The present study is rooted in the DCT, which is an outgrowth of the Resource-Based View (RBV) of the firm. The RBV attributes firms’ performance variances to diversity in resources owned and controlled by the enterprise, noting that resources that are rare, inimitable, organisable, and valuable in a special way generate superior firm performance (Hesterly & Barney 2015). However, recent scholars have criticised the RBV for being internally oriented and static, thereby failing to account for performance differences in turbulent, dynamic environments (Chinakidzwa & Phiri 2020; Zhou et al. 2019). Moreover, merely possessing resources is not sufficient; capabilities are needed to profitably transform resources into value (Chinakidzwa & Phiri 2020). Dynamic capabilities underscore the ‘key role of management in appropriately adapting, integrating, and reconfiguring internal and external experience, resources, functional competencies within a changing environment’ (Gregory, Ngo & Karavdic 2019:12).

The DCT is based on the premise that firms can rapidly and speedily adjust and change their internal resource arrangement to facilitate organisational processes after witnessing lucid signals from the marketplace (Morgan 2012). Dynamic capabilities are organisational capabilities that ‘enable the firm to implement strategies using new and different combinations and transformation of resources’ (Teece et al. 1997:8) to adapt the altering business conditions. We take the view that dynamic capabilities – in this case, OA, will enable social enterprises to develop and identify capabilities that will enable them to respond to social problems and market opportunities (Gregory et al. 2019). Since DCT has the power to explain performance differences in volatile, unpredictable, and dynamic markets like today’s turbulent, exacting marketplace caused or exacerbated by the COVID-19 pandemic, it forms the theoretical bedrock for our arguments.

The quest for a better way of achieving change in today’s competitive business and economy has sparked an accelerating research interest in OA during the recent two decades (Tarba et al. 2020). Scholars employed the human characteristic of ambidexterity as a symbol to represent organisations. Ambidexterity literally means ‘right on both sides’ (Simsek 2009:599). Therefore, organisations that are ‘capable of exploiting existing competencies as well as exploring new opportunities with equal dexterity are akin to humans employing both hands equally’ (Simsek 2009:599) Thus, OA is defined as a firm’s capability to concurrently orchestrate exploitative strategies and activities that streamline present operations to maximise value and explorative activities that accomplish future business opportunities and growth (Venugopal et al. 2020). Ambidextrous enterprises are not merely effective and efficient in managing current business requirements but also sufficiently flexible to adjust to the changes in the progressively dynamic, volatile, and unpredictable business environment (Limaj & Bernroider 2019), which in turn, leads to sustainable competitive advantage (Mazzelli et al. 2020).

Generally, the literature suggests two main forms of ambidexterity: exploitative ambidexterity and explorative ambidexterity. Exploitative ambidexterity permits organisations to respond to disruptions and changes by adjusting or adapting to current competencies and business demands, while explorative ambidexterity allows firms to respond to market changes by developing novel products and services and exploiting novel markets (Cegarra-Navarro, Sánchez-Vidal & Cegarra-Leiva 2011; Lubatkin et al. 2006). Exploration looks at ‘search, variation, risk-taking, experimentation, play, flexibility, and discovery’ (p. 71), while on the other hand, exploitation looks at ‘refinement, choice, production, efficiency, selection, implementation, and execution’ (March 1991:71). Empirical evidence shows that these two dimensions are important drivers of firm performance (O’Reilly & Tushman 2013), although the prevailing studies are limited to the for-profit context, particularly in SMEs (Lee & Kreiser 2018). Exploitative and explorative activities demand significantly diverse, often even conflicting cultures, capabilities, processes, and structures; thereby, inconsistencies, contradictions, and conflicts are predictable (O’Reilly & Tushman 2004). Whereas earlier research often deemed the trade-off required to accomplish high levels of exploitation and exploration as not possible to put into operation, other studies pointed out a series of solutions to implement OA, such as structural separation (O’Reilly & Tushman 2004) and non-structural, context-related factors like values, mindset or culture (Eisenhardt & Martin 2000).

