SIC is an organization’s capability to build new products or frameworks that deliver environmental, economic and social benefits. DS impact SIC through enhanced knowledge sharing via digital platforms, accelerated R&D and improved stakeholder co-creation through digital channels ([11]; [90]). Research indicates that sustainable digital initiatives directly improve SIC, which in turn elevates SOP ([11]; [55]). Therefore, digitally enabled SMEs adapt faster, fostering sustainable innovations that confer competitive advantages.

DS pushes companies to use digital resources to advance environmental and social goals. It helps companies to create routines that companies need to sense sustainability opportunities, reconfigure processes, and scale eco-innovations, thereby strengthening sustainable innovation capability ([53]; [78]). Previous studies consistently link digitalization to sustainability-oriented innovation. In manufacturing firms, in particular, digital transformation drives improvements in sustainable innovation capability ([32]). This means that digital capabilities are significantly raising green innovation performance, especially when paired with green supply-chain and top management’s awareness of environmental issues ([20]). Extending this to a strategic orientation, recent literature finds that DS initiatives positively affects green digital innovation, with capability reconfiguration acting as a driver underscoring DS as an upstream enabler of sustainability-focused innovative capacity ([90]).

Therefore, literature supports the hypothesis that DS is an integrative capability that systematically builds sustainable innovation capability in SMEs.

H1: Digital Sustainability positively impacts Sustainable Innovation Capability.

SOP balances economic, environmental, and social achievement, and is empowered by digital technologies to enable real-time monitoring, process optimization, and stakeholder transparency ([12]; [59]). [16] state that digital exploitation focuses on leveraging existing digital assets and exploring new digital frontiers, which in turn sustain the performance. Hence, the adoption of sustainable digitalization strategies positively affects SOP by reducing operational costs through automation, improving product quality via digital platforms, and facilitating responsiveness to market and regulatory shifts ([17]). DS has created tangible means for SMEs to facilitate SOP through higher resource efficiency, data transparency, and responsiveness towards stakeholders ([53]; [78]). Digital technologies at a micro level (IoT, data analytics, cloud) enable SMEs to eliminate waste, become resilient to sustainability standards, and develop resilience skills, with better operational and financial performance ([62]; [82]).

In recent literature, sustainability has become a related concept to digitalization, which resulted in the emergence of the concept of DS as a crucial concept, highlighting the role of technologies adoption in achieving sustainability outcomes ([41]). Literature highlights the growing role of sustainable digitalization on promoting organizational performance across various dimensions, and how integrating advanced technologies enhances the efficiency of organizations within their sustainable business strategies ([72]).

Existing literature highlights the growing role of sustainable digitalization in enhancing organizational performance across multiple dimensions. Organization’s strategic integration of digital technologies within their sustainable business strategies enhances operational efficiency and sustainable organizational outcomes ([72]).

Moreover, digitalization plays a key role in supporting sustainability goals by enhancing the organization’s processes and practices, which in return supports sustainability performance, especially when digital technologies are integrated with human resource and operational strategies to improve resource optimization, and environmental performance ([75]). Therefore, DS along with digital transformation, acts as strategic drivers promoting innovation, operational efficiency and sustaining business performance when integrated within the organization’s strategies. Using digital technologies strategically promotes not only productivity and value creation but also focuses and supports sustainable practices ([22]; [36]), especially in sectors going through transition towards sustainability.

To sum up, literature suggests that if companies connect digital efforts to sustainability objectives, they transform their digital capabilities into better sustainable performance at the firm level. ([53]; [78]; [85]; [91]).

H2: Digital Sustainability positively impacts Sustainable Organizational Performance.

SBMI involves reconfiguring value creation logic to incorporate environmental and social values in business model configuration. DS drives SBMI by facilitating platform-based ecosystems that enable sharing and circularity, optimizing resource use through IoT and analytics, and enabling new revenue streams via data monetization ([16]; [87]). Market-driven SBMI mediates the relationship between digital capabilities and SOP, since firms with strong digital strategies adopt innovative business models that enhance their sustainability outcomes ([15]; [17]). Integrating digital transformation with sustainable innovation is critical, especially with the emerging digital technologies in the market serving as strategic drivers for sustainability ([33]). This integration promotes the development of new business models incorporating not only digital technologies but also prioritizing sustainability ([5]).

