Green entrepreneurship involves identifying and capitalising on economic opportunities arising from environmental market failures (Dean & McMullen, 2007). It requires a proactive, risk-taking approach to promoting green products or services (Awuni et al., 2025), reflected in practical activities through a green entrepreneurial mindset (Jiang et al., 2018). This orientation aims for a balance between economic, environmental and social development, treating environmental factors as business opportunities. Business model innovation, a source of competitive advantage, modifies existing business models using new technologies and innovative thinking, enhancing efficiency through complementarity (Amit & Zott, 2012).

From an RBV perspective, GEO can be viewed as a strategic, difficult-to-imitate organisational resource that guides how firms allocate attention and resources toward environmental opportunities. However, GEO does not automatically translate into BMI unless it triggers changes in how firms design value creation and value capture. Specifically, GEO can foster: (1) opportunity framing (treating environmental constraints as market opportunities), (2) experimentation and risk acceptance, and (3) stakeholder-oriented value propositions, all of which are central to reconfiguring business model elements.

Green entrepreneurial orientation, marked by innovation, foresight and risk-taking (Jiang et al., 2018), provides a foundation for sustainable business development. On the one hand, companies that embrace a GEO demonstrate a spirit of continuous innovation, integrating green innovation into their company culture and strategy. For instance, Pacheco et al. (2010) suggest that companies introducing green products, services, or technologies to meet customer needs are more likely to explore and capture new markets by adopting a proactive approach to pursuing green growth. Fernando and Wah (2017) emphasise that green entrepreneurship considers the needs of nature, the economy and society, prioritising environmental protection and efficiency whilst engaging in green innovation practices for sustainable development. On the other hand, green entrepreneurs are also willing to undertake riskier ventures and pursue more business opportunities compared to their counterparts in traditional sectors. Chen and Chang (2013) argue that a GEO places higher ecological demands on companies, encouraging the search for low-pollution, low-waste approaches to conducting activities and creating green products, processes and services to address environmental challenges. Building on this, Shirokova et al. (2016) suggest that green entrepreneurship-oriented companies are more receptive to risk-taking strategies and maintain an optimistic outlook even when investing in uncertain projects. Alwakid et al. (2021) propose that the green entrepreneurship sector values eco-efficiency, enabling companies to pursue both economic and ecological impacts, leading to BMI. In resource-constrained start-ups, such orientation is especially relevant because BMI often represents a ‘leapfrogging’ route to compete without relying on scale advantages. Therefore, we propose the following hypothesis:

The mismatch between internal resources and external needs can hinder a company’s development, preventing it from maintaining a competitive advantage if it relies solely on its capabilities. However, the development of digital technology has helped companies open their organisational boundaries and obtain resources from outside sources (Yang et al., 2021b). Rosenkopf and Nerkar (2001) first introduced the concept of BSS, which is a process through which companies search for heterogeneous knowledge across existing organisational boundaries to solve innovation problems. From a strategic management perspective, BSS is defined as the process of capturing heterogeneous knowledge from external networks and reconstructing existing knowledge systems to overcome path dependency in a dynamic environment. In this context, companies must navigate the challenges of technological innovation and market evolution.

Organisational search theory suggests that when firms face uncertainty and novelty demands, they engage in search beyond local knowledge domains to access non-redundant ideas (Kruger & Steyn, 2024). For BMI, such a search is particularly important because business model design requires recombining technological, market, regulatory and stakeholder knowledge that often resides outside the focal firm. At the same time, BSS entails costs, and the benefits of search depend on the firm’s ability to absorb and recombine what it finds. This benefit–cost tension motivates our later focus on BDC as a key boundary condition.