As well, research underscores different types of strategies adopted by firms with regard to OA. For instance, scholars draw on a time-dependent perspective to differentiate between two approaches of OA as manifested in organisations: sequential ad simultaneous. On the one hand, the sequential ambidexterous approach occurs when a firm systematically uses a temporal cycle by either, firstly, implementing exploration or exploitation, followed by seasons of rotations (Venkatraman et al. 2007). On the other hand, simultaneous ambidexterity takes place when an organisation synchronously pursues exploitation and exploration simultaneously (O’Reilly & Tushman 2013). Other scholars also differentiate between structural ambidexterity and contextual ambidexterity that may affect a firm’s exploitative and explorative strategies. Structural ambidexterity addresses the overall organisational structure in distinct, discrete subunits that may influence the concurrent development of exploitation and exploration (Diaz-Fernandez, Pasamar-Reyes & Valle-Cabrera 2017; Simsek 2009). Research demonstrates that strategic structural ambidexterity enhances contexts for appropriate exploitation and exploration (Raisch et al. 2009), facilitates differentiation of tasks (Venugopal et al. 2018), and promotes diverse cultures and processes (O’Reilly & Tushman 2004). On the other hand, contextual ambidexterity involves the incorporation of both activities into one business unit, performed by all workers with no additional specialisation.

Social enterprises are organisations that pursue a social mission and utilise innovative revenue generation approaches to fulfill their mission (Kilpatrick et al. 2021; Weaver 2020). Thus, social enterprises are characterised by social mission and value creation, social innovation, social change, and entrepreneurial spirit. Farhoud et al. (2021) stress that social entrepreneurs, as a result of the increasingly dynamic, volatile business environment, coupled with the unpredictable nature of government programmes and implementation of social interventions (Weaver & Blakey 2022), shift their enterprises’ resources and capabilities between novel and current products and services to execute social missions. Social entrepreneurs not only identify and solve social matters but also explore social needs, determine social opportunities, and develop strategies to meet them (Davis & Bendickson 2021; Weaver & Blakey 2022). Farhoud et al. (2021) postulate that social enterprises must move swiftly to identify novel business roles and opportunities in today’s exacting, turbulent marketplace.

Social enterprises are relevant modulators and facilitators of ambidexterity for several reasons. First, two diverse but mutually exacting systems interact within this context: the firm system and stakeholders’ expectations. The resulting feature of social enterprises, especially the influence of stakeholders (e.g. donors) in terms of expectation, allows them to reach an increased degree of exploration for growth and survival in the future and exploitation of existing processes and competencies simultaneously. In fact, the heightened requirements for accountability toward stakeholders demand not-for-profit enterprises to enhance resources, mobilise for the transformation and evolution, and at the same time guarantee frequent production to create social, economic, and environmental values. Through ambidextrous functioning, social enterprises can accomplish this (Peng 2019).

Social enterprises may build an environment that offers the necessary unique focus on performance and social support. Peng (2019:1) noted, ‘organizational ambidexterity might provide some strategic leads for balancing the possible paradoxes within different kinds of expectations of the stakeholders’ of not-for-profit organizations’. Due to the recent pandemic, stakeholders’ demands and resource tensions have led social enterprises to novel changes to secure augmented efficiency. In some instances, social enterprises should be able, with fewer means, to ‘do more and do it better’ (Peng 2019), which could create some kind of organisational tensions and organisational paradoxes (Bloodgood & Chae 2010); thereby requiring social enterprises to discover novel strategic visions to achieve a balance between organisational change and stability of social intervention efficiency. Thus, social enterprises must handle the contending pressures of distilling current routines and structures for efficiency with the need to innovate and grow. Secondly, because social enterprises develop and distribute social programmes, previously executed by governmental organisations, they should balance the efficient delivery of current products and services with alterations in the larger business context that might influence future demand (Peng 2019) in such areas as discontinued or new funding opportunities (Sherlock & Gravelle 2018) and changes in growth potential (Kaplan & Grossman 2010). Thirdly, social enterprises operate within a central mission which orchestrates the scope of their products, programmes, and services, making them unlikely to achieve growth via unrelated diversification. These conditions make social enterprises an exceptional context for investigating OA and exploring if and how social enterprise managers/owners attempt to engage in explorative and exploitative ambidexterity.