In addition, this complementary relationship is highlighted by the role of digital technologies on developing novel configurations of sustainable business model elements, such as integrative and blended value propositions, and multidimensional value capture ([40]). This emphasizes the pivotal role of digital technologies in cultivating sustainable business practices and SBMI ([5]). Moreover, this integration empowers organizations to learn from dynamic environmental engagements and implement necessary innovations in their business models, particularly concerning value creation in the context of sustainability ([83]). Similarly, [54] highlight the impact of the organization’s digital platform ecosystem’s characteristics on their sustainable business model factors, thereby leading innovation toward sustainability.

Overall, DS serves as key enabler for SBMI by supporting organization in achieving sustainability performance. It provides the technological foundation and strategic framework for implementing SBMI that create value for multiple stakeholders. The business model framework, enhanced with DS considerations, provides a systematic approach for SMEs to redesign business models. In addition, DS reduces coordination and transaction costs for SBMI, making circular and low-impact models viable for SMEs that otherwise lack scale ([9]; [70]).

H3: The organization’s digital sustainability positively influences its sustainable business model innovation.

SIC is refers to the firm’s capability to continually create and adopt environmental and social product, process, and organizational innovations. It has been repeatedly linked to superior SOP. Literature studies in manufacturing show that firms with stronger SIC realize better environmental outcomes alongside improved corporate and financial results, because innovation cuts resource use and opens differentiation avenues ([4]; [89]). Recent SMEs evidence confirms that innovation capability is a significant predictor of business sustainability, even in resource-constrained contexts, highlighting its role as a dynamic capability that translates pressures and opportunities into durable triple-bottom-line gains ([63]).

Moreover, literature emphasized the positive impact of SIC on SOP by fostering sustainable competitive advantage and resilience, since firms with strong SIC are more able to integrate stakeholder demands, reduce environmental impact, and enhance overall performance outcomes ([27]). Consequently, developing SIC is both a strategic and operational fundamental for organizations seeking to achieve sustainable performance outcomes. Hence, fostering SIC via knowledge integration, process redesign, and governance for eco- and social innovation is crucial to perform environmentally, socially, and economically.

H4: The Organization’s sustainable innovation capabilities positively influence its sustainable performance.

Innovative sustainable business models empower companies to adopt environment, social, and economic elements into their business operations. SBMI involves reconfiguring how organizations build, produce, and acquire value a way to address sustainability challenges and generate a sustainable competitive advantage. Previous research indicates that companies adopting SBMI practices see improvements in operational efficiency, stakeholder engagement, and market reputation, which together enhance their overall sustainable performance.

Moreover, the literature is increasingly focusing on SBMI that stems from its potential to provide sustainable competitive advantages within dynamic external environments ([67]). SBMI is highlighted as a driver for both financial and sustainability performance, and the mechanisms through which it translates into organizational performance are still not yet fully clarified ([37]). Furthermore, existing literature on SBMI focuses on integrating sustainability into conventional business model elements rather than investigating it as transformative paradigm ([93]). Therefore, a detailed analysis of how various dimensions of SBMI individually and collectively influence organizational outcomes is needed ([4]).

By integrating sustainability into their business models, organizations can foster innovation that supports resource efficiency, social responsibility, and environmental stewardship. ([10]; [37]).

H5: The organization’s sustainable business model Innovation positively influence its sustainable organizational performance.

Literature in sustainability emphasizes the mediation of sustainability-oriented innovation capabilities between the organization’s sustainability practices its firm performance ([14]; [18]; [23]). Within SMEs, sustainable innovation adoption is a critical driver to performance and competitiveness under sustainability ([61]). The impact of DS on SIC aligns with literature evidence stating that digital transformation and related sustainable capabilities define and build the company’s sustainable innovation capacity by orchestrating resources and knowledge for eco-efficiency and clean technologies ([17]; [20]; [32]). In addition, previous studies show that SIC are critical for converting digital investments into measurable performance outcomes ([18]; [19]; [23]).

SBMI is considered a primary organizational mechanism supporting the translation of DS initiatives into a sustainable performance. Literature review further argues that SBMI emerges when digital and non-digital components co-evolve, positioning it as the configurational bridge that turns DS from process-level improvements into firm-level and triple-bottom-line results affecting the performance ([8]; [66]). Therefore, in the context of sustainability, companies reach performance gains when digital investments are embedded in sustainability-aligned business model designs ([16]). Hence the following hypotheses.

