Social enterprises apply business approaches to address social problems such as poverty, access to potable water, and homelessness (Nicholls, Reference Nicholls2006). They are starting-up at a faster rate and creating more jobs in the United Kingdom, for example, than their commercial counterparts (Nicholls, Reference Nicholls2006; Social Enterprise, UK, 2017). Scholars view social enterprises as a mechanism for economic transformation, advocating that social benefit creation should be the primary goal of business and profit merely the means to achieve that goal (Crane, Palazzo, Spence, & Matten, Reference Crane, Palazzo, Spence and Matten2014; Porter & Driver, Reference Porter and Driver2012).
While many social enterprises are founded, very few scale-up beyond their initial community or region (Dees & Anderson, Reference Dees and Anderson2004; Evans & Clarke, Reference Evans and Clarke2011; Smith, Kistruck, & Cannatelli, Reference Smith, Kistruck and Cannatelli2016). Unfortunately, small-scale social enterprises are limited in effect, often unable to ensure their impact matches the magnitude of the social need or problem they seek to address (Dees, Reference Dees2008). Scholars therefore describe scale-up from one geographical context to another as one of the most important but least understood outcomes of social entrepreneurship (Alvord, Brown, & Letts, Reference Alvord, Brown and Letts2004; Smith, Kistruck, & Cannatelli, Reference Smith, Kistruck and Cannatelli2016).
Existing research is predominantly conceptual, speculating why scale-up may be challenging (Dees & Anderson, Reference Dees and Anderson2004; Bloom & Chatterji, Reference Bloom and Chatterji2009). For example, researchers say social enterprises emerge from exposure to local issues such that an insular perspective may limit scalability of solutions (Zahra, Gedajlovic, Neubaum, & Shulman, Reference Zahra, Gedajlovic, Neubaum and Shulman2009; Smith & Stevens, Reference Smith and Stevens2010). Perrini, Vurro, and Costanzo (Reference Perrini, Vurro and Costanzo2010) suggest that the social, political, and economic context surrounding a social enterprise affects both start-up and scale-up. Other research is more anecdotal, offering practitioners advice about funding (Bradach, Reference Bradach2003) and clarity of the underlying value proposition (Evans & Clarke, Reference Evans and Clarke2011).
Empirical research has been slow to surface, most likely because enterprises that have scaled beyond their local context are relatively rare (Perrini, Vurro, & Costanzo, Reference Perrini, Vurro and Costanzo2010) and data are difficult and expensive to collect (Short, Moss, & Lumpkin, Reference Short, Moss and Lumpkin2009). This lack of research constrains our knowledge of how social enterprises can be grown so that their impact matches the magnitude of the social need or problem they seek to address (Dees, Reference Dees2008). Empirical research may suggest mechanisms for scale-up of social enterprises that are effective across contexts, potentially informing practice and theory.
The purpose of this paper is to explore the scale-up of social enterprises. Our exploration focuses on resources because the mobilization of resources is a key challenge in scale-up of social enterprises (Dees & Anderson, Reference Dees and Anderson2004; Bloom & Chatterji, Reference Bloom and Chatterji2009; Desa, Reference Desa2012; Desa & Basu, Reference Desa and Basu2013). These enterprises often operate in developing regions or countries where quality resources are scarce or nonexistent (Di Domenico, Haugh, & Tracey, Reference Di Domenico, Haugh and Tracey2010; Desa, Reference Desa2012; Desa & Basu, Reference Desa and Basu2013) and institutions that facilitate access to resources are undeveloped (Mair & Marti, Reference Mair and Marti2009; Desa & Basu, Reference Desa and Basu2013). Given the importance of resource mobilization to scale-up, we apply the dynamic capabilities framework from the strategic management literature to understand how enterprises acquire and configure resources. Dynamic capabilities are defined as abilities enterprises have to amass, configure, and reconfigure resources (Helfat et al., Reference Helfat, Finkelstein, Mitchell, Peteraf, Singh, Teece and Winter2007). An investigation of dynamic capabilities can show how enterprises achieve scale-up (Perrini, Vurro, & Costanzo, Reference Perrini, Vurro and Costanzo2010) and other positive outcomes (Ambrosini, Bowman, & Burton-Taylor, Reference Ambrosini, Bowman and Burton-Taylor2007; Teece, Reference Teece2007). Consistent with dynamic capabilities, we address the research question: How are resources amassed and configured for start-up and scale-up of social enterprises? Specifically, the study examines geographically different start-up and scale-up contexts.
LITERATURE REVIEW
Social entrepreneurship
Scholars generally agree that social entrepreneurship can be distinguished from commercial entrepreneurship by its focus on creating social benefit or value instead of, or as well as, private wealth (Austin, Stevenson, & Wei-Skillern, Reference Austin, Stevenson and Wei-Skillern2006; Mair & Marti, Reference Mair and Marti2006). Research identifies different approaches to social entrepreneurship based on geographical disparities (Bacq & Janssen, Reference Bacq and Janssen2011) and sector differences such as public versus private sector (Austin, Stevenson, & Wei-Skillern, Reference Austin, Stevenson and Wei-Skillern2006; Mair & Marti Reference Mair and Marti2006). Given there is no consensus definition of social entrepreneurship (Zahra et al., Reference Zahra, Gedajlovic, Neubaum and Shulman2009), readers are referred to review articles for multiple definitions of this phenomenon (see Zahra et al., Reference Zahra, Gedajlovic, Neubaum and Shulman2009; Dacin, Dacin, & Matear, Reference Dacin, Dacin and Matear2010; Bacq & Janssen, Reference Bacq and Janssen2011).
