A Willingness–Propensity–Ability Framework for Innovation Capability in Agri-Food SMEs: Evidence from the Sardinian Sheep Dairy Sector

Abstract

Innovation is a central driver of competitiveness, resilience, and sustainability in the agri-food sector, particularly among small and medium-sized enterprises (SMEs). However, traditional science- and technology-based models may not fully grasp the innovation dynamics in this domain, and research explicitly addressing agri-food SMEs remains limited. This study adapts, integrates, and extends existing Innovation Capability (IC) and related constructs into a unified WI–PI–IA framework (Willingness to innovate–Propensity to innovate–Innovation Ability) for agri-food SMEs. The framework is empirically tested through a sectoral quantitative case-study based on structured questionnaires administered to twenty SMEs operating in the Sardinian sheep dairy industry. The findings confirm the framework’s validity, highlighting the role of contextual factors and revealing distinct innovation patterns between cooperatives and private firms. This study is, to our knowledge, the first to conceptualise IC in agri-food SMEs as the outcome of the three above constructs and offers a comprehensive and context-sensitive approach that contributes to academic research and directs policymakers towards factors that affect agri-food SME innovation outcomes, considering their unique structures and specific challenges they face.

1. Introduction

Agri-food systems play a dual role, both as a source of environmental pressure and a driver of sustainable development [1,2]. Rapid transformations in the sector highlight the need for innovation to enhance resilience, provide access to nutritious and sustainable food, and advance the transition toward a circular food system [3,4], other than enable businesses to enhance efficiency, productivity, competitiveness and economic growth [5,6,7,8,9,10]. However, while important in developed economies, the agri-food industry, with its specific characteristics, is often viewed as a low-technology and research-light sector [11]. As a result, traditional models based on science and technology may not fully explain innovation in this sector [10,12,13].

The role of innovation in enhancing competitiveness is increasingly acknowledged as crucial for small and medium-sized enterprises (SMEs) within dynamic and fast-evolving markets, e.g., [14,15,16,17]. SMEs drive the transition to sustainable food systems [18,19,20], are crucial to rural societies and play a key role in both local and national economies [21,22,23,24,25], also generating invisible and hidden innovation that often escapes statistical detection [26]. In this context, the significance of innovation in SMEs within the agri-food sector becomes evident.

The academic debate regarding this phenomenon has focused mainly on innovation capability (IC) [27,28]. Moreover, research on IC also in the context of SMEs has branched into two research paths: one focused on understanding the factors that determine IC and the other on analysing its impact and consequences [29].

Although extensive research on innovation and innovation management has been conducted, only a limited number of studies have explicitly focused on SMEs [14]. Even fewer have concentrated on the agri-food sector [30]. While it is recognised that innovation can help agri-food SMEs address various challenges and enhance sustainability and profitability [31], several gaps still need to be addressed.

There is a need, inter alia, for further studies that explore IC as a multidimensional construct and translate theoretical models into practical applications (also moving away from the few empirical studies that develop a theoretical understanding of IC but still operationalise it as a consequence) [32] and focused specifically on agri-food SMEs’ IC [33]. Additionally, it behoves case-based research designs to provide missing input on how IC is actually understood and implemented among SMEs, as well as to shed light on the role of different contextual factors in the development of small business IC [29].

Research conducted at the regional level can offer valuable insights into the functioning of regional innovation systems and their influence on firm-level IC [33]. Furthermore, studies should include external aspects, such as barriers to innovation, and focus not only on product innovation but also on other types of capabilities that SME should develop, e.g., [29,33,34].

In light of these considerations, the main objective of this study is to investigate the antecedents of IC in agri-food SMEs, adopting a multidimensional perspective. Specifically, the study first aims to adapt, integrate, and broaden existing IC and related construct frameworks into a unified conceptual model explicitly tailored to agri-food SMEs, encompassing three key components: willingness to innovate (WI), innovation ability (IA) and propensity to innovate (PI) [34,35,36].

Building on this framework, the study empirically explores the determinants and formation mechanisms of IC at the regional level, focusing on the Sardinian sheep dairy sector, a context predominantly composed of SMEs, characterised by strong territorial embeddedness, and strategic relevance for the regional economy. Additionally, the connection between innovation and agri-food SMEs is recognised as essential for the sustainable future of Mediterranean food systems [20]. In particular, the analysis examines firms’ willingness to engage in innovation-oriented activities, their propensity to innovate in relation to strategic, structural, and behavioural dimensions that are path-dependent and context-specific [37,38], and the resources influencing their innovation ability.

