Comparing TOPSIS and VIKOR to prioritize international communication interventions across digital platforms

Abstract

Digital platforms have become primary arenas for international and intercultural communication. Yet practitioners still lack a transparent way to prioritize competing interventions when goals such as reach, authenticity, trust, and safety conflict. This study develops and tests a multi-criteria decision framework that compares the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) and the VlseKriterijumska Optimizacija I Kompromisno Resenje (VIKOR). These two methods were selected because they represent complementary decision logics widely used in multi-criteria decision analysis: TOPSIS ranks alternatives according to their geometric distance from an ideal solution, while VIKOR identifies compromise solutions that minimize the maximum regret among competing criteria. Employing both methods enables the study to capture different trade-off perspectives that arise in international communication planning, thereby improving the transparency and interpretability of intervention prioritization. Using China-anchored outbound communication towards ASEAN, EU, and MENA audiences as illustrative contexts, we construct a criteria system that integrates validated intercultural outcomes (e.g., openness, empathy, interaction management, perceived authenticity) with operational and platform-related factors (e.g., algorithmic visibility, translation and subtitling quality, accessibility, moderation and civility, and cost). A multidisciplinary expert panel (N = 100) provides importance weights and performance assessments for a portfolio of interventions, including localization depth, creator partnerships, cross-platform sequencing, and community moderation design. From these data, we build cleaned, normalized decision matrices and apply TOPSIS (closeness to an ideal solution) and VIKOR (compromise under conflicting criteria), followed by unified robustness and sensitivity analyses. Across contexts, both methods converge on a similar top tier of interventions: strategies that combine deep localization, sustained partnerships with credible creators, and proactive community moderation consistently outperform cost or speed-optimized options. Method divergences are confined to mid-ranked alternatives and align with theoretical differences between ideal-distance and regret-minimizing logics. The findings demonstrate that a dual TOPSIS–VIKOR lens can provide defensible, practice-ready rankings and diagnostics for international communication planning, and the study offers a replicable workflow that other institutions can adapt to their own platforms, audiences, and strategic priorities.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-45798-2.

Introduction

Digital platforms now mediate much of the world’s intercultural contact^1–3. However, practitioners still lack a rigorous way to determine which interventions should be prioritized when goals such as reach, trust, authenticity, and sustainability collide^4–6. Multi-criteria decision analysis (MCDA) offers a principled approach that integrates heterogeneous evidence and expert judgment to produce defensible rankings^7–9. Among the wide range of MCDA techniques, distance-based and compromise-based approaches are particularly suitable for complex policy problems characterized by conflicting criteria. TOPSIS and VIKOR were selected in this study because they are among the most widely applied and methodologically complementary MCDA frameworks. Compared with hierarchical weighting approaches such as the Analytic Hierarchy Process (AHP), which focuses primarily on pairwise comparisons of criteria, both TOPSIS and VIKOR directly evaluate the relative performance of alternatives across multiple criteria in a normalized decision matrix. This makes them well-suited to empirical contexts where alternatives must be compared simultaneously across operational, social, and communication outcomes. Moreover, previous comparative studies recommend using these two methods together because they highlight different decision logics: TOPSIS rewards alternatives that perform consistently well across all criteria, while VIKOR emphasizes compromise solutions that minimize the most severe trade-offs. The combined use of these methods therefore provides a more transparent diagnosis of decision robustness than relying on a single MCDA technique. Within MCDA, TOPSIS selects options closest to an ideal solution and farthest from an anti-ideal^10–12, while VIKOR emphasizes compromise under conflict and stakeholder regret; using them together enables both “closeness-to-ideal” and “compromise” perspectives on difficult trade-offs in international communication^13–15. Foundational work by Hwang and Yoon systematized TOPSIS¹⁶, and Opricovic and Tzeng⁸ formalized VIKOR’s compromise logic; subsequent reviews demonstrate the breadth of these methods and recommend their comparative use for robustness¹⁷. This dual lens is well-suited to communication choices that must balance linguistic accessibility, narrative authenticity, algorithmic visibility, and resource constraints^16,18,19.

Recent research has increasingly explored fuzzy and hybrid multi-criteria decision-making (MCDM) frameworks to support complex decision environments characterized by uncertainty and conflicting criteria. For example, studies on hybrid energy systems demonstrate how fuzzy-logic control mechanisms can improve decision stability and system performance in nonlinear environments, highlighting the ability of fuzzy methods to manage uncertainty and optimize operational outcomes in complex systems²⁰. Similarly, recent sustainability research has adopted hybrid fuzzy-ranking approaches to construct evaluation frameworks; the development of an ESG evaluation model for Taiwan’s science parks integrates the Fuzzy Delphi Method and clustering techniques to identify and prioritize sustainability indicators, illustrating the usefulness of fuzzy-based decision frameworks for structured evaluation problems²¹. Beyond engineering and sustainability governance, broader interdisciplinary work, such as Sharma (2025) on regenerative organic agriculture, further demonstrates how structured evaluation frameworks can guide decision-making in complex socio-environmental systems²².

Research on public diplomacy and social media demonstrates the need for a multi-criteria approach, yet it is rarely implemented end-to-end^23–25. Sevin’s comparative analysis links digital campaigns to audience perceptions but falls short of providing a formal prioritization model for candidate interventions^25,26. Meanwhile, emerging studies treat platform choice itself as an MCDA problem, demonstrating feasibility for extensive criteria sets and multiple platforms, and information-quality work uses fuzzy TOPSIS to evaluate Rumor-refutation capacity, an issue central to intercultural trust online²⁷. These strands validate the potential of MCDA for communication strategy analysis but reveal a key gap: existing studies rarely provide a structured framework for prioritizing concrete international communication interventions amid multiple, often conflicting objectives. Intercultural communication strategies on digital platforms require organizations to balance reach, credibility, localization depth, governance requirements, and resource constraints, yet systematic prioritization tools for these trade-offs remain limited. As a result, practitioners often rely on ad hoc judgment or single-criterion evaluations when deciding which communication interventions to implement across regions^28–30.

The intercultural communication literature has also matured, making an MCDA study more actionable³¹. Recent validated instruments, such as the Intercultural Effectiveness Scale and new intercultural communicative competence scales, offer measurable constructs, such as openness, empathy, and interaction management, that can be translated into decision criteria and expert questionnaires^13,32,33. Incorporating validated outcomes alongside operational criteria such as translation quality or comment civility enables a decision model that is both theoretically grounded and practically applicable. Yet these instruments are seldom embedded in comparative MCDA designs for communication planning, creating a missed opportunity to join measurement science with strategic selection^34–36.

Few studies compare TOPSIS and VIKOR on the same portfolio of international communication interventions to reveal when compromise-based rankings diverge from ideal-distance rankings and why that matters for decision-makers. Second, existing applications often rely on single-domain experts; intercultural interventions demand multidisciplinary panels spanning communication, sociolinguistics, platform governance, culture studies, and policy. Third, there is room for China-anchored scholarship that tests globally relevant cases, for example, outbound campaigns from Chinese cultural or educational institutions targeting ASEAN, MENA, and EU audiences, while convening a diverse international expert group to complete structured questionnaires. By building a criteria system that integrates validated intercultural outcomes with platform and operational factors, and by eliciting weights from a cross-field expert panel, this study will conduct a dual evaluation using TOPSIS and VIKOR and perform a sensitivity analysis to assess stability. The result will be practice-ready rankings and diagnostics indicating when decision-makers should privilege compromise over ideal proximity and how that choice affects policy and strategy in real intercultural contexts²⁹.

