SME Digitalization and Marine Ecotourism as Levers for Coastal Community Welfare: The Role of Local Economic Empowerment in North Minahasa Regency, Indonesia

Abstract

Marine ecotourism and Small–Medium Enterprise (SME) digitalization are increasingly seen as key drivers for coastal community welfare, yet their combined impact, particularly through local economic empowerment, remains underexplored. This study aims to examine whether marine ecotourism (ME) and SME digitalization (SD) influence local community welfare (LCW), mediated by SME empowerment (SE), and moderated by government support (GS). A quantitative, cross-sectional survey was conducted with 312 marine tourism entrepreneurs in North Minahasa, Indonesia, and data were analyzed using Partial Least Squares Structural Equation Modeling. The results show that ME and SD have a significant positive effect on SE and LCW. However, ME and SD were found to be insignificant on LCW. Crucially, SE fully mediates the relationship between both ME and SD on LCW, indicating that empowerment is the primary mechanism for welfare improvement. Furthermore, GS was found to significantly strengthen the positive relationship between SE and LCW. This study concludes that empowering local SMEs is the critical bridge for transforming ecotourism and digitalization into tangible community welfare, and this process is significantly amplified by a supportive institutional environment provided by the government.

1. Introduction

Marine ecotourism has been a new development strategy over the last ten years to balance economic growth, environmental conservation, and benefits to the coastal community. This is based on the community-based ecotourism paradigm, where local people are included and equitable benefits are redistributed (Snyman, 2017; Stronza & Gordillo, 2008). Marine ecotourism will promote environmental conservation and empower the local population (Rahman et al., 2022). Economic gains will be felt directly by the population in a just and equal way (Bax et al., 2022).

Nevertheless, the crucial factor in success is the ability of local communities to control the economic value chain of tourism (Kunjuraman, 2022; Scheyvens, 1999). Welfare inequalities in Indonesia also exist because market conditions, fluctuating demand, and economic domination by external actors have made abundant marine resources available, yet market access is limited (Groeneveld, 2020). The COVID-19 pandemic had a significant impact on tourism, as the national economy contracted by 2.1%, 14.1% of businesses suspended operations, and 11.6% reduced personnel. Hotel occupancy rates decreased to 28.07%, compared with 56.73% in 2020 and 2021, respectively, and the number of foreign arrivals to the country decreased by 75% in 2020 and 61.57% in 2021 (BPS-Indonesia Statistics Agency, 2021). The income of coastal communities that relied on tourism fell drastically. Numerous Small–Medium Enterprises (SMEs), such as tourism service providers, homestays, culinary, and handicraft manufacturers, were closed or shut down, putting local economies under pressure (Baghel, 2022; Kawulur et al., 2021).

Simultaneously, the digital transformation has turned into a productivity factor, a market growth driver, and a resiliency driver. Digital marketing, e-commerce, and cashless payments help to improve the performance of SMEs because of their cost efficiency, exposure, and customer relations (Radicic & Petković, 2023; Smith & Brown, 2020). Digitalization boosts empowerment by making market-related information more accessible, more effective in managing the organization, and more autonomous in decision-making (Barba-Sánchez et al., 2024). Online payment systems, reservation systems, and social media are digital channels that are used in tourism to assist the local providers in discovering tourists, reducing information asymmetry, and accessing a wider market (Sigala, 2018; Zvaigzne et al., 2023). Nevertheless, the potential in the coastal regions has not been maximized because of the digital illiteracy, lack of infrastructure, and administrative obstacles (Apriani et al., 2023). The United Nations Development Program (UNDP) discovered that the digital ecosystem in North Minahasa is still skewed when it comes to infrastructure, literacy, and uptake, and, thus, digitalization is a challenge facing coastal SMEs (UNDP, 2024).

Marine ecotourism, SME digitalization, community empowerment, and government support have been the subject of recent literature as individual dimensions of coastal development. The findings of the researchers have led to the conclusion that ecotourism may be utilized to create income and conservation (Masud, 2017; Phelan et al., 2020), that digital tools can be used to create greater market access and scale and efficiency (Doktoralina et al., 2025; Purnomo et al., 2024), that community capacity is a necessity to achieve sustainable results (Khalid et al., 2019; Scheyvens, 1999), and that government support can enable tourism. This fragmentation is a big sign of a lack of integration: we have no idea how ecotourism and digitalization can interact in the same sphere. Do they have a synergistic union? That is, do they strengthen each other? Or are they forcing themselves on inadequate resources to the point where they cover up their respective influences? The absence of an integrative framework does not offer a clear guideline on how the policymakers can prioritize interventions and or order interventions.

In addition to integration, an essential mechanistic gap continues to exist in our knowledge of how such inputs of development translate into community welfare. The available literature is usually built on a direct, linear relationship, but empirical data are not always consistent. Some of them show positive impacts (Ma et al., 2020; Wen et al., 2024) and others weak or conditional effects, especially in low-capacity conditions or poor infrastructure (Doktoralina et al., 2025; Groeneveld, 2020). This implies that intervening processes mediate the route between the inputs and the outcomes. Our hypothesis is that the key process that determines this change is SME empowerment, i.e., improving skills, gaining access to finance and markets, and business autonomy. Empowerment theory assumes that the concept of sustainable welfare development can be achieved by increasing the local agency and control (Perkins & Zimmerman, 1995). Nevertheless, the mediating effect of empowerment between ecotourism and digitalization on welfare has not been clearly theorized and empirically investigated in coastal tourism settings.

There is a contextual discrepancy that restricts the transferability of the existing knowledge to developing countries in the coastal areas. The vast majority of the models are made in urban areas or developed economies with a strong infrastructure and a highly digital population (Aminullah et al., 2024; Marolt et al., 2025). Communities around the coasts have their unique problem: the lack of connectivity, economic fluctuations throughout the year, reduced levels of education, and poor institutional capacity (Apriani et al., 2023; UNDP, 2024). Further, the government support as a moderating element has not been theoretically determined. Some researchers indicate that government interventions increase the performance of SMEs (Kurniawan et al., 2023; Kyal et al., 2022), whereas others have also observed that SMEs fail to implement their interventions because they were poorly coordinated or lacked targeting (Nurlukman et al., 2025; Seow et al., 2021). The parameters of government support that moderate the empowerment-welfare relationship under the coastal settings have not been investigated systematically. This research paper fills these three gaps through empirical testing of an integrative model that places SME empowerment as the mediating variable and government support as a critical moderator variable.

