Abstract
Brands are increasingly required to be ‘woke’ and communicate their stance on various divisive sociopolitical issues and to do so particularly on social media platforms. However, research shedding light on the outcome of woke brand communication is in short supply; it does not compare the suasory effects of the latter with those achieved by traditional persuasive appeals; and it provides scant guidance on which brands ought to adopt this strategy. Combining language expectancy theory, the brands-as-intentional-agents framework, and the literature on consumer engagement in social media, this paper aims to fill these gaps by means of a multi-industry, text-mining-based study which investigates both the volume and the semantic virality patterns of traditional vs. woke persuasive appeals adopted by brands on social media platforms. The findings suggest that woke communication generates higher levels of consumer engagement than do traditional persuasive appeals; in particular, woke communication is more effective for warm brands. Moreover, when competent brands undertake woke campaigns, they tend to trigger more polarized reactions in consumers’ comments than warm brands do.
Similar content being viewed by others
Explore related subjects
Discover the latest articles, news and stories from top researchers in related subjects.Avoid common mistakes on your manuscript.
Introduction
The established body of corporate social responsibility (CSR) literature has provided strong evidence that brands which take and communicate a stance on issues that extend beyond their profit-making scope outperform competing brands that prefer to remain neutral (Du et al. 2010; Magee 2022; Saxton et al. 2019; Weinzimmer and Esken 2016). However, while traditional CSR issues, like social, economic, and environmental sustainability, have acquired a status of generalized acceptance, other issues on which brands have recently been increasingly required to take a stance are instead perceived as divisive and controversial (Schmidt et al. 2021). These include, to name but a few, the defense of LGBTQIA + rights, race non-discrimination, the right to abortion, and active support for people during major crises like the recent pandemic (Feng et al. 2021; Jungblut and Johnen 2021; Middleton and Turnbull 2021; Mirzaei et al. 2022; Schmidt et al. 2021; Sobande 2019; 2020; Vredenburg et al. 2020). The distinctive form of communication where brands publicize their direct support for these causes is called “woke communication” (Feng et al. 2021; Middleton and Turnbull 2021; Mirzaei et al. 2022), and it is framed as a form of brand activism (Moorman 2020; Pimentel et al. 2023; Swaminathan et al. 2020; Vrendeburg et al. 2021). A term of Afro-American origin, “woke” denotes the brand’s active effort to increase awareness about and encourage sociopolitical change toward, socially significant issues (Mirzaei et al. 2022; Sobande 2020). There are now countless examples of brands that have embraced woke communication. To mention only a few popular cases, Dove’s globally famous “Real Beauty” campaign was released to assert the brand’s rejection of the unreal aesthetic served up by most brands operating in the beauty care industry. Nike has taken a strong stance to support black people’s rights and Colin Kaepernick’s fight against racism with a commercial bearing the tagline “Believe in something, even if it means sacrificing everything”. More recently, the brewery brand BrewDog has launched on its Twitter account a divisive anti-sponsorship campaign for the 2022 Football World Cup to condemn the human rights abuses taking place in the hosting country, Qatar, before and during the event. However, because the sociopolitical initiatives and activities promoted by a woke campaign are closely connected to a brand’s higher purpose and values, are not necessarily tied to the brand’s core-business, do not target a broad and inclusive audience, and, above all, revolve around an intrinsically controversial issue, they differ from other forms of brand activism that have been previously investigated, like cause-related marketing (CRM), corporate social initiatives and corporate social and political advocacy (Austin and Geither 2016; Bhagwat et al. 2020; Hydock et al. 2019; 2020; Jungblut and Johnen 2021; Mirzaei et al. 2022; Weber et al. 2023). Yet, although the number of brands embracing woke communication is booming (Guha and Korschun 2023), research shedding light on the outcomes of this form of communication is only emergent (Feng et al. 2021; Mirzaei et al. 2022; Vredeburg et al. 2020), and it has often yielded conflicting results (Hydock et al. 2019; Weinzimmer and Esken 2016). While some argue that brands engaging in woke communication enjoy greater consumer support than their counterparts (Austin et al. 2019; Bravo and Lee 2020; Li et al. 2022; Schmidt et al. 2021), others posit that woke communication is risky for brands because it exerts an overall negative effect on stakeholders’ attitudes and behaviors (Abitbol et al. Seltzer 2018; Mukherejee and Althuizen 2020; Wang et al. 2022), brand image (Jungblut and Johnen 2021), brand perceptions (Klostermann et al. 2021) and stock market performances (Bhagwat et al. 2020). A close look at the extant literature on woke communication shows three major shortcomings (See Appendix A for an overview of this stream of literature.). First, no study to date has compared the persuasive effects of woke communication with those prompted by other more traditional forms of brand communication. Recent research has investigated which consumers’ reactions are more likely to be elicited by woke communication (Feng et al. 2021; Yang et al. 2021; Mukherjee and Althuizen 2020; Wang et al. 2022; Weber et al. 2023), and how woke messages should be framed to persuade their recipients (Ahmad et al. 2023; Di Russo et al. 2020; Song and Choi 2023; Zhou et al. 2023), but it has not provided any empirical findings on whether woke communication is more, less, or equally persuasive than other, more orthodox forms of brand communication (Milfeld and Flint 2020). Moreover, beyond inquiry into whether engaging in woke communication is beneficial or harmful for a brand (Mukherejee and Althuizen 2020; Wang et al. 2022), research is needed to understand for what type of brands this communication strategy is most effective (Weber et al. 2023). Second, despite recent calls to investigate the myriad of social issues that woke communication can advocate (e.g., Feng et al. 2021), current research is skewed toward a narrow set of popular and blatantly partisan campaigns, like “femvertising” and pro-black people’s rights campaigns (Champlin et al. 2019; Mirzaei et al. 2022), while other socially important causes that may give ground to woke communication rarely come under the spotlight in the literature (Feng et al. 2021). Third, since social media (SM) platforms provide built-in response options through which consumers can interact with brand-generated contents in real time (Kabadayi and Price 2014), studies to date have evaluated consumers’ online reactions to woke communication mostly via structural volume-based metrics, such as the cumulated number of likes, views, or followers (Schaefers et al. 2021; Wang et al. 2022), while other ‘thicker’ and finer-grained metrics, such as the semantic content of consumers’ comments, have been neglected (Unnava and Aravindakshan 2021; Zhou et al. 2023). Indeed, the analysis of user-generated comments can enrich the understanding of consumer reactions by revealing the existence, and heterogeneity, of both positive and negative topics (Swaminathan et al. 2022; Tirunillai and Tellis 2014).
In light of these gaps, the aim of this study is to answer the following research questions: (1) Does woke communication affect consumer engagement on social media (henceforth “CESM”) differently from traditional persuasive appeals adopted by brands? (2) Does woke communication affect conversation topics in consumers’ SM comments differently from traditional persuasive appeals adopted by brands? (3) For what type of brand is woke communication on SM most effective?
To address these questions, we developed a theoretical model based on a combination of language expectancy theory (LET) (Burgoon 1993; Burgoon et al. 2002), the brands-as-intentional-agents framework (BIAF) (Kervyn et al. 2012), and the thriving literature on CESM (de Oliveira Santini et al. 2020). We tested it by means of a multi-industry, automated text analysis (Humphreys and Wang 2018) which investigated both the volume and semantic patterns of traditional and woke persuasive appeals adopted by brands on SM platforms (Mangiò et al. 2021; Suddaby and Greenwood 2005).
The rest of this paper is structured as follows: First, it reviews the LET, BIAF, and previous CESM literature, on which basis it develops the theoretical model and provides justification for the hypotheses tested. This is followed by a methodological section which provides information about the research context as well as the data collection and the analytical procedure pursued. Finally, the paper presents a general discussion of the results and concludes with implications, limitations, and suggestions for further research.
Theoretical background
Language expectancy theory
Language expectancy theory (LET) (Afifi and Metts 1998; Burgoon et al. 2002; Burgoon 1993) is a theory of persuasion which aims to explain the effects of expectancy confirmation or disconfirmation in interpersonal communication. According to LET, language is a rule-based system, i.e., one that is based on a set of anticipated linguistic norms called “expectancies”, socially constructed by individuals once they agree that there is an appropriate language to use in specific circumstances. Expectancies thus affect interpersonal interactions and determine whether the message will be accepted or rejected by its recipients (Burgoon 1993, p.32). Expectancies depend on three interactional and pre-interactional factors. The first factor is the communicator’s characteristics. This factor includes salient features of the communicator, including its typical communication style. In the brand communication context, the notion of communicator’s characteristics resonates with the actual persuasive appeal that the brand adopts on a specific communication occasion (Lee et al. 2018; Liadeli et al. 2022). The second factor is relationship. The relationship factor encompasses all characteristics that describe the relationship between the communicator, i.e., the brand in our case, and the recipients of communication, such as consumers or SM users. In the brand communication context, the relationship between brands and consumers is effectively captured by the notion of brand stereotypes (Cuddy et al. 2008) because it requires a form of anticipated relationship. The third factor is context characteristics, i.e., the contingent situational factors that characterize a communication context. In the SM realm, context characteristics refer to platform affordances whereby communication happening in this environment is different from communication taking place in different communication settings (Boyd 2010). Whenever a message is exchanged, recipients cling to prior expectancies to determine whether the message conforms to such expectancies (expectancy confirmation) or does not (expectancy violation). Then, when an expectancy violation occurs, an arousal change in the form of attentional allocation takes place in the message’s recipients, heightening their attention toward characteristics of the sender and their relational implications, so that they can make sense of such violation (Afifi and Metts 1998). In other words, following a two-stage process, recipients first pay attention to the message and its content, such as the persuasive appeal adopted; then, if a violation occurs, they shift attention to the transgressor, such as the brand, in order to make sense of the violation and attune future behaviors (Afifi and Metts1998; Rocklage and Fazio 2020; Yang et al. 2020).
Expectancy violations can be either positively or negatively valenced (Afifi and Metts 1998; Burgoon 1993). Positive violations occur when the persuasive attempt is made through a communication exchange that overcomes the recipients’ expectancies. For this reason, positive violations induce the formation of a more positive attitude toward the message and its sender, and they trigger more enthusiastic behavioral responses by the receiver (e.g., higher CESM). Negative violations, conversely, occur when recipients’ expectancies are not met. Negative violations thus lead to reduced attitude and trigger a behavioral response opposite to what was intended by the source (e.g., lower CESM) (Jensen et al. 2013; Kronrod et al. 2012a, b). Rooted in a social meaning representation of interpersonal communication, LET is particularly helpful in unpacking what happens when multiple non-consensual interpretations of a communication exchange are possible (Burgoon 1993), as in the case of brand communication on SM (Kronrod et al. 2012a, b; Yang et al. 2020). In order to determine the conversational norms that underpin brand communication on SM, this paper draws on the literature dealing with brand communication and CESM and on the brands-as-intentional-agents framework, as now shown.
Brand communication and consumer social media engagement
Given the increasing ubiquity and importance of SM platforms (Alalwan et al. 2017), research shedding light on the suasory effects of brand communication via SM has gained academic traction in the past decade (Brodie et al. 2013; Hollebeek et al. 2021; Kumar et al. 2016; Voorveld 2019). Even though the persuasiveness of brand communication via SM has been conceptualized and measured in various ways (de Vries et al. 2012; Lee et al. 2018; Villarroel Ordenes et al. 2019), research has recently converged on CESM as an effective and readily available operationalization (Ashley and Tuten 2015; de Oliveira-Santini et al. 2020; Shahbaznezhad et al. 2021; Swani and Labreque 2020). CESM is a multilevel and multidimensional phenomenon involving varying levels of users’ commitment to and interaction with brands and their activities on SM. CESM can result from the specific experiences that consumers undergo while being exposed to brand communication on SM (Beckers et al. 2018; Eigenraam et al. 2018; Voorveld 2019). Thus, CESM stands at the crossroads between users’ engagement behaviors (Gummerus et al. 2012; Van Doorn et al. 2010) and brands’ communication features, such as content and media type, posting frequency, and posting time (Barger et al. 2016; Deng et al. 2021; Dolan et al. 2019; McShane et al. 2021; Shahbaznezhad et al. 2021). This study draws on a conception of CESM as a “consumer’s behavioral manifestations that have a SM focus beyond purchase, resulting from motivational drivers” (van Doorn et al. 2010, p. 265). Starting from the assumption that consumers are cognitively, affectively, and conatively affected by brands’ communication (Barry and Howard 1990), CESM is conceptualized as a cumulative phenomenon occurring in the three stages of relationship formation (cognitive dimension), creation of engagement (affective dimension), and contribution (conative dimension) (de Oliveira-Santini et al. 2020; Dolan et al. 2019; Moran et al. 2020; Swani and Labreque 2020; Vlachvei et al. 2021). Brands wishing to foster CESM can choose among different persuasive appeals ranging from some that are more traditional, i.e., are more frequently used by brands, to others that are instead less widely used, as we now show.
Traditional persuasive appeals
The persuasive appeals that brands use in communication are known as ethos, pathos, and logos. Ethos signifies the means of convincing others by signaling the persuader’s credibility and trustworthiness. Pathos refers to ways of convincing others by creating an emotional response to an impassioned plea or a convincing story, through the use of emotional and/or entertaining content. Logos refers to ways of persuading people by appealing to their rationality using facts and figures, and by conveying information-based and/or remunerative contents (Lee et al. 2018; Mangiò et al. 2021; Panigyrakis et al. 2020; Suddaby and Greenwood 2005). According to previous research, these three persuasive appeals help brands to trigger CESM across the three aforementioned stages, i.e., relationship formation, creation of engagement, and contribution (Vlachvei et al. 2021) (Fig. 1).
There is indeed a great deal of evidence that SM are used by brands to disseminate information (a form of communication which implies the usage of the persuasive appeal of logos), to create, reinforce and maintain an emotion-based relationship with consumers (which is based on the persuasive appeal of pathos), and to create or reinforce an image of the brand as reliable, credible or trustworthy (through a persuasive appeal of ethos). Regarding the first way to leverage SM, research has shown that the persuasive appeal of logos is used by brands to assist consumers in collecting useful pre-purchase, purchase, and post-purchase information such as product availability, prices, discounts, and promotions (Alawan, 2018; Bowen et al. 2022; Dwivedi et al. 2015; Heinonen 2011; Moro et al. 2018; Westerman et al. 2014), and it has also shown that highly informative brand communication via SM is particularly effective in prompting positive consumers’ outcomes (Araujo et al. 2015; de Vries et al. 2012; Dolan et al. 2019; Eigernraam et al. 2020; Kim et al. 2015; Muntinga et al. 2011; Pletikosa Cvijikj and Michahelles 2013; Swani and Milne 2017). Regarding the use of the persuasive appeal of pathos in SM, research has shown that brands substantially imbue their communication messages with emotions (Liadeli et al. 2023; Tellis et al. 2019; Rocklage and Fazio 2020) as consumers are ever keener to accept that emotions are implied in their interactions with brands (Ashley and Tuten 2015; Swani and Milne 2017). As a matter of fact, previous research has shown that emotionally charged brand-generated posts more likely generate CESM compared to other posts that are not emotionally charged (Akpinar and Berger 2017; Berger and Milkman 2012; de Oliveira-Santini et al. 2020; Lee et al. 2018; Liadeli et al. 2023; Rietveld et al. 2020; Swani and Milne 2017; Tafesse and Wien 2018; Tellis et al. 2019).
