should be especially the case for obfuscating disclosure strategies (i.e., no disclosure or concealing disclosure) where followers might question the integrity of the influencer’s content (Carr & Hayes 2014; Magnini 2011) and trustworthiness. Altogether, in line with prior literature demonstrating that if the origins of a message seem disingenuous, users are more likely to distrust the content (Dou etal. 2012), it would be sensible to expect that: H4: Influencer credibility, and in particular the trustworthiness component, will mediate the effects of various sponsorship disclosures on individuals’ attitudes toward the brand and intentions to purchase. Followers versusnon‑followers The rate at which influencers’ credibility might suffer from concealed disclosure strategies is likely to depend on whether the viewership of the post is followers or non-followers. Prior research investigating the effect of disclosure on consumer perceptions primarily relied on laboratory-style experiments with fictitious posts and influencers (e.g., De Veirman etal., 2017; De Veirman & Hudders, 2020; Evans etal., 2017). Thus, related work on the effect of various disclosure strategies does not distinguish between the opinions and evaluations of nonfollowers and followers of real influencers. This is surprising for two reasons: Firstly, prior work on influencer marketing suggests that followers and non-followers have different sensitivity levels regarding the product-influencer fit (Belanche etal. 2020) and downstream outcomes of the marketing campaign (Leung etal. 2022). Secondly, the content of Instagram authors targets first and foremost followers, and individuals generally trust more friends and acquaintances than complete strangers (Resnick & Zeckhauser 2002). Following an influencer over time allows followers to learn about them. As followers are continually exposed to details of an influencer’s life, followers form certain attitudes toward the influencer and even experience a sense of friendship (Boerman 2020). For some followers, it seems as if they have a long-distance friend in the influencer (Djafarova & Rushworth 2017). Moreover, by being able to interact with the influencer (e.g., by commenting on an influencer’s posts), followers will feel increasingly similar to the influencer (Schouten etal. 2020). Also, influencers tend to address their followers directly in their posts, implying some degree of closeness and making followers think they are peers (Erz & Heeris Christensen 2018; Gannon & Prothero 2018). These arguments are in line with the Parasocial Interaction Theory (Horton & Wohl 1956), which refers to the fact that mass media viewers or listeners come to build some psychological relationship (e.g., friendship) with media personalities, despite having no contact or interaction with them. Transferring the Parasocial Interaction Theory to social media suggests that an individual A (i.e., follower) following another person B (i.e., the influencer) enables person A to develop specific attitudes or even experience feelings of intimacy towards person B (Boerman 2020). The special emotional bond influencers build with their followers is the reason why influencers can sway their followers’ opinions and behaviors (Ki etal. 2020). As the bond between influencers and their followers strengthens, so does their psychological contract and the ability to exert influence over their followers (Ki etal. 2020). Infractions on influencers’ side, such as concealed disclosure strategies, represent psychological contract violations that followers might perceive as a type of betrayal that significantly damages their relationship and has adverse spillover effects on the brand (Ki etal., 2020). In contrast, non-followers’ relationship with the influencer is neither Electronic Markets (2024) 34:60 Page 15 of 27 60 Table 4 Correlation matrix Trust-worthiness Expertise Similarity Purchase intention Attitude tw. the brand Concealing disclosure Impartiality disclosure No disclosure Female Activity level on Instagram Explicit disc. × followers Concealing disc. × followers Impartiality disc. × followers No disc. × followers Trust-worthiness 1 Expertise 0.711 1 Similarity 0.579 0.546 1 Purchase intention 0.608 0.560 0.542 1 Attitude tw. the brand 0.736 0.638 0.525 0.659 1 Concealing disc. −0.004 0.017 0.006 −0.004 −0.004 1 Impartiality disc. 0.047 0.009 −0.004 0.007 0.059 −0.334 1 No disc. −0.021 −0.042 0.006 −0.029 0.003 −0.338 −0.339 1 Female 0.254 0.197 0.412 0.236 0.179 −0.015 −0.004 −0.019 1 Activity level on Instagram 0.196 0.174 0.189 0.146 0.173 0.007 −0.019 −0.006 0.302 1 Explicit disc. × followers 0.057 0.093 0.128 0.094 −0.004 −0.263 −0.264 −0.267 0.162 0.092 1 Concealing disc. × followers 0.136 0.157 0.172 0.102 0.113 0.765 −0.255 −0.259 0.132 0.062 −0.201 1 Impartiality disc. × followers 0.177 0.122 0.170 0.099 0.141 −0.260 0.779 −0.264 0.148 0.046 −0.206 −0.1989 1 No disc. × Followers 0.134 0.138 0.219 0.081 0.117 −0.247 −0.248 0.731 0.137 0.107 −0.195 −0.1891 −0.193 1 Marker variable 0.066 0.071 0.108 0.087 0.016 −0.009 −0.009 0.001 0.068 0.062 0.037 0.026 0.007 0.054 Electronic Markets (2024) 34:6060 Page 16 of 27 A 5-point semantic differential scale provided by Spears and Singh (2004) measured individuals’ brand attitudes. Participants were asked to evaluate the various characteristics of the brand MyProtein (the brand shown in the posting), using five bipolar characteristics: unappealing/appealing, bad/good, unpleasant/pleasant, unfavorable/favorable, and unlikable/likable. This scale also proved reliable, with α = .851. Individuals’ purchase intention