Relational Pathways to Inclusion: Transforming Service Systems Through Macro-Social Marketing

Systems Thinking and Inclusive Service Systems

Systems thinking is a paradigmatic approach that views phenomena not as isolated entities but as interconnected elements within a larger whole (Jackson 2009, 2016, 2019; Kim and Senge 1994). It emphasizes the dynamic relationships, feedback loops, and emergent behaviors that arise from the interactions between system components (Fowler 2003). Rather than focusing on linear cause–effect patterns, systems thinking seeks to understand how structures and interdependencies shape systemic outcomes over time (Maani and Maharaj 2004). It offers a holistic lens through which complexity, adaptation, and co-evolution can be interpreted (Grewatsch, Kennedy, and Bansal 2023). Although systems thinking itself constitutes a broad and diverse body of literature (for instance Jackson 2009; Mingers and White 2010), encompassing multiple philosophical traditions such as cybernetics, complexity theory, and general systems theory, engaging with its full ontological and epistemological debates lies beyond the scope of this study.

Individuals Experiencing Vulnerabilities

Building on the foundational definitions of service systems, we now turn to a brief discussion of what is meant by individuals experiencing vulnerabilities as one subsystem. This study adopts the term “individuals experiencing vulnerabilities” to emphasize that vulnerability arises from dynamic and contextual conditions rather than being an intrinsic or static characteristic (Baker, Stephens, and Hill 2002; Russell-Bennett, Bedggood, and Raciti 2024). In adopting this terminology, we depart from deficit-based labels such as “vulnerable customers,” which risk perpetuating stigmatization and diminishing recognition of individual capabilities (Baker, Gentry, Rittenburg 2005). ² Grounded in a strengths-based approach (Raciti, Russell-Bennett, Letheren 2022), this perspective highlights the structural and situational forces that engender exclusion, while simultaneously affirming the agency, resilience, and potential of those affected.

In this study, we acknowledge that individuals experiencing vulnerabilities encompass a wide range of groups, including those experiencing socioeconomic disadvantage, cultural marginalization, and diverse physical or cognitive conditions (see Hill and Sharma 2020). While recognizing the breadth and heterogeneity of these classifications, we specifically focus on individuals with physical vulnerabilities. Within this domain, our attention centers on the Deaf ³ community, whose experiences offer a particularly salient case for examining exclusion within service systems (see, for instance, Abney et al. 2017). This focus ⁴ is guided by the sustained engagement of the author(s) with Deaf communities and the rich experiential insights developed over time, allowing for a more nuanced, contextually grounded, and relational understanding of their lived experiences.

While the primary focus of this study is on individuals experiencing vulnerabilities, we also consider the role of widely represented individuals. Although this may initially seem tangential to the discussion of vulnerabilities, it aligns with our systems thinking perspective (Jackson 2009, 2016, 2019), which emphasizes the interdependence of subsystems within the overall service systems. Individuals experiencing vulnerabilities do not exist in isolation; rather, they are embedded within societal structures where interactions with widely represented individuals are inevitable and influential (Anderson et al. 2016; El Jurdi, Baktir, and Price 2025).

Viewing both groups as distinct yet interrelated subsystems (Supplemental Appendix 4) underscores the necessity of examining not only the experiences of marginalized individuals but also the perceptions, actions, and responsibilities of those who are more widely represented (Abney et al. 2017). From a systems perspective, fostering inclusive service systems requires an understanding of how these subsystems interact, influence one another, and collectively shape conditions of inclusion (Parkinson et al. 2017). Therefore, this study incorporates the perspectives of both individuals experiencing vulnerabilities and widely represented individuals, reflecting the systemic interrelations that must be navigated to advance meaningful service inclusion.

Macro-Social Marketing, Systems Thinking, and Inclusive Service Systems

Moving from service systems, we employ MSM as a framework for transforming structural and institutional conditions (Adkins and Jae 2010; Kennedy and Parsons 2012). MSM emerges from the convergence of macromarketing, which examines the societal impacts of marketing systems (Layton 2007, 2019; Layton and Grossbart 2006), and social marketing (Flaherty et al. 2020; Kemper and Ballantine 2020; Kennedy and Santos 2019; Kennedy and Smith 2022), which seeks to influence individual behaviors for collective benefit. Rather than focusing solely on changing individual behaviors, MSM addresses how systemic structures, institutional norms, and collective practices produce and sustain exclusionary outcomes (Domegan et al. 2020; Kennedy 2016). It frames marketing as a strategic tool for achieving holistic, system-level transformation. Recent studies (for instance, Jagadale and Kemper 2022; Shaw, Harris, and Ali 2024) highlight how MSM facilitates system-wide change by fostering strategic partnerships and reshaping institutional structures, reinforcing its role in advancing social well-being through durable, system-level transformations. We build on this foundation by highlighting how everyday communicative and relational practices at the micro and meso levels contribute to these broader transformations, showing that systemic change unfolds not only through institutional design but also through the accumulation of situated interactions.