Extensive research in SM also furnishes substantial evidence that credibility, i.e., the believability of information and of its source (Hovland et al. 1953), positively affects consumer-brand awareness, trust, attitude (Gvili and Levy 2018; Hung and Li 2007; Lou and Yuan 2019; Wang and Scheinbaum 2018), as well as engagement (Cao et al. 2021; Cosenza et al. 2015; Tsai et al. 2013). Source credibility, which is manifest in communication through the use of the persuasive appeal of ethos, is crucial on SM platforms, because the idiosyncratic functioning of their affordances makes it difficult for users to assess the veracity of the information shared therein (Di Domenico et al. 2021; Sundar 2008).
Instantiating and reflecting the dominant and time-persisting commercial logic of markets (Aboelenien and Nguyen 2023; Mangiò et al. 2021; York et al. 2018), brand communications on SM which appeal to consumers’ rationality (logos), which evoke consumers’ positive emotions (pathos), and which express the persuader’s credibility and trustworthiness (ethos) are widely used by brands on their platforms (Liadeli et al. 2022; Tellis et al. 2019). As a consequence, these persuasive appeals enjoy full legitimacy from consumers, to the extent that they are generally taken for granted compared to novel appeals (Lee and Mason 1999; Panigyrakis et al. 2020; Schneiberg and Clemens 2006). As such, in line with the principal tenets of LET, the traditional persuasive appeals of ethos, pathos, and logos represent a conversational norm on SM; consequently, they trigger a mechanism of expectancy confirmation among their recipients (Burgoon 1993). As we show next, the mechanism of expectancy confirmation is especially significant when less traditional or less taken-for-granted persuasive appeals—like the persuasive appeal underlying woke brand communication—are implied in brand communication. The intrinsic novelty of these latter will in fact induce audiences to perceive a misalignment with their expectancies, resulting in a form of violation (Althuizen 2021; Förster et al. 2010; Hilton et al. 1991).
Woke persuasive appeal
Woke communication has recently emerged as a new form of communication which brands employ to resonate with their audiences (Mizrei et al. 2022; Song and Choi 2023; Wannow et al. 2023). Acknowledging this trending phenomenon in brand communication, Mangiò et al. (2021) demonstrated that woke communication is characterized by a different persuasive appeal that they named social pathos. Social pathos is a persuasive appeal which purposefully leverages on social-sensitive issues to showcase a brand’s support for a divisive social cause, thus pandering to recipients’ positive emotions and to their sense of community (Abitbol and VanDyke 2023; Gershon and Cryder 2018). For this reason, social pathos is a persuasive appeal which underpins authentic woke values (Karpen and Conduit 2020; Sobande 2020). Compared to the aforementioned traditional appeals of ethos, pathos, and logos, social pathos reflects a community logic (Gambetti and Biraghi 2023; York et al. 2018; Mangiò et al. 2021). Having made its debut in brand communication only recently, woke communication and the persuasive appeal it underlies (social pathos) are still far from achieving full legitimacy or an institutionalized status (Aboelenien and Nguyen 2023; Schneiberg and Clemens 2006), also because their usage in brand communication is intrinsically risky for brands (Jungblut and Johnen 2021; Mukherjee and Althuizen 2020). For this reason, woke communication is still less frequently used in brand communication compared to other forms of communication that leverage on more legitimated persuasive appeals (Guha and Korschun 2023; Korschun and Smith 2018). Indeed, recent research conducted by Guha and Korschun (2023) on a large sample of 177 brands across 35 industries has found that the share of woke communication in SM still amounts to just about one percent. It is hence clear that, compared to other institutionalized and taken-for-granted forms of brand communication, the novelty of woke communication in the corporate realm makes the adoption of social pathos by brands a violation of consumers’ expectancies about brand communication on SM (Althuizen 2021; Förster et al. 2010; Lee and Mason 1999). Moreover, since emotional communications set to fulfill prosocial goals—like those underpinning brand communication imbued with social pathos—are heuristically evaluated more positively by individuals, compared to other forms of self-focused communications set to exclusively or primarily fulfill commercial goals—like those underpinning traditional rhetorical appeals (Barasch et al. 2014; Nelson et al. 2016)—it is likely that the adoption of social pathos by brands will positively breach consumers’ expectancies. When this happens, such positive violation is expected to foster persuasion, and to lead to the formation of a positive attitude in, and a consistent behavioral response by, the receiver (e.g., high CESM). Based on this reasoning, we therefore hypothesize the following:
H1. Brand posts using social pathos generate higher CESM through the three stages of relationship formation (H1a), creation of engagement (H1b), and contribution (H1c) than do the traditional persuasive appeals of ethos, pathos, and logos.
The brands-as-intentional-agents framework
Consistently with the tenets of LET, when an expectancy violation occurs, the recipients shift their attention from the message content to the message source in order to make sense of that violation and attune future behaviors (Burgoon 1993; Rocklage and Fazio 2020; Yang et al. 2020). Because brand communication works differently depending on the brand (Dubois et al. 2016), the overall effectiveness of brand communication depends on consumers’ perceptions of the brand undertaking it (Eigenraam et al. 2021). In accordance with this view, we contend that the brands-as-intentional-agents framework (BIAF) offers a valuable lens through which to evaluate the perceptions of consumers activated at this stage of the interpretation-evaluation process prescribed by LET (Afifi and Metts 1998). The BIAF hinges on the assumption that brands are not merely names or symbols conveying functional features of products and services; rather, brands are intentional social actors ontologically comparable to people (Fiske et al. 2002; Kervyn et al. 2022, 2012). Drawing on the influential perception content model (Fiske et al. 2002), the BIAF has been complemented to include two stereotypes that are naturally implied in the heuristic perception and evaluation of brands by consumers: warmth and competence. Perceptions of warmth depend on a brand’s ability to convey trustworthiness, sincerity, kindness, and friendliness, while perceptions of competence are associated with a brand’s efficiency, skill, confidence, and intelligence (Cuddy et al. 2008). Previous studies have shown that perceptions of warmth and competence arise together with consumers’ expectations of contextual brand actions (Eigenraam et al. 2021; Dubois et al. 2016; Magee 2022; Ren et al. 2023). Put briefly, warm brands are associated with the fulfillment of emotional, hedonic, and altruistic needs, while competent brands are associated with the fulfillment of functional, utilitarian, and self-oriented needs (Eigenraam et al. 2021; Gershon and Cryder 2018; Grazzini et al. 2021; Johnson et al. 2018; Tellis et al. 2019; Zawisza and Pittard 2015). Thus, consumers perceive emotional communication as more authentic when it is enacted by warm brands, and less authentic when enacted by competent brands (Eigernraam et al. 2021). We contend that this reasoning can be transposed to woke communication, since this latter is by its very nature more emotional than functional, and more other-oriented than self-oriented. In support of this view, previous research has underlined that woke brand communication is imbued with altruistic overtones (Ahmad et al. 2023) and that it displays a dimension of morality (Abitbol and VanDyke 2023; Coombs and Holladay 2018), evoking deep moral-emotional states such as compassion, inspiration, and gratitude among the recipients (Abitbol and VanDyke 2023; Wannow et al. 2023; Zhou et al. 2023). Thus, when a competent brand, which is expected to refrain from exploiting emotions in its brand communication, instead does so, it will negatively breach consumers’ expectancies, inhibiting the communication effort’s ability to generate positive communication outcomes like CESM. Lastly, because consumers expect any kind of brand to engage in informative and credible communication (Eigenraam et al. 2021; Ismagilova et al. 2020), these CESM initiatives will trigger expectancy confirmation regardless of whether the brand is perceived as warm or competent, causing no change of attitude or behavior among the recipients.
In line with this reasoning, this paper hypothesizes the following:
H2. Relationship formation (H2a), creation of engagement (H2b), and contribution (H2c) are higher for warm brands than for competent brands using pathos and social pathos persuasive appeal.
Conceptual framework and hypotheses
Data collection and analytical strategy
To answer our research questions, we designed and implemented an automated text analysis field study (Humphreys and Wang 2018) to analyze publicly available consumer online reactions to woke communication embedded in prosocial Covid-19 campaigns in 2020. Because the pandemic had enormous socioeconomic consequences for the broader society (Taylor 2020)—and in regard to which brands took a divisive stance by communicating their active public health and economic support in terms of donations, countermeasures, and advocacy (Mangiò et al. 2021)—pandemic-focused communication by brands has been described as one of the broadest and most ubiquitous brand woke communication campaigns ever (Abitbol and VanDyke 2023; Sobande 2020). It is therefore a suitable setting in which to study the phenomenon under investigation. We collected Facebook data from a representative sample of 24 Italian brands operating in seven industries during the whole of 2020 (see Table 1). Among the various SM platforms available, Facebook was selected because it is characterized by higher levels of active rather than passive engagement compared to other SM platforms (Kübler et al. 2020; Shahabaznezahd et al. 2021). Brands were carefully chosen on the basis of two criteria: (1) they had to be listed on reliable international brand value rankings (Kantar 2021), and (2) they had a verified Facebook public page for Italy which remained active during the period of investigation. Once the brands’ official accounts had been identified, we collected their posts published in the period considered. In the absence of a dedicated public API (Tromble 2021), we developed a custom scraping protocol to obtain both brand- and audience-generated SM content. This procedure was conducted three times a week until one month after the end of the period considered in different schedules in order to update the dataset and avoid biases associated with the day of collection. Although we selected Italian brands only, because many of the brands identified have a worldwide reputation and serve a global market, we automatically detected and translated posts and comments written in languages other than Italian via Google Translate API in order to ensure data homogeneity. For the same reason, all social media managers’ moderation responses were deleted from the corpus of comments. To comply with research ethical standards, all references to users’ identifiers like names and mentions were removed. The final dataset consisted of 3252 brand-generated posts and 262,019 user-generated comments (avg. post length = 10.83 words, σ = 13.31; avg. comment length = 14.70, σ = 22.01). Once the data wrangling stage was over, we prepared the data for automated text analysis and split them into two different corpora: one including brand-generated posts, and one including users’ comments. These were then preprocessed and analyzed respectively through both top-down and bottom-up protocols (Humphreys and Wang 2018). The first step consisted in a volume-based CESM analysis involving the automatic classification of brand-generated posts according to their persuasive appeal. Applying ad hoc dictionaries adapted and validated for the specific context (Mangiò et al. 2021), this analysis was therefore aimed at assessing the extent to which traditional (i.e., ethos, pathos, and logos) and woke (i.e., social pathos) persuasive appeals are associated with different levels of volume-based CESM. The second step consisted in a semantic CESM analysis of user-generated comments. Used for this purpose were a computational content analysis involving both the semantic and affective aspects of CESM applying topic modeling (Roberts et al. 2014; 2019) and controversy detection analysis (Garimella et al. 2018).
Volume-based CESM analysis
CESM was operationalized via three volume-based metrics achieved by each brand-generated post over the period considered: number of likes (relationship formation), number of comments (creation of engagement), and number of shares (contribution) (de Oliveira-Santini et al. 2020; Swani and Labreque 2020; Vlachvei et al. 2021). Persuasive appeals were operationalized through four pre-built lexicons used in previous research (See Appendix B for sample brand-generated communication for each persuasive appeal.) (Mangiò et al. 2021; 0.72 < a < 0.84; Krippendorff 2018). In line with methodological suggestions for top-down automated text analysis (Humphreys and Wang 2018), the presence of each construct was operationalized as the token-weighted proportion of target lemmas in each document in the corpus (i.e., brand-generated post). Brand perceptions were assessed by means of a survey administered to a convenience sample of active Facebook users (N = 68), who were asked to rate the degree of warmth and competence of each of the 24 selected brands using validated scales (Aaker et al. 2010). A t-test made it possible to identify 12 predominantly warm and 12 predominantly competent brands (Ren et al. 2023). Following previous studies on CESM (Araujo et al. 2015; Moran et al. 2019), nine control variables were also included: post timing, post phase, industry, readability, fan base, brand equity, media vividness, post length, and use of hashtags.
Before any computation was performed, all measures obtained by means of a lexicon-based approach were scaled and mean-centered (via z-score) and shifted so that the minimum was equal to zero (Aiken and West 1991). The descriptive statistics of each variable and correlations are detailed in Tables 2 and 3, respectively.
To test hypotheses H1-H2, three sets of stepwise negative binomial regressions with maximum-likelihood estimation were run. Generalized linear models such as negative binomial regression are better suited to accounting for the overdispersion of the dependent variables which are positively skewed, like count-data (skewLikes 6.29; skewComments 9.51; skewShares 8.21).
Semantic CESM analysis
To assess the semantic dimension of CESM generated by different persuasive appeals, the second step of our analytical procedure involved a topic modeling (Roberts et al. 2014) and a controversy detection analysis (Garimella et al. 2018) on user-generated comments. For the topic modeling analysis, we built a model of online comments in response to brand-generated posts employing one of the most recent generative models-based techniques of the family of latent Dirichlet allocation algorithms (LDA) (Blei et al. 2003): that is, the structural topic models (STM) (Roberts et al. 2014). Once that text preprocessing, model specification, topic interpretation, and topic validation were over, STM allowed identifying 33 unique topics representative of seven distinct thematic clusters discussed by consumers in their comments. Controversy was quantified by means of an aspect-based sentiment-analysis approach (Choi et al. 2010; Dehler-Holland et al. 2022), aggregating topic sentiment scores at thematic cluster level and computing each cluster’s sentiment variance (Garimella et al. 2018; See Appendix C for more details on each step of the analytical protocol followed during the semantic CESM analysis and for a detailed description of each thematic cluster.).
Results
As for the volume-based CESM analysis, the regression results along the three stages of CESM for both the direct and interaction effects are reported in Table 4. To facilitate interpretation, the incidence rate ratios (IRR) of the β coefficients from the regression results were computed. The results of the likelihood ratio test indicated a good fit for all the negative binomial regression models. No multicollinearity issues were detected, since VIF for all predictors in the direct effects models was lower than 5 (James et al. 2017).
To test H1, we first tested the impact of the three traditional persuasive appeals of ethos, pathos, and logos on the three stages of CESM. Regarding ethos, the results showed a significant and positive effect on relationship formation and contribution of engagement, while the effect was found to be non-significant for creation.
Pathos was found to be non-significantly associated with two of the three stages of CESM (relationship formation and creation of engagement), but not with the last stage (contribution), which was found to be significant but negative. Similarly, logos was found to be significantly and negatively related to the first (relationship formation) and last (contribution) stage of CESM, while it was non-significant for the second (creation of engagement) stage. These unexpected negative effects, which partly point in a direction opposite to that of previous studies that have underlined a positive connection between content’s emotionality and content’s virality (Akpinar and Berger 2017; Berger and Milkman 2012; Villarroel Ordenes et al. 2019), emphasize the importance of considering the contingency of persuasive appeals effects on online behaviors (Rocklage and Fazio 2020; Shahbaznezhad et al. 2021; Tellis et al. 2019). For instance, Lee (2021) demonstrated that greater emotionality in brand communication on SM generates lower brand status perceptions of high sociocultural status groups, while it increases brand status perceptions in low sociocultural status groups. Rocklage and Fazio (2020) showed that emotional communication can hamper perceived digital content helpfulness for utilitarian products, doing so through a reduction of perceived trust. Similarly, it has also been found that informative content can backfire on CESM as well, because the arid nature of arguments commonly framing informative appeals, especially when displayed in experiential SM platforms like Facebook, reduces their likeability, and they are enjoyed only when they concern new products and services (Tellis et al. 2019).
