Introduction

Four well-developed research traditions in the branding literature have served as the springboard for new thinking about consumer–brand interactions. First, there is a rich stream of insights into emotions consumers attribute to brands; e.g., the attribution of love (Bagozzi et al., 2017; Batra et al., 2012; Malar et al., 2011), hate (Zarantonello et al., 2016), commitment and loyalty (Hwang and Kandampully, 2012; Tsai, 2011; Khamitov et al., 2019), attachment (Park et al., 2010), trust (Chaudhuri and Holbrook, 2001; Delgado-Ballester and Munuera-Aleman, 2001), and repulsion (Dessart and Cova, 2021). Second, scholars advocate for fostering positive—versus indifferent or negative—consumer–brand interactions from which these emotional and cognitive attributions toward brands emerge (e.g., Fetscherin et al, 2019). Among other benefits that can accrue, positive consumer–brand relationships reportedly lead users to influence others’ brand choices (e.g., Taillon et al., 2020; Wiedmann and Walter, 2021). Third, there is scholarly attention devoted to behaviors; i.e., how consumers behave toward brands (Dimitriu and Guesalanga, 2017; Francioni et al., 2021; Romani et al., 2015), and how brands behave in the marketplace (Gammoh et al., 2021; Moulard et al., 2016). Fourth, the most underdeveloped area of inquiry relates to examination of the influence exerted by macro-environmental forces such as psychosocial epidemics, sociopolitical developments—on user–brand relationships (see Lim et al., 2021a, 2021b, 2022). For instance, scholars have examined how branding is effected by consumers’ sensitivities about gender (Mahmoud et al., 2021), sustainability (e.g., Copeland and Bhaduri, 2020; Pengji et al., 2021), and corporate social responsibility (CSR, Muniz and Guzman, 2021). There are reports of consumer–brand relationships shaped by socioemotional epidemics such as loneliness (e.g., Loh et al., 2021; Snyder and Newman, 2019), or by economic forces such as recessions (e.g., Mark et al., 2016). More recently, scholars have reported on the influence of COVID-19 pandemic on perceived authenticity of brands (e.g., Shoenberger et al., 2021) and on consumption of luxury services (Lim et al., 2022a). These streams of research deserve further investigation. A glaring gap exists in the theoretical understanding of how and why the COVID-19 pandemic, a key macro-environmental shift, shaped consumers’ thoughts, emotions, and behaviors toward brands in the ways it did. While some opportunities for pandemic-triggered branding (e.g., Knowles et al., 2020) and brand messaging (e.g., Shoenberger et al., 2021) are discussed, the gap in current theory is substantive and consequential. Substantive because user–brand relationships are central to the theory and practice of marketing; consequential because these relationships are reportedly changing without attracting much in the way of investigations devoted to producing generalizable, theoretical insights. While scholars have examined the impact of the COVID-19 pandemic on franchising (Bretas and Alon, 2020), tourism (Lim, 2021b), on supply chain management (e.g., Schleper et al., 2021), and pricing and revenue management (e.g., Basak and Chu, 2021), branding theory currently lags.

This article reports findings from a two-stage study we conducted in Spring 2021, a year into COVID-19 epidemic-triggered mandates for social distancing and office/campus closures. The first study was led by two research questions: (a) in what way, if any, has the experience of living through the COVID-19 pandemic changed the way consumers think, feel, and do toward brands they personally purchase and consume, and (b) what do consumers attribute as the reason for these changes in cognitions, emotions, and behaviors toward brands. The initial exploration was conducted via focus groups. The purpose was to identify an ontology, or the conceptual building blocks of a grounded theory (based on Laudan, 1977). Anchored in consumer voices, we identified new constructs and measurement scales and developed a framework and hypotheses. In the second stage, we tested the hypotheses via a nationwide survey of consumers in June 2021 (n = 786).

As a result of presenting our findings and implications, we make the following contribution to future theory building efforts. First, rooted in the voices of consumers we heard during our first-stage exploratory study, we make the case that the COVID-19 pandemic: (a) triggered high levels of tiredness, sadness, disconnection among consumers and (b) they were challenged by rescheduling activities and managing their time in ways that altered their cognition, emotions, and behaviors toward brands. We report that the felt disruption triggered three types of resulting cognitions, emotions, and behaviors; i.e., heightened intentionality devoted to brand choice making, heightened concern for ethical behaviors of brand sponsors toward their employees, and engagement in proselytizing and evangelizing about brands. Second, while the brand evangelism construct has received scholarly attention during the pre-COVID-19 period (e.g., Becerra and Badrinarayanan, 2013; Hsu, 2019), we delineate the epistemology of the brand evangelism construct in the present and emerging context by identifying and testing its antecedent cognitive and emotional processes. We then present empirical evidence of a newly tested scale for its future assessment. These contributions are significant because the changes in consumers’ thinking, emotions, and behaviors during the COVID-19 pandemic are likely to linger beyond the immediate term. In this context, we term it as the deliberate and conscious engagement in convincing and converting others to one’s way of thinking because one believes one knows better and can help others improve their lives. Based on survey data we collected to test our grounded hypotheses, we present evidence of the reliability and construct validity of new constructs and measurement scales, and make the case for the structural coherence of our proposed framework. Finally, we discuss implications for future theory development efforts devoted to understanding consumer–brand relationships.

