Introduction

Brands have been increasingly using social media platforms to speak out on divisive sociopolitical issues (Bhagwat et al. 2020; Gaines-Ross 2017). A prominent example is the social media response from many global brands on racial injustice. Shortly after the killing of George Floyd in 2020, sportswear brand Nike posted a statement on Twitter, “For once, don’t do it. Don’t pretend there’s not a problem in America”. Nike’s tweet garnered 827 replies, 95,000 retweets, and 218,500 likes. Not long after, several other brands broke their silence on the controversial issue of racial injustice in the USA with their tweets (Sherman 2020).

Notwithstanding the apparent uptick in sociopolitical activities, brands are often reluctant to take public positions on controversial sociopolitical issues (Korschun and Smith 2018), largely because of their provocative nature. Sociopolitical issues can be divisive, engendering strong disagreements between segments of the public (Nalick et al. 2016), so brands that issue activist content on social media or other channels may offend consumers who disagree (Kotler and Sarkar 2017), potentially harming purchase in the process (Hydock et al. 2020). Consumers may also accuse a brand of engaging in inauthentic activism when its messaging is inconsistent with its purpose, values, and practice (Sobande 2019; Vredenburg et al. 2020). For example, PepsiCo’s 2017 activist campaign “Live for Now Moments Anthem” for Pepsi was accused of trivializing institutional racism, appropriating the “Black Lives Matter” movement, and attempting to profit from activism despite having no established record of advancing racial justice through corporate practice (Hogan 2017). As a result, PepsiCo’s brand value plummeted by 4%. On the other hand, when sportswear company Nike launched the Colin Kaepernick campaign addressing systemic racism in America, it aligned well with their long-term commitment to racial justice. Therefore, Nike’s online sales increased by 31% despite backlash on social media from opposing groups (Martinez 2018).

As the examples above indicate, the responses to brand activism on social media can be varied and unpredictable. Perhaps unfortunately, brands often have insufficient information to predict the response of their diverse stakeholder bases (Richardson 2012). To better predict these responses, brands could rely on surveys or focus groups; yet these data sources are often time-consuming and expensive (Hagiu and Wright 2020). Even if a brand had the time and resources required to conduct traditional market research, social media users might be unwilling to share their unvarnished political opinions; According to a survey conducted by Pew Research Center, only fewer than ten per cent of consumers share their opinions on sociopolitical issues on social media (McClain 2021). Clearly, traditional forms of market intelligence are often not a viable option for brands.

In this article, we contend that brands sometimes deal with this information gap by surveilling the communications of other brands on the sociopolitical issue. Prior research suggests that managers sometimes mimic peers’ strategies when facing challenging situations (Lieberman and Asaba 2006; Scharfstein and Stein 1990). They do this to maintain competitive parity and tap into superior information they may believe a competitor has about the market. We extend this prior research by showing that such peer effects may extend considerably wider than previously thought. More specifically, we present evidence that brands monitor the two-way social media interactions of other brands with their respective stakeholder bases regarding their non-market strategies (Baron 2009; Lux et al. 2011).

We draw from existing strategic management theories that distinguish between “market” and “non-market” components (Baron 1995; Boddewyn 2003; Henisz and Zelner 2012). The market environment involves interactions between firms and other parties in the marketplace through economic transactions, exchange of property, or other activities mediated by markets or private agreements. In contrast, the non-market environment entails “social, political, and legal arrangements that structure the firm’s interactions outside of, and in conjunction with, markets” (Baron 2003). Thus, non-market strategies involve political and social leverage, using either private politics strategies (engaging in activism on sociopolitical issues) or public politics strategies (e.g. lobbying and engaging with regulators; Baron 1995, 2009; Henisz and Zelner 2012).

Consistent with this literature, brand activism can be conceptualized as a somewhat novel form of public non-market strategy involving messaging, public statements, modification of corporate policies, monetary donation, or other commitments to a sociopolitical cause (Nalick et al. 2016; Wettstein and Baur 2016; Kapitan et al. 2019; Vredenburg et al. 2020). In the current study, we spotlight social media communications about brand activism. This is a natural point of focus because it is common practice for brands to use social media to state opinions about sociopolitical issues as well as to announce more tangible commitments (e.g. Intel’s 2021 tweet saying, “We are pledging $1 million…to address social injustice and antiracism. #blacklivesmatter”). Thus, social media posts are an effective means to capture a wide range of brand activism activities.

We investigate peer influence of brand activism on social media in the context of Twitter, analysing 4,505 activist tweets from 177 global brands on three high-profile sociopolitical problems (Black Lives Matter, LGBTQIA, and COVID-19 policy), 55,121 user replies to those tweets, as well as 87,254 direct appeals from users to the brands on our list to make statements on the sociopolitical issues. We find that brands are indeed more likely to address a divisive topic on social media if their peers have recently engaged in activism on the issue. Moreover, we find that brands look at the consumer response to peer activism; the likelihood to tweet is higher when user responses are more positive, and the more there are direct social media appeals to engage on the sociopolitical issue.