As regards social pathos, the results show that this persuasive appeal has a significant and positive effect on all of the three stages of CESM, and that the impact of social pathos compared to the other persuasive rhetorical appeals on the three stages of CESM is by far the strongest. H1 is thus fully confirmed.
Regarding control variables, we found significant differences among posting time periods, days of the week, and industries, which provided further support for the view that content scheduling is vital in SM (Vlachvei et al. 2021) and that different industries resort to both different media types and different content orientations while reaching their online audiences (Lee et al. 2018; Swani and Milne 2017; Tafesse and Wien 2018). As expected, the readability of a brand-generated post was found to significantly affect the first and last stages of CESM, i.e., relationship formation and creation, confirming that the more readable a post is, the more it triggers CESM (Pancer et al. 2019). Similarly, we obtained empirical evidence that the relative strength of the brand has a positive effect on engagement behaviors (Araujo et al. 2015) because both fan base and brand equity positively affect all stages of CESM. As regards media vividness, pictorial posts enhance only the liking behavior, while highly vivid posts including videos enhance sharing behaviors (Shahabaznezahd et al. 2021). Lastly, our analysis also confirmed that the use of linguistic features and SM affordances that favor cognitive processing enhance CESM (Arujo et al. 2015; Deng et al. 2021), since post length and hashtags were found to significantly and positively impact on all CESM stages except for creation.
To test H2, we evaluated the interaction between the persuasive appeals adopted by brands and the brand perceptions perceived by consumers (Table 4). The moderation analysis evidenced that the interaction between social pathos and brand perception has a positive and significant effect on the first two stages of CESM, but a non-significant effect on the last one. Social pathos is more likely to trigger forms of engagement such as liking and commenting if this persuasive appeal is resorted to by warm rather than competent brands. Regarding the other appeals, pathos is more effective for warm brands in terms of liking and sharing, while the interaction between logos and ethos and brand perception is non-significant across all stages of CESM, as expected. Thus, H2 is partially confirmed.
As regards the semantic CESM analysis, Fig. 2 depicts the network with 33 topics discussed by consumers in reaction to brand-generated posts. Within this network we identified five discourses (Phillips et al. 2004) widely investigated in the branding literature, i.e., “brand-consumer co-creation” (cluster 1), “positivity towards brands” (cluster 2), “brand referral” (cluster 3), “negativity towards brands” (cluster 4), “nostalgia” (cluster 6), and two more idiosyncratic discourses that we named “mixed feelings” (cluster 5), and “covid complaints” (cluster 7).
STM topic network graph
The controversy detection analysis revealed that the seven clusters differed in terms of controversy, and in particular that cluster 5 (“mixed feelings”) had the highest sentiment variance, and thus contained the most polarized reactions (σ2 TS(1): 18.484, σ2 TS(2): 20.436, σ2 TS(3): 16.691, σ2 TS(4): 18.218, σ2 TS(5): 22.223, σ2 TS(6): 18.772; σ2 TS(7): 19.616; Bartlett’s K^2 (39) = 1000.1***). Including the post type variable as a covariate in the STM estimation made it possible to model how the different persuasive appeals used by brands triggered the prevalence of topics occurring in consumer-generated comments. In other words, this feature of STM made it possible to obtain the proportion of each topic associated with all of the four persuasive appeals used for the classification task in the volume-based CESM analysis, and to assess whether the association between topics and persuasive rhetorical appeal was statistically significant. The estimated differences in topic proportions for the four persuasive appeals are shown in Fig. 3.
Thematic cluster prevalence, by post type
As is apparent, the consumer reactions that we identified as intrinsically polarized (cluster 5) were statistically more prevalent below brand-generated posts imbued with social pathos, i.e., with the persuasive rhetorical appeal that characterizes woke communication. Brand-generated posts dominated by the informative persuasive appeal (logos) were associated with highly negative consumers’ reactions on the platform (cluster 4). As a matter of fact, negative online reactions like consumer complaints are more typical of informative brand communication on SM, because the information contained in the complaints can turn out to be instrumental for a wider range of consumers seeking to satisfy utilitarian needs (Johnene and Schnittka 2019). In addition, we tested for the moderating role of brand perceptions. The results further confirmed H2: polarized consumers’ reactions (cluster 5) are more prevalent when social pathos is resorted to by competent brands. Conversely, when warm brands use a woke appeal, they are more likely to trigger consumers’ positive discourses, like those included in cluster 6 (Fig. 4). We thus obtained empirical evidence that not only is woke communication controversial (Garimella et al. 2018), but it is especially so for competent brands.
Thematic cluster prevalence, moderation of brand perceptions
Theoretical implications
The findings of the analysis reported in this study have several theoretical implications which enlarge the current understanding of woke brand communication. Existing research has primarily focused on the design of woke brand communication and has principally examined how different framing techniques (Ahmad et al. 2023; Chu et al. 2023; Di Russo et al. 2022; Song and Choi 2023) and pictorial strategies (Zhou et al. 2023) impact the effectiveness of these initiatives, considering factors such as the type of commitment (e.g., rhetorical vs. financial; Ahmad et al. 2023) and the level of involvement (e.g., Song and Choi 2023) employed by brands. However, previous studies have mainly examined woke appeals in isolation, and not in tandem with other traditional appeals. This gap in the literature is noteworthy since woke brand communication is just one among the many forms of brand communication that brands can utilize on SM platforms (Ashley and Tuten 2015; Lee et al. 2018). Deciding if, and how, to use woke communication in conjunction with other traditional appeals a brand can make to its audiences can no longer be overlooked, especially in light of the increased penetration that woke communication is achieving in brand communication strategies. In fact, as evidenced by the unexpected negative effects of pathos and logos on CESM that we found in our analysis, a cognitive fatigue effect may occur in a specific communication context (Lang 2000). This could take the form of a “zero-sum game”, where the persuasive appeal underlying woke communication monopolizes attentional resources to the detriment of other appeals, inhibiting their expected ability to trigger CESM. Furthermore, we have addressed recent calls to broaden our understanding of boundary conditions theoretically relevant for the effectiveness of woke brand communication (Weber et al. 2023). Beyond brand characteristics such as size (Hydock et al. 2020), value versus result-orientation (Korschun et al. 2019), perceived credibility (Song and Choi 2023) and authenticity (Chu et al. 2022; Thomas and Fowler 2023; Vredenburg et al. 2020), we have demonstrated that brand stereotypes—warmth and competence—influence the overall effectiveness of woke communication on SM platforms. In this regard, our findings support the notion that the effectiveness of brand activism relies on different extents of fit (Vrendenburg et al. 2018) which do not just encompass the brand and the cause endorsed (Chu et al. 2022), but also involve a pre-interactional fit like the warm-communal orientation congruency prescribed by BIAF. In this regard, the combination of LET and BIAF makes it possible to explain woke brand communication not as a challenge of authenticity, a nebulous concept that eludes a precise definition and practical operationalization (Nunes et al. 2021; Thompson and Kumar 2022), but rather, as one of perceptions like warmth and competence and their fit with other-oriented emotional overtones embedded in woke communication.
This study also contributes to communication theory in two significant ways. First, it advances the application of LET within SM contexts (Lee and Yu 2020). Previous research based on the theoretical tenets of LET focused on technology-mediated contexts where communication exchanges occur between human actors, as in the case of peer-to-peer online reviews (Folse et al 2016; Jensen et al. 2013; Wu et al. 2017). Our research validates the openness of LET (Burgoon 1993) by demonstrating that its principles hold true even in technologically mediated contexts involving communication between human (i.e., consumer audiences) and non-human (i.e., brands) actors. More generally, we provide a theoretical and empirical confirmation that LET is a suitable lens through which to study brand communication and the mechanisms prompting audiences to either develop attention, and to engage with brands. Second, the greater ability of social pathos to prompt CESM compared to traditional appeals supports the validity of the “moral contagion” thesis, which posits that the social transmission of moral emotions plays a key role in determining how moral ideas spread through social networks (Brady et al. 2020; 2017), and extends its application also in the context of brand communication via SM.
Besides theoretical implication stricto sensu, this paper also offers methodological contributions, because it responds to recent calls to (1) illuminate the effectiveness of woke brand communication by conducting a field study in a real-life, ecologically valid setting, utilizing big data-friendly methodologies (Zhou et al. 2023), and (2) extend such investigations beyond company websites or microblogs, such as Twitter (Abitbol and VanDyck 2023). The analytical protocol implemented has numerous advantages for analyzing responses and social evaluations in regard to divisive stimuli in naturally occurring settings, going beyond traditional sentiment measures (Yang et al. 2021; Weber et al. 2023). The controversy detection analysis enabled us to monitor consumers’ online reaction polarization as the variance in highly emotional opinions, an aspect often obscured when relying solely on average values and, in contrast to purely qualitative content analysis, allowing us to quantify this phenomenon across a substantial dataset in a quasi-experimental fashion. Simultaneously, unlike purely quantitative methods (Zhao et al. 2023), the STM approach permits a dialogical assessment of polarizing debates (Gambetti and Biraghi 2023), so that we were able to explore the relationship between the source of a given brand-generated content and its subsequent user-generated reaction.
Managerial implications
Given the status of sociopolitical polarization and digital participation in contemporary society, the results of this study provide timely and crucial insights also for brands and marketing practitioners. Contrary to the widely circulated “polarize and conquer” mantra, which suggests that being divisive is unequivocally beneficial for brands (Luo et al. 2013; Maryott 2021), our results instead advocate for a more cautious and mindful approach. They show that employing woke communication on SM is particularly persuasive only for those brands perceived as warm by online audiences. However, this is not to say that competent brands should refrain from endorsing woke communication. Brand activism is, after all, rooted in a brand’s value orientation and in a brand’s higher purpose. If a brand is genuinely inclined to contribute to a specific issue which is considered salient to its identity, it should remain faithful to its mission and consider pursuing woke communication regardless of whether it is perceived as warmer or more competent by external stakeholders (Coombs and Holladay 2018). Nonetheless, when employing woke communication, competent brands should be prepared to identify, address, and manage the consumer-driven contestations and firestorms that are more likely to be triggered compared to their warmer counterparts. To achieve this, competent brands could intensify their investments in content moderation routines and protocols, ensuring a prompt and thoughtful crisis response as soon as polarization escalates among user reactions (Dineva et al. 2020).
The methodological contributions of this paper offer practical guidance as well. Since managers report that business intelligence will increasingly depend on SM data (Sprout Social 2023), in light of our results we invite practitioners to perform continuous and reiterative SM listening with the correct tools and measures. These tools should be validated for the specific context at hand and not used as they come “off-the-shelf” (Hayes et al. 2021). This entails not only relying on average sentiment measures but also tracking sentiment distribution or extremity (Rocklage and Fazio 2020) and the presence of heterogeneous evaluators (Ilia et al. 2023). This endeavor is urgent both because digital media have expanded the pool of actors aware of and eager to advocate for social issues (Pimentel et al. 2023) and because such issues are hardly ever solved. At most, they are “resolved”, meaning that they may become dormant for a while, and suddenly reemerge (Coombs and Holladay 2018).
Conclusions, limitations, and future directions
Brands are increasingly required to be ‘woke’ and communicate their stance on various divisive sociopolitical issues, and to do so particularly on SM platforms. Prior research has shed light on which consumers’ reactions are more likely to be elicited by woke brand communication, and it has shown which framings are more effectively used in this form of communication. However, no research to date has compared the effectiveness of woke communication with respect to other forms of brand communication featuring different, yet more institutionalized, persuasive appeals, i.e., ethos, pathos, and logos. Moreover, previous research has provided very little guidance about the brands that are more likely to enjoy the benefits stemming from this form of communication. Bridging the tenets of LET, the BIAF, and the literature on consumer engagement in SM, and by means of an automated text analysis field study involving multiple industries and multiple brands, this paper has shed additional light on this still under-researched area. First, the volume-based CESM analysis confirmed that brand woke communication on SM, operationalized by the social pathos appeal, triggers more CESM through the phases of relationship formation, creation, and contribution than can traditional ethos, pathos, and logos persuasive appeals. However, this is particularly true for warm rather than for competent brands. Second, the semantic CESM analysis further supported and furnished finer-grained understanding of these results, because brand-generated posts framed with social pathos were statistically associated with the most polarized consumer comments especially when they were posted by competent brands. Hence, our study also demonstrates that volume-based analyses alone cannot provide an all-encompassing picture of CESM dynamics. This study contributes to the emerging scholarly debate on brand woke communication by shedding light upon its impact on CESM and, most importantly, by identifying for which type of brand this appeal is most effective.
However, the results of this study have some limitations which open avenues for further research. Firstly, although the choice of Italy as the research context was not random, focusing on a single country may provide only a partial explanation of the phenomenon investigated. The brand-consumer interactions which took place on the Facebook pages of Italian brands could have been biased by cultural dynamics. Future studies should therefore test the identified relationship between social pathos and polarized consumers’ reactions in cross-country, cross-cultural settings. Secondly, computational methods such as those employed in this study have many advantages; however, because they blend aspects of multimodal communication phenomena, they necessarily gloss over nuances. Qualitative in-depth investigations could infer finer-grained shades of the brand polarization phenomenon on social media. In this regard, computational social media analysis does not allow investigation of this phenomenon at the micro level, so that the individual details and features of the commentors engaged in sharing polarized comments can be gauged. Indeed, we still know very little about who lies behind such behavior. Further studies adopting methods framed by more controllable settings like experimental design could determine what are the recurrent sociocognitive factors and personality traits most likely associated with consumers posting polarized reactions to brands on social media.
Appendix A: A summary of studies on brands’ public engagement with a partisan sociopolitical issue.
