Towards a typology of negative engagement behavior in social media

ABSTRACT

Extant literature on consumer engagement has focused on positive manifestations of the construct, rather than on its negative dimension. Yet, many brand interactions are negative in nature. The purpose of this conceptual study is to develop a typology of negative engagement behavior in social media by using the multi-grounded theory (MGT) approach on a sample of 12,429 tweets extracted from Twitter. The analysis shows that negative engagement behavior with a brand or service provider as the object focus or target can be categorized according to (a) the manifestation of the negative engagement and (b) the emotional intensity of the negative engagement. Four categories of negative engagement behavior (NEB) were identified. These include negative review writing, justice-seeking complaining, retaliation acts, and firestorming. The study concludes that an understanding of the different categories of negative engagement in social media is essential if service providers are to effectively address and respond to different forms of consumer sentiment.

摘要

现有的关于用户参与的文献往往关注其概念的积极层面，而不是它的消极维度。然而许多品牌的互动在本质上是消极的。本研究以12429条推文为样本，运用多重扎根理论方法(MGT)，研究社交媒体中的负面参与行为类型。分析表明，以品牌或服务供应商为关注对象的负面参与行为可分为(a) 负面参与的表现形式; (b)负面参与的情绪强度。负面参与行为被确认为四类, 包含‘写负面评论’、‘为寻求公正而引发的抱怨行为’、‘报复行为’和‘激烈抨击’。本研究的结论是，如果服务供应商想要有效地处理和回应不同形式的消费者情绪，了解社交媒体中不同类别的负面参与是必不可少的。

KEYWORDS:

• Consumer engagement
• negative engagement
• valence
• typology
• social media
• Twitter

关键词:

• 用户参与
• 负面参与
• 诱发力
• 类型学
• 社交媒体
• 推特

Introduction

Social media platforms have facilitated the development and dissemination of consumer-led user-generated content, and hence, they have altered consumer–brand relationships (Carlson et al., 2019; Hollebeek & Macky, 2019; Read et al., 2019; Shahbaznezhad et al., 2021). Social media enables consumers to engage in mutual, co-creative communications (Kaplan & Haenlein, 2010; Wong, 2021), whilst supporting rich networking opportunities between brands and consumers (Islam & Rahman, 2016; Labrecque, 2014; Pansari & Kumar, 2017; Rather, 2019). Social media also positively and significantly effects engagement between consumers and brands (Rather, 2021a). From the perspective of service provider, social networking provides the opportunity to leverage the constellation of connections and drive consumer engagement (Hollebeek et al., 2022; Ibrahim et al., 2017; Verma & Yadav, 2021).

Yet, not all brand interactions that occur within social media are positive. Negative consumer relationships are in fact more common than positive relationships (Fournier & Alvarez, 2013), and both positive and negative engagement towards a brand may even occur simultaneously (De Villiers, 2015). Negative information has been found to be increasingly common in social media and on online review sites (Zhao et al., 2020), and to weigh more heavily in consumers’ consumption judgements than positive information (Tan & Chen, 2022). Studies also show that the impacts of positive reviews are neglected compared to negative reviews and it is possible that positive reviews are more frequent than negative ones (Shi et al., 2021). Understanding how and why consumers negatively engage with brands is critical for preventing customer loss and maintaining brand equity (Hollebeek & Chen, 2014; Obilo et al., 2021), especially now when more customer incivility is present (Yoon, 2022). Moreover, as the length of the relationship is an important factor (Al-Hawari, 2022), long-term engagement between brands and consumers requires continuous learning, and innovative and adaptive approach towards evolving technology and social media (Hollebeek et al., 2019).

While it is not a new phenomenon for brands, research has recently examined the concept of negative engagement (Dolan et al., 2016; Hollebeek & Chen, 2014; Naumann et al., 2017; Van Doorn et al., 2010). Negative engagement is defined as ‘consumers' unfavorable brand-related thoughts, feelings and behaviors during focal brand interactions’ (Hollebeek & Chen, 2014, p. 63). Depending on its intensity, it can have significant damaging effects on brands through value co-destruction (Hollebeek et al., 2016; Naumann et al., 2017), declining financial performance (Juric et al., 2016), reduced consumer value (Van Doorn et al., 2010), and negative word-of-mouth recommendations (Hollebeek & Chen, 2014). It may be especially contagious within social media given the highly networked nature of the medium and given that informal consumer-generated information is often perceived to be more reliable when compared to formal brand communications (Richins, 1984).

Whilst prior research has expanded the domain of engagement and its valences, more research is required to examine the potentially heterogenous nature of negative engagement and the categorization of its behavioral, emotional, and cognitive dimensions (see e.g. Li et al., 2018; Naumann et al., 2020). This is an important gap in the literature since the inherent nature of it may differ from positive expression of engagement (Dolan et al., 2019). In addition, despite the adverse effects of negative engagement on brand performance, very few studies have attempted to catalog negative forms of engagement (e.g. Azer & Alexander, 2018, 2020; Dolan et al., 2016). This is important since Morgan and Hunt (1994, p. 33) note that ‘just as medical science should understand both sickness and health, marketing science should understand both functional and dysfunctional relationships.’

