Exploring the conceptual foundations of social impact measurement in social enterprises: a systematic review of the literature

While sustainability performance measurement in research and practice has often concentrated on metrics aimed at reducing negative impacts, such as carbon or water footprints (Kühnen et al., 2022), minimizing harm alone is insufficient for achieving sustainable development (Ergene et al., 2021). Social enterprises (SEs), which pursue a social mission through economic activity (Costa and Andreaus, 2021; Grieco et al., 2015; Kah and Akenroye, 2020), have emerged as key actors in proactively advancing sustainable development (Tomei et al., 2024). Social entrepreneurship is commonly defined as “providing business solutions to social problems” (Rotheroe and Richards, 2007, p. 35), and it can manifest across a range of organizational forms and business models (Kim and Ji, 2020; Nicholls, 2009).

To assess the extent to which SEs fulfill their social missions and contribute to social development, it is crucial to establish robust social impact measurement systems that address the expectations of stakeholders and society at large (Mulloth and Rumi, 2022).

Despite the growing emphasis on social impact measurement within the field, existing approaches exhibit significant limitations. Many tools privilege short-term, quantifiable outputs, thereby overlooking the multidimensional, systemic and long-term nature of social change (Antadze and Westley, 2012; Hervieux and Voltan, 2019). Furthermore, engagement with impact measurement is not always driven by a commitment to learning or performance improvement. Organizations may adopt such practices in response to institutional pressures or to strengthen their legitimacy and competitive positioning in resource-constrained environments (Lall, 2017, 2019; McLoughlin et al., 2009; Van Rijn et al., 2024). As nonprofit organizations and SEs increasingly compete for limited funding, impact evaluation becomes a mechanism for signaling efficiency and accountability to external stakeholders (Kah and Akenroye, 2020). This creates a risk that organizations prioritize what is measurable over what is meaningful, potentially aligning their goals with external stakeholder expectations rather than their intended social mission (Argiolas et al., 2024; Battilana, 2018; Ćwiklicki et al., 2025). At the same time, SEs must balance their ambition to assess social impact with constrained internal resources (Costa and Andreaus, 2021; Molecke and Pinkse, 2017), giving rise to a fundamental dilemma: while the measurement of non-financial impact is central to their social mission, it remains resource-intensive and inherently difficult to operationalize. Accordingly, the design and use of performance measurement systems for SEs are characterized by a high level of compromise among competing priorities (Banerjee et al., 2024).

Given these challenges, there is a need to critically examine the conceptual foundations underpinning social impact measurement. While previous research has focused on mapping and evaluating available measurement tools to guide practitioners (e.g. Grieco et al., 2015; Kah and Akenroye, 2020), this study shifts the focus toward the underlying conceptual debates within the method-building literature. By mapping and synthesizing these debates, the study seeks to clarify key areas of contention and contribute to the development of more coherent and theoretically informed approaches to social impact measurement.

Specifically, the following research question (RQ) is addressed:

Compared to similar research fields such as the field of corporate social responsibility (Dahlsrud, 2008), we have to be aware that also social impact is contextual and culturally embedded. Accordingly, social impact should be understood as a socially constructed phenomenon (Berger and Luckmann, 1966) that warrants careful, context-sensitive analysis. On this basis, we develop four propositions that specify how underlying conceptualizations of impact shape social impact measurement. First, differences in measurement approaches are rooted in underlying conceptualizations of impact. Second, these conceptualizations give rise to distinct measurement logics. Third, SEs respond to competing logics by engaging in bricolage. Finally, efforts toward standardization are inherently constrained by these underlying tensions.

The structure of the paper is as follows: Section 2 outlines the systematic literature review’s methodology; Section 3 presents the findings; Section 4 introduces the four propositions; and Section 5 concludes with a summary of key insights and implications.