Unlike earlier multidimensional views of IC that treat organisational dimensions as parallel and largely technology-driven, this study advances a context-sensitive framework that highlights the sequential and interactive nature of capability development in traditional agri-food SMEs. By integrating WI, PI, and IA, the model departs from prior approaches by distinguishing between dispositional orientation, strategic enactment, and resource-based implementation capacity. This structure interprets IC not merely as an aggregation of internal competences, but as a governance- and network-dependent process shaped by path-dependent strategic behaviour and territorially embedded resources. In doing so, the paper refines the concept for low-tech, resource-constrained environments, where innovation relies more on institutional linkages, cooperation, and incremental adaptation than on formal R&D. In this perspective, the WI–PI–IA framework provides a context-sensitive interpretation of innovation capability formation in agri-food SMEs, linking innovation orientation, strategic behaviour, and resource mobilisation within territorially embedded innovation systems, offering practical insights for researchers, practitioners, and policymakers.

The paper is structured as follows: Section 2 reviews the literature on the role of innovation in agri-food SMEs, the IC with its constructs and related antecedents, and outlines the research framework and hypotheses, Section 3 describes the methodology, Section 4 presents the results, and Section 5 discusses the results, concludes the paper, and highlights theoretical and practical implications, limitations, and future studies beyond the recommendations for practitioners, academics, and policymakers.

2. The Literature Review and Hypothesis Development

2.1. Innovation in Low-Tech Agri-Food SMEs

Innovation plays a pivotal role in national development, contributing to consumer welfare, business performance, and overall economic growth. It enables firms to improve their performance and pioneer new products, services, and processes, thereby generating value and securing competitive advantage [32].

The literature defines distinct innovation shapes classified based on various dimensions, each playing a crucial role in the evolution of businesses, organisations, and societies. These include, inter alia, innovation of product or process innovation [39,40], organisational [41] and marketing.

The agri-food sector has long been regarded as a paradigmatic low-technology industry, exhibiting comparatively slower innovation dynamics than other industries [3,11,42,43]. Innovation is mainly expressed through product-related developments and process-oriented improvements [44] and tends to emerge through incremental adaptation, experiential learning, and supply-chain-based interactions [45,46]. These characteristics are consistent with the broader literature on innovation in low-tech industries, which emphasises incremental change, learning-by-doing processes, and the importance of supplier and user interactions rather than formal R&D activities [45,46]. Consequently, innovation often materialises in the form of process optimisation, quality- and origin-based product differentiation, organisational upgrading, and new governance mechanisms, rather than radical technological breakthroughs [11,47].

Moreover, the agri-food sector is predominantly composed of SMEs—accounting for over 99% of agri-food businesses in the European Union [48]—which often face structural constraints and complexity that extend beyond technology adoption, and hinder the identification and implementation of effective innovation trajectories [49].

Empirical research shows that the fragmented nature of agri-food production systems, characterised by a high prevalence of micro-enterprises and weak horizontal and vertical associativity, limits collaborative learning and knowledge exchange, thereby constraining innovation pathways [50,51]. Limited financial and technical resources further constrain innovation investments and organisational upgrading, as firms often lack the capacity to engage in dedicated R&D or to absorb new technologies [52]. Moreover, structural path dependency arising from established infrastructures and production modalities reinforces existing routines and inhibits transformative changes [53]. Finally, the embeddedness of firms in diverse institutional environments plays a critical role in enabling or hindering innovation processes in agri-food SMEs [50,52,54].

As a result, innovation capability in agri-food SMEs is often rooted in relational resources, cooperative arrangements, and territorially embedded networks rather than in internal technological infrastructures alone [55,56]. Moreover, innovation in this sector frequently takes incremental and “hidden” forms, making it difficult to capture through conventional science- and technology-based classifications [5,45,46,57]. Innovations span the entire firm’s value system, from strategic choices to resource development, and involve various activities that originate from its IC [14], making the identification and conceptualisation of IC components at the firm level a central and increasingly prominent issue in innovation management research [29,34].