was measured with three items on a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree). The original scale was developed by Berens etal. (2005) and entailed the following questions: “If you were planning to buy a product of this type, would you choose this product?,” “Would you purchase this product?,” “If a friend was looking for a product of this type, would you advise him or her to purchase this product?.” In our study, this scale also proved reliable, with α = .817. To capture the attitudes and behavioral intentions of a broad group of individuals who did not necessarily have an attachment to the influencer and therefore might react differently to the post and the advertisement disclosure, the study link was circulated among followers of the influencer and non-followers. Moreover, individuals were contacted and invited to participate in the online survey via social media platforms. To expand the reach of the survey, we also used the snowball method and asked participants to post the invitation on their social media pages (Wagenaar and Babbie 2004). In total, 1175 individuals responded to our invitations to participate. While 826 individuals completed the survey, after checking for attention bias, rushing through the survey (i.e., all participants who completed the survey significantly faster (approximately 5 min) or significantly slower (about 20 min), and other inconsistencies in individuals’ responses, the final sample entails 657 subjects. The following descriptive statistics refer to the final sample. We kept the sample population as similar to the influencer’s target audience as possible. Eighty-two percent of the participants were female, and 17.9% were male. These results align with the fact that the influencer mainly attracts a female audience (as commented in the section discussing the study setting). Also, more females use Instagram than males (Tankovska 2020). The participants’ age range was between 15 and 38 years (Mage = 22.86, SDage = 3.48). This sample resembles the average Instagram follower of the influencer as her main audience (69%) is between 18 and 34 years old (see Fig.2 “Profile analytics of “@dyedblondpony”). This is also in line with the recent trend on Instagram that shows that Instagram is dominated mostly by younger users below the age of 34 (Tankovska 2020). On average, people have been using Instagram for 2 to 5 years. Ultimately, 65% of the participants reported following the account @dyedblondpony on Instagram. Empirical analyses andresults To explore the proposed relationships empirically while ensuring our results’ internal and external validity we undertook a series of tests. First, we performed a randomization test for the treatment administration. A one-way analysis of variance (ANOVA) corroborates that the randomization procedure was successful and shows no statistically significant differences between the experimental groups with respect to gender (prob > chi2 = 0.582), age (prob > chi2 = 0.196), and intensity of Instagram use (prob > chi2 = 0.862). Second, we checked whether our treatment administration itself was successful—i.e., whether the administered treatments successfully influenced participants’ recognition of the advertisements—crucial for evaluating the subsequent effects on attitudes and behaviors. To this end, we asked participants to recall whether the post they viewed contained an expression that the post was an advertisement. We compared recognition rates across the different disclosure types. If all participants, irrespective of their group, would recognize the shown post as an advertisement, our treatments would not have been appropriate to measure the outcomes of interest. However, our data shows that advertisement recognition levels are depending on the administered treatments. The participants (63.13%) who viewed the explicit disclosure (advertisement) and 30.30% who viewed the impartiality disclosure post recalled having perceived the cue that the viewed content was advertising. In contrast, only 15.24% of the participants who viewed the concealing disclosure (#ad) and 16.07% of the no disclosure condition reported seeing a label indicating that the content was advertising. One-way analysis of variance (ANOVA) corroborates that in terms of advertisement recognition, the groups are significantly different from each other (prob > chi2 = .00). Third, to ensure that we do not report spurious results that might arise from complex models (Armstrong 2012), we tested our hypotheses H1 to H5 via a multivariate analysis of variance (MANOVA), linear regressions, and mediation analyses. Before running the MANOVA we computed the correlation factors between the variables of interest (see Table4). None of the correlation factors are higher than .79. Notably, all hypotheses are formulated such that they postulate effects of different disclosure strategies compared to a clear tagging of advertising (i.e., explicit disclosure). Explicit tagging of sponsored content is desired from not only a legal but also from an ethical perspective. Therefore, in all analyses, we choose the explicit disclosure strategy as our baseline strategy and compare the other disclosure strategies with this baseline in order to understand how they compare to each other. With this in mind, we ran MANOVA using individuals’ purchase intention, their attitude towards the brand, and influencer credibility components as dependent variables (DVs). The various postulated disclosure strategies, as well as the Electronic Markets (2024) 34:60 Page 17 of 27 60 postulated interaction between them and followership status, were