This conceptual paper makes two main contributions. First, it examines the behavioral nature of negative engagement. This study suggests that negative engagement has its own, unique characteristics in terms of its manifestations.

Second, four new categories of negative engagement behavior (NEB) are proposed, which are categorized according to (a) visible manifestation (e.g. justice-seeking, complaining, discrediting) and (b) the emotional intensity of the negative engagement. To achieve this, the study employs multi-grounded theory (MGT) to develop a typology for different categories of NEB within social media. This contributes to engagement scholars’ needs from the perspective of theory formation as well as through novel categorization.

This paper is organized as follows. First, a detailed overview of the literature on positive and negative engagement is presented, providing the basis for the development of the conceptual model. Secondly, the methodology is presented. Next, the data set, which includes over 12,000 consumer tweets from 6 telecommunications brands across 2 cross-cultural contexts, namely Australia and Finland, is presented. Finally, four categories of negative engagement behaviors are proposed namely negative review writing, justice-seeking complaining, retaliation acts, and firestorming. The paper concludes with conceptual and managerial implications arising from this study as well as suggestions for future studies.

Theoretical framework for negative engagement

Consumer engagement is a dynamic and reciprocal concept that represents a consumer’s cognitive, affective, and behavioral investment in an organization’s offerings (Brodie et al., 2013; Hollebeek & Chen, 2014; Vivek et al., 2012). To date, the extant literature has focused largely on positive manifestations of consumer engagement; however, research has begun to question how the assumed static nature of engagement itself may exhibit different intensities and, in particular, different valences (Obilo et al., 2021; Palmatier et al., 2013; Pansari & Kumar, 2017; Rather & Hollebeek, 2021).

Negative engagement has received increasing attention in the literature (see, e.g. Azer & Alexander, 2020; Dolan et al., 2015; Vargo, 2016). Negative customer engagement is considered as a negative valence of customer engagement (Bowden et al., 2016; Dolan et al., 2016; Hollebeek & Chen, 2014; Naumann et al., 2017; Van Doorn et al., 2010). It is defined as a customer’s brand-related unfavorable thoughts, emotions, and behaviors (Hollebeek & Chen, 2014), that cause negative consequences for the brand or firm (Van Doorn et al., 2010), and which lead to a negative orientation towards the brand or firm (Bowden et al., 2016; Dolan et al., 2016; Naumann et al., 2017). Whilst the extant literature to date has tended to focus on how brands engage consumers/users within social media-led communications to create value (Rather, 2021b; Rather et al., 2022), research also notes that some brand-related thoughts, sentiment, activities or behaviors are intended to harm the brand (Rather & Hollebeek, 2021). Rather and Hollebeek (2021) cite examples of this including establishing anti-brand communities and disseminating negative brand-related word-of-mouth which has the effect of impairing brand health. Additional research is required to explore this important issue.

Negative engagement is noted to be a process-driven concept similar to positive consumer engagement, where certain triggers, such as perceptions of a lack of distributive, interactional, and procedural justice, lead to negative consumer reactions (Do et al., 2021; Van Doorn et al., 2010; Vivek et al., 2012). In this sense then, negative engagement involves activation, immersion, and passion but at the unfavorable, negative end of the spectrum (Hollebeek & Chen, 2014). It manifests itself through unfavorable cognitive (e.g. negative bias, cynicism), emotional (e.g. hatred, fear, resentment, shame, humiliation), and behavioral inclinations (e.g. negative reviews and boycotting) (Juric et al., 2016). In addition, it may be collective in nature (Naumann et al., 2017). Importantly, negative engagement differs from more passive states of engagement with negative triggers, such as disengagement (see, e.g. Bowden et al., 2015, 2016; Goode, 2012), as it is more deliberate, participative, and public in nature (Juric et al., 2016) and can have potentially devastating effects on brand equity, reputation, and short- and long-term financial performance (Hollebeek & Chen, 2014).

Research on negative engagement has tended to focus on the conceptualization of its cognitive and affective dimensions (Bowden et al., 2016), including the effects of online firestorms on organizational reputation (Pfeffer et al., 2014) and types of passive and active anti-brand behavior (Kucuk, 2008, 2010, 2015). Previous studies have shown that negative engagement may be context-specific (Naumann et al., 2020). Recently, some attempts have been made to classify and categorize negative engagement expressions into typologies (see, e.g. Azer & Alexander, 2020; Kucuk, 2008; Vargo, 2016).

The main contribution of this paper is to advance the nascent literature on negative engagement by examining (a) the visible manifestation of the NEB (e.g. justice-seeking, complaining, discrediting), and (b) the emotional intensity of NEB. In order to achieve this, this paper examines the inherently varied, heterogenous qualities of NEB for six telecommunications brands on the social networking site Twitter.

Towards typologized categories of NEB

This study adopts a tri-dimensional framework of negative engagement given that it is the most widely accepted conceptualization within the literature (Dessart et al., 2015; Hollebeek & Chen, 2014; Naumann et al., 2017, 2020). In other words, negative engagement shares the same drivers and dimensions (affect, cognition, and behavior) as positive engagement, but the operation of the dimensions is ultimately context-specific and distinct (Juric et al., 2016). Negative engagement may occur where consumers experience negative emotions arising from unpleasant surprises or where consumer expectations are not met to a surprising degree (Liu & Keh, 2015), and according to regulatory engagement theory (Higgins & Scholer, 2009), it may subsequently lead to attraction or repulsion behavior with regard to the focal object or target (Hollebeek & Chen, 2014).