However, despite the emergence of various theories over time, numerous gaps remain to be addressed. In particular, the concept remains loosely defined, with limited attention given to SMEs [14] and even less focus on the agri-food sector [30].

2.2. Innovation Capability (IC) in Low-Tech Agri-Food SMEs

IC refers to the potential to create new and valuable knowledge or products [58], which is a crucial element for SMEs’ competition with larger, more resource-rich competitors [29], with direct effects on their performance. This link is particularly relevant for agri-food SMEs, where IC contributes to competitiveness but also to sustainable value creation along traditional supply chains. Empirical studies show that eco-innovation and dynamic capabilities in agri-food firms are positively associated with firm performance and sustainability outcomes, reflecting the strategic importance of innovation for environmental and economic value generation [44,52,59]. Moreover, practices such as digital transformation further enhance this linkage by enabling value creation and performance improvement of agri-food SMES across food supply networks [60].

The literature on IC primarily revolves around five key perspectives: the knowledge-based view (KBV), resource-based view (RBV), dynamic capabilities (DCs), strategic management (SM), and the practice-based view (PBV). As noted by Daronco et al. [34], the literature employs a wide range of related or synonymous terms, necessitating a precise conceptual clarification of IC. This ambiguity is even more pronounced in low-tech and food-related sectors, where innovation often follows non-linear and practice-based trajectories that are insufficiently captured by conventional IC frameworks [61].

Among these, the RBV and DC frameworks emerge as the two dominant approaches, both of which are intimately linked to the strategic role of IC in enhancing firm performance [14]. In agri-food SMEs, strategic innovation resources are frequently relational and territorially embedded, as innovation is often enabled through cooperative arrangements, intermediaries, and institutional innovation ecosystems, particularly in fragmented and network-dependent agricultural contexts [62,63].

The RBV posits IC as a strategic organisational asset, essential for fostering continuous innovation and sustaining competitive advantage. According to Amit and Schoemaker [64], capability refers to “a firm’s capacity to deploy resources, usually in combination, using organisational processes, to effect a desired end”. In this context, Szeto [65] combines the RBV with innovation, defining IC as the firm’s ability to utilise its internal resources in ways that consistently lead to the development of new and innovative solutions to address changing environments and market demands and, in turn, to maintain a competitive advantage. In the food industry, the deployment of such resources is often constrained by structural barriers and fragmented value chains, which shape incremental rather than radical innovation strategies [66,67].

Complementing the RBV, the DC perspective elaborates on how firms renew and adapt their resource base over time to maintain competitiveness and avoid obsolescence. According to Teece [68], DC are essential for enhancing a firm’s IC, which refers to a firm’s ability to develop and implement new ideas, products, and processes that drive its competitive advantage. This enhancement is achievable through three key components: the ability to sense and shape opportunities, to seize them through decisive action, and to maintain competitiveness through continually adjusting, combining and reconfiguring their resources and capabilities in alignment with evolving market demands. Recent studies highlight that these DC are increasingly linked to sustainability transitions and digital transformation processes in agri-food systems, where SMEs must adapt within emerging innovation ecosystems [62].

The KBV further enriches the understanding of IC by emphasising knowledge as a critical organisational resource. Innovation, from this viewpoint, is a function of the firm’s capacity to generate, disseminate, and apply knowledge [69]. Lawson and Samson [70] argue that both internal and external knowledge assets are pivotal to innovation performance. A central concept here is absorptive capacity, defined by Cohen and Levinthal [71] as the firm’s ability to recognise the value of external knowledge and integrate it into its existing knowledge base. Importantly, the relevance of this perspective becomes even more pronounced in sectors where innovation relies less on formal R&D and more on experiential and practice-based learning like the agri-food contexts, where knowledge is often tacit, locally embedded, and developed through incremental doing–using–interacting processes [45,61].

The PBV emphasises the role of organisational practices in shaping innovation processes [72], while the strategic management perspective highlights the alignment between strategy, organisational structure and IC [73,74]. Accordingly, IC in agri-food SMEs must be understood within broader institutional and network-based environments, where policy instruments, innovation brokers, and ecosystem actors shape firms’ strategic innovation opportunities [62,63].