For this research, we define social entrepreneurship as processes like product/service development, enterprise founding, and capability building that involve the innovative use and combination of resources for addressing social problems, in order to meet the needs of marginalized and disadvantaged peoples (Mair & Marti, Reference Mair and Marti2006; Martin & Osberg, Reference Martin and Osberg2007; Zahra et al., Reference Zahra, Gedajlovic, Neubaum and Shulman2009). To further describe social entrepreneurship, we explain Zahra et al.’s (Reference Zahra, Gedajlovic, Neubaum and Shulman2009) three types of social entrepreneurial processes that are based on well-known theories of entrepreneurship. The first type involves a social bricoleur who perceives a local need and draws on local knowledge to address the need through a small-scale enterprise. The second reflects a social constructionist whose entrepreneurial alertness makes him or her aware of a widespread social problem. The social constructionist develops a formalized, potentially scalable solution that, in theory, can be transferred to new locations. The third type requires a social entrepreneur who becomes aware of societal level issues and creates new institutions that challenge the existing social order. All three types have substantive constraints to overcome when trying to mobilize resources (Zahra et al., Reference Zahra, Gedajlovic, Neubaum and Shulman2009).
Scholars in addition to Zahra et al. (Reference Zahra, Gedajlovic, Neubaum and Shulman2009) note that social enterprises have more significant resource constraints than do fledgling commercial ventures and collectively point to three reasons. First, social enterprises often operate in developing regions and economies where resources can be scarce and of poor quality (Mair & Marti, Reference Mair and Marti2009; Desa & Basu, Reference Desa and Basu2013). For example, human capital needs can be difficult to fulfill given lack of education and employment experience (Di Domenico, Haugh, & Tracey, Reference Di Domenico, Haugh and Tracey2010). Second, institutions that support resource acquisition in developed economies, such as financing mechanisms, are weak or nonexistent (Mair & Marti, Reference Mair and Marti2009; Di Domenico, Haugh, & Tracey, Reference Di Domenico, Haugh and Tracey2010). Third, social enterprises often do not implement the profit-sharing mechanism that commercial ventures use to entice investment of resources into projects (Desa & Basu, Reference Desa and Basu2013).
Scale-up of social enterprises
Our focus is on the scale-up of social enterprises. We define scale-up as growth in social value by expanding a current product/service to other geographic locations (Smith & Stevens, Reference Smith and Stevens2010). Geographical expansion is needed to ensure the impact of a social enterprise matches the magnitude of the social need or problem it seeks to address (Dees, Reference Dees2008). Many social problems, like poverty or limited access to potable water, are global in nature and scale-up of social enterprise is advocated as a solution (Porter & Driver, Reference Porter and Driver2012). Unfortunately, scale-up of social benefits is one of the most important but least understood outcomes of social entrepreneurship (Smith, Kistruck, & Cannatelli, Reference Smith, Kistruck and Cannatelli2016).
Initial research on scale-up is predominantly conceptual in nature, drawing on anecdotal evidence of what appears to have worked. For example, the Zahra et al. (Reference Zahra, Gedajlovic, Neubaum and Shulman2009) types reflect differences with respect to scaling-up social enterprises. Types include: small, localized enterprises; formally constructed regional/national enterprises; and enterprises that have a global effect by destroying society’s current institutions and creating new ones. Bloom and Chatterji (Reference Bloom and Chatterji2009) offer a SCALERS model conjecturing that scaling-up social enterprises requires the following: staffing, communicating, alliancing, lobbying, earnings generation, replicating, and stimulating market forces. Others hypothesize that social enterprise scale-up is more likely to occur if the entrepreneur feels an intense moral obligation to solve a wider problem (Smith, Kistruck, & Cannatelli, Reference Smith, Kistruck and Cannatelli2016).
Conceptual research also posits specific mechanisms for scaling-up social enterprise. Dees and Anderson (Reference Dees and Anderson2004) describe three mechanisms: dissemination, affiliation, and branching. Dissemination is providing information and technical assistance but letting others scale-up the social benefit created by the original founder. Affiliation is akin to alliances for commercial businesses, it involves scaling-up through relationships with other organizations. Finally, branching is the creation of multiple sites by one organization similar to a commercial company owning and operating numerous shops in many locations. Smith and Stevens (Reference Smith and Stevens2010) extend these mechanisms by matching them with types of social entrepreneurs. Bradach (Reference Bradach2003) identifies a mechanism he calls replication, similar to franchising in the for-profit sector, which includes standardization of key activities, identification of a standard context, and a standardized mechanism for ensuring the flow of money. Much of the literature agrees with Bradach (Reference Bradach2003), arguing that replication or reproducing carbon copies of an original social innovation is the key to scaling-up social enterprises (Evans & Clarke, Reference Evans and Clarke2011).