The affective dimension of negative engagement comprises an individual’s negative emotional reactions towards the engagement focus. Emotional triggers, such as disappointment and insecurity, may elicit negative consumer behaviors (Azer & Alexander, 2018). Naumann et al. (2020) argue that feelings of anger and dislike are a common affective expression of negative engagement when customers hold negative emotions towards a brand relationship. Azer and Alexander (2020) note that emotional expressions that are embedded within the narrative of online reviews act to powerfully shape other consumers’ evaluations. While Gruber et al. (2020) note that people often participate in online firestorms due to social and collective reasons, emotions also play a role in the process. In addition, Goode (2012) found that consumers who responded emotionally in dealing with brand transgression events were more likely to negatively engage with a brand. The emotional dimension of NEB is thus an important defining feature.

The cognitive dimension of negative engagement combines an individual’s experiences, interest, and attention in relation to a focal engagement object. Cognitive triggers, such as service failure, overpricing, and deception, have been found to cognitively trigger NEB (Azer & Alexander, 2018). Moreover, Brodie et al. (2013) note that engagement triggers are usually a result of direct or indirect experiences with regard to the target brand or product. In addition, Naumann et al. (2020) argue that when consuming negative brand information, higher levels of cognitive processing are dedicated to the brand by these individuals. Cognition thus represents a beneficial basis through which to understand the nature of NEB.

The behavioral dimension of negative engagement is commonly expressed as active, deliberate, and purposeful actions taken towards the engagement object (Naumann et al., 2020). Juric et al. (2016) argue that manifestations of negative engagement usually entail diverse states and behaviors in which dynamic and iterative relational-exchange processes are present. The value is usually co-destructed in relation to the brand, but also in terms of relationships with other engagement actors within an ecosystem. In addition, visible NEB may carry different consequences for focal engagement objects, depending on who engages negatively (Zhou et al., 2019). These focal engagement objects in social media often include brand, its products, and/or its services. In fact, social media has gained more importance in strategic portfolios of brands and organizations (Li et al., 2021). It plays a significant role in consumer behavior and likelihood of co-creation, and it also mediates the relationship and trust between consumers and brands (Rather, 2021a; 2019; Rather et al., 2019). Moreover, consumers can distort brands in social media by re-creating images in ways that harm brand reputations and through the exploitation and recruitment of other actors towards a collective, common cause (Gebauer et al., 2013). We argue that this dimensionality of NEB is important when considering the development of negative engagement typologies.

Materials and methods

This study adopts a multi-grounded theory (MGT) approach (Goldkuhl & Cronholm, 2010). MGT is a combination or dialectical synthesis of inductive and deductive coding (Axelsson & Goldkuhl, 2004). This approach is based on combining empirical grounded data, pre-existing theory, and internal grounding, which refers to ‘an explicit congruence within the theory itself’ (Goldkuhl & Cronholm, 2010, p. 192). The formulation of the typologized categories of NEB and the data analysis were guided by the interplay between empirically driven analysis (‘inductivism’) and theory-driven analysis (‘deductivism’), which eventually led to a combined view and synthesis of the finalized categories of NEB that arose from the analyzed data (Goldkuhl & Cronholm, 2010).

We commenced the process by generating a theory-guided analysis to better understand the separation between positive and negative engagement behaviors. This theoretical grounding (Goldkuhl & Cronholm, 2010) focused on previous literature and studies on positive consumer engagement, and negative consumer behaviors online (e.g. complaining, negative user reviews). After the establishment of the different valences of engagement behaviors, we continued by examining the empirical data by harvesting user-generated content via Twitter. We undertook inductive coding without initial categorization to ensure that this process was ‘as free as possible from precategorizations,’ as per the recommendations of Goldkuhl and Cronholm (2010, p. 194).

Sample

Telecommunications companies were selected due to their active customer service and poor reputations (an average score of 6.0/10) among consumers (Brand Finance, 2020). A total of 12,429 tweets were collected from 6 different telecommunications companies and 9 brand accounts across 2 cross-cultural and comparable contexts, namely Australia and Finland. Despite them both being high technology countries with great telecommunications infrastructures, Australia and Finland represent very different markets when it comes to language, population, coverage, and competition.

TweetArchivist, a tool for tracking tweets was utilized to collect the data, including the content of each tweet, who it originated from, and how many likes the tweet had. The data of the Australian telecommunications sample consisted of 11,126 tweets from 3 companies’ Twitter accounts: 2136 tweets from the account of company number 1, 8325 tweets from the account of company number 2, and a total of 665 tweets from the 2 accounts of company number 3. With regard to the Finnish telecommunications companies, the data consisted of 1303 tweets from 3 companies’ Twitter accounts: 575 tweets from the account of company number 4, a total of 509 tweets from the 2 accounts of company number 5, and 219 tweets altogether from the 2 accounts of company number 6.

