Scaling in social enterprises: The case of digital social innovations

Ozman, Müge; Gossart, Cédric

doi:https://doi.org/10.59876/a-axv5-1zpk

The rapid rise of digital platforms addressing social and ecological issues across the world reflects changes in the ways societies deal with these issues. Digital platforms span across various sectors like, among others, governance and civic engagement (Coleman & Blumler, 2014; Gilman, 2017), civic crowdfunding (Davies, 2015), crowd science (Franzoni and Sauermann, 2014), collaborative consumption (Möhlmann, 2015), or climate change (Pearce et al., 2019).

Digital platforms open new possibilities in enhancing society’s capacity to act, facilitate participation to generate new ideas, mobilise actors’ collaboration, and enable them to use their own resources and capabilities more efficiently. However, the extent to which digital social enterprises can leverage these possibilities depends on their capacity to scale social impact, which is defined as the “process of expanding or adapting an organization’s output to better match the magnitude of the social need or problem being tackled” (Desa & Koch, 2014, p. 48). While an important body of literature examines scaling for offline social enterprises, research is limited for those aiming to achieve social goals using digital technologies. In this paper we aim to fill this gap. Entrepreneurs involved in digital social innovations (DSI) have different sets of resources, skills and tools available at their disposal compared to their offline counterparts. Consequently, some strategies of scaling identified in the literature for offline social enterprises might be more challenging to undertake, while others can be more readily achieved thanks to digital technologies.

Research on scaling distinguishes between two overarching strategies (Desa & Koch, 2014; Islam, 2022, Westley et al, 2015). While the first one relies on reaching out to a wider audience through organizational growth, the second one is based on catalysing welfare-enhancing systemic change through institutional and/or community focused initiatives (Islam, 2022). However, the former does not guarantee the latter, as expanding the user base is seldom equivalent to the social innovation’s potential for deeper transformations (Uvin et al., 2000; Nardini et al., 2022). This dilemma can even be more pronounced in the case of digital platforms as they rely on specific algorithms, increasing their potential to reach a maximum number of people. On the other hand, the same processes that increase efficiency of diffusion can cause difficulty when adapting to local culture and institutions, hence reducing platforms’ efficacy. Consequently, for deeper societal transformations the scaling strategies of digital enterprises have to consider not only the material technology underlying the platform, but also the fit between this technology and the local context where it is used.

Based on these considerations, in this paper our research questions are the following: (1) How do digital technologies facilitate or hinder scaling social impact in digital social enterprises? (2) How do digital social enterprises scale social impact?

Addressing these questions will help managers develop strategies that leverage technology to diffuse on a larger scale, while not jeopardizing their potential for deeper societal transformations. It will also help them better evaluate their strengths, pinpoint their weaknesses, and develop new skills in line with their social goals.

In addressing these questions, we are particularly concerned with platforms generating social innovations. To unleash them, empowering actors is particularly important (UNDP, 2020, Chapter 4), which is why we are focusing in this study on civic engagement platforms. They have been influential in generating social innovations for a wide range of social and public issues (Saldivar et al., 2019).

To address the first research question above, we were guided by the following motive: how to apply offline scaling concepts to civic engagement platforms? For this purpose, an in-depth literature survey on scaling social impact in offline social enterprises was first carried out. This survey revealed three broad mechanisms to scale social impact: scaling out, scaling up, and scaling deep (Bolzan et al., 2019, Moore et al., 2015; Westley et al., 2014). Scaling out corresponds to the replication of solutions in different geographical locations. Scaling up refers to macro level transformations in institutional and policy frames. As for scaling deep, it is community oriented and refers to the extent to which social innovation transforms places by changing the beliefs, practices and culture of people in specific geographical places (Bolzan et al., 2019; Moore et al., 2015; Westley et al., 2014; Nardini et al., 2021). By drawing upon the past research on civic engagement platforms and on the economics of digital platforms, we developed 10 indicators of scaling social impact for these three types of scaling, in the case of civic engagement platforms.