However, several scholars suggest that scaling-up is difficult because social entrepreneurs underestimate the need to make modifications for different contexts (Perrini, Vurro, & Costanzo, Reference Perrini, Vurro and Costanzo2010; Evans & Clarke, Reference Evans and Clarke2011). Evans and Clarke (Reference Evans and Clarke2011) advocate customizing for new locales to allow for local barriers, mindsets, and other idiosyncrasies. Perrini, Vurro, and Costanzo (Reference Perrini, Vurro and Costanzo2010) encourage social entrepreneurs to distinguish between the essential elements that make a product/service successful in an initial location from contextual conditions and to implement only the essential elements in any new location.
The present study extends existing research by collecting and analyzing evidence about how social enterprises scaled-up to new geographical contexts, in essence focusing on Zahra et al.’ s (Reference Zahra, Gedajlovic, Neubaum and Shulman2009) second type – the social constructionists who are trying to achieve scale and impact beyond a single locale. In particular, we apply dynamic capabilities, examining how resources are amassed and configured for start-up and scale-up.
Dynamic capabilities
Dynamic capabilities evolved from the resource-based view which sees an organization as an idiosyncratic configuration of resources (Eisenhardt & Martin, Reference Eisenhardt and Martin2000). The resource-based view illustrated favorable resource configurations but was limited in its explanation of how these configurations were modified for changing environments (Priem & Butler, Reference Priem and Butler2001; Barreto, Reference Barreto2010). The notion of dynamic capabilities addressed this limitation because it reflected routines and processes used by enterprises to transform resources to create new sources of value (Teece, Pisano, & Shuen, Reference Teece, Pisano and Shuen1997). Dynamic capabilities were formally defined as abilities to create, configure, and reconfigure resources to address changing environments (Helfat et al., Reference Helfat, Finkelstein, Mitchell, Peteraf, Singh, Teece and Winter2007). They enabled for-profit businesses to either sustain a competitive advantage relative to rivals or construct a series of temporary competitive advantages to address a continuously changing environment (Eisenhardt & Martin, Reference Eisenhardt and Martin2000).
Most dynamic capabilities research focused on large, established commercial businesses (Barreto, Reference Barreto2010; Corner & Wu, Reference Corner and Wu2012) and investigated how firms adjusted existing resource bundles to reflect environmental change (Teece, Reference Teece2007). For example, Teece (Reference Teece2007) speculated about how established firms sensed new technology developed by others and reconfigured to exploit it.
However, a small but growing body of research looked beyond large, established businesses to small, start-up enterprises – to the context of entrepreneurship. The extension was a natural one because dynamic capabilities promote change (Helfat et al., Reference Helfat, Finkelstein, Mitchell, Peteraf, Singh, Teece and Winter2007). For instance, Eisenhardt and Martin (Reference Eisenhardt and Martin2000) suggested dynamic capabilities could be used to create new markets or industries. Corner and Wu (Reference Corner and Wu2012) induced, from qualitative evidence, the notion of dynamic entrepreneurial capabilities or capabilities that facilitate the amassing, configuring, and reconfiguring of resources needed to achieve enterprise viability. Additional empirical research explored dynamic capabilities developed by scientists when trying to commercialize technologies they invented (Newey & Zahra, Reference Newey and Zahra2009).
There is little research that examines dynamic capabilities in social entrepreneurship (Tashman & Marano, Reference Tashman and Marano2010) but scholars theorize that these capabilities are likely different for social as compared to commercial entrepreneurship (Short, Moss, & Lumpkin, Reference Short, Moss and Lumpkin2009; Dacin, Dacin, & Matear, Reference Dacin, Dacin and Matear2010). For example, scholars speculate that social enterprises require cross-sector alliancing capabilities like the ability to partner with nongovernment organizations (NGOs) in order to address social problems (Brugmann & Prahalad, Reference Brugmann and Prahalad2007). Tashman and Marano (Reference Tashman and Marano2010) describe dynamic capabilities for enterprises trying to serve those who live on less than $US 2 per day as decision-making processes and enterprise abilities and structures that enable enterprises to improve the value contribution of people suffering from poverty. We extend this nascent research by empirically examining dynamic capabilities in the context of social enterprises, where social mission is the primary goal of an enterprise (Mair & Marti, Reference Mair and Marti2006; Luke & Chu, Reference Luke and Chu2013). Specifically, we explore dynamic capabilities as social enterprises start-up and scale-up to new geographical contexts.
RESEARCH METHODS
The research design is a qualitative, multiple case, inductive study of social enterprises that scaled-up to a geographical context different from start-up. Such an approach is suitable for underresearched and poorly understood phenomena (Eisenhardt & Graebner, Reference Eisenhardt and Graebner2007) and for examining complex processes that unfold over time (Alvord, Brown, & Letts, Reference Alvord, Brown and Letts2004). Moreover, it has been recommended for understanding the microactivities underlying dynamic capabilities (Helfat et al., Reference Helfat, Finkelstein, Mitchell, Peteraf, Singh, Teece and Winter2007). All told, the design is well suited to the study of the scale-up of social enterprises.