In turbulent environments, firms confront uncertain development landscapes where their growth potential is severely tested. Relying solely on existing knowledge hinders effective responses to external changes. Boundary-spanning search enables companies to acquire commercially valuable knowledge, enhancing their ability to learn from external sources and becoming a vital competitive advantage (Sidhu et al., 2007). Corredoira and Banerjee (2015) confirm that BSSs foster innovative development by increasing technology patent output. Tippmann et al. (2017) argue that BSS offers innovative solutions for optimising internal resources, serving as a bridge between knowledge transfer and inventive solutions. By crossing organisational boundaries, companies can expedite business model design, access external knowledge, explore environmental opportunities and develop new network values and social relationships (Lee & Huh, 2016). Schweisfurth and Raasch (2018) emphasise the pivotal role of both internal and external knowledge resources, as well as internal and external market opportunities, in a firm’s innovation process. Firms should proactively search for and integrate external resources to accelerate innovation through knowledge exchange. Consequently, BSS becomes a strategic choice for enterprises to gain a competitive edge, enhancing their ability to detect and understand the external environment, compensating for resource deficiencies and fostering BMI. Therefore, we propose the following hypothesis:

Rooted in organisational search theory, BSS is a vital approach for companies to acquire external knowledge and enhance their existing organisational knowledge structure (Leone et al., 2022). Resource-based theory posits that an organisation’s competitive advantage stems from its possession of diverse resources and relationships (Varadarajan, 2023). In the context of green entrepreneurship, companies need varied knowledge resources to develop eco-friendly products and services. Given the focus on green entrepreneurship, internal resources often fall short of these demands. Hence, companies employ BSS to connect with key resource owners, compensating for deficiencies and enhancing operational efficiency (Schierjott et al., 2018). Limited internal resources in SMEs hinder the adaptation to changing technology and market demands for green start-ups. Hence, exploring beyond technology or market boundaries is essential to find valuable external resources (Shafique et al., 2021). Boundary-spanning integration allows the incorporation of external information and user groups, fostering synergies for market expansion and product development. With the influence of Internet technology, boundary-spanning integration has become crucial for corporate growth.

Beyond ‘resource shortage’, GEO also shapes the motivation and scope of search. Specifically, GEO sensitises firms to environmental opportunities and stakeholder or regulatory expectations, increasing the likelihood of problem-driven and opportunity-driven search across organisational boundaries. Because green opportunities often span technologies, supply chains, community actors and regulatory domains, GEO-oriented start-ups have stronger incentives to seek heterogeneous partners and knowledge sources outside their current network.

Additionally, in response to potential green business opportunities, green entrepreneurship-oriented companies tend to act faster than their competitors to gain a competitive edge, with resources being a key element in gaining that advantage. Green entrepreneurial orientation represents an innovative and pioneering behaviour that prioritises the acquisition of complementary or alternative knowledge from external sources. These diverse and heterogeneous resources can support business growth (Sofka & Grimpe, 2010). Singh et al. (2022) suggest that green entrepreneurship is more likely to seize green business opportunities and requires a keen awareness of the external environment and information exchange between different organisations, effectively enhancing businesses’ adaptive capacity. Building upon the work of Makhloufi et al. (2022), GEO is believed to have a behavioural tendency that influences firms’ access to resources from external sources. Therefore, we propose the following hypothesis:

Compared to traditional enterprises, green entrepreneurship faces more pronounced resource bottleneck problems. As companies increasingly rely on external resources and foster closer collaborative relationships with innovation agents, the ability to access information becomes a crucial factor in determining the success or failure of an enterprise’s innovation. Thus, intensifying market exploration and actively assimilating and exploiting potential external market opportunities provide companies with a sustainable impetus for innovating their business models. Wu and Shanley (2009) argue that a GEO possesses both social and environmental attributes, enabling companies to discover new combinations of knowledge, thereby providing a valuable knowledge base for green innovation and facilitating the creation of green value. Inkinen (2016) affirms that a GEO helps companies clarify their direction and search strategies, facilitating the acquisition of new knowledge and ultimately improving innovation performance. Moreover, a GEO drives firms to seek environmentally relevant knowledge and resources (e.g. environmental standards, green technologies, etc.) that enhance their understanding of government regulations for the environment and consumer demand for green products, thereby enhancing their ability to innovate green business models. Thus, green entrepreneurship orientation motivates firms to leverage external knowledge and novel technologies to innovate business models that generate value for the firm.