Author | Method | Data and Sample | Controversial issue | Theoretical background | Independent variable | Dependent variable | Findings |
|---|---|---|---|---|---|---|---|
Abitbol and VanDyke (2023) | content analysis | 352 CSA messages (written or videos);116 Fortune250 U.S. brands, 9 industries | various | self-transcendent media psychology | - | - | The most prevalent topics among companies' CSA stances are related to "hot-button" issues, highly salient to both consumers and media during the period of investigation, like race and social justice, and health and Covid-19; the most prevalent stance is supportive of a social issue. Interestingly, most issues covered did no align with the company's business purpose. Most messages contained emotional elicitors. In terms of transcendental emotional elicitors, beauty and excellence (communicating how great things had been done for a cause) was the one most prevalent |
Ahmad et al. (2023) | Auomated text analysis (ATA) (exploratory); experiments | Twitter; 5000 tweets and 5000 replies; 45 brands in multiple industries (study 1); students ‘data (study 1–4); Mturk respondents (study 5) | BLM; gun violence; LGBTQ rights | message framing literature; CSR, CSA previous literature | BA commitment (financial, non-financial, rhetorical); message framing (hope vs frustration); brand equity | brand authenticity (also as moderator); brand love | Whereas high equity brands should communicate a financial commitment to gain authenticity and love, low equity brands should emphasize a non-financial commitment. In general, brands must convey hope rather than frustration in their messages. However, brands should convey frustration when communicating a rhetorical commitment to a cause |
Guha and Korschun (2023) | ATA, regression analysis | Twitter; 14.504 tweets and 55,121 replies, 77 brands, 35 industries | BLM; LGBTQIA rights, covid-19 | peer influence, brand communication on social media, brand activism | industry peer and peer tweets, positive replies, mentions, average replies, google searches | likelihood of tweeting on a given issue (daily) | Brands' likelihood of tweeting about divisive sociopolitical issues is higher when their peers have recently done so, when consumer responses to such messages are positive, and when there are more SM appeals to engage with the issue |
Lee and Chung (2023) | survey | 505 respondents | gun control | CSR communication framework, social judgment theory (SJT), and the elaboration likelihood model (ELM) | attitude toward the company, new credibility, moderator: pre-existing issue stance | issue stance change, attitude toward CSA, skepticism about CSA initiatives | A consumer's pre-existing stance on an issue affects his/her reactions to CSA. When a consumer's stance was undecided, his/her attitude toward the company and news credibility significantly relate to change in issue stance, attitude toward the CSA campaign, and skepticism about the company’s motives |
Song and Choi (2023) | survey | 530 U.S. respondents | diversity; gun control; immigration | persuasion knowledge model (PKM) and ELM | antecedents: CSA message’s informativeness; factual and promotional tone; mediator: CSA credibility; moderator: level of involvement | organization-public relationship (OPR); issue advocacy | CSA message’s informativeness, message tone, and credibility successfully predicted both organizational and social outcomes. Perceived CSA credibility strongly mediates for OPR, but weakly for issue advocacy behavior. Promotional tone does not have any significant association with CSA credibility. Factual tone positively affects OPR and issue advocacy, while promotional tone does not have a significant effect on OPR or on issue advocacy |
Thomas and Fowler (2023) | SM analysis (pilot) and experiments | Instagram; SM posts by 54 social media influencers (exploratory study); 134 students (study 1); 275 participants (study 2); 227 participants (study 3) | Ukraine war, BLM | expectancy-disconfirmation theory | Presence and type (temporary vs sustained) influencer activism; perceived authenticity (mediator) | expectations about activism, citizenship behaviors and direct support behaviors; attitudes toward the influencer | When an influencer engages in activism, consumers will report stronger attitudes toward the influencer, higher expectations of citizenship behavior, and higher expectations of direct support behavior. Temporary influencer activism will result in weaker attitudes toward the influencer and low perceptions of authenticity compared to sustained activism |
Wannow et al. (2023) | experiments | 179 participants (study 1); 244 participants (study 2); 205 participants (study 3) | immigration; abortion | moral emotions | mediator: moral emotions (other-condemning and other-praising); moderator: consumer-brand identification | brand attitude; issue advocacy | Other-praising and other-condemning moral emotions mediate consumer-brand (dis)agreement and brand attitude and brand issue advocacy respectively. Consumer-brand identification moderates such effects: consumers with high levels of identification experience stronger moral emotions in response to brand activism messages. Brand activism cues disproportionally activate the advocacy behavior of brand's social issue stance opponents, compared to advocates |
Weber et al. (2023) | ATA (study 1); online experiments (study 2,3) | Twitter, 253,649 tweets, multiple brands (study 1); 3,914 respondents 1 fictitious brand (study 2); 504 respondents, 1 fictitious brand (study 3) | various | literature on value match-effect; political efficacy | CPA vs CSR message; mediators: perceived appropriateness and controversy; participant political orientation | message sentiment; brand attitude; purchase intention | Consumers' responses in the form of SM sentiment and attitude to CPA are less favorable and more polarized than responses to CSR, in part because CPA is seen as less appropriate and more controversial; more polarized. Reactions to CPA are driven by consumers low in political efficacy |
Zhou et al. (2023) | experiments | 334 participants, 1 fictitious brand (study 1); 320 participants, 1 fictitious brand (study 2); | gender equality; health care affordability | transcendent media experiences, color, and message framing theory | color (BW vs. color); moderators: promotion vs. prevention-framing; mediator: inspiration | ad attitude, purchase intention | Brand activism ads that use BW (versus color) images, when paired with promotion-framed (versus prevention-framed) messages, lead to more favorable attitudes toward the ads and elicit higher purchase intentions by triggering inspiration and transcendent experiences |
Batista et al. (2022) | experiments | 156 participants (study 1b); 229 participants (study 2b); 394 (study 2); 573 (study 3); 623 (study 4); | various | literature about incivility, assertiveness and sarcasm | antecedents: type of reply (sarcastic vs assertive); mediator: perceived aggressiveness of the reply; boundary condition: support for the brand's stance | attitude toward the brand | Sarcastic replies are perceived to be more aggressive, while assertive ones create a more favorable attitude toward the brand. This varies according to consumers’ level of support for the brand’s stance, so that even in the presence of aggressive sarcastic responses, consumers who support the brand’s stance will still have a positive attitude toward the brand |
Chu et al. (2022) | survey | 349 respondents | various | cognitive behavioral theory, congruity theory, attribution theory, social identity theory | brand factors (brand-cause fit; self-serving motives); social media factors (attitude toward BA in SM; trust in SM); mediators (i.e., brand activism authenticity and consumers’ eWOM intention) | brand image, purchase intention; brand loyalty | Brand-sociopolitical cause fit, self-serving motive, attitude toward brand activism, and trust in SM positively relate to the authenticity of brand activism in SM. Brand activism authenticity is positively associated with electronic word-of-mouth (eWOM) intention, which consequently enhances brand image, purchase intent, and brand loyalty |
Di Russo et al. (2022) | experiment | 508 participants, 1 brand | opioid addiction; pandemic preparedness | Limited Capacity Model of Motivated Mediated Message | arousal, valence, issue salience; mediators: positive vs negative emotions | memory, attitude toward the organization, purchase intentions, social media intentions, political.participation intentions | Negatively valenced CSA message are more persuasive. High-arousal language and issue salience do not differentially impact on attitudes or purchase intentions, but the high-salience issue and high-arousal language enhance political participation and social media intentions |
Li et al. (2022) | survey | 345 American citizens; 1 brand | diversity | consumer involvement theory and stakeholder theory | issue involvement factors (cognitive and affective), brand attachment factors (brand-self connection; brand attitude) | CSA attitudes, negative and positive WOM | Consumers that consider the sociopolitical stance taken by the brand to be important, meaningful, and/or positively emotional are more likely to show support for that campaign and the company that chose to advocate for the issue. A favorable attitude toward the campaign is driven by issue, personality, and value alignment between consumers and the brand. CSA can also attract consumers with an originally low attachment to the brand if the latter advocates a sociopolitical issue supported by the former |
Mirzaei et al. (2022) | unsupervised ATA | 46,000 and 34,000 UGC, 2 brands | femvertising and anti-racism | Brand authenticity | - | - | Woke authenticity is influenced by social context independence, perceived inclusion, profit sacrifice, actual practice, and underpinning motivation |
Feng et al. (2021) | supervised ATA (study 1); experiment (study 2) | YouTube; 125,481 UGC (study 1); 1,139 users (study 2); 1 brand | femvertising | Social Norms Theory; Social Identity Model of Deindividuation effects; Information Cascades | social norm conditions | consumer reaction (type of comment); ad attitude; brand attitudes; purchase intentions; demographics | When evaluating a YouTube-based woke advertisement, consumers without the social norms condition are more likely than those in the static social norms condition to generate positive ad attitudes, positive brand attitudes and high purchase intentions; consumers exposed to a dynamic social norms condition are more likely to be influenced by the prevailing norms than are those in a static social norms condition. Conservative men tend to post more negative comments; liberal women tend to post more positive comments. consumers’ responses on social media are more negative than those from self-report data |
Jungblut and Johnen (2021) | experiments | 158 respondents, 2 FMCG brands (study 1);805 respondents, 2 fictional brands (study 2) | immigration; gun control | political consumerism; Balance Theory | strength and valence of individual opinion toward the brand’s political brand communication, category involvement, consumer political interest, ad skepticism (control) | brand image (study 1); purchase intention (study 2) | When brands engage in political communication, the negative effects on brand image and the purchase intention of disapproving consumers (boycotters) outweigh the positive effects of approving consumers (buycotters), and the magnitude of this effect decreases for higher levels of consumers’ political interest and low levels of category involvement |
Klostermann et al. (2021) | event study | 106 CPA events | various | Brand-consumer overlap and negative effects of corporate political advocacy; self-brand similarity; online protests; effort; concurrence | CPA behavior (effort, concurrence), online protest (mediator), control variables (brand awareness, alignment, event time, controversy) | cumulated abnormal value of brand perception | CPA has a negative effect on consumers’ brand perceptions; this effect is stronger for customers relative to non-customers. Effort and concurrence moderate CPA’s effect on consumer perceptions |
Park (2021) | survey | 960 respondents, 1 fictitious brand | fictitious | Signaling theory | consumer-company identification, corporate issue identification, CSR skepticism; controls: age, gender, income, education, CSA familiarity, attitude toward the company) | brand trust, brand loyalty | A brand’s strong and clear identification with a controversial sociopolitical issue is positively associated with brand trust and loyalty, and it is mediated by reduced skepticism about corporate non-market activities, especially when consumers have a favorable attitude toward the company |
Schmidt et al. (2021) | focus groups (study 1), survey (study 2), experiments (study 3, 4) | 5 groups with an average of 10 college students each (study 1); 33 brand managers (study 2); 99 and 107 business students (study 3); 208 respondents (study 4) | diversity, gender rights | sociocultural perspective on brands and brand authenticity | sociopolitical brand or not | brand personality appeal, brand attitude, product use | Brands take sociopolitical stances for both cause-driven and consumer-driven goals. Authenticity is a key construct, and brands need to approach sociopolitical issues by translating them into actions that have meaning for consumers and remain consistent in the long run. Sociopolitically active brands are regarded more positively by consumers; women are more likely than men to think positively about a sociopolitically active brand. Sociopolitical activeness results in more positive levels of product use than does non-sociopolitical activism by a brand |
Villagra et al. (2021) | event study | stock prices of 33 listed companies, 1 event | hate speech | Corporate activism and brand boycott | - | stock price response (abnormal returns) | Corporate activism, when directed at a firm, has a negative effect on the stock market value of the target firm, but it does not have any effect on sponsoring companies, especially if this action is carried out as a group |
Xu et al. (2021) | experiment | 296 respondents, 2 fictitious brands | various | Construal Level Theory | perceived psychological distance, consumer-company identification (mediator), political partisanship | expectations for the company's CSA; attitudes toward company buycott and boycott intentions | Psychological distance from the brand affects consumers’ expectations about the brand’s commitment to the sociopolitical issue, but not their attitudinal responses to CSA. Greater perceived psychological distance decreases the intention to buycott and increases the intention to boycott the politically liberal brand, with boycott intention particularly salient among Republicans |
Yang et al. (2021) | supervised ATA | Instagram, 32.702 UGC; 110 accounts between brands and SM influencers | anti-racism | source-message fit; especially in the context of CSR; consumer engagement | type of SM account (brand vs. SM influencer; black vs. non-black) | consumer engagement ratios, sentiment, and topics | Woke content promoted by SM influencers generates more engagement than brand-promoted woke content. Criticism is most frequently observed for woke posts by brands, followed by non-black SM influencers. Non-black SM influencers register a higher percentage of negative comments than black influencers; purchase and boycott is present only under brand-generated posts. Anger/frustration in regard to racism, intention to share/engage, endorsement intention, and showing empathy through personal stories are observed only in influencer-generated posts |
Bhagwat et al. (2020) | event study | 293 CSA events initiated by 149 brands across 39 industries | various | Signaling and Screening theories | CSA event, form of support, announcement source stature, business interest communication, coalition size, deviation from customer values, deviation from employee values, deviation from government values, deviation from brand image; controls: industry and time- specific control variables; | stock price response (abnormal returns) | Investors’ reactions to CSA are on average negative. They deteriorate when CSA deviates from stakeholders’ political values, takes the form of actions rather than statements, is announced by the CEO rather than another actor within the firm, does not explicitly communicate any business interests, and is undertaken by a brand on its own rather than in coalition with others. CSA is rewarded when it closely resonates with stakeholders’ personal values |
Bravo and Lee (2020) | experiments | 288 students, 1 fictitious brand | health care and abolition of death penalty | SJT | Issue involvement, message agreement | attitude toward the ad, attitude toward the brand, intention to support the behavior advocated in the ad, purchase intention | Among millennials, message agreement mediates the effect of issue involvement on purchase intention and intention to support the behavior advocated in the advocacy ad. High issue involvement exerts a positive effect on attitudes toward the persuasive message in the advocacy ad, on purchase intention and intention to support the behavior advocated in the ad |
Milfeld and Flynt (2020) | phenomenological semi structured interviews | 24 respondents, 1 brand | femvertising | brand storytelling; narrative transportation theory | - | - | Social narrative cues embedded in brand videos create a polarizing effect able to both resolve tensions and create new ones among the audience at the same time. This polarizing effect depends on the (dis)connection between consumers and the brand-intended story |