Organizations engaged in international digital communication increasingly face complex strategic choices regarding how to allocate limited resources across multiple digital interventions, platforms, and regional contexts. The practical challenge in digital international communication is not whether tools such as localization, creator collaboration, or moderation exist, but how to configure them into scalable implementation models under resource, governance, and legitimacy constraints. Decision-makers must routinely balance competing objectives: maximizing impact and acceptance while containing deployment cost and time-to-benefit, and simultaneously meeting platform governance and safety requirements across culturally distinct regions. This study contributes at the application level by translating these communication-theoretic levers into a structured, replicable prioritization framework: we specify a criteria system grounded in international communication practice, define implementable alternative models (A1–A5) as configurations of those levers, and evaluate them through expert-informed MCDA across outbound contexts. In doing so, the study offers a decision-ready template that can support policy and operational planning for digital public diplomacy and international communication campaigns. A key advantage of this framework is the integration of two complementary MCDA methods, TOPSIS and VIKOR, which provide dual analytical perspectives on the same decision problem. While TOPSIS evaluates alternatives based on their geometric proximity to an ideal solution, VIKOR identifies compromise solutions that minimize the maximum regret associated with conflicting criteria. The comparison of these two approaches enhances transparency by revealing whether rankings are driven by balanced overall performance or by avoiding extreme weaknesses.

To address this gap, this study develops a structured and practitioner-ready decision framework for prioritizing international communication interventions across digital platforms by applying and comparing two complementary MCDA methods, TOPSIS and VIKOR, within a unified evaluation design. Concretely, we (1) build a criteria system that fuses validated intercultural outcomes (e.g., openness, empathy, interaction management) with operational platform factors (e.g., algorithmic visibility, translation and subtitling quality, moderation and civility, cost), (2) assemble a multidisciplinary expert panel with Chinese authorship and international reviewers spanning communication studies, sociolinguistics, platform governance, cultural policy, and creator practice, (3) evaluate realistic, China-anchored case studies targeting audiences in ASEAN, EU, and MENA, and (4) conduct sensitivity and stability analyses to show when compromise-seeking rankings (VIKOR) diverge from closeness-to-ideal rankings (TOPSIS) and what that implies for policy and strategy. The study’s novelty lies in its dual-method head-to-head design within the intercultural communication domain, its explicit integration of validated ICC constructs into the decision criteria, its cross-field expert elicitation that reflects real-world decision settings, and its delivery of transparent diagnostics that translate directly into platform-specific, content-level, and community-management recommendations (see Fig. 1). By integrating validated intercultural communication constructs with operational platform factors and applying a dual-method MCDA comparison, the study contributes a transparent prioritization framework that helps organizations systematically evaluate alternative digital communication strategies rather than relying on informal or experience-based decision processes.

Fig. 1.

Workflow for prioritizing international communication interventions with TOPSIS and VIKOR, case selection (ASEAN, EU, MENA), criteria integrating intercultural outcomes and operational factors, multidisciplinary expert elicitation (China-led global panel), data preparation, dual analyses (TOPSIS closeness to ideal; VIKOR compromise), robustness checks, and resulting insights and recommendations.

Methodology

This study adopts a comparative multi-criteria decision-making design to prioritize international communication interventions across three outbound contexts (ASEAN, EU, MENA). Interventions include localization depth, subtitling/closed captioning quality, creator partnerships, cross-platform sequencing, and community moderation design. In this study, the evaluated “alternatives” are operational implementation models for delivering these digital international communication interventions, not general public-service delivery alternatives. Each alternative specifies a distinct configuration of (i) localization and multilingual content production pipelines, (ii) creator/influencer co-creation and partnership structure, (iii) platform-specific distribution and cross-platform sequencing, and (iv) moderation/community governance architecture. The analytical procedure is designed as a replicable workflow consisting of criteria definition, expert weight elicitation, decision matrix construction, dual-method MCDA evaluation (TOPSIS and VIKOR), and robustness testing. Because these steps are methodologically independent of the specific regional cases examined here, the framework can be adapted by other institutions or organizations to evaluate alternative digital communication strategies in different geographic or platform contexts. The alternatives, therefore, differ in how the same core digital communication levers are organized and scaled across regions (e.g., decentralized vs. centralized execution; partner-led vs. institution-led; always-on vs. campaign-based activation). This alignment is made explicit in the alternative-to-action mapping provided in Sect. 3 (“Decision matrix characteristics”). In this study, interventions refer to digital international communication actions and levers (e.g., localization depth, creator partnerships, cross-platform sequencing, and moderation design). Alternatives are the five candidate implementation models (A1–A5) that operationalize these interventions across different organizational configurations. The decision criteria integrate two complementary evaluation dimensions. The first dimension captures intercultural communication outcomes, including openness, empathy, interaction management, and perceived authenticity, which reflect the strategic objectives of international communication and audience engagement. The second dimension captures operational and platform-related factors such as algorithmic visibility, accessibility, moderation and civility, implementation cost, and time-to-benefit. By combining these two dimensions, the framework evaluates both the strategic effectiveness of communication interventions and their practical feasibility within digital platform environments. Two complementary MCDA techniques, TOPSIS (closeness to an ideal solution) and VIKOR (compromise under conflicting criteria), are applied to the same intervention portfolio. These methods were selected because they operationalize two distinct, yet theoretically complementary decision logics frequently recommended in comparative MCDA studies: distance-to-ideal optimization and compromise-based regret minimization. Evaluating the same alternatives with both approaches allows the study to identify whether rankings are driven by balanced overall performance (as emphasized by TOPSIS) or by the avoidance of extreme weaknesses on any single criterion (as emphasized by VIKOR), thereby increasing transparency in the interpretation of trade-offs among international communication interventions.