Conceptually, GEO reflects an orientation and commitment, but it does not guarantee that a start-up already possesses the complementary knowledge required for BMI. Boundary-spanning search, therefore, serves as the central mechanism that converts GEO into new resource combinations: by sourcing heterogeneous external knowledge, firms can redesign value propositions and reconfigure value creation and capture activities. This mechanism also helps explain why some green-oriented start-ups successfully innovate their business models whilst others fail – differences in search intensity and effectiveness translate the same orientation into different innovation outcomes. Therefore, we propose the following hypothesis:

Big data refers to extensive volumes of structured and unstructured data accessible in real-time (Ferraris et al., 2019). Its true business value emerges when organisations elevate big data resources to the level of big data capabilities (Wang et al., 2018). Hence, accurately identifying significant data resources becomes crucial. These identified resources can then undergo in-depth mining and analysis, ultimately facilitating BMI. Big data capability encompasses processing, analysing and visualising digital resources. This empowers organisations to create data-driven plans, make informed decisions and conduct business operations (Dubey et al., 2019).

Consistent with an information-processing view, BDC reflects an organisational capability system that enables firms to acquire, integrate, analyse and transform data into actionable insights for decision-making and reconfiguration. This capability is especially important when BSS generates large volumes of heterogeneous information, which can otherwise overwhelm managerial attention and slow down integration.

Compared to traditional information management and analysis, big data capabilities open new horizons for companies. They facilitate the use of big data analytics to tap into a broader array of external resources, enabling companies to offer high-value products and services and transform their business models. Barrett et al. (2015) argue that big data capabilities accelerate the process of knowledge awareness, uptake and utilisation, playing a vital role in integrating external knowledge and designing business models. Similar views are expressed by Santoro et al. (2019), who emphasise that the retail industry can undergo business model shifts by leveraging big data or employing innovative methods to access data, information and knowledge. Sorescu (2017) underscores that big data capabilities capture changes in external market conditions, providing real-time insights for businesses. Mikalef et al. (2019) suggest that big data resources and technologies help companies acquire new market knowledge, aiding the development of value propositions that enhance customer experience and significantly contribute to BMI. Consequently, BDC has become a crucial strategic resource. It enables enterprises to identify, integrate and utilise external resources effectively, discover new market opportunities, transform operational processes and drive BMI. In other words, when BDC is high, firms are better able to filter noisy signals, identify valuable patterns across external sources and recombine knowledge into coherent design choices for BMI. Therefore, the positive effect of BSS on BMI should be stronger for firms with higher BDC. Therefore, we propose the following hypothesis:

Combining the above analysis, the conceptual model of this study is proposed, as shown in Figure 1.

This study employs a questionnaire survey method to collect data on GEO, BSS, BDC and BMI (Appendix 1). To test the hypotheses, start-ups from four regions in China – Shandong Province, Jiangsu Province, Fujian Province and Zhejiang Province – were selected. The primary respondents were founders, co-founders and key members of entrepreneurial teams. Drawing on Baum et al. (2011), firms less than 4 years old were defined as start-ups and included in the sample. Data collection was conducted from January 2022 to June 2022, primarily using online survey platforms and on-site distribution of paper questionnaires. To reduce common method bias, participation was voluntary and anonymous; respondents were assured there were no right or wrong answers; items were mixed and clearly separated by construct, and established; clearly worded scales were used and adapted to the start-up context.

A total of 483 questionnaires were distributed, resulting in the collection of 376 questionnaires. After eliminating incomplete questionnaires and those with evident errors in responses, 307 valid questionnaires remained, achieving a valid recovery rate of 63.56%. The descriptive statistics of the sample data are presented in Table 1.