Data analysis and coding

The data were compared with the results of theoretical grounding. In this stage of conceptual refinement, existing theories and the empirical data were examined for commonality. Conceptual refinement is characterized as a stage of actively working with clarifying the used concepts so that important concepts are ‘assessed and continually refined during theorizing’ (Goldkuhl & Cronholm, 2010, p. 194). During this stage, those tweets that were either positive or neutral in content were identified as not being NEB, and they were excluded. This refinement led to the inclusion of a total number of 4501 NEB tweets and the exclusion of 7928 tweets. The NEB tweets therefore accounted for approximately 36% of all the tweets analyzed within the data set.

The next stage of data analysis involved pattern coding (Goldkuhl & Cronholm, 2010). We examined how NEB manifested itself behaviorally by analyzing the message content of the negative tweets. NVivo was employed as a qualitative engagement tool to explore different NEB categories based on their emotional intensity. We also employed theoretical grounding to inform whether a categorization like this already exists. Explicit grounding was then employed in conjunction with specific analysis actions, including theoretical matching, explicit empirical validation, and the evaluation of theoretical cohesion, within our categorization. This was accomplished by examining the visible manifestation, object focus or target, and the orientation of the NEB in each tweet. This was also combined with the examination of emotional intensity from the prior stage of analysis.

The coding and analysis were finalized by naming the categories and explaining them. In this stage, we also identified and finalized the action patterns of the typology (Axelsson & Goldkuhl, 2004). As a result of the explicit grounding, we identified a typology, which consisted of four main categories of NEB, including negative review writing, justice-seeking complaining, retaliation acts, and firestorming. We were also able to condense the theory and propose a conceptual NEB definition, utilizing the data and final coding scheme. The results of this analysis are discussed next.

Results and discussion

The numerical count analysis demonstrates that the amount of NEB varies between the telecommunications brands and across countries. Table 1 presents the amount of consumer NEB on the Twitter accounts of the Australian (companies 1–3) and Finnish (companies 4–6) telecommunications companies. Due to the sensitive nature of the data, the names of the companies are not provided.

Table 1. Number of tweets and the amount of NEB within the data.

	TeleCom Company AUS 1	TeleCom Company AUS 2	TeleCom Company AUS 3	TeleCom Company FIN 1	TeleCom Company FIN 2	TeleCom Company FIN 3
NEB tweets as the percentage of all the tweets analyzed from an account	43.2%	38.7%	14.7%	22.4%	17.7%	19.2%
	(n: 922)	(n: 3220)	(n: 98)^a	(n: 129)	(n: 90)^a	(n: 42)^a
Negative review writing	12.5%	14.3%	3.2%	5.9%	6.1%	9.1%
	(n: 266)	(n: 1189)	(n: 21)^a	(n: 34)	(n: 31)^a	(n: 20)^a
Justice-seeking complaining	24.1%	20.0%	8.7%	15.3%	10.0%	9.6%
	(n: 515)	(n: 1661)	(n: 58)^a	(n: 88)	(n: 51)^a	(n: 21)^a
Retaliation acts	5.1%	2.3%	2.3%	0.9%	1.0%	0.5%
	(n: 108)	(n: 192)	(n: 15)^a	(n: 5)	(n: 5)^a	(n: 1)^a
Firestorming	1.5%	2.1%	0.6%	0.3%	0.6%	0%
	(n: 33)	(n: 178)	(n: 4)^a	(n: 2)	(n: 3)^a	(n: 0)^a

^a Data compiled from 2 corporate Twitter accounts of the same company.

Based on our empirical results, we established and proposed four typologized categories of NEB that are targeted at a brand and expressed within social media networks based on visible manifestation and emotional intensity. Justice-seeking complaining was the most common NEB across all companies. This was followed by negative review writing, retaliation acts, and firestorming, respectively, across both the Australian and Finnish telecommunications companies. The four typologised categories of NEB are explained in more detail in Table 2.

Table 2. Four categories of negative engagement behavior (NEB) of varying intensity that are targeted at brands in social media.

NEB Category	Category 1 Negative review writing	Category 2 Justice-seeking complaining	Category 3 Retaliation acts	Category 4 Firestorming
	* Low emotional intensity	* Moderate emotional intensity	* High emotional intensity	* Very high emotional intensity
	* Information focus	* Justice focus	* Revenge focus	* Destructive focus
Affective dimension	NEB reflects a less activated and aroused state focused on negative information about a brand within a social media network.	NEB reflects less activated dissatisfaction and disenchantment with a brand within a social media network.	NEB reflects a highly activated and aggressive state with the intention of ‘getting even’ with a brand within a social media network.	NEB reflects a highly activated state of anger and intense indignation within a social media network.
Orientation	Achievement of brand experience enhancement for both the brand and other engagement actors.	Achievement of personal goals against the brand.	Reprisal against the brand for personal gain.	General call to collective negative action towards the brand.
Cognitive dimension	Simultaneous provision of diagnostic or informative information (e.g. a picture of the situation, a link to news articles).	Content points towards a specific criticism concerning perceived misconduct by a brand.	Content does not always point towards a specific criticism with regard to the nature of the problem.	Content does not always point towards a specific criticism with regard to the nature of the problem.

The four categories of NEB identified ranged from less emotionally intense negative review writing and justice-seeking complaining to more intense retaliation acts and firestorming. Our proposed categories of NEB arose following the explicit grounding phase of our inductive and deductive analysis. A sample of tweets concerning the initial categories of the NEB typology were presented in conjunction with our pattern coding, which introduced the consumer NEB identified within our data sample. The next section presents the proposed categories of NEB emerging from the analyzed data.