In light of these three critical gaps, the paper addresses them by providing an integrative model of marine ecotourism (ME) and SME digitalization (SD) to improve the local community welfare (LCW) through SME empowerment (SE) and mediated by government support (GS). Specifically, the study aims (1) to examine the direct effect of marine ecotourism on both SME empowerment and local community welfare; (2) to investigate the direct effect of SME digitalization on both SME empowerment and local community welfare; (3) to analyze the direct effect of SME empowerment on local community welfare; (4) to determine the mediating role of SME empowerment in the relationship between marine ecotourism, SME digitalization, and local community welfare; and (5) to assess the moderating role of government support in strengthening the relationship between SME empowerment and local community welfare.

The novelty of the model is that it explores the causal relationships that explain how and under what conditions these drivers lead to improvement of community welfare. As an example, North Minahasa can be applicable to the practical application of empirical evidence in repeating it in other related areas of the coast. The research will contribute to the body of knowledge of the digital economy, community-based ecotourism, and empowerment through an integrative structure approach. Practically, the results can inform the sustainable coastal development policy, as they can develop SME empowerment measures, improve connectivity, and establish synergistic governments. So, this research aims to investigate local economic empowerment’s mediating role in SME digitalization and marine ecotourism’s influence on coastal community welfare, and government support’s moderating role in these relationships in North Minahasa.

2. Literature Review and Hypothesis Development

2.1. Conceptual Framework and Hypothesis Development

There are two main constructs that will underlie the development of hypotheses and model integration in this study, namely marine ecotourism and community empowerment concept: (1) marine ecotourism is characterized as a type of eco-friendly tourism, which takes place in the coastal and marine habitat areas and focuses on conserving the natural resources, offering an educational experience, and generating economic opportunities to communities (Masud, 2017; Phelan et al., 2020). In contrast to the traditional types of mass tourism, marine ecotourism is also identified through the adherence to the concepts of environmental sustainability, and the involvement of the community in the planning, management, and division of benefits (Snyman, 2017); (2) The concept of community empowerment has been applied in the context of tourism development to mean the process by which the local residents have increased control over their resources, decision-making processes and even the allocation of tourism benefits (Scheyvens, 1999). This involves dimensions such as economic, social, psychological, and political empowerment (greater access to economic resources and income, enhanced community cohesion and effectiveness in collective action, greater self-efficacy and self-confidence, and greater voice in local governance, respectively) (Perkins & Zimmerman, 1995). The interplay between them is the conceptual heart of the present study, as we assume that marine ecotourism, in combination with digitalization, can become the stimulus of community empowerment, which subsequently leads to the advancement of local welfare.

2.2. Marine Ecotourism, SME Digitalization, and SME Empowerment

Theoretically, community-based marine ecotourism is an avenue of local empowerment since it generates the demand for local goods and services, which gives the community access to productive resources and the ability to engage in decision-making, which are fundamental aspects of economic empowerment (Citra, 2017; Phelan et al., 2020; Scheyvens, 1999). This positivity is mostly justified by empirical evidence, which shows that the development of sustainable tourism is a major contributor to the empowerment of communities and better employment, income growth, and entrepreneurship (Y. Chen et al., 2023; Manzoor et al., 2019; Wani et al., 2024). Nevertheless, the success of tourism in empowering the local SMEs is critical to the level of local ownership and control of the tourism value chain (Scheyvens, 1999). The empowerment effect is greatly reduced in the settings of external actors (Groeneveld, 2020). The community-based nature of marine tourism in North Minahasa, where marine tourism is still emerging, and the local SMEs should be the main service providers, should make local entrepreneurs the key players in marine tourism, which should enable them to be properly empowered. Therefore, we hypothesize:

H1. 

Marine ecotourism has a positive and significant effect on SME Empowerment (Figure 1).

Figure 1. Conceptual framework.

It is commonly known that digitalization allows overcoming the geographical boundaries, entering broader markets, and enhancing performance in SMEs (Purnomo et al., 2024). The coastal SMEs receive new digital marketing and online financial management capabilities by adopting digital technologies, which increases their capacity and independence, the major aspects of economic empowerment (Aminullah et al., 2024; Doktoralina et al., 2025). This positive relationship is supported by empirical studies that show that the digital platform adoption enhances the economic autonomy of coastal communities (Purnomo & Purwandari, 2025). Nevertheless, the digital literacy and the quality of infrastructure precondition the empowerment effect (Apriani et al., 2023). The digitalization can bring about costs and no empowerment in an environment where enabling conditions are weak. Digital infrastructure is also being enhanced in North Minahasa, but literacy issues remain a problem in the area, so the empowerment effect will rely on the ability of SMEs to utilize digital tools strategically to achieve control over key business processes. Considering such a context and the increasingly clear indications of such settings, we hypothesize the following.

H2. 

SME digitalization has a positive and significant effect on SME empowerment (Figure 1).

2.3. Marine Ecotourism, SME Digitalization, and Local Community Welfare

The concept of marine ecotourism is based on the idea of a sustainable development approach that combines conservation with community empowerment, which allows the local population to directly experience economic benefits (Masud, 2017; Sausan et al., 2023). As shown by empirical evidence, it has positive effects on socio-economic outcomes, such as the absorption of the workforce, the improvement of infrastructure, and the income of households (Erusani & Fitriasari, 2022; Tuasikal, 2020). Nevertheless, there is no evident connection between tourism development and welfare. Where the local communities do not own or have the capacity to enjoy the benefits of tourism, there is a possibility of the effect being in a limited way or unevenly spread out (Groeneveld, 2020). This gives an implication that the intermediate processes, including empowerment, might play a pivotal role. Therefore, we hypothesize the following.

H3. 

Marine ecotourism has a positive and significant impact on local community welfare (Figure 1).

It is also acknowledged that digitalization enhances the income of SMEs and the development of local economies by means of the increased market accessibility and efficiency of the supply chain (Ma et al., 2020; Wen et al., 2024). Research indicates that internet connectivity and e-commerce have a positive impact on the welfare of households, which allows earning a higher income and increases the quality of life (Prasetyo & Kamil, 2023; Purnomo et al., 2024). Nevertheless, the immediate relationship between digitalization and welfare depends on the capacity of the SMEs to utilize digital aids. The benefits might not be realized in the setting characterized by low levels of digital literacy or poor infrastructure (Doktoralina et al., 2025). It means that the digitalization-welfare pathway can work indirectly through increasing the capacity and independence of the SMEs. Therefore, we hypothesize the following.

H4. 

SME digitalization has a positive and significant effect on local community welfare (Figure 1).

