The communication of sustainability on social media: the role of dialogical communication

Abstract

Purpose

This research analyses how dialogical communication on social media enhances the effectiveness of sustainability messages in terms of brand engagement and purchase intention. Dialogical messages generate social media engagement, which helps improve consumer responses.

Design/methodology/approach

Two experimental studies (2 × 2) were conducted, in which two factors were manipulated: the dialogical potential of the message (dialogical versus monological) and message orientation (sustainability versus commercial).

Findings

The dialogical potential of the message moderates the effect of sustainability messages on brand engagement and purchase intention. Results indicate that sustainability messages featuring dialogical elements generate greater brand engagement and purchase intention than commercial messages. Furthermore, social media engagement mediates these effects.

Practical implications

This study offers valuable insights into applying dialogical principles to sustainability communication on social media. Marketers should design dialogical messages to foster dialogue with customers and enhance engagement.

Originality/value

Few studies have focused on analysing the effects of applying dialogical strategies on social media to communicate sustainability. Thus, this study highlights the importance of dialogical communication beyond the inclusion of interactive elements when communicating sustainability on social media. The inclusion of dialogical features specifically benefits sustainability messages, given the transparency and honesty they demand.

Citation

Herrada-Lores, S., Palazón, M., Iniesta-Bonillo, M.Á. and Estrella-Ramón, A. (2024), "The communication of sustainability on social media: the role of dialogical communication", Journal of Research in Interactive Marketing, Vol. ahead-of-print No. ahead-of-print. https://doi.org/10.1108/JRIM-10-2023-0372

1. Introduction

The sustainability discourse has gained popularity in recent years, with pro-environmental sustainability initiatives receiving greater attention from business management (Kang and Sung, 2022). Indeed, the increasing importance of sustainability for consumers has prompted companies to address environmental issues within their global supply chains (Ponte, 2020) and communicate them to the market (Kim and Ferguson, 2018; Liu et al., 2020). In this context, social media has become one of the most widely used communication channels for environmental issues because of its interactive and two-way nature, which facilitates dialogue (Chen et al., 2023; Shin and Ki, 2022). Social media plays a crucial role in such communication, as it is based on collective intelligence, collaboration, voluntariness, transparency, and commitment (Castelló-Martínez and Ros-Diego, 2012). Therefore, companies need to identify the best way to communicate sustainability information on social media (Rathee, 2024; Yang et al., 2020).

Considering social media as the primary channel for sustainability communication today, it is essential to consider ‘what” to communicate about sustainability and ‘how” to communicate it (Go and Bortree, 2017). Regarding what to communicate, information specificity refers to which social issues are more relevant to consumers (Agarwal et al., 2024; Pérez et al., 2020). Sustainability messages may be oriented towards product, production process, image, or environmental facts (Carlson et al., 1993). More recently, Agarwal et al. (2024) posited that messages may be oriented to environmental (i.e. energy consumption) or social sustainability (i.e. zero waste). In this regard, the companies emphasizing green content or products receive a high number of likes and replies on social media (Crapa et al., 2024; Shin and Ki, 2022). Additionally, the relationship or congruence between the social initiative and the company’s main activity has been identified as a key factor affecting consumers’ reactions to CSR messages (Ginder and Byun, 2022; Ruiz de Maya et al., 2016). However, less research has been devoted to how to communicate sustainability. Therefore, it is necessary to explore strategic ways to communicate sustainability through social media (Go and Bortree, 2017). In this sense, message design is a crucial element, as it influences consumers’ attitudes towards communication (Pérez et al., 2020; Rathee and Milfeld, 2023). Rathee and Milfeld (2023) defined message design as the advertising message’s structure and content.

Unlike conventional media, social media messages incorporate textual, verbal, and/or visual information and allow users to access extensive information through hyperlinks and hashtags in their message design (Shin and Ki, 2022; Xu and Saxton, 2019; Wang and Yang, 2020). The interactive tools of social media can facilitate companies to listen to and respond to consumers, as well as provide a two-way communication platform for consumers to engage with Go and Bortree (2017), which is particularly relevant for sustainability communication. Therefore, it is important to analyse whether improving the interaction and dialogue with customers will be beneficial for sustainability communication, as the language used on social media is more immediate, direct, and accessible (D’amato et al., 2019). Previous research on communication has defined the concept of dialogical communication as the exchange of ideas and opinions within the context of a relationship (Abitbol and Lee, 2017; Kent and Taylor, 1998). Social media is currently one of the main channels for promoting dialogical communication, owing to its interactive nature (Wang and Yang, 2020; Abitbol and Lee, 2017). Thus, this study aims to provide a comprehensive understanding of the role of dialogical communication in enhancing the effectiveness of sustainability messages on social media. We propose that consumers’ engagement with sustainability messages on social media (i.e. social media engagement, brand engagement, and purchase intention) is higher when the message incorporates dialogical characteristics.

Through two experimental design studies, we demonstrate that sustainability messages generate more social media engagement when they have dialogical features, which in turn leads to increased brand engagement and higher purchase intention. This study significantly contributes to the sustainability communication literature. First, it highlights the importance of fostering dialogue with customers. This study extends beyond the concept of interactivity, as dialogical communication is not only related to the presence of interactivity features but also to the interest in maintaining a dialogue with consumers (Song and Tao, 2022). Second, this study showed that dialogical features are not as important for commercial messages as they are for sustainability messages, providing further insight into the differences between sustainability communication and commercial communication.

In subsequent sections, we review the key concepts, formally propose hypotheses, and present the methodology in detail. We then report the main results and discuss the theoretical and managerial implications.

2. Theoretical framework and hypotheses

2.1 Sustainability messages and social media

Traditionally, sustainability messages have been classified according to their orientation (Agarwal et al., 2024; Carlson et al., 1993; Shin and Ki, 2021). This refers to the primary focus of the message, which can be oriented towards the product (highlighting the environmentally or socially beneficial features of a product) or towards the process (emphasizing the eco-friendly manufacturing and operating practices of the company such as the zero waste). Product and process orientations are identified as substantive messages because they describe concrete environmental or social benefits (Shen et al., 2024). Furthermore, the message may be oriented towards image (when it links the company to an environmental cause) or highlights some environmental facts without a clear connection to the company’s effort. These are considered associative messages because they do not provide clear evidence of the company–sustainability connection. Thus, substantive messages are more objective in nature and include companies’ actual changes in sustainable behaviour, whereas associative messages provide vague, non-concrete and unintelligible information about sustainability (Wang et al., 2023).