H6: The impact of digital sustainability on the organization’s sustainable performance is mediated by H6a. sustainable innovation capability and H6b. sustainable business model innovation.

This study adopts a combination of well-established measurement scales developed in previous studies, which consolidates the results’ validity and reliability. The measurement scale of DS has thirteen items adopted from [84] and reflecting digital quality and sustainability are reinforced through the

Figure 2. Methodology flowchart.

synergy of digital technology empowerment, sustainable leadership, and strategic integration. SIC is measured using five items formulated and adapted from [86] and highlighting the organization’s capability to encourage risk-taking in support of sustainable innovation, harness creativity to drive the development of sustainable products and services, exhibit tolerance for failure in uncertain contexts, and strategically position itself at the forefront of sustainable market innovation.

SBMI is reflected by a three-dimensional scale of ten items adapted from [6]. These items are associated with sustainable value creation and proposition, in addition to advancements in value delivery and value capture innovation. The measurement scale for SOP has seven items adapted from [35] study, assessing the company’s reduction of production costs, control of waste, optimization of resource use, improvement in employee satisfaction and customer satisfaction, capability to implement environmental protection standards, and its improved e-waste management.

To ensure the robustness of our instrument, a systematic process was carried out. Key measurement items were first extracted from a thorough in-depth analysis of the literature, after which they were refined to align with the research objectives and contextual framework. For further development of the scale and to validate the constructs’ structure and relevancy, we conducted a pilot panel of a team comprised of ten experienced SME managers and 4 postgraduate researchers from the relevant field of strategic management and environmental sustainability. To guarantee the constructs were appropriately structured and accurately represented the key factors, minor modifications were proposed by the experts. A pilot study was not performed, given that the items were derived from established research ([60]).

This study used the Likert scale with 7 points to operationalize the constructs, from “Strongly Disagree” to “Strongly Agree”. This scale offers a structured framework for participants to express their agreement or disagreement and effectively reflect their attitudes toward each construct. The values represent the level of agreement or disagreement, where (Strongly Disagree = 1), while (Strongly Agree = 7). This systematic technique makes it possible to collect data that is both consistent and quantitative and facilitates rigorous statistical analysis and the testing of hypotheses ([24]).

The survey instrument of this study was carefully crafted to capture pertinent variables pertaining to our study variables. In structural equation modeling (SEM), SmartPLS is a statistical tool often used to look at complicated relationships in a study model. This method makes it possible to look at how all the latent variables are connected and gives a strong statistical framework for checking and validating the hypotheses ([44]).

Using SmartPLS in the research design shows that the interplay of the main variables’ DS, SIC, SBMI and SOP is being analyzed in detail. This method makes it easier to measure and quantify these relationships, which helps us get a better sense of how these factors work together in the setting of Moroccan manufacturing SMEs. This choice of quantitative method leads to results that can be applied across various contexts. This helps researchers learn more about these processes and has real-world implications for both academics and professionals in the field.

The study was conducted on Moroccan manufacturing SMEs. We used a database of compliant Moroccan SMEs in different manufacturing activities from Casablanca-Settat region which is the biggest region in Morocco where SMEs are highly concentrated presenting 60.5% of the total turnover of Moroccan SMEs in 2023 ([64]).

To be eligible, a firm had to meet four criteria. First, the SME should be qualified according to the Moroccan legal definition established by Law No. 53-00 forming the SME Charter, employing no more than 200 permanent employees and with an annual turnover not exceeding MAD 75 million (with very small enterprises with a turnover of MAD 10 million or less) ([64]). Second, the principal activity of the SMEs should fall within the manufacturing sector. Third, the SME should be registered and active in the Casablanca-Settat region at the time of data collection. Last, the SME should have a valid and publicly available contact information.

The sampling frame included SMEs identified via publicly available SMEs databases obtained from the Moroccan commercial register, public business directories and industry associations, which were consolidated into a single list. After removing duplicate records and entries that did not meet the eligibility criteria, the frame comprised 3150 eligible firms. Contact information was used solely for academic research purposes.