Conceptual background

The conceptual underpinnings of our study rest in the grounded theory we generated during our first-stage exploration; i.e., the hypotheses that drove the second stage of the study emerge from voices of consumers and not extant theory (aligned with Lang et al.’s (2022) work on inductive followed by deductive research). Grounded theory was explicated by the state of the art; the disruption in consumer lives was substantial and unprecedented. The pandemic rendered people food insecure (Schanzenbach and Pitts, 2020), changed the notions of health, safety, and well-being (Pffefferbaum and North, 2020), and fundamentally altered shopping behavior (Mehta et al., 2020; Sheth, 2020; Zwanka and Buff, 2021). Extant literature did not yield insights into cognitive, emotional, and behavioral changes toward brands as a result of the pandemic, and precluded a literature-derived hypotheses testing effort. The discussion of the first-stage exploration that follows therefore is devoted to explaining how we derived grounded theory; i.e., we discuss the new constructs we identified, how the constructs differed from extant and similar-sounding notions. We highlight the newly derived scales and the new hypotheses we framed based on a grounded framework. In other words, the discussion that follows is distinct from an effort aiming to reproduce exhaustive reviews of branding literature that have occurred elsewhere (see Osorio et al., 2020; Reitsamer and Brunner-Sperdin, 2021; Robson and Jillian, 2021).

First-stage exploration

Focus groups were conducted by a co-author in three classes he taught at an AACSB-accredited B-School located in Northeastern USA (undergraduate Marketing Principles (MP) and Marketing Research (MR), and graduate Marketing Strategy (MS)). All classes were taught via Zoom during the Spring 2021 term; all enrolled students had lived through a year of the pandemic. MR students were engaged in a learning-by-doing process; i.e., they conducted and participated in focus groups, learned to draw inferences and hypotheses from verbal protocols (i.e., inductively derive key constructs based on evidence, develop box-and-arrows frameworks, and learn about developing reliable and valid scales). The MR class is a marketing elective and not a required course; students remain enrolled knowing the requirements of ‘learning-by-doing.’ MP and MS students participated in focus groups during classroom discussions of ‘influence of macro-environmental forces (in this case, COVID-19 pandemic) on user–brand relationships.’

The instructor served as moderator of all focus groups. Attendees were explained the purpose of the focus groups and asked to volunteer (all present in the Zoom sessions volunteered, hence 7 focus groups of 7–8 participants were conducted with a total of 54 volunteers). There was no credit granted for participation, no one was penalized for non-participation. All were either traditional age undergraduate students (ages 19–22), and all MBA students were completing their fifth year of a 4 + 1 degree program (39 males, 15 females). All focus groups began with the following introduction by the instructor/moderator:

The purpose of this focus group is to explore how you, as consumers, were affected by the COVID-19 pandemic, particularly between the time that you were forced to wear masks and isolate yourself from others starting in mid-March 2020 up to now. Please tell others in this focus group about your thinking, feeling, and behaviors related to brands that you purchased in this period.

The bulk of the discussion that followed relied heavily on probing questions triggered by initial responses. For instance, participants were urged to elaborate and provide specific examples that illustrated the notions they voiced. Probes such as: ‘what led you to think or do that?’ and ‘what did you do when you felt that way?’ ‘What is an example of that? ‘If you were to put together all that you said, how would you encapsulate that in one sentence or two for others to understand?’—were frequently used in each focus group.

Analysis of qualitative data

The Zoom recordings were transcribed by the co-author who moderated the focus groups. While students in the MP and MS classes were done after participating in the focus groups, students enrolled in MR not only participated in and observed focus groups, they also submitted a written report highlighting: (a) their key observations, the key constructs identified from verbal protocols together with supporting evidence (quotes to the extent possible), boxes-and-arrows frameworks and working propositions based (an ungraded assignment for which they had received extensive instruction). All present students turned in their written work. The focus group transcripts and the written papers turned in by MR students served as the key data points; they were analyzed by the co-author (moderator of focus groups) based on the guidelines of Miles et al. (2014). The same transcripts and papers were independently analyzed by a second co-author.

Briefly, the co-authors began by identifying themes in the data, identified key constructs, and developed propositional statements. Each theme was supported by quotes from the data. A list of statements made by participants was compiled, and clustered into separate lists—each of which: (a) eventually helped identify and define the key latent constructs shown in Fig. 1, and (b) served as the basis for the measured or indicator variables or the items on the Likert scales (see Table 1). After the independent analyses, the co-authors met repeatedly to reach consensus. Figure 1 illustrates the results of this consensus effort; i.e., it identifies the key conceptual building blocks of the grounded theory we derive and the proposed epistemology of the proximal brand construct. The directionality of relationships, illustrated by the one-way arrows that connect the latent constructs in the figure make explicit our data-derived notions of convergence and discriminance; i.e., we posit that each of the latent constructs are independent and relate to others as shown in the figure. The Likert scales listed in Table 1 represent inductively derived operational definitions, they represent one-to-one correspondence with words used by focus group participants.