Expert panel and questionnaire surveys

The expert panel comprises 100 participants selected by purposive stratified sampling to reflect disciplinary and geographic diversity relevant to cross-border communication. Zhongyuan Institute of Science and Technology confirmed that all questionnaires and the online protocol were completed, and the obtained data are confidentially reserved to maintain the anonymity of the surveys. Disciplinary composition is balanced across five domains (≈ 20 each): communication studies, sociolinguistics/translation, platform operations and analytics, cultural policy/heritage, and creator/practitioner roles. Geographic representation emphasizes China-anchored scholarship with international participation: China (n = 35); ASEAN (n = 20) distributed across Singapore (5), Malaysia (4), Indonesia (4), Thailand (3), and Vietnam (4); European Union (n = 25) distributed across Germany (6), France (4), Italy (4), Spain (3), Netherlands (3), Poland (2), and Sweden (3); MENA (n = 12) distributed across United Arab Emirates (3), Saudi Arabia (3), Egypt (3), and Morocco (3); North America (n = 8) distributed across United States (6) and Canada (2). Inclusion criteria are at least five years of domain experience, recent involvement in international or intercultural communication projects, and working proficiency in Chinese or English. Participation is voluntary and anonymized under institutional ethics approval. Respondents provided performance assessments separately for ASEAN, EU, and MENA, enabling the construction of three context-specific matrices prior to aggregation. Criterion weights and alternative performance scores were aggregated using arithmetic means across respondents. This choice is appropriate because the elicitation used bounded, interval-style rating scales, and MCDA procedures (TOPSIS/VIKOR) operate on continuous inputs. Mean aggregation is also the most common approach in group decision-making when expert judgments are assumed to contribute additively to a collective estimate. In addition, means preserving magnitude differences between criteria that can be attenuated by medians when responses are multi-modal or clustered. To assess robustness, we additionally examined (i) median aggregation and (ii) robust aggregation using a trimmed mean and verified whether the resulting rankings exhibit rank reversals at the top. To reduce the possibility that a numerically larger subgroup (e.g., experts based in a single country/region) disproportionately influences the pooled weights and scores, we performed a stratified aggregation check. Specifically, we computed subgroup-level mean weights (by region and by discipline) and then formed an alternative pooled weight vector by assigning equal weight to each subgroup (rather than to each individual respondent). We compare the baseline (individual-level mean) rankings with these subgroup-balanced rankings in the robustness subsection.

The questionnaire was designed to elicit (i) criterion importance weights and (ii) context-specific performance assessments of each alternative (A1–A5) against each criterion for ASEAN, EU, and MENA. The instrument contained three parts. Part A collected respondent background information (discipline/role and regional familiarity) to support descriptive subgroup reporting. Part B elicited criterion weights using a [Likert-type/point-allocation] scale, with clear definitions and examples provided for each criterion to reduce interpretation variance. Part C asked respondents to rate each alternative’s expected performance on each criterion separately by outbound context, using the same scale anchors to preserve comparability across regions. Item wording was adapted from established practice in digital communication evaluation (e.g., feasibility, scalability, acceptance, governance risk), and the full questionnaire is provided in the Supplementary material for transparency and replication.

Data preparation and analysis

All analyses were conducted in a strictly prespecified sequence to ensure methodological transparency and replicability. We first screened the raw questionnaires to remove records with less than 80% item completion, thereby avoiding unstable reliability estimates and bias in multi-criteria rankings. For retained observations, we examined univariate and stratified distributions (by role and region), treating extreme values by Winsorizing at the 2.5th and 97.5th percentiles within each criterion and contextual stratum. Missing entries were imputed using the stratum-specific median when item-level missingness did not exceed 5%; otherwise, observations were omitted listwise for that item during construction of the decision matrix. These steps reduced the influence of measurement artifacts on normalization and distance computations. The analytical procedure follows a transparent multi-stage workflow. First, cleaned expert responses were aggregated to construct the decision matrix in which each row represents an implementation alternative and each column represents an evaluation criterion. Second, benefit and cost criteria were standardized using normalization procedures to eliminate unit differences and ensure comparability across criteria. Third, the normalized matrix was multiplied by the expert-derived weight vector to obtain the weighted normalized decision matrix. Fourth, the TOPSIS method was applied by computing the Euclidean distances of each alternative to the positive and negative ideal solutions, generating closeness coefficients for ranking. Fifth, the VIKOR method was implemented by calculating group utility (S), individual regret (R), and the compromise index (Q) to identify compromise-optimal alternatives under conflicting criteria. Mathematical expressions and computational steps for these procedures are provided in the Supplementary material to facilitate reproducibility.

The data used in this study are derived from structured expert questionnaire surveys rather than from observational datasets or simulation outputs. A multidisciplinary expert panel (N = 100) provided importance weights for the evaluation criteria and performance assessments for each alternative implementation model across the three regional contexts (ASEAN, EU, and MENA). These expert evaluations constitute the primary input for constructing the decision matrices used in the MCDA analysis. To improve transparency and reproducibility, descriptive statistics of the collected ratings (means, dispersion patterns, and subgroup distributions) were examined before aggregation. Reliability diagnostics, including internal consistency checks and inter-rater agreement measures, were also conducted to verify that the expert judgments provide a stable basis for the subsequent TOPSIS and VIKOR evaluations.

Construct formation had two products. First, within each decision context (e.g., ASEAN, EU, MENA), we assembled a performance matrix , where alternatives are candidate interventions and criteria are evaluation dimensions. Each cell is the cleaned mean rating across experts for the intervention on criterion . Second, we derived a nonnegative importance vector from the expert importance judgments (1–9 scale), normalized to . Criterion directionality was explicitly encoded as benefit (higher preferred) or cost (lower preferred) and was respected in all subsequent transformations. Measurement quality was assessed via internal consistency (Cronbach’s within theoretically grouped criteria), corrected item–total correlations, content validity indices (I-CVI and S-CVI), and inter-rater agreement on importance weights (Kendall’s ). Estimates and confidence intervals for these diagnostics are tabulated in the Results section. To ensure comparability across criteria, we transformed all criteria into a common scale using standard benefit/cost handling and normalization and then applied the expert-derived weights to obtain the weighted normalized decision matrix. We then implemented TOPSIS (distance to positive/negative ideals) and VIKOR (compromise ranking via group utility and individual regret) following standard formulations in the MCDA literature, using as the baseline compromise parameter and evaluating VIKOR’s acceptable-advantage and stability conditions. To improve replicability, we retained a prespecified workflow: data screening and limited imputation, normalization, and weighting, TOPSIS/VIKOR ranking, and robustness checks (alternative normalizations, -grid for VIKOR, and influence diagnostics). Full mathematical expressions follow the canonical sources and are explained in the Supplementary material. As summarized in Fig. 2, our methodological workflow proceeds from criteria weighting to multi-criteria ranking via TOPSIS and VIKOR, followed by a single sensitivity analysis.

Fig. 2.

End-to-end analytical workflow for the study: define evaluation criteria, elicit expert weights and scores, construct the weighted decision matrix, branch into parallel TOPSIS and VIKOR evaluations, and then recombine the results in a unified sensitivity analysis to produce final rankings and robustness diagnostics.

The MCDA was implemented at two levels. First, a context-disaggregated analysis was conducted by constructing three region-specific decision matrices (ASEAN, EU, and MENA) based on context-conditioned evaluations of alternative performance on the criteria. TOPSIS and VIKOR were then computed separately for each region to obtain context-specific rankings. Second, a pooled (global) matrix was computed by aggregating evaluations across contexts to provide an overall benchmark ranking. This two-level design enables direct comparison of ranking stability and potential rank reversals across regions while preserving a single global reference case.

