DOI: 10.1002/mar.

21273

RESEARCH ARTICLE

Negative brand meaning co‐creation in social media brand

communities: A laddering approach using NVivo
George Rossolatos MSc, MBA, PhD

Department of Humanities & Cultural Studies,

University of Kassel, Germany Abstract
Social media brand communities (SMBCs) have been heralded for their co‐creative,
Correspondence
George Rossolatos, Department of Humanities participatory potential whereby consumers actively contribute to the proliferation of
& Cultural Studies, University of Kassel, Kassel meaningful brand avenues in a virtuously circular relationship with brands. Elevated
34109, Germany.
Email: georgerossolatos123@gmail.com loyalty and enhanced brand equity have been posited repeatedly as likely outcomes
of a positively engaged community of brand aficionados. However, evidence to the
contrary as negative brand co‐creation or brand co‐destruction has been progres-
sively piling up in the extant literature. This paper contributes to the meaning co‐
creation in SMBCs literature primarily on two grounds: first, by offering a
methodological framework for adapting the laddering research technique in a mixed
methods vein to SMBCs data in a thread‐specific context, by leveraging the analytical
capabilities of NVivo CAQDAS software; second, by addressing bottlenecks in the
applicability of the proposed methodology in light of negative brand co‐creation.

KEYWORDS
brand communities, brand meaning co‐creation, laddering technique, NVivo, social media

1 | INTRODUCTION: PROGRESSIVE EMPHASIS potentially large scale, content and storytelling” (Sorensen &
ON CONTENT IN CONSUMER/BRAND Drennan, 2017). Spearheaded the community perspective in
INTERACTIONS IN SOCIAL MEDIA BRAND branding research (Heding, Knudtzen, & Bjerre, 2009), according
COMMUNITIES to whom a brand community “is distinguished from other
consumer collectives by three markers: consciousness of kind,
Social media brand communities (SMBCs) have been proliferating shared rituals and traditions, and a sense of moral responsibility”
over the years at an exponential rate. Nowadays, there is not even a (Sorensen & Drennan, 2017). SMBCs differ from their offline
single brand, at least among the 100 leading brands in the FMCG counterpart primarily in terms of “transparent social context (e.g.,
sector which does not maintain an active presence across the major no aliases as in web forums), flat structure, potentially large scale,
social media platforms, and that is not in possession of the requisite content and storytelling” (Sorensen & Drennan, 2017).
social listening skills for monitoring discussions and transforming Consumers have been joining en masse brand communities, as
them into actionable insights. the latter have been increasing their influence in shaping brand
Muniz and O’ Guinn (2001) spearheaded the community meaning (co‐creation and consumer empowerment; Wiegandt, 2009).
perspective in branding research (Heding, Knudtzen, & Bjerre, In online brand communities, consumers “communicate through
2009), according to whom a brand community “is distinguished postings, instant messaging, and chat discussions about special
from other consumer collectives by three markers: consciousness interests related to the company’s products and brands” (Keller,
of kind, shared rituals and traditions, and a sense of moral 2009). In this context, “co‐creation reinforces ties between commu-
responsibility” (Sorensen & Drennan, 2017). SMBCs differ from nity members. Through games and competitions, they build up
their offline counterpart primarily in terms of “transparent social intimate relationships because they share the same passion for the
context (e.g., no aliases as in web forums), flat structure, brand and common memories” (Cherif & Miled, 2013). The
---------------------------------------------------------------------------------------------------------------------------
This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium,
provided the original work is properly cited.
© 2019 The Authors. Psychology & Marketing published by Wiley Periodicals LLC
Psychol Mark. 2019;36:1249–1266. wileyonlinelibrary.com/journal/mar | 1249
1250 | ROSSOLATOS

co‐creative potential of brand communities spans a broad range of content in SMBCs is in fact negatively valenced (Hollebeek & Chen,
marketing activities, most eminently with regard to new product 2014), and moreover, potentially hazardous for brand reputation and
development, new promotions, consumer insights, but also cultural equity, has fueled this burgeoning substream. As Prahalad and
and lifestyle aspects, either directly or indirectly related to the brand Ramaswamy (2004) contend, not all co‐creative activities result in
host. Rialti, Caliandro, Zollo, and Ciappei (2018) identified four positive value. This form of non‐collaborative behavior has been
principal types of co‐created experiences, namely a brand’s individual identified as a significant feature of interactions between brand
usage experiences, that is experiences stemming from individual owners and stakeholders. Although the community branding mantra
consumption activities; autocelebrative experiences which derive that is suggestive of a streamlined centering around a focal brand is
from individual consumption activities, but are influenced by regularly evoked (Dessart, Veloutsou, & Morgan‐Thomas, 2015), often
community settings; communal usage experiences with regard to the specific and quite dissonant, contested (Lee & Soon, 2017; Weijo,
the use of products for participating in communal activities; Hietanen, & Mattila, 2016; Yeritsian, 2018) and outright messy ways
collective celebration experiences where consumers use a brand in whereby such focalizations are attained remain under‐explored.
the company of other community members. Among the various A closer reading of consumers’ comments to brand initiated
benefits that marketers may reap from consumer participation in threads in SMBCs more often than not points to thematic territories
brand communities are new ways of unlocking competitive advan- and semantic ramifications that not only are dissonant with the initial
tage (Prahalad & Ramaswamy, 2004), enhanced engagement, equity discussion topic (Brogi et al., 2013), but display such heterogeneity as
(Cova, Pace, & Park, 2007; Habibi, Laroche, & Richard, 2014; to posit severe constraints to the sustainability of thematic control
Rossolatos, 2018a,b) and loyalty (Laroche, Habibi, Richard, & on behalf of the marketer. Subsequently, co‐creation becomes a
Sankaranarayanan, 2012; Meister, 2012). process of co‐destruction of brand meaning, rather than mutual
Various antecedents in terms of attitudes and motivations for brand development (Kristal, Baumgarth, & Henseler, 2018). The bulk
participating in online brand communities have been identified in the of research that has been produced to this end offers largely
constantly augmenting literature. Brand communities allow their qualitative typological frameworks that outline the types of negative
members to assume multiple social identities (Zaglia, 2012). In their co‐creation actions undertaken by consumers in SMBCs and brands’
review of the relevant literature spanning the marketing research reactions, by discourse or content analyzing comments, as will be
terrain, as well as the social sciences on a broader scale, Kamboj and shown in the literature review in the following section. These studies
Rahman (2017) singled out two main social theoretic perspectives on adopt a micro‐analytical approach to SMBCs content by attending to
which researchers have been drawing for conceptualizing antece- specific aspects of situated interaction.
dents: the needs/benefits and the social integration theories. This paper complements the extant literature on negative brand
Additionally, Martinez‐Lopez, Anaya‐Sanchez, Aguilar‐Illescas, and meaning co‐creation by focusing on SMBCs interaction in an as yet
Molinillo (2016) highlight the prevalence of social identification and unaddressed manner, that is by analyzing data, on the one hand, in a
social comparison theories. The former is geared towards explaining highly context‐specific fashion that is incumbent on the thematic
the emotive participation of an “I” in a “we,” whereas the latter orientation of individual brand initiated threads. On the other hand,
towards explaining how people compare themselves to others while it dimensionalizes negative co‐creation even further, by introducing a
adapting to multiple environments. sharper distinction between relevantly negative comments, irrele-
Furthermore, research in brand communities has focused on vantly negative comments, and outright indifferent ones. To add
types of interaction. In this context, two major types have been interpretive depth to brand meaning co‐creation in SMBCs, I apply
identified, namely between a brand and consumers (community the laddering technique in a critical vein with a view to encapsulating
members), and among brand community members themselves (Muniz adverse aspects of co‐creation “from below” (Yeritsian, 2018). The
& O’ Guinn, 2001, cited in Martinez‐Lopez et al., 2016, p. 109). laddering technique is particularly attuned to the exploration of
Likewise, two major types of brand communities are currently instances of negative brand evaluation (Woodside, 2010), and
operative in social media, namely brand initiated, moderated and concomitantly of negative co‐creation, while it seeks to unearth
controlled communities and consumer‐to‐consumer ones. Salient how, why and in what situations a brand is rejected either on a
typologies with regard to the various roles performed by brand particular dimension with regard to attributes, benefits, values or
community participants and their level of involvement have been across the board.
furnished, such as Kozinets, Valck, Wojnicki, and Wilner's (2010) As will be discussed in greater detail in the ensuing section, a
macro‐typology consisting of devotees, insiders, newbies, and major difference in the formation of brand‐related memory as brand
minglers, and Pongsakornrungsilp and Schroeder’s (2011) typology associations between the offline and online enclaves consists in the
of creative posters, brand warriors and moderators (Caliandro & conditioning of individual associations by a tapestry of synchronically
Gandini, 2017). accessible comments that make up a brand memory (Fujita, Harrigan,
Although the bulk of research into SMBCs has focused on yielding & Soutar, 2017) as living social text. This calls for a more detailed and
insights as to the antecedents, outcomes, and types of interaction and nuanced exploration of how exactly this collective brand memory is
engagement, added focus has been laid progressively on the posted generated and on what levels. The exploratory orientation of the
content. The recognition that a significant portion of user‐generated laddering technique is well aligned with the exigencies of this
ROSSOLATOS | 1251