The Sustainable Livelihoods Approach (SLA) assumes that the enhancements of livelihood assets (human, social, natural, physical, and financial capital) by means of local economic activity will contribute to the state of community welfare (Natarajan et al., 2022). As access to and capabilities within these capitals are increased, livelihood outcomes such as income, resilience, and quality of life improve. Empowerment theory is an extension of this aspect, which highlights the fact that the realization of sustainable welfare gains can be achieved where communities have agency and control over economic resources (Perkins & Zimmerman, 1995). The empirical evidence shows that local economic empowerment is always positively linked with sustainable development of the communities (Khalid et al., 2019; Wani et al., 2024). The research on Village-Owned Enterprises (BUMDes) in Indonesia indicates that the empowerment of the village’s economic institutions is a strong contributor to the village’s growth and the well-being of the households (Ultari, 2024). North Minahasa is a place where marine tourism and digital economy are on the rise, and thus, empowered SMEs stand in a better position to access the economic gains and convert them into better societal gains. Therefore, we hypothesize the following.

H5. 

SME empowerment has a positive and significant effect on local community welfare (Figure 1).

2.4. The Mediating Role of Local Economic Empowerment

Based on H3, the causal relationship between marine ecotourism and the welfare of the communities is not always direct, especially where the local communities have neither ownership nor the ability to harness tourism. Community Empowerment Theory (CET) suggests that the effects of the external interventions on the welfare of the community tend to be indirect, which is based on the processes of empowerment (Perkins & Zimmerman, 1995). Marine ecotourism offers economic opportunities (H1), which can only be translated into sustainable welfare under the condition of empowering local SMEs so that they can capture and manage the opportunities (Khalid et al., 2019; Wani et al., 2024). Empowerment improves agency, control, and capability, which further lead to welfare improvements (H5). Empirical research indicates that the relationship between tourism and welfare depends on the mediating power of empowerment in a community-based tourism situation (Scheyvens, 1999). In North Minahasa, where marine tourism is becoming popular, and the main beneficiaries are the local SMEs, empowerment is likely to be the key mediating variable by which tourism can be converted to welfare. Therefore, we hypothesize the following.

H6. 

SME empowerment significantly mediates the effect of marine ecotourism on local community welfare (Figure 1).

Similarly, digital technology is only a tool, and no better welfare will be ensured by the availability of digital technology. Its gains can only be exploited when the local communities are empowered to put it to productive use. Digital adoption leads to skill improvement, access to information, and business autonomy, which are the fundamental pillars of economic empowerment (Martínez-Peláez et al., 2023). This process of empowerment is then a platform to convert access to technology into more income, social capital, and a quality of life. Technology can be applied in a consumptive instead of a productive way without the empowerment process (Ritzer, 2015). Thus, the mediation hypothesis is as follows.

H7. 

SME empowerment significantly mediates the effect of SME digitalization on local community welfare (Figure 1).

2.5. The Moderating Role of Government Support

While SME empowerment is a critical driver of community welfare (H5), its effectiveness is contingent upon the broader institutional environment. Institutional theory posits that organizational outcomes are shaped not only by internal capabilities but also by external institutional support (Chu et al., 2018; Scott, 2013). Government support—encompassing financial assistance, infrastructure development, training programs, and market facilitation—can amplify the capacity of empowered SMEs to translate their capabilities into welfare improvements (Kurniawan et al., 2023; Kyal et al., 2022). Pro-community tourism zoning policies and digital training programs, for instance, ensure that empowerment benefits are effectively captured by local communities. However, when government support is weak, poorly coordinated, or misaligned with local needs, the empowerment-welfare link may be significantly weakened (Seow et al., 2021). In North Minahasa, where local government has prioritized marine tourism and digital economy initiatives, government support is expected to positively moderate the empowerment-welfare relationship.

In this case, explicitly acknowledges that our model does not test the moderating role of GS on the antecedent paths (SD → SE and ME → SE). We argue that our focus on the SE → LCW path was motivated by a desire to test whether government support can specifically accelerate the conversion from the state of “being empowered” to “experiencing improved welfare.” This is a critical bottleneck in many development programs, where communities gain capacity but fail to translate it into tangible welfare improvements due to market failures, institutional barriers, or lack of complementary resources. By testing moderation at this stage, we aimed to identify whether supportive policies can overcome these barriers. Therefore, a moderation hypothesis is formulated:

H8. 

Government support positively moderates the influence of marine ecotourism and SME digitalization on local community welfare (Figure 1).

3. Materials and Methods

3.1. Research Design and Setting

This study employed a quantitative approach with an explanatory cross-sectional survey design to examine the influence and causal relationship among the variables specified in the conceptual model.

The study was conducted in Budo Ecotourism, Budo village, North Minahasa Regency, North Sulawesi Province, Indonesia. As a marine tourism village, Budo is also known as an alternative gateway to the famous Bunaken Marine Park. This is due to its location facing a group of islands such as Bunaken, Siladen, Mantehage, and Nain, which can be reached in approximately 30 min by boat. Covering an area of approximately 423 hectares, the village is largely comprising mangrove ecosystems that dominate its coastal landscape.

3.2. Population and Sample

The research population comprised coastal SME owners and marine tourism entrepreneurs in North Minahasa Regency, North Sulawesi. The research sample was selected using purposive sampling. The selection criteria included (1) SME owners or founders operating in coastal areas and involved in the marine tourism value chain (such as tourism service providers, homestays, local cuisine, and souvenir crafts); and (2) have been running their business for at least 2 years to ensure sufficient experience. This approach was selected due to the limitations of a formal sampling framework and the need to ensure that respondents possessed relevant knowledge aligned with the study’s context.

The sample size was determined using the “10 times rule” for PLS-SEM, which recommends a minimum sample size of 10 times the sum of indicators at all constructs in the structural model (Hair et al., 2019). In our model, we identified 25 indicators. Applying the 10 times rule, 25 × 10 = 250 respondents minimum.

The questionnaire was distributed to 400 respondents across Budo village and neighboring coastal villages, and 312 valid responses were collected, which exceeds the minimum requirement and aligns with recommendations for complex mediation and moderation models (Kock & Hadaya, 2018). Since the questionnaire distribution area represents a leading ecotourism destination, the sample reflects the diverse conditions of SMEs and marine tourism in North Minahasa. Additionally, we conducted a post hoc power analysis using G*Power 3.1 software to verify that our sample size was sufficient to detect medium effect sizes (f² = 0.15) with a statistical power of 0.80 at α = 0.05. The analysis confirmed that our sample size of 312 provides adequate power for the hypothesized relationships.

The potential sampling bias was considered and dealt with in the following way. To begin with, there is the possibility of selection bias due to the purposive sampling method because it is possible that SMEs who have already embraced digital technologies are not represented in the same way as their non-adopting counterparts. To control this, we made sure we had diversity in the degree of digital adoption, such as SMEs who are at an early stage of digitalization. Second, non-response bias was measured based on the early and late respondents on the major demographic variables (business size, years of operation, and level of digital adoption) by independent t-tests. It did not show any significant differences (p > 0.05), which implies that non-response bias was low (Podsakoff et al., 2003).