Mukherjee and Althuizen (2020) | experiments | 144 US participants (study 1a); 115 French students (study 1b); 197 participants (study 2); 210 US participants (3), 304 US participants (study 4) | immigration, abortion, freedom of speech | moral foundations theory; moral rationalization and (de)coupling, | stance agreement; consumer’s moral reasoning strategies; source of the stance (moderator); consumer-brand identification (mediator) | consumer attitudes, intentions, and behavior | brand activist statements have an asymmetric effect on consumer attitudes, intentions, and behavior. Consumer-brand disagreement about the brand’s stance negatively affects brand attitude, whereas the effect of agreement is not significant. Consumer-brand identification partially mediates this negative effect of brand activism on consumer attitudes. A mismatch between brands and consumers’ moral foundations underpins the negative effect of consumer-brand disagreement about the brand’s stance on consumer-brand identification |
Parcha and Kingsley Westerman (2020) | experiment | 677 millennials, 1 fictitious brand | gun control; transgender rights | ELM | involvement (outcome-relevant and value-relevant involvement), advocacy fit, corporate credibility, bandwagon heuristic | attitude change toward the corporation’s position | CSA changes consumers’ attitudes in four ways: a) the more a sociopolitical issue personally affects a consumer’s goals, the more a woke statement on a low-fit issue changes his/her attitude; b) the more a sociopolitical issue personally affects a consumer’s goals, the more a woke statement supported by a large number of other brands changes his/her attitude; c) the less a sociopolitical issue personally affects a consumer’s goals, the more a woke statement supported by only a few other brands changes his/her attitude; d) the less a sociopolitical issue is important for a consumer’s values, the more a woke statement supported by only a few other corporations changes his/her attitude |
Rim et al. (2020) | social network analysis, quantitative content analysis | Twitter, 17,821 tweets from 4 hashtags, 2 brands | immigration | situational theory of publics | - | - | Within polarized social media communities generated by woke stances endorsed by brands, disapproving consumers (boycotters) appear not only in the aggregated brand boycotting networks, but also in the approving consumers’ (advocators) networks. Boycotters’ activities target also other brands or organizations that have adopted stances similar to that of the target brand |
Austin et al. (2019) | survey | 1,214 participants, 3 brands | diversity, gun control, femvertising | Public Interest Research and CSR | Public support for brands advocating for social issues varies according to political viewpoint, age, income, education, and gender. Liberal and younger respondents are more likely to express support than are older and conservative respondents. Higher levels of income, education, and overall concern for social issues also play a role in perceptions of corporate commitment to social issues | ||
Champlin et al. (2019) | inductive qualitative analysis | 19 commercials | femvertising | target audience brand-cause fit | - | - | In their femvertising practices, brands instantiate different forms of ‘brand-cause fit’ that are built on three types of matches: a functional match, an image match, a target audience match |
Feng et al. (2019) | supervised ATA | YouTube; 20,419 UGC; 1 brand | femvertising | Reception theory | - | - | Consumers discuss both adversarial and supporting topics about the sociopolitical stance, namely ad skepticism, beauty definition, praise, and discussion of broader issues |
Abitbol et al. (2018) | mixed methods case study | Twitter; 226 tweets; 4 employees (interviews); historical secondary data (2012–2016); 1 brand | anti-racism | Company-cause fit | - | - | Consumer reactions toward the woke campaign are polarized but predominantly negative, criticizing its poor execution, political skewness, logistic and setting inappropriateness, and perceived hypocrisy. Employees’ reactions include complaints about a lack of communication between organizers and employees and about a lack of participation among employees, but also expressions of pride in being a partner. Employees' perception of the brand do not change. No evidence of financial effects and ambiguous effect on reputation is detected |
Austin et al. (2016) | content analysis | Twitter, 917 UGC from 200 brand-generated posts;1 brand | anti-obesity, femvertising, sustainability | Typology of CSI types | CSI types, CSR topics | favorability of public comments | Posts emphasizing socially responsible business practices gain the most favorable public response, while posts focused on cause promotion gain the most negative ones. Brand communication is less effective when the issue and the advocated behavioral change appear to be acting against the brand's interests |
Leak et al. (2015) | experiments | 161 students | same-sex marriage | literature about ideology, affect, cognitions and management halo error | consumer-held ideology; brand attitude; anger derived from evaluating manager’s ideological position (mediator) | brand attitude change | When consumers perceive conflicting ideological beliefs vis-à-vis a manager’s expression of their stance on a social issue, congruence with such ideological stance leads to less anger and increased brand attitude. Pre-existing brand attitudes reduce the amount of anger (halo effect) |
Appendix B Sample brand-generated posts for each persuasive rhetorical appeal
Persuasive appeal | Example (translated) |
|---|---|
Ethos | “The main global stock exchanges closed August with record sales. Trust or enthusiasm? Here's the opinion of V.G., asset management director of Banca Mediolanum, for Panorama.it” |
Pathos | “Big or small, simple or elaborate, expected or unexpected: no matter what your present is like, your love is what makes it perfect #Nutellawithlove” |
Logos | “A2A aims at reducing carbon emissions by 30% by 2030. This new target was analyzed by the #ScienceBasedTargets initiative to verify the alignment of industry and Paris COP21's goals. More information here: https bit.ly/2wvpDmR” |
Social Pathos | “This emergency has put us to the test by redefining the way we work, teaching us to communicate in a new way. We stepped up to the plate for the community, without giving up, we went forward: fast, united, and together because we are Fastweb. Simply, #ConnectedTogether” |
Appendix C: Semantic CESM analysis analytical protocol
Topic modeling
Topic modeling comprises a large group of algorithms aimed at soft-clustering and discovering thematic structures hidden in large unstructured textual datasets without an a priori classification being needed (Airoldi et al. 2015; Blei 2012). The advantages that make topic modeling a toolkit increasingly used to conduct social science research (Aranda et al. 2021; Hannigan et al. 2019) reside in its being explicit, automated, inductive, and naturally keen to navigate the relational nature of textual data (DiMaggio et al. 2013). Among the various techniques refined over the years, generative models-based techniques such as latent Dirichlet allocation (LDA: Blei et al. 2003) have been found to be particularly effective in analysis of social media users’ comments (e.g., Mirzaei et al. 2022). However, LDA has some limitations: model estimation occurs without account being taken of significant document-level covariates that affect the topical prevalence (i.e., the frequency with which a specific topic is discussed) and the topical content (i.e., the differences in the language used to discuss a given topic); and it does so without it being possible to detect correlated topics, i.e., themes that tend to occur in the same documents (Hu et al. 2019). To overcome these shortcomings, we built a model of online comments in response to brand-generated posts that employed the structural topic model (STM; Roberts et al. 2014). As a recent extension of LDA and correlated topic models (Blei and Lafferty 2006), STM has already been fruitfully used in a variety of research domains, like political science, journalism studies, management and organization studies; but it has been only marginally employed in brand communication research (Fresneda et al. 2021; Hannigan et al. 2019; Reisenbichler and Reutterer 2019). Compared to previous generative models, the advantage of STM is that it makes it possible to explore the relationships between the identified topics and other variables by including pertinent document-level metadata during the model estimation, and by weighting for between-topics correlations (Schmiedel et al. 2019). The fundamental steps of the STM approach are text preprocessing, model specification, topic interpretation, and topic validation (Aranda et al. 2021; Hannigan et al. 2019).
Text preprocessing and model specification. Before conducting STM, we performed a thorough document-preparation and text preprocessing. We removed invalid records (such as comments with no words or duplicates), performed tokenization at the word level, document cleaning (e.g., removing hyperlinks such as URLs and https from the comments), enrichment (adding relevant bi-grams and tri-grams collocations), stopword-removal, word normalization (e.g., lowercasing and spelling). We also removed highly infrequent words, setting the minimum document occurrence at the conservative threshold of five (Banks et al. 2018). The final corpus contained 45,020 unique comments, 6,984 unique terms, and 323,002 tokens. After data wrangling and corpus preprocessing, the model was set up by defining topic prevalence as a generalized linear function of post type, brand perception, industry, and time of publication:
[1] Topic prevalence = g (post type, brand perception, post type*brand perception, industry, s (time of publication))
Where post type and brand perception are dummy variables with factors corresponding to the persuasive rhetorical appeal most prevalent in the post with which each comment is associated and the brand perception perceived by consumers, obtained in the first step of this study; industry is a dummy variable with factors corresponding to the seven industries investigated; and time of publication is operationalized as a spline function of the week of the year during which the user-generated comment was posted in order to account for nonlinearity of the time effects. After setting up the model, we selected the number of topics (K) for the STM. K is one of the most important user-specified parameters for topic modeling, although the literature warns that there is no one-size-fits-all procedure for identifying a number of topics that is the best one from both the analytical and interpretative standpoints (Hu et al. 2019; Schmiedel et al. 2019). We thus initially ran estimates for K in between 10 and 100, with an increment of five topics at each step, given the intrinsic nature of our corpus, which comprised several thousand short user-generated documents (Banks et al. 2018; Lindstedt 2019). Then, we compared the models through STM-specific diagnostics, namely held-out likelihood, exclusivity, and semantic coherence (Roberts et al. 2014), which informed us that the best models occur when 40 ≤ K ≤ 50 as differences in terms of held-out likelihood are small and most importantly the trade-off between semantic coherence and exclusivity is most marked (Roberts et al. 2014). Even though the larger the number of topics, the higher the level of exclusivity, more informative solutions can be obtained if exclusivity and semantic coherence are balanced. As with other clustering algorithms (Reisenbichler and Reutterer 2019), relying solely on these quantitative diagnostics is not sufficient. For this reason we qualitatively inspected the solutions of the models between 30 and 50 to check for the stability of topics among neighboring models, and we finally selected the 40 topics solution.
Topic interpretation and validation Although topic modeling techniques greatly help researchers to computationally assess extremely large textual data quickly and effectively, the interpretation of the results obtained through such techniques requires an interpretative inferential task and demands expertise on the part of the researcher (Aranda et al. 2021; DiMaggio and Blei 2013). For this reason, two researchers with extensive knowledge of the branding literature were employed to assign a label to each emergent topic on the basis of the underlying meanings of each of the top words and most representative comments that were automatically grouped together under the same cluster (Hu et al. 2019). Top words were identified with the FREX criterion (Roberts et al. 2014). Overall, 33 topics expressed themes coherent and exclusive enough to be associated with a single and concrete concept. To internally validate the results of the structural topic model, two authors coded a sample of the most representative comments per topic in order to assess if the model discriminated adequately. External validity was assessed by inspecting each topic’s performance with respect to its time distributions and prevalence over the time period considered (Grimmer and Stewart 2013). After topic interpretation and validation, three authors grouped the 33 topics into seven distinct thematic clusters representative of as many second-order constructs based on previous literature and inter-topic correlations (Hu et al. 2019).
Controversy detection analysis
The aim of the controversy detection analysis was to identify controversial topics, i.e., topics able to generate significant online debate (Garimella et al. 2018). Among the different methods available, we quantified controversy by means of a text and sentiment-analysis approach (Choi et al. 2010). Owing to the fact that the brands included in the analysis were well-known and that their social media presence was professionally managed (Kübler et al. 2020), we opted for an aspect-based sentiment-analysis (Dehler-Holland et al. 2022) which relied on a combination of the distribution of words of each topic identified via the STM with a selected lexicon. For the latter, the Italian version of the NRC-Emolex lexicon (Mohammad and Turney 2013) was chosen, adapted, and validated to our domain by 43 trained coders (average pairwise agreement: 81.22%; Krippendorff’s alpha: 72.58%). Given the role attributed to paralinguistic non-textual cues such as emoticons and emojis (McShane et al. 2021), sentiment was operationalized also via a rules-based approach which enabled us to detect and give a score to representative static emoticons and emojis selected because of their embedded sentiment polarity (Kralj Novak et al. 2015). Thus, for each topic, sentiment was computed using Dehler-Holland et al.’s (2022) approach as the normalized sum of the overall sentiment score per topic weighted with the word occurrence probability for each topic estimated by the STM (βwt), where the overall sentiment score for each lemma was the difference between the positive emotion score and the negative emotion score according to the sentiment lexicon.
Appendix D: STM summary table
Thematic cluster | Description | Literature | Mean cluster sentiment (σ2) | Topics | FREX (top 10*) | Representative documents* |
|---|---|---|---|---|---|---|
(1) Brand consumer co-creation | This cluster contains topics discussed by consumers that reflect the active engagement of the users as members of the brand community, for example in the form of personal suggestions on how to improve products and services to other users | Vallaster and von Wallpach (2013) | − 1.04 (18.485) | (1) Creativity | too, top, good, fond, yummy, family, surprise, find, anna, share | "What a lovely product to be enjoyed with families, I have goosebumps!:D" |
(2) Co-creation activities | years, son, daughter, inside, children, gift, eggs, surprise, made, name | "Here's a tiny artefact made by my 8-year-old daughter for school < 3" | ||||
(22) Brand contest | branch, app, respond, access, bank transfer, account, earn, reward, code, credentials | "Good morning, I am attaching my code so that you can register and enter the *** code. You will earn stars and medals and you can win great prizes, such as food packs, entrance tickets to cinemas, gyms, restaurants and attractions and even collectible LPs? Remember that those who already have the app but have not entered a friend code can reinstall it to be allowed to log in?" | ||||
(2) Positivity toward brands | This cluster gathers consumers’ comments containing extremely positive emotions toward the brands and/or their activities, for example feelings of appraisal, gratitude and admiration of consumes for brands and their representatives | Batra et al. (2012) | 0.644 (20.437) | (7) Admiration | suspension, excellent, gentle, united, grand, ennio, need, employees, very good, courage | "Great Director?! You're one in a million, and one of us!" |
(21) Wishes | great, good, celebration, happy, greetings, hello, good wishes, recovery, Sunday, | "Good wishes to all the dads all over the world!" | ||||
(28) Gratitude | thank, teacher, proud, energetic, gesture, immense, extraordinary, honor, human, reconversion | "About thirty years of continuous work, to date there is no bank in Italy closest to the needs of companies. Personally we can only thank them, a very long partnership that has allowed us to grow by navigating in any weather conditions … Thanks:”-)" | ||||
(3) Brand referral | This cluster includes comments expressing varying consumer's instantiations of referral regarding the brand and/or its value proposition, not necessarily directly related to what the brand did amidst the pandemic, for example in the form of voluntary reviews of products and services advertised via social media | Shan and King (2015) | − 1.119 (16.691) | (3) Fashion enthusiasm | shoe, fashion, dress, perfection, clothes, collect, glasses, adorable, amazing, collection | "Adorable colours! I would like to buy the yellow top " |
(14) Appraisal | tipo, punto, panda, finally, version, thousand, oil, satisfied, cross, gpl | "Feel free not to believe it, but my natural power model has covered ONLY one million km, with only ordinary maintenance! < 3 < 3 < 3" | ||||
(20) Praise for food | great, idea, yummy, taste, jar, cereals, small, smooth, rocher, mango | "I tasted the mango and maracuja one, simply delicious. I'll taste the blueberry and cranberry one as well!" | ||||