research task. Moreover, laddering may be used as a method for has been largely concerned with identifying the psychological and
gauging the depth of consumer engagement, insofar as it offers a behavioral antecedents and outcomes of brand‐to‐consumer and
comprehensive outlook from latent attitudinal states to manifest consumer‐to‐consumer interactions in brand communities, rather
evaluative judgments concerning product attributes and benefits. than focusing on the interactional modes and content.
In a nutshell, this paper pursues the following expository path.
First, it offers a literature review with regard to the conceptual and
2.2 | Negative brand meaning co‐creation in focus
methodological differences between sentiment analysis approaches
in a big data vein and marketing research microanalytical approaches. The added focus on SMBCs content in digital marketing research has
The focus, then, shifts towards scrutinizing how negative co‐creation been driven by the increasing recognition of the potentially adverse
has been framed in the marketing literature to argue for the impact of negative brand co‐creation in terms of brand image erosion
relevance of applying the laddering technique to SMBCs data and and equity dilution (Bambauer‐Sachse & Mangold, 2011; Kristal et al.,
how it may contribute in furthering our understanding of distinctive 2018). Negative brand co‐creation has been framed in various
types of comments under a uniform typological rubric. Subsequently, manners in the extant literature. Indicatively, Rauschnabel, Kammer-
the research questions of the exploratory study are formulated, lander, and Ivens (2016, pp. 381‐382) coined the term collaborative
followed by an outline of the theoretical underpinnings and empirical brand attacks (CBAs), as “joint, event‐induced, dynamic, and public
applications of the laddering technique. The exposition of the offenses from a large number of internet users via social media
methodological framework that is adopted for showcasing the platforms on a brand that are aimed to harm it and/or to force it to
laddering technique follows suit. The findings from the analysis of change its behavior.” By adopting a multicase study analytic
the selected threads with the aid of NVivo 12 Plus are discussed with approach, the authors coined a typology of brand reactions to CBAs,
a focalization on contested meanings in light of negative and consisting of ignoring, censoring/legal steps, counter stating, appeas-
indifferent comments. Finally, methodological reflections on the ing, content bumping, and changing behavior.
applicability of the laddering technique on social media data are Hollebeek and Chen (2014) expanded the scope of brand
offered. The paper concludes with a discussion on the managerial engagement as a cognitive, emotional and behavioral investment in
implications of the proposed framework, the limitations of this study, specific brand interactions, to accommodate both positively and
as well as areas for further research. negatively valenced interactions. They culminated in a three‐pronged
definition, featuring immersion, passion, and activation. Immersion
designates the level of a consumer’s positively/negatively valenced
2 | CO ‐C R E A T I O N O F B R A N D ME A N I N G
brand‐related thoughts, concentration, and reflection in specific
I N SM B C S : O P P O R T U N I T Y O R NU IS A N C E ?
brand interactions. Passion refers to the degree of a consumer’s
positively/negatively valenced brand‐related effect, exhibited in
2.1 | Limitations of big data analytics in generating
particular brand interactions. Activation stands for a consumer’s
deep insights about brand meaning co‐creation and
positively/negatively valenced level of energy, effort, and time spent
consumer engagement
on a brand in particular brand interactions. Their analysis drew on
Disillusioned with big data and “bag‐of‐words” approaches that are consumer, rather than brand controlled SMBCs, namely on fans of
plagued by an insufficient attentiveness to the contextual use of Apple and fans of Samsung mobile (reflecting positively valenced
linguistic and multimodal components, as well as to the interactional engagement), and on Apple sucks and Samsung sucks (reflecting
context among participants, researchers have been calling for closer negatively valenced engagement).
readings of social media data in a small data vein (Latzko‐Toth, Rosenthal, Pereira, and Brito (2017) put forward a brand meaning
Bonneau, & Millette, 2017). As argued by Puschmann and Powell co‐creation framework between brands and fans on Facebook by
(2018), misclassification in big data analyses can occur for a number applying a case study methodology, with a view to identifying how
of reasons, most strikingly due to domain or context specificity. The brand communities may be curated by its managers. Their framework
authors conclude that the appropriate usage of sentiment analysis was based on the role content performs in fans’ lives and their
presumes detailed knowledge of the domain of application. Small potential reactions, over and above the projected brand meaning.
data readings that seek to unearth the multilayered construal of The offered typology drew on Kapferer’s brand identity prism model
brand meaning in co‐creative predicaments may be afforded by (see Cherif & Miled, 2013 for a similar approach that draws on the
applying semiotic and/or discourse analytic frameworks to social brand identity prism), while a grounded theoretical approach was
media data. More specifically, although consumer engagement with implemented during the codification process. The authors culminated
branded content, as well as the ubiquitous user‐generated content, in an allegedly nonrepresentative typology of brand actions and
have been heralded as indispensable facets of brand meaning co‐ brand fans’ reactions. The latter comprised doubting, liking, sharing,
creation, big data analytics have been merely scratching the surface suggesting, praising, criticizing, self‐expression. However, no further
of engagement by focusing on segmenting user‐generated comments information was supplied as to the relative incidence of negative
into generic types, such as positive, negative, or neutral comments versus positive co‐creation in the scrutinized dataset, thus emitting
for gauging brand sentiment. At the same time, consumer research the impression that all brand/fans interactions were geared towards
1252 | ROSSOLATOS

positive co‐creation, also relevantly so, that is in alignment with the whereby private memories make up a constantly augmenting
overarching brand initiated threads. The scrutinized brand commu- collective memory, while collectively assembled memories feed back
nities’ posts were treated en masse, that is regardless of their specific into private memories. Thus, brand memory formation is the outcome
meaning in the context of brand initiated thematic threads. The of a constant discursive struggle in a linguistic marketplace (Bouvier,
framework revolved around high involvement categories/brands 2015), yet whose vestiges are not necessarily apparent in a directly
(whiskey, running apparel), thus limiting its applicability to low dialogical form. Instead, it is indirectly traceable through a secondary
involvement categories where, as the authors contend, engaging elaboration and dimensionalization of individual comments. The
consumers may be a more demanding exercise. This study takes on nondialogical nature of comments may be gleaned by attending to
board the authors’ suggestion for future research that is geared the types that regularly succeed one another in the same thread.
towards indicating intrinsic limits to fan engagement and the ability Such types vary to such an extent and may include (Hagedoorn,
to generate meaningful interactions in social networking sites. 2015, p.588) personal memories from one’s own experience
Kristal et al. (2018) put forward the concept of observer‐based (including mediated experiences), personal and mediated memories
brand engagement to account for the impact of a few participants in through engagement with lieux de memoire or responses of praise and
social media blogs who engage in negative brand meaning co‐creation criticism.
activities, on the larger proportion of passive participants who are The intermingling of personal with publicly available memories
exposed to the process of co‐destruction. In this context, they in the context of SMBCs brand memory formation (Marlowe,
distinguished between brand play and brand attack. Brand play Bartley, & Collins, 2017; Van Dijck, 2007) is further compounded by
accommodates negative comments as parody and pranks, whereas dint of the medium’s interactional orientation. In such settings,
brand attacks encapsulate instances where a brand is hijacked for a communicative memory (Bartoletti, 2011) is also operative, namely
destructively negative reimagining of its meanings. the individual appropriation of readily available public information
Despite the above studies’ adoption of a microanalytical while interacting with other members’ comments. In the context of
approach to the classification of negative content, they neither Twitter, for example, this hybrid public/private memory has been
contextualize the analysis within the strict semantic contours of identified as a meta network of communicators (Cantey & Robinson,
individual threads nor dimensionalize them in terms of whether 2015).
negatively (and positively) valenced comments are relevant or The participatory, yet fragmentary and not necessarily stream-
irrelevant to the overarching brand initiated themes. Yet, according lined with a brand’s thematic intentions co‐creative brand memory in
to the laddering model, as will be shown in greater detail in the SMBCs settings calls for exploring how brand associations are
following section, both direct and indirect links between cognitive shaped and what sort of brand meaning is generated in this
and affective elements are important in a means‐end ladder predicament. From a consumer‐based brand equity point of view
(Reynolds & Gutman, 1988; Huber, Beckmann, & Herrmann, 2004). (Keller, 1998; Rossolatos, 2014a), a brand memory consists of
This entails that although the underlying theme of a comment (either primary and secondary brand associations. The former are product‐
with regard to attributes or benefits) may be irrelevant with bound and concern identifiable product attributes (e.g., smell, color),
reference to the overarching theme in a brand initiated thread, its whereas the latter are non‐product bound, such as consumer benefits
impact in terms of positively engaging a brand community may be (e.g., status), user and usage imagery. Brand memory shapes up
paramount. Moreover, negatively valenced comments are not through the interaction of brand associations with latent consumer
addressed as to whether they are congruent or incongruent, attitudes in a brand knowledge structure. Keller’s brand knowledge
axiologically speaking, with the projected brand image in individual structure, by virtue of positing brand‐related associations at the very
brand initiated threads. These salient facets will be focalized here heart of a brand’s structure, affords to yield a comprehensive
with the application of the laddering technique. account of the interrelations amongst attributes, benefits, and
attitudes.
To explore SMBCs comments in a comprehensive manner, it is
2.3 | The idiosyncratic nature of brand memory
suggested to systematically interpret the selected corpus’ comments
formation in social media
by drawing on the laddering technique. The laddering technique
The co‐creative potential embedded in brand communities runs much brings under a common rubric attributes, benefits, but also
deeper than specific and isolated marketing activities and concerns a consumers’ personal values (axiological framework) that determine
wholly new way whereby brand memory is formed as a collective to what extent the former are valorized. The “values” component of
consumer/brand memory. This diachronical molding of collective the laddering technique corresponds to attitudes in Keller’s brand
memory has been captured in digital media theorizing by the so‐ knowledge structure (or self‐knowledge structure, according to
called memory turn, whereby official narratives are challenged with Bagozzi & Pratibha, 1994). Although the co‐creation of values (e.g.,
increasing frequency from countless uploads and the circulation of intrinsic, altruistic, impression‐seeking, peace‐loving, family‐oriented)
incidences, reflections, and illustrations (Liew, Pang, & Chan, 2015). between consumers and brands and among consumers in SMBCs has
Brand memory or archival memory formation in a Web 2.0 been addressed in the literature (Sorensen & Drennan, 2017), it has
environment is evinced as participatory memory (Hagedoorn, 2015), not been explored comprehensively in tandem with attributes and
ROSSOLATOS | 1253