This particular multicase study design reflects a replication logic wherein cases are analogous to a series of experiments and each case either confirms or disconfirms findings from other cases (Yin, Reference Yin2003; Graebner & Eisenhardt, Reference Graebner and Eisenhardt2004). The design includes three cases that display very similar conditions. In addition to scaling-up, these three cases have products/services that provide potable water for impoverished groups in less-developed regions. In particular, the less developed context means that these cases faced similar resource poor contexts when starting and scaling-up. The fourth case is similar to the first three in that it too is a social enterprise that scaled-up. However, it varies with respect to an important condition – it started and scaled-up in the United States, a developed country context where resources are generally more prevalent. We include a case that varies conditions somewhat to assess the robustness of the findings across a wider set of conditions (Yin, Reference Yin2003; Graebner & Eisenhardt, Reference Graebner and Eisenhardt2004). As such, we employ a design similar to that used by Graebner and Eisenhardt (Reference Graebner and Eisenhardt2004) in a study that examined three cases of very similar acquisitions within an industry and one that varied acquisition conditions. Collectively, the cases enable identification of patterns in scaling-up social enterprises that are suitable for large sample, quantitative testing in future research (Yin, Reference Yin2003; Eisenhardt & Graebner, Reference Eisenhardt and Graebner2007).
Cases and data collection
Collecting data for social enterprises can be exorbitantly expensive (Alvord, Brown, & Letts, Reference Alvord, Brown and Letts2004) and poses perhaps the greatest challenge to researchers interested in the topic (Short, Moss, & Lumpkin, Reference Short, Moss and Lumpkin2009). This issue likely stems from social enterprises occurring across multiple national contexts, sizes of organizations, and varied product/service offerings (Bacq & Janssen, Reference Bacq and Janssen2011; Desa & Basu, Reference Desa and Basu2013). Short, Moss, and Lumpkin (Reference Short, Moss and Lumpkin2009) advocate creative solutions when sourcing data and we adopt one such solution, the use of existing accounts (Alvord, Brown, & Letts, Reference Alvord, Brown and Letts2004). We thus identified existing social enterprises that had scaled-up to new geographical contexts and developed cases as follows. First, cases were identified through the oikos Global Case Competition in the social entrepreneurship category supported by Ashoka in 2009 and 2010. We selected cases that scaled-up from one geographical context to another. As such, similar findings across cases could be considered themes or patterns (Eisenhardt & Graebner, Reference Eisenhardt and Graebner2007) showing how social enterprises amassed and configured resources for scale-up.
Second, we searched for data sources available in the public domain and found a variety of sources including the following: interviews with founders, media articles, company, and NGOs reports, educational materials, and so forth. Data from multiple sources counteracts any potential bias from a singular data source (Eisenhardt & Graebner, Reference Eisenhardt and Graebner2007). Table 1 describes cases and lists data sources. To classify the cases, we turn to the Four Lenses Strategic Framework (2017). This framework provides theoretical constructs that provide the basis for debate and critical thinking about social enterprises. Under this framework, all four cases can be classified as mission-centric social enterprises, given that their missions were the reason for founding and the sole focus of enterprises’ activity. In particular, the table shows the product/service developed by each case to address a social problem. The first three cases, Wello, Playpumps International, and World Health International (WHI), had products/services designed to provide potable water for impoverished groups. These cases started and scaled-up in less-developed countries. The fourth case, The ReUse People (TRP), developed a deconstruction service to keep building materials out of landfills. The enterprise also recycled building materials, and trained and employed at-risk youth and ex-convicts in deconstruction. It started up in one region of the United States and scaled-up to another. Finally, the cases reflect various organizational forms as follows: Wello a nonprofit, Playpumps a hybrid structure of for-profit and nonprofit entities; WHI a for-profit; and TRP a nonprofit.
Case descriptions
Note. NGO=nongovernment organization; WHI=World Health International; TRP=The ReUse People.
Data analysis
NVivo software was used as a data management tool. We began analysis by constructing chronologies from gathered evidence for each case (see the Online Appendix). Chronologies included milestones in the founding of enterprises upon which cases were based and corroborated two important events: start-up and scale-up into a new geographical context. Chronologies thus provided a temporal structure for evidence and we mapped patterns from the data onto the start-up and scale-up events as described below.
Next, we completed within-case analysis, developing constructs and relationships to describe the process whereby resources were amassed and configured for start-up and scale-up in each case. We began with coding and Table 2 documents codes and provides illustrative quotes. We first applied open codes which involved reading and re-reading data and identifying passages consistent with open codes informed by the literature and an initial reading of data. In this way, evidence was bracketed into a few, very broad categories (Creswell, Reference Creswell2007). Open codes are used as an initial step to abstract from qualitative data (Creswell, Reference Creswell2007) and in this study included: resources, business models/strategies, and partners/stakeholders (see Table 2). We then completed expanded coding whereby open codes were parsed into more fine grained and elaborate codes (Creswell, Reference Creswell2007). Coding also involved memoing or making notes about constructs and relationships between constructs (Creswell, Reference Creswell2007). When constructs and relationships began to emerge, we took note but did not further refine them until analysis of separate cases was complete.
Literature-derived open codes and associated expanded codes
Note. UV=ultraviolet; WHI=World Health International.
