On CSR-related identities, storytelling strategies and the communicative dynamics between organizations and the public
Cathrin Schlüter University of Amsterdam
Student Number: 10842055 Master's Thesis
Graduate School of Communication
Master's Program: Communication Science Corporate Communication
Supervisor: Associate-Prof. Dr. Friederike Schultz 26th of June, 2015
!
Abstract
This paper scrutinizes CSR-related identities and storytelling strategies of different organiza-tional types, and further elaborates on the dynamics between organizations and the public as a result of their communication. The empirical part drew on an automated content and semantic network analysis of corporate weblog and Facebook posts and public's perception in form of electronic word-of-mouth (eWOM) to investigate possible difference of the two objects with regard to actors and issues, pragmatic intentions and associative frames. The results show that expected differences of CSR-related identities failed to appear, whereas storytelling strategies were proved to be distinct, indeed framing different organizational purposes. Public's eWOM, however, portrayed CSR-related identities and storytelling strategies, whereby significantly reframing the latter ones.
Keywords: Corporate identity, CSR, organizational purpose, storytelling, eWOM,
!
On CSR-related identities, storytelling strategies and the communicative dynamics between organizations and the public
It is more than just a recently upcoming and short-term trending topic: corporate so-cial responsibility (CSR) and its role in the organizational context. CSR, rather is, by now, an established issue that arrived on the agenda of academia and practitioners (e.g., Carroll, 1991; Lee & Carroll, 2011; Pollach, 2014; Scherer & Palazzo, 2011). This development is rein-forced due to strategic advantages ascribed to CSR: increased profits, beneficial costumer relationships or maintaining legitimacy (e.g., Du, Bhattacharya, & Sen, 2010; Schultz, Castelló, & Morsing, 2013).
Being further related to societal problems such as environmental protection, weak-nesses in developing countries or, more generally, deficiencies worth striving for change (Jones, 1995; Morsing & Schultz, 2006; Scherer & Palazzo, 2011; Schultz et al., 2013), CSR is per se burdened with values grounded in society's morality. Nevertheless, as society is di-verse, human beings differ in the way they are intrinsically tied to these issues. Acting as human representatives and being characterized by identities and values, organizations equate with human beings (Christensen & Cornelissen, 2011; Hatch & Schultz, 2002). Playing with this metaphor, the outlined correlation between intrinsic CSR dedication and identity can also be applied to organizations: organizations differ in the way they jump on the CSR wave, ded-icating their corporate identity to CSR on deviating levels (Du, Bhattacharya, & Sen, 2007; Hess, Rogovsky, & Dunfee, 2002). Research classifies along a continuum lasting from social to business entrepreneurships (Alvord, Brown, & Letts, 2004; Mair & Martí, 2006; Peredo & McLean, 2006), when analyzing organizations' fundamental reason to exist. This continuum is, for the purpose of the study, used to determine different organizational types and their CSR dedications. Although several studies take aspects of CSR's interaction with organiza-tional factors into account (Becker-Olsen, Cudmore, & Hill, 2006; Du et al., 2007), an
analy-!
sis including the coaction of organizational purposes, corporate identities and CSR dedica-tions is lacking. This paper aims to close the gap by scrutinizing this interplay. Central atten-tion is paid to the analysis of expected differences of so-called CSR-related identities with regard to actors and issues.
The same way CSR-related identities have to harmonize with organizational purposes, corresponding corporate communication strategies have to be carefully adjusted (Christensen & Cornelissen, 2011; Cornelissen & Harris, 2001). Especially when adapting the constructiv-ist perspective on communication, which implies that reality is socially constructed in collec-tive sensemaking processes of organizations and the public (Putnam & Nicotera, 2009; Weick, 1979, 1995). This outlook leads to the credo that communication constitutes organiza-tions (CCO), while allowing for ambiguity, polyphony and disagreement. Communicative encounters, thus, frame and produce organizational meanings (Christensen & Cornelissen, 2011; Cornelissen et al., 2012; Schultz et al., 2013; Taylor & van Every, 2000). As a conse-quence, organizations need profound knowledge of applying effective frames that might guide this collective sensemaking process. According to Hallahan (1999) storytelling is a form of framing, used to transfer the salience of certain issues and attributes to the public agenda (see also Entmann, 1993). As stories are further seen as persuasive narratives that undermine organizations' raison d'être (Dowling, 2006), it is assumed that organizations uti-lize corporate stories to frame organizational purposes, leading to different storytelling strat-egies. To analyze whether storytelling strategies indeed differ, this paper relies on a semantic network analysis.
As the environment in which communication takes place has undergone, and continu-ously faces significant changes, the traditional way organizations act as communicators on a playing field of social media, public debates and raising complexity, has to modify. Commu-nication models insisting on senders providing input, which directly reaches receivers in a
!
predetermined way, are outdated and no longer meet the expectations the communicative environment requires (Castelló et al., 2013; Cornelissen et al., 2012; Luoma-aho & Vos, 2010). Consequently, organizations not only face the challenge of shifting their corporate communication to social media (Waters, Burnett, Lamm, & Lucas, 2009), but also of adjust-ing their way of communicatadjust-ing CSR-related identities content-wise. Pragmatic intentions, empirically rooted in linguistics, provide the opportunity to evaluate the literal meaning and the underlying intention of messages (Austin, 1962; Searle, 1969, 1976). Hence, they ideally serve as schemata of analysis for online communication messages. However, prior research only applied this theory on the micro level (e.g., Carr, Schrock, & Dauterman, 2012; Hassell & Christensen, 1996; Nastri, Peña, & Hancock, 2006), calling for an extension on the corpo-rate level. This paper follows this call, also due to scholars' interests in knowing how to bene-ficially use social media for corporate communication (e.g., Christ, 2007; Kent & Taylor, 1998; Waters et al., 2009), analyzing the online communication of CSR-related identities with regard to pragmatic intentions.
Only adapting an organizational viewpoint on CSR-related identities and storytelling strategies would undoubtedly not reflect their full academic potential, thus, the public's per-ception is attributed equal relevance in this paper. Whereas the public's communicative be-havior online (i.e. electronic word-of-mouth, eWOM), caught academic attention regarding, e.g., its motives (Hennig-Thurau, Gwinner, Walsh, & Gremler, 2004) or its influence on product evaluations (Doh & Hwang, 2009), little is known about its direct interplay with cor-porate communication. This paper further analyses public's perception of CSR-related identi-ties and storytelling strategies in terms of eWOM.
Combining the two perspectives introduced, the following research question (RQ) evolves:
!
RQ: How do CSR-related identities and storytelling strategies of different organizational types influence public's perception in terms of communicative behavior online (eWOM)?
To gradually answer the overarching RQ, two complementary sub-questions are formulated, each covering one of the perspectives introduced.
RQa: To what extent do CSR-related identities and storytelling strategies of different organi-zational types differ with regard to (a) actors and issues, (b) associative frames and (c)
pragmatic intentions?
RQb: To what extent are (a) actors and issues and (b) associative frames formulated in CSR-related identities and storytelling strategies of different organizational types reflected in the
public's eWOM?
Although the theoretical components independently caught academic attention, their not negligible relatedness remains unnoticed in the academic field. This paper closes this gap and further enriches this relation by elaborating on the resulting dynamics between organiza-tions and the public. Broadening the view on framing by including a semantic network analy-sis allows for profound insights into how online communication gathers its momentum of its own. As a whole, the paper not only contributes to scientific knowledge, but also provides benefits for practitioners to handle the communicative challenges they face.