NEB category 1: negative review writing

The first sample of NEB identified within our typologized categories was negative review writing. It was identified through three unique characteristics, namely (a) a less activated and aroused state (affective dimension), (b) which was focused on brand experience enhancement for both the brand and other engagement actors (orientation), and (c) which demonstrated the simultaneous provision of diagnostic or informative information (e.g. a picture of the situation, a link to news articles). As such, negative review writing was inherently low in emotional intensity.

Prior research suggests that consumers actively search for negative information in order to support informed decision making (Berezina et al., 2016). Sparks and Browning (2011) note that during this search process, the information garnered by consumers may contain both positive and negative valences. However, research also suggests that negative information is more potent and thus weighs more heavily in consumers’ assessments of brands and product suitability since consumers are more likely to be influenced by negative reviews (Park & Lee, 2009). The findings concerning negative review writing support prior research by Lee et al. (2008), who note that negative reviews have a powerful impact upon attitude formation, irrespective of the content, which may vary from lower quality reviews, lacking information richness, to more diagnostic and informative higher quality reviews.

Prior research has found NEB to manifest itself through entrenched negative emotional expressions (e.g. Vargo, 2016). We suggest negative emotions may range from weak to strong. Weak negative emotions have been found to exhibit themselves through feelings of disenfranchisement and dissatisfaction (see e.g. Einwiller & Steilen, 2015; Naumann et al., 2017). Negative affect was found to be present in negative review writing; however, this affect was weakly valenced. Mixed emotions were also found to be present within negative review writing, indicating that consumers may often feel a state of discomfort in reviewing their brand experience. The names of the telecommunications companies have been removed and replaced with (…) in the upcoming exemplary tweets due to the sensitive nature of the data:

‘Um. (…) The main entry point to your small business page is broken, and has been broken for quite a while now. Someone might want to pay attention here.’

‘Please advise when internet will be restored in our area. 20 hours without service. Second major outage in a week. Information on (…) outage website inadequate for those of us who need internet for work.’

In addition, recent studies emphasize the importance of understanding the qualitative characteristics of consumers’ reviews, including the helpfulness of reviews (Fang et al., 2016; Forman et al., 2008; Lee et al., 2011; Mudambi & Schuff, 2010). We found that negative review writing provided informative data (such as screenshots of the issues, hyperlinks to diagnostics) often directly linked to the brand itself, as well as to other engagement actors. Such information is considered beneficial in that consumers rely heavily on the content of reviews and use pertinent information to support their subsequent decision-making behaviors. Even though consumer reviews may be biased towards extremes of valences (i.e. positive and negative with fewer neutral reviews in between), consumers have been found to read reviews thoroughly as opposed to relying merely on summary statistics (e.g. Lee et al., 2017). The following tweets were accompanied by hyperlinks to web pages offering additional diagnostical or statistical information:

‘(…) Network performance is still pretty awful here in our area.’

‘According to their coverage map, this is what (…) consider ‘4G Outdoor’ speed (…) #(…) #Only10milesFromTheCity’

Negative review writing was also found to include the sharing of negative news stories about the brand, which is considered informative for other engagement actors within the social media ecosystem. For example, these tweets were accompanied by hyperlinks to news stories:

‘(…) mislead customers with national broadband network options. #(…) #nationalbroadcastnetwork’

‘Telephone scammers impersonating (…) are still active, but (…) appear to be unable to tackle the problem.’

In summary, negative review writing was considered to be a behavior that contains negative information about a brand, product, or service, whilst simultaneously being diagnostic or informative for either the brand or other consumers. This is in line with Lee et al.’s (2008) findings that negative online consumer reviews may vary from lower quality reviews to more diagnostic and informative higher quality reviews. Consumers engaging in negative review writing often clearly identify the cause of their dissatisfaction, making it easier for brands to address the issue if the review is written in time.

NEB category 2: justice-seeking complaining

The second sample of NEB identified within our typologized categories was justice-seeking complaining. This NEB was identified through three unique characteristics, namely (a) a less activated and aroused state (affective dimension), (b) which was focused on the achievement of personal goals and objectives, and (c) which expressed a specific criticism concerning perceived misconduct by the brand. As such, justice-seeking was inherently low in emotional intensity but of a much more specific and targeted nature than negative review writing.

Justice-seeking complaining was found to result from dissatisfaction with a specific brand or product experience. Satisfaction is traditionally conceptualized as a post-consumption emotional-cognitive process (Bartikowski & Llosa, 2004; Oliver, 1980; Wirtz et al., 2007). The disconfirmation of expectations paradigm proposes that customers form choice decisions based on their underlying attitudes, attribute beliefs, and perceptions regarding a brand. Cadotte et al. (1987) note that consumers may experience negative disconfirmation when a brand’s perceived performance is below expectations. As a result of expectations being violated, and from an affective perspective, justice-seeking consumers displayed afflicted emotions precipitating complaining behavior. We define affliction as a critical incident and a state of being burdened with something that causes suffering, such as loss (Do et al., 2019; Sloan & Oliver, 2013). We also argue that consumers experience a state of mental uncertainty and unease (Franzak et al., 2014).