This research combines three areas of brand-focused literature that have traditionally been separate: peer influence, online communication between brands and consumers, and reactions to brand activism. Figure 1 shows that these studies, whether at the brand or individual level, have tended to remain distinct from one another. The present research, in contrast, connects all three areas and sheds light on the interactions between brands in their non-market strategies of engaging in brand activism.

Fig. 1
figure 1

Mapping the existing literature on brand activism, online communication, and peer influence based on the actions and their impact on brands and stakeholders

Full size image

Moreover, this research contributes to the literature on peer influence, brand activism, and online communication between brands and consumers in several ways. First, we extend prior research on peer influence and propose that while brands monitor and imitate other brands to create market strategies in conflicting situations (Scharfstein and Stein 1990; Aral and Walker 2011), such behaviour may also be reflected in times of developing non-market strategies, such as a brand’s decision to take stand on a contentious sociopolitical issue.

Second, our approach contributes by addressing another side of brand activism. Much of the literature focuses on the consequences of brand activism. Some examine the impact on market performance (Hadani and Schuler 2013; Fossen and Schweidel 2019; Lacka et al. 2021, etc.). Other research examines the direct consequences of brand activism on purchase (Dodd and Supa 2015; Hydock et al. 2019, etc.), or the effects of the appropriateness and the authenticity of activism strategies on polarizing issues (Sobande 2019; Vredenburg et al. 2020; Mukherjee and Althuizen 2020; Sibai et al. 2021). While the findings of this body of research aid in evaluating the outcomes of brand activism and offer guidelines to distinguish authentic activism from “woke washing” (Sobande 2019; Vredenburg et al. 2020), prior research has tended to overlook the antecedents of brand activism and has generally not considered the two-way dialog that often occurs between brands and social media users. In contrast, we identify factors that contribute to brand activism, including the behaviours of peers as well as the reactions of social media users.

Furthermore, this study contributes to the literature on social media by suggesting that increased user appeals to brands on social media to address an issue may indicate heightened customer expectations surrounding the issue at a given time. Using social media mentions and hashtags, we propose a novel way to quantify the theoretical concept of increased stakeholder expectations of brands to address contentious issues in the sociopolitical domain (Hoppner and Vadakkepatt 2019). In addition, our approach stands in contrast with much of the literature on social media that tend to explore either firm-generated content (FGC) or user-generated content (UGC) individually (Kumar et al. 2016; Tirunillai and Tellis 2012; Sonnier et al. 2011, etc.). This paper explores both FGC (such as brands’ sociopolitical tweets) and UGC (such as stakeholders’ reactions to those tweets and direct appeals to brands) to understand the overall impact of brand activism strategies on social media.

For practitioners, the findings from this study may provide insights into opportunities for developing effective strategies during conflicting situations. Our study provides evidence that brands may analyse the narratives and the user reactions to peers’ recent sociopolitical statements on social media as they look for an appropriate strategy during a challenging time. We contend that the percentage of positive replies to a brand’s activist statement may signal the direction of public support on a polarizing issue at a given time. This is supported by prior research that the volume of positive comments on a firm’s online posts positively correlates with the market response (Sonnier et al. 2011).

Overall, this article adds insights to the existing literature on brand activism and aims to take the discussion further by exploring whether brands independently decide on their activism decisions or whether the recent relevant activism of the peer brands influences them.

When brands engage in activism in response to the activism of peer brands

Our theoretical development emerges from the fact that brands may go through several stages when publicly addressing a sociopolitical issue on social media (Fig. 2). The process starts with exposure when a sociopolitical issue becomes a public debate or discussion point. In deciding whether to make a statement on the sociopolitical issue, the brand manager must then weigh the risks both (a) of speaking out on the issue and (b) of staying silent. Finally, the brand may issue a public statement on the sociopolitical issue or decline.

Fig. 2
figure 2

The decision stages of a brand’s engagement in activism

Full size image

We now develop a series of hypotheses to empirically test whether brands react to the activism of other brands, which brands they pay the closest attention to, and what information is most important as they make the decision. The key effects of our conceptual approach are illustrated in Fig. 3.

Fig. 3
figure 3

Conceptual framework for the research model

Full size image

Other brands’ recent activism campaigns on social media

Brands routinely monitor their respective market environments to identify and satisfy market needs. Such a market orientation (Kohli and Jaworksi 1990; Slater and Narver 1994) can provide a competitive advantage. Such marketplace scanning may involve identifying market needs (Day and Wensley 1983; Kotler 1988), but it may also involve being attuned to the actions and strategies of competitors. Such scanning may be heightened further in situations where a strategic decision is risky. For example, Scharfstein and Stein (1990) find that brands mimic competitors more when facing a risky decision about market strategies such as making a corporate investment in assets and purchasing stocks.