Fig. 1
figure 1

A conceptual model of COVID-19 influenced brand evangelism

Table 1 Measurement parameters for the theoretical model and standardized solution for the hypothesized models

COVID-19 as a disruptive event

The clearest finding in focus group data relates to the disruption felt by consumers; all participants report their lives were inordinately disrupted as a result. This disruption is defined as a mix of felt physical and emotional weakness, tiredness, sadness, lost connections, and the destabilization of one’s structure in life. Although not all voices reported feeling disrupted with equal intensity, there was no voice which claimed exception. The latent construct is given shape by seven indicator variables (see Table 1 for 7-point Likert scales reflecting participant voices).

The near universal expression of felt disruption led the moderator to ask two types of probing questions in every focus group (roughly worded as): (a) ‘what did you think, feel and do as a result of experiencing this disruption, how did you deal with it, how did you cope,’ and (b) ‘in what ways did your thinking, and feeling shape your attitude and behavior toward brands, how did the brands factor into your coping responses’?

Participants spoke of their coping related cognition, emotion, and behaviors, i.e., how they overcame the felt disruption. This definition of coping is consistent with current branding literature; scholars define coping as ‘what consumers do to overcome negative emotions’ as coping. That is, current thinking about coping extends beyond cognitive deliberation and rational choice making, all of what consumers do in response to negative emotions is termed coping in branding contexts (see Bayarassou et al., 2021; Mayer et al., 2019; Schenebelen and Bruhn, 2018). In this context, the felt-disruption-triggered responses of heightened intentionality, concern for brand sponsors’ ethical behaviors toward their employees (CEBE), and the resulting brand evangelism emerge as coping responses.

Intentionality

As Fig. 1 shows, heightened intentionality about brand choices and purchase behaviors emerged as a coping response during the COVID-19 pandemic. As opposed to a mental state of drifting, or undirected thought, emotions and actions, or impulsive actions, intentionality in the present context is defined as a purposeful mental state associated with deliberate, focused, directed attention to gaining and evaluating information about brands prior to selection. This grounded definition is aligned with the literature’s view; i.e., that of intentionality as ‘deliberately thinking about something in a focused way,’ as opposed to cognitive drifting or ‘incidental thinking about nothing in particular in unfocused ways’ (e.g., McIntyre and Smith, 1989; Searle, 1991). Higher levels of felt disruption appeared to trigger higher levels of intentionality, and greater reported attention to research into suitability, value, and quality of brands. The latent construct of intentionality was given shape by four indicator variables; each was defined as a statement suitable for a 7-point Likert scale (see Table 1). Some of the representative voices:

It made me more mission driven, paying more attention . . . not impulsivity. My impulsivity was non-existent for the whole, it was in check. I paid careful attention to everything. Did I really intend to buy this or that? I thought about that a lot.


COVID-19 stressed me out. Unstable financial condition like how am I going to pay for . . . like where was I going to work? Structure of my life, like what to do when—because you are not doing with others. I think that is when I started paying a lot more attention to what I was doing, what I was buying . . . I focused a lot on ‘what am I doing, what information do I have about the brand, is it the right thing for me at this time?’


The probing questions asked by the moderator triggered multiple responses, and some reconstructed reasoning. Often, the reconstruction was worded as: ‘what I mean is..;’ followed by reformulated and additional information. In some instances, the reconstructed words were unclear to the moderator, hence he asked for clarifications roughly worded as ‘based on all that you just said, and explained what you thought and did, how would you put it in a way that other participants can understand?’ Such probes triggered some abstracted explanations of what they (the participant) were thinking and doing; not general observations about what others were doing, such as:

. . . I think there is a movement to ‘thoughtful consumption’ of brands that support social issues, are sourced responsibly, or environmentally friendly.


. . being intentional, premeditative and increasingly mindful during this time as they are making more purchases for specific reasons.


. . acting more consciously and placing more value on present, tangible moments . . . mindfulness include awareness, attention, effort, purpose, long-term goals, and prioritization.