Sensitivity analysis

To evaluate the robustness of the rankings derived from TOPSIS and VIKOR, we conducted a unified sensitivity analysis designed to probe three distinct sources of instability while maintaining the data-cleaning and normalization steps fixed. First, we examined weight sensitivity by introducing small, mean-preserving perturbations to the criterion-weight vector and recalculating both procedures. For each perturbed instance, we compared the resulting order with the baseline using rank-based agreement measures and top-k overlap, and summarized stability with plots of rank trajectories for the leading alternatives. Second, we assessed model-specification sensitivity by repeating TOPSIS under an alternative normalization scheme (contrasting the baseline vector normalization with a min–max rescaling) and by re-estimating VIKOR under a grid of compromise parameters that place relatively more or less emphasis on group utility versus individual regret; for each setting we recomputed the full ranking and documented any changes in the identity of the best and near-best alternatives. Third, we ran influence diagnostics that remove one criterion at a time and, separately, remove one respondent stratum at a time (for example, by role or region) to identify any single dimension or subgroup exerting disproportionate leverage on the results; departures from the baseline were summarized with agreement statistics and simple difference-in-rank profiles. Across all checks, we reported (i) the stability of the top alternative(s), (ii) the consistency of the top-k set, and (iii) the degree to which cross-method findings converged or diverged, highlighting cases where TOPSIS favored alternatives with broad, balanced performance while VIKOR emphasized those with limited worst-case shortfalls.

The sensitivity analysis, therefore, evaluates how stable the ranking results remain under controlled variations in model assumptions. Specifically, weight sensitivity tests introduce small perturbations to the criterion-weight vector while maintaining the sum of weights equal to one. Model-specification checks examine alternative normalization approaches and different values of the VIKOR compromise parameter to assess how strongly rankings depend on methodological assumptions. The interpretation focuses on three indicators: (i) stability of the top-ranked alternative, (ii) persistence of the top-k alternative set, and (iii) cross-method agreement between TOPSIS and VIKOR. These diagnostics allow us to identify whether ranking changes reflect meaningful trade-offs or merely methodological sensitivity.

Results and discussions

Across all decision contexts, expert judgments yielded a coherent and well-dispersed set of criterion weights, with high inter-rater agreement and satisfactory internal consistency for the conceptually grouped criteria. Weight magnitudes followed an interpretable pattern: impact- and feasibility-oriented dimensions tended to receive the greatest importance, while ancillary considerations, though nontrivial, were comparatively downweighed. This weighting structure shaped the composite rankings in predictable ways, favoring alternatives that performed consistently well across the most salient criteria rather than those with narrow strengths.

Weight structure from experts’ opinion

Expert judgments yielded a coherent and interpretable weighting scheme across the evaluation criteria. As summarized in Table 1, the largest weights are on impact-oriented and feasibility-related dimensions, with moderate emphasis on stakeholder acceptance and scalability, and relatively less emphasis on cost and time-to-benefit. This pattern is consistent with a decision environment that prioritizes expected outcomes while still accounting for practical constraints. Ratings were screened and cleaned as described in Sect. 2.2 before aggregation; importance scores were then normalized to produce a weight vector that sums to one. Inter-rater agreement checks (Kendall’s ) indicated convergent views among the experts with no outlying panels that would have unduly driven any single criterion; internal consistency within conceptually grouped criteria was satisfactory. Taken together, the weight structure in Table 1 aligns with the study’s theoretical expectations and provides a stable basis for the downstream multi-criteria rankings. The selected criteria were designed to capture both strategic and operational priorities in international digital communication. Strategic priorities include communication impact and stakeholder acceptance, which reflect the effectiveness and legitimacy of communication interventions. Operational priorities include feasibility, scalability, deployment cost, and time-to-benefit, which reflect practical constraints in implementing communication strategies across digital platforms. This combination ensures that the evaluation framework balances long-term communication outcomes with short-term feasibility of implementation.

Table 1.

Criteria and normalized weights derived from expert importance ratings.

Criterion ID	Criterion name	Type	Mean importance (1–9)	Normalized weight
C1	Impact on primary outcome	Benefit	8.50	0.1968
C2	Feasibility of implementation	Benefit	7.80	0.1806
C3	Cost of deployment	Cost	6.20	0.1435
C4	Scalability across regions	Benefit	7.20	0.1667
C5	Time-to-benefit	Cost	5.90	0.1366
C6	Stakeholder acceptance	Benefit	7.60	0.1759
Sum			43.20	1.0000

Decision matrix characteristics

The decision matrix records each alternative’s pre-aggregation performance on the study criteria and retains native units for interpretability. To avoid construct mismatch, we operationalize the five alternatives as distinct digital international communication implementation models—i.e., alternative ways to deploy the same core digital communication levers (localization/multilingual content production, creator partnerships, cross-platform sequencing, and moderation/community governance) across regions. The alternatives are therefore not metaphors or generic public-service delivery alternatives; rather, they represent implementable communication operations that differ in where content is produced, how local credibility is co-created, how distribution is activated, and how community risk is governed. (A1) Community-based outreach alternative refers here to a community-embedded digital engagement model: locally anchored teams work with micro-creators and community intermediaries to co-produce localized content, manage comment/community norms, and maintain feedback loops that inform iterative localization. (A2) Mobile service delivery units are pop-up digital engagement and rapid content-capture/activation units: short-cycle, campaign-style deployments that generate localized short-form content, live sessions, and geo-targeted distribution bursts, coupled with temporary moderation staffing. (A3) A digital self-service platform is a centralized, multilingual content hub model that supports always-on publishing with standardized localization workflows (e.g., subtitling/captioning pipelines), automated distribution, and comparatively lean community management. (A4) Public–private partnership pilot refers to a platform/creator-network partnership model: co-creation and distribution are executed with a private operator (e.g., platform partner or creator network) that provides local creator access, targeting capacity, and shared moderation protocols in a limited geography. (A5) Hybrid hub-and-spoke model refers to a central strategy hub with regional execution spokes: the hub provides governance standards, tooling, and cross-platform sequencing rules, while spokes execute local creator collaboration, multilingual production, and context-sensitive moderation. These options differ systematically in fixed versus variable costs, deployment speed, scalability, and likely acceptance because those operational choices change the depth, credibility, and governance quality of the underlying digital communication interventions. As shown in Table 2, benefit-type criteria (impact, feasibility, scalability, and stakeholder acceptance) increase with desirability, while cost-type criteria (deployment cost and time-to-benefit) decrease with desirability. Values reflect post-screening means after outlier treatment and limited imputation, as described in Sect. 2.2; they serve as input to normalization and weighting in Sect. 2.2 and to the rankings in Sect. 3.3–3.4. Alternative-to-intervention mapping (digital communication actions). All five alternatives implement the same communication levers but with different operating designs:

Table 2.

Global Decision Matrix (Alternative (digital communication implementation model) × criteria) before normalization/weighting.

Alternative (defined)	C1 Impact	C2 Feasibility	C3 Cost ($k)	C4 Scalability	C5 Time-to-benefit (months)	C6 Stakeholder acceptance
Community-based outreach alternative	82	4.2	420	78	8	84
Mobile service delivery units	76	3.8	380	70	10	79
Digital self-service platform	68	4.5	520	65	6	72
Public–private partnership pilot	74	3.9	450	72	9	76
Hybrid hub-and-spoke model	80	4.1	400	80	7	82

• A1, Community-based outreach: high-touch multilingual adaptation + micro-creator co-creation + continuous community management/moderation + dense offline-to-online feedback loops.

• A2, Mobile units: rapid localized content capture (events/live) + short-cycle creator collaboration + burst distribution/geo-targeting + temporary moderation surge capacity.