‘Hey (…) I am having a very real problem with your service and billing, the harassment and constant calls. This has been an issue for nearly two months. I would appreciate somebody contact me before I take my complaints further. The problem has been proven to be at your end.’

‘Hey (…), I’m trying to track an order that is now 10 days late. What is going on? I’ve been on live chat twice, and given 2 different numbers to ring. I tried that and they have told me to go back to the live chat. #customerservicefail’

In addition, we found that justice-seeking complaining consumers focused on the specific objective of drawing attention to perceived misconduct by a brand for the purpose of achieving personal or collective goals (see, e.g. Einwiller & Steilen, 2015; Huppertz, 2003). Since justice-seeking complaining is often targeted at achieving personal goals, consumers’ emotional intensity is higher than when compared to negative review writing.

‘(…) 60 mins on hold and you think this is acceptable. What a joke! Even your hold music has packed it in.’

‘I called (…) explained my contract has expired yet nobody really cared about me as a longtime customer.’

However, we argue that unlike negative review writing, justice-seeking complaining is likely to have a more subdued impact on other engagement actors’ views of the focal brand. This is because even though the complaining behavior appears to have a justifiable rationale behind it, the quality and quantity of information conveyed to other engagement actors in justice-seeking NEB is often less specific and therefore less valuable and less visible. This is in line with the findings of Maheswaran and Meyers-Levy (1990) who suggest that negative information is more diagnostic and credible when it is specific and delivers information pertaining to the failure of the brand to deliver a high-quality experience with regard to specific features or attributes. This is because such information is necessary to support the decision-making process.

‘Is this how you reward loyal customers? Just last week, my (…) 4G had an outage. Now, it’s my (…) home broadband.’

‘(…) It’s still not working, I’ve spoken with multiple advisors and have been down to store multiple times. Not happy with this service at all.’

Moreover, justice-seeking consumers were found to often seek help with regard to their own problems. This supports prior findings by Kahneman and Tversky (1979), who note that consumers have a tendency to avoid losses.

‘Hey (…) why’s my internet keeps on getting interrupted?? What’s going on?’

‘(…) why can I not make calls overseas? It has been 5 days already and I’m paying my bills.’

In summary, justice-seeking complaining results from dissatisfaction and the disconfirmation of expectations and involves raising one’s voice on social media to draw attention to perceived misconduct by a brand to achieve personal or collective goals (see, e.g. Einwiller & Steilen, 2015; Huppertz, 2003). Because expectations are violated, consumers most likely have afflicted and confused emotions when they seek justice and complain. They often aim to achieve personal goals. The amount of information available to other consumers is often less valuable or less visible. Complaining consumers often identify the cause of their dissatisfaction, making it easier for brands to address the issue.

NEB category 3: retaliation acts

The third sample of NEB identified within our typologized categories was termed retaliation acts. This NEB was identified through three unique characteristics, namely (a) a highly activated and aggressive state (affective dimension), (b) which was focused on the achievement of reprisal against the brand, and (c) which does not always align with a specific criticism of the focal brand. As such, retaliation acts are inherently high in emotional intensity when compared to negative review writing and justice-seeking complaining behavior.

Prior research has identified anger as a key affective driver of negative commentary online (Coombs & Holladay, 2007, 2012). Anger is defined as a highly activated and aroused negative emotional state, which is often targeted at a brand that has transgressed (Bougie et al., 2003). Hostility occurs when consumers’ feelings of self-esteem and self-value have been violated (Smith, 2013), leading them to engage in destructive, retaliatory behavior. We found that consumers engaging in retaliatory behaviors express a clear intention to ‘get even’ with the focal brand. This may occur through threats to leave the brand or recommendations of another brand.

‘Let me guess … no, don’t tell me … (…) internet is down again. Why do I bother supporting a company that over charged me for several months? Moreover, staff wouldn’t fix nor call back when promised and now there are constant service interruptions. Time for a change.’

‘I am actually quite happy I changed to #(…) every now and then. Nice outage @(…) #(…) #4 g #3g’

We also found that consumers who engage in this behavior frequently reveal disturbed emotions, as evidenced through their use of vulgar language:

‘Data drop outs, s*** reception everywhere. Also phone call drop outs, while @(…) and @(…) folks are doing fine everywhere. It is a great pleasure to say I am f****** off @(…) forever. Worst network by far.’

‘@(…) you guys suck so f****** much. @(…) lines were down country wide today. I’d still rather change company’s and go with them.’

Retaliatory actions may subsequently create additional costs for the brand. Consumers engaged in retaliation have been found to engage in public threats to leave the brand and intensely negative recommendations regarding it. In line with the findings of Huefner and Hunt (2000), we found that retaliation acts often involve the non-identification of the cause of the brand failure, thus making it difficult for the brand to take corrective action to rectify it.

‘If you think about connecting to the internet or your mobile with @(…) please don’t. For the sake of your sanity go with anyone else but @(…)’

This lack of information concerning the brand failure event within consumers’ retaliatory acts, coupled with the tendency to engage in intensely negative, often derogatory expression, may reduce consumers’ perceived credibility. This is because other engagement actors may consider the consumers’ comments to be irrational, unpleasant, and uninformative (Kim & Gupta, 2012). This implies that negative reviews with intensely expressed negative emotions and limited utilitarian information can decrease their informative effect.