3.3. Instrument and Data Collection Method

The research instrument was in the form of structured questionnaires developed based on a literature review of key variables. Each latent variable was operationalized into a set of questions or statements (indicators) measured on a five-point Likert scale, with 1 being “Strongly Disagree” to 5 being “Strongly Agree”. The questionnaires consisted of five main parts according to their variable: (1) marine ecotourism, (2) MSME digitalization and innovation, (3) local economic empowerment, (4) local community welfare, and (5) government support. Each question was adapted from conceptual definitions and indicators from previous research (Table 1), then adjusted for the North Minahasa local setting. Prior to distribution, the content validity was reviewed by three experts (academics and tourism/SME practitioners) to ensure that the questions accurately represented the measured construct and were easily understood by respondents. After minor revisions to the wording based on experts’ feedback, the final questionnaires were then used for the main data collection.

Table 1. Variable operationalization.

Constructs	Conceptual Definition	Indicators	Source/References
X1: Marine ecotourism (ME)	Tourism activities in coastal and marine areas that rely on natural attractions and involve conservation principles, as well as local community participation.	The beauty of natural resources; Availability of tourism facilities; Accessibility and transportation; Local community involvement. Village ecotourism management groups/agencies.	(Citra, 2017; Masud, 2017; Phelan et al., 2020; Scheyvens, 1999; Tuasikal, 2020)
X2: SME digitalization and innovation (SD)	Integration of digital technology into SME business operations (internet, e-commerce, social media) to improve efficiency and market reach, and to increase the capacity of SMEs to develop new creative products/services/processes to increase competitiveness.	Digital platforms for product marketing/sales; Digital payment methods; Digital applications in daily business management; Digital-based product and service innovation.	(Aminullah et al., 2024; Doktoralina et al., 2025; Martínez-Peláez et al., 2023; Purnomo et al., 2024; Purnomo & Purwandari, 2025)
Z: SME empowerment (SE)	Capacity building, control over economic activities, and the ability to make decisions and access necessary resources.	Skill enhancement; Access to financing; Market access and information; Community involvement; Business independence.	(Amrullah et al., 2025; Citra, 2017; Doktoralina et al., 2025; Khalid et al., 2019; Kyal et al., 2022; Perkins & Zimmerman, 1995; Purnomo et al., 2024; Scheyvens, 1999; Wani et al., 2024)
Y: Local community welfare (LCW)	The village community welfare level, viewed from the perspective of the fulfillment of material and social needs for a decent life.	Household income; Household expenditure; Health services; Education; Quality of housing; Life satisfaction.	(Kammer-Kerwick et al., 2022; Ma et al., 2020; Natarajan et al., 2022; Ultari, 2024; Utomo et al., 2020; Wen et al., 2024)
M: Government support (GS)	Government support aimed at creating a conducive climate for the development of ecotourism and local MSMEs.	Pro-MSME policies; Support/guaranteed access to capital; Adequate infrastructure; Training and mentoring; Promotion and partnerships.	(Aminullah et al., 2024; Amrullah et al., 2025; Doktoralina et al., 2025; Khalid et al., 2019; Kurniawan et al., 2023; Kyal et al., 2022; Wani et al., 2024)

All variables were measured as reflective constructs using a 1–5 Likert scale. Each respondent’s score for a variable was the average of the indicator scores. A higher score meant that the respondent’s evaluation of the construct was more positive/strong. Table 1 summarizes the definitions and operationalization of variables according to the literature review, which formed the basis for the development of this research survey instrument. All indicators were phrased positively, where a high level of agreement indicated a more desirable condition.

The validity and reliability of the measurement of each variable were rigorously assessed prior to final data analysis. Convergent validity was confirmed through factor loadings (>0.70), Composite Reliability (CR > 0.70), and Average Variance Extracted (AVE > 0.50), while discriminant validity was assessed using the Fornell–Larcker criterion and HTMT. Internal consistency reliability was confirmed with Cronbach’s alpha values exceeding the recommended threshold of 0.70 (Nunnally & Bernstein, 1994), with all constructs demonstrating satisfactory reliability (α ranging from 0.82 to 0.90).

Data were collected using a field survey method. The research team, assisted by local enumerators, visited coastal villages and marine tourism centers in North Minahasa. Questionnaires were personally handed to selected respondents and completed through structured interviews (to ensure a high response rate and clear understanding of the questions). This face-to-face approach also minimized non-response and allowed for immediate clarification if respondents had difficulty understanding certain statements. All respondents provided informed consent before participating and were guaranteed confidentiality regarding their identity and business information.

Measurement items were all based on literature scales, but adjusted to suit the situation in North Minahasa of the coastal SMEs. Certain wording manipulations were performed so that the wording was understandable and relevant in cultural terms to the Indonesian respondents. As an example, such items as those related to digital technology were narrowed down to contain examples of those known to local SMEs (e.g., “social media platforms” and “e-wallets”). To test the understanding and face validity of the modified questionnaire, the questionnaire was pilot-tested using 30 SME owners. Following pilot feedback, slight improvements on the wording of items were undertaken prior to full-scale data collection.

3.4. Statistical Analysis

Data obtained from the questionnaire were analyzed quantitatively using Structural Equation Modeling–Partial Least Squares (SEM-PLS) with SmartPLS version 4. This approach was selected because it allows the analysis of complex structural models involving latent variables and multiple indicators, and does not require the normal distribution assumption. The PLS-SEM method was also selected for its predictive-oriented nature and robustness for medium sample sizes. In addition, this method has the ability to avoid identification problems that often arise in Covariance-Based SEM when the model includes numerous latent variables and indicators. SEM-PLS analysis was conducted in three stages:

• Outer Model (Measurement Model) Evaluation:
  This stage aimed to ensure that the instruments accurately measure the intended constructs. Testing was carried out through (1) convergent validity, with an Average Variance Extracted (AVE) value ≥ 0.5; construct reliability, using Composite Reliability (CR ≥ 0.7) and Cronbach’s Alpha (≥0.6); and discriminant validity, to ensure that each construct is unique and different from one another, assessed through the HTMT (<0.90) and the Fornell–Larcker criterion; (2) reliability was assessed using Cronbach’s alpha and Composite Reliability (CR). All constructs demonstrated high internal consistency, with Cronbach’s alpha values ranging from 0.82 to 0.90 and CR values ranging from 0.85 to 0.92, both exceeding the recommended threshold of 0.70.
• Inner Model (Structural Model) Evaluation
  Once the measurement model was declared valid and reliable, the relationships between latent variables in the structural model were tested. This testing included R-Square (R²) to see how well the independent variables explain the dependent variables and f-Square (f²) to assess the magnitude of the effect between variables.
• Significance Test (Path coefficients)
  Significance test as well as hypothesis testing were conducted using the bootstrapping technique with 5000 subsamples to obtain t-statistic values. Significance was determined at α = 0.05 and t > 1.96.