This systematic literature review adheres to established guidelines for management and entrepreneurship research as outlined by Hiebl (2023) and Sauer and Seuring (2023). To ensure a comprehensive and objective selection of relevant literature, a database-driven approach was implemented (Hiebl, 2023). The databases selected (Web of Science, ScienceDirect, Wiley Online Library and EconLit) were chosen because they offer comprehensive coverage of both multidisciplinary research fields and economics-specific literature, ensuring that studies from management, social sciences and core economic journals were captured. In line with established guidelines for evidence synthesis, these databases are classified as primary sources suitable for systematic reviews (Gusenbauer and Haddaway, 2020).

The process was guided by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) framework (Moher et al., 2009), which is widely recognized in both social entrepreneurship (Aliaga-Isla and Huybrechts, 2018; Musinguzi et al., 2023) and sustainability performance measurement research (Damtoft et al., 2025).

To come up with the search string and relevant time frame, prior systematic literature reviews in the domain of social impact measurement served as scoping studies to inform the search strategy (Grieco et al., 2015; Kah and Akenroye, 2020; Rawhouser et al., 2019).

While Kah and Akenroye (2020) focus on publications appearing within a ten-year window (2009–2019), Rawhouser et al. (2019) examine a longer, 20-year period from 1996 to 2016. However, they report that more than three-quarters of the contributions on social impact measurement were published in the most recent decade. Consequently, our analysis covers the period from 2004 to 2024.

Developing an appropriate search string was challenging due to the considerable conceptual ambiguity surrounding the terms “social impact,” “measurement” and “social enterprise” in the literature (Grieco et al., 2015). First, the search strategy incorporated both “social enterprise” and “social entrepreneurship” to maximize retrieval sensitivity and minimize the risk of excluding relevant studies. Although the concepts are analytically distinct, stakeholders in the field frequently use both terms interchangeably (Mohiuddin and Yasin, 2023). While social entrepreneurship can be considered as “a social value generation activity practiced by a variety of economic actors ranging from individuals, micro-enterprises to large firms” (Ramani et al., 2017, p. 218), the review focuses on social impact measurement practices that can be applied by organizations (i.e. SEs) (Beer and Micheli, 2018).

Second, regarding variants of the term “measurement,” we used a search string that combines keywords applied by former systematic literature reviews on social impact measurement (Table 1) (Alomoto et al., 2022; Grieco et al., 2015; Kah and Akenroye, 2020; Rawhouser et al., 2019).

Third, the terms social impact and social value were included, as the literature frequently uses social impact and social value interchangeably (Hietschold et al., 2023; Kah and Akenroye, 2020; Scartozzi et al., 2025).

Consequently, we used the following search string: “social entrepreneurship” OR “social entrepreneur*” OR “social enterprise*” AND “measurement” OR “performance” OR “evaluation” OR “SROI” AND “social impact” OR “social value.” For Science Direct and Wiley, search terms were restricted to the title, abstract or keywords. In addition, a backward search was conducted to capture any studies that were not retrieved in the initial database search. In line with prior review work (Alomoto et al., 2022), we focused on social return on investment (SROI) as the only specific measurement tool in the search string because SROI is the most frequently used impact-measurement keyword in the literature, whereas other tools rarely appear as standardized search terms (see Alomoto et al., 2022, p. 236).

The review adopts a deliberately narrow focus on the social impact measurement method-building literature to trace and analyze the underlying conceptual debates. Consequently, it is designed to capture the conceptual development of measurement approaches, rather than exhaustively covering every existing framework.

Accordingly, we included only peer-reviewed, English-language journal articles that explicitly introduce, apply or substantively discuss social impact measurement methods.

Several categories of publications were excluded at the screening and eligibility stage (Figure 1). First, we omitted articles in which social impact measurement was only mentioned in passing or treated as a secondary issue (e.g. within broader examinations of organizational performance). Second, we excluded studies whose primary focus lay on performance dimensions other than social performance, or on antecedents and outcomes of social impact (such as critical success factors or the scaling of social impact), as these do not engage directly with the construction of measurement methods. Third, systematic literature reviews and related overview articles were not retained, since the purpose of this study is to analyze conceptual debates within primary method contributions rather than to summarize existing evidence. Finally, we excluded publications that dealt exclusively with non-profit or public organizations without an explicit entrepreneurial or hybrid orientation. The review centers on SEs understood as hybrid organizations that combine a social mission with financial logics (Zeng and Van Staden, 2024), and the selected articles were required to reflect this context.