Systems thinking provides a complementary perspective by making explicit the interconnections between micro- and macro-level dynamics. While MSM emphasizes the institutional and collective arrangements that foster inclusion, systems thinking highlights how exclusion emerges from interdependent subsystems and feedback loops (Machin et al. 2019). Bringing these perspectives together makes it possible to analyze how performative actions—subjective, voluntary behaviors such as advocacy, resistance, or inclusive practices—interact with institutional actions, involving codified rules, policies, and governance structures that formalize systemic change (Kennedy 2016). These dual mechanisms influence cultural-moral norms (shared values, beliefs, and ethical expectations) and economic-task norms (practices related to resource allocation, efficiency, and operational roles), both of which shape how service systems include or exclude different actors.

In line with prior TSR, which has emphasized access, well-being outcomes, and institutional arrangements for marginalized groups such as refugees or those in human services (Finsterwalder et al. 2020; Randle and Zainuddin 2020), this study adds an additional layer by focusing on the communicative and relational dynamics that operate within and across subsystems. By doing so, it complements existing work and underscores that pathways to inclusion involve not only structural resources but also the quality of interactional practices that shape everyday service engagement.

Individuals experiencing vulnerabilities and widely represented individuals are seen as interacting subsystems whose engagements co-create systemic conditions of inclusion or exclusion. Flaherty et al. (2020) systematically map how systems thinking and MSM complement one another by addressing multi-level complexity and non-linear causality in social change processes. Additionally, Wymer (2021) highlights the importance of building stakeholder coalitions across levels, reinforcing the view that systemic transformation requires coordinated actions among interdependent actors within service systems.

Building from the convergence of MSM and systems thinking, inclusive service systems can be defined as service environments intentionally structured to dismantle systemic barriers and promote equitable participation for both individuals experiencing vulnerabilities and widely represented individuals (Fisk et al. 2018). In this way, the study extends TSR by deepening attention to communicative mechanisms of well-being and complements MSM by showing how institutional and policy interventions are reinforced through micro- and meso-level adjustments over time. These systems deploy performative and institutional mechanisms to shift social norms, operational practices, and governance structures, embedding inclusion as a core design principle rather than a secondary consideration. Huff and Barnhart (2022) emphasize that addressing complex problems in service systems requires recognizing the entanglement of market, social, and institutional dynamics—an insight critical to building genuinely inclusive systems. By engaging both subjective experiences and objective systemic structures, inclusive service systems create conditions where diversity, empowerment, and collaboration are foundational to service interactions.

Toward Inclusive Service Systems Transformation

With the theoretical foundations established, we now conceptualize the process through which service systems progressively adapt toward greater inclusivity (Supplemental Appendix 5). Building inclusive service systems for individuals experiencing vulnerabilities requires more than isolated intervention points; it demands a fundamental systemic reorientation of values, practices, and institutional structures. Rather than treating inclusivity as an additive layer, this approach embeds fairness, equity, and accessibility as foundational properties of the service systems (Alkire et al. 2022). These systems must not only accommodate diversity but actively evolve through continuous recalibration to address emerging and often unforeseen forms of exclusion (Fehrer, Kemper, and Baker 2024).

Here, we build on existing TSR scholarship that has highlighted the importance of access and well-being outcomes for marginalized groups (Finsterwalder et al. 2020; Randle and Zainuddin 2020) by drawing attention to the communicative and relational dimensions through which these outcomes are enacted. Complementing this, MSM emphasizes structural and institutional interventions (Shaw, Harris, and Ali 2024), and systems thinking allows us to view these processes as interconnected and adaptive, emphasizing feedback loops and emergent change (Jackson 2009; Maani and Maharaj 2004). Together, these perspectives underline that inclusivity is not a static state but a process of iterative recalibration within complex systems.