Traditionally, companies have been criticized for their lack of transparency in sustainability communication (Higgins et al., 2020); therefore, previous studies have indicated that the orientation of sustainability messages influences how consumers perceive and respond to sustainability communication (Shin and Ki, 2022). Substantive messages are thus better received by consumers and generate favourable attitudes towards the brand (Musgrove et al., 2018; Shen et al., 2024). Consequently, many companies are beginning to shift towards using more substantive-oriented messages than associative ones (Teona et al., 2020; Yu et al., 2013).

Due to the increasing relevance of social media today (Shin and Ki, 2022; Wang, 2021), companies must identify how to communicate sustainability on these media (Yang et al., 2020). Social media serves as one of the most suitable channels for ethical and transparent sustainability communication because it facilitates the disclosure of useful information, the development of multidirectional communication and, ultimately, engagement in dialogue with the public (Chen et al., 2019). The unique features of social media, such as fast responsiveness and connectivity, enhance dialogue with customers (Wang, 2021). Thus, it is important to understand the particularities of sustainability communication on social media in comparison to commercial communication.

2.2 Dialogical communication on social media

Dialogic communication theory, proposed by Kent and Taylor (1998) in Web 1.0, aims to guide organizations in effectively using websites’ dialogic potential to foster positive relationships with citizens. Dialogical communication is an interaction type that seeks to develop mutual respect, understanding, and benefits between organizations and the public (Wirtz and Zimbres, 2018). This communication is based on five dialogical principles: (1) dialogic loop (the strategy to foster interaction/dialogue); (2) usefulness of information (providing useful information); (3) generation of return visits (the attractive features to make users revisit the platform); (4) ease of interface (the intuitive design of the interface); and (5) rule of conservation of visitors (encouraging users to stay on the platform) (Kent and Taylor, 1998).

Traditional communication channels based on a one-way communication approach (i.e. corporate reports or websites) do not allow companies to engage with internal and external stakeholders in a dialogic dimension as social media does (Esposito et al., 2024). With the advent of social media, Abitbol and Lee (2017) identified three dialogical strategies that are most beneficial in cultivating relationships through these platforms: disclosure, a user-friendly message format, and interactivity. The disclosure strategy involves providing useful information appealing to individuals, with transparency and openness of information being key elements. To achieve this goal, the company should ensure that the information disclosed is relevant and prominent. The disclosure of companies’ commitment to sustainable initiatives in social media is acquiring a pivotal role in activating stakeholder engagement (Esposito et al., 2024). The user-friendly message format refers to how information is presented. It should be presented in an accessible format to facilitate interaction with stakeholders. Online, a user-friendly message format includes posting links to external news items about the company or its causes, as well as posting photos, graphics, or videos from the company (Carrera et al., 2008). The use of visuals elements compared to verbal ones in green messages increases their effectiveness as they are more likely to draw consumers’ attention and interest (Shen et al., 2024). As Rathee and Milfeld (2023) posited subtle changes in words and phrases can dramatically impact messages’ effectiveness. Lastly, interactivity refers to the company’s capacity to engage with consumers, thereby fostering participation among interested parties (Abitbol and Lee, 2017). Such a strategy is crucial for promoting involvement and nurturing consumer relationships. Interactivity entails the use of strategies, such as asking questions or calls to action and sharing/commenting to promote interaction and two-way communication between the company and users (Saxton and Waters, 2014). Social media enables two-way communication, accelerating opportunities for interactivity (Esposito et al., 2024). In essence, a dialogical message is characterized by useful information, the use of multimedia format and interactive elements, such as links, hashtags, questions, mentions, or emojis, and using interactive strategies, such as asking questions or calls to action and sharing/commenting to build relationships with users (Abitbol and Lee, 2017; Wang and Yang, 2020).

Developing dialogical messages may yield several benefits to companies seeking to improve their relationships with consumers. Particularly, it ensures the company’s active presence in social media conversations, strengthening its connection with consumers (Zhang et al., 2022) and boosting public engagement in terms of likes, shares, and comments (Men et al., 2018; Wang and Yang, 2020). For instance, the use of rhetorical styles (alliteration and repetition) and cross-message compositions enhance consumer message sharing (Villarroel Ordenes et al., 2019). The use of multiple communication cues, such as hashtags, predicts higher levels of social capital, helping to build a diverse and strategically relevant network (Xu and Saxton, 2019). Specifically, for non-profit organizations, social media communication through questions, comments, or opinions contributes to the development of organization–public relationships (Namisango and Kang, 2019). Finally, according to He et al. (2021), promoting conversation, consumer interaction, and participation in social media content improves consumers’ perceptions of brand personality and brand attitudes. Similarly, Chen et al. (2023) found positive effects on brand trust and the intention to engage in positive word-of-mouth. More recently, Kumar and Hsieh (2024) observe that employing interactive strategies to stimulate participation not only fosters discussions among consumers but also generates a positive brand experience. In summary, the potential effect of developing a sustainability-dialogical message is significant for encouraging two-way communication and relationships with customers, generating social media engagement, and improving brand attitudes.

2.3 Social media engagement

Engagement is closely related to the development of relationships and dialogue (Watkins, 2017). The concept of engagement has been integral to the theoretical foundation of dialogue because “every dialogic interaction involves conversational engagement” (Taylor and Kent, 2014, p. 389). Thus, to comprehend the effects of applying dialogical communication on social media, understanding how individuals engage with the content in this media is paramount.

Engagement on social media can be defined as consumer behaviour in terms of the degree to which consumers interact with the message (e.g. consuming, contributing, and creating) (Muntinga et al., 2011; Schivinski et al., 2016; Vander Shee et al., 2020). For example, Muntinga et al. (2011) classified consumers’ engagement activities with brands on social media from content consumption (e.g. viewing videos and reading reviews) to content contribution (e.g. rating products or companies and engaging in post conversations). Consumption represents a minimum level of engagement in which users passively consume content, while contribution entails a higher level of engagement (Cao et al., 2021).