Research demonstrates that OA leads to superior performance since the firm demonstrates efficiency in executing present market demands, while simultaneously manifesting the flexibility required to adapt to new opportunities and challenges in the business environment (Peng 2019). Thus, we expect that OA will enhance social enterprise performance. While the cost related to achieving a greater level of both aspirations are high (Haibin and Atuahene-Gima, (2007), Ambidexterity in product innovation management: The direct and contingent effects on product development performance, and a score of businesses remain unsuccessful (O’Reilly & Tushman 2004), we believe that the unique context of social enterprises eases ambidextrous thinking and the implementation of it in a firm’s strategic orientation. Inspired by the DCT, we distinguish between exploitative and explorative ambidexterity and examine their relative impact on relative dimensions of social enterprise performance. Madden (2012) found that OA is positively related to the fiscal performance in a not-for-profit enterprise. Other studies (e.g. Peng 2019; Plimmer, Bryson & Teo 2017) found that OA can increase the performance of not-for-profit enterprises.

However, one crucial limitation of these studies is that they did not examine the relative effect of exploitative and explorative ambidexterity, nor did they use a multidimensional approach to capture social enterprise performance. However, the social entrepreneurship literature suggests that enterprise performance should be captured into multiple dimensions (Pearce, Fritz & Davis 2010; Van der Vaart 2021), because a multidimensional performance framework offers a healthier test of the focal relationship. Accordingly, we used two dimensions subdivided into four dimensions to capture social enterprise performance, namely social performance (1) social marketing achievement and (2) social value creation, and economic performance (3) commercial marketing achievement and (4) economic value creation (Liu et al. 2015). Thus, we expect that higher levels of exploitative and explorative ambidexterity will lead to a higher level of social enterprise performance, as reflected in the following hypotheses, informed by the DCT:

We examined the hypothesised relationships in a research context of Ghanaian social enterprises. A quantitative research design is employed, and the research instrument used is a survey questionnaire.

To test the study’s hypotheses, we designed a large-scale empirical instrument and sent it to the owners/managers of the social enterprises to obtain primary data. We designed the questionnaire, based on the methodologies and recommendations from social entrepreneurship literature (Liu et al. 2015), and observed methodological measures to reduce concerns about responses biases (Podsakoff 2003). Moreover, we conducted a pilot-test incorporating numerous social enterprises owners, researchers, and managers to boost our survey questionnaire further. We employed a convenience sampling method as the address base and addressed the survey instrument to 1200 individual respondents who were owners/managers of social enterprises in Ghana. We have included social enterprises in the analysis of the study based on the following two criteria: firstly, the enterprise generates income from business activities (i.e. income from divergent avenues like trading ventures or service offering for novel fundraising (Guo 2006; Liu et al. 2015); and secondly, the enterprise has been able to generate sufficient revenue from multiple sources and has a track record of trading activities in the past three years. We found enterprises with annual revenue above Ghs80 000 (approximately $10 000).

With the inclusion criteria established, we sent out emails in January 2022, followed by a bespoke letter outlining the research scheme and purpose. After we sent out two mailing lists, 417 enterprises returned the survey instrument, representing a response rate of 34.75%. We tested the quality of the dataset created (Tabachnick, Fidell & Ullman 2007) before we proceeded to the empirical analyses. Firstly, we examined every item’s means, standard deviations (SDs), and minimum and maximum values for plausibility. However, we did not find any error concerning the accuracy of our data entry.

Nevertheless, some of the questionnaires reported missing values in diverse aspects of the questionnaire. We solved this problem by deleting these cases from the created dataset (Draper & Smith 1998), reducing the usable sample to 317 social enterprises. In addition, we checked for potential biases because of the cross-sectional survey nature of the study, since this could decrease the validity of the findings. Here, a commonly reported problem is the characteristic difference between the respondents of a study and non-respondents (Armstrong & Overton 1977). To test for any potential discrepancies between early and late responses, we used the non-parametric Kolmogorov-Smirnov test and Mann-Whitney U-test. Our analysis did not reveal any statistically significant response biases between the early and late responses. Apart from non-response bias, the impact of common method bias has been a relevant issue in entrepreneurship and management research (Podsakoff 2003). This bias points out a variance in findings associated with the employed measurement approach instead of to the constructs of the study (Podsakoff 2003). Harman’s single-factor test – a method commonly used to assess common method bias in a single-method research approach – was used to this effect (Malhotra, Kim & Patil 2006). Our analysis again demonstrates that common method bias is no issue in our applied measurement method.