used as independent variables. For better readability and interpretation of the results, Table5 reports only Pillai’s trace statistics. In general, Pillai’s trace is considered robust against violations of some of the assumptions of MANOVA, such as normality and homogeneity of variances (Finch and French 2013), and is therefore a favorable choice in many cases. The tests statistics do not indicate a significant multivariate relationship between the various disclosure strategies (concealing disclosure, impartiality disclosure, or no disclosure) and the dependent variables of interest. However, the statistics do support the existence of a moderating effect of followership on these disclosure strategies, by showing statistically significant Pillai’s Trace statistics (p < 0.01) for the interaction terms disclosure strategies × followership status. Altogether, the MANOVA results suggest that H5, which states that followership status is moderating the effects of sponsorship disclosure, finds empirical support in the data. We use these insights to estimate parsimonious OLS models that test H1 to H3 and mediation analyses for H4 to H5. In H1 we expected that compared to explicit sponsorship disclosure, concealing disclosure and no disclosure will negatively impact influencer’s credibility. H2 postulated that compared to explicit disclosure, all other disclosure strategies will result in less positive brand attitudes, while H3 expects the same for individuals’ purchase intentions. To empirically test these three hypotheses, we formulate and estimate a parsimonious OLS with following specification: DV =𝛽0+𝛽1Female +𝛽2Activity level + 4 ∑ j=2 𝛽3jDisclosure strategy + 4 ∑ j =1 𝛽4jDisclosure strategy ×Followership status +𝛽5Marker variable + 𝜀 Table6 presents the corresponding estimation results for which all VIF values are lower than 3.3. The results indicate that, contrary to our expectations, there are no statistically significant effects of various disclosure strategies on the individual components of influencer credibility, individuals’ attitudes towards the brand, or purchase intentions. Compared to explicit disclosure, none of the other disclosure strategies have an impact on the dependent variables of interest. Therefore, we can conclude that H1 through 3 are not supported. However, the results partially support Hypothesis 5, which suggested that followership status moderates the impact of various sponsorship disclosures on influencer credibility and brand attitude, with followers expected to respond more positively to transparent strategies such as explicit disclosure and genuine recommendations. We find substantial evidence that followership status indeed moderates the influence of disclosure strategies on the components of influencer credibility, enhancing the impact of each strategy. However, contrary to our initial expectation that followers would favor transparency, the interaction terms reveal that genuine recommendations with impartiality disclosure generate the highest influencer trustworthiness among followers (β =.856, p < .001), followed by concealing disclosure, which unexpectedly will yield the second highest level of influencer trustworthiness (β = .822, p < .001). In utter contrast to our expectations, the interaction term of explicit disclosure with the followership status will lead to the lowest level of influencer trustworthiness (β = .437, p < .05). Similar patterns emerge for influencer’s perceived expertise and similarity. Regarding the moderation effect of various disclosure strategies with followership status on individuals’ attitude towards the brand and purchase intention, we find only Table 5 Pillai’s trace from MANOVA analysis with DVs: attitude towards the brand, purchase intentions and influencer credibility components (trustworthiness, perceived expertise, and similarity) e = exact, a = approximate Statistic df F (df1, df2) FProb > F Model .462 13 65 3215 5.04 .000 a Residual 643 Gender (female) .064 1 5 580 8.76 .000 e Activity level .069 5 25 3215 1.80 .008 a Disclosure strategies (baseline: explicit disclosure) Concealing disclosure .004 1 5 639 .45 .810 e Impartiality disclosure .002 1 5 639 .24 .946 e No disclosure .006 1 5 639 .78 .567 e Disclosure strategies × followership status Explicit disclosure × follower .054 1 5 639 7.33 .000 e Concealing disclosure × follower .085 1 5 639 11.88 .000 e Impartiality disclosure × follower .085 1 5 639 11.81 .000 e No disclosure × follower .120 1 5 639 17.43 .000 e Residual 643 Total 656 Electronic Markets (2024) 34:6060 Page 18 of 27 Table 6 OLS estimations DV Trustworthiness Perceived expertise Similarity Attitude tw. the brand Purchase intention Coef. SE p > z Coef. SE p > z Coef. DSE p > z Coef. SE p > z Coef. SE p > z Trustworthiness 0.428*** 0.031 0.000 0.306*** 0.040 0.000 Expertise 0.191*** 0.034 0.000 0.204*** 0.047 0.000 Similarity 0.126*** 0.028 0.000 0.243*** 0.039 0.000 Female 0.204 0.109 0.060 0.003 0.111 0.977 0.588*** 0.105 0.000 −0.073 0.073 0.315 0.091 0.091 0.318 Activity level on Instagram 0.062** 0.026 0.019 0.048** 0.024 0.046 0.012 0.024 0.618 0.019 0.015 0.198 0.001 0.019 0.947 Disclosure strategy (baseline: explicit disclosure) Concealing disclosure −0.185 0.151 0.221 −0.188 0.170 0.268 0.026 0.160 0.869 −0.045 0.114 0.694 −0.024 0.146 0.871 Impartiality disclosure −0.130 0.159 0.412 −0.118 0.174 0.498 −0.060 0.166 0.715 0.083 0.112 0.458 0.012 0.140 0.933 No disclosure −0.142 0.146 0.330 −0.261* 0.157 0.098 −0.046 0.168 0.784 0.052 0.109 0.632 0.030 0.140 0.828 Disclosure strategy × followership (interaction effects) Explicit disclosure × followers 0.437*** 0.136 0.001 0.498*** 0.143 0.001 0.809*** 0.148 0.000 −0.195** 0.094 0.039 −0.068 0.129 0.599 Concealing disc. × followers 0.822*** 0.157 0.000 0.833*** 0.157 0.000 0.908*** 0.142 0.000 −0.053 0.103 0.610 −0.132 0.128 0.306 Impartiality disc. × followers 0.856*** 0.164 0.000 0.688*** 0.158 0.000 0.978*** 0.148 0.000 −0.153 0.095 0.107 −0.185 0.122 0.131 Electronic Markets (2024) 34:60 Page 19 of 27 60 isolated support for direct effects. For individuals’ attitude towards the brand, we found only one significant moderation effect of followership status on disclosure strategies. Interestingly, the estimated effect is negative (β = −.195, p < 0.05). We expected that transparency, in the sense of explicit declarations of sponsorship, would have a positive impact on followers’ attitude towards the brand. Our data contradicts this expectation. In terms of purchase intentions, not disclosing sponsorship significantly reduces the intention to purchase among followers (β = −.252, p < 0.05). Synthesizing all OLS estimation results reveals the following insight: individuals’ attitudes towards the brand and purchase intentions are significantly influenced by influencer credibility components (p < .001), which, in turn, are significantly influenced by the interaction between disclosure strategy and followership status (p <.001). This insight begs for the question of whether influencer credibility components mediate the effect of disclosure strategies on these outcomes, as hypothesized in H4. To test this relationship, we performed a multivariate mediation analysis. Specifically, we estimate seemingly unrelated regressions in STATA18 for following general and simplified model structures (see Fig.6) where IV represents the set of independent variables composed of the various disclosure strategies, their interaction terms with the followership status, and DV represents a marketer outcome of interest (attitude towards the brand or purchase intention). Table7 reports the total indirect effects of IV on DV through the channels of the individual influencer credibility components (Trustworthiness, Perceived Expertise, and Similarity). All indirect effects of the IV through the mediator variable on the DV—e.g., the effect of concealing disclosure through influencers trustworthiness on the DV of interest (see Fig.7a)—were computed from estimations with bootstrapped and bias corrected intervals from 1000 Table 6 (continued) DV Trustworthiness Perceived expertise Similarity Attitude tw. the brand Purchase intention Coef. SE p > z Coef. SE p > z Coef. DSE p > z Coef. SE p > z Coef. SE p > z No disclosure × followers 0.770*** 0.150 0.000 0.864*** 0.140 0.000 1.097*** 0.153 0.000 −0.148 0.100 0.140 −0.252** 0.124 0.042 Marker variable 0.016 0.025 0.520 0.017 0.024 0.470 0.034 0.022 0.114 −0.024 0.016 0.130 0.018 0.019 0.354 __cons 2.252*** 0.157 0.000 2.532*** 0.168 0.000 1.362*** 0.161 0.000 0.976*** 0.137 0.000 0.443*** 0.159 0.005 VIF 2.84 2.84 2.84 2.75 2.75 * p < 0.10, ** p < 0.05, *** p < 0.01 Fig. 6 Mediation analysis model Electronic Markets (2024) 34:6060 Page 20 of 27 Table 7 Indirect effects of disclosure strategies through influencer credibility components on DVs of interest (attitude towards brand; intention to purchase) * indicates statistical significance, which is given if the bias corrected (BC) confidence intervals do not contain 0 IV Mediators: credibility components DV: attitude towards brand DV: intention to purchase Coef. Bias Boot. SE. BC: [95% conf. interval] Coef. Bias Boot. SE. BC: [95% conf. interval] Explicit disclosure Trustworthiness −0.080 0.002 0.065 −0.216 0.040 −0.057 −0.002 0.049 −0.156 0.038 Expertise −0.036 0.002 0.033 −0.104 0.025 −0.039 −0.001 0.037 −0.121 0.035 Similarity 0.003 0.001 0.020 −0.037 0.042 0.006 −0.001 0.041 −0.072 0.093 Total indirect effect −0.113 0.004 0.098 −0.314 0.065 −0.090 −0.004 0.101 −0.276 0.126 Impartiality disclosure Trustworthiness −0.056 −0.002 0.068 −0.180 0.090 −0.040 −0.001 0.050 −0.137 0.059 Expertise −0.022 0.000 0.034 −0.091 0.041 −0.024 0.000 0.038 −0.104 0.046 Similarity −0.007 0.000 0.020 −0.049 0.035 −0.015 −0.002 0.043 −0.102 0.068 Total indirect effect −0.086 −0.002 0.105 −0.278 0.140 −0.078 −0.003 0.108 −0.281 0.154 No disclosure Trustworthiness −0.062 0.002 0.064 −0.182 0.066 −0.044 −0.001 0.045 −0.131 0.044 Expertise −0.050 0.000 0.032 −0.126 0.004 −0.054 −0.001 0.035 −0.133 0.007 Similarity −0.006 0.000 0.021 −0.051 0.035 −0.012 −0.002 0.042 −0.096 0.068 Total indirect effect −0.119 0.003 0.098 −0.303 0.071 −0.110 −0.004 0.099 −0.285 0.100 Explicit disclosure × follower Trustworthiness* 0.189 0.004 0.059 0.078 0.301 0.134 0.002 0.047 0.049 0.238 Expertise* 0.096 0.001 0.034 0.040 0.174 0.103 0.000 0.038 0.044 0.194 Similarity* 0.101 0.001 0.032 0.048 0.173 0.200 0.000 0.049 0.114 0.299 Total indirect effect* 0.385 0.007 0.095 0.201 0.567 0.437 0.002 0.091 0.266 0.619 Concealing disclosure × follower Trustworthiness* 0.354 0.001 0.074 0.221 0.521 0.252 0.001 0.059 0.150 0.379 Expertise* 0.160 −0.001 0.042 0.094 0.253 0.171 0.001 0.054 0.082 0.300 Similarity* 0.114 0.001 0.031 0.059 0.181 0.224 0.001 0.049 0.141 0.337 Total indirect effect* 0.627 0.001 0.104 0.417 0.824 0.648 0.004 0.100 0.457 0.842 Impartiality disclosure × follower Trustworthiness* 0.368 0.003 0.077 0.231 0.537 0.262 0.003 0.061 0.159 0.394 Expertise* 0.132 0.001 0.038 0.066 0.213 0.141 0.000 0.047 0.069 0.260 Similarity* 0.122 0.001 0.034 0.061 0.194 0.240 0.002 0.056 0.142 0.361 Total indirect effect* 0.621 0.005 0.110 0.396 0.821 0.643 0.005 0.111 0.424 0.866 No disclosure × follower