While it is important to develop theoretically grounded indicators for scaling in digital enterprises, it is not sufficient. Such a study would not provide insights into actual scaling strategies used by digital social enterprises. What types of strategy mixes do they use? Is it possible to differentiate platforms based on their scaling strategies? Therefore to address our second research question above (How do digital social enterprises scale social impact?), we carried out an empirical analysis to explore their scaling strategies by collecting data on the 10 aforementioned indicators for 189 civic engagement platforms in Europe. We used clustering analysis, which enables the identification of qualitatively different subgroups within populations, with respect to a set of indicator variables (Weller et al., 2020). The clustering analysis revealed three broad classes of digital social enterprises, according to their scaling strategy. These are, data-driven, technology-driven, and grassroots-driven scaling.

The paper is organised as follows. In the second section we provide a definition of civic engagement platforms and present the theoretical framework in the context of previous research on scaling in offline social enterprises. We also highlight the differences between offline social enterprises and digital ones that might influence scaling activities. In the third section, we explain the methodology and data used in the study. The fourth section is allocated to results and discussion. The final section concludes with policy and managerial implications, and directions for future research.

Theoretical background

General context: civic engagement platforms

We take a “digital platform” as a governed technology, in the sense that its organisation and functionality (design), as well as data analyses and algorithms automating the management of stakeholders, are essentially embedded in the digital code (Kitchin, 2017). These digital platforms can be deemed specialised algorithms (ibid.) designed to enable people to exchange, access and provide data for local development; share opinions, views, resources and skills; participate in the governance of their communities; and address problems related to social and ecological problems faced in their living environments. We take DSI as a generic term concerning a variety of different sectors and social problem areas (Ozman & Gossart, 2017). In this paper, we are concerned with civic engagement platforms, which aim to increase the capacity of civic society to formulate a social or environmental problem, bring it to the fore of public arenas, and engage a variety of stakeholders to jointly frame and solve this problem (Ozman & Gossart, 2019). An example is “civic tech”, which refers to platforms that are connecting governments and citizens (e.g. https://get.flui.city ). In civic engagement platforms, solutions might emerge through the self-organisation of a community, the increased capacity of direct communication with other stakeholders, or the expansion of their reach (Gilman, 2017).

This definition of civic engagement platforms also builds upon the social innovation literature. Although social innovation is a contested concept (Ayob et al., 2016), there are regularities across definitions. First, social innovations explicitly aim at solving societal (social &/or ecological) problems. Second, like standard innovations they bring novelty to their local context. Third, in most cases they facilitate a redefinition of power relations between actors (Moulaert et al., 2005; Mulgan, 2006). What distinguishes civic tech is their reliance on digital technologies, which can serve users’ needs all over the world or meet specific needs tied to a local context.

Table 1

**Societal issues addressed by DSI enterprises**

Table 1 shows the main societal issues of civic engagement platforms that we take in this paper, with some examples from Europe. In particular, we distinguish between four fields of activity.

The first field is governance, where platforms enable citizens to participate in open debates and provide opinions and discussions. We take civic agency similar to Allen & Light (2015), as “the activities through which people pursue political efficacy”, without necessarily involving in formal political membership or having a legal status. As such, civic agency is “oriented towards how people live together –whether locally or globally–, shape their worlds together, especially in conditions of diversity, working both through and outside political institutions” (ibid, p. 5).

In the “Governance” societal issue are also included platforms focusing on local issues in a particular city or town, and facilitating idea generation and implementation in the local context through the participation of residents to local governance. Digital platforms matching associations searching for volunteers with candidate volunteers are also included in that issue (e.g. https://www.tousbenevoles.org). Finally, online petition platforms (Puschmann et al., 2017) are also included.

The second issue concerns digital platforms promoting transparency and information disclosure to address problems related to social, institutional and geographical issues (Fung et al., 2007) (e.g. https://alaveteli.org). Given that many governments put in place transparency policies to foster the performance of institutions, digital technologies have been widely used to disclose information in many spheres like food safety (e.g. https://fr.openfoodfacts.org), financial information (e.g. https://opencorporates.com), or physical accessibility conditions in geographical places for people with disabilities (e.g. https://wheelmap.org).

The third field concerns “Finance”, which includes civic crowdfunding platforms (e.g. https://www.spacehive.com). These DSi are mainly designed for civic projects:

“A civic crowdfunding project can be defined as one that develops a shared resource that is accessible to the community either as a public asset, a community-owned resource or a public-private partnership, and may or may not involve government.”
Davies, 2015, p. 83

Finally, social inclusion platforms facilitate inclusive living and working conditions for vulnerable people, as Mon-copilote for people with disabilities (cf. https://mon-copilote.com).

