“I like, comment and share social media fashion innovators' content”: investigating millennial consumers' engagement with native advertising of Facebook brand pages

Fashion innovators (FIs) or social media fashion influencers' (SMFIs) are the market mavens of self-producing content (Schouten et al., 2020) on social networking sites (SNSs) who have easily reached their large follower base (Martínez-López et al., 2020; Lou and Yuan, 2019). Along with the South Asian countries, social media influencers (SMIs) are influencing the purchasing decisions of a large number of followers, notably in Bangladesh (Ghosh and Islam, 2025). Among the 3.07 billion monthly active users worldwide, Bangladesh accounted for 37.8% of Facebook users aged 25 to 34 up to February 2025 (Napoleoncat.com, 2025), largely dominated by 98.5% of smartphone users (Ghosh and Islam, 2025). Exclusively, SMFIs are cultivating their bonds with millennial consumers based on content generation as the latest blogs, videos, posts and texts (Giertz, 2022; Lou and Yuan, 2019). Millennials, or Generation Y, born between 1981 and 2000, known as digital natives (Ghosh and Islam, 2025), aged 24 to 44, perceive peer-originated brand content on Facebook brand pages (FBPs) as explicitly credible (Ghosh, 2022). In influencer marketing, native advertising (NAD) is the practice of blending SMIs' sponsored content with non-sponsored content (Kay et al., 2020) to fit seamlessly without disrupting viewers' experience (Kim and Kim, 2021a). When interacting with NAD, millennials appreciate watching compelling brand content as an ad (Kim and Kim, 2021a) if the sponsorship disclosure is clearly labeled (Breves et al., 2021). FBP followers assess the entertainment and informational value of online content through SMFI's interaction level, which activates cognitive, affective and behavioral engagement (Munnukka et al., 2019; Tsai and Men, 2013). In Bangladesh, the expansion of Facebook-based fashion houses specializing in boutiques, lifestyle products, traditional outfits and fusion western clothing is exclusively shaping female millennials' purchase behavior following over 5,00,000 FBPs on fashion houses (Babu, 2024). These issues have driven the effects of SMFIs’ endorsement as an integral component in recent global promotional strategies to persuade fashion-conscious followers, especially for Bangladeshi female millennial FBP followers' engagement. This research adopts FIs of Cabeza-Ramírez et al. (2022, p. 1434) and SMFIs (Quelhas-Brito et al., 2020) regarding micro-fashion SMIs, whose acceptable follower size is between 100,000 and 1,000,000 followers (Kay et al., 2020), to determine the drivers of millennial consumer engagement (consumers' engagement (CE)) in FBP-based NAD.

Hence, Bangladeshi millennials' niche brand choices seamlessly integrated with the FBP-based SMFIs' endorsed content through the latest organic videos, shorts, reels, stories and photos (Ghosh and Islam, 2025), significantly improving a local plus global fashion brand's reach and engagement in NAD. Subsequently, fashion retailers are keen to understand how such large followers of FBP value Bangladeshi SMFIs' content's entertainment and information value, interaction level, positively activating users' engagement (Munnukka et al., 2019), besides the online content characteristics. Therefore, this study firstly attempts to explore the local fashion clothing industry by addressing two research objectives: which content characteristics of micro SMFIs on FBP-based NAD, along with the contents' perceived credibility (PC), impact exclusively on either Bangladeshi male or female millennial CE and bringing new insights for South Asian emerging countries’ fashion marketers.

Consequently, this research is necessary for three major reasons: first, FI marketing grew from 3.7 billion in 2022 to 5.03 billion in 2023 in terms of market size (Skyquest.com, 2024). Meanwhile, given the abundant potentiality that the antecedents of SMFIs' FBP-based NAD content reinforce gender-specific millennial CE, this has not yet been explored to benefit industry operators. Second, few studies have identified PC as a mediating construct in NAD (Reinikainen et al., 2020; Munnukka et al., 2019), but not from a Facebook-based NAD micro SMFIs' perspective in emerging countries like Bangladesh (see Table 1). Third, although an important topic with valuable theoretical and practical implications, a few studies have acknowledged SMI disclosure statements (DS) effects on native ads globally (Kim and Kim, 2021a; Kay et al., 2020; Dhanesh and Duthler, 2019), largely in developed countries (Breves et al., 2021; Kim and Kim, 2021a; Boerman and Reijmersdal, 2020) but have overlooked how this may affect CE with the significant antecedents of SMFIs' FBP-based native ad content academically.