Our data-derived view of intentionality as a COVID-19 triggered cognitive, emotional, and behavioral response deserves delineation from current thinking about the construct. First, current thinking about brand intentionality refers to an independent construct; i.e., about the intention of the brand. This view holds brands as intentional agents capable of acting independently and triggering cognitive and emotional responses from consumers (e.g., Puzakova and Kwak, 2017; Giovanis and Athanasopoulou, 2017; Kervyn et al, 2012a, Kervyn et al, 2012b). Scholars caution about the damage to consumer–brand relationships when consumers sense negative intentionality of the brand (Ward and Ostrom, 2006) such as intentional hypocrisy of brand sponsors (Jung et al., 2021), or intentional service failures (Saavedra et al., 2021). In sharp contrast, our qualitative data indicate heightened intentionality of consumers not brands; a notion that has invited no scrutiny in current branding literature. Second, current thinking regards intentionality, a notion associated with purposefulness and deliberation, as a dimension of trust consumers attribute to brands (Delgado-Ballester and Munuera-Aleman, 2002). In contrast, our finding about heightened intentionality of consumers is unrelated to notions of brand trust; instead the grounded construct reflects what consumers are thinking, feeling, and doing, as a response to a highly disruptive event. Similarly, scholars have spoken of intentionality as inseparable from moral judgment about brands (Huang et al., 2020), as a consequence of nostalgia (Wen et al., 2019), and as a driver of social media usage (Lim and Schumann, 2019). We find intentionality unrelated to moral judgment. In our study, participants made deliberative, intentional choices as a way of coping with the feelings of disruption.

Our definition of intentionality resonates, however, with current thinking in three inter-related ways. First, current thinking regards intentionality as a response of people in stressful situations, as a form of coping (see Bloom, 2020). Second, scholars agree that stress and disruption can lead people to act in ways that instill purpose and meaning in their lives (e.g., Vignoles et al., 2006). Third, heightened intentionality is viewed as an act of taking control (e.g., Banks and Welhaf, 2022; Cardoso et al., 2019). In other words, scholars have addressed issues of intentionality as inseparable from issues of control, a context that our study also finds inseparable; i.e., the feelings of lost control heightened intentionality among consumers (e.g., Yao and Siegel, 2021). However, that external, disruptive forces would lead consumers to mindful, purposeful evaluation of brands, and brand evaluation as a form of response to COVID-19-induced distress—are new to the branding literature and deserve testing.

Concern for brand sponsors’ ethical behaviors toward employees (CEBE)

The COVID-19-induced disruption, and heightened intentionality triggered heightened CEBE as a clear cognitive and emotional response. Heightened levels of CEBE, in turn, triggered brand evangelism behaviors (see Fig. 1). We define CEBE as consumers reported attention to the brand sponsors’ reputation for paying their employees a fair living wage, for upholding high ethical standards in the workplace, and for promoting diversity and inclusion in the workplace. In other words, the concerns for ethical behaviors are defined in a narrow context and a narrower question: ‘is the brand sponsor a good employer?’.

The CEBE construct deserves delineation from the currently popular construct of corporate social responsibility (CSR); the latter is associated with positive brand-related attributions (see Diallo et al., 2021; Won-Moo et al., 2020). Our findings about the CEBE construct, when contrasted with current discussions of the CSR construct in the literature, point to distinctive epistemologies; i.e., with differing antecedents, separate definitions, and distinct outcomes. First, CEBE, unlike CSR, is triggered by COVID-19-induced stress. Current literature suggests that antecedents of CSR are expansive and include—among other things—community voices, consumer power, shareholder demands, and employee power (see Yang and Rivers, 2009). Second, CSR is a multidimensional construct associated with people’s perception of financial performance of the firm, quality of ethical statements, reputation, reliability, and risk reduction, trust, and loyalty; i.e., it is a more expansive, multidimensional evaluation of a firm’s conduct by a large number of constituencies—not just buyers of their brands (see the 42-item scale presented by Turker, 2009; also see Oberseder et al., 2014; Stanaland et al., 2011 for evidence of expansive and multidimensional conceptual domain of CSR). CEBE is a narrower domain focused on three key issues related to perceptions of the brand sponsors’ ethical behavior toward employees (and excludes concerns about ethical behavior toward others). Third, they trigger separate outcomes. If a reputation for positive CSR leads consumers to think favorably about the firm and its brands, CEBE is primarily triggering brand evangelism behaviors. CSR is about perceptions of the firm in terms of what it does as beneficial for a large mosaic of constituencies; CEBE is about ‘is this brand sponsor empathetic to my concerns at this time of disruption toward employees.’ Consider the words:

“I wanted to know what the firm was doing during COVID. Are they acting responsibly like they know what we are going through, you know, losing our jobs, and tips . . .?’ expect (firms) to act socially responsibly.”


“. . . did they make a statement of inclusion? Starbucks did. Some did, they said after the BLM started that they were for it. I looked at that. I said yes, that they understand the anger I am feeling, they are inclusive. I thought more positively of them.’

Brand evangelism behaviors

Engagement in brand evangelism emerges as the key and ultimate brand-related response to the COVID-19-consumer–brand interactive context. It is shaped directly by the disruption of the pandemic, the resulting increase in intentionality and CEBE. Ultimate in that it serves as the dependent variable of the conceptual model we propose and test. Brand evangelicals tell others about the great brands they are buying, sharing insights into and advocating for the brands to improve other people’s lives. Consider the voices:

I went through this like really. I know what I am talking about this brand for. I looked into it, what it does, why it is better. When I say ‘buy’ this, I know it is to help them, like I am not getting a commission or anything. (I think that) on some level they get this, they can see that I wouldn’t be doing this otherwise, like telling them what to do . . . they know I care. So I say, ‘don’t buy that, buy this, it’s better and good for you. You’ll thank me (later).