benefits or with how personal values determine the valencing of model, consumers are always projecting abstract values onto
attributes and benefits. concrete consumption instances, albeit in an indirect fashion. All
consumption choices have consequences. Means‐ends theory as-
cribes valences to these consequences, thus accommodating product
2.4 | Research questions
choice under a comprehensive axiological rubric.
In light of the preceding discussion, and against the identified gaps in The laddering technique generates pathways between A‐C‐Vs. As
the negative co‐creation literature in SMBCs, the main research an example from the salty snacks category, Reynolds and Gutman
objective consists in identifying whether and how the laddering (1988, p.13) cite the following ladder: (A) flavored chip →(A) strong
technique may be applied to consumers’ comments with a view to taste → (C) eat less → (C) don’t get fat → (C) better figure → (V)
obtaining a comprehensive picture of consumer engagement along- self‐esteem. The typical procedure for effectively applying the
side attributes, benefits, values. This application calls for a highly laddering technique to data gathered through in‐depth interviews
contextualized, microanalytical approach that takes into account (a) comprises the following steps: Data are collected on each component
the bespoke thematic contours of individual brand initiated posts (b) of the ACV triad through constant probing that attempts to generate
the valencing of consumers’ posts with regard to such bespoke as much in‐depth understanding of the links between the compo-
settings. Subsequently, to identify how laddering may be applied to nents as possible. The research may aim at producing either single
SMBCs data, given the ubiquity of negative co‐creation, we must first brand ladders or multibrand comparative ones (Saaka, Sidon, & Blake,
clarify what negative and positive valencing mean in specific thematic 2004). ACVs, then, are classified according to a typical content
contexts. To attain the stated research objectives, the following analytic coding or discourse analytic procedure. The resulting
research questions (RQs) will be addressed: summary table features the ACVs (in columns) and the informants
(RQ1): How can we operationalize the laddering strata in the (in rows), as well as the pathways between the ACVs. The summary
context of SMBCs comments in such a way as to effectively account table is then transformed into a hierarchical value map (HVM) in the
for both positive, as well as negative brand meaning co‐creation form of an orthogonal 1 × 1 matrix (also called an implication matrix;
aspects? Amatulli, Pino, De Angelis, & Cascio, 2018).
(RQ2): How can the identified bottlenecks in the operationaliza- The strength of the pathways is assessed by calculating the
tion of the laddering technique in the context of SMBCs be frequency whereby each pair in a ladder is mentioned by a project’s
circumvented with a view to yielding managerially salient insights? informants. “A HVM is gradually built up by connecting all the chains
The methodological framework that will be used for addressing that are formed by considering the linkages in the large matrix of
these RQs will be displayed after the ensuing section which provides relations among elements. The most typical approach is to try to map
an overview of the laddering technique in light of the extant all relations above several different cutoff levels (usually from three
literature, aimed at identifying its salient facets and their relevance to five relations, given a sample of 50–60 individuals). The use of
for consumer research in SMBCs. multiple cutoffs permits the researcher to evaluate several solutions,
choosing the one that appears to be the most informative and most
stable set of relations” (Reynolds & Gutman, 1988, p. 13). The
3 | THE LADDERING TE CHNIQUE AS components of a ladder were further qualified by Abeele and Zaman
B R AND ME AN IN G GE NER AT I O N (2010) into concrete and abstract attributes, functional and
PROC EDURE psychosocial consequences.
The laddering technique has many applications, and is particularly
The laddering technique is a qualitative method for synthesizing data useful for consumer segmentation (Orsingher, Marzocchi, & Valentini,
gathered mostly through in‐depth interviews in a comprehensive 2011), brand positioning (Gutman, 1982), adaptive selling (Wansink,
manner by incorporating information on product attributes, con- 2003), and advertising development purposes (Reynolds & Gutman,
sumer benefits, and consumer values. A key benefit that stems from 1988; Saaka et al., 2004). In the following pages I showcase how the
the application of the method consists in its ability to bypass laddering technique may be fruitfully applied in SMBC settings.
avoidance behavior on behalf of informants, especially in the face of
sensitive or taboo research topics.
4 | RESEARCH DESIGN
As outlined by Reynolds and Gutman (1988), the conceptual
backbone of the laddering technique consists of A‐C‐Vs, that is
4.1 | Mixed‐methods methodology
associative networks or hierarchically linked conceptual ladders
comprising product attributes (A), consequences of brand usage as This study adopts an integrative mixed methods research design
consumer benefits (C), and personal consumer values (V) as (Schoonenbooma, Burke, Johnson, & Froehlich, 2017), seeking to
components of a higher‐order axiology. According to this model, Vs conjoin the insights that are generated across the different parts of
are always mediated by Cs, while Cs stem from a determinate set of the research. “Integration can be said to occur to the extent that
A’s. The laddering technique was developed by drawing on the different data elements and various strategies for analysis of those
means‐ends theory. According to Gutman’s (1982) means‐ends chain elements are combined throughout a study in such a way as to
1254 | ROSSOLATOS

become interdependent in reaching a common theoretical or Kellogg: For over 70 years, South African families have started their
research goal, thereby producing findings that are greater than the day the Kellogg's® Corn Flakes way! #NourishYourGreat
sum of the parts” (Bazeley & Jackson, 2013, p.196). In this context, Domesto’s: Domestos and Unilever are proud to support the
the quantitative analytics are supportive of the ensuing interpretive #GlobalGoals for a #brightFuture where everyone, everywhere
endeavor (Herring, 2004). A case study methodology with cross‐case has access to a clean, safe toilet.
comparative orientation is applied to the thematic discovery
procedure and the analysis of findings. “One aim of studying multiple Each thread counts as an individual sampling point, while each
cases is to increase generalizability. At a deeper level, the aim is to comment as a single analytical unit. The data were scraped from
see processes and outcomes across many cases, to understand how Facebook by using NVivo’s NCapture, saved in pdf format and
they are qualified by local conditions, and thus to develop uploaded on NVivo. No ethical issues emerged, as these are publicly
more sophisticated descriptions and more powerful explanations” available and accessible data.
(Bazeley & Jackson, 2013). Case study methodology is also
particularly well aligned with theory building, in general, as well as
4.4 | Data analysis
with the objectives and components of the laddering technique in
particular (Woodside, 2010). Each individual post (comments and replies) within the selected
threads counts as a unit of analysis. The discovery procedure consisted
of reading each comment simultaneously on three layers, namely
4.2 | Sampling frame
description, contextualization, and signification, the so‐called process
The sampling frame criteria for data collection consist in that the of data “thickening” (pace Geertz). This process is in line with the
brand community (a) must have been initiated by the marketer (rather interpretivist paradigm of qualitative inquiry which recognizes that a
than a consumption group), as the topics in marketer initiated threads social phenomenon can only be understood in context through fine‐
versus consumer‐initiated threads, on the one hand, span all marketing grained accounts (Latzko‐Toth, Bonneau, & Millette, 2017). The coding
functions, from new products to promotions; while, on the other hand, procedure took place against an a priori coding scheme (in terms of
they correspond more strictly to the concept of co‐creation given the brand association type: attribute, benefit, personal value, and valence:
participation of both marketers and consumers, as against consumer‐ negative, positive; neutral valence was excluded as maximum polarity
initiated brand communities where the participation of the marketer is was sought; Table 1). All comments were coded multiply based on the
not guaranteed (and rather tends to be considered as intrusive). The a priori coding scheme (Angus, 2017) in a qualitative content analytic
offered analytic focuses strictly on brand initiated and moderated vein, against the background of the three laddering strata, dimensio-
communities in Facebook (Vitak, 2017), for the sake of consistency in nalized according to (a) the comments’ relevance to the brand initiated
terms of technological platform, mode of interaction, and type of user‐ theme (b) their valencing (c) in the case of values, whether they are
generated comments (b) must have a long‐standing presence in congruent or incongruent with the specific brand value that is
Facebook of at least 5 years to ensure a sufficiently large user base, conveyed in the brand initiated theme of the focal threads. The
allowing for richness of brand associations against the background of CAQDAS software NVivo 12 Plus was used for coding, analyzing, and
multifarious topics (c) each brand initiated thread must be sufficiently synthesizing the data in an iterative fashion (Salmons, 2017).
sizeable (i.e., must have garnered at least N = 30 comments), to be A bespoke code “indifferent” was used for coding comments
amenable to quantitative analysis. Let it be recalled that, as highlighted which were neither relevant to the brand initiated theme in terms of
by Miles and Huberman (1994), the relationship between analytical/ attributes/benefits, nor irrelevant to the brand initiated theme (yet
interpretive depth and the sample size is inversely proportionate. still relevant to the brand). These exploratory areas were found to be
missing from the coding approaches in the extant literature, as shown
earlier, although being highly relevant from a depth of engagement
4.3 | Data collection
point of view. As above stressed, despite the speculative remarks in
The preparatory stage consisted of screening the 100 biggest FMCG the marketing literature about the equity, loyalty, and involvement
brands' (in terms of global sales turnover) Facebook communities enhancing potential of SMBCs co‐creative activities, the bleak truth
with regard to the sampling frame criteria (Kozinets, 2002; Mayr & is that user‐generated comments not only often tend to be utterly
Weller, 2017; Meister, 2012). Three threads were ultimately selected dissonant from and negatively tinged with regard to marketers’
from Pepsi’s, Kellogg’s, Domesto’s (Unilever) Facebook brand intentions, but also blatantly indifferent. “Management failures can
communities, respectively. Positing at least three brands as the take a number of forms. Fans can simply express boredom and
analytical corpus was deemed necessary to minimize selection bias, indifference” (Yeritsian, 2018, p.7). “Online customer and brand
as well as allow for cross‐case comparisons. The selected brand communities rarely exist in substantial or meaningful ways and there
community threads are as follows: is little evidence of their practical functionality for marketing
management” (Bradshaw & Zwick, 2016, p.3). Rather than constitut-
Pepsi: ‘Pepsi LOVE IT. LIVE IT. LOVE IT LIVE IT! Painting the world ing solid communities of an integrationist orientation, brand
blue with Messi, Marcelo, Kroos, Lloyd and Dele’ communities often resemble situational social collectives whose
ROSSOLATOS | 1255