However, most research in this domain has conceptualized social media engagement as the intention to interact with the message (Sicilia et al., 2020). In other words, they have focused on the outcome of social media engagement in the behavioural aspect (Cao et al., 2021). These interactions not only explain social media engagement but also foster the development of meaningful consumer relationships (Saikia and Bhattacharjee, 2024). Thus, consistent with prior research, this study conceptualizes social media engagement as the consumer’s actions or reactions to brand messages through participation and interaction in the form of liking, sharing, or creating a comment (i.e. García-de-Frutos and Estrella-Ramón, 2021; Lee et al., 2018).

Social media engagement is characterized by high levels of participation and interaction (Hollebeek et al., 2023; Swaminathan et al., 2020). It is a mutually beneficial process in which consumers and companies create brand-related content on social media (Barger et al., 2016). The act of interacting with a brand on social media makes users more engaged, trusted, and satisfied with it (Men and Tsai, 2012). Furthermore, social media engagement is particularly relevant due to the social multiplier effect that a user’s reaction can have on other users. This fact contributes to developing a ‘connective action between like-minded individuals” (Hopke and Paris, 2022, p. 3) and facilitates the dissemination of the product or brand (Bapna and Umyarov, 2015). In this sense, engagement is closely related to the development of relationships and dialogue (Watkins, 2017).

2.4 The effect of applying dialogical communication to sustainability messages on social media engagement

Currently, consumers demand transparent and easily verifiable sustainability information (Ginder and Byun, 2022) to make more conscious and responsible consumption decisions (Byrd and Su, 2020). The use of social media to communicate sustainability enables the development of more ethical, empathetic, honest, and transparent communication, thus providing useful information that encourages engagement in dialogue with the public (Chen et al., 2019; Kent and Taylor, 2016).

More specifically, dialogical messages generate more reactions from users, encourage their participation, and facilitate two-way communication (Abitbol and Lee, 2017) than messages that do not include them. Some studies suggest that using dialogical strategies, such as disclosure of useful content on social media, including a user-friendly message format (i.e. links or multimedia format), and interactivity (i.e. call-to-action, sharing/commenting), contribute to relationship building and facilitate dialogue (Men and Tsai, 2012). Similarly, the use of dialogical strategies that promote interactivity and two-way communication, such as hashtags, questions, mentions (Cheung et al., 2020), or emojis (Valenzuela-Gálvez et al., 2023), helps to increase social media engagement (Hamzah et al., 2021). Thus, dialogical strategies are considered antecedents of social media engagement (Tong and Chan, 2023).

Previous research has demonstrated that sustainability messages generate more social media engagement than commercial messages (Aguirre et al., 2023; Crapa et al., 2024; Shin and Ki, 2022). Based on this rationale, when a sustainability message includes dialogical features, it will generate more social media engagement than when it does not. As a result, the positive effect of sustainability messages on engagement may outweigh that of commercial messages when they include dialogical characteristics. Hence, we formulate the following hypothesis:

H1.

Sustainability messages on social media will generate more engagement than commercial messages when the message is dialogical.

2.5 The mediating role of social media engagement in the relationship between message characteristics and behavioural outcomes

Social media engagement is an indicator of users’ interest in interacting with companies, potentially predicting overall brand engagement. Specifically, it reflects the affective dimension, driven by the subconscious feelings and emotional reactions elicited by exposure to social media messages (Dessart, 2017; Eisend and Tarrahi, 2016). Additionally, social media engagement is a determinant factor that favours purchase intention (Tajvidi et al., 2020).

Previous studies have shown that incorporating dialogical strategies into messages encourages user participation and, consequently, contributes to increasing overall brand engagement (Cheung et al., 2020) and purchase intention (Tajvidi et al., 2020). Likewise, orienting messages towards sustainability with specific and tangible information is crucial for generating overall engagement (Crapa et al., 2024; Ngai and Singh, 2021) and influencing how consumers perceive and respond to sustainability communication. Accordingly, social media engagement will mediate the relationship between sustainability messages, brand engagement, and purchase intention when the messages are dialogical. Therefore, we formulate the following hypothesis:

H2.

The effect of sustainability messages on (a) brand engagement and (b) purchase intention is mediated by social media engagement, and it depends on the dialogical features of the message.

Figure 1 depicts the proposed research model.

3. Study 1

3.1 Methodology

3.1.1 Study design and data collection

A 2 × 2 between-subjects experimental design was developed, in which two factors were manipulated: the dialogical potential of the message (dialogical versus monological) and the sustainability orientation of the message (sustainability versus commercial). Both factors were manipulated through the image and text of the message. Data were collected in June 2022 from participants selected through an online panel of 1,236 people, with a response rate of approximately 54%. Finally, 314 Instagram users, comprising women aged 18–48 years, participated in the study.

To develop the stimuli, different decisions needed to be made (Eshghi et al., 2017). First, a sector and a brand had to be selected for the study. We selected the fast fashion industry because of its reputation as one of the most polluting and socially irresponsible sectors (Bhatia and Devraj, 2017), with H&M being one of the most sustainable fast fashion retailers globally (Kwon and Lee, 2021). The second decision pertained to social media platforms. We selected Instagram based on the following important characteristics: (1) H&M’s substantial Instagram following, (2) the fashion industry’s propensity for using visual-centric platforms to communicate with their customers and (3) it use as a platform to communicate sustainability activities (Kwon and Lee, 2021).

Following Abitbol and Lee (2017), the dialogical potential of the message was manipulated through the image and text of the Instagram publication. The dialogical message incorporated dialogical strategies, such as hashtags, emojis, and links, to the company website and other websites related to sustainable materials. Additionally, the characters that appeared in the picture were mentioned, the brand invited followers to participate in the conversation through a question, and the message encouraged reactions, such as likes. In contrast, the monological message lacks dialogical strategies. The stimuli are presented in Figures 2 and 3.