To reflect upon the research questions, the author, first, reviews relevant literature. To reach empirical proof, the author continues conducting an automated content and semantic network analysis of corporate weblog and Facebook posts of three exemplary mobile phone companies, namely Fairphone, OnePlus and Blackphone, all expected to mirror different CSR-related identities and storytelling strategies. After presenting and discussing the results, the author concludes with limitations of this paper, leading to avenues for future research.
!
Corporate CSR-related Identities
Organizational Type as the Principal Reason for Organizations' Existence
Performing as constitutive actors, organizations have to make decisive strategic choices, leading them through their daily operational affairs. Which choices are predominant, also depends on the organizational type and its purpose. Research, for instance, differentiates between social and business entrepreneurships (e.g., Austin, Stevenson, & Wei-Skillern, 2006; Peredo & McLean, 2006; Seelos & Mair, 2005). Although a variety of fuzzy defini-tions is used when differentiating between social or business entrepreneurships, researchers share the common belief that social entrepreneurships strive to start social movements in or-der to improve social and environmental deficiencies and to raise societal awareness of these unsatisfied needs (e.g., Alvord et al., 2004; Austin et al., 2006; Mair & Martí, 2006; Seelos & Mair, 2005). Overarching goal is to create social value, instead of fulfilling shareholder value or maximizing wealth as for business enterprises (Austin et al., 2006; Peredo & McLean, 2006). Additionally, catalyzing social change by mobilizing the crowd is not only a short-term goal. Rather, long-short-term social transformation, including fundamental structures of think-ing and behavior, is aimed (Alvord et al., 2004; Mair & Martí, 2006). That is, regardless of any obstacles, social entrepreneurships continuously aim to achieve their social missions, which, as being the fundamental reason to exist, notably influences their corporate identities (Austin et al., 2006; Gamson, 1989). In most of the cases, social entrepreneurships make true business moves (i.e. selling products or services) to first reach financial independence, which in turn allows them making progress in the underlying social purpose (Di Domenico, Haugh, & Tracey, 2010; Kroeger & Weber, 2014; Mair & Martí, 2006).
Although differences become clear, both organizational types remain intertwined and unique characteristics of one type inspire the operations of the other one. Consequently, being classified as either a social or a business entrepreneurship does not follow a strict pattern.
!
Organizational types are instead characterized along a continuum, ranging from pure social to pure business orientations. In the same way extreme cases are possible, divergences towards and away from both ends of the continuum are common in today's business landscape (e.g., Austin et al., 2006; Peredo & McLean, 2006; Seelos & Mair, 2005).
What is meant by 'social'? Equal to the blurred lines between both organizational types, the question of what is being assessed as 'social' arises (Kroeger & Weber, 2014; Mair & Martí, 2006; Seelos & Mair, 2005). Indeed, the ideal of setting the highest priority on serv-ing society's needs is a first hint, but still lacks of clarification. A document published by the European Commisson (2002) accentuated the importance of incorporating CSR in corporate strategies, in order to create sustainable social welfare and to meet society's needs, which is related to the "social" aspect of social entrepreneurships. This view is substantiated by re-search: Seelos and Mair (2005) stressed 'social's' relatedness with CSR issues and sustainable development (i.e. environmental challenges), especially when acting as social entrepreneurs. Respectively, Christensen, Morsing, & Thyssen (2013) underscored that communicating CSR issues has the power to initiate social change, a quality which is also part of the social entre-preneurship definition (Alvord et al., 2004; Mair & Martí, 2006). Thus, acting as social en-trepreneurs can further be analyzed along CSR issues, which, corresponding to the continu-um, assume deviating intensities.
Using the Organizational Type to Further Define Corporate Identities
Assessing the organizational type as the principal foundation for organizations' exist-ence, the point of departure to further develop strategic steps is an organization's identity (e.g., Abratt & Kleyn, 2012; Cornelissen, Haslam, & Balmer, 2007; de Chernatony, 1999). Corporate identities encompass organizations' most decisive strategic choices and their cor-porate expressions (Abratt & Kleyn, 2012). Reflecting upon questions like 'Who are we?' 'What do we stand for?' and 'What do we want to be in the future?' are a starting point to
for-!
mulate corporate identities and integral identity issues (Abratt & Kleyn, 2012; Balmer & Gray, 2003; Urde, 2003). As the construct is multidimensional, such strategic choices are only a first step of identity development and are subdivided and manifested in strategic mis-sions, visions and corporate values (Abratt & Kleyn, 2012). Corporate core values that enrich corporate identities are of great significance, as they play a key role in the
consumer-organization identification process (Balmer & Gray, 2003). Moreover, and of utmost gravity when recalling the implications of the CCO approach, core values serve as main themes for organizations' communication strategies, guiding the communication process with internal and external stakeholders (Urde, 2003; van Riel, 1995).
Finalizing Corporate Identities: Institutionalizing CSR
Although a clash of opinion regarding a precise and complete definition of CSR (Christensen et al., 2013), its scope of validity (Castelló et al., 2013; Scherer & Palazzo, 2011), its managerial implications and outcomes (Bhattacharya & Sen, 2004; Brown & Dacin, 1997), sets the tone in academic literature, the combination of different viewpoints leads to an integral, yet imperfect clarification of CSR. This perspective includes four dimen-sions of CSR, to wit an economic, an ethical, a legal and a philanthropic dimension (Carroll, 1991; Lee & Carroll, 2011). Each of the four qualities comprises its individual characteris-tics, but at the same time, enables the subsequent one. That is, starting as being economically responsible (i.e. maximizing financial assets), via obeying the law and fulfilling ethical re-sponsibility by behaving right and fair, organizations gradually reach the highest level, which is philanthropic responsibility. It includes producing resources and contributing to societal welfare (Carroll, 1991; Lee & Carroll, 2011).
Even though society does not per se impose the role of being philanthropically re-sponsible on organizations (Carroll, 1991; Lee & Carroll, 2011), to date, CSR issues arrived on the public agenda, strengthening society's awareness of these concerns (Uscinski, 2009).
!
In case of such issues, the causality of influence is a different one: commonly, the media agenda determines the issues on the public's agenda (e.g., Carroll & McCombs, 2003; McCombs & Shaw, 1972; Wu & Coleman, 2009). But in this case, public's CSR awareness exerts pressure on organizations to incorporate CSR in corporate identities (Carroll, 1991; Lee & Carroll, 2011; Uscinski, 2009). Consequently, with the business environment being on its peak of competitiveness, organizations feel the urge to follow this call, in order to fulfill raising stakeholder demands (Lee & Carroll, 2011; Pollach, 2014). As this burden seems too powerful to ignore, it appears obvious that organizations follow and accept this pressure of institutionalizing CSR in corporate identities differently.