Cross-case analysis involved comparing findings from one case with findings revealed in the second case and so forth, across all the cases (Eisenhardt & Graebner, Reference Eisenhardt and Graebner2007). Our purpose was to identify constructs and relationships that were similar, or replicated, across cases. For example, Wello sought and successfully used volunteers so we looked across the other cases to see if the finding, ‘use of volunteers,’ was replicated and it was. We used tables and other analytical devices to facilitate this comparison (Miles & Huberman, Reference Miles and Huberman1994; Eisenhardt & Graebner, Reference Eisenhardt and Graebner2007). The approach kept us strongly connected to the data so that we did not prematurely abstract from evidence. Finally, we mapped constructs and relationships revealed by cross-case analysis onto the two major events over time for these cases: start-up and scale-up into a new geographical context.
Trustworthiness of evidence
We addressed trustworthiness following techniques recommended by Lincoln and Guba (Reference Lincoln and Guba1985). Transferability was dealt with by providing descriptions of cases so that readers could assess the extent to which findings generalize to other social enterprises (see Table 1 and the Online Appendix). To ensure credibility, we used multiple data sources and peer debriefing wherein a peer not involved in the research was asked to probe our approach to analysis and findings. The peer was an experienced qualitative researcher employed by QRS International to train researchers in the use of NVivo. Specifically, she reviewed coding of the data and the identified patterns, thereby also ensuring dependability of results.
FINDINGS
How are resources amassed and configured for start-up and scale-up of social enterprises? We identify patterns across cases that address this research question and suggest dynamic capabilities required for social enterprises. Figure 1 provides an overview of findings, reporting particular patterns and dynamic capabilities revealed for start-up and scale-up of enterprises. The top half of the figure depicts dynamic capabilities and the bottom half shows the process of start-up and scale-up as well as outcomes achieved in these two phases. Details are given in the following sections and provide insight into why scaling-up social enterprises is challenging and relatively rare.
Dynamic capabilities involved in start-up and scale-up of social enterprises
Dynamic capabilities pre-startup
Studied enterprises enacted and reenacted common approaches that helped amass resources prior to start-up. We call these patterns, the ‘anteceding routines and processes’ (Tashman & Marano, Reference Tashman and Marano2010) that created the conditions necessary for more formal enterprise founding (Vohora, Wright, & Lockett, Reference Vohora, Wright and Lockett2004). We suggest the patterns are dynamic capabilities in the pre-startup phase as shown in the first box of Figure 1. These dynamic capabilities included the following: ideating business models, refining technology, and forming enterprises. Ideating business models involved imagining or conceiving of a product/service and the business processes needed to support it. Ideating was an emergent capability in that business models changed somewhat for all the enterprises. For instance, Wello’s first idea was to source water transportation devices from other enterprises and get them into the hands of the impoverished groups it was trying to help. Similarly, WHI’s first idea was to sell its ultraviolet (UV) water filter as a product to rural Indian families. Playpumps’ original business model was focused narrowly on manufacturing merry-go-rounds and did not include mechanisms to gather the large donations needed to install these water systems.
Figure 1 depicts these false start business models. Specifically, founders began enterprises with their original ideas for business models as represented by the arrow that emanates from the ‘dynamic capabilities pre-startup’ box. Founders discovered, in some cases, that initial business models did not work. They then returned to the pre-startup phase, as illustrated by the feedback loop emanating from the business model diamond in Figure 1. Founders again enacted dynamic capabilities characteristic of pre-startup, particularly ideating business models and (re)forming enterprises. For both Wello and WHI, the original business models were financially unsustainable. Wello learned that other organizations’ water transportation devices were far too costly and WHI discovered that one of its UV water filters was too expensive for an economically disadvantaged family to purchase. Both enterprises returned to the pre-startup phase depicted in Figure 1 to ideate business models again. Wello’s successful business model idea involved manufacturing and selling its own water transportation device, the water wheel, and WHI’s entailed housing the water filter in a kiosk that could provide water inexpensively for an entire Indian village.
Refining technology involved honing the mechanical and skill-based processes required to address targeted social problems. For example, the inventor of WHI’s UV water filter, Ashok Gadgil, refined the filter several times to make it extremely low cost to operate. Field-testing pointed out additional issues and the filter was refined once again. Playpumps merry-go-round was refined several times, relative to the original water system the founder saw at an agricultural fair. The best-known refinement was based on feedback from children who used the first two merry-go-rounds installed and wanted them to rotate in both directions. Wello designed and tested prototype water wheels in regions targeted for future sales.
Forming enterprises included all the legal, financial, and administrative tasks included in establishing enterprises as legal entities. Wello’s founder spent considerable time applying for a US 501c3, nonprofit status, believing that this form of organization would ensure the receipt of patient capital for the enterprise. WHI’s founder envisioned a for-profit company and raised large amounts of capital from private sector partners to help fund start-up costs. Playpumps’ founder quit his job after successful installation of the first two merry-go-rounds and courted investors to create a for-profit company.
The pre-startup dynamic capabilities ultimately enabled the start-up of the social enterprises, as represented by the arrow emanating from the ‘dynamic capabilities pre-startup’ box in Figure 1. The capabilities at this pre-startup phase were more about amassing resources and less about configuring resources. For example, resources were acquired to complete the development of what became WHI’s UV water filter and Wello’s founder drew on her own time and skills to apply for US 501c3 nonprofit status. Playpumps founder secured volunteered work-time from his wife and the inventor of the merry-go-round. Even the case from the more resource rich context, TRP, enacted these patterns. The founder secured a nonprofit, US 501c3, status for his enterprise and figured out how to make deconstruction economically viable.