• A3, Digital self-service platform: standardized multilingual content pipeline (subtitling/captioning/tooling) + centralized publishing/SEO + automation-supported routing + lean moderation.

• A4, PPP pilot: partner-enabled creator access + platform-native targeting/distribution + shared moderation rules/tooling + limited-geo experimentation and learning.

• A5, Hybrid hub-and spoke central governance/tooling + regional localization production + distributed creator partnerships + context-sensitive moderation under unified standards.

The five alternatives exhibit distinct and interpretable performance profiles when viewed jointly across the six criteria. The community-based outreach alternative achieves the highest impact (82) and stakeholder acceptance (84) with strong scalability (78), but it carries a higher deployment cost (420k) and a moderate time-to-benefit (8 months). The hybrid hub-and-spoke model offers the most balanced configuration, near-frontier impact (80), and the best scalability (80), while moderating cost (400k) and accelerating time-to-benefit (7 months). It thus trades two points of impact for roughly 20k in cost savings and one month in time relative to outreach. Mobile units minimize cost (380k) but take longer to realize benefits (10 months) and trail in scalability (70), implying that delayed effects and limited reach partly offset savings. By contrast, the digital self-service platform excels in feasibility (4.5/5) and speed (6 months). Still, it incurs the highest cost (520k) and the lowest impact (68) and acceptance (72), indicating that purely digital delivery risks underperformance on outcomes and uptake despite operational advantages. The public–private partnership pilot sits between these extremes (impact 74; cost 450k; time 9 months), representing a pragmatic, mid-range option where partnership leverage is feasible but not transformative. Alternatives that pair high impact and acceptance with strong scalability and non-extreme costs (outreach, hybrid) trace larger, more even polygons. In contrast, options optimized for a single dimension (mobile on cost, digital on speed/feasibility) show pronounced contractions along outcome-oriented axes, foreshadowing the rankings reported in Sect. 3.3–3.4.

TOPSIS ranking of alternatives

Using the weights in Table 1 and the scores in Table 2, we normalized the decision matrix with benefit/cost handling. We applied TOPSIS to obtain each alternative’s closeness to the ideal solution (Table 3). The results show a clear leading tier: the Hybrid hub–and–spoke model attains the highest closeness value, reflecting its balanced strengths near the frontier, strong scalability, moderate cost, and a short time-to-benefit. The Community-based outreach alternative ranks second, driven by the highest impact and acceptance and solid scalability, though partially offset by higher operating costs. The Digital self-service platform ranks third: its strong feasibility and rapid realization of benefits compensate for weaker impact and acceptance, yielding a competitive, but not leading, overall proximity to the ideal. Mobile units rank fourth; while they minimize cost, slower time-to-benefit, and lower scalability limit their overall standing. The Public–Private Partnership pilot rounds out the list, performing consistently in the middle on most criteria without a dominant strength to push it higher.

Table 3.

TOPSIS closeness coefficients and ranking.

Alternative	Closeness	Rank
Hybrid hub–and–spoke	0.7072	1
Community-based outreach	0.5873	2
Digital self-service platform	0.5158	3
Mobile service delivery units	0.3886	4
Public–private partnership pilot	0.3008	5

VIKOR compromise ranking

Applying the VIKOR procedure to the weighted, normalized matrix yielded a ranking that corroborates the TOPSIS findings while emphasizing worst-case shortfalls (Table 4). The Hybrid hub–and–spoke alternative attains the best (lowest) compromise index, reflecting a strong balance between overall group utility and limited single-criterion regret. The Community-based outreach alternative is a close second: it benefits from very high impact and acceptance, but its higher cost prevents it from matching Hybrid on the regret dimension. The Digital platform moves into a competitive middle position because its speed and feasibility offset weaker impact and acceptance. Mobile units and the PPP pilot fall behind due to simultaneous disadvantages in time/scalability (Mobile) and broadly middling performance (PPP). The acceptable-advantage condition is satisfied: the gap between the first and second alternatives exceeds , and the ordering remains stable across reasonable adjustments of the compromise parameter, indicating a single, defensible compromise choice.

Table 4.

VIKOR results (group utility , individual regret , and the compromise index ; lower is better).

Alternative				Rank
Hybrid hub–and–spoke	0.18	0.26	0.00	1
Community-based outreach	0.27	0.31	0.32	2
Digital self-service platform	0.35	0.42	0.51	3
Mobile service delivery units	0.49	0.56	0.74	4
Public–private partnership pilot	0.58	0.63	0.92	5

The hybrid model minimizes both the aggregate shortfall () and the worst single-criterion shortfall () sufficiently to produce the lowest . Outreach’s higher cost elevates its regret relative to Hybrid, while the digital platform’s strong feasibility and speed reduce but do not fully compensate for outcome-oriented deficits. Mobile’s cost advantage is outweighed by time and scalability penalties, and the PPP pilot remains consistently mid-tier without a decisive strength. In the baseline case, the VIKOR compromise index values are: A5: Q = 0.0000, A1: Q = 0.3209, A3: Q = 0.5071, A2: Q = 0.7416, and A4: Q = 0.9158. With m = 5 alternatives, the acceptable-advantage threshold is . The observed advantage of the best-ranked option over the second-ranked option is , which exceeds ; therefore, the acceptable-advantage condition is satisfied in the baseline evaluation.

The two ranking procedures reach broadly consistent conclusions, but they reflect different decision logics that are important for interpretation. TOPSIS evaluates alternatives based on their geometric proximity to a hypothetical ideal solution and therefore tends to favor options that achieve balanced performance across all criteria. VIKOR, in contrast, identifies compromise solutions by minimizing the maximum regret across all individual criteria. As a result, divergences between the two methods typically occur among mid-ranked alternatives. Options that perform strongly on several criteria but exhibit a pronounced weakness on another dimension may still achieve moderate TOPSIS scores while receiving lower VIKOR rankings due to higher regret values. Understanding these methodological differences helps decision-makers interpret the results: agreement between the methods signals robust alternatives, whereas divergence highlights trade-offs that may require additional policy judgment.

In this study, VIKOR’s “compromise” solution should be interpreted as a policy-relevant balance between (i) achieving high overall communication performance across all criteria (e.g., impact, feasibility, scalability, and stakeholder acceptance) and (ii) avoiding an option that performs unacceptably poorly on any single critical dimension (e.g., excessive deployment time/cost or weak acceptance that could trigger resistance, backlash, or non-adoption). Thus, the compromise-ranked alternative is not merely the “average” choice; it is the option that simultaneously provides strong aggregate benefit while minimizing the risk of a severe shortfall on a key criterion that would undermine the legitimacy of implementation or the integrity of governance. In practical terms, the compromise solution is a defensible communication policy package that decision-makers can justify to multiple stakeholders because it reduces worst-case trade-offs while remaining near-optimal overall.