Shaw, Harris, and Ali (2024) and Flaherty et al. (2020) emphasize that system-level change is not linear but adaptive, with inclusive service systems developing through feedback loops, incremental shifts, and iterative stakeholder re-engagements. Integrating systems thinking into this framing reinforces the view that exclusion emerges from interdependencies rather than isolated barriers, and that transformation requires continuous monitoring of feedback and dynamic coordination among actors. Accordingly, fostering inclusivity demands sustained, reflexive processes that build upon both performative and institutional actions over time, rather than relying on discrete, one-time interventions (Wymer 2021).

Hence, we conceptualize this progress through four interrelated modalities, representing dynamic zones of transformation within service systems. The term modality is used to emphasize that the processes of transformation within service systems are not fixed, linear, or sequential. Instead, modalities capture dynamic, overlapping, and recurrent patterns of systemic change, allowing for multiple points of entry, feedback, and reconfiguration rather than prescribing a rigid developmental pathway. These modalities do not follow a fixed or sequential pathway; rather, they emerge and recur in iterative cycles, deepening systemic engagement with inclusivity through continuous recalibration (Sönnichsen et al. 2024). By grounding these modalities in systems thinking, we highlight how adaptive, recursive feedback cycles drive change; by aligning them with TSR, we stress that such change is oriented toward well-being and equity; and by linking to MSM, we illustrate how cross-sectoral and institutional interventions can amplify localized practices into durable transformation. Each modality spans one or more levels of the service system—micro (individual), meso (organizational/community), and macro (institutional/policy)—and together, they delineate how systems adapt and evolve. Together, they delineate how service systems progressively move from surface-level awareness of exclusion toward the institutionalization of inclusive norms, with each cycle enhancing resilience against new exclusionary structures.

The first modality, Awareness of Exclusion, marks the emergence of initial recognition of systemic inequities. At the micro level, actors begin identifying marginalized experiences and questioning entrenched assumptions, often through symbolic or performative actions that initiate ethical shifts (Boenigk et al. 2021a). Although early-stage responses may lack deep structural impact, they are vital for opening discursive spaces and mobilizing broader system sensitivity to exclusionary dynamics (Sönnichsen et al. 2024).

Interrogation of Norms introduces more critical and reflective engagement. Actors turn their attention to the underlying institutional routines, cultural logics, and systemic structures that normalize exclusion. This occurs predominantly at the meso level, where organizational practices and community norms come under scrutiny. Critical interrogation of these norms enhances the system’s capacity to detect and challenge institutionalized barriers, particularly those that disadvantage individuals experiencing vulnerabilities (Dean and Indrianti 2020; Tanouri, Kennedy, Veer 2022). This modality is crucial for shifting from superficial recognition to deeper systemic awareness.

Participatory Realignment represents a pivotal shift from analysis to redesign. It involves collaborative reconfiguration efforts, where those historically excluded—especially individuals experiencing vulnerabilities—actively co-create inclusive structures, practices, and governance mechanisms. Drawing on the transformative ethos of MSM and participatory approaches (Rahman, Finsterwalder, and Kuppelwieser 2024), this modality sees performative and institutional actions converge, creating durable systemic shifts through stakeholder collaboration across diverse sectors (Boenigk et al. 2021a). This modality operates across meso and macro levels, engaging organizations, public institutions, and multi-actor coalitions in the structural co-design of inclusive service delivery.

Finally, Institutionalization of Inclusivity signals the embedding of inclusive values into formalized policies, operational routines, and collective service norms. Inclusivity becomes codified and reinforced through systemic feedback mechanisms that maintain alignment across cultural, organizational, and governance domains (Fehrer, Kemper, and Baker 2024; Russell-Bennett, et al. 2024). At the macro level, this institutionalization shapes policy frameworks and standard-setting mechanisms. However, this institutionalization is not a static endpoint; as demonstrated by Sönnichsen et al. (2024) and Huff and Barnhart (2022), it is inherently adaptive, requiring the system to remain alert to emergent exclusions and adapt accordingly.

Research Approach

This study adopts an abductive approach (Timmermans and Tavory 2012; Vila-Henninger et al. 2024) to investigate the research questions, commencing with a preliminary conceptual model—outlined in the Theoretical Background—centered on four modalities of inclusive service systems transformation (Supplemental Appendix 6). This initial model operates as a heuristic guide, refined iteratively through sustained engagement with empirical data sourced from individuals experiencing vulnerabilities. To achieve this, we implement a multi-method case study design (Yin 2018) that synthesizes multiple empirical layers to inform theory development (Jaakkola 2020; Ulaga et al. 2021).