Applying these criteria in the initial screening of titles, abstracts and keywords for the 593 identified records reduced the pool to 68 publications. In a final step, we screened the full text of the remaining 68 articles against the exclusion criteria outlined above and removed non-peer reviewed items, resulting in a final sample of 27 publications ( Appendix 1). This sample size is consistent with other recent systematic literature reviews in the field of social impact measurement in SEs (Beer and Micheli, 2018; Kah and Akenroye, 2020; Musinguzi et al., 2023).

We began our analysis by carefully reading all articles in the final sample to develop an initial understanding of the substantive foci and empirical contexts of social impact measurement studies. In this first iteration, we constructed a structured codebook in Excel and systematically extracted information on each article, including author(s), publication year, journal outlet, methodological approach, empirical context (such as sample size and business models of the SEs) and main findings. This step enabled us to provide a transparent overview of the social impact measurement research domain and to delineate the variety of contexts and perspectives from which social impact measurement has been examined.

In a second iteration, we conducted an inductive, theme-focused analysis of the full set of articles. For social impact definitions, we compiled all definitions reported in the studies and highlighted text segments that followed similar definitional logics, thereby identifying recurring conceptual patterns. Synthesizing these understandings in the field, we clustered the patterns into four conceptual dimensions that we consider the most meaningful way of organizing the literature, as they offer a new but still compatible structuring with established social entrepreneurship research (e.g. discussions on locus in Lumpkin et al., 2018 and depth in Hietschold et al., 2023). The results are illustrated in  Appendix 2^[1]. Since nine out of 27 publications did not contain a definition of social impact (André et al., 2018; Lee et al., 2021; Millar and Hall, 2013; Moody et al., 2015; Mulloth and Rumi, 2022; Nicholls, 2009; Rotheroe and Richards, 2007; Studer, 2022; Taušl Procházková et al., 2021), the conceptualization of impact focused on the 18 remaining publications.

For social impact measurement methods, we first tabulated the different methods and grouped studies that used or advanced the same approach (Table 2). We then synthesized the main findings for each method in separate analytic memos and coded these summaries for recurrent themes ( Appendix 3). On this basis, we derived social impact measurement logics, as illustrated in Table 3. Combining the findings of Table 3 with the findings regarding the conceptualization of social impact, we developed the conceptual framework as illustrated in Figure 3.

Social impact measurement in SEs has gained attention over the past years: 56% of the publications have been published within the past five years and 78% within the past ten years. This is consistent with previous scholars’ findings (Kah and Akenroye, 2020; Rawhouser et al., 2019). Most of the articles (59.3%) on social impact measurement follow a qualitative approach. Only a few (7.4%) pursue a quantitative approach. The most common data collection method is a case study based on one company (25.9%), as illustrated in Figure 2.

In total, 67% of the studies are published in an American or British journal. Only one journal (Business Strategy and Development) originates from a developing country. This shows that there is a dearth of quantitative studies and studies that investigate social impact measurement in an emerging and developing economy context.

The coding results ( Appendix 2) indicate that social impact is not defined uniformly across the literature but varies systematically. Our analysis indicated a three-layer conceptual model that progressively addresses the RQ. The first layer consists of the four core dimensions: normativity, causality, locus and temporality. These dimensions reflect the underlying assumptions about what social impact is. The second layer refers then to the measurement logics (managerial, learning and accountability), which are influenced by the conceptual foundations of the first layer. Finally, the third layer (bricolage) provides explanations how measurement approaches manifest in SEs in various contexts and under competing logics informed by the underlying conceptual foundations.