Scaling social impact in social enterprises

Scaling has been deemed a major indicator of performance for social enterprises (Bloom & Smith, 2010; Dees et al., 2004; Islam, 2022). Desa & Koch (2014) define scaling as “process of expanding or adapting an organization’s output to better match the magnitude of the social need or problem being tackled” (p. 48). Research about scaling social impact for social enterprises spans across various disciplines, including management and entrepreneurship (Bauwens et al, 2020; Desa & Koch, 2014), marketing (Bloom & Smith, 2010; Epstein & Yuthas, 2012), third sector studies (Lyon & Fernandez, 2012; Uvin et al., 2000), and social innovation (Westley et al., 2014; Mair et al., 2023; Bolzan et al., 2019). This literature broadly highlights three types of scaling (Moore et al., 2015; Nicol, 2021), which are scaling out, scaling up and scaling deep. In the next sections, we examine each of these in more detail, in relation to digital platforms.

Scaling out

Scaling out refers to various mechanisms used by social enterprises to replicate a successful model (Bolzan et al., 2019; Moore et al., 2015; Westley et al., 2014), so that small enterprises operating at local scale evolve into national or global organizations (Nardini et al, 2022; Islam, 2022, Lyon & Fernandez, 2012). Various terms have been used to refer to this type of scaling including organizational growth (Islam, 2020), horizontal integration (Uvin et al., 2000), breadth scaling (Desa & Koch, 2014), quantitative scaling, or organizational scaling (Uvin,1995; Andre & Pache, 2016). This type of scaling has been criticised on the grounds of undermining social enterprises’ capacity to address societal problems, by focusing solely on the growth of the social enterprise itself (Lyon & Fernandez, 2012; Nardini et al., 2022; Westley et al., 2014; Uvin et al., 2000).

Common strategies for scaling out: Scaling out usually involves strategies like diversifying product and services, franchising (Tracey & Jarvis, 2007; Lyon & Fernandez, 2012), or duplicating programs to serve beneficiaries in a broader geographic area (Islam, 2022; Bauwens et al., 2020). Here, the organization increases its efficiency of diffusion by leveraging economies of scale (Desa & Koch, 2014).

The case of digital social enterprises: These can scale out fast thanks to their technology, which helps them leverage economies of scale. Digital platforms classify, sort, categorise, and manage interactions through the algorithmic management of the data they produce (Gillespie, 2014; Steiner, 2012). The more data are collected, the more efficient is algorithmic management. This explains why platforms favour the replication of standardised algorithms in different locations: thanks to digital technologies diffusion is faster and yields higher efficiency gains, sometimes at the expense of local practices and institutions. As a result, DSI scaling bears important differences compared to offline social innovations. By definition, the central role of digital technology facilitates scaling out, but it also incurs costs of adaptation to local contexts, thereby reducing its potential to scale up and scale deep.

In the case of digital platforms, there are additional mechanisms that can make diffusion even faster, notably by leveraging network externalities. The performance of most platforms depends on the extent to which they facilitate interactions between participants, by matching users who benefit from what others offer (Evans & Schmalensee, 2016; McIntyre et al., 2021; Gawer, 2009; Piskorski, 2011). Most digital platforms rely on a multisided model, where there are typically more than one group of users engaging in mutually beneficial interactions through the platform (Rochet & Tirole, 2003). Therefore, the utility obtained by one group of users depends on the number of participants in the other group. For example, for a platform matching association with volunteers, the more associations that can be reached through the platform, the more volunteers it will attract, and vice versa. This mechanism, called network externalities (Arthur, 1989; David, 1986), initiates a self-reinforcing growth cycle of the platform, which can be strategically exploited to rapidly increase its number of users.

Another mechanism that can facilitate scaling out is crowdsourcing data from many different sources as in the case of open data. Many people can participate to such data pooling because of low barriers to entry, and the data is open to anyone wishing to access information resources. The more data is available, the more new users are drawn to it, maintaining a self-reinforcing growth mechanism. Such data pools are relevant to tackle problems that can effectively be solved by collecting dispersed pieces of information, as well as problems that require collective intelligence. User commitment can range from collaborative to spot users (e.g. Wheelmap, Openfoodfacts).