T A B L E 1 Codebook with a priori NVivo codes

Code name Definition
ATT_POS_REL Positive product attribute: The comment’s thematic focus rests on a positively valenced attribute that is relevant to the
thematic orientation of the brand initiated thread
ATT_NEG_REL Negative product attribute: The comment’s thematic focus rests on a negatively valenced attribute that is relevant to the
thematic orientation of the brand initiated thread
ATT_POS_IRREL Positive product attribute: The comment’s thematic focus rests on a positively valenced attribute that is irrelevant to the
thematic orientation of the brand initiated thread
ATT_NEG_IRREL Negative product attribute: The comment’s thematic focus rests on a negatively valenced attribute that is irrelevant to the
thematic orientation of the brand initiated thread
BEN_POS_REL Positive consumer benefit: The comment’s thematic focus rests on a positively valenced consumer benefit that is relevant to
the thematic orientation of the brand initiated thread
BEN_ NEG_REL Negative consumer benefit: The comment’s thematic focus rests on a negatively valenced consumer benefit that is relevant
to the thematic orientation of the brand initiated thread
BEN_POS_IRREL Positive consumer benefit: The comment’s thematic focus rests on a positively valenced consumer benefit that is irrelevant
to the thematic orientation of the brand initiated thread
BEN_ NEG_IRREL Negative consumer benefit: The comment’s thematic focus rests on a negatively valenced consumer benefit that is relevant
to the thematic orientation of the brand initiated thread
VAL_CONG Congruent personal values: The comment’s thematic focus rests on a personal value that is congruent with the specific brand
value that is projected in the brand initiated thread
VAL_INCONG Incongruent personal values: The comment’s thematic focus rests on a personal value that is incongruent with the specific
brand value that is projected in the brand initiated thread
IND Indifferent: The comment is utterly indifferent to the thematic orientation of the brand initiated thread, e.g., random
remarks, gibberish, misplaced humor

members coexist as imaginary communities (Gruzd, Wellman, & on first order lexemes. Theming also allows for disambiguating the
Takhteyev, 2011). This is why it is particularly important to focus on meaning of words based on the context wherein they are embedded.
indifferent comments, inasmuch as on negative ones, not only To this end, memos were regularly annotated to the coded segments
because of their discrepant orientation with regard to brand initiated in an attempt to identify links between the three laddering strata.
themes, but equally importantly because they are indicative of a
noninteractional attitude (in peer‐to‐peer communication terms),
4.5 | Data synthesis
given the discrepant conversational (posting) turns.
As regards irrelevant posts with regard to the thematic orientation Pursuant to the preparatory stage in the laddering analytical
of a thread, the data thickening approach that is adopted here views procedure whereby the data are screened for salient thematic
them as indispensable signs of negative co‐creation. Despite the fact that patterns, the data were interrogated via frequency analyses, lexical
marketers occasionally seek to contain relevantly negative and/or clouds, clustering exercises, and HVMs with the aid of NVivo
dissonant comments by inviting consumers to send a message privately (Rossolatos, 2014b). The output of these exercises is discussed in
in an attempt to contain the communicative “fire” from spreading virally the following section.
to peers, more often than not irrelevant comments are not tackled at all.
They are left intact in the archival brand memory as vestiges of
4.5.1 | The benefits from using NVivo in mixed
ambivalent semantic orientation, even though they appear to be
methods research designs
compromising both the integrative character of a thread, as well as
the salience of other, perhaps more seriously intended comments. The employment of CAQDAS software (including NVivo) is particu-
Although such comments may not be strictly reflective of the driving larly useful in research designs of both exploratory and confirmatory
theme of a brand initiated thread, they may be expressions of personal orientation that feature a mixed‐methods approach, i.e., a combina-
values that are incongruent with a brand’s DNA. Given the sheer volume tion of qualitative and quantitative methods of analysis and
of this type of comments across brand communities, the application of synthesis, based on different coexisting objectives that may be
the laddering technique in unearthing comprehensively co‐created brand embedded in the same research design. On the one hand, it allows for
meaning looms like a most pertinent methodological avenue. qualitatively inquiring a set of data and producing preliminary
Finally, theming was applied to specific comments to match the insights at an initial exploratory stage. On the other hand, it allows
focalization requirement in the data interrogation process, as it allows for content analyzing a data pool based on both a priori and
for contextualizing data based on underlying dimensions, and hence is emergent codes. By dint of its dynamic nature, the software allows
amenable to more salient reporting, compared to analyses that draw for deleting, merging, or coining new codes as the analysis
1256 | ROSSOLATOS

FIGURE 1 The NVivo project workbench. [Color figure can be viewed at wileyonlinelibrary.com]

progresses, as well as for the generation of hypotheses that may be elaboration of posted comments during the coding process along-
further explored through SPSS. Throughout the process, the side the dimension of relevance to the overarching thread imposes
embedded memo annotation function allows for commenting on more stringent constraints on the co‐creative potential of con-
each part of the analysis/synthesis, thus keeping track of how a sumers’ comments. Subsequently, it yields a more fine‐grained
reading strategy deploys while addressing the same data in multiple picture of consumer engagement in SMBCs. Furthermore, the
waves. manual, line‐by‐line coding that is afforded by the use of CAQDAS
Moreover, NVivo Plus (cf. Figure 1 for a screenshot from this software allows for a nuanced extrapolation of the distinctive
project’s workbench) is uniquely equipped with social media data laddering strata, as well as for data cleaning in terms not only of
analytic functionalities. These range from the initial stage of data
scraping where the software’s Chrome extension NCapture may be T A B L E 2 The laddering table
used for saving and importing data from platforms such as Facebook,
Total, Pepsi, Kellogg, Domestos,
Twitter, YouTube, up to the ability of producing social network
N = 132 N = 38 N = 40 N = 54
graphs (sociograms) for gauging the relationship, for example,
Codes
between influencers and followers in social media communities.
ATT_NEG_IRREL 7% 5% 0% 13%
The software is equipped with various reporting capabilities and
ATT_NEG_REL 28% 5% 83% 4%
templates, most importantly in the context of social media analysis of
multimodal orientation, i.e., allowing for the inclusion in the same ATT_POS_IRREL 2% 3% 3% 0%

research project of both verbal and audiovisual data (videos, pictures, ATT_POS_REL 3% 8% 0% 2%

social memes, and infographs), and the application of a uniform BEN_ NEG_IRREL 3% 3% 0% 6%
coding scheme across modalities. BEN_ NEG_REL 3% 0% 5% 4%
Last, but not least, NVivo allows for a highly nuanced BEN_POS_IRREL 0% 0% 0% 0%
elaboration of primary social media data and the expansion of BEN_POS_REL 2% 3% 0% 2%
sentiment analysis’ interpretive layers. This is afforded by adopting
IND 40% 63% 10% 46%
a highly contextualized stance in interpreting the generated
VAL_CONG 3% 5% 0% 4%
comments’ valencing against the semantic background of individual
VAL_INCONG 10% 5% 0% 20%
threads, rather than an acontextual distinction into positive,
100% 100% 100% 100%
negative, and neutral comments. As shown in Table 1, the additional
ROSSOLATOS | 1257

FIGURE 2 Hierarchical chart of codes’ distribution across the corpus. [Color figure can be viewed at wileyonlinelibrary.com]

irrelevant but of utterly indifferent comments which might pass capabilities, aimed at inquiring the data at a preliminary thematic
unnoticed by an automated coding procedure. level. First, a frequency analysis of the most often recurring words
(including their synonyms) for each brand was produced to identify
potentially salient themes (Table 4). Then, word clouds (Figures 3–5)
5 | D IS C U S S IO N O F F I N D I N G S were produced from the three brand data sets, aimed at theme
discovery based on the relative distribution of lexemes in their
(RQ1) How can we operationalize the laddering strata in the context respective clouds. By referring the high ranking lexemes that were
of SMBCs comments in such a way as to effectively account for both spawned from the first exercise and the central lexemes that
positive, as well as negative brand meaning co‐creation aspects? emerged from the word clouds back to the posted comments, the
As may be gauged from the Laddering Table (Table 2), the bulk of following preliminary insights were yielded by brand:
comments (81%) partake of either the indifferent or the negative As regards Pepsi, the negative themes that emerged most
types, with indifferent comments accounting for 40% of the entire frequently (e.g., plastics recycling or the employment of palm oil in
corpus. This picture is also vividly portrayed in Figure 2 where ¾ of the product’s manufacturing) are ranking quite prominently in the
the hierarchical chart area (NVivo output) is covered by the collective brand memory’s associative inventory. Although these
indifferent/negative types. The incidence of indifferent comments is attributes are irrelevant to the brand initiated theme, their negative
boosted by the Pepsi and Domestos threads, whereas Kellogg’s valencing is sufficiently strong, resulting in brand avoidance in terms
comments are negatively relevant almost in their entirety. Indicative of engaging positively with the brand’s blue vision at a higher level of
comments per type may be found in Table 3. abstraction up the brand’s ladder.
The operationalization of the laddering technique in this largely As regards Kellogg, the underlying themes beneath the most
negative co‐creation brandscape was geared towards identifying and centrally distributed and most often recurring lexemes concerned a
interpreting (a) partial links or absence thereof between laddering massive plea for restoring key sub‐brands’ (such as Rice Crispies and
strata in individual comments (b) links and absences thereof between Coco Pops) sugar levels to earlier product formula versions. This
laddering strata throughout each thread (c) reasons for negatively largely negative meaning co‐creation mode (83% as per Table 2) that
engaging with the brand initiated topic based on the coded rests on an attribute level is responsible for the reluctance on behalf of
comments’ content (d) instances of comments that reflected all disgruntled consumers to engage positively with the brand’s heritage
three laddering strata. as core brand image attribute that is highlighted in the brand initiated
As soon as the coding procedure was finalized, a series of reports theme. Subsequently, any further brand claim on the value dimension
and graphs were produced by leveraging NVivo’s analytical failed to generate positive associations due to bottlenecks at the lower
1258 | ROSSOLATOS