They should benefit from all the effort I’ve put into this, and tell them to trust me, why else would I say anything if I didn’t have their good intents at heart . . . their future in my heart? It’s only their lives that are going to get better. I say, ‘just buy this, because you don’t know, I know and this is way better than what you are getting now.’

Q: have you persisted if the people you were trying to convince were defensive?

A: Defensive? Like what I did if they resisted? Of course I kept on and on until they pretty much gave in (laughs).


Brand evangelism as a response to COVID-19-induced stress is new to the literature; the notion has attracted attention in other contexts. For instance, the notions that brands have evangelists and that fostering evangelism among consumers is beneficial to brands are not new to the literature—yet they relate to a pre-COVID-19 context (see Harrigan et al., 2021; Nyadzayo et al., 2020). They are alternatively referenced as brand apostles (Jones and Sasser, 1995), champions (Bhattacharya and Sen, 2003), and advocates (Chung and Darke, 2006). Current thinking suggests that evangelicals not only buy the brand but also provide positive brand referrals (Doss, 2015; Matzler et al., 2005; Swimberghe, Astajkhova and Wooldridge, 2014; Wallace. Buil, and de Chernatony, 2014), and engage in pro-brand behaviors (Beccara and Badrinarayanan, 2013; Hsu, 2019). Evangelical behaviors seem more likely to emerge when brands are highly differentiated rather than homogeneous (Doss, 2015; Doss and Carstens, 2014). Current theory holds that brand evangelicals signal an intent to purchase a brand, provide positive word of mouth, and try to convince others to not purchase competing brands (Becerra and Badrinarayanan, 2013; Hsu, 2019).

Consumer voices offer a different view relevant to the present context; brand evangelicals aim to help others become better consumers, advocate on behalf of brands that have served them well, and aim to convert others to their way of thinking while meeting them in person (see the measurement scale we used in Table 1). Similarly, questions can arise about similarities with purveyors of opinions, WOM and eWOM; i.e., is a brand evangelist an opinion leader or an influencer? Here too, we can draw epistemic distinctions based on comparison of our findings about brand evangelism and the literature’s view of WOM. In our context, brand evangelism is uniquely about users’ evangelical fervor, the interest in proselytizing and converting others’ to one’s point of view—as a response to COVID-19-induced stress. Conversion and fervor are unassociated with the way eWOM or influencer constructs are measured; current definitions of eWOM and influencers are devoid of concerns about a stress-induced response and heightened intentionality or CEBE. Second, currently popular notions of influencers, eWOM purveyors, and influencers are firmly embedded in a media context of communication, social media, or otherwise. Brand evangelism behaviors in the present context relate to interpersonal interactions; when consumers meet others personally, they want to improve their lives, convert them to brands they are evangelizing about.

Third, brand evangelism behaviors are anteceded not just by feelings of disruption but also by activated cognitive and behavioral responses related to acting in more intentional ways and thinking about the salaries and working environment of servers fits with one’s sense of ethics. Converting others to one’s faith, using the brand as a prop, appears more about gaining a sense of control (as a result of controlling others). Opinion leaders seek acceptance of others, aim to present themselves in enhanced ways (Winter and Neubaum, 2016), or gain extrinsic (monetary) rewards (Shi and Wojnicki, 2014). Purveyors of eWOM are motivated by the prospects of self-enhancement and enjoyment (Hu and Kim, 2018). Brand influencers are motivated by money (Jin et al., 2019). Brand evangelical behaviors, in sharp contrast, emerge as selfless coping responses to environmentally induced disruption (COVID-19 pandemic), and as a result of heightened intentionality and CEBE.

Hypotheses:

The hypotheses that guided the data collection of the next stage of our study were as follows; all relate to a specific context of consumer–brand interactions during the COVID-19 present reality of March 2020-April 2021 period (grounded, inductively derived latent constructs in italics):

H1: Higher the level of COVID-19-induced disruption reported by the consumer:

  1. a.

    Higher the level of reported intentionality in brand choice behaviors.

  2. b.

    Higher the reported concern for brand sponsors’ ethical behaviors toward their employees (CEBE).

  3. c.

    Higher the reported brand evangelism behaviors.

H2: Higher the level of reported intentionality in brand choice behaviors:

  1. a.

    Higher the reported concern for brand sponsors’ ethical behaviors toward their employees (CEBE).

  2. b.

    Higher the reported brand evangelism behaviors.

H3: Higher the reported concern for brand sponsors’ ethical behaviors toward their employees (CEBE), higher the reported brand evangelism behaviors.