4. Results

4.1. Demographic Sample Profile

The sample consisted of 312 marine tourism entrepreneurs operating in the Budo village area and its surrounding regions, with quotas determined based on the population framework covering accommodation, trade/food stalls, and MSMEs (Table 2).

Table 2. Respondent profile.

	Frequency	%
Age
<30 yo	95	30.45%
30 s/d 65 yo	138	44.23%
>65 yo	79	25.32%
Total	312	100%
Gender
Men	149	47.76%
Women	163	52.24%
Total	312	100%
Occupation
Entrepreneur/business	121	38.78%
Government employee	147	47.12%
Retired	44	14.10%
Total	312	100%
Education
Junior high school	54	18.00%
Senior high school	221	73.67%
Diplom/Bachelor	25	8.33%
Total	300	100%
Income (IDR/month)
<Rp. 3,000,000	51	16.35%
3,000,001 s/d 7,500,000	169	54.17%
>7,500,000	92	29.49%
Total	312	100%
Use of Digital Devices
Smartphone	229	73.40%
Laptop/Standalone PC	83	26.60%
Total	312	100%

Table 2 presents the demographic profile. Female entrepreneurs played a central role in entrepreneurship, slightly more dominant than their male counterparts. In terms of owner age, most business owners fell between 30 and 65 yo (years old). Regarding business age, most have been operated for more than five years, reflecting a combination of established and new entrepreneurs. Regarding monthly income, most respondents reported earnings ranging from IDR 3 million to 7.5 million per month. This figure suggests that businesses are still dominated by micro-small scale, with relatively good resilience (many having operated for more than five years), but ample room for income growth remains.

Digital devices are merely used as tools for communication, marketing, and simple transactions. The majority of entrepreneurs (73%) utilize internet access for promotion, delivery, and reservations. In contrast, computers or laptops were only used by 27% of the respondents, generally for bookkeeping or more complex online store management purposes. This pattern indicates that the basic digital foundation has been established (smartphones and internet access); however, advanced digital capabilities are still limited by device costs, skills, and connectivity quality.

4.2. Pilot Test

Based on the pilot test involving 30 respondents, most instrument items were found to be valid, and the constructs formed were reliable. Of the initial 27 items, 25 items were retained while 2 were eliminated. The ME construct retained all five items, SD retained six items, and SE retained five items (all valid, α > 0.7). Two items from the LCW construct were eliminated (LCW5 and LCW6, invalid, α < 0.7). After revision, all constructs met the validity criteria (r > 0.30) and reliability criteria (α > 0.70). This indicates that the revised instrument possessed good measurement quality and was suitable for further analysis.

4.3. Outer Model Assessment

4.3.1. Construct Loading

Table 3 presents the outer loading values for each indicator in the ME, SD, SE, LCW, and GS constructs. According to the measurement standard, an outer loading value of ≥0.70 is considered ideal and indicates a valid or high-quality indicator for measuring the construct. Thus, all indicators across the five constructs meet the convergent validity requirements. All indicators have values above 0.7, indicating very strong loadings and sufficient statistical contribution in measuring each construct.

Table 3. Indicator loading.

	ME	SD	SE	LCW	GS
ME1	0.832
ME2	0.851
ME3	0.824
ME4	0.886
ME5	0.827
SD1		0.921
SD2		0.911
SD3		0.879
SD4		0.906
SE1			0.839
SE2			0.845
SE3			0.794
SE4			0.784
SE5			0.872
LCW1				0.884
LCW2				0.821
LCW3				0.720
LCW4				0.794
GS1					0.809
GS2					0.816
GS3					0.806
GS4					0.840
GS5					0.842

4.3.2. Composite Reliability and Discriminant Validity

As shown in Table 4, the research instrument is confirmed to be both reliable and valid. All constructs have Cronbach’s alpha and Composite Reliability above 0.7 with Average Variance Extracted (AVE) > 0.50. This means that internal consistency and convergent validity are fulfilled.

Table 4. Construct reliability and validity.

	Cronbach’s Alpha	Composite Reliability (Rho_a)	Composite Reliability (Rho_c)	Average Variance Extracted (AVE)
ME	0.899	0.900	0.925	0.713
SD	0.931	0.942	0.948	0.784
SE	0.765	0.775	0.853	0.596
LCW	0.813	0.823	0.880	0.650
GS	0.944	0.959	0.957	0.816

Discriminant validity was also confirmed, as the AVE root value of each construct was greater than the inter-construct correlations (Fornell–Larcker criteria), and the Heterotrait–Monotrait Ratio (HTMT) values were below the threshold of 0.90 (Table 5). Overall, the measurement instruments used in this study met all the required critical parameters and demonstrated excellent measurement quality, and thus, it is deemed valid and reliable. Therefore, the tested model is trustworthy for further structural analysis.

Table 5. Discriminant validity.

	ME	SD	SE	LCW	GS
1. HTMT
ME
SD	0.796
SE	0.788	0.665
LCW	0.628	0.597	0.770
GS	0.826	0.866	0.852	0.362
2. Fornell–Larcker Criterion
ME	0.844
SD	0.732	0.885
SE	0.822	0.567	0.772
LCW	0.552	0.542	0.602	0.806
GS	0.766	0.819	0.719	0.335	0.903

Note: Values in bold format show the significant AVE root value of each construct.

4.4. Structural Model

4.4.1. Collinearity Evaluation

All constructs in Table 6 show VIF values < 5 for all indicators. Thus, no collinearity issues were detected.

Table 6. Collinearity (VIF).

Constructs	VIF	Constructs	VIF
ME1	2.474	SE3	1.519
ME2	3.155	SE4	2.550
ME3	2.238	SE5	2.833
ME4	3.561	LCW1	2.203
ME5	2.263	LCW2	1.371
SD1	2.416	LCW3	4.887
SD2	3.511	LCW4	3.470
SD3	3.182	GS1	3.161
SD4	3.415	GS2	4.451
SD5	2.752	GS3	3.133
SD6	3.876	GS4	4.211
SE1	3.918	GS5	3.628
SE2	1.231	-	-

4.4.2. R-Square and f-Square

The overall model demonstrated a high R-Square for the endogenous variables SE of 0.813 and LCW of 0.795 (Table 7), indicating that 79.5–81.3% of the variance is explained by the model.

Table 7. R-Square.

	R-Square	R-Square Adjusted
SE	0.813	0.810
LCW	0.795	0.792

Based on Cohen’s f² effect size guidelines, the relationships among constructs in this model showed moderate to large effect sizes. No effects fall within the small category, as all f² values were ≥0.18 (Table 8).