‘@(…) Do you have another outage? F*** you guys, @(…) here I come.’

It is also possible that retaliation acts have no specific reasoning or justification provided by the consumer, again reducing the quality of information conveyed in the retaliatory message:

‘well off to @(…), bye @(…)’

‘@(…) is really starting to suck. Thinking of moving on from then when my mobile plan is up too … ’

‘@(…) I believe the message is clear #stayawayfrom(…)’

In summary, retaliation acts differ from the aforementioned NEB as they are more emotionally intense. These behaviors are often accompanied by the intention to get even with the brand in different ways. At the same time, these actions have the potential to create additional problems and costs for the brand, including threats to leave the brand or recommendations of another brand. Consumers engaging in retaliatory acts often do not identify the cause of the initial issue, thus making it difficult for brands to take corrective action.

NEB category 4: firestorming

The fourth sample of NEB identified within our typologized categories was termed firestorming. This NEB was identified through three unique characteristics, namely (a) a highly activated state of anger and intense indignation (affective dimension), (b) which was focused on a generalized call to collective action against the brand, and (c) which does not always align with a specific criticism of the focal brand. As such, firestorming inherently has high emotional intensity, even when compared to other activated NEBs, such as retaliatory acts. Consumers engage in firestorming to express public disapproval and stir up potential action in others.

Prior research has found that collective online aggression directed towards focal engagement objects or targets, such as brands and products, is an increasingly common phenomenon. Social media offers a platform to consumers wishing to express their discontent and broadly promote it (Dolan et al., 2016; Kaplan & Haenlein, 2010). This is especially true for content communities, such as Twitter (Rost et al., 2016). Pfeffer et al. (2014) found that online firestorms involve waves of negative indignation and that social media dynamics offer a rich platform for opinion-forming between the members of the network. In fact, pre-existing networks are more likely to facilitate online firestorms than spontaneous forms of new networks, and Twitter stands out as the fastest social media platform, playing a critical role in helping firestorms to spread (Pfeffer et al., 2014).

We found that firestorming result from higher levels of collective anger and reflect agitated and disturbed emotions. This NEB was found to arise in response to a transgression or violation of consumers’ perceived rights. It is directed at eliciting collective action through emotional contagion and often involves the consumer displaying disturbed emotions which refer to uneasiness that troubles the mind of a person and facilitate the potential display of emotional illness in social media (Bishop, 2014; Hardaker, 2010). This phenomenon of aggressive, offensive, and inflammatory commentary within social media has been labeled flaming and toxic online disinhibition (Rost et al., 2016).

‘@(…) Stop d**** the guys and fix ya s***!’

‘@(…) I don’t know how, but you have managed to do it again. Get your s*** together you w******.’

We also found that firestorming involves a collective call to action to partake in derogatory, crowd-based outrage. One of the defining features identified in our data was the tendency of online firestorms to involve intensely negative, emotive, impulsive, and sensation-seeking commentary targeted at a focal brand. This NEB is value destructing in the sense that it can lead to widespread and broadly perpetuated, snowballing commentaries from other networked engagement actors within a very short timeframe.

‘@(…) what the h*** is wrong with your internet in this area? It has been s*** for the last 5 days.’

‘I’m sick of being nice to you @(…) listen to your f****** customers or otherwise you are going to lose them!’

The lack of social norms present in digitally mediated environments enhances the propensity for firestorms to occur. That is, visual and verbal cues in Twitter are absent, removing the presence of interpersonal norms. In addition, the perceived focal engagement target (e.g. the brand) is subsequently dehumanized. This was found to reduce the empathy of a consumer engaging in firestorming. We therefore found that it frequently manifests itself through intense, obscene, and vulgar language:

‘@(…) why the f*** is your service so s***?’

‘@(…) I heard that someone suggested that (…) should change its name to S******.’

‘@(…) your support is f****** trash.’

Even though a consumer might share brand or product information with their audience, the extensive emotional intensity of the message was often found to outweigh the benefits derived from the shared information.

‘@(…) you idiots. I am overseas and you revert to (…) … their service is a joke and 4G is not working AGAIN. WHAT THE F*** @(…)’

‘@(…) Seriously, my national broadcast network has been f***** all the time now. Also my pay TV is f*****. You are seriously one incompetent corrupt company. How do you your executives sleep at night? I’ll tell you, on a big pile of money you have effectively stolen off your local consumers!’

In summary, firestorming was found to harness the collective power of social media and elevate individual consumer voices, providing a platform for public discontent. This enabled consumers to communicate their perceived brand transgressions to large, geographically dispersed audiences, giving firestorms a unique characteristic. Unlike physically constrained settings, aggressive and collective outrage was found to occur with greater intensity and frequency in digitally mediated environments. Firestorming contained high numbers of negative messages and/or intense indignation expressed against a person, brand, or group. However, the behavior did not always point to an actual specific criticism. Consumers engaging in firestorming often identified the cause of their dissatisfaction, making it easier for brands to address the issue, but their strong emotional state could hinder brand collaboration and scare off other consumers.