T A B L E 3 Indicative comments by code/brand

ATT_NEG_IRREL
Pepsi You may be aware that there is another large petition going round to pressure Pepsi on your use of unsustainable palm oil in the
food products owned by the Pepsi corporation
Pepsi PepsiCo needs to update its priorities. Instead of focusing on marketing its newest product, PepsiCo needs to address its massive
conflict palm oil problem. There’s nothing “bubly” about the pattern of exploitation by PepsiCo’s notorious
Domestos How are you helping the world when you are using palm oil? Shameful
Domestos If your mission is to help people live better lives then you would think you would encourage people to be environmentally friendly—
not using products which are poisonous to humans, other animals, and the environment
ATT_NEG_REL
Kellogg Hi, I have a question regarding rice krispies. I am craving rice krispies, has the original recipe been brought back? If not, when will it
be brought back?
Kellogg Used to yes! Hate the new tastes! Not buying Kelogg's anymore!!!! Bring back old tastes!
Pepsi Can you not do a campaign with Joan Crawford. Pepsi Cola had such a better image than this
Pepsi Still, Ronaldo is better than Messi
Domestos How can you say that you care about the environment when you've stopped selling the refill pack?? I hate having to carry those big
bottles home every time:(
ATT_POS_IRREL
Kellogg When I was in school I wrote a letter to the Kellogg’s company telling them how much I loved Coco Pops. I even wrote the lyrics to
the Coco Pops song from the commercial. They kindly sent me back a free Coco monkey which I still have and love. I still know the
words to the song
Pepsi Please never cuck out and put ANY sweeteners in your original flavor Pepsi. Cola tastes horrible when not made with sugar
ATT_POS_REL
Pepsi Its been long Pepsi actually had an array of stars that promotes and encourage young talent. Good old days: Pepsi all‐star. (Beckham,
Figo, Edgar Davis, Carlos, Zidane, Shevchenko, Ronaldinho, Ronaldo, etc)
Pepsi Lionel Messi my best
Domestos Very good. Only toilet product been using for too many years to remember
BEN_ NEG_IRREL
Pepsi Please never cuck out and put ANY sweeteners in your original flavor Pepsi. Cola tastes horrible when not made with sugar
Domestos have you tried using chemicals that aren't poisonous?
BEN_ NEG_REL
Kellogg I still don't understand why the original flavor was not kept. I mean most companies keep the original product but still introduce new
flavors. Keep the original for those who love it. The sugar in the new ones is a lot. Don't think it's even good for consumption on a
daily.
Kellogg You don’t know what you’ve got till it’s gone. How true those words ring. Will I ever enjoy a bowl of Coco Pops again in my life?
Domestos How about funding organic sustainable products that sanitize and clean without harming the environment really its a long time
investment to your company don't you think > safety standard says use bleach but doesn't describe there are certain procedures to
use bleach every time you use it…. http://ucfoodsafety.ucdavis.edu/files/26437.pdf
BEN_POS_REL
Pepsi Great event!
IND
Kellogg WHY don't you answer your loyal (now ex) customers?
Kellogg I have sent an email directly regarding a specific cereal, guess what, no response from them. I've seen many people post complaints
on this fb notice, and not once have I seen Kelloggs actually respond to anyone. So this must b normal for them.
Pepsi Please, export Mountain Dew in Italy:(
Pepsi Do hope to see progress on the reduction of plastics this year
Domestos All want an UN peacekeeping force in Arakan, Burma imm to protect Rohingya who are long been persecuted and are subject to
state cleansing Paul Polman
VAL_CONG
Pepsi Hate has changed all that now..#peaceindworld… Pepsi blue world
Domestos Superb idea… Unilever Foods also help for poor families…I'm always a fan of Unilever Foods products
VAL_INCONG
Pepsi Shame on you Pepsi https://www.greenpeace.org.uk/picturesmassive…/
Pepsi In Pictures: Massive deforestation linked to major consumer brands | Greenpeace UK
Domestos Stop testing your products on animals
Domestos What has toilets to do with healthy living. Healthy starts from within the mind, not from the outside. Or do you think different?
ROSSOLATOS | 1259

T A B L E 4 Frequency analysis of most often recurring words

Domestos Kellogg Pepsi

Word Count Word Count Word Count
Environment 6 Rice 18 Product 10
Sustainable 3 Back 17 Public 7
Application 1 Kellogg 14 Life 6
Conditions 1 Krispies 15 Love 6
Continuous 1 Pops 12 Need 6
Corporation 1 Coco 11 Celebrate 5
Deportation 1 Bring 9 Make 5
Discomfort 1 Original 9 Plastic 5
Environmentally 1 Still 8 Coming 4
Everywhere 1 Old 6 Feel 4
Immediately 1 Recipe 6 Greenpeace 4
Impossible 1 Andre 5 Messi 4
Investment 1 Change 5 Oil 4
Marketplace 1 New 5 Palm 4
Netherland 1 Even 4 Put 4
Peacekeeping 1 Kelloggs 4 Save 4
Persecuted 1 Know 4 Still 4
Procedures 1 Loved 4 Used 4
Remarkable 1 Now 4 Become 3
Responsive 1 Products 4 Massive 3
Terminated 1 South 4 New 3
Breizionerd 1 Strawberry 4 Nice 3
#Brightfuture 1 Sugar 4 Wont 3
#Globalgoals 1 Take 4 Blue 2
Globalgoals16 1 Buying 3 Borneo 2
State cleansing 1 Cereal 3 Bottle 2
Ucfoodsafety 1 Complaints 3 Business 2
Umweltzerstoerung 1 Consumers 3 Change 2
Unileverer 1 Flavor 3 Conflict 2
Love 3 Deforestation 2
Much 3 Dew 2
Need 3 Eest 2
People 3 Elephants 2
Product 3 Filled 2
Something 3 Food 2
Taste 3 Free 2
Tastes 3 Great 2
Help 2
Hope 2
Iframe 2
Indofood 2
Lionel 2
1260 | ROSSOLATOS

F I G U R E 5 Pepsi word cloud [Color figure can be viewed at

F I G U R E 3 Domestos word cloud [Color figure can be viewed at wileyonlinelibrary.com]
wileyonlinelibrary.com]

“I gather this is an obituary notice for corn flakes if your

strata of the laddering funnel. This diametrically opposed evaluative
rice krispies strategy is anything to go by….” (Kellogg)
posture between the brand’s projected image and its codestruction by
consumers is reflected quite forcefully in the comment “Why the hell did you change the coco pops????” (Kellogg)
“70 years—yes and you ruined a perfectly good recipe.” (Kellogg)
As regards Domestos, a widespread value incongruence with “Still Ronaldo is better than Messi” (Pepsi)
regard to the projected cause‐related marketing initiative is at stake.
This value incongruence is attributed to consumers’ perceptions As regards the utterly disruptive comments in the corpus, that is
about the brand that revolves around its unsustainable draining of the indifferent ones, it is suggested to term them as instances of
environmental resources. The permeating perception of the mother coindifference, in line with the term codestruction. Whereas codes-
brand (Unilever) as not being environmentally friendly overshadows truction refers to negatively valenced comments and brand image
the ability of the sub‐brand Domestos to generate positive dilution, coindifference designates the utter dissolution of brand
associations with its cause‐related marketing initiative. meaning by dint of being trajectorized in totally random ways (with
On a cross‐case comparative level and with regard to the regard to the thematic contours of specific brand initiated posts).
distinction between brand play and brand attack that was discussed
in section 2, significant differences emerged in terms of the mode “When is Pepsi next coming to Denmark??????” (Pepsi)
whereby relevantly negative comments are voiced in each case. In
the case of Kellogg, relevantly negative comments are voiced in “All will be Clean in God's New World” (Domestos)

frontal attack mode, and collaboratively so, whereas in the case of

“Global Trade Start Here!” (Domestos)
Pepsi they surface in a more playful manner:

F I G U R E 6 Cluster analysis of coded comments for the entire

F I G U R E 4 Kellogg word cloud [Color figure can be viewed at corpus (Pearson’s correlation) [Color figure can be viewed at
wileyonlinelibrary.com] wileyonlinelibrary.com]
ROSSOLATOS | 1261