Variations across Organizational Types: Different CSR-related Identities
Touching on the introduced definitions of organizational types, corporate identities and CSR, several crucial ties can be delineated. First, the qualities of being philanthropically responsible coincide with the underlying purpose of social entrepreneurships. However, Car-roll (1991) further indicated that not all organizations reach the highest level of CSR respon-sibility, leading to the proposition that various CSR integration level exist (see also Du et al., 2007), which in turn, corresponds with the continuum of dedicating organizations from social to business purposes (e.g., Austin et al., 2006; Peredo & McLean, 2006; Seelos & Mair, 2005). By interpreting the CSR dedication as the aspirational choice regarding organizations' positioning, Brown and Dacin (1997) highlighted the integrative character of CSR into corpo-rate stcorpo-rategies, primarily coined by business types and secondly, by corpocorpo-rate identities. This line of reasoning is also confirmed by Hess et al. (2002), arguing that in order to be success-ful, CSR issues have to arise from and be connected with organizational core values, deter-mined in corporate identities. Taken together, this leads to the assumption that business types and corporate identities predetermine CSR identifications, resulting in the so-called
CSR-!
related identities. As the determinants vary, CSR-related identities are expected to differ, re-sulting in RQa1.
RQa1: To what extent do CSR-related identities of different organizational types differ with regard to actors and issues?
Making Sense of Organizational Communication: Corporate Storytelling and Associa-tive Framing
Corporate Storytelling
Concluding that expressing core values is storytelling, and, vice versa, storytelling is expressing core values, Jensen (1999) uncovered the roots of corporate stories: playing with characteristics of fairytales, stories communicate organizations' personalities and correspond-ing core values to the public. By becorrespond-ing theoretical rooted in framcorrespond-ing (Hallahan, 1999) and by further relying on cognitive psychology, stories have the power to change people's minds and, on a corporate level, to influence public's attitude towards organizations (Dowling, 2006; Gill, 2014; Lundqvist, Liljander, Gummerus, & van Riel, 2012). Thus, by catching public attention and melting corporate reality with narrative elements, stories possess the unique power of what is referred to as 'narrative transportation' (Escalas, 2004; Green & Brock, 2000): being exposed to corporate stories, public's attitude towards organizations "is left changed" (Lundqvist et al., 2012, p. 286), also regarding the understanding of organiza-tions' true purpose and benefits.
As corporate storytelling raises less skepticism and negative feelings towards organi-zations, it gains notable importance in the context of CSR. To lower suspicion towards rate CSR issues, organizations have to find a balanced way of integrating CSR in their corpo-rate communication stcorpo-rategies (Bhattacharya & Sen, 2004; Du et al., 2010; Morsing, Schultz, & Nielsen, 2008; Morsing & Schultz, 2006; Ziek, 2009). As CSR issues originate from
cor-!
porate core values (Colleoni, 2013), storytelling is thought to be an effective narrative tech-nique to communicate CSR issues (Urde, 2003).
Associative Framing
Frames. By defining frames as "schemata of interpretation", Goffman (1974, p. 21) not only paved the way for an extensive stream of research, but also sparked long lasting dis-cussions about what framing actually means and causes (e.g., Benford & Snow, 2000; Entmann, 1993; Scheufele & Tewksbury, 2007; Scheufele, 1999; Snow, Rochford, Worden, & Benford, 1986). Entmann (1993), seeking to clarify the concept, defined framing as focus-ing on selected attributes in order to increase public awareness, i.e. makfocus-ing certain attributes more salient, while discounting others, leading to a distinct interpretation of reality. This cor-responds with Scheufele's (1999) explanations, stating that framing socially constructs reality by burdening issues and events with meaning.
Especially within the context of social movements, the concept of framing is of severe relevance: to initiate social change, organizations have to apply "collective action frames" (Benford & Snow, 2000; Snow et al., 1986) . By doing so, they provide a structure to create shared meaning, which in turn leads to collective identities (Benford & Snow, 2000; Gamson, 1989). Further developing this idea, Snow and Benford (1988) categorized frames into three types: diagnostic, prognostic and motivational frames to define their underlying intention more precisely (see also Benford & Snow, 2000). Diagnostic frames indicate the cause of the problem, attributing its responsibility. Defining strategic solutions for the problem identified is the core quality of prognostic frames. Lastly, motivational frames aim to propel collective actions, encouraging the public to actually follow the call to actions provided by shared struc-tures of meaning (Benford & Snow, 2000; Snow & Byrd, 2007).
Associative frames. Associative frames built on the cognitive paradigm (D'Angelo, 2002) and are theoretically rooted in framing theory. By expanding framing's scope and
over-!
coming its weaknesses, they emphasize the interrelatedness and mutual, non-hierarchical associations between issues, actors and objects in semantic networks, by considering public's cognitive information processing (Ruigrok & van Atteveldt, 2007; Schultz, Kleinnijenhuis, Oegema, Utz, & van Atteveldt, 2012). Hence, associative frames take the complexity of
communicative structures into consideration, by observing the fact "that the meaning of objects or concepts derives from their multiple relations and integration in meaning net-works." (Schultz et al., 2012, p. 99; based on Cassirer, 1944; Geertz, 1973).
Framing Organizational Purposes: Different Storytelling Strategies
Depending on the organizational type, it is anticipated that different storytelling strat-egies are prevalent, also because they basically derive from corporate core values. This as-sumption finds further proof in Gamson's (1989) thoughts. He underlined that entrepreneur-ships, striving for social change, "are engaged in battles over meaning" (Gamson, 1989, p. 457). Although this statement originates from a different time, its implications can be applied to, and even gain more remarkable importance in today's media landscape: by being present in social media, organizations are exposed to an online crowd, heightening the battle over meanings in online arenas, where reality and organizations are commonly constituted by communicative encounters between organizations and stakeholders (i.e. CCO approach). Consequently, influencing the online agenda to catalyze social change and setting the right frames is even more important when communicating online, as the battle over meaning as-sumes a new dimension. And this might, indeed, not only account for social entrepreneur-ships, but also for business entrepreneurentrepreneur-ships, albeit differences are assumed.
Social entrepreneurships, for instance, do not necessarily know from the beginning whether their mission to catalyze social change is successful, also because of the interde-pendence with the public, who automatically co-engages in the creation of social entrepre-neurships' future. This leads to the proposition that, following Snow and Benford's (1988)
!
typology of frames, social entrepreneurships might include causal and responsibility meaning structures (i.e. diagnostic frames) in their corporate stories, also because they do not know if their stories close successfully, leaving the consequences and solutions open and apart. Vice versa, being able to continuously expect resource stability, business entrepreneurships are in the position to also present consequences and solutions (i.e. prognostic frames), resulting in closed stories. This assumption is further stressed by the fact that social entrepreneurships per se depend on resources, not only financially-wise, but more importantly, also strategically-wise in terms of a supportive crowd (e.g., Alvord et al., 2004; Austin et al., 2006; Gamson, 1989). By constantly attributing causes and responsibility, i.e. convicting a scapegoat, social entrepreneurships might trigger public's accusation and zest for action, leading to collective identities, supporting the social mission. Consequently, and as opposed to business entrepre-neurships, social entrepreneurships might rely on diagnostic, instead of prognostic frames. Using this argumentation as the underlying logic for associative frames, different associative networks are expected to evolve. Whether these differences indeed hold empirical proof needs to be further investigated, leading to RQa2.
RQa2: To what extent do storytelling strategies of different organizational types differ with regard to associative frames?