The Deaf community is adopted as a salient case to illustrate systemic dynamics of inclusion and exclusion, offering a compelling lens into how individuals experiencing vulnerabilities engage with service systems. By multi-method, we refer to the integration of diverse data collection and analysis techniques conducted within a unified epistemological and methodological orientation: a qualitative, constructivist, and exploratory paradigm. Rather than seeking to confirm pre-specified hypotheses, the objective is to construct a contextually grounded conceptual framework, emergent from the interpretive synthesis of lived experiences, digital discourse, and situated practices.

Anchored in this approach, our empirical strategy draws upon four key sources: (1) interviews with members of the Deaf community; (2) netnographic analysis of Deaf-oriented practices and discourses in digital environments; (3) interviews with widely represented (hearing) individuals who have engaged with Deaf individuals; and (4) rapid ethnography conducted across four events that featured interactions between Deaf and hearing individuals ⁵ (Supplemental Appendix 7). This ensemble of data sources affords both breadth of perspective and depth of experience—two complementary dimensions critical to capturing the complexity of inclusive service phenomena (Supplemental Appendix 8). Breadth is attained through interviews with a range of actors and the expansive coverage enabled by netnographic inquiry, which encompasses digital narratives from diverse contributors within the online Deaf community. Depth, in contrast, is generated through immersive rapid ethnographic engagement and in-depth interviewing, both of which uncover nuanced, situated experiences that are not always accessible via mediated discourse.

Research Context

The study is situated in Indonesia, where the Deaf community—estimated at approximately two million individuals (Kementerian Pendidikan Dasar dan Menengah 2023)—faces exclusion not because of deafness itself but due to societal and institutional structures that insufficiently recognize their linguistic and cultural identity. Despite this sizable demographic, public awareness of Deaf culture remains limited, illustrated by incidents such as officials proposing whistles as communication aids (Christine and Kristian 2021), reflecting persistent deficit-based misconceptions that overlook the community’s cultural and linguistic richness.

Within this context, Bahasa Isyarat Indonesia (BISINDO) functions as the community’s preferred sign language, grounded in lived experience (Palfreyman 2020), whereas institutional promotion of SIBI (Sistem Isyarat Bahasa Indonesia), a constructed system, often feels misaligned with Deaf communicative norms (Wirajuda 2022). Such institutional preferences, combined with impairment-focused views, reinforce communication barriers and contribute to ongoing stigma, demonstrating how formal structures can exacerbate exclusion even when framed as supportive.

At the same time, the Deaf community in Indonesia has built strong grassroots movements through advocacy, social media engagement, and Deaf-centered digital content, alongside emerging technological initiatives that support visual communication. These developments highlight the community’s resilience and signal growing momentum toward more inclusive service systems—ones that address structural inequities while affirming Deaf linguistic and cultural identity as foundational to equitable service design.

Data Source 1: Engagement with Deaf Individuals

The author(s) engaged directly with the Deaf community in Indonesia, grounding their interaction in both linguistic preparation and cultural sensitivity. Drawing on functional fluency in Bahasa Indonesia and foundational exposure to BISINDO, they undertook a certified 10-week BISINDO training program in 2024 led by a Deaf instructor. This experience not only provided essential signing skills but also fostered a deeper appreciation for the communicative norms, cultural practices, and lived experiences of the Deaf community. Through this deliberate preparation, the author(s) sought to build respect and understanding—positioning themselves to collaborate with, rather than merely study, the community.

Following the completion of formal BISINDO training, the author(s) conducted in-depth interviews directly with 12 Deaf individuals in Indonesia (Supplemental Appendix 9), without relying on a sign language interpreter. This direct engagement enabled the collection of firsthand insights into how Deaf participants navigate familiar service environments, particularly in relation to the study’s conceptual model of inclusive systems transformation. By conducting the interviews in BISINDO—the participants’ preferred mode of communication—the author(s) ensured dialogic authenticity and mutual intelligibility, even as minor asymmetries in fluency occasionally emerged.