Further, the differences between cultural settings are not very substantial in the results. Whereas Australian consumers utilized more NEB in total, Finnish consumers most often resorted to justice-seeking complaining and negative review writing, which was also in line with the behavior of Australian consumers. This could be due to cultural differences beyond the context of this study but could also signal the maturity of the market and the quality of the competition available.

Implications and conclusions

The framework developed in this paper provides a number of theoretical insights into the nature of NEB, as well as how it occurs within social media. The proposed framework is a novel contribution and continues the previous work of scholars such as Dolan et al. (2015), Vargo (2016), Azer and Alexander (2020), and more recently Do et al. (2021) by examining consumers’ visible manifestation and emotional intensity of NEB. This is significant because the framework proposed in this study offers insights into its inherent characteristics, orientation, and foci, thus advancing the literature and implications for scholars. Our study confirms that whilst social media is an avenue for consumers looking to express positive forms of engagement (Hollebeek & Macky, 2019), it is also replete with negative forms. In addition, the extent to which NEB manifests itself and the intensity with which it is expressed can significantly impact brand performance (Azer & Alexander, 2018).

This study suggests that service brands should utilize automatic sentiment analysis systems (e.g. Mediatoolkit; Qualtrics XM) to monitor reviews within social media and extract and catalogue customer complaints to be sensitive to the specific category of NEB manifested by the consumers. Various forms of NEB may pose different levels of risk to brand reputation and, therefore, have variable impacts in terms of financial loss, switching behavior, and value co-destruction (Hollebeek et al., 2016; Naumann et al., 2017; Plé & Chumpitaz Cáceres, 2010). This data can then be maintained in a real-time customer complaint database to assess service performance and consumer sentiment. Service firms should proactively and regularly engage in satisfaction surveying of their customer database in order to identify different types of NEB. Customers could then be classified and mapped according to their level of dissatisfaction allowing for prioritization of NEB. Segmentation of the customer base according to type and level of NEB will enable management to provide a more tailored and commensurate recovery process to prevent certain categories of customers from switching and engaging in further negative engagement.

Second managerial implication of this study proposes that less intense NEB, such as negative review writing may be beneficial for the service provider because of the valuable information revealed by consumers in the review process. In fact, consumers engaging in negative review writing may be more effectively handled through recovery and restoration processes than other consumers expressing more vehement and intense forms of NEB. They may also be more willing to repurchase from the brand in the future (Bijmolt et al., 2014). Thus, when addressing negative review writing, it is critical for the service provider to take note of the details of the review in order to enhance and improve the brand experience. Paying attention to detail in a consumer’s review can also assist the brand in responding appropriately and steering the tone of the discussion in a positive direction.

Third managerial implication of this study proposes that dealing with somewhat more intense NEB such as justice-seeking complaining requires understanding of the reasons and motivations behind the consumers’ complaints. These consumers display moderate emotional intensity in their NEB and are more likely to complain if economic benefits are apparent, for example, if the perceived costs, perceived benefits, and the probability of the complaint being successful weigh in their favor (Oliver, 2015). Service providers are advised to provide justice-seeking consumers with a recovery response process, which focuses on listening to the consumer and taking actions that aim to solve the issue at hand. A carefully tailored response and the proper handling of a complaint can create positive emotions for an upset consumer (Argyris et al., 2021).

Fourth managerial implication of this study encourages service providers to act in an unprejudiced manner. With regard to consumers who display most intense forms of NEB, such as retaliation acts and firestorming, it is advisable for service providers to be cautious but not always afraid to respond to such commentary. Consumers are more likely to experience heightened emotional intensity and rage when brand failures remain unresolved. As a result, residual negative emotions are carried forward, and a firm’s resources are threatened (Surachartkumtonkun et al., 2015). These consumers are best handled by an approach that focuses on mitigating the amount of anger. More often than not, emotional intensity levels begin to diminish if proper actions are taken. Whilst these consumers are likely to resist and challenge restorative actions, treating them as fairly as possible is advisable given that, for most brands, long-term future profits and revenues are an important consideration. Choosing not to attempt restoration could avoid short-term losses but risks long-term brand equity if consumers perceive themselves to be engaged in an unfair exchange relationship. After all, affective commitment is significant and relevant mediator in consumer–brand relationships and have an impact on engagement (Rather et al., 2018; Van Tonder & Petzer, 2018). Increasing the amount of trust is also a significant contributor towards consumer loyalty (Rather, 2019). Overall, the findings of this study suggest that the understanding the intensity of NEB is crucial.

This study provides a novel attempt to develop a typology of categories of NEB targeted at brands based on (a) the visible manifestation of the NEB and (b) the emotional intensity of the NEB. As such, there are several limitations to this study. First, the data collection period was restricted. The use of alternate timeframes and multiple data collection periods would be beneficial in identifying the context-specific nature of NEB. Second, the findings are limited to the telecommunications sector. Future research should examine other sectors in order to explore the generalizability of the model. Third, the cultural settings of the study may have ramifications for the findings beyond our understanding, and future studies are needed to test the different NEB categories in other cultures. Given the speed with which the literature is evolving, future research should empirically test competing frameworks and typologies of NEB, such as the one proposed in this study, with the aim of stabilizing and validating the negative engagement construct.

Disclosure statement

No potential conflict of interest was reported by the author(s).