Furthermore, the message orientation towards sustainability was manipulated through the image and text. The sustainability message featured an image of three characters from behind, dressed in neutral colours. The message also referred to the ecological characteristics of the product (e.g. conscious material) and the sustainable production process (e.g. solvent-free). Conversely, the commercial message featured an image of three characters from the front, dressed in eye-catching prints. The stimuli are presented in Figures 2 and 3. To establish the four conditions of the experiment, we combined these images and texts.

To check whether the message was perceived as dialogical or monological and as a sustainable or commercial message, we conducted a pre-test involving 110 participants. Participants were asked to indicate how dialogical they perceived the Instagram publication and to what extent the publication contained sustainable content. The dialogical potential of the message was measured using two items with a 7-point Likert scale adapted from Go and Bortree (2017): (1) the company invites to converse and (2) the company is open to dialogue. To determine whether the message was perceived as a sustainability message or commercial message, one item with a 7-point Likert scale was applied: ‘The Instagram publication provides information about the use of environmentally friendly production processes and materials’. Finally, to check if the image was perceived as sustainable or commercial, an item with a 7-point Likert scale was used: ‘Seeing this picture gives me the feeling that the garments are sustainable and environmentally friendly’.

The pre-test results for the dialogical potential of the message confirmed our expectations (Mdialogical = 5.27, Mmonological = 2.96; F (1) = 63.41, p = 0.00). We observed significant differences between the sustainability and commercial content of the message (Msustainability = 5.00, Mcommercial = 2.05; F (1) = 28.76, p = 0.00). Additionally, the results showed differences between the sustainable/environmental perception of the image (Msustainability = 5.00, Mcommercial = 3.45; F (1) = 10.03, p = 0.00). Finally, we employed a scale adapted from Verhagen et al. (2012) to test whether both images were perceived as equally attractive. The results indicated no significant differences between the sustainability and commercial images (Msustainability = 4.49, Mcommercial = 4.27; F (1) = 1.57, p = 0.20).

3.1.2 Procedure and measurement

Regarding the procedure, participants were first asked whether they had an Instagram account and whether they followed fashion brand accounts. If so, the following instruction was provided: “Imagine you have just logged into your Instagram account and you see the following post from H&M”. Participants were then randomly assigned to one of the four experimental conditions, showing them the corresponding stimulus. Scenario 1 showed an Instagram publication featuring a sustainability message and dialogical message, Scenario 2 showed an Instagram publication with a sustainability message and monological message, Scenario 3 showed an Instagram publication with a commercial message and dialogical message, and Scenario 4 showed an Instagram publication with a commercial message and monological message. Next, participants proceeded to complete the rest of the questionnaire.

Several scales were adapted from the literature to measure the constructs used in this research (see Table 1), and all of them used 7-point scales. Brand engagement was assessed by focusing on the affective dimension, using the adapted scale from Hollebeek et al. (2014), while social media engagement was measured using a scale adapted from Schivinski et al. (2016). We evaluated purchase intention based on Onofrei et al. (2022). Additionally, CSR scepticism was included to control for potential effects (Skarmeas and Leonidou, 2013). Measures to control for manipulation checks were also included, as in the pre-test conducted.

3.2 Results

3.2.1 Manipulation checks

The mean age of participants was 30.09 years, ranging from 18 to 45 years old. Participants displayed a positive attitude towards the brand, scoring 4.97 on a 7-point scale.

Manipulation checks revealed that the dialogical message was perceived as more interactive/inciting to dialogue than the monological one (Mdialogical = 4.84, Mmonological = 4.07; F (1) = 44,97, p = 0.00). Additionally, individuals believed that the sustainability message provides information on environmentally friendly production processes for their products (Msustainability = 5.53, Mcommercial = 4.02; F (1) = 82.66, p = 0.00). Regarding the image of the Instagram publication, the sustainability message was perceived as showing more sustainable/environmentally friendly garments than the commercial message (Msustainability = 4.94, Mcommercial = 4.19; F (1) = 19.06, p = 0.00). Finally, most participants remembered the scenario to which they had been assigned, both in terms of the dialogical potential of the message (95.5% for dialogical and 94.9% for monological) and its content (97.4% for sustainability and 93.1% for commercial). These retrieval rates are acceptable, as previous research reported similar results (Sicilia et al., 2020).

3.2.2 Common method variance

In quantitative studies, there are frequent methodological problems when self-reported measures are used in a single survey, such as the common method variance (CMV). To avoid this problem, procedural and statistical techniques can be applied (Malhotra et al., 2017). For this, procedural techniques were used, such as (1) a pre-test of the questionnaire, (2) informing participants of the confidentiality of their answers, (3) informing participants that there are no right/wrong answers, and (4) encouraging participants to answer honestly (Malhotra et al., 2017; Podsakoff et al., 2003). In terms of statistical techniques, all items were modelled as indicators of a single factor representing the common method effect (Malhotra et al., 2006). The results of the confirmatory factor analysis (CFA) showed a low fit (χ2[135] = 2122.559, p = 0.000, NNFI = 0.545, IFI = 0.600, CFI = 0.599, RMSEA = 0.217), confirming that CMV is not a problem. We also conducted an exploratory factor analysis (EFA) extracting two factors, and the variance explained by the most significant factor was 48.3%. The result showed that no single factor dominated the total variance, as there is no single general factor that accounts for more than 50% of the covariance among the measures; hence, common method is not an issue (Podsakoff et al., 2003).

3.2.3 Scale validation

The psychometric properties of the scales were measured via CFA using the R package lavaan (Rosseel, 2012). The fit of the measurement model was acceptable [X2 (120gl) = 232.761, p = 0.000, X2/df = 1.939, SRMR = 0.043, CFI = 0.970, NNFI = 0.962, GFI = 0.894, NFI = 0.940] (Iacobucci, 2010).

Firstly, we evaluated the reliability of each indicator. The minimum criterion is that the load size is equal to or greater than 0.70. As shown in Table 1, in addition to being significant, most of the loadings exceeded 0.70. Social media engagement and sustainability message items with weaker loadings (slightly below 0.7) were retained for their contribution to content validity (Hair et al., 2009).