As regards positively valenced comments, either as attributes or as Finally, as regards instances of comments that reflect all three
benefits, they were marginally encountered in the entire corpus. As laddering strata, only two were discovered in the entire corpus, in the
per Table 2, they are almost equally split between relevant and Pepsi and Domestos threads respectively. The former is a case of
irrelevant ones. This is vividly portrayed in the cluster analysis positive, whereas the latter of negative co‐creation:
(Pearson’s correlation) dendrogram that was produced with NVivo
(Figure 6), in quest for similarities among the coded data. “Clustering Ooh so cool… I really love the product.. Its been long pepsi
data based on similarities in content or coding of qualitative data is actually had an array of stars that promotes and
generally best used in an exploratory manner, to provoke ideas, rather encourage young talent.. Good old days: pepsi all star.
than as explanatory evidence of an association” (Bazeley & Jackson, (beckam,figo,Edgar davis,carlos,zidane,shevchenko,ronal-
2013, p.237; Corter, 1996; Jolibert & Baumgartner, 1997; Macia, dinho,ronaldo, etc) hate has change all that now..#-
2015). The cluster analysis yielded a 3 cluster solution (Figure 6), with peaceindworld… Pepsi blue world (Pepsi)
clusters 2 and 3 branching off to subordinate groupings. Cluster
1 features indifferent comments alongside positive, yet irrelevant This comment is a remarkable instance of a full‐fledged ladder. It
benefits which is suggestive of the minimal semantic distance between is an attestation of this individual’s peace‐loving personal value that
positively irrelevant comments and utterly indifferent ones. This is conditions him into valorizing positively the endorsers who have
further corroborated in the context of cluster 3 where positive, yet paraded in Pepsi’s brand history, also motivated by the benefit of the
irrelevant comments concerning attributes have little difference from brand’s support for young talent. Figure 7 portrays which parts of a
negative, yet relevant ones. Cluster 2 groups negative attributes and brand ladder contribute in shaping up the immediately superior one
benefits with incongruent values, thus attesting to the interdepen- in a HVM. Notice that the relationship between personal values and
dency of the three laddering strata in yielding a concise picture as attributes is symmetrical and not one‐way.
regards comments’ elaboration and valencing. Ultimately, only a small
subcluster in cluster 3 groups positive comments that are fully How about funding organic sustainable products that

relevant with congruent values. sanitize and clean without harming the environment really

This is a particularly important finding that lends further its a long time investment to your company don't you

credence, on the one hand, to the hypothesis that positive, yet think? (Domestos)

irrelevant comments differ marginally from negative ones, while, on

the other hand, being affirmative of the small contribution of As mentioned earlier, the main perceptual barrier in Domestos’
genuinely positive comments in a collective attitudinal state that is case that mediated negatively in the recognition of its cause‐related
marked by negativity and indifference. In this context, consumers marketing initiative, concerned the brand’s nonenvironmentally
appear to be unmotivated and far from willing to participate in friendly orientation. In the Domestos HVM (Figure 8), the consumer
positive co‐creation (cf. Schau‐Hope, Muñiz, & Arnould, 2009), while attempts through negative co‐creation to identify ways whereby this
clinging to issues that pertain to the brand and corporate barrier may be overcome by suggesting that Domestos produces
responsibility, and unfavorable product formula changes. organic products that do convey the benefit of sanitization, albeit in

F I G U R E 7 Pepsi HVM based on a single Facebook brand F I G U R E 8 Domestos HVM based on a single Facebook brand
community comment. HVM, hierarchical value map [Color figure can community comment. HVM, hierarchical value map [Color figure can
be viewed at wileyonlinelibrary.com] be viewed at wileyonlinelibrary.com]
1262 | ROSSOLATOS

an environmentally friendly manner. In this case, negative co‐ 6.3 | Differences in the data collection stage
creation at the apex of a HVM results in positive co‐creation at
Since social media data have not been generated on demand, but
lower levels of abstraction.
constitute secondary data, they do not conform strictly to the
analytical exigencies of a traditional laddering exercise. The main
6 | METHODOLOGICAL RE FLECTIONS O N discrepancy between primary in‐depth interview data and second-
T H E AD A P T A T IO N O F TH E L A D D E R I N G ary social media comments with reference to a laddering blueprint
TEC HN IQUE I N THE AN ALYSIS OF SMBC S consists in that the latter have not been produced through constant
COMME NTS probing. Therefore, from a traditional laddering point of view, they
are bound to be lacking in comprehensiveness as regards the
Although the application of the laddering technique in a Web 2.0 setting spectrum of AVC’s. Nevertheless, this does not entail that they are
may sound plausible on a programmatic level, there are significant not usable. It is just that their usefulness is not represented in full‐
differences compared to its employment in in‐depth interviews. These fledged triads and triadic relationships amongst the components of
differences are punctuated and qualified in the ensuing subsections. ladders. Again, this is an area where the marketer’s intervention is
crucial with a view to promoting positive brand meaning co‐
creation, while minimizing destructive and/or indifferent co‐
6.1 | Differences as regards the elicitation process creation.
and context
Whereas in the context of in‐depth interviews the relationship
between memory contents and the subject of elicitation (the elicitee) 7 | MAN AG ER IA L IMP LI CA TIO NS
is monological, i.e., between the subject and himself, in the context of
SMBCs the relationship not only is dialogical, but what might be (RQ2) How can the identified bottlenecks in the operationalization of
termed polydialogical (as an adaptation of Bakhtin’s concept of the laddering technique in the context of SMBCs be circumvented
dialogism in a Web 2.0 environment). This designates that a subject is with a view to yielding managerially salient insights?
enmeshed by default in a social text that involves other subjects Laddering in the context of SMBCs may be used for addressing a
whose memories are copresent in synchrony with the memory‐ broad range of marketing objectives, from tactical to strategic ones.
retrieval process of the individual elicitee. This predicament also calls It may be used as an acid test for gauging perceptual barriers at an
for a recalibration of the role performed in typical interview settings axiological level which may be thwarting consumers from engaging
by the researcher/interviewer who is represented in this instance by positively with the attributes and benefits that are projected by a
an impersonal other (the community) whose collective voices brand in its posted threads. For example, the majority of negative
perform indirectly the probing function. This is an indirect and comments that were posted in Kellogg’s thread that is included in
largely subconscious probing, albeit a probing that is not marked by this study concerned attributes, with very little to no information
latent intentionality for generating ladders between attributes, with regard to benefits and values. This partial information is inviting
benefits, and values. It is a peculiar sort of subconscious probing as towards further research with regard to less manifest (than taste or
demand for participating in the brand community by contributing sugar levels) drivers, of axiological nature, that may be thwarting
with one or more laddering components to a dynamically and consumers from viewing positively the brand’s long‐standing heri-
collectively morphing ladder that crystallizes as a brand’s ongoing tage. It may also be used as a strategic consumer perception
social text. Thus, from a methodological point of view, and in the management tool over time by setting Key Performance Indicators
absence of a follow‐up interview stage, eliciting latent axiologies by (KPIs) in terms of each ladder stratum, as well as target conversion
starting with concrete product attributes in the context of a social rates between a ladder’s strata.
media brand community, is tantamount to a reconstruction of the Furthermore, it may be applied comparatively across different
individual comments (and their replies) against a laddering pattern. types of brand initiated threads to gauge variable engagement
This process is equivalent to transforming individual brand associa- patterns by type of theme, e.g., new product launch announcements,
tions into a collectively produced brand memory. seasonal promotions, invitations to engage with brand symbols on
various occasions, user‐generated content, competitions, etc. Types
of posts may be segmented alongside a ladder’s engagement levels
6.2 | Differences as regards the research design
and actions taken to boost engagement for low performing posts or
orientation
re‐evaluate their effectiveness in a content marketing plan. “Man-
Whereas the laddering technique has been traditionally embedded in agers may evaluate whether the interactions on their brand pages
qualitative research designs, the sheer volume, the coding complex- are embedding brand meanings in a manner that is consistent with
ities, and the concomitant interpretive exigencies placed upon the brand objectives. Managers can fine‐tune the content that is posted
researcher in the context of a social media data pool call for a mixed on their brand pages to obtain desired meanings” (Rosenthal et al.,
methods research design orientation. 2017).
ROSSOLATOS | 1263