However, amassing resources was challenging for these nascent enterprises. Many resources were acquired through two activities that we present in Table 3: mustering volunteers and scrounging donations. Mustering volunteers involved finding individuals or other organizations to donate skills needed for start-up. For example, another organization designed Wello’s website for free and engineers volunteered to design the water wheel. Scrounging donations was seeking funding through others’ generosity as Playpumps did by securing funding for the two prototype playpump installations from another organization, the Umgeni Water Company. The entries in Table 3, taken as a whole, show how others provided resources critical to enacting the basic conditions that enabled starting-up these social enterprises. Moreover, the resources others contributed weregiven without expectation of financial returns.
Patterns of microactivities for resource acquisition at start-up
Note. UV=ultraviolet.
Dynamic capabilities start-up
Social enterprises started-up in a single geographical context, attempting to turn a good idea into a product/service that could provide a social benefit to local people. It is important to note that the boundary between the pre-startup and start-up phases is not always clear-cut, similar to commercial start-ups (Vohora, Wright, & Lockett, Reference Vohora, Wright and Lockett2004). Nevertheless, founders enacted an important dynamic capability that cemented the start-up of enterprises as shown in Figure 1. Below, we describe this dynamic capability and describe outcomes from the start-up phase.
Experimenting with and building resource configurations
The products/services designed and built by the studied enterprises had to be encapsulated within resource configurations that effectively delivered the social benefit inherent in the product/service. Stated differently, resources had to be configured in order to fund, manufacture, and distribute products/services. Founders looked for stakeholders that had resources the fledgling enterprises needed and worked to incorporate those resources into an overall resource configuration. For example, Cynthia Koenig of Wello applied for several fellowships so she could be financially supported while starting up her enterprise. WHI secured investment from large, multinational corporations like Dow Chemical to fund the start-up of WHI’s water filtration kiosks for Indian villages and secured the services of NGOs to educate villagers on the health benefits of clean water. Playpumps formed a public/private partnership with South Africa’s Department of Water Affairs and Forestry to obtain access to ground water across the nation. Even TRP, the case in the more resource-rich US context, pursued a grant from a government agency in order to add a retail component to its deconstruction enterprise, a warehouse that sold recycled building materials.
There was some trial and error involved in assembling initial resource configurations. For example, Wello developed portable manufacturing equipment that fit in a shipping container but later located manufacturing in one city in Africa. Playpumps initially located sites for water systems but then outsourced this task to another organization. WHI recruited an NGO to help educate villagers on the benefits of using clean water because, initially, some villagers continued to get water from traditional sources instead of purchasing the very low-priced water from the WHI water kiosks.
Finally, Table 4 reports details of the resource configurations built by each social enterprise at start-up. There are two overarching points to be made about resource configurations for the start-up context. First, configurations included the resources of others such as donations (e.g., Wello, Playpumps, WHI) and grants (e.g., TRP), as already stated. These resources were given by others without the expectation of financial return and illustrated the ability of social enterprises to mobilize resources on behalf of their social mission, an ability that commercial enterprises do not have (Russo, Reference Russo2010). Arguably, enterprises would not have been able to start-up without these resources. Second, social enterprises, even after they had a basic resource configuration in place, had to fine-tune configurations for practical reasons (Wello, Playpumps) or to better serve their target population (WHI).
Resource configurations for start-up and scale-up
Note. NGO= nongovernment organization; WHI=World Health International; TRP=The ReUse People.
Outcomes from start-up
The start-up phase yielded outcomes for social enterprises as shown in Figure 1. First, enterprises created social benefit. However, it is important to remember that benefit was created not just by clever products/services but also through building resource configurations that supported these products/services (see Table 4, for descriptions of configurations). In sum, configurations across the cases reflected common imensions. All enterprises amassed resources to ensure they were able to do the following: manufacture/produce the product or service, fund activities, and disseminate/distribute products/services. The second outcome was a core product/service. Descriptions of these core products/services can be seen in Table 1 under the heading of ‘product/service.’ These are core products/services because they serve as a base for geographic expansion given their demonstrated viability in the initial geographic context.
Dynamic capabilities for scale-up
Identifying differences in the new context
Social enterprises began scale-up by identifying the ways in which the new geographic context was different from the initial one. Table 5 identifies common differences encountered by enterprises when trying to scale-up to new geographic contexts. Differences are listed in the first column and included dissimilarities in the following: governing structures; potential distribution channels; available funders and partners; and even the location of groups needing help – urban versus rural.
Key differences in start-up and scale-up geographic contexts
Note. NGO= nongovernment organization; WHI=World Health International; TRP=The ReUse People.
These differences had clear implications for the studied enterprises, the same resources simply were not available in both startup and scale-up contexts. Enterprises therefore required different stakeholders and resources as compared to start-up. The table also illustrates why enterprises did not have the scale-up options hypothesized in the literature. Scholars advocate branching (Dees & Anderson, Reference Dees and Anderson2004) or replication (Bradach, Reference Bradach2003) which involves enterprises duplicating their resource configurations in multiple geographic locations. Table 5 shows why duplication of resource configurations was not possible for these enterprises. All told, the table creates a rich picture of why social enterprises can be challenging to scale.