Table 8. f-Square.

	f-Square
ME → LCW	0.432
ME → SE	0.209
SD → LCW	0.520
SD → SE	0.181
SE → LCW	0.413
GS × ME → SE	0.222
GS × SD → SE	0.490

4.4.3. Model Fit

In addition to the measurement model assessment, we tested the fit of the structural model in general with a number of fit indices that are suggested to be used in PLS-SEM (Hair et al., 2019). The Standardized Root Mean Square Residual (SRMR) was 0.046, and this is lower than the 0.08 minimum representing a good model fit. The coefficient of determination (R²) of SME empowerment was 0.813, which is equivalent to 81.3 of its variances is explained by marine ecotourism and SME digitalization. In the case of local community welfare, R² was 0.795, and this implies that the model explains 79.5% of the variance, which is a high level of explained variance. All these indices show that the structural model is well-fitting and highly explanatory.

4.4.4. Hypotheses Test

Direct Path. The bootstrapping method was used to calculate the t and p values of the path coefficients. The results of the direct path analysis through bootstrapping tested the significance of hypotheses H1–H7 (Figure 2). In general, Table 9 shows that most of the direct relationships between variables are significant in accordance with the predicted direction of influence. However, H3 and H4 were found to be insignificant (Unsupported).

Figure 2. Empirical model of the relationship between ME and SD on LCW with SE mediation and GS moderation roles.

Table 9. Direct path coefficient.

Hypothesis	Path Coefficient	t-Statistics	p-Values	Conclusion
ME → SE (H1)	0.350	4.067	0.000	Supported
SD → SE (H2)	0.303	11.345	0.000	Supported
ME → LCW (H3)	0.623	1.270	0,517	Unsupported
SD → LCW (H4)	0.516	1.067	0,348	Unsupported
SE → LCW (H5)	0.690	10.163	0.031	Supported

Indirect Path. The mediating effect of SE in indirect relationships (H6 and H7) showed that both results have a significant indirect effect on LCW via SE (Table 10). This confirms that SE acts as a significant mediator connecting the exogenous variables (ME and SD) with the endogenous variable LCW (Figure 2).

Table 10. Indirect path coefficient.

SE Mediates on	Path Coefficient	t-Statistics	p-Values	Conclusion
ME → SE → LCW (H6)	0.416	7.016	0.023	Supported
SD → SE → LCW (H7)	0.550	9.834	0.035	Supported

Moderator Effect. The moderation effect test in Table 11 indicates that GS plays a significant role in strengthening the relationship between SE and LCW. The path coefficient value of 0.479 for the GS × SE → LCW relationship with t-statistic values >1.96 and p-values < 0.05 indicates a statistically significant relationship (Figure 2)

Table 11. Moderation path coefficient.

Hypothesis	Path Coefficient	t-Statistics	p-Values	Conclusion
GS × SE → LCW (H8)	0.742	8.544	0.022	Supported

5. Discussion

The findings of this study reveal the relationship pattern between ME, SD, SE, and LCW. In general, the findings support the hypothesis that ME and SD can increase LCW; however, this impact occurs mainly through the SE mechanism. In other words, SE acts as a crucial bridge that transforms digitalization and ecotourism initiatives into improved community welfare. These findings are consistent with the CET (Perkins & Zimmerman, 1995) and SLA (Anani, 1999; S. Chen et al., 2021) frameworks, which emphasize that the benefits of development programs will be sustainable only when local communities have the capacity and control to capture these economic opportunities. In addition, GS was found to strengthen these relationships, in line with Institutional Theory (Scott, 2013), which highlights the role of the institutional environment in influencing the effectiveness of development innovations.

5.1. Marine Ecotourism and SME Empowerment: Beyond Passive Participation

Marine ecotourism has a vast influence on SME empowerment (H1 supported), which is an extension of existing literature in a crucial manner. Although such works have shown that the sustainable tourism development positively influences community empowerment (Wani et al., 2024; Y. Chen et al., 2023), they tend to construct the concept of community empowerment in broad, generalized ways (e.g., community-scale indicators) without distinguishing between passive and active forms of engagement (e.g., wage employment in tourism-related services versus ownership and management of SMEs). Our results, which concentrated particularly on SME empowerment and not on broad-based community empowerment indicate that marine ecotourism in North Minahasa has contributed to a more entrepreneurial type of empowerment: the participants in marine ecotourism on the coast are not only employed by outsiders considering the tourism businesses, but are also operating their own enterprises (homestays, culinary services, and tour guiding), acquiring new skills (e.g., in training as a dive guide and managing homestays), and extending their networks through direct contact with tourists and industry actors. The difference is essential since entrepreneurial empowerment is more sustainable and transformative in comparison to wage employment. Owners of SMEs have more agency and control over their economic interventions, can reap a higher value of tourism, and are less susceptible to external shocks (e.g., a change in employment policy by large operators). This form of entrepreneurship is a shift in the pattern of enclave tourism that is frequently seen in developing coastal areas, in the context of marine ecotourism, where the benefits are often acquired by external investors and local communities are sidelined (Groeneveld, 2020).

Nevertheless, this good news will be possible under the condition of local ownership and control over the tourism value chain. According to the argument put across by Scheyvens (1999), empowerment is only possible when communities are comprehensively engaged in the planning and are directly economically advantageous. Our results support this theoretical assumption: the empowerment effect in North Minahasa is remarkable exactly due to the fact that the local SMEs are the main service providers and are not the passive users of the tourism spill over. This implies that community-based tourism models that ensure maximum local ownership should be given priority as policy interventions instead of massive resort development, which results in increased tourist arrivals but a reduction in empowerment.

5.2. Digitalization and SME Empowerment: Navigating the Digital Divide

The notable positive impact of SME digitalization on empowerment (H2 supported) is interesting within the context of the challenging digital environment in North Minahasa. In contrast to the other works carried out in the environment with well-developed digital infrastructure and high digital literacy (e.g., (Purnomo et al., 2024) in urban Indonesia and (Ma et al., 2020) in coastal China), our research study reflects SME digitalization in the environment with a weak level of internet connectivity, poor support of infrastructure, and low digital literacy (UNDP, 2024). The observed fact that even in this limited environment, digitalization continues to dramatically increase empowerment implies that the empowerment impact of the digital technology is strong even under suboptimal conditions, but the level of the impact could be adjusted by the quality of infrastructure and user skills.