Across the reviewed studies, “impact” is predominantly endowed with a positive connotation, referred to as benefit, beneficiaries or improvement ( Appendix 2). Only a small number of studies adopt a more neutral vocabulary, referring instead, to effects or change ( Appendix 2). This positivity bias suggests that social impact is often treated as a normative category rather than as an analytically open one, leaving little room for ambivalent or negative outcomes (Musinguzi et al., 2023; Scartozzi et al., 2025; Vázquez Maguirre et al., 2018). A more critical perspective, however, would require measurement frameworks that also capture unintended effects and tradeoffs across social, environmental and economic dimensions (Gigliotti and Runfola, 2022; Green et al., 2024) and among different stakeholders (Costa and Pesci, 2016; Golbspan Lutz et al., 2025; Stroehle et al., 2026; Vázquez Maguirre et al., 2018).

A dominant strand in the literature conceptualizes social impact as an effect attributable to the activities of a social enterprise ( Appendix 2). Conceptually, this view rests on a causal–attribution logic, defining impact as effects that occur “above and beyond what would have happened anyway” (Grieco et al., 2015, p. 1175). Social impact is thus framed as a discrete, isolable consequence of organizational action. A growing body of work problematizes this linear causality and instead foregrounds the relational and systemic nature of impact, positing that outcomes are co-produced within networks of actors and institutional structures (Barraket and Yousefpour, 2013; Ćwiklicki et al., 2025; Gigliotti and Runfola, 2022; Hervieux and Voltan, 2019; Stroehle et al., 2026; Vázquez Maguirre et al., 2018). From this perspective, impact is not a stable “effect” that can be neatly allocated to a single organization but an emergent property of complex social systems, exemplifying the need to understand impact through an “ecosystem strategy” (Stroehle et al., 2026, p. 6). At a more granular level, individual organizations struggle to establish clear relationships between complex input factors and the social impacts to which they contribute (Nicholls, 2009). Taken together, this points to a broader theoretical tension between a positivist understanding of impact as a measurable, organization-specific outcome and a relational understanding of impact as contingent, distributed and difficult to disentangle from its context (Antadze and Westley, 2012; Hervieux and Voltan, 2019; Nicholls, 2009).

A third conceptual dimension concerns the locus of impact, that is, the level at which impact is understood to materialize: individual beneficiaries, communities, society at large or the natural environment ( Appendix 2). Rather than merely representing different units of analysis, these levels encode distinct assumptions about the nature of social change, whether it is primarily actor-centered and micro-level or multi-scalar and embedded in interdependent systems. Lumpkin et al. (2018) argue that the community and societal levels are the most appropriate loci of impact, as SEs typically pursue missions that reach beyond the boundaries of the organization and its members. Yet, this view sits uneasily with the dominant traditions in entrepreneurship and management research, which have historically concentrated on phenomena and outcomes at the individual, group and organizational levels (Ebrahim and Rangan, 2014; Lumpkin et al., 2018; Walsh, 2003). Ciccarino et al. (2024) find that value created for individuals and organizations frequently spills over into broader societal benefits, blurring the boundaries between micro-, meso- and macro-level outcomes and making it difficult to separate actor-specific value from wider social welfare. In this regard, Bassi and Vincenti (2015, p. 22) conceptualize the distinctiveness of nonprofit organizations and SEs as residing not only in what they produce but also in how, with whom and for whom value is created. They argue that such organizations generate different forms of value across levels: fostering responsibility among individuals at the micro level, producing relational goods for organizations and local communities at the meso level and generating social capital for broader social systems at the macro level. Taken together, these perspectives suggest that social impact should be understood as a multidimensional phenomenon that emerges across interconnected levels, rather than as an outcome located exclusively at any single locus.

Temporality is the least developed dimension in the reviewed literature ( Appendix 2). While some studies equate impact with long-term transformations, others treat both short-term and long-term changes as relevant ( Appendix 2), sometimes granting short-term effects the status of impact only when they cross a threshold of significance (Ebrahim and Rangan, 2014; Molecke and Pinkse, 2017). These temporal distinctions are not just operational choices; they reflect deeper assumptions about whether impact is conceived as an event, a process or an enduring shift in underlying structures.