We extend prior research and contend that brands attend to, and react to, the non-market actions (engaging in brand activism on sociopolitical issues) of any other brand regardless of their industries. When facing the prospect of making a social media statement on a potentially controversial sociopolitical issue, brands may monitor the actions of other brands as a means of assessing risks. More specifically, we predict that a brand will be more likely to make such public statements the more any other brand has done so because it will provide critical signals that may inform the brand’s decision about making such a public statement themselves. For example, ride-sharing brand Uber took a stand on the mask mandate on social media in the middle of the COVID-19 pandemic in 2020 and tweeted, “Masks keep our communities safe” while using the hashtag #wearamask. Within a few hours, one of its non-industry peers, tech giant Twitter, used the same hashtag and tweeted, “The only thing we want going viral is this Tweet”, which garnered 34,081 retweets and went viral. This paper investigates whether a brand’s (such as Twitter’s) decision to speak out on a controversial topic is made independently or influenced by other brands’ (such as Uber’s) recent related activity. We argue that waiting for others to make statements first provides data on those brands’ strategic decisions in a conflicting situation. Consequently, it may reduce risks that arise from a lack of customer information.

In addition, when a brand chooses to speak up on a sociopolitical issue, target customers may look for other major brands weighing in on the subject. Finally, we argue that the brands are also aware of the possibility of increased customer expectations of them to address an issue. As a result, more brands will be willing to engage in activism on a sociopolitical issue when there are recent activism campaigns from peer brands on the issue. We thus make the following prediction:

H1

A brand is more likely to engage in activism on social media on a given day the more other brands have recently engaged in activism on the issue.

Recent brand activism campaigns by industry peers on social media

In H1, we explored the influence of recent activism on social media on a contentious sociopolitical issue from any other brand, regardless of industry. However, it is reasonable to predict that brands might glean the most relevant and valuable information from industry peers. Brands often compete fiercely with brands in the same industry as they may share potential customers (Berendt et al. 2018). Since peers within an industry share the same customer groups, their strategic decision may be informative to a brand during a conflicting time. As a result, when industry peers make a public statement on a sociopolitical issue, it should, in theory, carry more relevant information than statements from brands in other industries.

This is consistent with the prior literature that suggests that a brand’s tendency to mimic the strategic decisions of its rivals may come from a combination of information-based motives and rivalry-based motives (Lieberman and Asaba 2006). First, competitors sharing standard technology, market orientation, and a similar customer base may be deemed as informative to each other. Therefore, although managers rely on a multitude of public and private information sources, they may look at industry peers in the same industry for the most relevant information (Bustamante and Frésard 2021). Second, competitors are more likely to take similar strategies to maintain their relative position or mitigate rivalry (Lieberman and Asaba 2006). In addition, Scharfstein and Stein (1990) find that an unprofitable decision is not as bad for a brand’s reputation when peer brands make the same mistake. As a result, brands are often more favourably evaluated for their financial decisions when they follow their peers versus behaving independently in challenging times. Thus, brands may engage in activism as a defensive strategy to maintain their relative position and stay competitive.

Therefore, we argue that speaking up on a contentious issue in challenging times may be a more favourable strategy for brands when their industry peers have recently addressed it online.

H2

A brand is more likely to engage in activism on social media on a given day the more industry peer brands have recently engaged in activism on the issue.

The volume of positive user responses to peers’ brand activism

The first two hypotheses predicted that brands monitor the activities of peer brands as they decide whether to make a statement on a sociopolitical issue themselves. However, the information a brand may glean from such peer activism on social media may not be limited to the activist statements themselves. Our approach suggests that brand managers will also examine how the public responds to activist statements on social media to assess the risk of engaging in activism on the same issue. Thus, to understand the overall audience sentiment on peers’ activist statements, brands may monitor brand-related online chatter or user-generated content (UGC) (Tirunillai and Tellis 2012) posted as replies to the activist statements on social media. For example, following the same-sex marriage ruling in the USA in 2015, when Domino’s Pizza tweeted the hashtag #LoveWins (Cipoletti 2015), it gained hundreds of positive replies from Twitter users such as “This is the greatest thing ever!!”, as well as negative replies like “I prefer my pizza without political affiliations. No more Domino’s pizza”. We contend that a higher volume of positive responses in such a polarizing situation may suggest that the risk of engaging in activism is relatively low. Thus, one indicator of how well consumers and other stakeholders receive a brand’s public statement on a polarizing issue may be the volume of positive (or, conversely, negative) replies to the activist content.

Therefore, we contend that brands can anticipate the possible audience reaction to a contentious issue by considering the percentage of positive responses to peer brands’ recent public statements on the subject on social media. When a peer brand’s recent activism on a divisive sociopolitical issue garners a lot of positive responses on social media in general, it may indicate that the brand’s viewpoints on the issue are well aligned with the audience. As a result, the remaining brands may interpret it positively, reducing their fear of ambiguity on social media users’ overall sentiment on the topic as well as the risk of speaking up on the issue. Therefore, we argue that a brand may be more likely to address a contentious sociopolitical topic when the relevant prior statements from peer brands have received a lot of positive replies on a social media platform.