Accordingly, grounded within the SMI marketing literature, the originality of the research lies in proposing a conceptual framework as “FIIMCEB”, which is unexplored by earlier academics. This study pioneers and extends empirical understanding of the disclosure statements from the persuasion knowledge model (PKM) to Friestad and Wright (1994), parasocial interaction (PSI) theory by Horton and Wohl (1956), entertainment (entertainment value (ENT)) and informativeness (INF) value from the advertising value model by Ducoffe (1996), PC from source credibility by Ohanian (1990) and consumer engagement (CE) by Hollebeek et al. (2014) and Chen et al. (2021). Thus, this study investigates the content characteristics of SMFIs endorsed for Indigenous fashion clothing FBPs, especially benefiting the fashion marketers seeking modern fashion retailing to engage fashion-conscious millennials (see Figure 1).

NAD emerged in 2011, allowing prescribed format and location (Kim and Kim, 2021a), similar to SNSs' nonpaid content on news feeds, interchanged with sponsored content (Boerman et al., 2018), best leveraging an SMI's content credibility (Hayes et al., 2019), activating responses as likes and shares on FBPs (Kim and Kim, 2021a). CE is the combination of cognitive, affective and behavioral endeavors proposed by Hollebeek et al. (2014) (Giertz, 2022; Hayes et al., 2019) regarding the SMIs' performance beyond purchase (Ghosh, 2022), attachment sentiment towards a focal entity (Chen et al., 2021). The engagement criteria included liking, commenting and sharing micro SMIs' sponsored content (Wies et al., 2023; El-Naga et al., 2022; Reinikainen et al., 2020; Lou et al., 2019; Tsai and Men, 2013). Consumer brand engagement should be differently considered from an online perspective, considering the consumer's sensory connection (cognitive), expression (conation or affection) and behavioral activity to brand interaction within a community setting (Chen et al., 2021, p. 4; Hollebeek et al., 2014), which has been classified into three types: consumption, contribution and creation (Chen et al., 2021, p. 5).

SMFIs, as content generators (Lou et al., 2019; Lou and Yuan, 2019), are a quick solution to mitigate ad fatigue in native ads (Breves et al., 2021). The Federal Trade Commission (FTC) has prescribed disclosures to reduce distrustful confusion (Hayes et al., 2019; Evans et al., 2017; FTC, 2015), affected by consumers' advertising recognition as the transparency tool or ”sponsorship disclosure” of paid partnerships (“#brand name” or “#ad”) initiated by the advertising regulators (Boerman et al., 2018, p. 3, 2012).

For fashion marketers' consumer long-term engagement (Ao et al., 2023; Reinikainen et al., 2020), brand alignment is ensured using location-based sponsored content (Hayes et al., 2019), DS with niche consumers' preferences and micro SMIs' brand advocacy (Conde and Casais, 2023; Schouten et al., 2020). Academics have affirmed that by using Facebook features (Ghosh and Islam, 2025), SMIs direct consumers' level of interaction toward sponsored content engagement (Wies et al., 2023).

Evaluating these contents' PC, consisting of trustworthiness, expertise and attractiveness (Lou and Yuan, 2019; Schouten et al., 2020; Evans et al., 2017; Ohanian, 1990), young consumers recognize this characteristic as SMFIs' content trustworthiness (Cabeza-Ramírez et al., 2022) in triggering cognitive and emotional engagement levels (Martínez-López et al., 2020) for future brand adoption identified as informativeness, entertainment and an interactive way (Munnukka et al., 2019). Although prior studies focused solely on female consumers (Reinikainen et al., 2020; Djafarova and Rushworth, 2017), this study offers a broader perspective by including all genders. Hence, SMFIs fail to establish closeness with millennial users; product placement with DS remains ineffective (Ghosh, 2022) because peer approval regarding fashion brands highly engages millennials' persuasive power and product adoption (Ao et al., 2023) compared to celebrities (Belanche et al., 2021; Schouten et al., 2020) and mega- and macro-SMI endorsement (Conde and Casais, 2023).