Organizational Communication via Social Media Using Social Media as a Corporate Communication Tool
In line with society's adaptation of social media (Peters, Chen, Kaplan, Ognibeni, & Pauwels, 2013), it is a logical consequence that organizations integrate social media in their communication strategies, as they represent the arena where the public is present and reacha-ble (Kent & Taylor, 1998; Waters et al., 2009). Distinct characteristics such as an egalitarian power structure, not being restricted by time and space, as well as social media's key function in providing organizations with the opportunity to actively engage in dialogue, cause
corpo-!
rate communication's shift to take place in online issues arenas (Hennig-Thurau et al., 2004; Peters et al., 2013; Waters et al., 2009). Although it seems social media bear advantages for organizations, the different voices in online issue arenas should warn them: reinforced by the CCO perspective, online arenas nowadays get more dynamic, complex and polyphone (Castelló et al., 2013; Cornelissen et al., 2012; Luoma-aho & Vos, 2010), also specifically regarding CSR issues (Colleoni, 2013; Kent & Taylor, 1998). Hence, organizations not only have to respect the influence the public holds when using social media as a corporate com-munication tool, but also have to strategically design their online agenda when communi-cating CSR-related identities.
Communicating CSR-related Identities with Pragmatic Intentions
Corporate communications are the means to communicate CSR-related identities to public agendas (Christensen & Cornelissen, 2011; Cornelissen & Harris, 2001; Cornelissen, Haslam, & Balmer, 2007), and as the shift towards social media as a corporate communica-tion tool implies, to online agendas. As stated, organizacommunica-tions ought to feed their online agen-das with favorable content in order to at least partly determine public' perception and the con-tent development.
As a strategic answer, pragmatic intentions are, very recently, becoming the center of attention. Rooted in linguistics, pragmatic intentions manifest themselves in form of speech acts. According to Austin's (1962) speech act theory, they are defined as an utterance, a speaker's intention and the corresponding effect of these words meant to reach the receiver (see also Searle, 1969, 1976). Speech acts can be analyzed regarding three different levels: locutionary, perlocutionary and illocutionary, whereby the last level, the speakers' true inten-tion, further classifies speech acts into five different patterns. Assertive speech acts transfer meaning in order to make factual and updating announcements. Directive speech acts request the public to perform particular actions, whereas commissive speech acts commit the speaker
!
to future actions. Uttering attitudes, emotions and feelings about issues or state of affairs is classified as an expressive speech act. Lastly, declarative speech acts are characterized by using a particular issue to illustrate an attitude or point of view (Searle, 1976).
Translating CSR-related Identities: Differences with regard to Pragmatic Intentions Conceptualizing the analysis of online communication with the speech act theory al-lows to directly investigate the meaning senders aim to transfer. Although the theory was applied in several studies regarding online communication on the micro level, including the specific context of Facebook (Carr et al., 2012; Hassell & Christensen, 1996; Nastri et al., 2006), research regarding speech acts' corporate meaning is still outstanding. However, the theoretical explanations above provide a foundation to assume differences between organiza-tional types regarding their use of pragmatic intentions when communicating CSR-related identities. In the context of social movements, the intended meaning of words might be even of greater relevance than for business entrepreneurships, due to the mutual dependence with the public (Benford & Snow, 2000). However, if this argumentation holds true is further in-vestigates with RQa3.
RQa3: To what extent do CSR-related identities of different organizational types differ with regard to pragmatic intentions?
Dynamics between organizations and the public: Public's Perception of CSR-related Identities and Storytelling Strategies
Public's Perception in Terms of eWOM
Changing the perspective and focusing on the public's point of view, corporate online communication portrays and humanizes organizations, enabling the public to identify itself with organizations (de Chernatony, 1999; Urde, 2003). Due to this relationship between or-ganizations and the public and linked with the CCO perspective, it is assumed that the public reacts to formulated CSR-related identities and storytelling strategies, in order to take part in
!
the collective sensemaking process (McCombs, Shaw, & Weaver, 2014; Morsing & Schultz, 2006; Schultz et al., 2013), leading to a public agenda.
In online settings, this communicative behavior articulates itself in form of eWOM. eWOM is further defined as any statement, no matter which valence, "[…] made by potential, actual, or former consumers about a product or company […]." (Hennig-Thurau, Gwinner, Walsh, & Gremler, 2004, p. 39). As soon as such a statement is online, it is available for an indeterminable and uncontrollable amount of people (Hennig-Thurau et al., 2004). As eWOM leads to a significant shift of the public, relying on consumer-generated information rather than on any other corporate source, it influences public's evaluations of products and brands, having long-term consequences for organizations (e.g., Doh & Hwang, 2009; Hennig-Thurau et al., 2004; Lee & Youn, 2009).
eWOM with regard to actors and issues. Developing a framework with antecedents and consequences of CSR, Bhattacharya and Sen (2004) underlined the importance of under-standing consumers' internal (i.e. awareness and attitudes) and external (e.g., eWOM) reac-tions towards CSR communication. This call for attention, combined with the CCO perspec-tive, forces organizations to develop a sense of how CSR-related communication is per-ceived, and whether it gathers a momentum of its own, being exposed to an to an uncontrol-lable dynamic and intensity in online arenas. To this end, scholars further claim that CSR communication is in urgent need of more scientific attention (Golob et al., 2013), also in or-der to keep pace with the raising importance of the topic itself (Ziek, 2009) or to avoid possi-ble negative consequences of communicating CSR (Du et al., 2010). This paper expands sci-entific knowledge by analyzing whether communicated CSR-related identity actors and is-sues even reach the public, leading to RQb1:
RQb1: To what extent are actors and issues formulated in CSR-related identities of different organizational types reflected in public's eWOM?
!
eWOM with regard to associative frames. Once concluding that the media "put[s] pictures inside people's head of the world beyond their reach", Lippmann (1922, p. 31) high-lighted media's power on public's scope and creation of reality. Although the statement did not reflect the correlation of the present context, its implications can still be extrapolated. As organizations continuously forward corporate online stories, the public is exposed to different associative frames. On this basis, the public ideally develops its reality interpretation, prede-termined by organizations. However, by communicating via social media, organizations are, by all means, able to easily reach out to the public, but also run the risk that their associative frames lose the intended meaning structure in the dynamics of the public discourse. An anal-ysis of how associative frames are perceived, is thus of great importance, also because these associations in public's long-term memories partly determine reactions and behavioral inten-tions towards organizainten-tions (Franzen, 2009). This leads to RQb2.
RQb2: To what extent are associative frames formulated in storytelling strategies of different organizational types reflected in public's eWOM?
Method Research Design
For the purpose of this study, a primary focus was set on the communicative dynamics between organizations' corporate online communication and public's perception in terms of communicative behavior online (i.e. eWOM). A case study design was employed to answer the research questions. Fairphone, Blackphone and OnePlus were the exemplary organiza-tions taken under consideration when analyzing corporate communication in terms of corpo-rate weblog and Facebook and posts and public reactions in form of comments.
Case Study
Fairphone. Stating that its raison d’être "as a social enterprise is to use commercial strategies to create social impact" ("Fairphone fact sheet", 2014, p. 1), Fairphone clearly
indi-!
cates its underlying organizational purpose: contributing to society's need and creating social welfare. Although Fairphone starts building up its social entrepreneurship with a true busi-ness move (i.e. selling phones), this action is only seen as a first step, gradually fulfilling the long term goal of strengthening consumers' awareness of CSR issues ("Fairphone fact sheet", 2014). Taken together, these characteristics reflect the description of entirely acting as a so-cial entrepreneur, having CSR itself as the main purpose of origin. Thus, Fairphone is chosen as a representative of the extreme case being a social entrepreneur, which is expected to de-termine its CSR-related identity and storytelling strategy.