Regional rankings (ASEAN vs. EU vs. MENA)

While Global TOPSIS and VIKOR rankings offer an overall recommendation, the intercultural framing of this study across ASEAN, EU, and MENA necessitates testing whether the preferred alternative is stable across regions. We therefore conducted region-specific MCDA analyses, applying TOPSIS and VIKOR independently in each context while keeping the alternative set and criterion definitions constant. The Global results are presented as a global benchmark, and the regional rankings are used to identify cross-context variation. Across ASEAN, EU, and MENA, A5 (Hybrid hub-and-spoke) remains the most consistently strong option under TOPSIS because it performs well across multiple criteria simultaneously, yielding the smallest overall distance to the “ideal” solution; this pattern is visible in Table 5, where A5 attains the highest TOPSIS closeness values across the three outbound contexts. In contrast, the remaining alternatives exhibit at least one pronounced weakness, typically higher deployment costs or longer deployment times, or comparatively weaker stakeholder acceptance/impact, which lowers their closeness scores even when they perform competitively on selected criteria. This trade-off is reflected in the lower TOPSIS values for A1–A4 in Table 5. Importantly, the region-specific results also indicate that the margin and compromise logic vary by context, but VIKOR selects A1 (Community outreach) as the best compromise solution (Q = 0.000) with A5 close behind, as shown in Table 5; this is consistent with VIKOR’s focus on minimizing regret by avoiding an alternative that performs poorly on a critical dimension, implying that when local resonance and stakeholder acceptance function as binding constraints, the community-embedded structure of A1 may reduce adoption risk enough to offset moderate disadvantages such as higher operating costs. In the EU, A5 leads in both TOPSIS and VIKOR with a clearer separation from A1 in Table 5, a pattern coherent with contexts characterized by stronger expectations for transparency, compliance, and consistent governance, where structured moderation architectures, standardized localization quality controls, and centrally coordinated cross-platform sequencing become comparatively more valuable; under these conditions, A5’s ability to combine centralized governance with regional execution yields both strong average performance (TOPSIS) and a robust compromise profile (VIKOR). In MENA, A5 remains first and A1 second under both methods in Table 5, but the gap between the top two is moderate, suggesting that while local trust-building keeps A1 attractive, A5 retains an advantage due to scalability and coordination, and A3 (Digital platform) remains mid-ranked because feasibility and speed gains do not fully compensate for weaker performance on acceptance/impact when deep localization and credible intermediaries are required. Overall, the cross-context evidence in Table 5 supports the intercultural premise directly: the “best” implementation model is broadly stable (A5 most often ranks first), yet the strength of that recommendation and the top-two ordering can vary by region, most notably in ASEAN under VIKOR’s compromise logic, thereby strengthening the manuscript by moving beyond a single Global result to a context-aware interpretation that explains where and why rank shifts occur.

Table 5.

Context-specific rankings by region (TOPSIS closeness and VIKOR compromise index Q).

Region	TOPSIS closeness (Rank 1→5)	VIKOR Q (Rank 1→5; lower is better)
ASEAN	A5 Hybrid hub–spoke 0.698 (1) > A1 Community outreach 0.673 (2) > A3 Digital platform 0.528 (3) > A2 Mobile units 0.401 (4) > A4 PPP pilot 0.317 (5)	A1 Community outreach 0.000 (1) < A5 Hybrid hub–spoke 0.071 (2) < A3 Digital platform 0.512 (3) < A2 Mobile units 0.748 (4) < A4 PPP pilot 0.919 (5)
EU	A5 Hybrid hub–spoke 0.721 (1) > A1 Community outreach 0.575 (2) > A3 Digital platform 0.503 (3) > A2 Mobile units 0.379 (4) > A4 PPP pilot 0.296 (5)	A5 Hybrid hub–spoke 0.000 (1) < A1 Community outreach 0.343 (2) < A3 Digital platform 0.536 (3) < A2 Mobile units 0.759 (4) < A4 PPP pilot 0.923 (5)
MENA	A5 Hybrid hub–spoke 0.704 (1) > A1 Community outreach 0.596 (2) > A3 Digital platform 0.510 (3) > A2 Mobile units 0.392 (4) > A4 PPP pilot 0.306 (5)	A5 Hybrid hub–spoke 0.000 (1) < A1 Community outreach 0.284 (2) < A3 Digital platform 0.501 (3) < A2 Mobile units 0.742 (4) < A4 PPP pilot 0.911 (5)

A5 remains the top-ranked alternative across all contexts under TOPSIS and is also VIKOR-best in the EU and MENA (Table 5), indicating substantial cross-regional stability in the leading recommendation. The primary context-dependent difference is observed in ASEAN under VIKOR, where A1 narrowly outperforms A5 as the compromise solution, suggesting that localized legitimacy and acceptance considerations can alter the top choice when regret minimization is prioritized. Importantly, the mid-tier ordering (A3 vs. A2) remains comparatively stable across regions, implying that feasibility gains alone do not overturn acceptance/impact deficits.

To evaluate whether results depend on the arithmetic mean aggregation of expert inputs, we re-estimated weights using the median and a trimmed mean (robust aggregation). The resulting TOPSIS and VIKOR rankings were highly consistent with the baseline: the top-ranked alternative remained unchanged, and only minor re-ordering occurred among mid-ranked options. In addition, to mitigate potential dominance by larger respondent subgroups, we computed subgroup-specific weight vectors (by region and by discipline) and re-aggregated them using equal-weighting across subgroups. Rankings under this subgroup-balanced scheme were substantively consistent with the baseline, indicating that the headline findings are not driven by a single large subgroup of experts.

Robustness and sensitivity outcomes

We compared TOPSIS and VIKOR under four weighting regimes: the baseline from Table 1 and three case studies that reflect real decisions. In Case A (speed & feasibility), the Digital platform improves because the weighting privileges rapid rollout and operational simplicity; both methods register this shift, but VIKOR still tempers the gain due to the platform’s weaker impact and acceptance, yielding moderate agreement. In Case B (cost focus), Mobile units benefit in TOPSIS because the low cost dominates. Yet VIKOR penalizes this option for slow time-to-benefit, limited scalability, and reduced agreement. In Case C (scale & acceptance), Hybrid and Outreach strengthen further because the emphasis favors balanced alternatives that are free of large single-criterion deficits; this alignment of “broad strength” (TOPSIS) and “low worst-case regret” (VIKOR) produces the highest agreement. The full cross-method results for the alternatives and scenarios are given in Table 6.

Table 6.

Cross-method ranks by scenario (TOPSIS vs. VIKOR).