This argument is consistent with the existing research that suggests that brands may monitor the social media users’ response to their peers’ brand activism campaigns to determine whether their position aligns with the customers’ expectations. For example, prior research found that firms may gauge whether the rewards of gaining new customers outweigh the risk of losing existing ones (Hydock et al. 2020). In addition, on social media, brands sometimes monitor peer brands to determine whether target customers are pleased with a competitor’s service (Slater and Narver 1994). We argue that when a brand receives a lot of positive user replies to its activism on social media, it signals a favourable reaction on the part of potential customers to the brand activism. This is consistent with prior research where Sonnier et al. (2011) observed that the volume of positive user comments on a firm’s online post positively correlates with its daily sales performance. In addition, when most consumers are aligned with a brand’s stance, the net effect of brand activism is more likely to be positive at the market level (Hydock et al. 2020).

Therefore, we propose that the publicly available information on the number of positive replies to a peer brand’s activist content on social media may reduce other brands’ risk of speaking up on the issue.

H3

A brand is more likely to engage in activism on a given day the higher the percentage of positive replies to peer brands’ recent social media statements on the issue.

Users’ direct appeals to brands on social media to address an issue on the platform

While publicly available data on social media users’ responses to a peer brand’s public statement on a contentious issue may indicate the broader audience sentiment at the moment, a brand may continue to seek direct nudges from stakeholders before speaking up on the issue. For example, Disney posted a tweet on racial justice in 2020, saying, “We stand against racism. We stand for inclusion”. In response, when a Twitter user publicly addressed Disney to take more tangible actions by saying, “Open your purse Mickey Mouse”, it gained 51,000 favourites in 24 h. Moreover, it was retweeted even more than the original tweet by Disney. We argue that such UGC in the form of online mentions or direct appeals to brands to take action may build up social pressure. In addition, it may instantly reveal to brands what trending sociopolitical problems matter the most to the target audience at a specific time. To capture the consumer expectations on brands to address an issue at a specific time, we propose considering the cases in which online users resort to social media and explicitly mention brands to address a sociopolitical issue on the platform. We argue that the number of mentions or direct appeals from users to the brands on social media hints at the market expectations growing on them. More importantly, brands face higher risks from staying silent in such a scenario, as customer demands would continue to grow.

This is consistent with the prior literature that when a company fails to speak up under social pressure, the demands from customers, employees, and industry peers continue to rise (Gaines-Ross 2017). In addition, while taking a political stand can be risky for brands, abstaining from a stand amid high stakeholder expectations can present substantial risk (Korschun et al. 2019). As a result, when a brand receives more social media mentions from consumers to address a contentious issue, the brand is more likely to take a stand on the platform to satisfy the growing customer demands. Thus, we hypothesize:

H4

A brand is more likely to engage in activism on social media on a given day the more online mentions there are to the brands to address the issue.

Methodology

Selection of brands and sociopolitical issues

We choose Twitter as the social media platform for our context to explore how brands monitor their peers online while taking a stand on a divisive sociopolitical issue. According to the Statista Research Department (2021), Twitter is one of the leading social media platforms used by B2B (54%) and B2C (46%) brands worldwide. Therefore, Twitter provides a fruitful context for the current research because it enables an analysis of company statements as well as consumer responses to those statements.

To create our data set, we first collected all brand names that appeared on at least one of three leading listings of leading global brands (2020 rankings by InterBrand, Brand Finance, and BrandZ). To ensure that we specifically captured brands’ strategic response to divisive topics rather than the regular news updates on ongoing sociopolitical issues from news channels, we excluded three news channels (CBS, NBC News, and Fox News). The final dataset consisted of 177 global brands representing 35 diverse industries.

The analysis examines three sociopolitical issues that have polarized society in recent years: Black Lives Matter, LGBTQIA, and COVID policies. Each is a prominent and divisive issue. For example, in August 2020, Ipsos conducted a poll on the Black Lives Matter movement, revealing that 43% of Americans support it, and 38% oppose it. In 2021, Ipsos conducted another poll on COVID-19 policy issues, which showed that 72% of respondents support mask mandates, 61% support mandatory vaccinations with medical or religious exemptions, and 39% oppose it (Jackson and Newall 2021a, 2021b). In addition, a poll conducted by the Public Religion Research Institute (PRRI) in 2021 found that 76% of respondents support LGBTQIA rights, while 19% oppose it (PRRI Press Releases 2021). Table 1 summarizes our dataset on brands’ tweets, user replies, and Twitter mentions for the brands from users on the three sociopolitical issues of our analysis.