We adapted the Lubatkin et al.’s (2006) (from Gottschalck et al. [2021] scale to assess exploitation and exploration. Depending on their enterprise’s OA in the previous three years, the owners/managers of the social enterprises were asked to evaluate six (6) statements each on exploitation and exploration on a scale of 1 (strongly disagree) to 5 (strongly agree). Regarding the social enterprise performance, we measured it using the respondents’ subjective assessment for the reasons following. Firstly, obtaining and appropriately interpreting a social enterprise’s performance is daunting because there is scarcely any dependable access to data on performance for privately held enterprises (Dess & Robinson 1984). Secondly, social enterprises are not required by law to publish their financial performance and/or non-performance (Liu et al. 2015). Thirdly, research shows that managerial actions and decisions are driven principally by perceptions or views of organisational performance (Morgan, Kaleka & Katsikeas 2004). This way, strategic management scholars have made extensive use of perceived performance indicators, which have shown relevant reliability and validity applicable to objective performance assessments.

To assess the perceived performance of the social enterprises based on the four dimensions, we asked the respondents to rate the extent to which they have achieved their social and economic performance. Firstly, social performance is examined via social marketing achievement and social value creation by adapting the scale of Liu et al. (2015). Social marketing achievement involves the enterprise’s marketing programme’s effectiveness concerning whether or not it has accomplished its preset donor and volunteer contribution-based targets, and it was measured using five items. On the hand, the social value creation measures how well the enterprise is able to create social value, based on critical social performance indicators like serving more beneficiaries, bidding for public service contracts, expanding social service to more locations, among others (Gainer & Padanyi 2005). It was measured using five-scale items (Liu et al. 2015). Likewise, we examined economic performance based on economic marketing achievement and economic value creation, adapting the scale of Liu et al. (2015). Commercial marketing achievement involves the effectiveness of the enterprise’s marketing programmes vis-à-vis preset market-based goals (Cano et al. 2004). In contrast, economic value creation involves the ability of the enterprise to create economic value from its social mission programmes based on key economic performance measures in the past year. Accordingly, we measured the commercial marketing achievement using five items, while six items were used to capture the economic value creation of the social enterprises (Liu et al. 2015).

Previous studies have used size as a control variable (e.g. Alegre-Vidal & Chiva-Gomez 2013), stressing that large enterprises are likely to possess a strong resource-base of well-established reputation and brand to garner business opportunities in both social and commercial contexts. Thus, an enterprise’s size, measured in this study by the total revenue of the enterprises, may potentially influence its ability to achieve high social and economic performance. Therefore, we used turnover instead of staff (number of employees) because the total number of staff might not be a true reflection of the size of social enterprise, due, in part, to the fact that social enterprises tend to consist of paid and volunteer employees, and so there is a considerable variation in the hours of each volunteer (Liu et al. 2015; Liu & Ko 2012). Accordingly, we measured the size of the social enterprise using a 5-point Likert scale (1 = very small, 5 = very large) that allows for the differentiation of the total revenues of the social enterprises. An interval of Ghs 80 000 (approximately $10 000) was maintained between each point scale. By employing a Likert scale, we sought to quell the conventional problem that sometimes arises from firms’ reluctance to disclose financial information, coupled with the fact they may not provide accurate figures when they choose to disclose financial data (Zahra, Neubaum & El-Hagrassey 2003). Three hundred and seventeen (62% men, M_age = 48 years, SD = 12.91) participated in an online survey. The research study was carried out according to the principles of the Declaration of Helsinki.