Some authors respond to these tensions by introducing terminological distinctions that seek to map conceptual differences onto separate labels. For instance, Ebrahim and Rangan (2014) distinguish between outcomes as medium- and long-term micro-level changes and impact as long-term meso- and macro-level changes, thereby reserving “impact” for more deeply institutionalized forms of change. By contrast, Hietschold et al. (2023) define social value as short-term micro-level changes and social change as long-term macro-level transformations. Rather than merely proposing alternative measurement schemes, these distinctions can be read as competing conceptual models of how social value manifests over time and across levels of analysis. Taken together, these findings suggest that social impact is not a self-evident empirical category, but a contested theoretical construct shaped by assumptions about value, causality, scale and temporality.

The review shows that social impact measurement is characterized by considerable methodological diversity and the absence of generally accepted standards (Banerjee et al., 2024; Savall Morera et al., 2022; Tomei et al., 2024). As a result, the assessment of impact often remains largely contingent on the preferences and objectives of investors and investees (Lee et al., 2021).

Table 2 illustrates the impact measurement methods mentioned by the 27 publications included in the systematic literature review.

Among the reviewed studies, SROI is the most frequently used method. Rooted in cost-benefit analysis, SROI expresses social impact as the ratio of value created to resources invested and thus operationalizes impact in monetized terms (Lingane and Olsen, 2004). Its emphasis on stakeholder engagement, materiality, causal mapping and deadweight adjustment makes it one of the most structured approaches to impact measurement (Rotheroe and Richards, 2007).

The logic model, or impact value chain, is another widely used framework. Originating in program evaluation, it conceptualizes impact as a sequence from inputs and activities to outputs, outcomes and final impact (Ebrahim and Rangan, 2014). It is especially useful for aligning mission and performance measurement, and later versions extend the model by incorporating temporal and stakeholder dimensions (Costa and Andreaus, 2021).

The literature also includes a range of complementary methods. Adaptations of Kirkpatrick’s four levels of evaluation provide a resource-efficient way to assess participant outcomes and program effectiveness (Lee et al., 2021). Triple bottom line indicators offer structured measures of social, economic and local performance (Taušl Procházková et al., 2021). More recent data-driven approaches include Big Data Analytics-based models (Tamym et al., 2023), which estimate impact through environmental proxies and data envelopment analysis (Dia and Bozec, 2019), which compares input–output relationships across organizations to assess relative efficiency.

Accounting-oriented frameworks such as blended value accounting (Nicholls, 2009) and compromising accounts (André et al., 2018) combine financial and non-financial reporting, while qualitative approaches emphasize stakeholder dialogue and contextual adaptation. Process-oriented frameworks, including the SIMPLE model (McLoughlin et al., 2009) and the organizing framework for understanding value creation (Ormiston, 2023), structure impact management as a sequence from problem definition to decision-making. Systems-oriented approaches and methods such as outcome harvesting (Tomei et al., 2024), systems map (Khare and Joshi, 2018) and system approach (Hervieux and Voltan, 2019) address complexity by focusing on interdependencies and non-linear change.

Overall, the literature shows that social impact is increasingly understood as dynamic, relational and context-dependent. At the same time, recent work suggests that organizations often respond to measurement pressures through bricolage, combining available tools, practices and narratives rather than relying on standardized methodologies alone (Helleputte and Périlleux, 2025; Molecke and Pinkse, 2017).

The coding results ( Appendix 3) reveal that the measurement methods cluster around three distinct measurement logics (managerial, accountability, learning/systemic), defined by recurring patterns in comparability, adaptability, stakeholder engagement, causality assumptions and purpose, as illustrated in Table 3. These logics correspond closely to prior conceptual distinctions in the literature between accountability and organizational learning (Schillemans and Smulders, 2015), managerial and deliberative conventions (Studer, 2022) and primary functions of social impact measurement, namely, management, communication and strategy (Bellazzecca et al., 2025). More specifically, managerial logics align with the management function and emphasize control, quantification and internal decision-making. Accountability logics correspond to the communication function and foreground transparency, legitimacy and communication with external audiences. Learning and systemic logics, by contrast, are more closely associated with the strategy function and are better suited to complexity, reflection and organizational adaptation (Bellazzecca et al., 2025).