Platforms that scale out can operate anywhere in the world thanks to digital technologies, hence gaining access to a very large pool of potential users. This requires capabilities to reach out to a global audience as well as specific algorithmic practices. However, not all social entrepreneurs wish to scale out. Contributing to institutional change and empowering their local community may suffice them (Andersson, 2020).

Scaling up

The second type of scaling is concerned with the extent to which a social enterprise can develop and sustain a supportive ecosystem at macro or meso levels around the problem tackled (Islam, 2022). Termed “scaling up” (Westley & Antadze, 2010), catalysing institutional and policy-level change through networks with ecosystem actors is at the heart of this scaling process. Other terms have also been used for this type of scaling, including political scaling (Uvin, 1995) or depth scaling (Desa & Koch, 2014). Scaling up a social innovation to a supporting ecosystem refers to the effectiveness with which social innovations contribute to transformative social change (Avelino et al., 2017; Westley et al., 2014). Here the aim is catalysing institutional and policy-level change at the macro level. Scaling up strategies support the ecosystem as a whole (Islam, 2022), by “bringing in other organizations, corporations, and/or institutions to help spread the social impact” (Weber, 2012).

Common strategies for scaling up: Social enterprises carry out advocacy activities to influence policy and legislations around the cause, organize campaigns that involve other actors to influence public policy makers and other organizations (Islam, 2022; Uvin et al., 2000; Uvin, 1995; Bloom & Smith 2010; Bauwens et al., 2020), as well as participate in awareness raising campaigns around a given cause (Islam, 2022). Another way to scale up involves developing infrastructure for the benefit of other actors in the ecosystem (Islam, 2022). This might include developing a welfare-enhancing technology made available to other organizations, providing open-source materials (Lyon & Fernandez, 2012; Bhatt et al, 2016), or developing ICT infrastructure to be used by ecosystem actors (Bhatt et al., 2016).

The case of digital social enterprises: Empowering actors is particularly important to unleash transformation (UNDP, 2020, chapter 4), which is easier when one has access to technological tools. Digital social innovations’ building blocks are the material technologies aiming to address a social or ecological problem. Making these technologies available to other organizations contributes to building and growing an ecosystem around the given problem, by helping other organizations address the same problem. While they can scale up by providing their technology to other organizations, whether the latter can adapt the solution to their own context highly depends on whether the technology is open source or not.

Digital platforms based on centralised proprietary algorithms tend to adopt a closed participation model (Eisenmann et al., 2009), and give little room to the participation of users (Gümüsay et al., 2022), and to other social enterprises so that they can adapt the software to their own needs and context. On the other hand, open source projects have another model, seen as a paragon of user-as-innovator collaborative projects (von Hippel & Krogh, 2003). Being more transparent and realised through the participation of a distributed set of actors, they can be used to prevent central platform owners to decide alone what is revealed or kept secret when the platform is adopted by different parties. The governance of digital platforms based on open source codes tends to be more democratic, in the sense that they permit developers and local stakeholders to incorporate locally anchored knowledge into the algorithm. This knowledge is highly relevant to address local problems or find local solutions to grand challenges. Such objectives tend to be more easily reached with open-source technologies that are available to all, since “Collective agency has the greatest potential to change social norms” (UNDP, 2020, Chapter 4, p. 145).

In addition to technology provision and open sources choices, digital technologies can help for advocacy and awareness raising, notably by enabling the diffusion of information on a large scale and establishing communication channels between different actors (Allen & Light, 2015). Some civic engagement platforms facilitate advocacy by helping other organizations diffuse information, or access information and resources that support advocacy activities. In addition, beyond technical support to its users, a social enterprise can be involved in advocacy work itself. For example, Opencorporates is a digital platform that supports corporate transparency by providing open data on corporations, which can be used by other organizations to tackle corruption and criminality. But Opencorporates does not only support other actors to accomplish their missions, it also carries out “data-based advocacy” (Charalabidis et al., 2018).[1] In the case of the French DSI OpenFoodFacts, in which users provide data about the ingredients present in industrial food products, it is not only backed by a dynamic and international open-source community, but it is also involved in several movements promoting healthy nutrition. It can do so through a very active network strategy, enabling it to work with many local civic collectives.