A DS is a clear representation of the nature of the post sponsorship, earlier researched as disclosure of sponsorship in the PKM (Kay et al., 2020; FTC, 2015). Sponsorship labels recommended by the FTC, such as “sponsored” and “paid ad,” as the DS intensify consumers' persuasion efforts (Ghosh and Islam, 2023; FTC, 2015) in recognizing the SNSs' editorial content (Evans et al., 2017) and “coping mechanisms” (Friestad and Wright, 1994, p. 3) comprehending SMI's posts as an advertisement (Lee and Kim, 2020; Boerman et al., 2017; Boerman et al., 2012). Additionally, SMI's ad content value stems from its followers' voluntary consumption of entertaining and informative content (Herrando and Martín-De Hoyos, 2021; Lou and Yuan, 2019), leading to faster brand content acceptance in NAD (Kim et al., 2019). Young followers' cognitive and emotional attachment is influenced through SMI-generated valuable brand information (El-Naga et al., 2022; Sánchez-Fernández and Jiménez-Castillo, 2021), which reduces negative feelings toward ads (Lou and Yuan, 2019). Furthermore, PSI differs from social interaction (Breves et al., 2021; Horton and Wohl, 1956). Initially, it was proposed by Horton and Wohl (1956) to measure television audiences' one-sided imaginary closeness with their desirable media personalities (Chen et al., 2021). It is interchanged with parasocial relationship between the followers and SMIs (Reinikainen et al., 2020; Tsai and Men, 2013), positively affecting consumers' frequent engagement and perceived feelings of SMI content and reducing ad literacy (Boerman and Reijmersdal, 2020).

The number of likes in endorsed content is evaluated by higher ad-skeptical consumers (Ghosh and Islam, 2023); peer comments reduce the associated risks but increase users' persuasion knowledge (Chen et al., 2021; Reinikainen et al., 2020), impacting CE. In SMI's video content (Sánchez-Fernández and Jiménez-Castillo, 2021), this increases millennial CE through likes, comments and shares (Hughes et al., 2019) and content consumption (Chen et al., 2021), expressed as the “sense of belongingness” in PSI (Tsai and Men, 2013, p. 79), induced by Facebook bloggers' expertise and discounts (Hughes et al., 2019). Also, necessary brand information reduces searching time, easing problem-solving ideas (Tolunay and Ekizler, 2021) and intended decisions increase intention to click on the ads (Kim et al., 2019). Hence, micro SMIs are forming a higher level of PSI with a stronger effect on credibility perception (Conde and Casais, 2023), mainly proven with female followers (Reinikainen et al., 2020). Consequently, the number of occurrences and time on each interaction (Sánchez-Fernández and Jiménez-Castillo, 2021), comments and personal messages (Reinikainen et al., 2020) impact the interaction to the engagement expressed in the form of likes, comments and shares (Schouten et. al., 2020), parallel with the findings of El-Naga et al. (2022) of YouTube influencers, shaping PSIs (Chung and Cho, 2017). Based on this evidence, the study hypothesizes:

On Facebook-based NAD, vaguely presented website-based native ad information develops ad skepticism as per PKM (Lee and Kim, 2020); hence, a highly credible SMI effectively stimulates followers' PC of the unfamiliar brand content as a “media vehicle” diminishes consumers' ad deception (Lee and Kim, 2020, p. 4). Consumers' ad recognition increases coping strategies (Boerman et al., 2012) by directing ad disbelief (Kim and Kim, 2021a; Lee and Kim, 2020) of DS (Boerman and Reijmersdal, 2020), but Hayes et al. (2019) have confirmed that SNS ad credibility can reduce this. Likewise, SMFIs' INF clarity, they are perceived as entertaining content generators and “trusted tastemakers” shaping followers' brand behavior (Breves et al., 2021, p. 1), interacting credibly compared to non-followers' product-endorser fit (Schouten et al., 2020). Besides, Ohanian's (1990) source credibility model works best to meaningfully influence followers' avoidance of ad information regarding SMI's interactions as friends (Chung and Cho, 2017).

SMFI content on FBP-based NADs can closely connect followers' emotions (Hayes et al., 2019) compared to Instagram and pseudo-interactions in Twitter engagement (Kim and Kim, 2021a), although higher website credibility reduces Instagram users' deceptive product review perception in NAD (Lee and Kim, 2020; Boerman et al., 2017). Higher PSI determines SMIs to be perceived as credible content creators, driving FBP's positive brand credibility (Ghosh and Islam, 2023), than the skeptical non-followers (Breves et al., 2021; Belanche et al., 2021). If the followers have presumed SMIs' posts as exaggerated content, the communal interaction worsens (Dhanesh and Duthler, 2019; Evans et al., 2017) and young users feel cheated (Mabkhot et al., 2022) compared to trusted brand demonstration (Wies et al., 2023; Reinikainen et al., 2020). So, SMIs' PSI is harmed more than brands.