Secondly, we analysed the internal consistency of the constructs. The criterion used for this assessment was that both composite reliability (Jöreskog, 1971) and Cronbach’s alpha exceeded 0.70. Cronbach’s alpha, which also assesses internal consistency, maintains a threshold of approximately 0.70. Thirdly, we evaluated convergent validity using the average variance extracted (AVE). As shown in Table 1, the results equal or exceed 0.5, which is the minimum suggested by Fornell and Larcker (1981). These findings support the reliability of the multi-item measures used in this study. Lastly, we analysed the discriminant validity using two criteria. First, AVE was greater than the square of the correlation between the construct pairs. Second, the confidence interval was calculated at ±2 SE around the correlation between the factors; none of the confidence intervals included 1 (Hair et al., 1994). As shown in Table 2, the measurement instrument did not exhibit any discriminant validity issues.

3.2.4 Results of the hypotheses testing

In H1, we predicted that the effect of sustainability messages on social media engagement is moderated by the dialogical characteristics of the message. Furthermore, H2 proposed that the effect of sustainability messages on brand engagement and purchase intention is mediated by social media engagement and is contingent on the dialogical characteristics of the message. We used Model 7 of Hayes’s (2022) PROCESS Macro (version 4.2) in SPSS to test for moderation and moderated mediation. A 95% bias-corrected bootstrapped confidence interval (N = 10,000) was estimated (Preacher et al., 2007).

Rather than using the dichotomous variable based on the scenarios, we incorporated dialogical and sustainability messages into the model using manipulation. This involves introducing a metric variable that measures the degree to which users perceive a publication as dialogic and/or sustainable. This approach has been used in previous studies (Sicilia et al., 2020) because metric variables provide richer information in regression analysis than nominal variables based solely on scenarios (Hair et al., 2009).

The result revealed that the main effects of sustainability and dialogical messages on social media engagement are insignificant. Social media engagement has a positive effect on brand engagement (B = 0.54, p < 0.00) and purchase intention (B = 0.30, p < 0.00). Interestingly, we observed a significant and positive interaction effect between sustainability messages and dialogical messages (B = 0.06, p < 0.04) on social media engagement, confirming H1 (see Table 3). To better understand the interaction effect of the dialogical message on the sustainability message, we calculated the conditional effects at values of the moderator (m ± 1 SD) by following Preacher et al. (2007) recommendations. The results are presented in Table 4. The analysis revealed that the sustainability message generated greater social media engagement when the dialogical level of the message increased (Blow = 0.21, SE = 0 0.07, 95% CI [0.06, 0.35]; Bmedium = 0.29, SE = 0.07, 95% CI [0.15, 0.43]; Bhigh = 0 0.37, SE = 0.09, 95% CI [0.20, 0.54]). In other words, when consumers perceive a message to be more dialogical, the sustainability message becomes more effective. CSR scepticism was introduced as a covariable.

We tested the moderated mediation using the index proposed by Hayes (2022), which provides an inferential test to determine whether the indirect effect depends on the moderator. Moderated mediation exists when this index is not zero (Hayes, 2022). The moderated mediation index yielded significant results for both brand engagement (index = 0.035, SE = 0.015, 95% CI [0.006, 0.065]) and purchase intention (index = 0.019, SE = 0.009, 95% CI [0.003, 0.038]), thus confirming H2 (see Table 4). The results showed that the effect of sustainability messages on brand engagement and purchase intention through social media engagement was moderated by the dialogical level of the message (see Table 5). The indirect effect of sustainability messages on brand engagement and purchase intention through social media engagement was more pronounced at higher moderator levels.

4. Study 2

4.1 Methodology

4.1.1 Study design and data collection

To reinforce the external validity of the results, we conducted an additional study involving another product category suitable for consumption by both men and women. We used a fictitious new bar and restaurant called “Food Corner”. The novelty of the service ensured that our participants held no prior attitude towards the company (Sicilia et al., 2020). Bars and restaurants, known for their adoption of sustainability policies, have been employed in previous studies in this domain (i.e. Aguirre et al., 2023).

Similar to Study 1, a 2 × 2 between-subjects experimental design was developed in which two factors were manipulated: the dialogical potential and sustainability orientation of the message. Data were collected in February 2024, and the participants were selected through the same online panel. Finally, 302 Instagram users, comprising 50% women and 50% men aged 18–55 years, participated in the study.

Consistent with Study 1, the dialogic message incorporated dialogic strategies, such as hashtags, emojis, links to the company’s website, mentions, questions, and invitations to participate in the conversation and react with likes. To manipulate the sustainability aspect of the message, we highlighted the sustainable characteristics of the food (e.g. sustainable local products) and the elaboration process (e.g. reduces food waste). We also used a restaurant image with a natural decoration style. The stimuli are presented in Figures 4 and 5.

A pre-test was conducted with 42 participants to assess the efficacy of the manipulation. The results for the dialogical potential of the message confirmed our expectations (Mdialogical = 4.54, Mmonological = 2.27; F (1) = 25.47, p = 0.00). To test whether the message was perceived as a sustainability message or a commercial one, one item with a 7-point Likert scale was applied: “The Instagram publication provides information on making sustainable cuisine”. The results showed significant differences between the two messages (Msustainability = 5.32, Mcommercial = 2.40; F (1) = 21.29, p = 0.00). Finally, we checked whether the images employed were perceived as equally attractive (Msustainability = 3.50, Mcommercial = 4.10; F (1) = 1.01, p = 0.32).

4.1.2 Procedure and measurement

First, individuals were asked whether they had an Instagram account and if they used to go out to eat at bars and restaurants. If so, the following instruction was provided: “Imagine you have just logged into your Instagram account and you see the following publication of a new BAR/RESTAURANT that is going to open in your city called FOOD CORNER”. Next, participants proceeded to complete the rest of the questionnaire. Regarding the measurement of the constructs, we employed the same scales as in Study 1, with the subsequent adaptation to the new product category (see Table 6). We also included the attitude towards Food Corner (Kim et al., 1998; Batra and Stayman, 1990). Additionally, to measure the perceived sustainability of the message, participants were asked to write down everything they remembered after exposure to the Instagram publication (Sicilia et al., 2005). This process is called thought elicitation. For this task, they were provided with a space for writing.