Participants were initially recruited through informal interactions at a local coffee shop where Deaf baristas work, providing a comfortable and accessible environment for initiating conversations. After obtaining consent to participate, the author(s) employed a snowball sampling technique to reach other potential participants. This engagement process occurred over approximately 10 months spanning from August 2024 to May 2025, spanning five informal meetings, allowing for the gradual and respectful building of trust within the Deaf community.

Each interview was guided by a semi-structured protocol (see Supplemental Appendix 10), lasting 20 to 30 minutes to balance conversational flexibility with thematic focus. These interactions revealed nuanced, context-rich accounts of participants’ service experiences, highlighting both systemic constraints and their own adaptive responses. Complemented by field notes capturing contextual observations and researcher reflections, this bottom-up approach centers lived realities and communicative agency over prescriptive, top-down service design solutions.

Although the absence of a professional sign language interpreter represents a methodological limitation, the author(s) deliberately chose this path to preserve direct engagement and cultivate trust. Informants demonstrated flexibility in communication, navigating the conversations through a combination of visual and verbal cues. Their willingness and ability to adapt not only sustained the flow of dialogue but also reinforced the importance of recognizing communicative diversity as a strength rather than a deficit (AccessForce 2023).

The analytical process began with descriptive coding to surface salient features within the interview data (see Supplemental Appendix 11). This initial phase employed a set of predefined categories—such as “social media awareness,” “public event,” and “community event”—that were conceptually aligned with the study’s preliminary framework. Subsequently, thematic analysis (Braun and Clarke 2012) was applied to synthesize these descriptive codes into broader, interpretive themes including “initial awareness,” “institutional initiative,” and “community movement.” In conducting the coding process, we adopted a multi-step approach that integrated both theory-driven and data-driven strategies.

Descriptive coding was initially guided by our conceptual framework, with codes derived from predefined constructs related to service accessibility, public perception, and institutional response (Saldaña 2021). In parallel, in vivo coding was employed to capture participants’ own language and expressions, preserving the emic meanings embedded in their accounts. This dual approach enabled the identification of both anticipated and emergent dimensions of meaning. The coded data were then thematically organized through a process of iterative refinement and constant comparison, allowing for the generation of higher-order themes that reflected patterns across the interviews and aligned with broader socio-institutional dynamics.

Data Source 2: Netnographic Method

This study employed a netnographic approach (Kozinets, Dolbec, and Earley 2024) using YouTube to capture self-authored narratives of Deaf individuals, enabling analysis of naturally occurring discourse while maintaining analytical distance and transparency in interpreting socially situated experiences. The sampling process employed a purposive strategy, guided by three inclusion parameters: (1) the content must foreground personal or communal experiences of Deaf individuals; (2) the format must reflect conversational genres such as interviews, talk shows, or news reports; and (3) the language of delivery was limited to English or Bahasa Indonesia, regardless of the availability of closed captions, due to the author(s)’ linguistic competencies and the practical constraints of accurately interpreting content beyond these two languages.

A total of 44 videos published between 2013 and 2024 were selected using specific keywords ⁶ representing diverse national contexts including Indonesia, the United States, the United Kingdom, Austria, Australia, Russia, the Netherlands, and Japan (Supplemental Appendix 12). These selections captured a heterogeneous spectrum of themes ranging from sign language advocacy and educational barriers to familial dynamics and Deaf entrepreneurship. Of the 39 captioned videos included in the netnographic analysis, additional non-captioned materials were excluded from formal coding but used contextually to enrich interpretation without compromising rigor.

Data were acquired using Python scripts in Jupyter Notebooks, with the YouTube Data API retrieving structured metadata and comment threads and the Transcript API providing closed-caption transcripts in English or, when unavailable, Bahasa Indonesia (Supplemental Appendix 13). All transcripts were stored as text files, organized into a unified DataFrame with language metadata, and supplemented with documented fallback procedures to ensure transparency. The dataset comprised 10,218 comments from 8,902 users (9,979 unique entries, totaling over 1 million characters; Supplemental Appendix 14), with videos lacking comments retained for analysis through their metadata and transcripts. This process yielded two complementary corpora—audience discourse and speaker-authored captions—subsequently preprocessed to ensure textual consistency, analytical tractability, and alignment for qualitative and computational analysis.