Table 1 Information on the dataset on brand activism and Twitter mentions
Full size table

We investigate brands’ and users’ day-to-day online responses on a 1769-day timeline from January 1, 2017, to November 5, 2021, for “Black Lives Matter” (BLM) and LGBTQIA rights-related tweets. Since the first COVID-19-related tweet from brands in our list appeared in 2020, we use a 674-day timeline from January 1, 2020, to November 5, 2021, for our analysis of tweets related to COVID-19 policies. We scrape all the tweets our 177 global brands posted on Twitter over the specific timeline for each of the three chosen sociopolitical issues. We found 975 BLM-related tweets from 124 brands, 1312 LGBTQIA rights-related tweets from 123 brands, and 2217 COVID-19 policy-related tweets from 101 brands from our list. For each of these three sociopolitical issues, tweets were filtered using dictionaries of related hashtags (e.g. #BLM, #LGBT, #Covid19, etc.) and phrases (e.g. police brutality, LGBTQIA rights, mask mandates, etc.) specific to the sociopolitical issues (Appendix). We gathered these hashtags using the online social media toolkit “best-hashtags.com” and “RiteTag.com”, which offer Twitter’s top trending hashtags on a given topic based on users’ real-time hashtag engagement. The hashtags and phrases represent user engagement from both sides of the political spectrum to portray the polarized environment surrounding all three sociopolitical issues.

Empirical model and key variables

This study aims to investigate whether brands engage in activism as a response to the recent activism of their peer brands on social media. To test the predictions of the study in the context of Twitter, we examine whether or not the prior tweets from peer brands on a divisive sociopolitical issue affect a brand’s decision to engage in activism on the issue on a given day of our timeline. Our study uses a fixed-effects logistic regression model that allows for unobserved time-invariant individual heterogeneity across the brands (such as variation in the brands’ objectives and core values). The model has a binary dependent variable, yi,d, that records whether a focal brand, i, from our list of 177 brands tweets on a specific sociopolitical issue (for example, BLM) on a given date d of the timeline. Thus, for parameter p = P(yi,d = 1), we can present our fixed-effects logistic regression model as the following:

$$ \begin{gathered} \log \frac{p}{1 - p} = \alpha_{i} + \beta_{1} {\text{Peer}}\;{\text{tweets}}_{i,d} + \beta_{2} {\text{Industry}}\;{\text{peer}}\;{\text{tweets}}_{i,d} + \beta_{3} {\text{Positive}}\;{\text{replies}}_{i,d} \hfill \\ \quad \quad \quad \quad \; + \beta_{4} {\text{Twitter}}\;{\text{mentions}}_{i,d} + \beta_{5} {\text{Average}}\;{\text{reply}}\;{\text{count}}_{i,d} + \beta_{6} {\text{Google}}\;{\text{Search}}_{d} \hfill \\ \end{gathered} $$

Here, \({\alpha }_{i}\) represents a shorthand for a set of dummy variables we used to incorporate the fixed effects for each brand i of our list that are each multiplied by their respective regression coefficients (except for one reference brand to avoid perfect collinearity). This accounted for brands’ idiosyncratic characteristics (such as existing commitment to racial equity embedded in its core values) that could motivate them to tweet on the focal sociopolitical issue (irrespective of what their peers are doing).

For a focal brand i and a given date d of our timeline, we have the following independent variables of interest with respective coefficients denoted as β1 through β6:

\({\mathrm{Peer\, tweets}}_{i,d}\) is the count of activist peer tweets on a particular sociopolitical issue that were posted within the last three days from a given day by any peer brand from the list of 177 brands. This includes tweets from both industry peers and non-industry peers during the 3-day period. The 3-day window is based on an initial exploratory analysis relating peers’ activist tweets to the likelihood of a focal brand’s activism on a given day. We created three separate variables reflecting the count of peers’ activist tweets based in the last 24 h, the last 25–72 h, and the last 73–120 h prior to a given day. Across all three sociopolitical issues, the data suggest an exponential decay in peer effects. While peer influence is the highest during the preceding 24 h from a given day (effect size across the three issues ranges from a statistically significant 11% to 17% increase in the odds of a brand engaging in activism), the influence exponentially decreases for peer tweets posted during 25–72 h prior to the given day (only a 2% to 7% increase in the odds of a brand engaging in activism). Beyond the preceding 72 h (i.e. 3-day window), the peer influence levels off rapidly, presumably due to the rapid pace at which Tweets are posted and responded to. Based on this initial analysis, we use a preceding 3-day window from a given day to study the influence of peer brands’ prior tweets, user replies on those tweets, and user mentions on a brand’s decision to engage in activism on a given day.

\({\mathrm{Industry\, peer\, tweets}}\,_{i,d}\) is measured by the percentage of the last three days’ tweets on a sociopolitical issue that specifically came from industry peers. We identify industry peers based on the industry as identified by the brand rankings.

\({\mathrm{Positive\, replies}}\,_{i,d}\) is the percentage of positive user replies to the prior sociopolitical tweets posted during the last three days by any of the 177 peer brands regardless of their industries. By implementing sentiment analysis, we measured the percentage of positive replies to peers’ tweets on the sociopolitical issue. The analysis scores the polarity of users’ replies on a scale from -1 to + 1.