Accountability and learning logics reflect a more participatory and deliberative orientation, whereas the managerial logic remains tied to attribution, quantification and standardization (Studer, 2022).

Across these logics, the coding reveals four recurring tensions.

Most methods focus on micro- and meso-level outcomes, while macro-level impact remains comparatively underrepresented ( Appendix 3). Stakeholder engagement is central in accountability and learning approaches but limited in managerial approaches. Comparability and adaptability also appear in tension: standardized methods improve benchmarking, while tailored methods increase contextual relevance. Finally, approaches differ in how they handle causality, with managerial methods simplifying causal relations and learning-oriented methods emphasizing process and complexity.

The purpose of measurement also varies. In most cases, methods are used for financing, standardization or management; only a minority are explicitly oriented toward societal change ( Appendix 3). This suggests that measurement remains primarily internal or organizational in orientation, despite the broader social ambitions of the field (Hervieux and Voltan, 2019).

No single method addresses all of these dimensions at once. Instead, organizations combine different approaches to manage tradeoffs and respond to external demands. This pattern indicates that bricolage (Lévi-Strauss, 1962; Mair and Marti, 2009) is not marginal but a common mode of social impact measurement practice.

Overall, the findings show that social impact measurement is shaped by a set of interconnected conceptual and methodological tensions. Social impact is framed in different ways, operationalized through diverse methods and organized around distinct measurement logics that privilege different purposes, stakeholders and assumptions about causality, scale and temporality.

Rather than converging on a single standard, the literature reveals a field marked by plurality, selective adaptation and ongoing compromise. These patterns suggest that social impact measurement cannot be understood solely as a technical exercise. It is also a conceptual and organizational practice shaped by competing logics and by the need to reconcile legitimacy, learning and comparability (Barraket and Yousefpour, 2013). The discussion section therefore turns to what these findings imply for how social impact should be understood, how measurement logics can be theorized and why bricolage plays a central role in mediating between them.

The identified tensions in social impact measurement are not merely methodological but reflect more fundamental disagreements about the nature of social impact itself. The literature reveals four core conceptual divides: whether impact is inherently positive or includes negative and unintended effects; whether it can be causally attributed to individual organizations or emerges from complex systems; whether it materializes at the level of individuals, communities or broader societal structures; and whether it should be understood as short-term outcomes or long-term transformations. These conceptualizations shape and partly constrain, the design of measurement approaches. Building on this insight, this study proposes a framework in which measurement logics (managerial, learning and accountability) are understood as manifestations of underlying conceptual assumptions, while bricolage explains how SEs navigate and reconcile these tensions (Figure 3).

Within this framework, bricolage describes the mechanism through which SEs work across competing assumptions and associated measurement logics, to construct workable, hybrid approaches in practice.

While early work has largely understood bricolage as “making do with what is at hand” (Baker and Nelson, 2005; Lévi-Strauss, 1962), Molecke and Pinkse (2017) show that, in the domain of impact measurement, SEs engage in both material and ideational bricolage: they draw on data, narratives and indicators that are actually available and simultaneously develop alternative methodological and conceptual frames that can accommodate these forms of information in their impact accounts. In this process, formal, often accounting- and finance-inspired methodologies are first delegitimized, that is, critically questioned in terms of their underlying assumptions and practical fit. This delegitimization opens up interpretive flexibility: the individual elements of formal approaches, metrics, causal attribution logics, evidence requirements, are loosened from the rigid structure of an integrated methodology and turned into “building blocks” that can be recombined into new, context-appropriate understandings and practices of impact evaluation (Molecke and Pinkse, 2017).

Helleputte and Périlleux (2025) show, in their in-depth case study of a formal, randomized controlled trial-inspired social impact assessment, that bricolage remains central even when SEs explicitly commit to formal methods. In their case, considerations of feasibility, efficiency, ethics, legitimacy and purpose collide with the demands of formal procedures. Bricolage operates here as a navigation mechanism: through multidimensional, participative, culturally embedded, mixed-methods and adaptive principles, elements of formal logics (e.g. standardization, comparability, evidence orientation) are combined with alternative, field-embedded logics (e.g. narrative and experiential forms of evidence). The result is “formally driven” SIA bricolage, in which SEs do not simply adopt formal requirements but selectively adapt, supplement and blend them with their own conceptions of impact.