Survey, hypotheses testing

Questionnaire design

The co-authors initially compiled all statements made during the focus groups that pertained to each of the latent constructs shown in Fig. 1. The statements were translated into items for Likert scales for measuring the latent constructs. In other words, there is a one-to-one correspondence between statements during focus groups and the Likert scale items (cleaned up to remove colloquialisms and grammatical errors).

The resulting questionnaire with scales reported in Table 1 was circulated to a nationwide sample of consumers using Amazon’s MTurk service in mid-June 2021. The findings reported below refer to 786 responses. Participants were offered a dollar for completing the questionnaire; only completed questionnaires were used for analysis. The sample skewed toward males (65%) of respondents; females represented 35% of responses; age and gender are independent in the sample (based on a Chi-square test). There is also no significant relationship between gender and each of the hypothesized latent constructs (based on t tests of equivalence in means of male and female respondents). Half of the sample was 34 years old or younger; most (47%) of respondents were between ages of 25–34. One-way ANOVA indicates that people ages 65 and older reported lowest disruption (mean = 3.2), whereas those aged 25–35 reported the highest levels of disruption (mean = 3.86). People aged 35–44 emerged as the strongest brand evangelicals (mean = 4.06); those aged 55–64 emerged as the least likely to evangelize (mean = 3.34). Ten percent of the sample were foreign nationals living in the USA; the most represented states were California (14.6%), Indiana (12.7%), Texas (8.7%), Florida (8.4%), and New York (5.9%).

  • Step 1. Testing structural coherence and hypotheses

    We used EQS 6.2 software, and followed the two-step process for constructing a structural equation model (SEM) based on guidelines of Anderson and Gerbing (1988). The first-stage testing of the measurement model focused on testing whether any structural coherence—as hypothesized—existed in the data. In other words, we aimed to test whether our grounded notions of latent constructs—COVID-19-induced disruption, intentionality, CEBE, and brand evangelism as specified by their measured, indicator variables—were structurally sound and satisfied concerns about goodness of fit. In this step, we examined whether a confirmatory factor analysis produced the results as hypothesized; i.e., whether the latent constructs were orthogonal and significant. We specified the robust estimation procedure to overcome some of the problems that can arise as a result of non-normality (based on Yuan and Zhang, 2012; Salmones et al., 2021). Based on the Lagrange multiplier test, we removed three indicator variables from the CEBE construct because they cross-loaded on other factors (see Table 1 for details on the items removed). The resulting CFA showed excellent goodness of fit (NFI = 0.942, BNNFI = 0.966, IFI = 0.972, RMSEA = 0.034). We used SPSS to calculate Cronbach’s alphas for latent constructs, and the SEM-produced factor loadings of each indicator variable to assess the construct reliability and construct (discriminant and convergent) validity from the average variance extracted measure (CR, AVE, see of reports in Table 2). The factor loading of the final CFA model and the parameters of the theoretical model are reported in Table 1 (all reported statistics are significant at 95%).

  • Step 2. Testing for common methods variance.

    We tested for the presence of common methods variance (CMV), given that all measured variables were obtained from a single rater. Lindell and Whitney (2001) suggest that the lowest and second-lowest correlations among indicator variables are reasonable proxies for CMV; i.e., if the method contributed to relationships not otherwise present, they are likely captured by these correlations. In our data, the lowest and second-lowest correlation are 0.137 (I4-CID6) and 0.178 (I4-CID1); hence, our initial concerns of CMB were not heightened.

    We tested for the significance of CMV in the following ways based on Podsakoff et al., (2003) and Malhotra et al. (2006). First, high correlations (0.9 or greater) between latent constructs signal the presence of CMV (Pavlou et al., 2007). Table 1 shows that the correlations do not exceed 0.49. Second, we conducted Harman’s (1976) single-factor test (based on Fuller et al., 2016). We conducted an EFA using the ‘principal axis factoring’ as the extraction method, and setting the number of factors to one. We found that one factor (eigenvalue greater than 1) explains 38.962% of all the variance; or that a single factor does not explain more than 50% of all variance in the sample (Podsakoff et al., 2003). That is, we inferred that CMV is not a significant source of concern. We further tested Harman’s notion while conducting the CFA; we loaded all measured variables on a single factor to examine its structural coherence and fit. If there is notable CMB, the CFA would yield acceptable fit parameters (Malhortra et al., 2006). Our model had an unacceptable fit (NFI = 0.678, NNFI = 0.65, RMSEA = 0.138); we inferred that common methods bias was unsubstantial. Finally, we used a common latent factor (CLF) test. We introduced a new latent construct (with all indicator variables) and ran the CFA (variances of all latent factors were constrained to one, covariances between hypothesized latent constructs and the CLF were constrained to zero, and all paths between the CLF and indicator variables were constrained to equal each other). The regression equations thus obtained for each of the measured variables were then compared with the regular model (i.e., one with the CLF, one without). The biggest magnitude of difference between beta weights of hypothesized constructs (as independent variables)—upon comparison of the two models—is 0.11; i.e., significantly lower than 0.2 we used as a cut-off (based on Lindell and Whitney, 2001; Serrano et al., 2018). After multiple tests, we ruled out the concern with CMV.