Trustworthiness* 0.333 0.002 0.070 0.209 0.487 0.237 0.002 0.054 0.146 0.357 Expertise* 0.166 0.001 0.044 0.096 0.272 0.178 0.001 0.051 0.090 0.291 Similarity* 0.138 0.001 0.039 0.068 0.225 0.272 0.003 0.061 0.170 0.412 Total indirect effect* 0.636 0.004 0.107 0.445 0.866 0.687 0.005 0.108 0.486 0.908 Electronic Markets (2024) 34:60 Page 21 of 27 60 replications. The total indirect effect in Table7 represents the cumulative indirect effect of an IV through all influencer credibility components on the DV (see Fig.7b). Ultimately, please note that in Table7 statistical significance of a coefficient is established when the bias-corrected confidence intervals do not include zero. For improved readability, we have highlighted significant entries in the table in gray. To recap, in H4 we proposed that influencer credibility, particularly its trustworthiness component, would mediate the impact of various sponsorship disclosures on individuals’ attitudes toward the brand and their purchase intentions. The indirect effects presented in Table7 offer partial support for this hypothesis by showing that only the indirect effects of the interaction terms between disclosure and followership status are statistically significant. Other indirect effects are not statistically significant. When looking closer at the estimated indirect effects of various disclosure strategies on individuals’ attitude towards the brand, it is noticeable that the indirect effect mediated by the trustworthiness component is almost double than the mediation through other components (e.g., through influencers’ perceived expertise, similarity). For explicit disclosure, for instance, the indirect effect through trustworthiness yields βExplicitDisc × follower | Trustworthiness = .189 (BC [95% conf. interval] = [.078; .301]). In contrast, the indirect effect through perceived expertise and similarity was estimated to βExplicitDisc × follower|Perceived expertise =. 096 (BC [95% conf. interval] = [.040; .174]) and βExplicitDisc × follower|Similarity = .101 (BC [95% conf. interval] = [.048; .173]), respectively. The same pattern applies also to the other significant indirect effects for the interaction terms on attitude towards the brand. In terms of individual’s purchase decision, this pattern does not replicate, leading us to conclude that H5 is supported only partly. Discussion The objective of our study was to investigate and identify the most effective disclosure strategy from the viewpoints of influencers and marketers. In pursuit of this objective, we specifically examined the comparative efficacy of explicit sponsorship disclosure and impartiality disclosure against deviant practices such as concealed disclosure or no disclosure. Influencers are driven by the need to find an optimal disclosure strategy that preserves their relevance and attracts a larger following Wies etal. (2022), while marketers are focused on a strategy that reinforces brand messaging, fosters consumer trust, and ultimately drives sales (Lou etal. 2019). Based on data from two field experiments and an online survey with an experimental design, this article identifies the optimal disclosure strategies from influencers’ and marketers’ perspectives. From theinfluencers’ perspective One of the most “important benefit of influencer marketing is its ability to stimulate engagement with the sponsored content” (Boerman 2020, p. 200). For influencers, it is essential to continuously observe individuals’ engagement with content and thus decide not only what type of content users like but also what disclosure strategy is best suited to keep the users’ engagement up. Amidst the current uncertainties around social media sponsorship disclosure and the plethora of possibilities to explicitly disclose and even try to conceal paid advertisements, it is more than ever critical to be informed about the implications of different disclosure strategies on influencers’ image and subsequent ramifications. This is particularly the case for nanoand micro-influencers, who in relation to macro-influencers have a relatively small (but usually dedicated) follower base (Campbell & Farrell, 2020; Tafesse & Wood, 2021). After all, due to the relatively small follower base, losing followers or their attention can have severe consequences for the influencers’ reach and, ultimately, their careers. The insights presented in this article indicate that the optimal disclosure strategy depends on whether influencers seek to improve their credibility levels or improve the engagement of their viewership with their posts. Influencers are often viewed as opinion leaders who first and foremost inform (Boerman 2020; Djafarova & Trofimenko 2019). Thereby, it seems that individuals generally are okay with influencers being paid or compensated for their recommendation if they are transparent about it. Assuming that the Fig. 7 a (left) Indirect effect of IV through influencer’s trustworthiness on DV; b (right) total indirect effect of IV on DV Electronic Markets (2024) 34:6060 Page 22 of 27 influencers’ goal is to increase viewership engagement with the promoted content, the results in Study 1 indicate that influencers can enjoy higher levels of user engagement with their content as long as they label their content accordingly. As long as influencers provide information on sponsorship or non-sponsorship, the influencers do not have to worry about decreasing engagement rates due to the employed disclosure strategy. As Study 1 shows, the lowest engagement rates occur for content with no disclosure. In contrast, if influencers are keen on increasing their