\({\mathrm{Twitter\, mentions}}\,_{i,d}\) is the count of Twitter users’ public appeals to all the 177 brands in our list to address an issue during the 3-day window. We measure Twitter mentions by counting how often consumers directly mention one or more brands from our list in their tweets while also using specific hashtags for a sociopolitical issue.

\({\mathrm{Average\, reply\, count}}\,_{i,d}\) is the average count of replies to the peers’ prior tweets on the issue. We were also concerned that a bias might occur due to possible variations in the volume of users’ replies to the sociopolitical tweets of the brands. Therefore, we included the average number of replies to the last three days’ sociopolitical tweets as a control variable for our model.

\({\mathrm{Google\, Search}}_{d}\) measures the relative popularity of a search term or topic on Google during the 3-day window. We included this control measure due to concerns that the model could reflect an overall uptick in public attention to the sociopolitical issue rather than the effect of peer activism. To account for this, we include the Google search trends data on the focal issue to reflect the popularity of a topic on Google’s search engine during a given timeline (for example, Google searches for “Black Lives Matter” increased in the days immediately following the killing of George Floyd in 2020).

Results

Table 2 presents two models for each of the three sociopolitical issues. The first model for each respective sociopolitical issue is a simple model which tests only the effects of the count of prior sociopolitical tweets on the likelihood of a focal brand tweeting on a specific issue on a given day. The simple model includes brand fixed effects and the number of Google searches during the last three days. The second (full) model includes all hypothesized variables and controls.

Table 2 Logistic regression results on a simple and a full model
Full size table

Hypothesis 1 predicted that a brand would be more likely to post a sociopolitical tweet on a given day the more peers had tweeted on the issue in the three days prior. We find robust support for this prediction. Across all three issues and in each model, the effect of peer tweets is positive and significant (coefficients range from 0.111 to 0.030; p < 0.01 in each case.) This suggests that all else being equal, a brand is more likely to tweet on a divisive sociopolitical issue on a given day when there have been prior tweets on the issue from peer brands in the last three days. The results provide evidence for our main argument that brands react to each other on social media on divisive sociopolitical issues.

Hypothesis 2 explored the effect of activist tweets from only the industry peers of our focal brand on its likelihood to engage in brand activism. We do not find support for this hypothesis. Across the three sociopolitical topics, the coefficients range from 0.004 to 0.002, and none of them reach significance at the p < 0.05 level. This suggests that it is as likely for a brand to look beyond its industry as it is to respond to an industry competitor when weighing whether to engage in brand activism.

Having established that brands do indeed respond to peer brands’ activism, we looked for evidence that managers may weigh the risk of speaking up or the risk of staying silent. Hypothesis 3 predicted that the percentage of positive user replies to the peer brands’ prior sociopolitical tweets would signal that it may be safe to engage in brand activism. The data support this notion. The effect of positive replies is positive and significant for all three sociopolitical issues, although the coefficients are quite small. (Coefficients range from 0.022 to 0.008; p < 0.01 in each case.)

To investigate the notion that expectations heighten the risk of staying silent, we investigate the Twitter mentions to all brands on the three sociopolitical issues for Hypothesis 4. We find support for this hypothesis across all three issues; as Twitter mentions increase, so does the likelihood that a brand will engage in activism. This suggests that brands are sensitive to rising expectations or pressures from consumers around a sociopolitical issue. The size of the effect was significant, yet varied widely by issue (bBLM = 0.115; bLGBTQIA = 2.159; bCOVID = 0.033; all effects p < 0.01).

Post Hoc Analysis

This article uses the count of sociopolitical tweets from peer brands as an antecedent of the brand activism of a focal brand. However, a potentially interesting follow-up question is whether the mere presence of a peer tweet on a sociopolitical issue could provide sufficient information to allay concerns that brand activism is a viable strategy. To gain more insight into these dynamics, we conducted a post hoc analysis bifurcating the count of peer tweets into two new variables: (a) mere presence of a peer tweet and (b) additional count of peer tweets. The former is a binary indicator showing the presence of at least one related tweet from the peer brands in the three-day window. The latter is the total count of additional tweets from the peer brands during the time window; this ranged from 0 to a maximum of 90 additional tweets in the last three days from a given day.

Models 7–9 in Table 3 display the results from the post hoc analysis, which includes the additional two variables. We find evidence that both variables have explanatory power. The mere presence of peer tweets is significant for all three sociopolitical issues. The coefficients range from 2.132 to 0.558 (all p < 0.01). This suggests that a single peer tweet can provide enough information about the risk involved in speaking up on an issue. However, we also find evidence that, beyond the first tweet, each additional activist tweet provides incremental explanatory power. In all three sociopolitical issues, the additional count of peer tweets has a positive and significant effect on a brand’s likelihood to engage in activism. The coefficients range from 0.073 to 0.029 (all p < 0.01). This suggests that as more and more other brands tweet on an issue, the focal brand continues to monitor and becomes more likely still to tweet themselves.