Scaling deep

If scaling up aims at global and macro level changes, scaling deep is local and community oriented (Bolzan et al., 2019). It refers to the extent to which social innovations transform local places (Moore et al., 2015). With this strategy, DSI empower local communities by involving them in the transformation process by “changing relationships, cultural values and beliefs” (Moore et al., 2015, p. 75). This enables them to tackle social or ecological problems, since shaping established worldviews, values or identities tends to reinforce “environmentally significant individual behaviors” (Stern, 2000).

Here, learning and connectedness are critical (Desa & Koch, 2014). Scaling deep involves second order learning (Geels & Schot, 2007; Schot & Steinmueller, 2018; van Mierlo & Beers, 2020), by which actors change their worldviews, values or identity. Second order learning practices deepen the capacity to reshape established practices, beliefs, and routines that triggered social problems at first hand (Seyfang & Haxeltine, 2012), such as the belief that agency to solve grand challenges only lies in the hands of governments. These are important, as they can ultimately change people’s practices regarding how they deal with socio-ecological problems in the long run.

Common strategies for scaling deep: These include aiming at social change in physical places (Moore et al., 2015), fostering collaboration between community members, and building community cohesion through social capital (Emery & Flora, 2006). These are essential because creating social bonds within neighbourhoods helps effective community organizing, and social change begins locally (Mandell, 2010). In addition to fostering exchanges between community members (Bloom, 2021), transformative learning opportunities can be provided through joint training centres, training and apprenticeship programs (Lyon & Fernandez, 2012; Islam, 2022). It has been highlighted that scaling deep requires a very good understanding of localities to generate social change (Nardini et al., 2022; Desa and Koch, 2014).

The case of digital social enterprises: For digital social enterprises, scaling deep can be difficult to achieve, because digital technologies are by definition related to “killing the distance” (Cairncross, 1997). While digital and real spaces are now very much intertwined and can be taken as a “social fact” à-la-Durkheim (Ling, 2012), when it comes to social and ecological problems the distinction is sharp. Indeed, problems such as ecological damages, poverty, health or inequalities are experienced in physical spaces. Social entrepreneurs need to work actively with target populations in the field to develop solutions and encourage the participation of stakeholders at local level. However, in the case of DSI, social entrepreneurs’ risk is encapsulated in technological bubbles isolated from other actors operating in real life, which eventually undermines their capacity to reach the social goals they aim for. In this sense, their digital attributes might turn into a disadvantage. It is therefore essential to better understand the benefits and risks yielded by using digital technologies to scale digital social enterprises. To balance this adverse effect of their technology, social entrepreneurs need to learn how to deepen social capital and infrastructural capacity in the local social economy, by forming new networks or strengthening existing ones in offline contexts, and not be confined to the online sphere.

While there are difficulties in scaling deep for digital platforms, there are also different ways for them to leverage digital technologies to address community problems and foster community cohesion. For one thing, a platform can focus on social projects seeking to create opportunities for local revival. For example, some civic crowdfunding platforms focus exclusively on local revival projects (e.g. https://miimosa.com). In addition platforms can strengthen community cohesion, as in the case of social inclusion platforms that help neighbours exchange resources, skills, or support socially excluded community members (e.g. https://www.entourage.social). In the latter case, physical exchanges between people are fostered, which is important for community cohesion (Quan-Haase et al., 2002). In some cases, platforms facilitate topic-specific gathering activities, around which individuals organise social relations. These are local foci, the latter being a “social, psychological, legal or physical entity around which joint activities are organized” (Feld, 1981), which can generate social capital (Quan-Haase et al., 2002). These events attract people sharing similar interests and facilitate social integration, as by attending similar foci, actors deepen common domains of interest. Moreover, such event organisations can reach communities not using digital platforms.