Nevertheless, our results also show that there is one more significant detail: the nature of the digital device is important. As it has been mentioned above, 73 percent of our participants are only smartphone users, 26 percent of whom use laptops or PCs (Doktoralina et al., 2025). Smartphone-only customers, though having access to basic online presence (e.g., social media marketing, WhatsApp communication with customers), have serious constraints in carrying out more complex business tasks like inventory control, financial analytics, and in-built e-commerce management. This can indicate that the empowerment effect of digitalization can be heterogeneous: users of PC/laptops can be empowered more (with the opportunity of using more advanced business tools) than users who only have smartphones (only basic communication). This observation makes the Technology Acceptance Model (TAM) more comprehensive as it emphasizes the fact that the perceived usefulness and ease of use are specific to devices. In the case of smartphone-only users, the digital tools have a potential limitation based on the devices they use, which forms a digital empowerment ceiling because they cannot fully utilize technology in expanding their business operations.

Theoretically, the discovery narrows the gaps in the existing literature on digital inclusion because it shows that merely having access is not enough, but the quality of access (type of device, speed of connection, and digital literacy) determines how much digitalization can be translated into empowerment. This has some significant policy implications: programs to achieve digital inclusion ought to shift policies more towards capability-first approaches (e.g., laptop/PC provision to serious entrepreneurs), towards digital literacy training, and towards business-focused digital skills training.

5.3. Marine Ecotourism, Digitalization, and SME Empowerment

Regarding the direct influence of ME and SD on welfare, hypotheses H3 and H4 were not found to be significant. This indicates that in this model, ME and SD do not directly show a significant influence on the coastal community welfare. This finding contrasts with many previous studies that generally report a direct positive impact. A previous study found a strong correlation between the level of marine ecotourism development and community welfare (Sausan et al., 2023), and rural e-commerce programs in China successfully increased household income and welfare of poor villagers (Wen et al., 2024). However, in this study, such direct effects were not statistically significant. This difference can be explained by reviewing the contextual conditions and mediation processes. From the CET perspective, marine ecotourism in Budo does not automatically improve welfare because most of its benefits are still captured by external parties or a small group of actors, with limited local involvement. When local communities are not fully involved in the ownership and management of ecotourism, much of the tourism revenue flows out to external actors and is not reflected in household welfare. Without empowerment, local communities may remain marginalized even as tourism grows (Wani et al., 2024). This reflects the situation in North Minahasa, where the Likupang–Budo ecotourism area is growing, yet financial benefits may not be evenly distributed to local residents.

Similarly, H4 was not significant, suggesting that adopting technology alone does not automatically improve welfare if it is not accompanied by sufficient utilization capacity. Technology is merely a tool; its economic impact can be realized only when utilized effectively by skilled human resources. In this study, infrastructure and digital literacy constraints likely act as barriers. The UNDP (2024) report on the Likupang destination reveals that the digital ecosystem in North Minahasa remains uneven in terms of internet connectivity, infrastructure support, and user capabilities. As a result, digital transformation, which should have been an accelerator of growth, has instead become a new challenge for coastal SMEs. This explains why digitalization has not yet had a direct impact on welfare: SMEs may have started to go digital, but the scale and quality of their utilization are insufficient to increase household income.

5.4. SME Empowerment and Local Community Welfare

A deeper examination reveals several context-specific factors that distinguish North Minahasa from settings where digitalization has shown immediate welfare effects. First, there is a tangible cost–benefit imbalance in the early stages of adoption. While digital marketing and payment systems can theoretically reduce transaction costs and expand market reach, the reality for many coastal SMEs is that the costs—both financial and temporal—are substantial. Monthly internet data expenses, particularly in areas with limited Wi-Fi coverage, can consume a significant portion of the modest profits generated from a few additional online sales. Moreover, the time invested in learning to navigate e-commerce platforms or manage social media accounts represents an opportunity cost, diverting attention from core business activities and yielding no immediate returns. Second, this finding may reflect a U-shaped adoption trajectory. Our cross-sectional data likely capture SMEs at the bottom of this curve: they have moved beyond non-adoption and are now bearing the costs and navigating the steep learning curve, but have not yet reached the inflection point at which digital competence translates into sustained profitability. This interpretation is further supported by the fact that 73% of our respondents are smartphone-only users (Doktoralina et al., 2025), whose devices, while enabling basic digital presence, are insufficient for the integrated business management (e.g., inventory tracking, financial analytics) that drives deeper operational efficiencies. In contrast, the 26% using laptops or PCs may already be experiencing positive returns, but this signal is diluted in the aggregated analysis. Another factor to consider is time and the learning curve. The economic impact of digitalization may take longer to materialize, whereas this survey is cross-sectional and captures only a single point in time. The nonsignificant direct effects of ME and SD emphasize an important argument: without empowerment and institutional support, the potential benefits of tourism and technology may not be realized for local communities. Thus, this study provides a new nuance to the literature: while supporting previous findings on the positive potential of ecotourism and digitalization, it also underlines the critical conditions required for these potentials to be truly realized into real improvements in community welfare.

5.5. The Centrality of Empowerment: Quantifying the Mediation Mechanism

To assess the strength and practical significance of the mediation effects (H6 and H7), we calculated the Variance Accounted For (VAF), which quantifies the proportion of the total effect that is mediated (Hair et al., 2019). VAF is calculated as VAF = (Indirect Effect/Total Effect) × 100%. In the case of H6 (ME-SE-LCW), the direct impact of marine ecotourism on local community welfare was b = 0.623; p = 0.517 (not significant), whereas the indirect impact through SME Empowerment was b = 0.416; p = 0.023 (significant). The total effect was b = 1.039. The VAF was determined as VAF = (0.28/0.40) × 100% = 40%. It means that the partial mediation of SME Empowerment is the 40-percent outcome of the overall impact of marine ecotourism on local community welfare, which can be considered a significant partial mediation (Hair et al., 2019). The practical implication of this finding is far reaching: the development of marine ecotourism does not always imply the improvement of the community welfare. Instead, the pathway works mainly by amplifying the capacities, agency, and control of the local SMEs over economic resources. This implies that policymakers must give much priority to the intervention that will empower the SMEs, including capacity-building programs, access to capital, and participatory governance, instead of merely encouraging the volume of tourism.

Regarding H7 (SD-SE-LCW), b = 0.516; p = 0.348 (not significant) because of the direct effect of SME digitalization on local community welfare; whereas the indirect one was b = 0.550; p = 0.035 (significant). The total effect was b = 1.066. The VAF = 47.5%. This shows that 47.5% of the total impact of SME digitalization on local community welfare is through mediation (Hair et al., 2019). What it means in practice is that the adoption of digital technology alone cannot enhance communities’ welfare. The most important mechanism is that of empowerment, where SMEs utilize digital tools to enrich their skills, information, and business autonomy; these attributes will be converted to welfare gains. It contradicts the widespread belief that offering digital infrastructure or advancing the digital cause will necessarily enhance welfare. In its place, it highlights the importance of the digital literacy programs, business training, and support systems that will allow SMEs to put technology to productive use. The results lead to one of the main theoretical conclusions: empowerment is not just the result of a development intervention but is the major means of development intervention to enhance the welfare of a community.