    Bagozzi et al. (2017) note that common method bias—to the extent detected (see Lindell and Whitney, 2001), can hurt discriminant validity of latent construct, a notion that was further discussed by Conway and Lance (2010). In other words, if we do not detect CMV after multiple tests because the model contains many latent constructs (see Malhotra et al., 2006), Conway and Lance (2010) advocate for rigorous testing of construct reliability (each latent factor should report CR > 0.7), and discriminant validity (average variance extracted or AVE for each latent construct should exceed 0.5, and exceed the squared correlations among hypothesized latent constructs). Aligned with their suggestions, Table 2 reports the measures we calculated to test for such reliability and validity, i.e., (a) composite reliability (CR) of all latent constructs exceeds 0.7; i.e., the observed variables correlate well within each parent factor, (b) the average variance extracted (AVE) exceed all squared correlations among latent constructs (see highlighted diagonal in the table), and maximum shared variance (MSV) is 0.497, i.e., lower than the lowest calculated AVE. Based on evidence of convergent and discriminant validity, and tests for CMV, we infer that findings have acceptable levels of reliability and validity, and an insignificant concern about CMB.

  • Step 2. Path analysis, hypotheses tests, and findings

    In the next step, we specified the hypothesized paths in the SEM procedure and tested for goodness of fit. All hypothesized paths are significant (see Fig. 2 for betas, t-statistic, and fit indicators including RMSEA of the hypothesized model). As Fig. 2 shows, hypothesized relationships leave no room for additional testing between constructs; there was no basis for testing the existence of other relationships among latent constructs (i.e., Wald’s test was redundant). Hence, we find significant evidence to suggest that the COVID-19 pandemic-triggered significant disruption among consumers, and that this disruption triggered intentionality behavior, and concern about brand sponsors’ ethical behaviors, as well as brand evangelical behaviors. Similarly, both intentionality and BSEB served as partial mediators in the link between COVID-19 caused disruption and brand evangelical behaviors.

Table 2 Reliability and construct validity statistics (n = 786, June 2021)
Fig. 2
figure 2

A theoretical model of COVID-influenced brand evangelism

Theoretical implications

The clearest implication of the study is that the COVID-19 pandemic changed the way consumers think, feel, and do about brands. We find three distinct cognitive, emotional, and behavioral responses specific to this context, including brand evangelism behaviors. Hence, the epistemology and framework we present are uniquely relevant to the present reality; it is distinct from pre-COVID-19 views of WOM, eWOM, opinion leadership, influencers, or brand evangelicals as those who signal an intent to purchase the brand, and exhibit and pro-brand behaviors—as these extant constructs are currently discussed (see Hsu, 2019). Similarly, CEBE emerges as a distinct coping mechanism with a narrower conceptual domain than that associated with CSR—a comparison that is bound to arise. We discuss two key implications of our findings that deserve to inform new theories of consumers’ cognition, emotions, and behaviors toward brands in an emerging reality that is shaped by the disruption of the COVID-19 pandemic.

Consumer–brand relationships shaped by macro-environment induced stress

To the extent managing stress is termed ‘coping,’ branding literature offers multiple insights into relationships with brands as coping responses (see Bayarassou et al., 2021; Mayer et al., 2019; Schenebelen and Bruhn, 2018; Xiao and Lee, 2014; for review), and coping responses to service failures (Sengupta, et al., 2015). For instance, consumers who feel unpleasant emotions cope, or repair damage done to them by interacting with brands with pleasant personalities (Trump and Newman, 2021), respond to materialism by loving brands they own (Ahuvia et al., 2021), to their low incomes by showing more loyalty to brands (Hamilton, 2012; Murilo et al., 2021), to their ambiguous identities by favoring old or retro brands (Hemetsberger et al., 2011), and to their feelings of hate toward brands (Fetscherin and Sampedro, 2019). Most of this attention relates to coping with what firms do or should do (e.g., Rindell et al., 2011; Hutchinson et al., 2013), or to factors intrinsic to the customers (e.g., Diallo et al., 2021); the branding literature is largely insensitive to consumer choice making, and to the activated cognition, emotion, and behaviors in the consumer–brand interactive context as a result of macro-environmental shifts beyond control of consumers and firms. Lim’s (2021a; 2021b) reports on impact of COVID-19 related changes in tourism are notable exceptions.