As a method for gauging consumer engagement in brand stage, with regard to demographics, as well as usage and loyalty
communities, laddering may be used in the form of a funnel of status (category and brand wise). Segmentation bases may prolifer-
engagement, from indifferent to irrelevant and negative to irrelevant ate in various directions, depending on the type of consumer
and positive to relevant and negative to relevant and positive feedback that is sought, and the marketing issue at hand. Not all
comments. The more consumers engage co‐creatively with brands segments’ feedback is equivalent or of equal value to specific
across the laddering strata, the more brand relevance is likely to be marketing problems. Unfortunately, this consumer profile informa-
enhanced. Norms for each part of the engagement funnel may accrue tion is not readily available in social media settings. Therefore,
overtime, thus allowing for benchmarking against competitors, as comments may not be taken at face value, as their source may be
well as on an intrabrand level by comparing and contrasting different unduly skewed in favor of one segment at the expense of others,
posts. Monitoring how the dynamics between the strata shift perhaps more valuable ones.
overtime against the background of set KPIs may be a viable route The usefulness of the laddering technique in comprehensively
for devising content and moderation strategies. In this context, mini dimensionalizing SMBCs data may hardly be overstated in light of the
laddering sessions may be conducted through probing on the spot, highly sought after congruence between consumers’ personal values
while members may be recruited for more extensive in‐depth and a brand’s values. As seen repeatedly in SMBCs, the incidence of a
interviews. fundamental incongruity at an axiological level overshadows the
As became evident at the initial screening phase of multiple brand potentially positive valencing of either attributes or benefits at lower
communities, the marketer’s involvement in brand meaning co‐ levels of a brand ladder. For example, when Mondelez moved some of
creation is at an all‐time low. In the majority of cases, the marketer its Oreo manufacturing operations from the US to Mexico in 2015,
vanishes after posting a new topic for discussion. Wherever market- Facebook user‐generated comments were packed with avowals to
ers make their presence felt, most likely they are seeking to contain stop purchasing the brand, regardless of the thematic orientation of
consumer satisfaction issues by urging consumers to send their individual threads. These comments continued surfacing for years in
messages in private. In the case of brands where co‐creation with the a row, a phenomenon which is at least alarming as to how exactly
active involvement of marketers is standard practice (e.g. Starbucks), negative meaning co‐creation may be tapped by brands and perhaps
this is evinced as flattering consumers or thanking them for their favorably reversed. The application of the laddering technique in
comments. Such interventions are far from the ongoing probing that SMBCs is geared towards achieving this principal task.
is mandated in the application of the laddering technique. Since a
customary response pattern to consumers’ comments in a brand
owned SMBCs consists of replying in private, a potential avenue for 8 | L IM I T AT IO N S AN D AR E A S FO R
furthering positive co‐creation with genuinely engaged consumers FUR THE R RE SE AR CH
consists of probing on posted comments privately and in real time,
for example through Messenger. Failure on behalf of marketers to Although the quantitative results of this study with regard to the
perform their co‐creative role in social media is bound to perpetuate distribution pattern by code (Table 2) are by no means representa-
the incidence of indifferent, but also contextually irrelevant tive of a wider population, such as the featured brands’ total SMBC
comments, ultimately leaving brand communities in a state of posts or even on a category level, they are theoretically general-
semiotic frivolity, rather than constituting resourceful pools for izable, as regards the applicability of the microanalytical framework
distilling actionable insights. to other brand cases. “In analytical generalization, it is the concepts
However, there is also a downside as regards the level of control or theories that are generalizable, not the specific context or
marketers may exert in SMBCs. Deleting comments may backfire in populations” (Smith, 2018, p.141). As stressed earlier, but also as
terms of retaliatory actions on behalf of silenced consumers, such as repeatedly noted in the extant literature on negative brand co‐
launching accusations about censorship. However, not omitting some creation in SMBCs where qualitative designs have been largely
comments, especially those that partake of the indifferent and applied, as well as in the laddering‐related literature (e.g. Hakkio &
irrelevantly negative types, may distort the brand’s social memory Laaksonen, 1998), it is customary in case‐study research to employ
and invite further negative co‐creative actions by dint of silently smaller samples which allows for delving deeper into the meaning of
legitimating such posts, i.e., by failing to take action against them. The a dataset. This trade‐off is both inevitable, and a source of
optimal route, in this case, is not readily available. However, the competitive advantage of small versus big data approaches. Ladder-
emerging double‐bind should not lead to paralysis. The bifurcated ing is a well‐tried and tested qualitative and mixed methods approach
options should be weighed on an individual comment basis against across various samples and research objectives. In this sense, issuing
the background of strategic communication objectives as they a call for confirmatory studies geared towards testing the model’s
emerge. conceptual robustness may as well be a pleonasm. The featured
Finally, it should be stressed that SMBCs data are neither a brand threads aimed at exemplifying and discussing the bottlenecks
panacea nor a substitute for traditional consumer research. On the that emerge while applying the laddering technique on SMBCs data,
one hand, in traditional interview settings, specific screening and especially in the face of a permeating negative co‐creative
questions are usually asked during the participants’ recruitment brandscape. Provided that the identified bottlenecks are effectively
1264 | ROSSOLATOS

tackled, the methodology may be strategically leveraged for Bazeley, P., & Jackson, K. (2013). Qualitative data analysis with NVivo.
generating longitudinal insights about levels of consumer engage- London: Sage.
Bouvier, G. (2015). What is a discourse approach to Twitter, Facebook,
ment in SMBCs. Any well known issues of representativeness in
YouTube and other social media: Connecting with other academic
social media studies notwithstanding (primarily stemming from a fields? Journal of Multicultural Discourses, 10(2), 149–162. https://doi.
fundamental unknowability as regards the distribution of brand org/10.1080/17447143.2015.1042381
community participants compared to the distribution of brand users Bradshaw, A., & Zwick, D. (2016). Biopolitical marketing and social media
brand communities. Theory, Culture & Society, 0(0), 1–25.
in a total population), extensive studies on both intrabrand, as well as
Brogi, S., Calabrese, A., Campisi, D., Capece, G., Costa, R., & Di Pillo, F.
on a category‐wide level may yield useful benchmarks for identifying (2013). The effects of online brand communities on brand equity in
consumers’ level of engagement with brand initiated themes along- the luxury fashion industry. International Journal of Engineering Business
side the laddering strata. Management, 5, 32. https://doi.org/10.5772/56854
Caliandro, A., & Gandini, A. (2017). Qualitative research in digital
This study drew on brand controlled and moderated SMBCs. The
environments: A research toolkit. London: Routledge. https://doi.org/
scope of laddering’s applications may be expanded to peer‐to‐peer 10.4324/9781315642161
SMBCs, while comparisons with brand controlled SMBCs may be Cantey, N. J., & Robinson, C. (2015). #BlackTwitter: Making waves as a
undertaken, thus obtaining a more comprehensive picture as to social media subculture. In N. Rambukkana (Ed.), Hashtagpublics: The
differences in engagement along a ladder’s strata. Additionally, power and politics of discursive networks (pp. 219–228). New York:
Peter Lang.
Facebook was opted for as the main platform for sampling purposes.
Cherif, H., & Miled, B. (2013). Are brand communities influencing brands
The technique may be extended to other social platforms, in an through co‐creation? A cross‐national example of the brand AXE: In
attempt to account for differences and additional bottlenecks. Last, France and in Tunisia. International Business Research, 6(9), 14–29.
but not least, the dataset used in this study comprised mainly verbal https://doi.org/10.5539/ibr.v6n9p14
Corter, J. E. (1996). Tree models of similarity and association. London: Sage.
data. The technique may be extended multimodally to encompass
https://doi.org/10.4135/9781412986380
visual and sonic data, especially in brand initiated posts that request Cova, B., Pace, S., & Park, D. (2007). Global brand communities across
of consumers to submit visual materials. borders: the Warhammer case. International Marketing Review, 24(3),
313–329. https://doi.org/10.1108/02651330710755311
Dessart, L., Veloutsou, C., & Morgan‐Thomas, A. (2015). Consumer
A C K N O W L E D GM E N T engagement in online brand communities: A social media perspective.
Journal of Product & Brand Management, 24(1), 1–34. https://doi.org/
I would like to thank the anonymous reviewers of Psychology & 10.1108/JPBM‐06‐2014‐0635
Fujita, M., Harrigan, P., & Soutar, G. (2017). A netnography of a
Marketing for their constructive comments to earlier versions of this
university’s social media brand community: Exploring collaborative
paper. co‐creation tactics. Journal of Global Scholars of Marketing
Science, 27(2), 148–164. https://doi.org/10.1080/21639159.2017.
1283798
OR CID Gruzd, A., Wellman, B., & Takhteyev, Y. (2011). Imagining Twitter as an
imagined community. American Behavioral Scientist, 55(10), 1294–1318.
George Rossolatos http://orcid.org/0000-0002-8884-646X https://doi.org/10.1177/0002764211409378
Gutman, J. (1982). A means‐end chain model based on consumer
categorization processes. Journal of Marketing, 46(2), 60–72. https://
REFERENC ES doi.org/10.1177/002224298204600207
Habibi, M. R., Laroche, M., & Richard, M. O. (2014). Brand communities
Abeele, V. V., & Zaman, B. (2010). Laddering with young children in user based in social media: How unique are they? Evidence from two
eXperience evaluations: theoretical groundings and a practical case. exemplary brand communities. International Journal of Information
Proceedings of the 9th International Conference on Interaction Management, 34, 123–132. https://doi.org/10.1016/j.ijinfomgt.2013.
Design and Children, Barcelona, Spain, 156‐165. 11.010
Amatulli, C., Pino, G., De Angelis, M., & Cascio, R. (2018). Understanding Hagedoorn, B. (2015). Towards a participatory memory: Multiplatform
purchase determinants of luxury vintage products. Psychology & storytelling in historical television documentary. Continuum, 29(4),
Marketing, 1–9. https://doi.org/10.1002/mar.21110 579–592. https://doi.org/10.1080/10304312.2015.1051804
Angus, D. (2017). Theme detection in social media. In L. Sloan, & A. Quan‐ Hakkio, S., & Laaksonen, P. (1998). Relationships in marketing channels:
Haase (Eds.), The SAGE handbook of social media research methods (pp. Examining communication abilities through cognitive structures.
530–544). London: Sage. https://doi.org/10.4135/9781473983847.n31 Psychology & Marketing, 15(3), 215–240. https://doi.org/10.1002/
Bagozzi, R. P., & Pratibha, A. D. (1994). Decisions to recycle: A means‐end (SICI)1520‐6793(199805)15:3<215::AID‐MAR2>3.0.CO;2‐D
chain analysis. Psychology & Marketing, 11(4), 313–340. https://doi. Heding, T., Knudtzen, C. F., & Bjerre, M. (2009). Brand management:
org/10.1002/mar.4220110403 Research, theory and practice. London: Routledge. https://doi.org/10.
Bambauer‐Sachse, S., & Mangold, S. (2011). Brand equity dilution through 4324/9780203996171
negative online word‐of‐mouth communication. Journal of Retailing and Herring, S. C. (2004). Computer mediated discourse analysis: An approach
Consumer Services, 18, 38–45. https://doi.org/10.1016/j.jretconser.2010. to researching online behaviour. In S. A. Barab, R. Kling, & J. H. Gray
09.003 (Eds.), Designing for virtual communities in the service of learning (pp.
Bartoletti, R. (2011). Memory and social media: New forms of 338–376). New York: Cambridge University Press. https://doi.org/10.
remembering and forgetting. In B. M. Pirani (Ed.), Learning from 1017/CBO9780511805080.016
memory: body, memory and technology in a globalizing world (pp. Hollebeek, L. D., & Chen, T. (2014). Exploring positively versus negatively‐
82–111). Newcastle‐Upon‐Tyne: Cambridge Scholars Publishing. valenced brand engagement: A conceptual model. Journal of Product &
ROSSOLATOS | 1265