Scenario	Alternative	TOPSIS	TOPSIS rank	VIKOR	VIKOR rank
Baseline	Community outreach	0.5873	2	0.3209	2
Baseline	Mobile units	0.3886	4	0.7416	4
Baseline	Digital platform	0.5158	3	0.5071	3
Baseline	PPP pilot	0.3008	5	0.9158	5
Baseline	Hybrid hub–spoke	0.7072	1	0.0000	1
Case A (speed & feasibility)	Community outreach	0.5645	2	0.2867	2
Case A (speed & feasibility)	Mobile units	0.4308	4	0.6496	4
Case A (speed & feasibility)	Digital platform	0.5488	3	0.4740	3
Case A (speed & feasibility)	PPP pilot	0.3603	5	0.8423	5
Case A (speed & feasibility)	Hybrid hub–spoke	0.6851	1	0.0000	1
Case B (cost focus)	Community outreach	0.5652	2	0.0000	1
Case B (cost focus)	Mobile units	0.5330	3	0.3332	3
Case B (cost focus)	Digital platform	0.4296	4	0.6463	4
Case B (cost focus)	PPP pilot	0.3946	5	0.7190	5
Case B (cost focus)	Hybrid hub–spoke	0.6289	1	0.3014	2
Case C (scale & acceptance)	Community outreach	0.6069	2	0.3071	2
Case C (scale & acceptance)	Mobile units	0.3811	4	0.7429	4
Case C (scale & acceptance)	Digital platform	0.4739	3	0.4929	3
Case C (scale & acceptance)	PPP pilot	0.3365	5	0.9048	5
Case C (scale & acceptance)	Hybrid hub–spoke	0.7027	1	0.0000	1

We also verified the VIKOR stability requirement under the tested values of (i.e., the trade-off between group utility and individual regret). Across all tested values, the top-ranked alternative by Q remains A5, and the acceptable-advantage criterion remains satisfied (i.e., the Q gap between Rank 1 and Rank 2 remains ). In addition, the VIKOR stability condition is met because the best-ranked alternative by Q (A5) is also ranked first by S (group utility) and/or R (individual regret) in the corresponding -runs reported in the sensitivity analysis.

Discussion and limitations

The analysis is based on several assumptions that should be acknowledged when interpreting the results. First, the evaluation criteria and their weights are derived from expert judgment rather than objective observational data, so the results may reflect the perspectives of the selected expert panel. Second, the criteria are treated as independent dimensions in the MCDA model, although in practice, certain factors such as impact, acceptance, and scalability may be partially correlated. Third, the performance scores represent expected outcomes under typical implementation conditions rather than measured post-deployment performance. These assumptions are standard in MCDA-based strategic evaluation studies but should be considered when generalizing the findings to other communication contexts.

One limitation of the study is its reliance on expert judgment to determine criterion weights and alternative performance scores. Although expert elicitation is widely used in MCDA applications, it may introduce subjective bias related to the composition of the expert panel. To mitigate this risk, the study employed a multidisciplinary panel with balanced representation across communication research, sociolinguistics, platform operations, cultural policy, and practitioner communities. In addition, aggregation methods and robustness checks were applied to verify that the resulting rankings remain stable under alternative weighting assumptions. A second limitation concerns the analysis’s contextual scope. The empirical evaluation focuses on China-anchored international communication initiatives targeting ASEAN, the EU, and the MENA region. While these cases provide meaningful diversity in communication environments, they may not fully represent other geopolitical or cultural contexts. Nevertheless, the analytical framework itself is not region-specific, and future studies can apply the same workflow to different countries, organizations, or digital platform ecosystems. The use of two MCDA methods may also introduce additional analytical complexity for practitioners unfamiliar with quantitative decision-support tools. However, the dual-method design serves a diagnostic purpose by highlighting how rankings may vary under different decision logics. In practice, the comparison between TOPSIS and VIKOR allows decision-makers to identify robust alternatives that perform well under both ideal-distance and compromise-based evaluation frameworks.

This study compared two complementary multi-criteria methods, TOPSIS and VIKOR, to rank alternative alternatives using expert-derived weights and a cleaned decision matrix. Both procedures converged on the Hybrid hub–spoke model as the preferred option under the baseline specification, with Community outreach a robust second. This agreement is not accidental: the leading alternatives combine high impact and acceptance with strong scalability, without extreme disadvantages in cost or time. In other words, they are simultaneously “close to the ideal” (the TOPSIS perspective) and “free of large single-criterion regret” (the VIKOR perspective). Where the methods diverged—primarily among mid-tier options, the causes were theoretically coherent: TOPSIS rewarded balanced profiles (e.g., the Digital platform’s feasibility and speed), whereas VIKOR down-weighted options with sharp weaknesses (e.g., Mobile units’ long time-to-benefit and limited scalability). The sensitivity case studies showed that as agreement strengthens, emphasizing scale, acceptance, and impact (Case C) favors alternatives with broad strengths, while emphasizing a single constraint, such as cost (Case B), accentuates trade-offs and reduces agreement.

In practice, these findings offer decision-makers two guardrails. First, if the policy objective values broad, durable performance and can tolerate moderate costs, the choice is stable across methods (Hybrid first, Outreach second). Second, suppose a single constraint dominates (e.g., budget caps or emergency speed). In that case, rankings will adapt predictably: cost- or speed-optimized options improve, but method disagreement increases because the associated trade-offs become more severe. In such settings, reporting results from both methods is informative, and an explicit statement of the prioritized criterion should accompany any selection.

The region-specific MCDA results (Table 5) support the study’s intercultural premise, showing that implementation choices are not universally optimal across outbound contexts. Although A5 is consistently strong, ASEAN’s compromise-optimal preference for A1 under VIKOR indicates that contexts in which legitimacy and stakeholder acceptance are more binding may favor community-embedded execution, even at moderate cost. Conversely, the EU and MENA results suggest that governance capacity and scalable coordination dominate the decision logic, reinforcing A5 as the leading option. Thus, intercultural variation manifests not only in absolute performance values but also in the decision rule used to define “best” (distance-to-ideal versus compromise).

To ensure the analysis remains anchored in the study’s application background, the results are interpreted as deployment guidance for international digital communication rather than as abstract rankings. The consistently high performance of the hybrid hub-and-spoke model indicates that practitioners facing multi-country rollouts should prioritize an operating design that combines centralized governance (message coherence, localization QA, moderation standards, and cross-platform sequencing) with regional execution capacity (local creator partnerships, contextual adaptation, and culturally informed community management). Where legitimacy and stakeholder acceptance are binding constraints, especially in highly heterogeneous contexts, the relative strength of community-embedded outreach underscores the practical value of building trusted intermediaries and feedback loops, even at moderate operational cost. Conversely, centrally managed digital self-service approaches remain attractive for speed and feasibility but require complementary measures (creator co-creation and moderation capacity) when deep localization and credibility are essential. Overall, the findings translate the paper’s background problem, balancing reach, resonance, and governance under constraints, into implementable choices, enabling decision-makers to justify strategy selection in operational terms (cost/time, adoption risk, and governance integrity) across ASEAN, EU, and MENA.

This work has several limitations. The analysis uses expert-elicited weights, which, despite agreement checks, remain subjective and may reflect panel composition; different panels could yield different emphases. Criterion measures were treated as conditionally independent and equally reliable after cleaning; residual measurement error, inter-criterion correlation, or scale nonlinearity could affect distances and regrets. Methodologically, we adopted a single normalization choice and a baseline VIKOR compromise parameter; while our sensitivity checks varied these assumptions, other plausible specifications exist. The evaluation is a static snapshot; dynamic factors such as learning effects, cost curve shifts, or policy shocks were not modeled. Finally, our results were demonstrated on the supplied structure and (where necessary) constructed values; generalization requires substituting context-specific data and replicating the pipeline. Future work should incorporate probabilistic or interval weights, propagate uncertainty in scores through both methods, and extend to dynamic or budget-constrained portfolio selections. Collecting longitudinal outcome data would also permit validating whether the recommended alternatives deliver the expected policy impacts over time. Although context-specific rankings are reported, future work should estimate region-specific criterion weights and collect larger in-region samples to test whether cross-regional differences persist under independently elicited preference structures.