4.2 Results

4.2.1 Manipulation checks

Manipulation checks revealed that the dialogical message was perceived as more interactive and inciting to dialogue than the monological one (Mdialogical = 4.83, Mmonological = 4.08; F (1) = 34,09, p = 0.00). Furthermore, the total number of sustainability thoughts indicates that individuals perceived the sustainability message as providing more information on sustainable cuisine than the commercial one (Msustainability = 1.33, Mcommercial = 0.06; F (1) = 8190.66, p = 0.00).

4.2.2 Common method variance

We analysed the CMV following the same procedure as in Study 1. The results of the CFA indicated a low fit (χ2[189] = 1288.571, p = 0.000, NNFI = 0.659, IFI = 0.694, CFI = 0.693, RMSEA = 0.189), In the EFA extracting two factors, the variance explained by the most significant factor was below 50%. Therefore, common method is not an issue (Podsakoff et al., 2003).

4.2.3 Scale validation

The psychometric properties of the scales were measured via CFA. The fit of the measurement model was acceptable [X2 (168gl) = 342.490, p = 0.000, X2/df = 2.03, SRMR = 0.049, CFI = 0.959, NNFI = 0.948, GFI = 0.853, NFI = 0.925] (Iacobucci, 2010). As Tables 6 and 7 show, the results support the composite reliability, convergent validity, and discriminant validity of the measures used in the study.

4.2.4 Results of the hypotheses testing

To corroborate the hypotheses, we employed Model 7 of Hayes’s (2022) PROCESS Macro (version 4.2) in SPSS. A 95% bias-corrected bootstrapped confidence interval (N = 10,000) was estimated (Preacher et al., 2007). As in Study 1, the dialogical message and the sustainability message were included in the model using manipulation rather than the dichotomous variable based on the scenarios.

The result showed that the main effects of sustainability and dialogical messages on social media engagement are insignificant. Social media engagement has a positive effect on brand engagement (B = 0.36, p < 0.00) and purchase intention (B = 0.27, p < 0.00). We observed a significant and positive interaction effect between sustainability message and dialogical message (B = 0.06, p < 0.03) on social media engagement. Thus, H1 is supported in Study 2 (see Table 8). As in Study 1, the analysis revealed that the sustainability message generates greater social media engagement when the dialogical level of the message increased (Bmedium = 0.14, SE = 0.06, 95% CI [0.01, 0.26]; Bhigh = 0 0.22, SE = 0.07, 95% CI [0.07, 0.37]) (see Table 9). In other words, when consumers perceive the message as more dialogical, the sustainability message becomes more effective. In fact, the positive effect of the sustainable message was insignificant for low levels of dialogical communication. The attitude towards Food Corner was introduced as a covariable. CSR scepticism was excluded because there were no significant differences between conditions (M = 4.65, p > 0.481).

The moderated mediation index was significant for both brand engagement (index = 0.025, SE = 0.010, 95% CI [0.006, 0.044]) and purchase intention (index = 0.019, SE = 0.008, 95% CI [0.005, 0.034]), supporting H2 as in Study 1 (see Table 8). The results showed that the effect of the sustainability message on brand engagement and purchase intention through social media engagement is moderated by the dialogical level of the message (see Table 10). The indirect effect of the sustainability message on brand engagement and purchase intention through social media engagement is more pronounced at higher moderator levels. We observed no significant indirect effect when the message is perceived as having low dialogical attributes.

5. Discussion

Given the growing importance of sustainability for consumers and the fact that social media is one of the main communication channels today, companies face challenges in finding an effective way to communicate sustainability on social media (Yang et al., 2020). In this sense, it is essential to analyse whether improving interaction and dialogue with customers will be beneficial for sustainability communication (D’amato et al., 2019). Therefore, this study aims to provide a comprehensive understanding of the role of dialogical communication in enhancing the effectiveness of sustainability messages on social media.

Through two experimental studies, this research demonstrates the benefits of including dialogical features in sustainability messages in terms of social media engagement, brand engagement, and purchase intention. Previous studies have shown that consumers highly value sustainability messages that include concrete environmental benefits (Ginder and Byun, 2022; Teona et al., 2020). Moreover, the inclusion of dialogical features, such as links, hashtags, mentions, or calls to action, generates greater social media engagement (i.e. liking, sharing, and commenting on publications) (Abitbol and Lee, 2017). Based on this knowledge, this study establishes that the inclusion of dialogical characteristics favours consumers’ engagement with sustainability messages. Both experimental studies revealed that when the message is perceived as more dialogic, the sustainability message becomes more effective.

The results showed that when the message is dialogical, the effect of the sustainability message on brand engagement and purchase intention is mediated by social media engagement. In other words, fostering brand engagement and purchase intention involves the generation of social media engagement. Positive consumer reactions on social media are crucial to achieving overall brand success (Dessart, 2017).

In summary, it has been corroborated that social media is an effective means of communicating about sustainability (Ginder and Byun, 2022), especially when implementing dialogical features in the message design.

5.1 Theoretical contributions

Firstly, this study contributes to dialogical theory (Kent and Taylor, 1998) by highlighting the importance of including dialogical features in sustainability communication on social media. The development of sustainable messages with dialogical characteristics encourages interactions, facilitates two-way communication, and fosters dialogue with users, thus affecting users’ behavioural intentions. Therefore, this study expands the previous knowledge about the importance of the dialogic principles to the social media domain (e.g. Abitbol and Lee, 2017; Wang and Yang, 2020). Our findings are in line with the recent results obtained by Shen et al. (2024), who found the benefits of design messages with visual aspects to increase the effectiveness of green messages.

Secondly, this study enriches the interactive marketing field and confirms that using two-way interactive communication with customers (Wang, 2021), represented by the inclusion of dialogical features in a message, results in increased engagement with the message. Thus, we extent the previous knowledge regarding the higher engagement of sustainability messages on social media compared to commercial ones (e.g. Aguirre et al., 2023; Crapa et al., 2024), considering that the inclusion of dialogical features will be a key element in the design of the message to increase its effectiveness in terms of likes, comments or sharing.