The comment dataset was extracted via the YouTube Data API, yielding 10,218 comments from 8,902 users (9,979 unique entries; 1,047,570 characters). Closed-caption transcripts were retrieved using the YouTube Transcript API, organized into a DataFrame by video ID and language, and cleaned through a custom function that removed timestamps and formatting artifacts. Non-English transcripts were translated into English using a Google Translate API pipeline with preprocessing and a throttled retry mechanism. This multistage process—de-duplication, cleaning, annotation, and translation—produced a harmonized dataset suitable for thematic extraction, linguistic modeling, and cross-cultural interpretation. For transcripts, we analyzed a pre-processed corpus of 39 translated videos, normalized the text, and vectorized it into a 5,000-term document-term matrix before applying a semi-supervised CorEx model guided by four anchors (“awareness,” “engagement,” “collaboration,” “institutionalization”).

For comments, 10,218 entries (including four previously excluded videos) were cleaned, language-detected, and stemmed, yielding 1,867 English comments aligned with the same anchors. These were analyzed using a semi-supervised CorEx model with CountVectorizer (bigrams; 5,000-term cap), where some stemmed anchors (e.g., “exposur,” “institute”) triggered sparsity warnings. Applying a 0.05 relevance threshold retained 752 comments (40.3%) and excluded 1,115 as diffuse. Retained comments averaged 369 characters (median 264; range 2,360), with a right-skewed distribution—25% below 160 characters and the upper quartile above 469—and higher-liked entries (mean 11.67; max 713) tended to contain richer reflections on awareness, critique, and gratitude.

Principal component analysis (PCA) was applied to a document-term matrix of 1,000 high-frequency lexical features after stopword removal and zero-variance filtering, with 241 components explaining 90% of variance and 352 retained to preserve finer lexical structure. MANOVA tested topic separability using these 352 components as dependent variables and semi-supervised topic labels as the independent factor, while PERMANOVA—using Euclidean distance matrices derived from term-frequency vectors—relaxed parametric assumptions and ensured robustness for sparse, high-dimensional text. To assess semantic distinctiveness, Jensen-Shannon Divergence (JSD) was computed between normalized topic-level word distributions, with significance evaluated through 1,000-label permutation tests and Benjamini–Hochberg corrections, confirming statistical separability and coherence across topics. Temporal analyses standardized timestamps into year–month bins, constructed time-indexed contingency tables, and applied chi-squared tests before modeling topic series through ARIMA, with stationarity verified using the Augmented Dickey–Fuller test and parameters optimized via stepwise AIC, enabling 12-month forecasts that illuminate retrospective and projected patterns in discursive.

Data Source 3: Engagement with Widely Represented Individuals

We focused on widely represented individuals who had direct engagement with Deaf individuals—recruited through snowball sampling—to capture how their perceptions, misunderstandings, and systemic responses illuminate subsystem dynamics and inform the design of more inclusive service systems. A total of 28 informants participated between August 2024 and May 2025, comprising eight customers, eight business owners, seven industry professionals, and five government officials (Supplemental Appendix 9). Data collection involved semi-structured interviews of 20–30 minutes, supported by field notes, with informed consent obtained from all participants. Following the same approach as Data Source 1, descriptive coding and thematic analysis were conducted, integrating predefined framework categories with emergent in vivo codes to generate higher-order themes (Healy and Perry 2000; Miles, Huberman, and Saldaña 2020).

Data Source 4: Rapid Ethnography with Deaf and Hearing Individuals

We incorporated rapid ethnography to complement interviews and digital analysis, grounding the study in real-time, situated practices. Rapid ethnography condenses ethnographic immersion while retaining its interpretive depth (Handwerker 2001; Millen 2000). Researchers in other marginalized contexts have used it to surface hidden dynamics (e.g., Galea et al. 2020; Janda, Shainesh, and Hillebrand 2021; Vo 2020), and we applied it here to examine systemic exclusion and resilience within Deaf–hearing interactions. Rather than replacing other methods, we used rapid ethnography to extend them, generating experiential insights into how Deaf and hearing individuals negotiate inclusion in everyday settings. While interviews and online discourse revealed retrospective accounts, our fieldwork captured embodied and co-present interactions, giving us a closer view of relational, communicative, and spatial dynamics in practice.