OnePlus. Contrary, OnePlus follows the mantra "#never settle", aiming to become "the next global technology company" ("About us", 2015). In line with this overarching goal, OnePlus aims to offer the best smartphones, reaching higher technological standards than competitors. CSR issues are far away from being the company's reason to exist, instead, OnePlus strives for maximum business assets ("About us", 2015). Following the theoretical conceptualization above, OnePlus is assessed as the extreme business entrepreneurship, with-out any discernible CSR integration into its identity, which is again thought to predefine OnePlus' CSR-related identity and storytelling strategy.
Blackphone. To complete the spectrum, Blackphone is included in the research. Blackphone's main priority is to put privacy first, securing consumers' private data and usage information as "Blackphone is more than just hardware and software. It's your opportunity to take control of your privacy." ("Company", 2015). Although the value of privacy is not as evidently related to CSR as Fairphone's social value, depending on the perspective, privacy can be assessed as a social value, serving both societal and individual needs (Margulis, 2003). Ergo, Blackphone is chosen as the median between Fairphone and OnePlus, partly integrating CSR, in terms of privacy values in its CSR-related identity. Congruent with the other two
!
companies, Blackphone's storytelling strategy is thought to be influenced by its organization-al purpose.
Units of Analysis
Units of analysis consisted of corporate weblog and Facebook posts and public reac-tions in form of comments. Representing one of the traditional forms of social media, web-logs, i.e. webpages with chronological reverse posts, providing the possibility for the public to leave comments (Kaplan & Haenlein, 2010), are nowadays commonly used as a corporate communication tool (Kaplan & Haenlein, 2010; Kolbitsch & Maurer, 2006; Rosenbloom, 2004). Fieseler, Fleck, & Meckel (2010) further argued that corporate weblogs serve as use-ful CSR communication platforms, leading organizations to certainly communicate CSR is-sues on weblogs. Facebook was chosen due to its main function, providing organizations and the public the possibility to communicate on equal terms (Waters et al., 2009), which reflects the correlation anticipated in the present context.
Each corporate post was treated as a single unit of analysis, whereas on public's side, all the comments on one corporate post were analyzed as one aggregated unit of analysis. Corresponding to the research questions, each unit of analysis was analyzed with regard to actors and issues, associative frames and pragmatic intentions. However, the latter ones were, due to their nature, only studied on the corporate agenda.
Data Collection
The data collection took place between 30th of March and 5th of April 2015. All exist-ing weblog posts of the three organizations under investigation, as well as correspondexist-ing public comments, were downloaded. The timeframe for the blog posts was closed on the 1st
of March 2015. As the Facebook data of the cases showed different scopes, not possible to ignore, and in order to avoid an overrepresentation of one of the cases (Herring, 2004), a cross-sectional sampling was applied, including the last 100 Facebook posts and their public
!
comments from 1st of March 2015 backwards. To assure that only corporate posts were inte-grated in the sample, either Fairphone, OnePlus or Blackphone had to be the designated au-thor of the post. Table 1 presents an overview of the sample.
Table 1
! ! !
Overview of the Sample
! ! !
Case Retrieved from Number of
Corporate Posts
Number of Public
Com-ments
Fairphone Weblog http://www.fairphone.com/blog/ 91 91*
Facebook https://www.facebook.com/Fairphone 100 100*
N = 191 N = 191
OnePlus Weblog https://blog.oneplus.net 48 48*
Facebook https://www.facebook.com/oneplusofficial 100 100*
N = 148 N = 148
Blackphone Weblog https://blog.blackphone.ch 28 28*
Facebook https://www.facebook.com/BlackphoneOfficial 100 100*
!! !! N = 128! N = 128
* Aggregated answers towards one corporate post, excluding comments in other languages than Eng-lish and corporate responses.
Data Analysis
To analyze the communicative dynamics between organizations and the public, this research built on qualitative and quantitative research methods. The software AmCat (Amsterdam Content Analysis Toolkit) (van Atteveldt, 2008) was chosen to conduct an au-tomated content and semantic network analysis. To run further statistical tests, the data was imported in SPSS.
Preliminary analysis. The author first conducted a qualitative sample reading to evaluate the data's quality and to detect for the context of the mobile phone industry and of CSR relevant actors, issues and pragmatic intentions. This procedure paved the way for a search string with predominant actors, issues and pragmatic intentions, which served as a foundation for a following exploratory analysis with AmCat. The search string was build on the Apache Lucene query language, including Boolean Operators (OR, AND, NOT), wild cards (* and ?, as in technolog*, implying that for instance technological or technology could be coded) and keyword proximity conditions (i.e. in the context of improving social
condi-!
tions, the words "fair", "social" or "improve*" should appear within a maximum distance of three words from "conditions", "approach" or "venture") (van Atteveldt, 2008). AmCat's "keywords in context" and "summary list" functions enabled to author to thoroughly verify the search string. After these verification and reliability procedures, a more precise search string could be proved. Ultimately, as argued by Schultz et al. (2012), its overall validity needed to be assessed in order to avoid mistakes of first and second order ("text elements wrongly taken as a 'hit'" and "text elements wrongly skipped as a 'hit'" p. 4). The final key-word list was then used for subsequent analyses.
Actors, issues and frames. The abovementioned approach revealed the following actors: company representatives, such as the CEOs or the online community managers, indus-trial actors (e.g., producer, manufacturer), political representatives (e.g., government, court), public actors, including the community or customers, independent actors with keywords like NGO or civil society and the media (e.g., newspaper, journalist).
Benford and Snow's (2000) systematization of diagnostic and prognostic frames was used as an underlying foundation for the qualitative reading, revealing context-relevant, pre-dominant issues. Diagnostic frames were operationalized according to Schultz and Raupp's (2010) definition, stating that "actors make attributions if they are saying who (attribution sender) makes whom (attribution addressee) for what (attribution object) in which way (eval-uation) and in relation to which cause (locus) responsible" (p. 3) (built on Gerhards,
Offerhaus, & Roose, 2007). Further highlighting the linkages between attribution theory and diagnostic frames, Schultz and Raupp (2010) assembled the issues of responsibility and cause to diagnostic frames (see also Benford & Snow, 2000). Hence, in the present context, diag-nostic frames were covered by cause and responsibility issues, with keywords such as "ex-ploitation" or "political instability" and "responsibility" or "guilt", respectively. Prognostic frames included ascriptions with regard to consequences of and solutions for the problem
!
identified (Schultz & Raupp, 2010). Hence, different consequences and solutions deriving from being a mobile phone company were operationalized as issues. Consequences were subdivided into the following categories: business, social and environmental consequences. Business consequences influence daily business operations and their development (e.g., "ac-celerating prices", "stimulated economy"). While producing an industrial good, mobile phone companies cause environmental consequences, for instance "resource depletion" or "climate change". Social consequences affect society's constitution and quality (e.g., "working condi-tions", "child labor").
Following the same logic, solutions were systematized as social, business or environ-mental solutions. Business solutions refer to issues indicating an improvement of the business landscape, such as "industry improvement" or "better quality". Reducing pollution ("recycle phones") or fairly respecting the resources that nature provides ("raw materials") were cate-gorized as environmental solutions. Social solutions refer to CSR issues with the overall goal to improve society's welfare and to ameliorate societal deficiencies (e.g., "fair principles" or "increased wages"). An overview of all the keywords covering actors and issues is presented in Appendix A.