Tweaking core products/services
Data indicate that social enterprises, like their commercial counterparts (Russo, Reference Russo2010), tweaked or made adjustments in their core products/services for new contexts. For example, WHI housed its UV filter within aqua stores or small ‘mom and pop’ shops in the Philippines, in contrast to the kiosks it developed for Indian villages. The change was made because the impoverished groups needing access to very inexpensive water were located in cities, not in remote rural villages like India. Playpumps experimented with locations for its merry-go-rounds such as schools, clinics, and village centers. Wello modified its water wheel slightly for India and partnered with Seva Mandir, an Indian NGO, to field test it. Even TRP had to tweak its core service to scale from one region to another within the United States. It elected to train and license other contractors to operate in the new context instead of providing its own deconstruction crews. It also used Habitat for Humanity’s Restore shops to sell recycled building materials instead of constructing and operating recycling warehouses like the enterprise did in California. The tweaking pattern again illustrates why social enterprises did not have the scale-up options of branching or replication when expanding geographically, the core product/service had to be adapted for the new context.
Engaging with context-specific resources
Social enterprises engaged with stakeholders and resources specific to the scale-up context. For example, WHI arranged loans for owners of the aqua stores in the Philippines through the Rotary Club instead of providing loans out of WHI’s funds as was done in India. The Rotary Club was very active in the Philippines but did have a presence in India, WHI’s start-up context. Playpumps looked to Unicef to donate the funds for playpump installations in Swaziland when the World Bank prize had provided the needed donations in South Africa. Wello partnered with different NGOs to test the water wheel in India versus Africa and planned for distribution channels unique to each context. At first glance, this finding may seem like common sense, suggesting that different resources would have to be used when crossing national borders. Even TRP had to engage with different resources when expanding to a new region within the United States. However, further consideration of this finding begs the question of why theorized scale-up strategies emphasize standardization. Why is the notion of duplication of products, business models, and resources so persistent in the existing literature?
Scale-up outcomes
Data provide evidence of three outcomes from scaling-up social enterprises and these are reported in Figure 1. First, enterprises ultimately built resource configurations idiosyncratic to the scale-up context. Table 4 provides details of the resource configurations created for scale-up (see last column) along with configurations built for start-up so a comparison can be made. Different configurations were necessary because enterprises were unable to replicate configurations from initial geographic contexts. We see this as a second act of social entrepreneurship given this phenomenon is defined as the innovative use and combination of resources to address social needs (Mair & Marti, Reference Mair and Marti2006). In particular, scaling-up the studied enterprises involved innovation beyond that required for start-up and it is innovation that characterizes social entrepreneurship (Luke & Chu, Reference Luke and Chu2013). The implications of this finding are discussed in the following section. Second, enterprises enhanced the robustness of their core products/services through tweaking them for the new context. Successful tweaking demonstrated product/service hardiness and ability to extend product/service reach. Third and most importantly, the social enterprises accomplished the geographical dispersion of the social benefit they initially created. Such dispersion is a widely touted as a key mechanism for addressing social problems but is rarely achieved (Dees, Reference Dees2008).
Discussion
The paper empirically explored the scaling-up of social enterprises from the perspective of dynamic capabilities, a framework from the strategic management literature that focuses on patterns in acquiring, configuring, and reconfiguring resources for an enterprise. We thus addressed the research question: How are resources amassed and configured for start-up and scale-up of social enterprises? Scale-up, the expansion of a social enterprise from one geographic context to another, requires understanding the mobilization of resources in both start-up and scale-up contexts. Findings shed light on scale-up – which is one of the least understood but most important outcomes of social entrepreneurship (Alvord, Brown, & Letts, Reference Alvord, Brown and Letts2004; Smith, Kistruck, & Cannatelli, Reference Smith, Kistruck and Cannatelli2016). The outcome is important because it is necessary for the impact of social entrepreneurship to match the magnitude of the social problems being addressed (Dees, Reference Dees2008).
Findings specifically identified many dissimilarities between the scale-up and start-up geographical contexts. Scale-up therefore involved enterprises tweaking their products/services for new contexts and building novel resource configurations that reflected scale-up contexts’ idiosyncrasies. As such, our findings offer some reasons why enterprises rarely scale-up beyond their initial community or region (Dees & Anderson, Reference Dees and Anderson2004; Smith, Kistruck, & Cannatelli, Reference Smith, Kistruck and Cannatelli2016), extending knowledge of scale-up beyond existing research which is predominantly conceptual and anecdotal. Findings also revealed the substantive differences in resource configurations assembled in the start-up and scale-up contexts as shown in Table 4.
Findings have three implications for the wider literature. First, findings call into question the emphasis in existing research on standardization in order to scale-up social enterprises to new contexts. For example, standardization is advocated through scale-up strategies like branching (Dees & Anderson, Reference Dees and Anderson2004) and replication (Bradach, Reference Bradach2003) both of which call for the duplication of a social enterprise’s business model in multiple geographic contexts. Bradach (Reference Bradach2003) even recommended identifying a standard context into which social enterprises can be expanded. However, current findings call into question the extent to which standardization is possible. Products/services had to be tweaked and entire new resource configurations had to be constructed to support them when scaling-up to a new context. Stated differently, findings show that social enterprises could not standardize their business models – how they manufactured, delivered, educated customers about, and funded their core products/services – due to the idiosyncratic characteristics of the scale-up context. Future research could identify social enterprises that are able to standardize business models for scale-up and compare them to enterprises that require new business models. Such research could begin to clarify when and how more standardization might be possible.