Hence, DS, ENT, INF and PSI are closely associated research areas in celebrity endorsement (Conde and Casais, 2023; Kay et al., 2020; Munnukka et al., 2019), but a lack of focus is given to SMFIs' content characteristics validation in PC on FBPs. Moreover, how NADs affect consumer behavior in diversified media environments is a matter to be researched intensely (Hayes et al., 2019), which has been academically researched in developed countries (Boerman et al., 2012). Therefore, the study hypothesizes:

SMIs' PC leveraged with sponsored content (Belanche et al., 2021; Martínez-López et al., 2020) is undertaken as trustworthiness (Lou and Yuan, 2019), exhibiting brand expertise in problem minimization in an NAD context (Hayes et al., 2019). Moreover, the follower's future persuasion process in rational and affective engagement (Martínez-López et al., 2020) generates a high confidence level (Lou and Yuan, 2019) with DSs' positive effect (Weismueller et al., 2021) mainly through FBP followers' voluntary consumption of INF and ENT content (El-Naga et al., 2022; Herrando and Martín-De Hoyos, 2021) in an NAD context. Based on the mediating effect of credible characteristics (Lee and Kim, 2020), to engagement, SMIs are the relationship mavens in endorsing multi-layered relationship marketing (Schouten et al., 2020). A similar finding is justified by Behl et al. (2024) in the data-driven innovation study. PSI reinforcing followers' comments identified as the “internal identification” of CE perceived as unbiased on the SMI-sponsored post (Chen et al., 2021, p. 9; Reinikainen et al., 2020). Thus, it is hypothesized,

This quantitative-descriptive research aims to test existing theories and models previously revealed by academics rather than constructing a theory. The data gathered from the estimated 389 samples are above 200 through applying 24 reflective scale items, indicating that covariance-based structural equation modeling (SEM) is more suitable than partial least squares SEM (Hair et al., 2014; Nunnally and Bernstein, 1995; Fornell and Larcker, 1981). Moreover, the association between dependent and independent variables of the present study's conceptual framework and multivariate techniques can be better examined by covariance-based structural equation modeling (CB-SEM) (Nunnally and Bernstein, 1995; Fornell and Larcker, 1981) through AMOS 23.

This study was conducted in Bangladesh through an online survey, reaching 37.8% of the entire population (Napoleon.com, 2024), earning around $97 million globally (Ghosh and Islam, 2025) with advertisements reaching 68.4% of the highest local Facebook user base aged 25 to 34 following over 500,000 FBPs (Babu, 2024). To generate awareness of fashion clothing, local fashion designers utilize Facebook as the most popular fast-growing platform, with users 67,179 200 up to February 2025 (Napoleoncat.com, 2025) for online fashion stores 9.7 million ad reach up to January 2024, targeting 35,100,000 millennials (Ghosh and Islam, 2025).

As of December 2025, 173.17 million Bangladeshi millennial Facebook users aged 24 to 44 following FBPs of fashion products endorsing micro SMIs accessed through mobile devices and personal computers (www.businesspostbd.com., 2026) have been specified to meet research objectives (Hair et al., 2014). This research applies a purposive sampling rather than a convenience sampling technique because it can't analyze the entire millennial population's everyday engagement patterns of FBPs in Bangladesh (Malhotra and Dash, 2016) and the participants chosen as the “informants” other than the convenience (Malhotra and Dash, 2016) to estimate complex predictors (Kim et al., 2019) of SMFIs' FBP-based content in native ads (Ghosh and Islam, 2023).

The data was collected from January 2025 to April 2025. The researchers distributed Google Form survey links through Facebook Messenger, WhatsApp and Gmail to individual eligible respondents, preserving confidentiality and obtaining their consent in the online structured questionnaire developed in English. The questionnaire was sent to the 389 selected respondents living in Bangladesh, with a total of 348 responses gathered with an 89.46% response rate. Completing the questionnaire took about 10 min.

In the questionnaire design, a clear definition of NADs on Facebook, micro SMFIs (followers of 10,000 to 10,00,00) endorsing solely for individual FBPs and answering questions based on favorable SMFIs' sponsored content uploaded on fashion clothing FBPs seen on news feed labeled with sponsorship disclosures '(“#sponsored” “#paid ads” and “#in collaboration”) were included for respondents. The demographic variables for this study are categorized into gender, age, education and occupation, and the chosen scales are designed as five-point Likert scale questions (1 as “strongly disagree” to 5 as “strongly agree” range) to determine millennial CE with FBP-based NAD.