Thirdly, it contributes to the literature on sustainability communication by highlighting the importance of fostering dialogue with consumers. Analysing this communication on social media is particularly interesting, as the language used on social media is more immediate, direct, and accessible to broader audiences compared to traditional channels (D’amato et al., 2019). Thus, this study extends beyond the concept of interactivity and the inclusion of visual appeals in the message design (i.e. use of emoticons/emojis) (Matthes et al., 2014), showing the relevance of including dialogical features to sustainability communication compared to commercial messages.

5.2 Managerial implications

A significant challenge organizations face is how to communicate sustainability on social media. This study offers valuable insights into applying dialogical principles to sustainability communication. Marketers should design dialogical messages to initiate dialogue with customers and enhance engagement. Therefore, it is recommended that brands incorporate into their messages –both text and image– the dialogical strategies identified and applied in this study. These strategies include using hashtags (#), mention of the photo characters using @, emojis, or links to the company’s website and other sustainability-related websites, and calls to action. Additionally, when designing the message, it should invite followers to participate in the conversation through a question (e.g. which is your favourite?) and encourage them to react with likes.

5.3 Limitations and future research

Although this study fills an important research gap and provides relevant implications for sustainability communication management, it has some limitations that may drive future research. First, it was conducted in a specific research context: fashion brands and bar/restaurant. Future research would benefit from broadening its scope to encompass other brands and sectors, such as grocery products or the hospitality industry. Moreover, it is recommended to expand the study to other social media platforms because of their different characteristics (e.g. Twitter, TikTok, etc.). The second limitation is the use of the scenario method in a simulated environment where individuals cannot interact with the brand. To overcome this limitation, future research could design a scenario in which participants can interact with the brand or other followers in real-time, or a field study may be conducted. Third, other factors may influence consumers’ intention to dialogue with the company about sustainability, such as their concern about the planet’s sustainability or other psychological variables, such as self-disclosure.

Figures

Figure 1

Research model

Figure 2

Example of images for the sustainability-dialogical message and the commercial-dialogical message

Figure 3

Example of text for the sustainability/dialogical message, the sustainability/monological message, the commercial/dialogical message, and the commercial/monological message

Figure 4

Example of images for the sustainability-dialogical message and the commercial-dialogical message

Figure 5

Example of text for the sustainability/dialogical message, the sustainability/monological message, the commercial/dialogical message, and the commercial/monological message

Table 1

Internal consistency reliability and convergent validity

Construct	Item	Loading		SD	CA	CR	AVE
Purchase intention	Give me the chance, I would consider shopping at H&M in the future	0.948	***	0.012	0.962	0.962	0.895
	I am likely to shop at H&M in the near future	0.926	***	0.014
	Given the opportunity, I intend to purchase from H&M	0.963	***	0.014
Brand engagement	I felt very positive when I saw the H&M post	0.900	***	0.018	0.936	0.936	0.787
	Seeing this H&M post has made me happy	0.891	***	0.016
	I felt good to see this post from H&M	0.909	***	0.018
	I felt proud to see this post from H&M	0.848	***	0.025
Social media engagement	I would like to “Like” this H&M post	0.854	***	0.027	0.823	0.801	0.575
	Comment on this H&M post	0.663	***	0.039
	I would share this H&M post	0.747	***	0.030
Dialogical message	The company invites to converse	0.794	***	0.045	0.804	0.805	0.674
	The company is open to dialogue	0.846	***	0.045
Sustainability message	The Instagram publication provides information about the use of environmentally friendly production processes and materials	0.587	***	0.052	0.677	0.7031	0.551
	Seeing this picture gives me the feeling that the garments are sustainable and environmentally friendly	0.871	***	0.048
CSR scepticisms	It is doubtless/doubtful that H&M is a socially responsible company	0.916	***	0.015	0.940	0.940	0.780
	It is certain/uncertain that H&M is concerned about improving the welfare of society	0.882	***	0.022
	It is sure/unsure that H&M follows high ethical standards	0.907	***	0.017
	It is unquestionable/questionable that H&M acts in a socially responsible way	0.866	***	0.031

Note(s): SD: standard deviation; CA=Alfa de Cronbach; CR=Composite reliability; AVE = Average variance extracted, ***p < 0.001

Source(s): Authors’ own work

Table 2

Discriminant validity

	Purchase intention	Brand engagement	Social media engagement	Dialogical message	Sustainability message	CSR scepticims
Purchase intention	0.946	0.416	0.304	0.229	0.204	0.252
Brand engagement	0.571; 0.719	0.887	0.750	0.239	0.374	0.356
Social media engagement	0.458; 0.646	0.808; 0.924	0.758	0.274	0.409	0.375
Dialogical message	0.357; 0.601	0.377; 0.601	0.406; 0.642	0.821	0.366	0.305
Sustainability message	0.340; 0.564	0.510; 0.714	0.538; 0.742	0.479; 0.731	0.742	0.540
CSR scepticism	0.400; 0.604	0.513; 0.681	0.517; 0.709	0.445; 0.661	0.631; 0.839	0.883

Note(s): Diagonal elements are the square root of average variance extracted (AVE) between the constructs and their measures. Above the diagonal are squared correlation between constructs. Below the diagonal are confidence intervals

Source(s): Authors’ own work

Table 3

Moderated mediation analyses for social media engagement brand engagement and purchase intention

	M (social media engagement)				Y (brand engagement)				Y (purchase intention)
	Coef	SE	t	p	Coef	SE	t	(p)	Coef	SE	t	(p)
Sustainability message (SM)	0.004	0.150	0.028	0.977	0.033	0.056	0.595	0.552	0.095	0.069	1.386	0.166
Dialogical message (DM)	−0.068	0.159	−0.431	0.666
SM × DM	0.065	0.032	1.978	0.048
M (social media engagement)					0.548	0.045	11.969	0.000	0.305	0.056	5.449	0.000
CSR scepticism	0.389	0.067	5.769	0.000	0.255	0.056	4.514	0.000	0.282	0.069	4.050	0.000

Source(s): Authors’ own work

Table 4

Conditional effects at values of the moderator

Conditional effects of X (sustainability message) on M (social media engagement) at values of the moderator (dialogical message)
	Effect	BootSE	LLCI	ULCI
Low dialogical level	0.211	0.074	0.063	0.358
Medium dialogical level	0.293	0.069	0.155	0.430
High dialogical level	0.375	0.087	0.203	0.546