In these cases, digital platforms can change the way people perceive their agency and efficacy and actually exercise it, strengthening what Cantijoch et al. (2015) describe as “community efficacy” – i.e. people’s beliefs that they can actually make a difference in the community they live in, and thus contribute to the creation of social capital (Quan-Haase et al., 2002). In the case of civic political engagement, searching for information about problems and ways to solve it can raise awareness and improve citizens’ understanding about the political process, and be used to solve other problems (spillover effect). So can simple actions of online activism. For example, “After the September 11, 2001 terrorist attacks (…) some people felt politically active for the first time—because they were on the Internet.” (McCaughey & Ayers, 2003, p. 6).

In addition, physical exchanges can enable second order learning (changes in worldviews, values and identities). Learning consists of the acquisition of new knowledge and knowledge is sticky (Szulanski, 2003), in the sense that it is very much tied to a specific person. Unless that person makes the effort of codifying her tacit knowledge, physical exchanges are key to knowledge transmission, since “tacit knowledge can be acquired only through shared experience, such as spending time together or living in the same environment” (Nonaka et al., 2000, p. 9). The same goes for the acquisition of new life experiences, whose acquisition through face-to-face interactions mobilises emotions that are essential to learning processes:

“Emotions (…) involve ‘psychological energy, transmitted by feelings, emotions, attitudes and motivations which both mobilize and, at the same time, are conditions that may be influenced and developed through learning’ [Illeris (2002, p. 18)].”
cited in Merriam et al., 2020, p. 97

While learning effects can be achieved through building community, some platforms also offer training services to beneficiaries about the tackled societal issue. Such training can be carried out directly in physical locations, but also facilitated by digital technologies. For example, the France-based e-petition platform i-boycott.org aims to “raise citizens’ awareness about the benefits of more ethical consumption patterns by offering training to young people through the civic service or in business schools.”[2](training).

Summary

To analyse the scaling strategies of digital social enterprises, the 10 indicators we developed based on the literature survey were highlighted in bold above and are summarised in Table 2. We first address scaling out, where technology facilitates diffusion in different locations. Second, we consider scaling up to examine the ways in which civic engagement platforms support their ecosystem. Third, we examine scaling deep to highlight the sustainable change elements related to local cultures and communities.

In the next section, we apply our 10 indicators to a sample of 189 civic engagement platforms. A priori, it is difficult to understand whether civic engagement platforms use these three scaling strategies simultaneously, or whether certain strategies can be identified in relation to specific indicators. It is the objective of the cluster analysis, presented in the next section, to explore the scaling strategies of the cases in our sample to address our second research question.

Data and Methodology

Sample

Our research draws upon 189 civic tech cases in Europe. This sector includes a range of organisations aiming to build and strengthen capacity towards achieving sustainable development goals (grassroots initiatives, non-profit and for-profit organisations, public sector). In selecting these cases, we relied upon two methods used to study online platforms (Stiver et al., 2014; Wu Song, 2009). First is the review of academic and non-academic media. We followed a problem-led search, in which we first identified the range of social issues that civic engagement platforms were likely to tackle (cf. Table 1). We scanned magazines focusing on the social economy from both digital and non-digital sources, as well as websites of major sponsors of digital innovations (we focused on social innovations in these sites), as well as social innovations (we focused on digital innovations in these sites). We also screened journals and online sources focusing on the digital sector and social entrepreneurship, prize nominations granted by public and private organisations as well as by public organisations involved in the third and digital sectors. This review enabled us to identify the most significant keywords in the field, which we used in online search functions of major news outlets. Table 3 shows the distribution of cases per country, and Table 4 per domain of activity , and table 5 platforms' age.

table 2

**Scaling strategies & platform characteristics**

Methodology

Data collection

Based on our literature survey we developed a set of indicators deemed important for the scaling out, scaling up, and scaling deep of DSI. These are presented in Table 2. For each case included in our sample, we collected additional data on these indicators through their websites. Table 6 provides tdescriptive statistics for the data collected for each scaling strategy. These 10 variables are all categorical and take the value of 0 or

In developing these indicators, we were guided by the idiosyncratic characteristics of each scaling strategy revealed by our literature survey. For scaling out, these characteristics centred upon organizational growth; for scaling up on institutional and/or organizational support to broader ecosystem actors; and for scaling deep, which has stronger individual and societal orientations, it is the interactions between people in a local community that take a central stage. The corresponding indicators to each of these scaling types are summarized bellow (cf. Table 6).

Table 4

**Cases per domain of activity**