The analysis suggests that measurement practices are guided by implicit assumptions about impact. The prevalence of positive framings, attribution-based reasoning and short-term operationalization indicates that measurement is often shaped by the need to demonstrate results, remain feasible and maintain legitimacy (Barraket and Yousefpour, 2013; Hervieux and Voltan, 2019; Studer, 2022).

This means that measurement does not simply capture impact; it also shapes how impact is defined and communicated within SEs and to their stakeholders:

The findings indicate that the three measurement logics represent structured responses to these conceptual assumptions. SEs do not select measurement tools randomly; rather, their approaches reflect different organizational priorities and stakeholder expectations (Costa and Andreaus, 2021; Costa and Pesci, 2016).

Managerial approaches are oriented toward control, quantification and internal decision-making. Accountability approaches focus on transparency, legitimacy and communication with external audiences. Learning and systemic approaches, by contrast, are better suited to complexity, reflection and organizational adaptation (Bellazzecca et al., 2025):

SEs rarely rely on a single measurement logic. Instead, they combine elements from different approaches to balance competing demands, such as comparability and contextual sensitivity, or accountability and learning (Helleputte and Périlleux, 2025; Molecke and Pinkse, 2017). This can be understood as bricolage: organizations selectively assemble and adapt tools, indicators and frameworks to fit their specific circumstances (Lévi-Strauss, 1962). In this sense, bricolage is not simply a pragmatic compromise but a central mechanism that enables SEs to make measurement workable in practice:

The findings also highlight the limits of standardization. While standardized approaches promise comparability and clarity (Kim and Ji, 2020; Mulloth and Rumi, 2022; Savall Morera et al., 2022), they cannot fully resolve the tensions that arise from different understandings of social impact: “numbers are never neutral concepts, but […] they inadvertently carry ideas of morality and normativity, which we must critically consider” (Stroehle et al., 2026, p. 6).

For SEs, this means that standardization should be seen as one possible way of organizing measurement, rather than as a universal solution. Its use is constrained by conceptual tensions and by the need to adapt measurement to local and organizational contexts (Costa and Andreaus, 2021; Lumpkin et al., 2018; Millar and Hall, 2013; Moody et al., 2015):

Overall, the findings support a processual understanding of social impact measurement. Rather than being a fixed system, measurement is an ongoing practice in which understandings of impact and measurement routines evolve together over time (Moellmann et al., 2025; Moody et al., 2015; Nicholls, 2009; Tomei et al., 2024). For instance, impact measurement may change in purpose and perception over time, shifting from a predominantly legitimacy-oriented function toward a more learning-oriented one (Lall, 2019).

This suggests that the field is unlikely to converge on a single standard model. Instead, social impact measurement will probably remain characterized by plurality, adaptation and negotiation across different organizational settings. For social entrepreneurship research and practice, this should be seen not as a weakness, but as a reflection of the diversity and complexity of the field: SEs are established to fill institutional voids that are created because conventional reporting practices fail to capture full value creation opportunities (Molecke and Pinkse, 2017; Nicholls, 2009). In this sense, the usefulness of impact measurement depends less on adherence to a single method than on the robustness of the underlying practices. Akwetey et al. (2025, p. 282) highlight five key stages ^[2], arguing that organizations that undertake these stages “will find impact measurement worth the time and effort,” regardless of whether they rely on bricolage or a more standardized method. These practices can help SEs scale and combine different measurement approaches to meet their needs. As Hervieux and Voltan (2019, p. 273) note: “Measurement is about getting from point A to point B and our work is about traveling all over the map.”