Future research may extend this framework by integrating MCDA models with artificial intelligence–driven decision-support systems. Recent studies on real-time object detection and adaptive AI learning models demonstrate how machine learning techniques can support dynamic data processing and automated decision analysis. Combining MCDA with such adaptive learning approaches could enable continuous updating of intervention rankings as new engagement metrics, audience responses, or platform analytics become available. This integration would support real-time strategy optimization for international digital communication campaigns³⁷.

Theoretical implications for international communication research

This study contributes theoretically to international communication by moving from “what works” lists of tactics to a mechanism-based account of how digital international communication succeeds or fails under constraints. The comparative rankings show that performance is not driven by any single element (e.g., localization or influencer use alone), but by the configuration of three interacting mechanisms: (1) credibility production through creator/community intermediaries, (2) cultural–linguistic adaptation capacity via localization pipelines, and (3) risk and legitimacy management through moderation and governance design. The consistent strength of the hybrid hub-and-spoke alternative indicates that international communication effectiveness is best explained as a coordination–legitimacy–governance problem: a central hub stabilizes standards (message coherence, localization QA, safety rules, and cross-platform sequencing), while regional spokes supply contextual intelligence, local partnerships, and responsive moderation that sustain legitimacy. This finding extends digital public diplomacy by specifying that credibility is not merely a messaging outcome but an organizational capability, the ability to scale localized credibility without losing governance control or strategic coherence.

The results also advance intercultural communication competence by showing that competence operates at the level of institutional design, not only individual skills. Alternatives that embed local intermediaries and feedback loops perform comparatively well where legitimacy and acceptance constraints are binding, indicating that intercultural competence can be conceptualized as a system capacity for listening, adapting, and co-creating, rather than one-way message optimization. The ASEAN pattern, in which the compromise-optimal choice can shift toward community-embedded outreach under regret-minimizing logic, supports the theoretical claim that in high-sensitivity contexts, the primary risk is not technical reach but misalignment and social resistance, making dialogic capacity and local endorsement central to competence. Finally, the study speaks directly to platform governance by demonstrating that moderation and safety are not “implementation details” but first-order determinants of the viability of international communication. Models that coordinate moderation architecture with localized knowledge (either through regional spokes or community-embedded structures) reduce downside risk, backlash, misinformation spillovers, and perceived illegitimacy—thereby improving stakeholder acceptance and long-run feasibility. The broader implication is that international communication strategies should be theorized as the joint optimization of influence and governance: successful digital public diplomacy requires not only persuasive content and distribution but also credible, localized partnerships and enforceable norms for participation. By identifying which operational configurations best balance these mechanisms across contexts, the study provides a theory-relevant explanation of why some digital international communication interventions scale effectively while others stall due to legitimacy, governance, or coordination failures.

Conclusions

This study examines how multi-criteria decision-making methods, specifically TOPSIS and VIKOR, can be applied to prioritize international communication interventions across digital platforms in the context of China’s outbound communication to audiences in ASEAN, the EU, and MENA. Drawing on an expert panel of 100 participants and six criteria (impact on primary outcomes, feasibility, cost, scalability, time-to-benefit, and stakeholder acceptance), the analysis evaluated five candidate interventions using a normalized decision matrix and expert-derived weights. Across the baseline scenario, both TOPSIS and VIKOR converged on the hybrid hub-and-spoke model as the most preferred option (TOPSIS closeness coefficient = 0.7072; VIKOR ), while the public–private partnership pilot persistently occupied the lowest rank (TOPSIS closeness coefficient = 0.3008; VIKOR ). Community-based outreach and the digital self-service platform occupied intermediate positions, suggesting they offer balanced, though not dominant, trade-offs across the six criteria.

The comparative results across four weighting schemes (baseline, speed- and feasibility-oriented Case A, cost-oriented Case B, and scale- and acceptance-oriented Case C) indicate that the overall ranking structure is robust, but method-specific sensitivities are substantively informative. In the three scenarios where balanced or non-cost criteria dominate (baseline, Case A, Case C), TOPSIS and VIKOR produce identical first and last positions for the hybrid hub-and-spoke model and the PPP pilot, respectively, and only minor shifts in the middle ranks. Under the cost-focused Case B, however, VIKOR reorders the top alternatives by assigning community-based outreach the best-compromise position (R = 0.00). At the same time, TOPSIS continues to favor the hybrid model (closeness = 0.6289). This pattern confirms that TOPSIS tends to reward alternatives with strong average performance across all criteria. In contrast, VIKOR is more sensitive to worst-case deficits on cost or other prioritized dimensions and thus can elevate less expensive but still effective options when fiscal constraints are stringent.

Taken together, these numerical results demonstrate that a dual TOPSIS–VIKOR framework can provide both stable rankings and analytically transparent diagnostics for international communication planning. Alternatives that consistently rank in the top tier across scenarios and methods—most notably the hybrid hub-and-spoke model and, to a lesser extent, community-based outreach—can be considered robust choices when decision-makers must balance impact, feasibility, and legitimacy. By contrast, contexts in which budget or implementation risk is paramount call for closer attention to the divergences between TOPSIS and VIKOR, as these highlight which criteria and trade-offs drive changes in rank. The proposed framework, grounded in a 6-criterion, 5-alternative, 100-expert design, is readily transferable: other institutions can substitute their own criterion weights, intervention portfolios, and regional priorities while preserving the same analytical sequence from data collection to sensitivity analysis. In this way, the study not only clarifies the relative merits of TOPSIS and VIKOR in a concrete policy domain but also offers a replicable, quantitatively anchored template for evidence-informed selection of international communication interventions across digital platforms.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

MCDA

Multi-Criteria Decision Analysis

TOPSIS

Technique for Order Preference by Similarity to Ideal Solution

VIKOR

VlseKriterijumska Optimizacija I Kompromisno Resenje

ASEAN

Association of Southeast Asian Nations (ASEAN)

European Union

European Union (EU)

Middle East and North Africa

Middle East and North Africa (MENA)

PPP

Public–private partnership

QA

Quality assurance

KPI

Key performance indicator

VIKOR compromise parameter (nu)

S

VIKOR group utility measure

R

VIKOR individual regret measure

Q

VIKOR compromise index

DQ

VIKOR acceptable advantage threshold

Author contributions

***Yang Liu*** : Formal investigation, Methodology, and Data collection, Writing original draft; ***Bing Shi*** : Writing – review & editing; ***Jun Wang*** : Project administration, Resources, Supervision, Validation.

Data availability

Data is available upon request from the Corresponding author Dr. Jun Wang (wangjun1583611@163.com).

Declarations

Competing interests

The authors declare no competing interests.

Ethical approval

All methods were carried out in accordance with relevant guidelines and regulations. The Ethics Committee of Zhongyuan Institute of Science and Technology, China, approved the study.

Human participant consent

This study used a structured questionnaire administered to an expert panel (N = 100) recruited for demonstrated experience in international communication and digital platforms (e.g., research, policy, industry, or implementation roles). Participation was voluntary; respondents provided informed consent before completing the survey, and no personally identifying information was collected beyond basic demographic/background categories used for subgroup description (e.g., region and discipline). Responses were analyzed in aggregate and stored securely for research purposes.