For instance, current theory and models of brand choice making are based on utility maximization as a rational option (Chintagunta, 1999), or on brand attributes (Singh et al., 2005), or on assigned utilities to brand attributes (Matsatsinis and Samaras, 2000), or as a result of manipulation of gross rating points (Ban et al., 2011), or price expectations (Kalwani et al., 1990), or more recently based on life event changes (Koschate-Fisher et al., 2018), or signaling theory (e.g., Oh et al., 2021). The process by which consumers make brand choices, however, is also shaped by forces outside of the control of firms, brands, brand attributes, or prices. Only recently has branding scholarship focused on the impact of loneliness (Murthy, 2017) and social disconnection (Putnam, 2000), in terms of conspicuous consumption (Liu et al., 2020), and alleviation by formation of brand communities (see Snyder and Newman, 2019; Sullivan and Richardson, 2020, also see Loh et al., 2021). Other psychosocial epidemics that can shape brand choices and brand-related behaviors remain relatively unaddressed; e.g., branding literature has yet to explain the influence of emerging epidemics of heightened stress, mental illnesses and depression, unwanted weight gain among a majority of the adult population (see APA 2020, 2021), or explain the role of branding in the wide and deep opioid crises—many linked to brands—faced by 1.7 million Americans (Brand, 2018; NIDA, 2021), or of climate change crisis (e.g., Roderick, 2017). Similarly neglected are growing influence of psychosocial and cultural changes associated with burnout—evident in multiple professions even prior to the pandemic (e.g., Bakhamis et al., 2019) and continues to persist (Robinson, 2021); or with the epidemic of sedentary lifestyles and obesity that similarly existed prior to the pandemic (e.g., Morabia and Costanza, 2005) and continues to persist (Srinivas et al., 2020). Moreover, branding theories—not sufficiently shaped by macro-level forces that define contexts of consumption, and relatively incognizant of bounded-rational and extra-rational consumption behaviors cannot adequately explain, for instance, how and why the popularity of brands has shifted during the COVID-19 crisis; e.g., brands popular prior to the pandemic, such as Ford, Jeep, and BMW, were replaced by Google, YouTube, and Toyota (Jones, 2020). We learn that consumers’ deep personal sadness and disconnection from others along with loss of agency triggered distinct cognitions and emotions, and triggered brand evangelism behaviors; i.e., the growing impact of uncontrollable forces deserve adequate accounting in future branding theories (see Lim et al., 2022a, b as example).

A perspective into organizational ethics and branding

As we note earlier, CEBE relates to a coping response to felt disruption of the COVID-19 pandemic, and it is specific to the concern about how brand sponsors are behaving toward their employees. CSR is a significantly more expansive construct; CEBE is specific to the context of our study. Other distinctions are worth noting. Current thinking holds that positive evaluation of CSR by consumers triggers positive responses toward the firm (e.g., Bogan and Dedeoglu, 2019; Edinger-Schons et al., 2019; Koch et al., 2019; Thorne et al., 2017). The CFA we conducted, and the model purification process based on the Lagrange multiplier test highlights the key theoretical and empirical distinctions between CSR and CEBE. Our original scale, based on verbal protocols, included six items (see Table 1, the original six-item Likert scale with the question: After making it this far in the COVID-19 vaccine present and available environment I am much more likely to buy brands produced by firms that). During the purification, we discarded three items because they cross-loaded and weakly related to the parent construct of CEBE (items related to: (a) engage in eco-friendly practices, (b) take pains to protect the personal information I share with them, and (c) share profits with employees). In other words, our CFA procedure indicates that cognitive retrieval triggered by our measurement scale was not reliably associated with these concerns about eco-friendliness, privacy, and profit sharing—issues otherwise central to definitions and measures of CSR (see Turker, 2009). The extent to which CSR and CEBE are overlapping constructs, or whether the latter is distinctly activated while experiencing personal distress, as we find in our study, deserve further inquiry.

Conclusion

The contributions of our study are: (a) the COVID-19 pandemic-triggered disruption, stress and a host of cognitive, emotional, and behavioral coping mechanisms among consumers, (b) the disruption led to heightened intentionality and concerns about brand sponsors’ ethical behaviors as a coping response, and (c) as the result of disruptions and other activations, consumers evangelized about brand as coping behaviors. The study highlights two multiple new directions for branding research including: (a) new research that examines the influence of large scale, macro-environmental forces on user–brand relationships (e.g., Lim et al., 2022a, b), (b) further development and refinement of measurement scales designed to reflect user–brand relationships—some of which we identify in our study (e.g., intentionality about brand purchases, CEBE, and brand evangelism). A host of emerging factors in the macro-environment (such as the COVID-19 pandemic) are shifting the nature of user–brand relationships and their outcomes; these developments deserve to inform future theory. Finally, we caution against broad generalization of findings given that this is the first study that identified the antecedents of brand evangelism in the COVID-19 context of user–brand interactions. Similarly, we used SEM procedure not to draw causal inferences but to test hypotheses in one-shot (e.g., Fornell and Larcker, 1981). Arguments about why the arrows in Figs. 1 and 2 point the way they do rest in qualitative data and not current literature. One study conducted in the COVID-19 present environment, exploring consumers’ brand-related emotions, cognitions, and behaviors cannot entirely describe the phenomenon of consumer–brand interactions.