Brand Management, 23(1), 62–74. https://doi.org/10.1108/JPBM‐06‐ Orsingher, C., Marzocchi, G. L., & Valentini, S. (2011). Consumer (goal)
2013‐0332 satisfaction: A means‐ends chain approach. Psychology & Marketing,
Huber, F., Beckmann, S. C., & Herrmann, A. (2004). Means–end analysis: 28(7), 730–748. https://doi.org/10.1002/mar.20409
Does the affective state influence information processing style? Pongsakornrungsilp, S., & Schroeder, J. E. (2011). Understanding value co‐
Psychology & Marketing, 21(9), 715–737. https://doi.org/10.1002/mar. creation in a co‐consuming brand community. Marketing Theory, 11(3),
20026 303–325. https://doi.org/10.1177/1470593111408178
Jolibert, A., & Baumgartner, G. (1997). Values, motivations, and personal Puschmann, C., & Powell, A. (2018). Turning words into consumer
goals revisited. Psychology & Marketing, 14(7), 675–688. https://doi. preferences: How sentiment analysis is framed in research and the
org/10.1002/(SICI)1520‐6793(199710)14:7<675::AID‐MAR3>3.0. news media. Social Media + Society, 4, 205630511879772. https://doi.
CO;2‐D org/10.1177/2056305118797724
Kamboj, S., & Rahman, Z. (2017). Customer participation in brand Prahalad, C. K., & Ramaswamy, V. (2004). Co‐creation experiences: the
communities on social media: A systematic literature review. next practice in value co‐creation. Journal of Interactive Marketing,
International Journal of Web Based Communities, 13(4), 437–467. 18(3), 5–14. https://doi.org/10.1002/dir.20015
https://doi.org/10.1504/IJWBC.2017.089349 Rauschnabel, P. A., Kammerlander, N., & Ivens, B. S. (2016). Collaborative
Keller, K. L. (1998). Strategic brand management: Building, measuring and brand attacks in social media: Exploring the antecedents, character-
managing brand equity. New Jersey: Prentice Hall. istics, and consequences of a new form of brand crises. Journal of
Keller, K. L. (2009). Building strong brands in a modern marketing Marketing Theory and Practice, 24(4), 381–410. https://doi.org/10.
communications environment. Journal of Marketing Communications, 1080/10696679.2016.1205452
15(2‐3), 139–155. https://doi.org/10.1080/13527260902757530 Reynolds, T. J., & Gutman, J. (1988). Laddering theory, method, analysis
Kozinets, R. V. (2002). The field behind the screen: Using netnography for and interpretation. Journal of Advertising Research, 28, 11–31.
marketing research in online communities. Journal of Marketing Rialti, R., Caliandro, A., Zollo, L., & Ciappei, C. (2018). Co‐creation
Research, 39, 61–72. https://doi.org/10.1509/jmkr.39.1.61.18935 experiences in social media brand communities: Analyzing the main
Kozinets, R. V., de Valck, K., Wojnicki, A. C., & Wilner, S. J. S. (2010). types of co‐created experiences. Spanish Journal of Marketing ‐ ESIC,
Networked narratives: Understanding word‐of‐mouth marketing in 22(2), 122–141. https://doi.org/10.1108/SJME‐03‐2018‐0011
online communities. Journal of Marketing, 74, 71–89. https://doi.org/ Rosenthal, B., Pereira, E., & Brito, Z. (2017). The brand meaning co‐
10.1509/jmkg.74.2.71 creation process on Facebook. Marketing Intelligence & Planning, 35(7),
Kristal, S., Baumgarth, C., & Henseler, J. (2018). “Brand play” versus 923–936. https://doi.org/10.1108/MIP‐09‐2016‐0171
“Brand attack”: The subversion of brand meaning in non‐collaborative Rossolatos, G. (2014a). Brand equity planning with structuralist rhetorical
co‐creation by professional artists and consumer activists. Journal of semiotics. Kassel: Kassel University Press.
Product & Brand Management, 27(3), 334–347. https://doi.org/10. Rossolatos, G. (2014b). Exploring the rhetorical semiotic brand image
1108/JPBM‐01‐2017‐1405 structure of ad films with multivariate mapping techniques. Semiotica,
Latzko‐Toth, G., Bonneau, C., & Millette, M. (2017). Small data, thick data: 2014, 335–358. https://doi.org/10.1515/sem‐2014‐0012
Thickening strategies for trace‐based social media research, Califor- Rossolatos, G. (2018a). A sociosemiotic approach to consumer engage-
nia: Sage. ment in user‐generated advertising. Social Semiotics, 28(4), 555–589.
Liew, K. K., Pang, N., & Chan, B. (2015). Remembering digitally: new media https://doi.org/10.1080/10350330.2017.1381452
and the communicating of new memories. Continuum, 29(4), Rossolatos, G. (2018b). Interdiscursive readings in cultural consumer
509–5010. https://doi.org/10.1080/10304312.2015.1051801 research. Newcastle upon Tyne: Cambridge Scholars Publishing.
Laroche, M., Habibi, M. R., Richard, M. O., & Sankaranarayanan, R. (2012). Saaka, A., Sidon, C., & Blake, B.F. (2004). Laddering: a “how to do it”
The effects of social media based brand communities on brand manual – with a note of caution. Methodology Series, Cleveland State
community markers, value creation practices, brand trust and brand University. Retrieved from http://academic.csuohio.edu:8080/cbrsch/
loyalty. Computers in Human Behavior, 28(5), 1755–1767. https://doi. home.html
org/10.1016/j.chb.2012.04.016 Salmons, J. (2017). Using social media in data collection: Designing studies
Lee, M. S. W., & Soon, I. (2017). Taking a bite out of Apple: Jailbreaking with the qualitative e‐research framework. In L. Sloan, & A. Quan‐
and the confluence of brand loyalty, consumer resistance and the co‐ Haase (Eds.), The SAGE handbook of social media research methods
creation of value. Journal of Product & Brand Management, 26(4), (pp. 177–197). London: Sage. https://doi.org/10.4135/9781473983
351–364. https://doi.org/10.1108/JPBM‐11‐2015‐1045 847.n12
Macia, L. (2015). Using clustering as a tool: Mixed methods in qualitative Schau, H. J., Muñiz, A. M., Jr., & Arnould, E. J. (2009). How brand
data analysis. The Qualitative Report, 20(7), 1083–1094. community practices create value. Journal of Marketing, 73(Sep),
Marlowe, J. M., Bartley, A., & Collins, F. (2017). Digital belongings: The 30–51. https://doi.org/10.1509/jmkg.73.5.30
intersections of social cohesion, connectivity and digital media. Schoonenbooma, J., Burke, R., Johnson, M., & Froehlich, D. E. (2017).
Ethnicities, 17, 85–102. https://doi.org/10.1177/1468796816654174. Combining multiple purposes of mixing within a mixed methods
Martinez‐Lopez, F. J., Anaya‐Sanchez, R., Aguilar‐Illescas, R., & Molinillo, S. research design. International Journal of Multiple Research Approaches,
(2016). Online brand communities: Using the social web for branding and 10(1), 271–282. https://doi.org/10.29034/ijmra.v10n1a17
marketing. London: Springer. https://doi.org/10.1007/978‐3‐319‐24826‐4 Smith, B. (2018). Generalizability in qualitative research: Misunderstand-
Mayr, P., & Weller, K. (2017). Think before you collect: Setting up a data ings, opportunities and recommendations for the sport and exercise
collection approach for social media studies. In L. Sloan, & A. Quan‐ sciences. Qualitative Research in Sport, Exercise and Health, 10(1),
Haase (Eds.), The SAGE handbook of social media research methods (pp. 137–149. https://doi.org/10.1080/2159676X.2017.1393221
107–124). London: Sage. https://doi.org/10.4135/9781473983847.n8 Sorensen, A., & Drennan, J. (2017). Understanding value‐creating
Meister, S. (2012). Brand communities for fast moving consumer goods: An practices in social media‐based brand communities. The Service
empirical study of members' behavior and the economic relevance for the Industries Journal, 37, 986–1007. https://doi.org/10.1080/02642069.
marketer. Wiesbaden: Springer. 2017.1373098
Miles, M. B., & Huberman, M. A. (1994). Qualitative data analysis. London: Van Dijck, J. (2007). Mediated memories in the digital age. Stanford:
Sage. https://doi.org/10.1007/978‐3‐8349‐4055‐1 Stanford University Press.
Muniz, A. M., Jr., & O' Guinn, T. C. (2001). Brand community. Journal of Vitak, J. (2017). Facebook as a research tool in the social and computer
Consumer Research, 27(4), 412–432. https://doi.org/10.1086/319618 sciences. In L. Sloan, & A. Quan‐Haase (Eds.), The SAGE handbook of
1266 | ROSSOLATOS

social media research methods (pp. 627–645). London: Sage. https://doi. Consumer Culture, 0(0), 146954051878757. https://doi.org/10.1177/
org/10.4135/9781473983847.n37 1469540518787576
Wansink, B. (2003). Using laddering to understand and leverage a brand’s Zaglia, M. E. (2012). Brand communities embedded in social networks.
equity. Qualitative Market Research: An International Journal, 6(2), Journal of Business Research, 66, 216–223. https://doi.org/10.1016/j.
111–118. https://doi.org/10.1108/13522750310470118 jbusres.2012.07.015
Weijo, H., Hietanen, J., & Mattila, P. (2016). New insights into online
consumption communities and netnography. Journal of Business Research,
67(10), 2072–2078. https://doi.org/10.1016/j.jbusres.2014.04.015
Wiegandt, P. (2009). Value creation of firm‐established brand communities. How to cite this article: Rossolatos G. Negative brand
Wiesbaden: Gabler. https://doi.org/10.1007/978‐3‐8349‐8460‐9 meaning co‐creation in social media brand communities: A
Woodside, A. G. (2010). Case study research: Theory, methods and laddering approach using NVivo. Psychol Mark.
practice, Bingley: Emerald.
2019;36:1249–1266. https://doi.org/10.1002/mar.21273
Yeritsian, G. (2018). Participation from above and below: Brand
community and the contestation of cultural participation. Journal of