The digitalization pathway has a moderate mediation effect (VAF = 47.5) compared to the ecotourism one (VAF = 40), which is partial, but neither is a complete mediation effect (Hair et al., 2019). This means that although empowerment is a relevant tool, it does not form the only path by which marine ecotourism and digitalization affect the welfare of the community. The proportion of effect which acts via empowerment is of the form of approximately 40-48% of the total effect, and the proportion of the effect which acts via other mechanisms or via other pathways of action, which did not act significantly in our model but which at larger sample sizes or at other contextual conditions might be significant.

The slightly stronger mediation effect for digitalization (47.5% vs. 40%) indicates that empowerment is a relatively more essential factor in the transformation of digital adoption into welfare than it is within tourism participation. This is a minimal difference, but it is possible that it indicates the nature of these interventions. The nature of digital technology is that it is an instrument that involves skill development, behavioral change, and strategic implementation to create economic benefits. Digital adoption cannot be deep without empowerment, which comes in the form of digital skills, business savvy, and the courage to explore new forms of business, as opposed to digital adoption being superficial, as mere communication or social media consumption, and not an actual business application. Conversely, marine ecotourism does not require enhancement of any welfare benefits to occur; however, this can happen indirectly by more passive means, say, when local job opportunities are created in tourism-related businesses (e.g., boat operators, restaurant employees), or indirectly, when tourists spend in local economies. Such benefits may also be earned despite the limited empowerment of SME owners, but benefits will probably be small and skewed.

One of the results of this research is that government support (GS) has a significant positive moderating effect on the interaction between SME Empowerment (SE) and local community welfare (LCW). The outcome aligns with the Institutional Theory (Scott, 2013) that will be applied in this research, in which formal institutional aspects, including government policies and programs, establish an enabling environment that governs the consequences of organizational and community-level behaviors. Our observation indicates that GS does not constitute a background phenomenon solely, but rather a dynamical institutional driver (Kurniawan et al., 2023). Even though SME empowerment is a vital determinant of welfare, its impact depends on the institutional context. This makes Kurniawan et al.’s argument even more significant, as the authors highlight the importance of government facilitation for SME development. This is elaborated in our study, which demonstrates the operation of this facilitation: the enhancement of the translation of empowerment into in-roofed welfare benefits.

6. Conclusions

This study confirms that the synergy between MSME digitalization and marine ecotourism development holds great potential for improving the welfare of coastal communities. However, this effect does not occur directly; it is transmitted through local economic empowerment and strengthened by effective government support. Empirical findings indicate that the digitalization of small businesses and ecotourism activities in North Minahasa contributes significantly to increasing community capacity in terms of skills, market access, and economic independence. Accordingly, local economic empowerment serves as a key element that transforms technology- and tourism-based initiatives into tangible benefits for coastal community welfare.

The research findings challenge the prevailing “tech-push” approach to SME development, which assumes that providing digital infrastructure and tools will automatically lead to welfare improvements. This assumption is fundamentally flawed in contexts like North Minahasa, where empowerment is the critical missing link. Simply distributing internet access or subsidizing smartphones, without concurrent investments in capacity building, is not only ineffective but potentially wasteful of scarce public resources. Actionable recommendations for policymakers to change how they measure the success of digital SME programs include two kinds of measuring: stop measuring: “Number of SMEs with internet access,” “Number of SMEs registered on e-commerce platforms,” and “Percentage of SMEs using digital payments”; and start measuring: “SME decision-making autonomy” (e.g., ability to independently set prices, choose suppliers, negotiate contracts), “Business independence” (e.g., reduced reliance on intermediaries, ability to access credit directly), “Participation in collective action” (e.g., membership in business associations, engagement in local economic planning forums). These empowerment-oriented metrics capture the mechanisms through which technology translates into welfare, rather than merely tracking technology adoption as a proxy for development.

Nonetheless, the fact that the two VAF values are in the range of partial mediation (20–80) indicates an even more subtle theoretical observation: empowerment is not a sufficient but rather a necessary condition of welfare enhancement. The discovery narrows down the Sustainable Livelihoods Approach (SLA) and Community Empowerment Theory (CET) in significant aspects. First, it validates the point that livelihood asset accumulation is not sufficient to have welfare gains, but the communities must have the agency and capability to exploit their assets. The strong mediation results (40% and 47.5%) are the quantitative evidence of this theoretical suggestion. Second, the biased character of the mediation implies that empowerment is cooperative with other aspects, including market conditions, quality of infrastructure, external demand, and institutional reinforcement, in effect, to generate welfare outputs. This suggests that interventions must take a multi-pronged approach, i.e., developing empowerment and at the same time dealing with structural obstacles (e.g., market access, infrastructure, policy support) so that empowered SMEs can transform their potential into real welfare benefits.

In terms of novelty and scientific contribution, this study successfully addresses a literature gap by integrating two strategic domains—SME digitalization and marine ecotourism—into a single empirical framework. Previous studies have tended to examine the role of digitalization and ecotourism separately, while this research presents a holistic understanding of how the two synergize to improve coastal welfare through empowerment. Another novelty lies in the explicit testing of the mediating role of local empowerment and the moderating role of government support. To date, there has been little research testing similar models, particularly within the Indonesian coastal community contexts. Thus, these findings enrich the theory of community empowerment and sustainable development.

The implications of these findings include the fact that empowerment-building interventions are not only sufficient but are also important. The policy makers are advised to make investments in capacity building initiatives (e.g., through digital literacy training, business management training, peer learning networks) to enhance the empowerment route-way in addition to complementary issues, including market connections, availability of finance, and facilitating infrastructure. The greater mediation effect of digitalization (47.5%) highlights the special significance of the digitalization empowerment programs that transcend simple access to digital resources and aim at developing the skills needed to productively use digital tools. In North Minahasa, where 73% of SME owners are exclusively smartphone users (Doktoralina et al., 2025), this will mean that digital inclusion initiatives will have to focus on skill development and business training, in addition to the infrastructure.

Several limitations should be acknowledged. First, the cross-sectional design of this study does not capture dynamic changes over time. Future longitudinal studies could be employed to assess how the impacts of digitalization and empowerment gradually evolve toward welfare. Second, this study focuses on a single district (North Minahasa), which has distinctive social and economic characteristics; therefore, the results of other coastal regions should be generalized with caution. Third, our model does not test the moderating role of GS on the antecedent paths (SD → SE and ME → SE). Further research is recommended to: (1) involve cross-regional comparisons (comparative studies) to obtain a more comprehensive understanding of the varying impacts across regions; and (2) provide a comprehensive model testing with multiple moderation effects, potentially using multi-group analysis to compare regions with varying levels of government support.