Table 3 Logistic regression results for the post hoc analysis
Full size table

Overall, the results from our empirical models provide evidence for our argument that a brand is more likely to engage in brand activism on a divisive issue on social media as a response to the recent relevant activism of its peer brands on the platform. In addition, we provide evidence of external validity for our model by using a wide range of heterogeneous brands for three different polarizing topics over five consecutive years.

General discussion

Managers face considerable risk when deciding whether or not to engage in social media activism on divisive issues such as BLM, LGBTQIA rights, or COVID-19 policies. Prior research has examined the consequences of those actions; however, there has been a lack of attention to how managers make these decisions in the first place. This research examines brand behaviour as a function of market signals both from peer brands and their respective consumers. The data support the notion that managers attend to the activism of peers as a way to gain market-driven intelligence about the potential risks and benefits of taking a stand on such issues; a brand is more likely to tweet on an issue on any given day to the extent that other brands have recently tweeted on that issue. The evidence also suggests that managers attend to consumer social media activity in response to the issue; the more consumers react positively to activist tweets from other brands, and the more consumers call out brands for not tweeting, the more likely a brand is to issue an activist tweet.

Peer influence for non-market strategies

The extant literature has mainly emphasized peer influence on business-related decisions (i.e. market strategies) such as product design or financial policies (e.g. Lieberman and Asaba 2006; Aral and Walker 2011; Leary and Roberts 2014). We extend this prior research to the non-market strategy domain, specifically, posting public statements on a sociopolitical issue on social media. This insight invites research on other forms of non-market strategy. For example, it seems plausible that a company may lobby the government more if the lobbying expenditures of its competitors increase. Future research may also examine whether the extent to which a company’s political position is liberal or conservative is affected by the position of its peers.

Interestingly, while we find significant peer effects in all three issues in the study, the effect sizes vary somewhat, with the strongest effects in BLM (bBLM = 0.09) and LGBTQIA rights (bLGBTQIA = 0.08) and a somewhat smaller effect for COVID policy (bCOVID = 0.03). This implies that the intrinsic nature of an issue may have an important role to play in brands engagement in activism. For example, it is plausible that managers may rely on information from peer activism less for COVID policy because there were more regulatory constraints and federal guidelines on that particular issue. Prior research suggests that firms often adapt their public messaging in a regulated environment and rely on federal policies when they address a topic (Kolsarici and Vakratsas 2010). In our context, some brands may address COVID-19 policies on their social media platforms following Centers for Disease Control and Prevention (CDC) guidelines (such as offering incentives for employees to get vaccinated), regardless of what their peers are doing (Miller 2021).

On the other hand, the differences in peer effects could also be due to the fact that managers may deem some issues to be riskier than others. As a result, the additional risk could lead them to rely more on social media signals regarding growing customer expectations surrounding the topic. This may also explain why the brands show varying levels of sensitivity to consumers’ online mentions or direct appeals to them to address each of the issues (bLGBTQIA = 2.159; bBLM = 0.115; bCOVID = 0.033). The variation in peer effects invites future research on whether differences in the effect sizes are due to constraints or the need to assess the risks of addressing an issue.

Peer brand-customer interactions as sources of market information

This article also builds upon existing literature on brand activism. For example, much of the literature on brand activism assesses the consequences of political engagement, such as financial returns, market performance, or consumers’ attitudes toward the brand (Hadani and Schuler 2013; Lacka et al. 2021; Bhagwat et al. 2020; Dodd and Supa 2015, etc.). Other research on the consequences of brand activism focuses on the characteristics of the activism itself, such as its perceived authenticity (Sobande 2019; Vredenburg et al. 2020; Mukherjee and Althuizen 2020; Sibai et al. 2021). In contrast, we focus on the antecedents of brand activism decisions on social media. Our findings suggest that, given the short decision period and the difficulty in procuring market data from other means, brand managers use social media as a tool to gather intelligence about the behaviour of peers and customers. Notably, they consider not only the behaviours of peers and customers separately but also the dynamic interplay between those peers and their customers.

Using online mentions data to evaluate growing customer expectations

We find that brands have a higher propensity to engage in activism when consumers (i.e. members of the public on the social media platform) appeal directly to brands to become more outspoken on a sociopolitical issue. This suggests that managers are keenly aware that consumers’ expectations raise not only a brand’s risk of taking a stand but also the risk of not taking one. We construct a novel measure to assess these expectations, social media mentions and hashtags, and find that they do, in fact, affect brand behaviour. Thus, the present research moves beyond prior research that has looked at firm-generated content or user-generated content (Kumar et al. 2016; Tirunillai and Tellis 2012; Sonnier et al. 2011 etc.). Our approach suggests that direct social media appeals to brands may be used to examine customer expectations of product development or the design of service encounters. For example, one could envision a study that examines whether appeals to improve a product results in more significant changes to the product the next year that it is released.