Source(s): Authors’ own work

Table 5

Conditional indirect effect through social media engagement

	Effect	BootSE	LLCI	ULCI
Conditional INDIRECT effect of X (sustainability message) on Y (brand engagement at values of the moderator (dialogical message)
Low dialogical level	0.115	0.042	0.035	0.202
Medium dialogical level	0.160	0.043	0.078	0.249
High dialogical level	0.205	0.052	0.106	0.311
Conditional INDIRECT effect of X (sustainability message) on Y (purchase intention) values of the moderator (dialogical message)
Low dialogical level	0.064	0.025	0.018	0.116
Medium dialogical level	0.089	0.027	0.038	0.147
High dialogical level	0.114	0.034	0.052	0.185

Source(s): Authors’ own work

Table 6

Internal consistency reliability and convergent validity

Construct	Item	Loading		SD	CA	CR	AVE
Purchase intention	Give me the chance, I would consider going to Food Corner in the future	0.901	***	0.018	0.945	0.946	0.854
	I am likely to go to Food Corner in the near future	0.935	***	0.015
	Given the opportunity, I intend to go to Food Corner	0.936	***	0.013
Brand engagement	I felt very positive when I saw this food Corner publication	0.830	***	0.024	0.932	0.935	0.784
	Seeing this Food Corner publication has made me happy	0.928	***	0.012
	I felt good to see this publication from Food Corner	0.931	***	0.013
	I felt proud to see this publication from food Corner	0.853	***	0.020
Social media engagement	I would like to “Like” this Food Corner publication	0.775	***	0.031	0.867	0.876	0.702
	Comment on this Food Corner publication	0.875	***	0.021
	I would share this Food Corner publication	0.864	***	0.020
Dialogical message	The company invites to converse	0.880	***	0.024	0.882	0.883	0.792
	The company is open to dialogue	0.902	***	0.022
Sustainability message	The Instagram publication provides information on making sustainable cuisine	0.724	***	0.043	0.794	0.801	0.669
	Seeing this photo gives me the feeling that it is a sustainable and environmentally friendly bar/restaurant	0.912	***	0.026
Brand attitude	Negative/positive	0.890	***	0.029	0.913	0.914	0.781
	Unfavorable/favorable	0.855	***	0.042
	I don´t like it/I like it	0.906	***	0.019
CSR scepticisms	It is doubtless/doubtful that Food Corner is a socially responsible company	0.846	***	0.025	0.931	0.931	0.771
	It is certain/uncertain that Food Corner is concerned about improving the welfare of society	0.883	***	0.020
	It is sure/unsure that Food Corner follows high ethical standards	0.886	***	0.021
	It is unquestionable/questionable that Food Corner acts in a socially responsible way	0.898	***	0.016

Note(s): SD: standard deviation; CA=Alfa de Cronbach; CR=Composite reliability; AVE = Average variance extracted, ***p < 0.001

Source(s): Authors’ own work

Table 7

Discriminant validity

	Purchase intention	Brand engagement	Social media engagement	Dialogical message	Sustainability message	Brand attitude	CSR scepticism
Purchase intention	0.924	0.632	0.476	0.385	0.352	0.576	0.331
Brand engagement	0.741; 0.849	0.885	0.611	0.403	0.423	0.555	0.370
Social media engagement	0.614; 0.766	0.714; 0.850	0.838	0.442	0.381	0.519	0.312
Dialogical message	0.521; 0.721	0.545; 0.725	0.589; 0.741	0.890	0.451	0.434	0.502
Sustainability message	0.487; 0.703	0.551; 0.751	0.528; 0.708	0.564; 0.781	0.818	0.547	0.565
Brand attitude	0.675; 0.843	0.663; 0.827	0.661; 0.781	0.567; 0.751	0.646; 0.834	0.884	0.457
CSR scepticism	0.462; 0.690	0.485; 0.733	0.433; 0.675	0.591; 0.827	0.646; 0.858	0.566; 0.786	0.878

Note(s): Diagonal elements are the square root of average variance extracted (AVE) between the constructs and their measures. Above the diagonal are squared correlation between constructs. Below the diagonal are confidence intervals

Source(s): Authors’ own work

Table 8

Moderated mediation analyses for social media engagement brand engagement and purchase intention

	M (social media engagement)				Y (brand engagement)				Y (purchase intention)
	Coef	SE	t	p	Coef	SE	t	(p)	Coef	SE	t	(p)
Sustainability message (SM)	−0.167	0.152	−1.103	0.271	0.127	0.045	2.807	0.005	0.073	0.052	1.383	0.167
Dialogical message (DM)	0.081	0.161	0.500	0.617
SM × DM	0.068	0.031	2.171	0.030
M (social media engagement)					0.366	0.039	9.265	0.000	0.271	0.046	5.865	0.000
Brand attitude	0.574	0.070	8.153	0.000	0.331	0.054	6.053	0.000	0.504	0.064	7.882	0.000

Source(s): Authors’ own work

Table 9

Conditional effects at values of the moderator

Conditional effects of X (sustainability message) on M (social media engagement) at values of the moderator (dialogical message)
	Effect	BootSE	LLCI	ULCI
Low dialogical level	0.057	0.073	−0.087	0.201
Medium dialogical level	0.140	0.653	0.012	0.269
High dialogical level	0.224	0.078	0.071	0.379

Source(s): Authors’ own work

Table 10

Conditional indirect effect through social media engagement

	Effect	BootSE	LLCI	ULCI
Conditional INDIRECT effect of X (sustainability message) on Y (brand engagement at values of the moderator (dialogical message)
Low dialogical level	0.020	0.026	−0.034	0.069
Medium dialogical level	0.051	0.024	0.001	0.097
High dialogical level	0.082	0.027	0.025	0.135
Conditional INDIRECT effect of X (sustainability message) on Y (purchase intention) values of the moderator (dialogical message)
Low dialogical level	0.015	0.019	−0.024	0.051
Medium dialogical level	0.038	0.018	0.002	0.074
High dialogical level	0.061	0.021	0.019	0.104

Source(s): Authors’ own work