To evaluate the quality of our construct measures regarding construct dimensionality, reliability, and convergent/discriminant validity, we carried out a confirmatory factor analysis (CFA) via AMOS version 28. As a first step, we calculated the Kaiser-Meyer-Olkin (KMO) value to measure the sampling adequacy. As a rule of thumb, the KMO value must exceed 0.50 before one can proceed to the factor analysis (Janssens et al. 2008). Table 1 shows that the fit indices of the model fit are very satisfactory (χ² = 512.068 [degrees of freedom {df} = 315; p < 0.001]; RMSEA = 0.06; NNFI = 0.95; CFI = 0.96; GFI = 0.87; RMR = 0.06; KMO = 0.95). Moreover, Table 1 shows that all of the average variance explained (AVE) exceeds the 0.5 thresholds (Fornell & Larcker 1981), while all of the composite reliabilities also exceed the 0.7 thresholds (Hair et al. 2014), indicating that our model meets the convergent validity and construct reliability threshold, respectively.

Table 2 shows that the discriminant validity, a measure of non-linear correlation, is also confirmed (Hair et al. 2014; Henseler & Chin 2010).

This study makes numerous theoretical contributions. Firstly, this quantitative survey concurrently investigates the two components of OA in association with the social and economic performance of social enterprises and extends the existing work on OA (Lubatkin et al. 2006; O’Reilly & Tushman 2013; Raisch et al. 2009; Simsek 2009) to the social enterprise context. Our focus on social enterprises in emerging economies adds theoretical value as it enables us to evaluate OA in the social entrepreneurship context. Contrary to expectations, we found that OA does not decrease but increases social enterprises’ social and economic performance. Secondly, the present study allows greater comparability and generalisability by being grounded in a well-established theoretical model (the DCT), which highlights how firms can integrate, realign, and readjust their strategies to adapt to the changing environment; thereby showing the applicability of the theory to the social enterprise setting. Thirdly, our study responds to calls by numerous field scholars for urgent, large-scale quantitative analysis of social enterprises (Liu et al. 2015; Meyskens et al. 2010). Fourthly, our results provide valuable insights into how both exploitative and explorative ambidexterity independently impacts diverse social enterprise dimensions; thereby exposing social managers to the performance effects of OA. Finally, our study, extended to an emerging market, unlike previous studies, allows for cross-cultural and cross-country validation of the OA model.

The findings of the research also put forth several relevant managerial implications. The first implication is in regards to the relative, joint effect of exploitative and explorative ambidexterity. This implies that managers can develop and implement them jointly to improve performance, notwithstanding the conventional thinking that it is impossible to put them both into operation because of the trade-off necessary to accomplish high levels of exploitation and exploration. Therefore, the result offers valuable insights to social enterprise owners/managers on achieving the amalgamation of the steadiness of organisational transformation and social service deliverance via OA. The point is that although balancing the two may not be achieved on a silver platter, our results show that it is possible and that through solutions such as structural separation (O’Reilly & Tushman 2004) and non-structural separations (Eisenhardt & Martin 2000), firms, including social enterprises, can leverage the strategic gains from both exploration and exploitation simultaneously. Hence, social enterprises should not be side-minded, whether because of their culture, mission, perceived lack of resources and capabilities, or philosophy. Instead, they must engage in some level of exploration and exploitation to reap the benefits associated with the OA model.

Moreover, in a resource-constrained context like social enterprises, managers must manage the trade-offs between exploitation and exploration by focusing on such factors as efficiency, cost, and incremental innovation for exploitation and flexibility, speed, and radical innovation for exploration. Therefore, social enterprises must realign and readjust their competencies, structures, leadership skills, and strategies. Again, social enterprises’ managers should foster employee work engagement behaviours through social awareness (organisational awareness, attunement, empathy) and relationship management (teamwork, inspiration, influence) to facilitate a shared vision and a collective comprehension of how employees and leaders will manage the exploitation-exploration trade-off.

Because OA can only be nurtured in a supportive context and environment, where employees are motivated to make their own decisions in terms of the amount of exploration and exploitation since management is not involved in single decisions but responsible for shaping the context, social enterprises that aim to be ambidextrous must harness ambidextrous employees who can understand the demands of both exploitative and explorative activities and ultimately manage to accommodate its tension. Finally, our comprehensive analysis of the OA – social enterprise performance relationship may represent a solid foundation for policy recommendations to governments and national decision-makers regarding technological and institutional capacity and systems that must exist to enable social enterprises to achieve a relative balance between exploration and exploitation.