Managerial implications

Due to the increasing political polarization and rising rhetoric related to brand activism, our findings are timely and relevant to marketing scholars and practitioners. Our results provide evidence that brands may benefit from examining publicly available comments on prior activism campaigns to gauge the overall market sentiment on the issue in general. Therefore, managers may extract valuable information on the risk of taking a stand on a divisive issue from the positive and negative chatter from the audience. And while doing so, brands may avoid the myopic approach of focusing only on their industry peers. Although monitoring industry peers is a standard approach when brands decide on their market or investment strategies (Scharfstein and Stein 1990; Aral and Walker 2011), brands may benefit from gathering as much information as possible from peer brands across industries when weighing the risk of speaking up and the risks of staying silent in part. Therefore, they may look for all the social media brands engaged in relevant conversation across industries when deciding on addressing a sociopolitical issue.

Furthermore, our findings demonstrate that social media users’ online remarks can assist brands in measuring customer expectations at the time and deciding when to make a statement on an issue. Overall, our findings offer significant insights into how, despite the risks, brands can leverage data from peer brands’ recent activist statements on social media to determine how to respond to a contentious sociopolitical topic with their own activism strategies.

Limitations and future directions

This paper has sought to map the varied factors that may affect a brand’s likelihood of addressing a divisive sociopolitical issue on social media on a given day. Based on the effect size of the statistical results supporting our arguments, we find that the likelihood of a brand’s decision to conduct social media activism was substantially larger for the increased risk of inaction than the reduced risk of speaking up. This invites further research to determine whether one of the two forces has a bigger influence on a brand’s decision to speak up on a contentious sociopolitical issue.

While this study used multiple control variables and fixed effects in the model to minimize potential biases and concerns, further opportunities to strengthen the generalizability of the findings may exist. While we conducted the research in the context of Twitter, future research might expand the analysis on other platforms such as Facebook, Instagram, and TikTok. According to a Pew Research Center survey conducted in 2021, the demographics of different social media sites’ users are vastly different (Auxier and Anderson 2021). This means that brands’ digital marketing strategies on each platform may slightly differ depending on their target audience. Future researchers may address these differences and expand the analysis on other online platforms.

Moreover, this study examines the peer influence of brand activism by exploring their commitment to divisive sociopolitical issues via public social media statements (the intangible form of activism). Future research might explore whether the influence of peers’ such activism on social media spills over to other firms’ tangible brand activism initiatives as well, such as contributing money to a sociopolitical cause or changing corporate practices.

In addition, other factors that operate in tandem with peer influence and brand activism may be interesting to further investigate the competitive dynamics of brand activism. For example, future work might examine whether brands lookout for specific features in their peers before deciding whether they should imitate their strategies. Examples of such factors include market share, social media presence, prior history of a brand being cancelled on social media, etc.

A final question may be about minimizing brands’ risks when speaking out on contentious issues. While this study does not further investigate peer brands’ role in managing the risks, researchers may be interested in studying whether recent peer activities could help other brands reduce the risks of alienating certain customer groups. For example, practitioners might be particularly interested in determining the extent to which they may choose one side of a polarizing issue that may align with one segment of their target customers while alienating others.

Appendix

Topic: Black Lives Matter

  1. (1)

    Dictionary of relevant phrases: Black community, racial injustice, systemic racism, police brutality, amplify Black voice, Black history, racial equality, George Floyd, say her name.

  1. (2)

    Dictionary of relevant hashtags: #blacklivesmatter, #blm, #icantbreathe, #stopracism, #sayhername, #saytheirname, #blmmovement, #georgefloyd, #takeaknee, #blackvoices, #untilweallwin, #blackhistorymonth.

Topic: LGBTQIA rights

  1. (1)

    Dictionary of relevant phrases: lgbt, gay rights, love is love, lesbian, queer, bisexual, transgender, gay pride, lgbt pride, non-binary, pride month, lgbt community, marriage equality.

  1. (2)

    Dictionary of relevant hashtags: #lgbt, #transrightsarehumanrights, #lgbtcommunity, #pridemonth, #loveislove, #lgbtqpride, #equalityact, #gaypride, #lovewins, #marriageequality, #bornthisway, #transrights.

Topic: COVID-19-related policies

  1. (1)

    Dictionary of relevant phrases: COVID-19 mask mandate, anti-mask, vaccine propaganda, coronavirus mandate, mask propaganda, COVID-19 vaccine, get vaccinated, COVID vaccine, COVID-19, lockdown, plandemic.

  1. (2)

    Dictionary of relevant hashtags: #maskssavelives, #antimasker, #maskpropaganda, #maskon, #masksdontwork, #maskskill, #nomask, #maskup, #covidmask, #fightagainstcovid, #wearamask, #fightagainstcoronavirus, #vaxxie, #getvaccinated, #covidvaccine, #scamdemic, #plandemic, #novaccine, #covidiots, #stayhome, #covid19, #socialdistancing.