(36) Automotive enthusiasm | aston, martin, car, rear, grill, design, sports car, supercar, iconic, colour | "I love it! The rear looks like a Corvette, while the back is a mixture of Aston Martin and Porsche" | ||||
(37) Praise for taste | always, choco, cocco, family, all, chosen, smile, take, best, nice | "Too good…. I always keep the "mini" in my pocket…for moments of weakness, not for gluttony…. yeah, no-one could believe it!:P" | ||||
(4) Negativity towards brands | This cluster gathers all those topics that manifest varying degrees of consumers’ dissatisfaction with brands, expressed through online complaints, overt criticism, and verbal protests. For instance, consumers use the brand’s social media page to complain about service failures, product malfunctioning or customer care inadequacies, but also to perform anti-brand activism and other forms of consumer resistance | Zarantonello et al. (2018) | − 1.586 (18.219) | (4) Indignation | car, enzo, shame, leclerc, bignotto, vettel, mercedes, pilot, track, drive | "Unwatchable …. you have the best pilots, and you give them the worse car … zero evolution, zero updates … simply disgusting and embarrassing … F. go away, you are destroying a myth …" |
(9) Customer care | question, service, why, write, read, understand, ask, work, say, not at all, | "Too bad you do not answer the phones and do not call back !!! so I would like to know how to make an appointment with you. People have deadlines and you have to respect them, because delays do not affect the bank, but US !!! Answer or listen to the messages left on your answering machine" | ||||
(13) Made in Italy | pasta, wheat, glutin, eat, lactose, italian, jars, packaging, provenance, slice, | "I invite all consumers to stop buying XX pasta made with Canadian grain full of glyphosate which causes many very serious diseases, most of them fatal. The wording "only Italian wheat" on packages is a scam!" | ||||
(15) Complaints (telecom) | fiber, modem, bill, connection, giga, fixed, bill, adsl, activation, unlimited | "Dear XX, your contribution would be that our fixed and mobile lines worked properly but since yesterday morning I have had no fixed line, and the mobile line is not even reliable." | ||||
(17) Service Failure (telecom) | private, solve, feedback, contacted, report, certified email, bad, reply, problem | "After sending you the requested data, I have been waiting for an answer for about two weeks" | ||||
(26) Service Failure (website) | can, must, none, purchase, having, problems, lament, pro, interested, discount | "Speaking of TIME, FOR HEAVEN’S SAKE INCREASE THE WEB SESSION TIMEOUT! INCREASE THE WEB SESSION TIMEOUT! INCREASE THE WEB SESSION TIMEOUT! INCREASE THE WEB SESSION TIMEOUT! INCREASE THE WEB SESSION TIMEOUT!" | ||||
(27) Customer care (bank) | need, rest, comments, few, not even, negative, atm, instead, advance, take off | "If you manage not to permanently close the branches, customers could also access them. See for example Corso Moncalieri and Via Val della Torre in Turin." | ||||
(31) Anti-brand activism | immediately, sanpaolo, want, carbon, fossils, finance, sources, stop, climate, #dirtyalliance | "XX, if you want to defend the climate you must immediately stop financing coal and fossil fuels! We do not want #dirtalliance Renounce to finance Adani and the controversial project to exploit a coal field in Australia, where millions of animals and entire forests have been engulfed by fires." | ||||
(32) Service Failure (bank) | card, credit, info, step, site, web, canon, order, home, reload | "The app never recognizes the payment QR code…and I have an Iphone pro, not a low-quality smartphone" | ||||
(35) Complaints (automotive) | time, issues, loose, price, old, capital, damage, errors, guaranteed, system | "Speaking of injustice … even your prices don't allow people with lower incomes to afford them? I don't think your prices are justified by their production costs. Your brand benefits from social differences (brands as status symbols) and justifies this system by making your stuff accessible only to people with higher incomes. It would be easy for your business to produce a cheaper line and make it more accessible." | ||||
(5) Mixed feelings | This cluster contains polarized social media users’ reactions to social media communication that brands undertook during the pandemic, ranging from expression of deep admiration, skepticism, perceived opportunism and even disgust | Schneider and Schwarz (2017) | − 2.223 (22.215) | (10) Spot | ad, masterpiece, amazing, spot, proud, bad, people, chaplin, rich, evil | "A memorable speech by Charlie Chaplin was used to advertise the coffee … it's really embarrassing …" |
(11) Safety concerns | part, work, employees, proud, important, must, central, excellence, yet | "A proper cleaning of the branches.. Sanitization in the rooms of firms????? No??? !!!!" | ||||
(29) Backlashes | moment, south, masks, lombardy, hospitals, govern, shame, suspend, hard, difficult | "But do you know the situation in Lombardy ????? Do you fucking read how many infections there are in Lombardy? In my opinion you don't even look, and talk as usual just to let your breath out … Even in difficult moments" | ||||
(6) Nostalgia | This cluster gathers consumers' expression of nostalgia for the brand and/or its activities and contexts | Heinberg et al. (2020) | − 0.106 (18.773) | (16) Memories | memories, santorini, magnifique, delicious, greece, fabulous, unforgettable, islands, go back, | "I visited both of them…. Palma and Santorini…Santorini has had a special place in my heart since then < 3" |
(25) Missing travels | miss, sea, balcony, hikes, buffet, restaurant, fun, relax, shows, pleasure | "I miss everything about the cruise!!!The halls, the swimming pools, the restaurants, the samsara, the parties, I miss the love of the crews of every sector!!!" | ||||
(34) Desire for normalcy | hope, soon, can't wait, restart, go back, end, marvelous, miss, #restarttogether, jump on | "I hope it will happen very soon:) we all need to start again.. and to find you on board again " | ||||
(7) Covid complaints | This cluster gathers topics that manifest varying degrees of consumers' dissatisfaction with brands, but contextual to Covid-19 | – | − 2.02 (19.607) | (33) Edginess | now, less, area, cases, continue, desire, availability, countermeasures, coronavirus, none | "Dear friends of XX, I have been in the red zone since February, the very first town that was closed. I have not received the infamous message to have the possibility to use the giga in an unlimited way!" |
(5) Travel restraints | cruising, holiday, book, hope, miss, decision, cancelled, news, anxiety, positive | "Stop it! Stop and be done with it, you're making fools of yourselves. You and your decisions really let me down. I won't travel with you any more in the future! You are irresponsible!" | ||||
(8) Uncertainty | certainly, must, passengers, possibility, bad, remain, next, certainty, host, situation | "My wife and I have to go on a cruise in March. Oman and Jordan have already issued orders not to accept Italians. If other states forbid us from disembarking, do we risk spending all the time on the ship?" | ||||
(19) Refunds | know, voucher, client, penalties, closed, refund, avoid, date, distance, right | "Manager, you should allow your clients to freely choose between refunds and vouchers according to their needs " | ||||
(38) Booking issues | smeralda, booking, grandiosa, caribbean, leave, september, august, emirates, may, route | "I have a reservation for the April departure with XX: even if the departure from Savona is confirmed, will there be any changes of itinerary considering the closures of France and Spain?" | ||||
(40) Dissatisfaction | exchange, again, happen, visit, fear, treated, expire, decline, visibility, suggest | "From this emergency I have truly understood that you are highly disorganized, and that information does not pass correctly between you. As soon as everything ends, I will do the subrogation of the mortgages, just to have no more to deal with you! " | ||||
Excluded | This cluster gathers consumer topics which were not theoretically relevant (i.e., topic 18) or that were deemed not interpretable by the research team | – | – | (18) San Remo festival | song, great, congrats, diodato, festival, gabbani, sanremo, winner, amadeus, deserve | "The true winner of the Festival, XX the best! Great song and great music, enjoyable, joyful, nice, bravo XX!:):):)" |
(6) undefined | something, think, should, sure, doubt, worse, was, guess, worst, sorry | – | ||||
(12) undefined | things, world, god, sure, mister, bless, fan, have, world, learn | – | ||||
(23) undefined | well, want, see, words, maybe, go, thought, moving, one, hear | – | ||||
(24) undefined | never, better, this way, that is, maybe, hard, sorry, late, change, unfortunately | – | ||||
(30) undefined | people, do, say, really, by the way, understand, page, other, pay, look for | – | ||||
(39) undefined | same, thing, more, value, suggestion, happening, say, sky, guys, imagine | – |
*Translated from Italian to English.
References
Aaker, J., K.D. Vohs, and C. Mogilner. 2010. Nonprofits are seen as warm and for-profits as competent: Firm perceptions matter. Journal of Consumer Research 37 (2): 224–237.
Abitbol, A., and M.S. VanDyke. 2023. Analyzing the nature of self-transcendent emotional elicitors in corporate social advocacy messages. Public Relations Review 49 (4): 102364.
Abitbol, A., N. Lee, S. Lee, and T. Seltzer. 2018. # RaceTogether: Starbucks’ attempt to discuss race in America and its impact on company reputation and employees. Public Relations Journal 12 (1): 1–28.
Aboelenien, A., and Nguyen, C. M. 2023. From Dr. Seuss to Barbie’s cancellation: brand’s institutional work in response to changed market logics. Journal of Brand Management 1–18.
Afifi, W.A., and S. Metts. 1998. Characteristics and consequences of expectation violations in close relationships. Journal of Social and Personal Relationships 15 (3): 365–392.
Aggarwal, P., and A.L. McGill. 2012. When brands seem human, do humans act like brands? Automatic behavioral priming effects of brand anthropomorphism. Journal of Consumer Research 39 (2): 307–323.
Ahmad, F., F. Guzmán, and B. Kidwell. 2022. Effective messaging strategies to increase brand love for sociopolitical activist brands. Journal of Business Research 151: 609–622.
Aiken, L.S., and S.G. West. 1991. Multiple Regression: Testing and Interpreting Interactions. Newbury Park: Sage Publications.
Airoldi, E.M., D.M. Blei, E.A. Erosheva, and S.E. Fienberg. 2015. Introduction to mixed membership models and methods. Handbook of Mixed Membership Models and Their Applications 100: 3–14.
Akpinar, E., and J. Berger. 2017. Valuable virality. Journal of Marketing Research 54 (2): 318–330.
Alalwan, A.A. 2018. Investigating the impact of social media advertising features on customer purchase intention. International Journal of Information Management 42: 65–77.
Alalwan, A.A., N.P. Rana, Y.K. Dwivedi, and R. Algharabat. 2017. Social media in marketing: A review and analysis of the existing literature. Telematics and Informatics 34 (7): 1177–1190.
Althuizen, N. 2021. Revisiting Berlyne’s inverted U-shape relationship between complexity and liking: The role of effort, arousal, and status in the appreciation of product design aesthetics. Psychology & Marketing 38 (3): 481–503.
Aranda, A.M., K. Sele, H. Etchanchu, J.Y. Guyt, and E. Vaara. 2021. From big data to rich theory: Integrating critical discourse analysis with structural topic modeling. European Management Review 18 (3): 197–214.
Araujo, T., P. Neijens, and R. Vliegenthart. 2015. What motivates consumers to re-tweet brand content? The impact of information, emotion, and traceability on pass-along behavior. Journal of Advertising Research 55 (3): 284–295.
Ashley, C., and T. Tuten. 2015. Creative strategies in social media marketing: An exploratory study of branded social content and consumer engagement. Psychology & Marketing 32 (1): 15–27.
Austin, L.L., and B.M. Gaither. 2016. Examining public response to corporate social initiative types: A quantitative content analysis of Coca-Cola’s social media. Social Marketing Quarterly 22 (4): 290–306.
Austin, L., B. Gaither, and T.K. Gaither. 2019. Corporate social advocacy as public interest communications: Exploring perceptions of corporate involvement in controversial social-political issues. The Journal of Public Interest Communications 3 (2): 3–31.
Banks, G.C., H.M. Woznyj, R.S. Wesslen, and R.L. Ross. 2018. A review of best practice recommendations for text analysis in R (and a user-friendly app). Journal of Business and Psychology 33 (4): 445–459.
Barasch, A., E.E. Levine, J.Z. Berman, and D.A. Small. 2014. Selfish or selfless? On the signal value of emotion in altruistic behavior. Journal of Personality and Social Psychology 107 (3): 393.
Barger, V., J.W. Peltier, and D.E. Schultz. 2016. Social media and consumer engagement: A review and research agenda. Journal of Research in Interactive Marketing 10 (4): 268–287.
Barry, T.E., and D.J. Howard. 1990. A review and critique of the hierarchy of effects in advertising. International Journal of Advertising 9 (2): 121–135.
Batista, J.M., L.S. Barros, F.V. Peixoto, and D. Botelho. 2022. Sarcastic or assertive: How should brands reply to consumers’ uncivil comments on social media in the context of brand activism? Journal of Interactive Marketing 57 (1): 141–158.
Batra, R., A. Ahuvia, and R.P. Bagozzi. 2012. Brand love. Journal of Marketing 76 (2): 1–16.
Beckers, S.F., J. Van Doorn, and P.C. Verhoef. 2018. Good, better, engaged? The effect of company-initiated customer engagement behavior on shareholder value. Journal of the Academy of Marketing Science 46 (3): 366–383.
Berger, J., and K.L. Milkman. 2012. What makes online content viral? Journal of Marketing Research 49 (2): 192–205.
Bhagwat, Y., N.L. Warren, J.T. Beck, and G.F. Watson IV. 2020. Corporate sociopolitical activism and firm value. Journal of Marketing 84 (5): 1–21.
Blei, D.M. 2012. Probabilistic topic models. Communications of the ACM 55 (4): 77–84.
Blei, D., and J. Lafferty. 2006. Correlated topic models. Advances in Neural Information Processing Systems 18: 147–155.
Blei, D.M., A.Y. Ng, and M.I. Jordan. 2003. Latent Dirichlet allocation. Journal of Machine Learning Research 3: 993–1022.
Bowen, M., X.H. Wen, and S. Kim. 2022. A lure or a turn-off: Social media reactions to business model innovation announcements. Marketing Letters 34: 1–21.
Boyd, D. 2010. Social network sites as networked publics: Affordances, dynamics, and implications. In A networked self (pp. 47–66). Routledge.
Brady, W.J., J.A. Wills, J.T. Jost, J.A. Tucker, and J.J. Van Bavel. 2017. Emotion shapes the diffusion of moralized content in social networks. Proceedings of the National Academy of Sciences 114 (28): 7313–7318.
Brady, W.J., M.J. Crockett, and J.J. Van Bavel. 2020. The MAD model of moral contagion: The role of motivation, attention, and design in the spread of moralized content online. Perspectives on Psychological Science 15 (4): 978–1010.
Bravo, O.S.A.C., and J. Lee. 2020. The mediating effects of message agreement on millennials’ response to advocacy advertising. Journal of Marketing Communications 26 (8): 856–873.
Brodie, R.J., A. Ilic, B. Juric, and L. Hollebeek. 2013. Consumer engagement in a virtual brand community: An exploratory analysis. Journal of Business Research 66 (1): 105–114.
Brown, S., R.V. Kozinets, and J.F. Sherry Jr. 2003. Teaching old brands new tricks: Retro branding and the revival of brand meaning. Journal of Marketing 67 (3): 19–33.
Burgoon, J.K. 1993. Interpersonal expectations, expectancy violations, and emotional communication. Journal of Language and Social Psychology 12 (1–2): 30–48.
Burgoon, M., P.V. Denning, and L. Roberts. 2002. Language expectancy theory. In The persuasion handbook: Developments in theory and practice, ed. J.P. Dillard and M. Pfau, 117–137. Thousand Oaks: Sage.
Cao, D., M. Meadows, D. Wong, and S. Xia. 2021. Understanding consumers’ social media engagement behaviour: An examination of the moderation effect of social media context. Journal of Business Research 122: 835–846.
Chatzidakis, A., and J. Littler. 2022. An anatomy of carewashing: Corporate branding and the commodification of care during Covid-19. International Journal of Cultural Studies 25 (3–4): 268–286.
Choi, S., S.Q. Liu, and A.S. Mattila. 2019. “How may i help you?” Says a robot: Examining language styles in the service encounter. International Journal of Hospitality Management 82: 32–38.
Choi, Y., Jung, Y., and Myaeng, S. H. 2010. Identifying controversial issues and their sub-topics in news articles. Pacific-Asia Workshop on Intelligence and Security Informatics: 140–153.
Chu, S.C., H. Kim, and Y. Kim. 2023. When brands get real: The role of authenticity and electronic word-of-mouth in shaping consumer response to brands taking a stand. International Journal of Advertising 42 (6): 1037–1064.
Coombs, W.T., and S.J. Holladay. 2018. Social issue qua wicked problems: The role of strategic communication in social issues management. Journal of Communication Management 22 (1): 79–95.
Cosenza, T.R., M.R. Solomon, and W.S. Kwon. 2015. Credibility in the blogosphere: A study of measurement and influence of wine blogs as an information source. Journal of Consumer Behaviour 14 (2): 71–91.
Cuddy, A.J., S.T. Fiske, and P. Glick. 2008. Warmth and competence as universal dimensions of social perception: The perception content model and the BIAS map. Advances in Experimental Social Psychology 40: 61–149.
de Oliveira Santini, F., W.J. Ladeira, D.C. Pinto, M.M. Herter, C.H. Sampaio, and B.J. Babin. 2020. Customer engagement in social media: A framework and meta-analysis. Journal of the Academy of Marketing Science 48 (6): 1211–1228.
de Vries, L., S. Gensler, and P.S. Leeflang. 2012. Popularity of brand posts on brand fan pages: An investigation of the effects of social media marketing. Journal of Interactive Marketing 26 (2): 83–91.
Dehler-Holland, J., M. Okoh, and D. Keles. 2022. Assessing technology legitimacy with topic models and sentiment analysis–The case of wind power in Germany. Technological Forecasting and Social Change 175: 121354.