Second, findings have implications for Dees & Anderson’s (Reference Dees and Anderson2004) other two strategies for scaling-up, namely dissemination and affiliation. The social enterprises in this study did not use either as an overarching strategy to scale-up to a new context. The differences between initial and scale-up contexts precluded a generic resource configuration that could be disseminated or spread through affiliation. However, we did see affiliation used to scale-up a constituent of an enterprise’s resource configuration. Playpumps initially located sites for its water installations but, with expansion, relied on another organization to do so. Current findings thus suggest that scale-up is a complex phenomenon that may not be accomplished through singular, broad-brush strategies. Future research could take a more fine-grained approach, considering the separate constituents of resource configurations such as resources needed for product/service distribution. Scholars could explore the efficacy of strategies such as dissemination or affiliation to scale-up constituents of resource configurations to new contexts.
Third, the dynamic capabilities perspective applied in this study extends our understanding of social entrepreneurship. In particular, findings reveal how each enterprise had to develop two different resource configurations given the dissimilarities between start-up and scale-up contexts. Arguably, the construction of two different configurations could be seen as two acts of social entrepreneurship, given this phenomenon is defined as the innovative use and combination of resources to address social needs (Mair & Marti, Reference Mair and Marti2006). Evidence documenting the differences in resource configurations across the two contexts provides support for this idea and suggests a reason why so few social enterprises scale-up – perhaps scaling-up reflects multiple acts of social entrepreneurship. Future research could examine this idea from the point of view of social entrepreneurs who have scaled-up enterprises from one geographical context to another. To what extent do such entrepreneurs see scale-up as an additional act of entrepreneurship? Are there similarities between serial entrepreneurs in the commercial context and social entrepreneurs who have scaled-up their businesses to multiple geographic contexts?
Findings also imply that social entrepreneurs engage in bricolage or the use of resources that are free or inexpensive (Baker & Nelson, Reference Baker and Nelson2005; Desa & Basu, Reference Desa and Basu2013). The founders in this study employed the bricolage-like activities of mustering volunteers and scrounging donations for start-up. Under what conditions would bricolage be applied in scale-up? Bricolage was also relied on most for Wello, a nonprofit, and least for WHI, a for-profit company. So, another research question that could be explored is to what extent does the organizational form of a social enterprise influence the use of bricolage to mobilize resources?
In addition to scholarly implications, the research offers insights for practitioners wishing to found and scale-up social enterprises. Evidence suggests using an initial context to refine product/service ideas and to adjust and prove the business model. Most importantly, our findings call into question the prevailing advice to practitioners which is to standardize products and business models. Evidence instead encourages social entrepreneurs to accept that new geographical and initial contexts will likely be different and their focus should be on identifying differences, and seeking to further adjust business models and resource configurations accordingly. They might also be warned that expansion could be construed as a second act of social entrepreneurship, warranting additional innovation, commitment, and effort.
Like all research, our study has limitations to consider when interpreting findings. The first is the use of existing accounts to construct cases. The cases provide benefits like availability of data and generalizability (to theory) beyond a single national context and single product/service (Alvord, Brown, & Letts, Reference Alvord, Brown and Letts2004). Nevertheless, they limit data to secondary sources and our ability to gain precisely comparable data across cases. Although secondary data revealed some nuances with respect to scaling-up of social enterprises, it is possible that primary data would have provided a thicker description that yielded further nuances. A second limitation is that three of the cases were about providing potable water to impoverished peoples. In part, external validity is enhanced by the widespread nature of this social problem globally and the extent to which this fundamental need is being tackled through social entrepreneurship. The fourth case that varies conditions somewhat enhances the generalizability of findings to theory. Nevertheless, findings are best thought of as induced theory suitable for testing via large sample, quantitative research designs.
In conclusion, this paper takes a promising initial step in empirically examining the scaling-up of social enterprises. Findings provide insight as to why it is difficult for founders to implement the scale-up strategies advocated in existing literature. On a positive note, however, findings elaborated patterns that did ultimately enable scale-up. These patterns suggest mechanisms that can be tested in future research as well as provide practical advice for social entrepreneurs seeking to scale-up to realize wider social benefit.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/jmo.2018.38
Acknowledgement
None.
Conflicts of Interest
None.
Funding
This research received no specific grant from any funding agency, commercial, or not-for-profit sectors.
About the authors
Patricia (Trish) Doyle Corner is the Professor of Strategic Entrepreneurship at Auckland University of Technology in Auckland, New Zealand. Her current research interests include commercial and social entrepreneurship, innovation, and advancing human knowledge.
Kate Kearins is the Professor of Management and Dean of the Faculty of Business, Economics and Law at Auckland University of Technology, New Zealand. Kate is interested in business and sustainability, entrepreneurial enactment, and stakeholder collaboration.