A sample of 342 SMEs was selected using systematic random sampling, and the list of these companies was screened to assess whether they were available and agreed to participate in the study. SMEs in the frame were ordered alphabetically by their names, and a random starting point was generated as a random integer between 1 and 9, and every 9^th firm was selected thereafter until the target of 342 firms was reached. This study respects ethical research guidelines of protecting the respondents’ confidentiality and anonymity.

The study’s data were collected through the design of a structured questionnaire and the administration of a survey. To ensure cultural adaptation, questionnaires were available in both English and French, since French is the second language in Morocco. To ensure precise translation and prevent potential linguistic ambiguities for non-English speaking managers, the parallel translation method was utilized ([51]). The questionnaire was developed using Qualtrics and contained two sections: the First section was to gather data of the respondent’s personal information (Gender, Age, Educational level, Job position, years of experience) and the respondent’s company information (Industry field, company age, Company size) (Table 2). The second section included the four constructs from our theoretical model. 342 questionnaires were sent via email to the selected companies, with reminder emails sent in two weeks intervals. Moreover, data

Table 2. Demographic profile of research respondents.

were cleaned to remove irrelevant and short response time answers and verified manually to strengthen the quality of the data. We received 240 answers, 38 questionnaires with incomplete answers and 202 complete questionnaires for further analysis, yielding a response rate of 70% and a validity rate of 84%.

Common method and non-response bias were assessed for this study following the suggestions of [68]. In order to control common method bias, this study used several measures during the design of the questionnaire and the process of data collection. First, clarified definitions for each variable of this study were presented to minimize the respondents’ understanding bias. Second, the accuracy of the questionnaire was enhanced by adapting it to experts’ feedback. In addition, common method bias was evaluated using Harman’s one-factor as mentioned in previous studies ([85]; [86]), and the first factor accounted for less than 50% of the total variance. This indicates that common method bias did not pose a substantial issue in this study.

18.8% of the respondents were top managers, 65.8% mid-level managers and 15.3% were low-level managers working in quality, maintenance and financial departments. Most respondents are between 26 and 45 years old (61.4%), with an undergraduate educational level (70.3%), and with over 10 years of work experience (58.4%). Company profiles data show that the Moroccan manufacturing SMEs in our sample are operating in a broad range of sectors: Construction materials (8.9%), Chemicals, paints & detergents (9.4%), Agro-food processing (11.4%), Furniture, wood & joinery (8.4%), Leather & footwear (9.9%), Textiles & apparel (12.4%), Plastics & packaging (6.9%), Automotive components (24.8%), and Metalworking & fabricated metals (7.9%).

The predominant proportion of enterprises is medium-sized (51 - 200 employees) (88.1%), while small-sized enterprises with 10 - 50 employees presented 8.9% and micro-sized enterprises with less than 10 employees were 3% of the total sample. The sample of this study was drawn from multiple industries and included respondents with adequate educational and experiential backgrounds, thereby allowing respondents to answer effectively.

Partial least squares structural equation modelling (PLS-SEM) was selected in preference to covariance-based SEM (CB-SEM) for four reasons. First, the aim of this study is theory-development-oriented rather than purely confirmatory, a setting in which the variance-based PLS estimator is recommended. Second, the structural model is comparatively complex, incorporating two mediators (SIC and SBMI) and several simultaneous paths. Third, the analytical sample (n = 202) and the non-normal distribution of several indicators favor the distribution-free PLS algorithm, which does not require multivariate normality. Fourth, PLS-SEM performs robustly under this combination of model complexity, sample size and data characteristics ([45]; [74]). All four constructs (DS, SIC, SBMI and SOP) were specified as reflective: their indicators were drawn from previously validated reflective scales, are conceptually interchangeable manifestations of a single underlying latent variable, and are expected to be highly correlated and to covary. This reflective specification is precisely why each construct is evaluated through indicator loadings, internal-consistency reliability (Cronbach’s alpha, rho_A and CR), convergent validity (AVE) and the HTMT discriminant-validity criterion.

The constructs demonstrated strong internal consistency, with both CR and Cronbach’s Alpha exceeding the accepted threshold of 0.70 (Table 3). AVE values for some construct items were below the advised threshold of 0.50, therefore, the removal of these specific items was necessary to establish a robust and valid measurement model, thereby ensuring the integrity of the subsequent structural model analysis. DS4, DS8, DS9, DS10, DS12, SIC4, SOP1 were dropped in this study. The final, refined model meets all criteria for reliability and validity, and the study’s results are depicted in the following sections, thereby ensuring good convergent validity of the subsequent structural model analysis. To examine the structural validity, confirmatory factor analysis (CFA) was performed, and the results showed that the model structure is valid. Items with factor loadings greater than 0.4 were retained (Table 4) according to [77] who emphasize that items should have a factor loading of 0.40 or more to be significant.