Deng, Q., Y. Wang, M. Rod, and S. Ji. 2021. Speak to head and heart: The effects of linguistic features on B2B brand engagement on social media. Industrial Marketing Management 99: 1–15.
Di Domenico, G., J. Sit, A. Ishizaka, and D. Nunan. 2021. Fake news, social media and marketing: A systematic review. Journal of Business Research 124: 329–341.
DiMaggio, P., M. Nag, and D. Blei. 2013. Exploiting affinities between topic modeling and the sociological perspective on culture: Application to newspaper coverage of US government arts funding. Poetics 41 (6): 570–606.
Dineva, D., J. Breitsohl, B. Garrod, and P. Megicks. 2020. Consumer responses to conflict-management strategies on non-profit social media fan pages. Journal of Interactive Marketing 52: 118–136.
DiRusso, C., C. Buckley, P. Diddi, F.E. Dardis, M. Vafeiadis, and N. Eng. 2022. Designing effective corporate social advocacy campaigns using valence, arousal, and issue salience. Public Relations Review 48 (3): 102207.
Dolan, R., J. Conduit, C. Frethey-Bentham, J. Fahy, and S. Goodman. 2019. Social media engagement behavior: A framework for engaging customers through social media content. European Journal of Marketing 53 (10): 2213–2243.
Du, S., C.B. Bhattacharya, and S. Sen. 2010. Maximizing business returns to corporate social responsibility (CSR): The role of CSR communication. International Journal of Management Reviews 12 (1): 8–19.
Dubois, D., D.D. Rucker, and A.D. Galinsky. 2016. Dynamics of communicator and audience power: The persuasiveness of competence versus warmth. Journal of Consumer Research 43 (1): 68–85.
Dwivedi, Y.K., K.K. Kapoor, and H. Chen. 2015. Social media marketing and advertising. The Marketing Review 15 (3): 289–309.
Eigenraam, A.W., J. Eelen, A. Van Lin, and P.W. Verlegh. 2018. A consumer-based taxonomy of digital customer engagement practices. Journal of Interactive Marketing 44: 102–121.
Eigenraam, A.W., J. Eelen, and P.W. Verlegh. 2021. Let me entertain you? The importance of authenticity in online customer engagement. Journal of Interactive Marketing 54: 53–68.
Ertimur, B., and G. Coskuner-Balli. 2021. Brands expressing compassion and care through advertising. Journal of Advertising 50 (3): 230–239.
Feng, Y., H. Chen, and L. He. 2019. Consumer responses to femvertising: A data-mining case of Dove’s “Campaign for Real Beauty” on YouTube. Journal of Advertising 48 (3): 292–301.
Feng, Y., H. Chen, and H.Y.A. Ahn. 2021. How consumers react to woke advertising: Methodological triangulation based on social media data and self-report data. Journal of Research in Interactive Marketing 15 (4): 529–548.
Fiske, S.T., A.J.C. Cuddy, P. Glick, and J. Xu. 2002. A model of (often mixed) perception content: Competence and warmth respectively follow from perceived status and competition. Journal of Personality and Social Psychology 82: 878–902.
Folse, J.A.G., M. Porter III., M.B. Godbole, and K.E. Reynolds. 2016. The effects of negatively valenced emotional expressions in online reviews on the reviewer, the review, and the product. Psychology & Marketing 33 (9): 747–760.
Förster, J., J. Marguc, and M. Gillebaart. 2010. Novelty categorization theory. Social and Personality Psychology Compass 4 (9): 736–755.
Fournier, S. 1998. Consumers and their brands: Developing relationship theory in consumer research. Journal of Consumer Research 24 (4): 343–373.
Fresneda, J.E., T.A. Burnham, and C.H. Hill. 2021. Structural topic modelling segmentation: A segmentation method combining latent content and customer context. Journal of Marketing Management 37 (7–8): 792–812.
Gambetti, R.C., and S. Biraghi. 2023. Branded activism: Navigating the tension between culture and market in social media. Futures 145: 103080.
Garimella, K., G.D.F. Morales, A. Gionis, and M. Mathioudakis. 2018. Quantifying controversy on social media. ACM Transactions on Social Computing 1 (1): 1–27.
Gershon, R., and C. Cryder. 2018. Goods donations increase charitable credit for low-warmth donors. Journal of Consumer Research 45 (2): 451–469.
Grazzini, L., D. Acuti, and G. Aiello. 2021. Solving the puzzle of sustainable fashion consumption: The role of consumers’ implicit attitudes and perceived warmth. Journal of Cleaner Production 287: 125579.
Grewal, D., A.L. Roggeveen, R. Sisodia, and J. Nordfält. 2017. Enhancing customer engagement through consciousness. Journal of Retailing 93 (1): 55–64.
Grimmer, J., and B.M. Stewart. 2013. Text as data: The promise and pitfalls of automatic content analysis methods for political texts. Political Analysis 21 (3): 267–297.
Guha, M., and Korschun, D. 2023. Peer effects on brand activism: evidence from brand and user chatter on Twitter. Journal of Brand Management, 1–15.
Gummerus, J., V. Liljander, E. Weman, and M. Pihlström. 2012. Customer engagement in a Facebook brand community. Management Research Review 35 (9): 857–877.
Gvili, Y., and S. Levy. 2018. Consumer engagement with eWOM on social media: The role of social capital. Online Information Review 42 (4): 482–505.
Hang, H., L. Aroean, and Z. Chen. 2020. Building emotional attachment during COVID-19. Annals of Tourism Research 83: 103006.
Hannigan, T.R., R.F. Haans, K. Vakili, H. Tchalian, V.L. Glaser, M.S. Wang, and P.D. Jennings. 2019. Topic modeling in management research: Rendering new theory from textual data. Academy of Management Annals 13 (2): 586–632.
Heinberg, M., C.S. Katsikeas, H.E. Ozkaya, and M. Taube. 2020. How nostalgic brand positioning shapes brand equity: Differences between emerging and developed markets. Journal of the Academy of Marketing Science 48 (5): 869–890.
Heinonen, K. 2011. Consumer activity in social media: Managerial approaches to consumers’ social media behavior. Journal of Consumer Behaviour 10 (6): 356–364.
Hilton, J.L., J.G. Klein, and W. von Hippel. 1991. Attention allocation and impression formation. Personality and Social Psychology Bulletin 17: 548–559.
Hollebeek, L. D., Sharma, T. G., Pandey, R., Sanyal, P., and Clark, M. K. 2021. Fifteen years of customer engagement research: a bibliometric and network analysis. Journal of Product and Brand Management. In press
Hovland, C.I., I.L. Janis, and H.H. Kelley. 1953. Communication and persuasion. Yale University Press.
Hu, N., T. Zhang, B. Gao, and I. Bose. 2019. What do hotel customers complain about? Text analysis using structural topic model. Tourism Management 72: 417–426.
Humphreys, A., and R.J.H. Wang. 2018. Automated text analysis for consumer research. Journal of Consumer Research 44 (6): 1274–1306.
Hung, K.H., and S.Y. Li. 2007. The influence of eWOM on virtual consumer communities: Social capital, consumer learning, and behavioral outcomes. Journal of Advertising Research 47 (4): 485–495.
Hydock, C., N. Paharia, and T.J. Weber. 2019. The consumer response to corporate political advocacy: A review and future directions. Customer Needs and Solutions 6 (3): 76–83.
Hydock, C., N. Paharia, and S. Blair. 2020. Should your brand pick a side? How market share determines the impact of corporate political advocacy. Journal of Marketing Research 57 (6): 1135–1151.
Iglesias, O., and N. Ind. 2020. Towards a theory of conscientious corporate brand co-creation: The next key challenge in brand management. Journal of Brand Management 27 (6): 710–720.
Ismagilova, E., E. Slade, N.P. Rana, and Y.K. Dwivedi. 2020. The effect of characteristics of source credibility on consumer behaviour: A meta-analysis. Journal of Retailing and Consumer Services 53: 101736.
James G, Witten D, Hastie T, and Tibshirani, R. 2017. An introduction to statistical learning: With applications in R. 1st ed. 2013, Corr. 7th printing 2017 edition. Springer.
Jensen, M.L., J.M. Averbeck, Z. Zhang, and K.B. Wright. 2013. Credibility of anonymous online product reviews: A language expectancy perspective. Journal of Management Information Systems 30 (1): 293–324.
Johnen, M., and O. Schnittka. 2019. When pushing back is good: The effectiveness of brand responses to social media complaints. Journal of the Academy of Marketing Science 47: 858–878.
Johnson, Z.S., Y.J. Lee, and M.T. Ashoori. 2018. Brand associations: The value of ability versus social responsibility depends on consumer goals. Journal of Brand Management 25 (1): 27–37.
Jungblut, M., and Johnen, M. 2021. When brands (don’t) take my stance: the ambiguous effectiveness of political brand communication. Communication Research 1–26
Kabadayi, S., and K. Price. 2014. Consumer – brand engagement on Facebook: Liking and commenting behaviors. Journal of Research in Interactive Marketing 8 (3): 203–223.
Kantar. 2021. Kantar BrandZTM Most Valuable Italian Brands, 6 May, http://www.kantar.com/campaigns/brandz/italy
Karpen, I.O., and J. Conduit. 2020. Engaging in times of COVID-19 and beyond: Theorizing customer engagement through different paradigmatic lenses. Journal of Service Management 31 (6): 1163–1174.
Kervyn, N., S.T. Fiske, and C. Malone. 2012. Brands as intentional agents framework: How perceived intentions and ability can map brand perception. Journal of Consumer Psychology 22 (2): 166–176.
Kervyn, N., S.T. Fiske, and C. Malone. 2022. Social perception of brands: Warmth and competence define images of both brands and social groups. Consumer Psychology Review 5 (1): 51–68.
Kim, J.E., and K.K. Johnson. 2013. The impact of moral emotions on cause-related marketing campaigns: A cross-cultural examination. Journal of Business Ethics 112: 79–90.
Kim, D.H., L. Spiller, and M. Hettche. 2015. Analyzing media types and content orientations in Facebook for global brands. Journal of Research in Interactive Marketing 9 (1): 4–30.
Kim, S., and H. Rim. 2019. The role of public skepticism and distrust in the process of CSR communication. International Journal of Business Communication 1–21.
Klostermann, J., C. Hydock, and R. Decker. 2021. The effect of corporate political advocacy on brand perception: An event study analysis. Journal of Product and Brand Management 31 (5): 780–797.
Korschun, D. and Smith, N. C. 2018. Companies can’t avoid politics and shouldn’t try to. Harvard Business Review (Digital Article), https://hbr.org/2018/03/companies-cant-avoid-politics-and-shouldnt-try-to
Korschun, D., Rafieian, H., Aggarwal, A., and Swain, S. D. 2019. Taking a stand: Consumer responses when companies get (or don’t get) political. Available at SSRN 2806476.
Kralj Novak, P., J. Smailović, B. Sluban, and I. Mozetič. 2015. Sentiment of Emojis. PLoS ONE 10 (12): 1–22.
Krippendorff, K. 2018. Content analysis: An introduction to its methodology. Sage publications.
Kronrod, A., A. Grinstein, and L. Wathieu. 2012a. Enjoy! Hedonic consumption and compliance with assertive messages. Journal of Consumer Research 39 (1): 51–61.
Kronrod, A., A. Grinstein, and L. Wathieu. 2012b. Go green! Should environmental messages be so assertive? Journal of Marketing 76 (1): 95–102.
Kübler, R.V., A. Colicev, and K.H. Pauwels. 2020. Social media’s impact on the consumer mindset: When to use which sentiment extraction tool? Journal of Interactive Marketing 50: 136–155.
Kumar, A., R. Bezawada, R. Rishika, R. Janakiraman, and P.K. Kannan. 2016. From social to sale: The effects of firm-generated content in social media on customer behavior. Journal of Marketing 80 (1): 7–25.
Lang, A. 2000. The limited capacity model of mediated message processing. Journal of Communication 50 (1): 46–70.
Leak, R.L., O.P. Woodham, and K.R. McNeil. 2015. Speaking candidly: How managers’ political stances affect consumers’ brand attitudes. Journal of Product & Brand Management 24 (5): 494–503.
Lee, J.K. 2021. Emotional expressions and brand status. Journal of Marketing Research 58 (6): 1178–1196.
Lee, S.Y., and S. Chung. 2023. Publics’ views of corporate social advocacy initiatives: Exploring prior issue stance, attitude toward a company, and news credibility. Management Communication Quarterly 37 (2): 281–309.
Lee, Y.H., and C. Mason. 1999. Responses to information incongruency in advertising: The role of expectancy, relevancy, and humor. Journal of Consumer Research 26 (2): 156–169.
Lee, C.H., and H. Yu. 2020. The impact of language on retweeting during acute natural disasters: Uncertainty reduction and language expectancy perspectives. Industrial Management and Data Systems 120 (8): 1501–1519.
Lee, D., K. Hosanagar, and H.S. Nair. 2018. Advertising content and consumer engagement on social media: Evidence from Facebook. Management Science 64 (11): 5105–5460.
Li, J.Y., J.K. Kim, and K. Alharbi. 2022. Exploring the role of issue involvement and brand attachment in shaping consumer response toward corporate social advocacy (CSA) initiatives: The case of Nike’s Colin Kaepernick campaign. International Journal of Advertising 41 (2): 233–257.
Liadeli, G., F. Sotgiu, and P.W. Verlegh. 2023. A meta-analysis of the effects of brands’ owned social media on social media engagement and sales. Journal of Marketing 87 (3): 406–427.
Lindstedt, N.C. 2019. Structural topic modeling for social scientists: A brief case study with social movement studies literature, 2005–2017. Social Currents 6 (4): 307–318.
Lou, C., and S. Yuan. 2019. Influencer marketing: How message value and credibility affect consumer trust of branded content on social media. Journal of Interactive Advertising 19 (1): 58–73.
Love, E., Sekhon, T., and Salinas, T. C. 2022. Do well, do good, and know your audience: the double-edged sword of values-based CSR communication. Journal of Brand Management: 1–17.
Lucisano P, and Piemontese, M.E. 1998. GULPEASE: a formula to predict readability of texts written in Italian language. In School and Town, La nuova Italia, Brescia (Italy), pp. 3–31.
Luo, X., M.A. Wiles, and S. Raithel. 2013. Make the most of a polarizing brand. Harvard Business Review. 91 (11): 29–31.
Mafael, A., S.A. Gottschalk, and H. Kreis. 2016. Examining biased assimilation of brand-related online reviews. Journal of Interactive Marketing 36: 91–106.
Magee, R.G. 2022. Understanding worldview beliefs to allay skepticism toward CSR advertising. Journal of Brand Management 29: 1–18.
Mangiò, F., G. Pedeliento, and D. Andreini. 2021. Branding rhetoric in times of a global pandemic: A text-mining analysis. Journal of Advertising 50 (3): 240–252.
Maryott, K. 2021. Polarization – why, for brands, it’s a good thing. WARC. January 22, https://www.warc.com/newsandopinion/opinion/polarization--why-for-brands-its-a-good-thing/en-gb/4022
McShane, L., E. Pancer, M. Poole, and Q. Deng. 2021. Emoji, playfulness, and brand engagement on twitter. Journal of Interactive Marketing 53 (1): 96–110.
Merchant, A., and G.M. Rose. 2013. Effects of advertising-evoked vicarious nostalgia on brand heritage. Journal of Business Research 66 (12): 2619–2625.