We carried out fieldwork across four sites between May 2024 and May 2025: (1) a weekly Jakarta community event over 4 weeks, where we observed facilitated cross-modal participation; (2) an international sign language celebration, where hearing and Deaf participants demonstrated linguistic inclusivity in advocacy and celebratory spaces; (3) an inclusive performing arts workshop, where Deaf and hearing audiences redefined the boundaries of sound-based art; and (4) six visits to a Deaf-staffed café, where we engaged in informal dialogue with baristas and observed how visual communication, assistive tools, and customer familiarity shaped service encounters. These engagements provided grassroots insights into barriers, adaptive strategies, and resilience (Supplemental Appendix 16).

We analyzed fieldnotes with a systems orientation, focusing on relational dynamics and interdependencies. We constructed causal loop diagrams (CLDs) from observations of communicative patterns, interaction flows, and environmental enablers, iterating sketches to capture emergent system behaviors (Supplemental Appendix 17). By applying CLDs, we visualized nonlinear feedback and systemic change, aligning with earlier ethnographic applications in health and social research (Burrell et al. 2021; Carr et al. 2022). In doing so, we reinforced the methodological and theoretical contribution of ethnographic insights to our multi-method design.

Validity, Reliability, and Data Saturation

We ensured validity and reliability following Healy and Perry’s (2000) criteria, with thematic convergence across interviews, netnography, and ethnography providing consistency and contextual depth (Miles, Huberman, and Saldaña 2020) and reinforcing the credibility of the findings (Supplemental Appendix 11). Data saturation was evaluated across all sources rather than relying solely on Deaf interviews, and ethical integrity was maintained through informed consent, institutional approval, and adherence to established guidelines to ensure confidentiality, respect, and non-maleficence (Dodds et al. 2023).

The Four Modalities of Inclusive Service Systems Transformation

Drawing from the empirical findings, we refined the conceptual modalities initially proposed into four empirically grounded modalities that more accurately reflect the relational, communicative, and adaptive processes observed in situated practice. These refined modalities are: Surface-Level Engagement, In-Depth Exploration, Co-Creative Participation, and Sustained Involvement (Supplemental Appendix 18). While the original conceptual terms captured the progression toward system-level transformation in abstract terms, the revised modalities offer a more practice-oriented vocabulary (Supplemental Appendix 19). The rearticulation allows us to better illustrate not only the structural dimensions of inclusion but also its ongoing negotiation in everyday service encounters.

Conversational Dynamics: Netnographic Support of the Four Modalities

The MANOVA results suggest that the four modalities ⁷ —Surface-Level Engagement, In-Depth Exploration, Co-Creative Participation, and Sustained Involvement—are statistically distinct in their linguistic composition. MANOVA on 352 PCA components revealed a significant topic effect (Wilks’ Lambda = .093, F = 1.355, p < .001), supported by Pillai’s trace, Hotelling–Lawley, and Roy’s root (Supplemental Appendix 20). Topic 1 contained general interaction terms such as “understand” and “hear,” whereas Topic 4 used system-level language such as “institutionalization” and “embedded,” confirming that each modality forms a coherent semantic cluster aligned with differing levels of system engagement. PERMANOVA further confirmed discursive separability across all pairwise comparisons (p = .001), with the strongest contrasts between Topic 1 and Topic 2 (pseudo-F = 7.60) and between Topic 1 and Topic 3 (pseudo-F = 5.87), and the closest overlap between Topics 4 and 3 (pseudo-F = 1.54), reflecting shared structural emphases (Supplemental Appendix 21).

JSD results reinforced these findings, showing high divergence (JSD > 0.57) and significant differences (FDR-adjusted p < .05) for five topic pairs—including (4,3), (2,4), and (2,3)—with the greatest contrast between Topic 4 (Sustained Involvement) and Topic 3 (Co-Creative Participation) (Supplemental Appendix 22). Topic visualizations, including bar charts of the top 20 terms (Supplemental Appendix 23) and proportion plots (Supplemental Appendix 24), highlight clear thematic structures and reveal the dominance of certain modalities within the broader discourse.

Temporal analyses indicate that Surface-Level Engagement (Topic 1) consistently dominates over time (Supplemental Appendix 25), while Topics 3 and 4 remain marginal. A chi-square test (χ² = 379.899, p = .884) shows no significant temporal shift; regressions reveal slight, non-significant increases for Topics 2 and 3 and near-zero change for Topic 4 (Supplemental Appendices 26–27). ARIMA modeling (Supplemental Appendix 28) shows Topic 1 exhibits strong temporal dependence (ARIMA(1,0,1), AR = –0.861, p < .001; intercept = 0.984, p < .001), Topic 2 follows an ARIMA(2,0,0) pattern (intercept = 0.313, p < .001), and Topics 3 and 4 fit intercept-only models (Topic 3 = 0.108; Topic 4 = 0.058; both p < .001). Collectively, these results confirm that the modalities are semantically and temporally distinct, with Surface-Level Engagement persistently dominating public discourse and more advanced modalities appearing only sporadically.