The research model primarily encompasses four variables: GEO, BSS, BMI and BDC. To ensure measurement validity, this study adopts established scales from prior international studies and adapts them to the start-up context. A 5-point Likert scale is employed for data collection:

Common method bias was assessed using Harman’s one-factor test, yielding a Kaiser–Meyer–Olkin (KMO) value of 0.881. The approximate chi-square value was 2924.333, with 351 degrees of freedom (df) and a significance level of 0.000. The first common factor extracted in the unrotated state accounted for 24.724% of the total variance, which falls below the 50% threshold. This result suggests that common method bias is unlikely to be a serious concern in this study.

The reliability test results for the four variables, namely GEO, BSS, BDC and BMI, are presented in Table 2. The findings indicate that all four variables have Cronbach’s α coefficients greater than 0.8, and their composite reliability (CR) values exceed 0.8, indicating high reliability for each scale. Reliability and validity were assessed by examining convergent and discriminant validity. The factor loading values for all four variables were above 0.7, and the average variance extracted (AVE) ranged from 0.502 to 0.558, all surpassing the threshold of 0.5, indicating strong convergent validity. Furthermore, the square root of the AVE for each variable exceeded the correlation coefficients between variables, confirming the presence of discriminant validity.

In addition to Harman’s one-factor test, we conducted a confirmatory factor analysis (CFA)-based model comparison to assess the potential severity of common method bias. As reported in Table 2, the hypothesised four-factor measurement model (GEO, BSS, BDC and BMI) showed a good fit (χ² = 431.445, df = 318, χ²/df = 1.357, comparative fit index (CFI) = 0.958, Tucker-Lewis Index (TLI) = 0.953 and root mean square error of approximation (RMSEA) = 0.034), whereas a one-factor model in which all items loaded on a single factor fits the data poorly (χ² = 1553.212, df = 324, χ²/df = 4.794, CFI = 0.540, TLI = 0.501 and RMSEA = 0.111). This substantial deterioration in fit suggests that common method variance is unlikely to be a serious concern in this study.

Descriptive statistics and inter-variable correlation analysis were conducted for each variable, and the results are presented in Table 3. Significant positive correlations were observed between GEO and BSS (r = 0.297, p < 0.01), and between GEO and BMI (r = 0.345, p < 0.01). Boundary-spanning search was also positively correlated with BMI (r = 0.240, p < 0.01). These correlation patterns are consistent with the proposed hypotheses and provide preliminary support for subsequent regression analyses. Multicollinearity was further assessed in the regression analysis stage, where all variance inflation factor (VIF) values were below the conservative threshold of three, indicating no serious multicollinearity.

This study advances the literature on green entrepreneurship and BMI in three ways. Firstly, it extends green entrepreneurship research by shifting the focus from performance outcomes to BMI as a strategic reconfiguration outcome, and by clarifying why a green entrepreneurial posture does not uniformly translate into business model change (Makhloufi et al., 2022). Conceptualising GEO as an intangible strategic resource, we argue and show that its value lies not only in ‘being green’ but in enabling start-ups to redesign value creation and capture when complementary knowledge is mobilised, thereby responding to calls for more empirical work on the antecedents and micro-foundations of BMI (Appiah et al., 2025; Foss & Saebi, 2017).

Secondly, the study contributes to organisational search theory by specifying BSS as the conversion mechanism through which GEO is transformed into BMI. Importantly, we foreground the benefits–cost tension inherent in external search: whilst BSS provides access to non-redundant knowledge needed for business model redesign, it can also impose attention and integration burdens on resource-constrained start-ups (Yang et al., 2021a). By empirically establishing partial mediation, our findings help explain heterogeneity in outcomes amongst green-oriented ventures, by clarifying why some ventures translate sustainability intent into BMI whereas others do not.