Middleton, K., and S. Turnbull. 2021. How advertising got ‘woke’: The institutional role of advertising in the emergence of gender progressive market logics and practices. Marketing Theory 21 (4): 561–578.
Milfeld, T., and D.J. Flint. 2020. When brands take a stand: The nature of consumers’ polarized reactions to social narrative videos. Journal of Product and Brand Management 30 (4): 532–548.
Mirzaei, A., D.C. Wilkie, and H. Siuki. 2022. Woke brand activism authenticity or the lack of it. Journal of Business Research 139: 1–12.
Mohammad, S.M., and P.D. Turney. 2013. Crowdsourcing a word–emotion association lexicon. Computational Intelligence 29 (3): 436–465.
Moorman, C. 2020. Commentary: Brand activism in a political world. Journal of Public Policy and Marketing 39 (4): 388–392.
Moro, S., G. Pires, P. Rita, and P. Cortez. 2020. A cross-cultural case study of consumers’ communications about a new technological product. Journal of Business Research 121: 438–447.
Mukherjee, S., and N. Althuizen. 2020. Brand activism: Does courting controversy help or hurt a brand? International Journal of Research in Marketing 37 (4): 772–788.
Muntinga, D.G., M. Moorman, and E.G. Smit. 2011. Introducing COBRAs: Exploring motivations for brand-related social media use. International Journal of Advertising 30 (1): 13–46.
Nelson, S.K., K. Layous, S.W. Cole, and S. Lyubomirsky. 2016. Do unto others or treat yourself? The effects of prosocial and self-focused behavior on psychological flourishing. Emotion 16 (6): 850.
Nunes, J., and C., Ordanini, A., and Giambastiani, G. 2021. The Concept of Authenticity: What It Means to Consumers. Journal of Marketing 85 (4): 1–20.
Pancer, E., V. Chandler, M. Poole, and T.J. Noseworthy. 2019. How readability shapes social media engagement. Journal of Consumer Psychology 29 (2): 262–270.
Panigyrakis, G., A. Panopoulos, and E. Koronaki. 2020. All we have is words: Applying rhetoric to examine how social media marketing activities strengthen the connection between the brand and the self. International Journal of Advertising 39 (5): 699–718.
Parcha, J.M., and C.Y. Kingsley Westerman. 2020. How corporate social advocacy affects attitude change toward controversial social issues. Management Communication Quarterly 34 (3): 350–383.
Parhankangas, A., and M. Renko. 2017. Linguistic style and crowdfunding success among social and commercial entrepreneurs. Journal of Business Venturing 32 (2): 215–236.
Park, K. 2021. The mediating role of skepticism: How corporate social advocacy builds quality relationships with publics. Journal of Marketing Communications 28: 1–19.
Phillips, N., T.B. Lawrence, and C. Hardy. 2004. Discourse and institutions. Academy of Management Review 29 (4): 635–652.
Pimentel, P. C., Bassi-Suter, M., & Didonet, S. R. (2023). Brand activism as a marketing strategy: an integrative framework and research agenda. Journal of Brand Management: 1–23.
Pletikosa Cvijikj, I., and F. Michahelles. 2013. Online engagement factors on Facebook brand pages. Social Network Analysis and Mining 3 (4): 843–861.
Ramírez, S.A.O., C. Veloutsou, and A. Morgan-Thomas. 2019. I hate what you love: Brand polarization and negativity towards brands as an opportunity for brand management. Journal of Product and Brand Management 28 (5): 614–632.
Reisenbichler, M., and T. Reutterer. 2019. Topic modeling in marketing: Recent advances and research opportunities. Journal of Business Economics 89 (3): 327–356.
Ren, S., S. Karimi, A.B. Velázquez, and J. Cai. 2023. Endorsement effectiveness of different social media influencers: The moderating effect of brand competence and warmth. Journal of Business Research 156: 113476.
Rietveld, R., W. Van Dolen, M. Mazloom, and M. Worring. 2020. What you feel, is what you like influence of message appeals on customer engagement on Instagram. Journal of Interactive Marketing 49: 20–53.
Rim, H., Y. Lee, and S. Yoo. 2020. Polarized public opinion responding to corporate social advocacy: Social network analysis of boycotters and advocators. Public Relations Review 46 (2): 1–10.
Roberts, M.E., B.M. Stewart, D. Tingley, C. Lucas, J. Leder-Luis, S.K. Gadarian, and D.G. Rand. 2014. Structural topic models for open-ended survey responses. American Journal of Political Science 58 (4): 1064–1082.
Roberts, M.E., B.M. Stewart, and D. Tingley. 2019. Stm: An R package for structural topic models. Journal of Statistical Software 91: 1–40.
Rocklage, M.D., and R.H. Fazio. 2020. The enhancing versus backfiring effects of positive emotion in consumer reviews. Journal of Marketing Research 57 (2): 332–352.
Saxton, G.D., L. Gómez, Z. Ngoh, Y.P. Lin, and S. Dietrich. 2019. Do CSR messages resonate? Examining public reactions to firms’ CSR efforts on social media. Journal of Business Ethics 155 (2): 359–377.
Schaefers, T., T. Falk, A. Kumar, and J. Schamari. 2021. More of the same? Effects of volume and variety of social media brand engagement behavior. Journal of Business Research 135: 282–294.
Schmidt, H.J., N. Ind, F. Guzman, and E. Kennedy. 2021. Sociopolitical activist brands. Journal of Product and Brand Management 31 (1): 40–55.
Schmiedel, T., O. Müller, and J. Vom Brocke. 2019. Topic modeling as a strategy of inquiry in organizational research: A tutorial with an application example on organizational culture. Organizational Research Methods 22 (4): 941–968.
Schneiberg, M., and E.S. Clemens. 2006. The typical tools for the job: Research strategies in institutional analysis. Sociological Theory 24 (3): 195–227.
Schneider, I.K., and N. Schwarz. 2017. Mixed feelings: The case of ambivalence. Current Opinion in Behavioral Sciences 15: 39–45.
Shahbaznezhad, H., R. Dolan, and M. Rashidirad. 2021. The role of social media content format and platform in users’ engagement behavior. Journal of Interactive Marketing 53: 47–65.
Shan, Y., and K.W. King. 2015. The effects of interpersonal tie strength and subjective norms on consumers’ brand-related eWOM referral intentions. Journal of Interactive Advertising 15 (1): 16–27.
Shoenberger, H., E. Kim, and Y. Sun. 2021. Advertising during COVID-19: Exploring perceived brand message authenticity and potential psychological reactance. Journal of Advertising 50 (3): 253–261.
Sobande, F. 2019. Woke-washing:“Intersectional” femvertising and branding “woke” bravery. European Journal of Marketing 54 (11): 2723–2745.
Sobande, F. 2020. ‘We’re all in this together’: Commodified notions of connection, care and community in brand responses to COVID-19. European Journal of Cultural Studies 23 (6): 1033–1037.
Sprout Social. 2023. The Sprout Social Index, Edition XVII: Accelerate. August 28, https://sproutsocial.com/insights/data/index-accelerate/
Song, B., and Choi, M. 2023. Testing Publics’ Perceptions of Corporate Social Advocacy Messaging: Linking Organizational and Social Outcomes. International Journal of Strategic Communication, 1–20.
Stubbersfield, J. M., Dean, L. G., Sheikh, S., Laland, K. N., and Cross, C. P. 2019. Social transmission favours the ‘morally good’over the ‘merely arousing’. Palgrave Communications 5(1).
Suddaby, R., and R. Greenwood. 2005. Rhetorical strategies of legitimacy. Administrative Science Quarterly 50 (1): 35–67.
Sundar, S. S. (2008). The MAIN model: A heuristic approach to understanding technology effects on credibility, 73–100 MacArthur Foundation Digital Media and Learning Initiative
Swaminathan, V., A. Sorescu, J.B.E.M. Steenkamp, T.C.G. O’Guinn, and B. Schmitt. 2020. Branding in a hyperconnected world: Refocusing theories and rethinking boundaries. Journal of Marketing 84 (2): 24–46.
Swaminathan, V., H.A. Schwartz, R. Menezes, and S. Hill. 2022. The language of brands in social media: Using topic modeling on social media conversations to drive brand strategy. Journal of Interactive Marketing 57 (2): 255–277.
Swani, K., and Labrecque, L. I. 2020. Like, Comment, or Share? Self-presentation vs. brand relationships as drivers of social media engagement choices. Marketing Letters 31(2): 279–298.
Swani, K., and G.R. Milne. 2017. Evaluating Facebook brand content popularity for service versus goods offerings. Journal of Business Research 79: 123–133.
Tafesse, W., and A. Wien. 2018. Using message strategy to drive consumer behavioral engagement on social media. Journal of Consumer Marketing 35 (3): 241–253.
Taylor, C.R. 2020. Editorial: Advertising and COVID-19. International Journal of Advertising 39 (5): 587–589.
Tellis, G.J., D.J. MacInnis, S. Tirunillai, and Y. Zhang. 2019. What drives virality (sharing) of online digital content? The critical role of information, emotion, and brand prominence. Journal of Marketing 83 (4): 1–20.
Thomas, V.L., and K. Fowler. 2023. Examining the outcomes of influencer activism. Journal of Business Research 154: 113336.
Thompson, C.J., and A. Kumar. 2022. Analyzing the cultural contradictions of authenticity: Theoretical and managerial insights from the market logic of conscious capitalism. Journal of Marketing 86 (5): 21–41.
Tirunillai, S., and G.J. Tellis. 2014. Mining marketing meaning from online chatter: Strategic brand analysis of big data using latent Dirichlet allocation. Journal of Marketing Research 51 (4): 463–479.
Tromble, R. 2021. Where have all the data gone? A critical reflection on academic digital research in the post-API age. Social Media+ Society 7(1): 2056305121988929.
Tsai, W.H.S., and L.R. Men. 2013. Motivations and antecedents of consumer engagement with brand pages on social networking sites. Journal of Interactive Advertising 13 (2): 76–87.
Unnava, V., and A. Aravindakshan. 2021. How does consumer engagement evolve when brands post across multiple social media? Journal of the Academy of Marketing Science 49 (5): 864–881.
Vallaster, C., and S. von Wallpach. 2013. An online discursive inquiry into the social dynamics of multi-stakeholder brand meaning co-creation. Journal of Business Research 66 (9): 1505–1515.
Van Doorn, J., K.N. Lemon, V. Mittal, S. Nass, D. Pick, P. Pirner, and P.C. Verhoef. 2010. Customer engagement behavior: Theoretical foundations and research directions. Journal of Service Research 13 (3): 253–266.
Villarroel Ordenes, F., D. Grewal, S. Ludwig, K.D. Ruyter, D. Mahr, and M. Wetzels. 2019. Cutting through content clutter: How speech and image acts drive consumer sharing of social media brand messages. Journal of Consumer Research 45 (5): 988–1012.
Vlachvei, A., Notta, O., and Koronaki, E. 2021. Effects of content characteristics on stages of customer engagement in social media: investigating European wine brands. Journal of Research in Interactive Marketing. Vol. ahead-of-print. https://doi.org/10.1108/JRIM-12-2020-0275.
Voorveld, H.A. 2019. Brand communication in social media: A research agenda. Journal of Advertising 48 (1): 14–26.
Vredenburg, J., S. Kapitan, A. Spry, and J.A. Kemper. 2020. Brands taking a stand: Authentic brand activism or woke washing? Journal of Public Policy and Marketing 39 (4): 444–460.
Wang, S.W., and A.C. Scheinbaum. 2018. Enhancing brand credibility via celebrity endorsement: Trustworthiness trumps attractiveness and expertise. Journal of Advertising Research 58 (1): 16–32.
Wang, Y., M.S. Qin, X. Luo, and Y. Kou. 2022. Frontiers: How support for black lives matter impacts consumer responses on social media. Marketing Science: In press.
Wannow, S., Haupt, M., and Ohlwein, M. 2023. Is brand activism an emotional affair? The role of moral emotions in consumer responses to brand activism. Journal of Brand Management, 1–25.
Weber, T.J., J. Joireman, D.E. Sprott, and C. Hydock. 2023. Differential response to corporate political advocacy and corporate social responsibility: Implications for political polarization and radicalization. Journal of Public Policy & Marketing 42 (1): 74–93.
Weinzimmer, L.G., and C.A. Esken. 2016. Risky business: Taking a stand on social issues. Business Horizons 59 (3): 331–337.
Westerman, D., P.R. Spence, and B. Van Der Heide. 2014. Social media as information source: Recency of updates and credibility of information. Journal of Computer-Mediated Communication 19 (2): 171–183.
Wojciszke, B., R. Bazinska, and M. Jaworski. 1998. On the dominance of moral categories in impression formation. Personality and Social Psychology Bulletin 24 (12): 1251–1263.
Wu, L., Shen, H., Fan, A., and Mattila, A. S. 2017. The impact of language style on consumers′ reactions to online reviews. Tourism Management: 590–596.
Xie, C., R.P. Bagozzi, and K. Grønhaug. 2019. The impact of corporate social responsibility on consumer brand advocacy: The role of moral emotions, attitudes, and individual differences. Journal of Business Research 95: 514–530.
Yang, A., A. Saffer, and Y. Li. 2020. Managing stakeholder expectations in a politically polarized society: An expectation violation theory approach. Journal of International Crisis and Risk Communication Research 3 (2): 275–299.
Yang, J., P. Chuenterawong, and K. Pugdeethosapol. 2021. Speaking up on Black lives matter: A comparative study of consumer reactions toward brand and influencer-generated corporate social responsibility messages. Journal of Advertising 50 (5): 565–583.
York, J.G., S. Vedula, and M.J. Lenox. 2018. It’s not easy building green: The impact of public policy, private actors, and regional logics on voluntary standards adoption. Academy of Management Journal 61 (4): 1492–1523.
Zarantonello, L., S. Romani, S. Grappi, and M. Fetscherin. 2018. Trajectories of brand hate. Journal of Brand Management 25 (6): 549–560.
Zawisza, M., and C. Pittard. 2015. When do warmth and competence sell best? The “Golden quadrant” shifts as a function of congruity with the product type, targets’ individual differences, and advertising appeal type. Basic and Applied Social Psychology 37 (2): 131–141.
Zhao, P., Z. Ma, T. Gill, and C. Ranaweera. 2023. Social media sentiment polarization and its impact on product adoption. Marketing Letters 34: 497–512.
Zhou, X., Lou, C., and Huang, X. 2023. Transcendent Brand Activism Advertising: Explicating the Roles of Color and Message Framing in Advertising Effectiveness. Journal of Advertising, 1–17.
Acknowledgements
The authors wish to thank all participants for their involvement in this research and colleagues as well as attendees at the 15th Global Brand Conference 2022 (4th–6th May 2022, Sheffield Hallam University) for comments on earlier versions of this research. The empirical sections are partly based on data collected for the first author’s (unpublished) PhD dissertation at the University of Bergamo, XXXV graduate programme in Business and Law.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of Interest.
On behalf of all the authors, the corresponding author states that there is no conflict of interest.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Mangiò, F., Pedeliento, G., Andreini, D. et al. How persuasive is woke brand communication on social media? Evidence from a consumer engagement analysis on Facebook. J Brand Manag 31, 345–381 (2024). https://doi.org/10.1057/s41262-023-00347-4
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1057/s41262-023-00347-4