This study adapted validated multi-item scales from previous studies to measure the six constructs of the proposed conceptual model, using 24 items on a 5-point Likert scale (see Table 3). A pilot study of 29 respondents guided the researchers to conduct a preliminary screening of the 24 adapted items from 6 scales, each containing four items, adapted from Dhanesh and Duthler (2019) and Boerman et al. (2018). ENT and INF have been adapted from Ducoffe (1996) and Tolunay and Ekizler (2021). The scale PSI has been adapted from Horton and Wohl (1956), Giertz (2022) and Chung and Cho (2017). Next, the PC scale has been adapted from Ohanian (1990) and Reinikainen et al. (2020). Lastly, the CE scale has been adapted from Hollebeek et al. (2014) and Giertz (2022).

A final sample of 317 respondents was gathered after eliminating 31 questionnaires that were deemed unacceptable. Table 2 shows the respondents' demographic profile, confirming no selection bias, which resembles the E-commerce Association of Bangladesh (Babu, 2024). Respondents are mostly women (76.03%) aged 25 to 29 (62.46%), 51.11% completed graduation and post-graduation having private services (38.1%), justifying that Bangladeshi SMFIs are majorly engaging around 36% female Facebook users (24,24,512 in number) (https://napoleoncat.com, 2024), in the NAD context.

To determine the applicability of CB-SEM to the reliability and validity of the measurement model, the normality test is conducted via SPSS 26. Applying Mahalanobis outliers (Pallant, 2016), 31 responses were excluded from a total of 348 obtained responses; exploratory factor analysis and Kaiser–Meyer–Olkin analysis resulted in acceptable values in the remaining 317 responses in Table 3 (Pallant, 2016). This research initially conducted confirmatory factor analysis (CFA) on AMOS 23, applying the convergent validity following the discriminant validity justified by Straub (1989). In Table 3, the final 20 items are retained from the 24 items of 6 constructs, affirming the reliability threshold values reflected in Cronbach's alpha being above the threshold value of 0.70 (Hair et al., 2014), and 4 items are excluded due to scale purification through CFA (Hair et al., 2014). Moreover, loadings more than 0.5 or above and maximum shared variance less than average variance extracted, confirming the strength of the internal consistency, reflecting undoubted discriminant validity (Nunnally and Bernstein, 1995; Fornell and Larcker, 1981).

The value of each construct resulted in a higher correlation value, as shown in Table 4, and the heterotrait-monotrait ratio of correlations (HTMT) in Table 5 recommended that all the criteria are less than 0.9 (Henseler et al., 2015), confirming the established discriminant validity. Considering that the maximum likelihood estimation method is best suited to SEM, the model fit is confirmed by CFA (Hair et al., 2014).

Next, the hypotheses are examined using path analysis in SEM. The bootstrap method is applied to mediation analysis to examine the indirect and direct effects of SMFIs' sponsored content characteristics on millennial CE via PC. Table 6 represents the good fit of both the direct and indirect effects of the structural model of the path coefficient (Byrne, 2013), where χ2/df = 1.261, comparative fit index = 0.989, root mean square error of approximation = 0.030, Tucker–Lewis index = 0.987 and goodness-of-fit index = 0.961 (Hair et al., 2014).

The standardized regression weight estimation in Table 7 and Figure 2 indicates that DS has both an indirect relationship with PC (β = 0.495; p < 0.001) and a direct relationship with CE (β = 0.470; p < 0.001). Thus, H1 and H2 are accepted. This supports the earlier discussed literature in which skepticism is reduced through SMI sponsorship disclosure (Ghosh and Islam, 2023; Dhanesh and Duthler, 2019; Lou et al., 2019) but contrasts with Herrando and Martín-De Hoyos (2021). Surprisingly, the ENT is in a direct relationship with CE (β = 0.396; p < 0.001) but has a non-significant relationship to PC (β = 0.227; p > 0.001), not supporting H3, but supporting H4, indicating that SMFIs ENT characteristic in the native ad need not always be negative to CE justified by previous academics (Tsai and Men, 2013). INF, the indirect relationship to PC, is significantly positive (β = 0.328; p < 0.001), supporting H5, whereas non-significantly affect CE (β = 0.174; p > 0.001), not supporting H6. Although the positive effect is consistent with previous studies (Lou and Yuan, 2019), the negative impact aligns with the findings of Ghosh (2022) and Tsai and Men (2013), unlike Ducoffes' (1996) findings. Furthermore, PSI to PC has both an indirect relationship (β = 0.579; p < 0.001) and a direct relationship with CE (β = 0.503; p < 0.001). Thus, H7 and H8 are accepted, contributing to CE-based (Ghosh and Islam, 2025; Lou et al., 2019; Tsai and Men, 2013), earlier NAD studies (Munnukka et al., 2019; Hayes et al., 2019). In contrast, this has a negative correlation for Romanian Instagram SMIs' content in young followers' product involvement (Balaban et al., 2022). To conclude, PC is positively associated with CE in the NAD (β = 0.292; p < 0.001), supporting Hypothesis 9, as a mediation relationship on purchase intention (Mabkhot et al., 2022; Belanche et al., 2021) and on brand attitude (Munnukka et al., 2019). However, it is dissimilar to the negative result of FBPs' PC and CE in a developed country (Tsai and Men, 2013).