credibility levels in terms of perceived trustworthiness, expertise, and similarity with the viewership, it depends on whether influencers seek to address their follower base or non-followers with the respective advertisement campaign. Assuming that followers are the main group influencers seek to address by marking genuine recommendations with impartiality labels, influencers can enjoy higher levels of trustworthiness on the part of followers. In contrast, influencers must take into account that for sponsored content, explicit sponsorship disclosure will harm their trustworthiness image in front of their followers. Hence, if most of the influencer’s recommendations are genuine, influencers will enjoy high trustworthiness in front of their followers by using the impartiality disclosure. If the influencer does not frequently provide genuine recommendations and is heavily involved in sponsorship, they should anticipate that complying with the current regulations, which require explicit sponsorship disclosure, will impact their trustworthiness. On the contrary, assuming that influencers’ strategic priority is to be favorably perceived by non-followers, transparent disclosure—specifically, explicit disclosure on sponsored content—is again the preferred choice. This approach yields benefits, especially for influencers whose content predominantly involves sponsorship relations with a brand rather than genuine recommendations. Our experiments indicate that non-followers tend to be particularly skeptical about impartiality disclosure. From marketers’ perspective Influencer marketing is driven by the idea that influencers are relatable individuals who can educate and inform a wide audience (Boerman 2020; Djafarova & Trofimenko 2019), particularly, their followership. In general, followers commonly tend to believe that influencers are trustworthy sources who express their honest recommendations. Accordingly, followers are also more likely to be persuaded by influencers than by other advertising channels. Our studies indicate that explicit sponsorship behavior is less favorably received by followers when compared to other disclosure strategies. Specifically, we found that genuine recommendations with impartiality disclosure generate the highest influencer trustworthiness among followers, followed by concealing disclosure. Additionally, no disclosure emerges as the strategy with the most positive impact on followers’ attitudes toward the brand and purchase intentions. While one might be tempted to interpret these findings as an indication that influencers should provide no disclosure or conceal their sponsorship, it is essential to consider that non-disclosure violates the current regulations and has potential ethical implications and the long-term effects on trust between influencers and their followers (Borchers & Enke 2022). After all, influencers and brands have a social responsibility towards their audience. In line with the concept of Corporate Digital Responsibility, which emphasizes transparency and ethical conduct in the digital realm (< blinded for review > 2021), it is crucial for brands to maintain transparency, honesty, and ethical standards in their digital marketing efforts. Consequently, building authentic relationships with influencers and openly disclosing any compensated partnerships is also the optimal strategy from the marketers’ perspective. Marketers engage in influencer marketing to leverage the strong bond between influencers and followers (Ki etal. 2020). Therefore, from a marketer’s perspective, the optimal disclosure strategy would be if influencers market products as genuine recommendations and use impartiality disclosure. Our analyses showed that for followers, impartiality disclosure would translate into higher levels of influencer trustworthiness and thus significantly improve followers’ attitudes toward the brand. Hence, from the marketers’ perspective, it is advisable that instead of paying influencers to produce brand-related content, marketers need to target the influencers themselves and genuinely convince them of the companies’ products and brands. If companies manage to do so, they ultimately increase the chances that influencers impartially recommend the products and brand without directly being paid. Theoretical implications From a theoretical standpoint, our work highlights the need for new theoretical frameworks to better understand the complex effects of disclosure strategies on key outcomes. Our results reveal a gap between theory and practice, where many hypotheses based on existing theories did not materialize as expected, suggesting previously overlooked factors. Such key factors influencing the effect of sponsorship disclosure might include influencers’ size (e.g., mega, macro, and nano) and type (e.g., lifestyle and beauty, fitness and health, and gaming), the product promoted, and audience status (follower versus non-follower). In our study with a micro-influencer in the fitness domain, we found no direct effect of sponsorship disclosure on influencer credibility, brand attitude or intention to purchase. Although this outcome contrasts our initial expectations in the first three hypotheses, it somewhat aligns with findings from Giuffredi-Kähr etal. (2022) who observed that larger influencers—mega and macro influencers—invoke higher Electronic Markets (2024) 34:60 Page 23 of 27 60 skepticism than smaller influencers—micro and nano influencers—when sponsorship is not disclosed. Smaller influencers are generally perceived as more authentic, rendering their content more credible (Giuffredi-Kähr etal. 