Theoretical Implications

This study extends systems thinking, TSR, and MSM by conceptualizing inclusion as an iterative, adaptive process rather than a linear trajectory. TSR positions service systems as platforms for well-being (Anderson et al. 2013; Anderson and Ostrom 2015; Blocker, Davis, and Anderson 2022), while MSM emphasizes systemic and institutional interventions (Flaherty et al. 2020; Layton 2019). Our four modalities show how service systems move between symbolic gestures, relational adjustments, collaborative redesign, and adaptive institutionalization across micro, meso, and macro levels, advancing systems thinking in service research by foregrounding feedback loops and interdependencies (Chandler and Lusch 2015; Jackson 2009).

A first key contribution concerns the paradox of symbolic recognition. Awareness campaigns and accessibility symbols can entrench exclusion when they remain at Surface-Level Engagement and are not supported by deeper communicative and relational change, thereby reproducing inequities under an “inclusive” veneer (Fisk et al. 2018; Russell-Bennett, Bedggood, and Raciti 2024). Likewise, Sustained Involvement reveals that institutionalization is fragile: inclusive policies and structures require continuous monitoring and recalibration to remain responsive to emergent exclusions, aligning with systems thinking’s emphasis on non-linearity and adaptation (Domegan et al. 2020; Kennedy 2016; Maani and Maharaj 2004; Shaw, Harris, and Ali 2024).

A second contribution refines the understanding of co-creation in TSR. While co-creation is often treated as inherently empowering (Polonsky et al. 2024; Rahman, Finsterwalder, and Kuppelwieser 2024), our findings show that marginalized actors may still bear disproportionate communicative and emotional burdens, even in ostensibly collaborative arrangements, complementing work that focuses on resource access for communities experiencing vulnerabilities (Finsterwalder et al. 2020; Randle and Zainuddin 2020). By integrating MSM and systems thinking, we further extend MSM beyond macro-level campaigns to show how micro-level relational practices—such as everyday sign language use and communicative adjustments—accumulate into system-level transformation (Flaherty et al. 2020; Kennedy and Parsons 2012; Wymer 2021). The four modalities thus offer a practice-oriented vocabulary that links TSR’s well-being focus, MSM’s institutional strategies, and systems thinking’s adaptive insights into a unified account of inclusive service systems transformation.

Managerial Implications

Managers can advance inclusion by shifting from symbolic gestures toward practices that mirror the relational and adaptive dynamics identified in the findings and theoretical implications. This requires embedding communicative adaptation—such as basic sign language skills, visual interaction techniques, and structured training on Deaf communication norms—into everyday operations rather than treating them as one-off initiatives. Co-design mechanisms, including advisory groups or participatory workshops with Deaf stakeholders, help redistribute the burden of adaptation and reflect the collaborative ethos central to Co-Creative Participation. These structures convert micro-level relational adjustments into meso-level organizational routines, ensuring that inclusion is practiced rather than merely declared.

To maintain alignment with Sustained Involvement, managers should institutionalize inclusion through reflexive systems that support continuous recalibration. Integrating inclusion metrics into organizational evaluation, establishing ongoing feedback loops with marginalized stakeholders, and partnering with community organizations reinforce adaptive governance and guard against regression into surface-level efforts. Such mechanisms reflect the non-linear, iterative nature of system transformation emphasized in the theoretical framing and ensure that inclusive practices remain resilient as service environments evolve.

Limitations and Future Research

This study’s focus on the Deaf community within predominantly a single national context offers depth but limits the transferability of findings to other groups and settings, while the reliance on English and Bahasa Indonesia YouTube content narrows linguistic diversity. The rapid ethnographic design provided situated insights but constrained prolonged immersion, and the semi-supervised modeling captured discursive patterns without accounting for multimodal communication central to Deaf experiences. Future research should therefore examine how the four modalities manifest across different forms of vulnerability, employ multilingual and cross-context digital ethnographies, and adopt longitudinal or multi-site ethnographic designs to trace how inclusion evolves, stabilizes, or regresses over time.