Pragmatic intentions. To conceptualize pragmatic intentions, Austin's (1962) speech act theory, in combination with Searle's (1976) further development, was applied. The quali-tative reading uncovered relevant keywords, identifying the conceptualized speech act. How-ever, also Searle's (1976) initial suggestions were indicatory when designing the search string. Assertive speech acts were covered by keywords such as "release" or "call", directive speech acts by "share" or "subscribe", commissive ones by "later today" or "soon", and ex-pressive speech acts by "apologize" or "greet". Appendix A presents a full overview.
Frequency analysis. To assess the amount of attention spent to actors, issues and pragmatic intentions, a frequency analysis for each unit of analysis was performed, using the
!
developed Lucene search string. The data was reported in an aggregated way for weblog and Facebook posts, as well as separately. However, as Blackphone's corporate weblog did not allow the public to leave comments, only public's comments towards Blackphone's Facebook posts were part of the frequency analysis. In case one of the keywords of the Lucene search string was found in the right context, it was considered as one hit.
Semantic network analysis. While attributing associations between different actors and issues, semantic network analysis relies on Tversky's conditional probability approach (1977). Its main thought ascribes associations between objects an asymmetrical relation, de-fined in terms of conditional probabilities. Put differently, the probability that concept A will be expressed after concept B was mentioned, is asymmetrical (Tversky, 1977). For this study, a semantic network analysis was performed to analyze associative frames on the corporate and on the public agenda. The results were displayed collectively for blog and Facebook posts and included the weight for the number of hits. However, public's eWOM towards Blackphone's Facebook posts was too weak to expose conditional probabilities. Hence, only a comparison between Fairphone's and OnePlus' public agendas was possible. To further re-duce the complexity of the network graphs, only the three most frequent actors, namely pub-lic actors, the industry and the relevant company representatives were displayed. Arrows showed the relative conditional probability (i.e. p) that concept B was discussed after concept A was mentioned. The more striking the arrows, the stronger the association between the concepts.
Correlation analysis. To reveal the strength of the relationship between the corporate and the public agenda, a correlation analysis was performed. The corporate agenda was un-derstood as the accumulated blog and Facebook agenda, except for Blackphone. As there no corresponding public agenda with respect to blog comments existed, only Blackphone’s Fa-cebook agenda was used for the correlation analysis. Public agendas were understood in the
!
same way, however, differently for Blackphone. A distribution analysis, including kurtosis and skewness, revealed that the variables measuring the corporate and the public agenda were not normally distributed, which could be intensified due to the fact that they were imported as count variables in SPSS. Consequently, a logarithmic transformation had to be performed, a procedure to correct distributional abnormalities. Hereafter, Pearson's correlation measure was performed.
Regression analysis. To predict the possible influence of actors, issues and pragmatic intentions on public’s evaluation of the corresponding Facebook post in terms of likes, sever-al simple linear regression ansever-alyses were performed. Following the same logic as above, a logarithmic transformation had to be performed in order to correct the distributional abnor-malities of the relevant variables.
Results
CSR-related Identities and Storytelling Strategies of different Organizational Types Actors and issues. Fairphone attached the highest importance to public actors: in 60% of the documents under investigation, the social entrepreneurship addressed direct rep-resentatives of society, such as consumers or the community supporting the company (Ap-pendix B). This was followed by industry actors (42%) and Fairphone actors (30%). Fair-phone comparably seldom mentioned independent actors (16%), such as NGOs. Political actors (13%) and the media (6%) were rarely mentioned.
With regard to the operationalized issues, Fairphone, remarkably, was most concerned with solutions. Specifically, business solutions caught the highest value (45%), followed by environmental (31%) and social (26%) solutions. Fairphone, thus, relied less on causal mean-ing structures (cause: 40%; responsibility: 15%) and least on consequences (business: 12%; social: 11%; environment: 7%).
!
Comparing actors and issues formulated in Fairphone's CSR-related identity to OnePlus' and Blackphone's, an interesting picture evolved (Appendix B). The three compa-nies paid the highest attention to public actors. However, Fairphone relatively highlighted public actors the most (60%), followed by OnePlus (40%) and Blackphone (30%). The same patterns held true for industry actors (Fairphone: 42%, OnePlus: 16%; Blackphone: 13%). OnePlus placed itself on the third position (12%), whereas Blackphone on the second (28%). Comparable to Fairphone, OnePlus and Blackphone gave least attention to independent ac-tors (2%, 4%), political acac-tors (1%, 9%) and the media (3%, 2%).
In line with Fairphone, OnePlus and Blackphone also took most notice of solutions, followed by causes (cause: 8%, 14%; responsibility: 1%, 5%) and consequences (business: 3%, 2%; environment: 2%, 0%; social: 0%, 2%). Interestingly, among all issues, the social enterprise scored the highest value for business solutions (45%), whereas the exemplary business enterprise only scored 28%. Blackphone displayed the lowest value for business solutions (19%). Social solutions again exposed expected attention: Fairphone having the highest value (26%), OnePlus the lowest (7%) and Blackphone the median (13%).
Summarized, only small differences of CSR-related identities among different organi-zational types with regard to actors and issues were found. For both issues and actors, the differences were solely true for the relative intensities with which they were mentioned and not for the highest frequencies.
Associative frames. Fairphone strongly accused the industry for being responsible for social evils and deficiencies (p(Industry|Responsibility) = 0.93) (Fig. 1, Appendix D). Fur-ther, Fairphone consistently mentioned consequences regarding, for instance, poor working conditions, environmental pollution or accelerating prices for goods or services in relation to the industry (p(Industry|ConsequenceSocial) = 0.90, p(Industry|ConsequenceEnvironment) = 0.84, p(Industry|ConsequenceBusiness) = 0.88). Notably, the social enterprise saw itself as
!
solvent for business solutions (p(SolutionBusiness|Fairphone) = 0.62), as apparent in the fol-lowing: "we [Fairphone] could […] help hold the entire industry to a higher standard." (van Abel, 2013b). While aiming to improve the supply chain, Fairphone also associated the industry to business solutions (p(SolutionBusiness|Industry) = 0.64), demanding manufac-tures to be open for the "[…] input in specifying the components and vendors […]" or to work with their "preferred subsuppliers" (van Abel, 2015).
Fundamental differences became obvious when comparing these results to OnePlus' and Blackphone's associations (Fig. 2 and Fig. 3, Appendix E, F). OnePlus covered the strongest associations around public actors as for example
p(PublicActors|ConsequenceBusiness)= 0.96 or associated itself with the public
(p(PublicActors|OnePlus)= 0.90), as in the following statement "[…] we [OnePlus] invest our time in our users. By engaging with them on the forums and social media, we have real con-versations with people interested in OnePlus." (Cat, 2014). Although the associations with which Fairphone was related to public actors showed a comparable value
(p(PublicActors|Fairphone) = 0.87), public actors, in general, did not have such a centralistic role as in OnePlus' online communication. These findings were also consistent for Black-phone (p(PublicActors|BlackBlack-phone) = 0.68). As stated above, FairBlack-phone related the responsi-bility to the industry, whereas OnePlus did not judge the industry responsible
(p(Industry|Responsibility) = 0.00) and Blackphone only to a lower degree
(p(Industry|Responsibility) = 0.45). When comparing the association with which OnePlus (p(SolutionBusiness|Industry) = 0.53) and Blackphone (p(SolutionBusiness|Industry) = 0.44)
Fig. 1. Associative frame Fairphone,
!
related the industry to business solutions, Fairphone's value (p(SolutionBusiness|Industry) = 0.64) was outstanding, as the social enterprise covered the strongest association.