2022) and muting the impact of sponsorship disclosures on consumer responses. Additionally, varying findings from other research studies suggest that also factors such as influencer type, product appeal, and influencer-product congruence matter. For example, in a study with sustainability influencers (“sinnfluencers”), Schorn etal. (2022) report that products with personal benefit appeals (highlighting enjoyment or aesthetics) enhance influencer credibility and positively impact consumer attitudes and purchase intentions more effectively than the ecological appeals of a product. Other studies with lifestyle influencers promoting a fictive toothpaste (Lee and Kim 2020) or fictitious influencers in food (Giuffredi-Kähr etal. 2022) or literature (De Cicco etal. 2021) report varying effects of sponsorship disclosure on influencer credibility or consumer perceptions. Another important factor that needs to be investigated is the audience status—i.e., followers versus non-followers. In our study, we did not find support for direct effects of sponsorship disclosure on the variables of interest for non-followers, but we did observe significant effects for followers. Although our data does not allow us to delve deeper in the exact mechanisms driving our observations, prior literature presents various potential reasons. One potential reason for the observed difference is that followers and non-followers exhibit varying levels of sensitivity to the influencer’s messaging. Followers are typically more invested in the influencer, having developed a sense of familiarity or even a “friendship” through continuous exposure to the influencer’s content (Boerman, 2020; Djafarova & Rushworth, 2017). This stronger psychological bond makes them more sensitive to sponsorship transparency or concealment, which can impact their perceptions of the influencer’s credibility, brand attitude, and purchase intention (Ki etal., 2020). In contrast, non-followers do not share this depth of relationship with the influencer and, therefore, may approach disclosure strategies with a more neutral or detached perspective. Without this personal connection, non-followers are less likely to view concealed disclosure as a breach of trust, which could explain the lack of significant effects for this group. The difference in emotional investment and expectations between these two audience types drives the divergent outcomes in how disclosure strategies influence influencer marketing campaigns. Thus, the effectiveness of a disclosure strategy may depend not only on the content itself but also on the relationship between the influencer and the audience. Our work offers a more nuanced understanding of how these two user groups—followers versus non-followers—respond to various disclosure strategies. It shows that the influence of disclosures on followers and non-followers differs, allowing us to draw distinct conclusions depending on the audience analyzed. While existing literature acknowledges the importance of this distinction (e.g., Breves etal. 2021; Lou 2022), our results emphasize that this differentiation must be central to theoretical discussions in influencer marketing. Altogether, the lack of support of our first three hypotheses reflects the complex interplay of factors such as influencer type, product congruence, and appeal type, which may not align as uniformly as predicted across contexts. Hence, our results emphasize the need for further exploration into how specific influencer, product, and other contextual variables mediate consumer responses in terms of influencer and brand-related outcomes. Limitations Despite the value of this article’s insights on the optimal disclosure from an influencers’ and marketers’ perspective, we must also note diverse limitations that apply. A potential limitation of the study is the use of a single influencer, as different influencers with varying follower bases, engagement strategies, and niche expertise may produce different outcomes, which may affect the generalizability of the results. Future research should explore the impact of sponsorship disclosure across a broader range of influencers, taking into account variations in follower demographics, engagement strategies, and niche markets, to better understand how these factors influence real-world consumer behavior. Another limitation of our work is the repeated exposure of participants to the same post throughout all studies. While this approach aimed to minimize confounding factors and focus on the effects of interest, it introduces the possibility of order effects. Participants may have been influenced by their prior exposure to the post, leading to potential biases in their responses. Although we made efforts to address this limitation by spacing out the posts every 2 weeks, our study design cannot eliminate the potential for order effects entirely. For example, individuals who liked the first post might be less inclined to react to the same post again, potentially impacting their engagement and evaluations. In contrast to the potential order effects arising from participants’ repeated exposure to the same post, we hoped that using the same post would ensure consistency across studies. We recognize the danger that changes in an advertising’s appearance, wording, or framing can lead to different outcomes. Nevertheless, we urge future research to examine the effects of different disclosure strategies by using different posts across experiment sessions. Furthermore, we note two limitations inherent to Study 2. Firstly, since Instagram does not permit external randomized experiments without