This study opens several avenues for future research on social impact measurement in SEs. First, the conceptual framework linking normativity, causality, locus and temporality to managerial, accountability and learning logics invites quantitative and qualitative empirical analysis in diverse organizational and national contexts. How do SEs in different institutional settings, such as Europe, North America or emerging economies, navigate these conceptual tensions and to what extent does bricolage vary across contexts? Second, while this study identifies bricolage as an empirical pattern, the conditions under which bricolage becomes productive versus counterproductive remain unclear. Under what circumstances does combining multiple logics enhance measurement quality and organizational legitimacy, and when does it lead to fragmentation or learnability loss? Third, the processual nature of measurement suggested by the findings calls for longitudinal research that tracks how conceptualizations of impact and measurement practices co-evolve over time. How do SEs adapt their measurement systems as they scale, shift missions or respond to changing stakeholder expectations? Finally, the limits of standardization highlighted in this study point to a need for research on hybrid governance models that balance comparability with contextual sensitivity. What design principles enable measurement systems to serve multiple purposes, internal learning, external accountability and financing, without sacrificing conceptual clarity or adaptability? One promising avenue for future research on this question is the use of artificial intelligence-enabled measurement systems that can dynamically integrate heterogeneous data sources, tailor indicators to different stakeholder audiences and learn from past assessments over time. Building on recent work that explores artificial intelligence as a way to overcome model-related barriers (e.g. Abad-Itoiz et al., 2025), such systems could help reconcile the multiple purposes of impact measurement.

This study addresses a critical gap in social entrepreneurship research by moving beyond the practical classification of impact measurement methods to interrogate the conceptual debates that underpin the field. Our study yields three key theoretical contributions. First, it reconceptualizes social impact measurement as being rooted in deeper assumptions about normativity, causality, locus and temporality rather than as a purely technical exercise. Second, it shows that managerial, accountability and learning logics are not neutral measurement options, but different expressions of these underlying conceptualizations of impact. Third, it extends theory on hybridity and organizational adaptation by showing how bricolage enables SEs to navigate and reconcile competing logics in practice.

More broadly, the findings suggest that tensions in social impact measurement are not simply problems of implementation or standardization. Instead, they reflect fundamental disagreements about what social impact is, how it should be assessed and at what level and over what time horizon it should be understood. This shifts the analytical focus from choosing the “best” measurement tool to examining the conceptual assumptions that shape measurement choices in the first place. Taken together, the framework contributes to the literature by linking conceptualizations of impact, measurement logics and bricolage into a single explanatory model.

Our practical implications reveal that for SEs, choosing a measurement approach is not only a technical decision, but also a strategic one. Organizations need to be explicit about whether measurement is intended primarily for internal learning, external accountability or managerial control, because these purposes imply different design choices (Barraket and Yousefpour, 2013; Lall, 2019). Hybrid approaches may be more realistic than attempts to enforce a single standard. Rather than treating plurality as a problem to be eliminated, SEs may benefit from designing measurement systems that combine comparability with flexibility and that allow different audiences to be served in different ways (Hervieux and Voltan, 2019; Lingane and Olsen, 2004).

The results also carry important implications for public policy. They challenge the assumption that standardization of impact measurement is inherently desirable. While policymakers often promote unified frameworks to enhance comparability and accountability, enforcing a single measurement approach may overlook the diverse purposes and operational realities of SEs. Instead, public policy could play a more supportive role by enabling pluralistic measurement ecosystems. This might involve endorsing flexible guidelines rather than rigid standards, funding capacity-building initiatives and recognizing multiple forms of evidence depending on whether the primary aim is learning, accountability or control. Such an approach would better align with the hybrid nature of SEs and reduce the risk of mission drift caused by overly prescriptive measurement regimes.

We recognize that our review may not capture all potentially relevant literature and social impact measurement methods, due to the applied search string and the inconsistent terminology surrounding social impact. Additionally, this systematic literature review is limited by its focus on a selected number of databases and its inclusion of only English-language sources; to identify cultural differences, future reviews should incorporate non-English literature. Several avenues for further research have been identified, as discussed above. From a methodological perspective, given the contextual embeddedness and social constructionist nature of the phenomena under investigation, qualitative research approaches are particularly promising, as they facilitate the development of in-depth and nuanced understanding.