This study's findings extend the theoretical scope implied within SNS-based fashion clothing NAD: Firstly, regarding the SMFIs' sponsored content characteristics sourced from FBPs in the NAD settings seen on the news feeds through “FIIMCEB,” contributing to the applicability of the CE theory of Hollebeek et al. (2014) and Chen et al. (2021), where research is unexplored. To date, only some studies have revealed how various spokespersons produce diversified marketing effects regarding SMIs' followers' persuasion knowledge on Instagram (Ghosh and Islam, 2023; Lou et al., 2019) and little research on SNS-based NAD (Kim and Kim, 2021a; Kim et al., 2019; Hughes et al., 2019; Boerman et al., 2017).

Second, the originality of this research exhibits a novel mediating relationship between the PC of SMFIs' content characteristics and millennials’ CE, where sufficient research has resulted in a negative effect of SMI disclosure (Evans et al., 2017; Boerman et al., 2012); this study augments existing SNS literature, ensuring DS need not always be negative (Ghosh and Islam, 2023; Breves et al., 2021; Hayes et al., 2019; Kay et al., 2020), in the FBP-based NAD perspective.

Third, while DS is included, PSI as SMFIs' content characteristics resulted as the leading characteristics, enhancing the findings of Ghosh and Islam (2025), Munnukka et al. (2019) and Reinikainen et al. (2020), but challenging the prevailing literature on SNSs' celebrities (Kay et al., 2020; Schouten et al., 2020).

Fourth, the ENT of the content negatively affects SMFIs' PC in millennial consumers' doubtful perceptions regarding NADs' credibility. The negative result of INF on CE is that SMFI's videos and other visual content often appear superficial, similar to organic-like NADs, if unclear disclosure labels do not drive credibility perception, which is academically uncommon.

Fifth, past studies have mainly focused on female consumers (Reinikainen et al., 2020), but this study includes all genders and makes a broad contribution to the native ad literature by specifying sponsored content.

By applying “FIIMCEB”, FBP-clothing retailers can integrate group analyses using demographic variables through segment-detailed graphically enriched aesthetic content strategies instead of a “one-size-fits-all” technique for young millennials. Whereas older millennial demographics should be incorporated by leveraging SMFI credibility-heightened content adaptation, tailoring gender-receptive strategic customization can blend female millennials' engagement with emotionally provoking content and functional features for males, thereby increasing the effectiveness of fashion retailers' NAD through group resonance. Furthermore, economic realities should be differently considered for diverse income millennials, for example: value-based positioning cues sourced from credible SMFIs targeting higher-income groups, whereas engaging NAD content from friend-like SMFIs for the lower-income groups ensure highest effectiveness. FBP managers can upgrade their own brands' engagement metrics by clarifying disclosures and implementing data-driven multi-group analysis tools, while aligning SMFIs with millennials' congruence.

As this study is based on millennials' self-experience with SMFIs' sponsored content characteristics exposed to biased response, which can differ considerably, future academics should conduct cross-cultural and qualitative experimental research in the SMI domain, with no disclosure. Focusing only on a few millennial respondents on FBPs confines generalizability; future academics could apply longitudinal research investigating larger samples of usage behavior in SMI categories on Instagram and TikTok to track behavioral change for improved theoretical relevance.

Consistent with the earlier studies, SMFIs' content characteristics in NAD proved to have a meaningful effect on millennial CE. Contributing to the existing NAD literature, this research provides a direction for the developing countries' enthusiastic fashion followers, SMIs and retailers that these influential attributes must be considered critically with PC in terms of SMIs' endorsement in NAD.