The illustrations above underscored that indeed disparities in storytelling strategies of different organizational types with regard to associative frames existed. In essence, Fairphone strongly blamed the industry for various consequences, whereas OnePlus arranged all its online communication around public actors in the center of attention. Blackphone's storytell-ing strategy emerged as mixture of both, as the industry as well as public actors were part of the associative frame, both having mid-centered positions.
Pragmatic intentions. Fairphone mostly made use of directive speech acts (65%) such as "Share our story" (Mier, 2013) or "Join us" (Wernink, 2013) (Appendix B). This was
Fig. 3. Associative frame Blackphone,
cutoff level = 0.60.
Fig. 2. Associative frame OnePlus,
!
followed by expressive speech acts (48%), assertive and commissive speech acts (40%, 38%). OnePlus and Blackphone followed the same patterns (Directive: 41%, 34% Expres-sive: 34%, 12%; Assertive: 24%, 21%; CommisExpres-sive: 20%, 12%) (Appendix B).
The three companies under investigation only showed small deviations in their CSR-related identities with regard to pragmatic intentions. The differences were restricted to the relative attention the companies spent to the corresponding speech act, but did not account for the highest frequency with which they were applied.
Dynamics between Organizations and the Public: Public's Perception of CSR-related Identities and Storytelling Strategies of Different Organizational Types
Actors and issues. Also on Fairphone's public agenda, public actors caught most at-tention (56%) (Appendix C), only slightly less than on Fairphone's corporate agenda (60%). Likewise, the industry (26%) was assigned to the second position, and the company itself (24%) to the third (Fairphone's agenda: 42%, 30%). Equivalent to the numbers on Fair-phone's agenda, independent actors (7%), political actors (16%) and the media (6%) rated the lowest values.
Regarding issues, it became obvious that business solutions (40%) were predominant, comparable to the corporate agenda (45%). Although both environmental (14%) and social solutions (18%) were reflected on the public agenda, the values were lower than on the cor-porate agenda (31%; 26%). Despite the weaker intensity, also causal meaning structures (Cause: 29%; Responsibility: 6%) were present on the public agenda. Similar distributions between public and corporate agenda were also true for consequences (business: 7%, 12%; environment: 11%, 7%; social: 12%, 11%).
When associating the results to OnePlus' and Blackphone's public agenda, outstanding findings emerged. Unique among the triplet, the intensity paid to actors on OnePlus' public agenda was higher than on the corporate agenda (Appendix C), revealing peak values for
!
public actors (78%, corporate agenda: 40%). This indicated, that the OnePlus' online com-munity seemed to develop its own dynamic, still re-communicating and taking over the actors predetermined by OnePlus, but with a higher intensity. Compared to Fairphone, none of the actors reached more attention on the public than on the corporate agenda. This was also true for Blackphone's case, which showed the only value for Blackphone actors (2%) on the pub-lic agenda. In OnePlus' case, the pubpub-lic agenda spent the second highest attention to the in-dustry (22%), which corresponds with Fairphone's public agenda (26%). Yet, for OnePlus, the public agenda again reached a higher level than the corporate agenda (16%).
Also concerning issues, OnePlus public's eWOM developed an incomparable dynam-ic. Again, all issues played a more important role on the public than on the corporate agenda. OnePlus' public was mostly concerned with business solutions (47%, corporate agenda: 28%) and causes (37%, corporate agenda: 8%). On Blackphone's side, causes were the only issues, which were retold by the public, however, only once (1%).
With regard to RQb1 it can be concluded that actors and issues formulated in CSR-related identities of social and business entrepreneurships were indeed reflected in public's eWOM. However, public's acceptance differed among organizational types: in Fairphone's case the numbers on the public agenda were approximately comparable to the ones on the corporate agenda, whereas in OnePlus' case, the relative weighting spent to actors and issues was higher on the public, than on the corporate agenda.
Associative frames. In Fairphone's case, it became apparent that public actors signifi-cantly moved into the focus of communication (Fig. 4, Appendix D). Public Actors, compa-rable to Fairphone’s corporate agenda (p(PublicActors|Responsibility) = 0.83), made them-selves responsibly for causing problems (p(PublicActors|Responsibility) = 0.85), as in the following sentence "And we of the community should have a responsibility not to exert even more pressure." (Lämmert, 2013). This is linked with the probability that public actors were
!
stated in eWOM that mentioned causes (p(PublicActors|Cause) = 0.82). By stating "Many (most) customers have bought the Fairphone for the social impact it aims to achieve." (Heiner, 2014), a Fairphone supporter spoke for the crowd behind Fairphone, first, being aware of their reliability within the movement and second, making themselves responsible for social solutions (p(PublicActors|SolutionSocial) = 0.82). This leads to the conclusion that instead of primarily displaying strong associations between responsibility and consequences and their relationship to different actors and issues, as Fairphone did, public's eWOM devel-oped its own dynamic, rather showing strong associations all centered around public actors.
Scrutinizing OnePlus' public eWOM, comparable structures became apparent. In general and similar to OnePlus' corporate agenda, public actors were remarkably positioned in the center of attention (Fig. 5, Appendix E). For Fairphone, this structure only evolved within the process of the public leaving comments on the public agenda, but was not as evi-dently found on the corporate agenda. In line with the findings on OnePlus' corporate agenda (p(PublicActors|Responsibility) = 0.88), the relation between responsibilities and public ac-tors was found among the strongest on the public agenda (p(PublicAcac-tors|Responsibility) = 1.00). This corresponds with Fairphone’s values (Fig. 1, Fig. 4). The OnePlus' community indeed understood OnePlus' organizational business purpose, seeing the company as solvent for business solutions p(SolutionBusiness|OnePlus) = 0.62) as in the following statements: "In 5 years OnePlus will be the biggest tech company. No doubt." (ZdecydowanieTy, 2014) or "[…] OnePlus delivers beyond expectations. […] So, please, please: never settle."
Fig. 4. Associative frame
!
(Chronos, 2014). Besides that, also the social entrepreneurship, Fairphone, on the corporate and on the public agenda, was related to business solutions (p(SolutionBusiness|Fairphone) = 0.62; p(SolutionBusiness|Fairphone) = 0.68), even with stronger association than the busi-ness entrepreneurships was.
In essence, the extend to what associative frames formulated in storytelling strategies of different organizational types were reflected in public's eWOM, partly differed for organi-zational types. In total, the inter-agenda comparison demonstrated that, for Fairphone and OnePlus, the public agendas developed their own dynamics. In Fairphone's case, public ac-tors first reached the center of attention after being discussed on the public agenda, whereas for OnePlus, this central position even got intensified on the public agenda, compared to the corporate agenda. However, inter-company-wise, expected differences due to the different organizational purposes almost failed to appear. For both corporations, associations related to business solutions, as well as the ascription of responsibility, seemed to be predominant on the public agendas.