Table 3. Reliability and validity measures.

Table 4. Retained measurement items.

According to Table 5, the data display a reasonably symmetric distribution. Skewness values range from −0.730 to −0.048 and kurtosis values from −0.897 to 0.484, both of which lie within acceptable limits, supporting the assumption of normality. Furthermore, standardized outer loadings exceed 0.70 (p < 0.001). Furthermore, the standardized outer loadings of the retained indicators reported in Table 5 range from 0.622 to 0.814 and therefore all satisfy the 0.50 retention threshold adopted in this study ([45]; [50]).

An assessment of multicollinearity using variance inflation factor (VIF) shows that all values are below the threshold of 5.0, with a maximum of 2.121, indicating that no multicollinearity concerns in the model ([45]).

Table 5. Cross loading, kurtosis, skewness and VIF values.

Discriminant validity was assessed using the heterotrait-monotrait ratio of correlations (HTMT), which [49] recommend as a more sensitive criterion than the Fornell-Larcker and cross-loadings approaches. Discriminant validity is established when HTMT values fall below the threshold of 0.85 for conceptually distinct constructs ([52]) or, more leniently, below 0.90 for conceptually related constructs ([39]; [49]). The HTMT ratios were computed directly from the indicator correlations of the retained items (n = 202) and are reported in Table 6.

Table 6. Discriminant validity: Heterotrait-monotrait ratio (HTMT).

As shown in Table 6, all HTMT ratios are below the liberal 0.90 cutoff. These constructs are conceptually adjacent rather than identical, the 0.90 criterion ([39]; [49]) is adopted as the primary benchmark; under this criterion discriminant validity is supported.

For inferential statistical analysis, SmartPLS was used to examine the significance of the theoretical model through bootstrapping with 5000 resamples. The model fit indices showed SRMR = 0.077 (<0.08), NFI = 0.762, dG = 0.69 (<0.95), and dULS = 0.772 (<0.95). These findings indicate that the model achieves both acceptable and significant fit ([48]).

The R² values represent the proportion of variance in the endogenous constructs accounted for by the model. The model explains a considerable portion of variance in SIC (R² = 0.484), SBMI (R² = 0.609), and SOP (R² = 0.621). These values are considered strong, suggesting that the model has high explanatory power.

DS has strong positive effects on both mediators (SIC and SBMI). Both mediators positively and significantly predict SOP, with SBMI demonstrating the larger effect. The indirect effect of DS on SOP via SIC and SBMI is sizeable, because the direct DS → SOP path remains significant and the findings indicate partial mediation, primarily through SBMI. The results demonstrate that DS significantly and positively affects SIC (β = 0.695, p < 0.001), SBMI (β = 0.780, p < 0.001), and SOP (β = 0.293, p < 0.001), with the strongest effect observed on SIC and SBMI.

SBMI significantly and positively impact SOP (β=0.442, p < 0.001), while SIC positively, but insignificantly, affects SOP (β = 0.114, p = 0.053) (Table 7). This implies strong support for H1, H2, H3 and H5, and H4 was not supported, which is in line with existing literature ([42], [43]).

Table 7. Hypothesis testing.

Using the PLS-SEM bootstrapping two-step approach, we tested the indirect effects DS on SOP through SIC and SBMI ([69]).

The indirect effects analysis reveals how the effects of DS on SOP performance are mediated through the mediation constructs SIC and SBMI. Results indicate that the mediation effect of SIC between DS and SOP is not significant (β = 0.079, t = 2.01, p = 0.057), while SBMI demonstrate a highly significant effect (β = 0.355, t = 5.10, p < 0.001). Therefore, the study supports the H6a, while H6b hypothesis is not supported (Figure 3 and Figure 4).

The direct path from DS to SOP remains statistically significant (β = 0.261, t = 3.48, p < 0.001), and partial mediation of with SBMI in the DS-SOP relationship is accounting for roughly 48% of the total effect.

Figure 3. Structural model results.

Figure 4. Relationship path.