Thirdly, the study advances the big data and digital capability literature by theorising and testing BDC as a conversion-enabling boundary condition, rather than merely a direct driver of innovation outcomes. The results indicate that BDC strengthens the BSS–BMI link by enhancing firms’ ability to filter, integrate and recombine heterogeneous external information into coherent business model design choices (Mikalef et al., 2019; Suoniemi et al., 2020). This reconceptualisation opens avenues for examining digital capabilities as moderators in sustainability-oriented entrepreneurship, particularly in emerging-market contexts, such as South Africa and many African economies, where knowledge ecosystems are fragmented and digital capability is uneven, making the returns from BSS more contingent.

Against the backdrop of climate and sustainability pressures, the findings offer practical guidance for start-ups seeking to translate green entrepreneurial intent into BMI, particularly in emerging-market settings such as South Africa and many African economies, where resource constraints, fragmented ecosystems and uneven infrastructure can make execution challenging.

Firstly, make ‘green’ a business model design problem, not only a CSR initiative. Founders can operationalise GEO by explicitly redesigning value propositions, partner configurations and revenue–cost logic around environmental value. A useful practice is to run short ‘minimum viable green’ experiments that test both customer willingness-to-pay and operational feasibility (e.g. piloting cleaner inputs with one supplier, testing a repair or take-back scheme with a single channel partner, or trialling pay-per-use models that reward durability). In contexts with tighter financing constraints, framing sustainability as measurable customer value (cost savings, reliability, compliance readiness and waste reduction) helps avoid overinvestment in symbolic green practices that do not change the underlying business model.

Secondly, treat BSS as a focused portfolio, anchored in the local ecosystem. Start-ups should deliberately combine near sources (customers, suppliers, distributors and competitors) with a small set of distant sources that provide non-redundant knowledge. For South African/African ventures, high-leverage distant sources often include universities and applied research centres, industry associations, standards and certification bodies, large incumbents’ supplier-development programmes and public or municipal initiatives related to sustainability. To prevent search from becoming unfocused, teams can assign an ‘owner’ to each search channel, set a monthly cadence for harvesting insights and use a simple decision rule (e.g. ‘keep/kill/pivot’ after two review cycles) so that external ideas are quickly translated into concrete business model choices.

Thirdly, build a ‘minimum viable’ BDC to convert external information into decisions. Because BSS can generate noisy and heterogeneous information, managers should prioritise light-weight information-processing routines before expensive technology investments. A practical sequence is: (1) define a small set of decision-critical metrics (unit economics, customer adoption and retention, partner performance and compliance costs), (2) implement basic data governance (data ownership, definitions and quality checks), (3) establish cross-functional information sharing (regular reviews that link data to decisions) and (4) add simple analytics tied to specific choices (pricing, channel selection, supplier switching and service design). This approach is especially useful where data availability and quality are uneven and teams are small: the goal is not to ‘do big data’, but to build enough capability to filter signals, integrate external inputs and accelerate business model experimentation. Managers should also recognise pitfalls: overly broad search can overwhelm attention, and data initiatives without clear decision use-cases risk becoming costly technology projects rather than capability building.

Whilst this study yields meaningful findings, several limitations should be acknowledged. Firstly, although the sample size meets the requirements for empirical analysis, the data were collected from a limited number of regions, which may constrain the generalisability of the findings. Future research could extend data collection to broader geographical contexts and incorporate qualitative approaches, such as case studies or longitudinal designs, to enhance external validity. Secondly, this study focuses on GEO as an overall strategic orientation and does not disentangle the potentially heterogeneous effects of its specific dimensions. Future research could examine how different components of GEO or additional contextual factors influence BMI. Finally, the study relies on self-reported survey measures collected at a single point in time. Future studies could employ multi-source data, alternative operationalisations, or mixed-method approaches to further strengthen measurement robustness.