Further findings
Agenda interplay. Among all companies and agendas, the only significant result was found for Fairphone's relationship between the corporate and the public agenda with regard to actors (Appendix G). There was a negative, significant relationship between the actors men-tioned on the corporate agenda and the ones on the public agenda, r = -.32, p < .001, inferring
Fig. 5. Associative frame OnePlus’
!
that in case actors on Fairphone's agenda caught more attention, the attention on the public agenda got less.
Post evaluation. Only the significant results are presented hereafter, for a full over-view see Appendix H. For Fairphone, the regression models with likes as dependent variable and business solutions or directive or expressive speech acts as independent variables were significant: Business solutions: F(1, 81) = 4.85, p = .030 with R2 = .06; Directive: F(1, 81) = 5.47, p = .022 with R2 = .06; Expressive: F(1, 81) = 6.51, p = .013 with R2 = .07. Per
an-nounced business solution, the predicted number of likes decreased by 2.60, per directive speech act by 2.53 and per expressive speech act it increased by 3.20. In OnePlus' case, the regression model with likes as dependent variable and commissive speech acts as independ-ent variable was significant: Commissive: F(1, 41) = 7.40, p = .010 with R2 = .15. Per com-missive speech act the number of likes increased by 21.13. For Blackphone, the models with likes as dependent and directive or commissive speech acts as independent variables were significant: Directive: F(1, 55) = 4.26, p = .044 with R2 = .07; Commissive: F(1, 55) = 6.08, p = .017 with R2 = .10. Per directive speech act the number of likes increased by 0.56 and per commissive speech act in decreased by 1.40.
Discussion
The underlying purpose of this study was to investigate CSR-related identities and storytelling strategies of different organizational types, and above that, to analyze the com-municative dynamics with the public emerging as a result of their communication. The re-search relied on an automated content and semantic network analysis of corporate weblog and Facebook posts of three exemplary organizations, namely Fairphone, OnePlus and Blackphone, and of corresponding eWOM towards these posts in form of comments.
The first sub-questions asked to what extent CSR-related identities and storytelling strategies of different organizational types differed with regard to actors and issues,
associa-!
tive frames and pragmatic intentions. Concerning actors, the three companies showed the same patterns, scoring the highest values for public actors, followed by the industry and the corresponding company actors. By appreciating public actors the most, Fairphone follows its organizational purpose: according to theoretical derivations, social entrepreneurships depend on a supportive crowd in order to successfully achieve their social mission (Alvord et al., 2004; Mair & Martí, 2006). That is, by continuously addressing the "[…] amazing network of followers, buyers […]." or the "[…] strong and lively community […]." (van Abel, 2014) Fairphone underpinned its respect for the public, ideally leading to its support. That also OnePlus and Blackphone highlighted public actors the most, could have been intensified by the nature of the units of analysis: being posted on social media, the highly praised medium to foster relationships and to engage in dialogue with the public, corporate posts per se target public actors (Kent & Taylor, 1998; Waters et al., 2009), regardless of the organizational purpose. Comforting, research claims that also business enterprises use social media as a pub-lic relationship management tool (Zeng, Chen, Lusch, & Li, 2010). This rationalization might also account for the fact that all three companies mostly relied on directive speech acts, when communicating CSR-related identities with regard to pragmatic intentions. By doing so, they could have aimed to involve the community, gaining beneficial support, which again, is nec-essary, no matter which underlying purpose organizations hold. An additional analysis re-vealed that in Blackphone's case, directive speech acts indeed had the power to significantly predict the popularity of a Facebook post: the more directive utterances, the higher the amount of likes. However, it has to be acknowledged that the opposing case arose for Fair-phone. This paradox indicates that demanding the public to further engage with organizations is able to negatively influence the evaluation of a Facebook post. This might be, for instance, intensified due to the information and mental overload the public is exposed to on social
me-!
dia (Zeng et al., 2010), which could cause a refusal of further active participation as asked by directive speech acts (Jones, Ravid, & Rafaeli, 2004).
Considering issues, a far more critical argumentation sparked. Notably, the three dif-ferent companies all took most notice of business solutions, the social enterprise Fairphone even reaching peak values. Theory, however, implied that the interplay between organiza-tional purposes, corporate identities and CSR dedications, determines different CSR-related identities with regard to issues. The anticipated effect failed to appear. Obviously, issues are not as evidently linked to organizational purposes and corporate identities as anticipated (Brown & Dacin, 1997; Hess et al., 2002). Theory also claims that the characteristics of so-cial and business enterprises are intertwined, and also soso-cial entrepreneurships often start their social missions with business moves (e.g., Kroeger & Weber, 2014; Mair & Martí, 2006). This is true for Fairphone, itself proclaiming in a blog post about its social impact that they are a "[…] social enterprise, applying commercial strategies to maximize social impact." (van Abel, 2013a). Yet, this causing a predominant focus on business solutions could not have been expected. Adapting a broader perspective, this finding partly questions the theoret-ical distinction between social and business entrepreneurships. Obviously, many grey areas in which both organizational types act the same way exist. Although literature is aware of this fact, they are still studied as black and white opponents (e.g., Mair & Martí, 2006; Seelos & Mair, 2005). The results theoretically implied that an adjustment of the perspective on the two organizational purposes is necessary, as the grey area gets superior to their differences.
Opening this line of argumentation for the results of the semantic network analysis, new valuable insights emerged. Noteworthy, Fairphone was the only company among the triplet, which obviously related the responsibility to the industry. As theorized, the social enterprise thus convicted a scapegoat, possibly triggering public's zest for action which in turn leads to its necessary support (Austin et al., 2006; Gamson, 1989). However, this fact
!
also provided the basis for a paradoxical finding: Fairphone also associated the industry with solutions, mostly with business solutions, almost on the same level than itself, and, above that on a higher level than the business enterprise, OnePlus. Thus, the social enterprise framed its storytelling strategy around the industry, judging it, on the one hand, responsible for causing problems, but on the other hand also solvent for them. By doing so, Fairphone lowers and partially hands over its responsibilities of actually unfolding solutions. This could be seen as a confession to the organizational purpose 'social entrepreneurship' itself: while aiming to improve society's well-being, social enterprises might per se depend on the support of other influential drivers, such as the industry. Asking others for help appears to be a realistic, nei-ther overhasty, nor activist approach, which might be more successful than to solely rely on the supportive crowd behind the movement, as literature demands (e.g., Alvord et al., 2004; Benford & Snow, 2000; Mair & Martí, 2006). Aside from this argument, the finding contra-dicts theoretical assumptions, as the social enterprise was not expected to present a closed story in terms of solutions. Whether this can be assessed as successful is doubtable, as an additional analysis also showed that presenting business solutions in Facebook posts signifi-cantly lowered their evaluation in terms of likes. Conspicuously, in OnePlus' storytelling strategy, public actors were assigned a centralized role. If actors were excluded from the net-work graph, almost no associations would be left. This could indicate the strong individualis-tic orientation of the business enterprise: the company itself could frame the phone as a mean for self-expression, satisfying personal needs of hedonistic beings. In sum, this leads to a product-, consumer-focused storytelling strategy. As the descriptions above underlined, the social entrepreneurship, in contrast, manifested a production-, societal-focused storytelling strategy. Blackphone indeed framed around the industry and public actors, however, a little stronger around the latter ones, displaying a consumer-focused storytelling strategy.