partnering directly with the platform, Study 2 required participants to follow a link from Instagram to an external website. To reduce biases and frictions from leaving Instagram, the website was designed to resemble a blog, closely mirroring Instagram’s aesthetic, aiming to Electronic Markets (2024) 34:6060 Page 24 of 27 create a seamless transition and make the experience feel familiar for participants. However, despite our best efforts to minimize friction, we acknowledge that asking participants to leave Instagram and access an external website may have influenced the data collection through selection bias or slight changes in user experience and engagement. Secondly, in Study 2 we cannot distinguish between followers and nonfollowers. However, given that followers and non-followers react differently to influencers’ content, investigating differences in sentiment between followers and non-followers might be a promising path forward. This is especially the case since research on the effect of disclosure strategies and user sentiment in real-world settings is relatively scarce. In relation to Studies 1 and 2, we note that we could not apply text analysis to extract consumers’ sentiments and had to rely on visual cues. Although prior literature suggests that the used emoji scale is appropriate to capture the sentiment and our results corroborate results from prior literature, future research should try to investigate the link between disclosure and consumer sentiment by relying on both textual and visual cues. Ultimately, a notable limitation of our third study is that the sample consisted solely of German participants. Recognizing that cultural differences may influence how individuals respond to various disclosure strategies, it is important for future research to extend this work by replicating our study across diverse cultural backgrounds. This broader approach will help determine the generalizability of our findings and potentially reveal nuanced insights into the effects of disclosure strategies in different cultural contexts. Concluding remarks The insights presented in this article have broader implications for influencers, policy-makers, marketers, and brands. For influencers, which are key intermediaries between consumers and brands in digital markets, influencers play a critical role in shaping consumer behavior, our results suggest that each influencer should get their priorities and strategic goals right. They need to balance maintaining trust with their followers while adhering to disclosure regulations. The challenge lies in aligning their marketing strategies with transparency requirements to sustain long-term credibility and engagement. Since our data indicates that content—i.e., whether influencers issue many genuine recommendations versus sponsored content—but also followership status matters, the optimal disclosure strategy, in the end, depends on whether they seek to improve their credibility in front of their followers or non-followers. For both marketers and influencers who strongly focus on their followership, genuine recommendations with impartiality disclosure are the best way of marketing. Thus, influencers should consider issuing more genuine recommendations than sponsored content. However, a shift from sponsored to organic content might require influencers to find new ways and business models to generate income that can at least partly replace paid collaborations with brands. For marketers and brands, who work with influencers to reach target audiences in an authentic way, our findings suggest that marketers should focus on fostering genuine relationships with influencers, encouraging impartiality disclosures to build trust. Although social media platforms are responsible for enforcing disclosure rules and maintaining a marketplace where users can confidently engage with, it is also marketers’ and brands’ responsibility to ensure their campaigns remain compliant with the current regulations. From a policy perspective, the insights presented in this article reinforce efforts toward more transparency of sponsored content in influencer marketing. Our findings point to the need for clearer and more enforceable guidelines to ensure that sponsorships are transparently communicated, preventing potential harm to consumer trust and platform credibility. If more precise regulation is not feasible, and monitoring compliance with the rules can be difficult due to the ever-growing number of influencers, transparency registers in which companies must list their influencer collaborations help retrace the blurred lines between organic and sponsored content. In the end, while stricter regulations on sponsorship disclosure may seem challenging for some, transparency of influencer marketing is critical to maintaining trust and credibility in digital markets. Funding Open Access funding enabled and organized by Projekt DEAL. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. References Abendroth, L. J., & Heyman, J. E. (2013). Honesty is the best policy: The effects of disclosure in word-of-mouth marketing. Journal of Marketing Communications, 19(4), 245–257. Agung, N. F. A., & Darma, G. S. (2019). Opportunities and challenges of instagram algorithm in improving competitive advantage. International Journal of Innovative Science and Research Technology, 4(1), 743–747. Ángeles Oviedo-García, M., Muñoz-Expósito, M., CastellanosVerdugo, M., & Sancho-Mejías, M. (2014). Metric proposal for customer engagement in Facebook. Journal of Research in Interactive Marketing, 8(4), 327–344. https:// doi. org/ 10. 1108/ JRIM0520140028