IC conceptualisation can also be classified based on whether innovation is viewed as a process or an outcome. With a focus on the SME context, in the innovation conceptualisation as a process, IC is seen both as a one-dimensional [70,75,76,77,78] and multiple-dimensional [72,79] phenomenon encompassing actions enforced to generate innovative outputs and improve the performance of SMEs. The second line of research views innovation as an outcome and defines IC as the ability to produce various types of innovation, e.g., [29,33]. As a result, having a strong IC leads to the creation of innovative outcomes that contribute to an organisation’s success, growth, and competitiveness.

2.3. Propensity to Innovate (PI) as Strategic Enactment

In a comprehensive literature review, Daronco et al. [34] conceptualise the IC through the DC approach, viewing it as both the propensity to act and the ability to act on that propensity, thereby positioning IC as a function of intentional and effective actions related to innovation.

The literature provides several definitions of PI [80] and synonymous concepts like innovativeness and organisational innovation [81]. PI is essential for fostering innovation within an organisation as “if organisations don’t display a propensity for innovation (…), innovation cannot and will not occur” [82]. It is closely related to the firm’s DNA and reflects a firm’s openness to depart from established practices and its readiness to pursue innovation, shaped by factors such as corporate culture, organisational strategy, and leadership [83]. Moreover, it is understood as the assessment of innovation’s role in the firm’s performance improvement, and as an outcome of recognising the need for internal change aligned with organisational goals and resources [84]. In other words, it involves perceiving innovation opportunities, being willing to act on them, and having the capability to implement change [85]. However, exploring PI is complex, since it depends on the specific path and context that shape the organisation’s strategy, structure, and behaviour [37,38]. In particular, recent research highlights that sunk costs and existing infrastructures create structural rigidities, reinforcing path dependency and often hindering technological and organizational change [53]. Thus, PI in agri-food SMEs is not purely forward-looking, as innovation behavior is shaped by past investments, infrastructures, and institutional histories. Evolutionary economic geography shows that industrial development follows branching trajectories, where past competences and structural conditions strongly influence future innovation, limiting strategic flexibility [86]. Consequently, firms’ PI reflects both managerial intent and the cumulative constraints and opportunities of their regional and sectoral histories [87].

Building on the conceptualisations of PI by Stanislawski [84], Iranmanesh et al. [37], and Liu et al. [38], key factors contributing to PI can be identified and categorised into two groups: those inherent to the business itself and those related to the specific context in which the business operates.

Regarding the specific antecedents in business, previous strategies shape firms’ PI over time as past decisions influence organisational routines and constrain future innovation pathways [88]. PI therefore reflects not only forward-looking intentions but also historically embedded, path-dependent dynamics as past decisions shape organisational routines, resource allocations, and the strategic context within which future innovation activities occur [89].

Based on the study conducted by Grando et al. [90] focused on agri-food SMEs, nine relevant macro-classes of strategies (along with their subclasses) can be identified as affecting future business decision processes and outcomes: intensification and upscaling, technological innovation, market orientation, financialization, multifunctional diversification based on the framework of [91], risk management, coping with firming decline and externalization (While Grando’s [90] approach addresses the macro-strategic category of “Blurring firm borders,” which includes externalizations and partnerships, this work aligns with De Martino and Magnotti’s [33] perspective, which views partnerships as external resources). Alongside these strategies, pricing should be included. It is crucial as it determines not only the product pricing but also helps portray value, impacts on market competitiveness, attracts customers, facilitates customer value creation, and meets their expectations. However, it can also inhibit profitability and, in turn, affect the business’s IC [92,93].

The business’s perceived strengths and weaknesses represent key business-specific antecedents of PI, as they are essential for interpreting competitive conditions, guiding decision-making across different business situations, and shaping strategic orientations. In this sense, SMEs are more likely to engage in innovation activities when they recognise internal competencies and organisational capabilities as strategic assets, whereas perceived constraints—such as high innovation costs, limited financial resources, or lack of qualified personnel—tend to reduce their innovation propensity [94,95].

Considering from its specific-context dependence point of view, emerges that firms’ PI also reflects their perceptions of the external environment—including industry conditions, competitive dynamics, and market opportunities and threats [96].

To explore the factors—both internal and external—that shape a firm’s PI, the SWOT (Strengths, Weaknesses, Opportunities, and Threats) analysis serves as a robust and insightful tool, offering a clear understanding of a firm’s perception of its strategic landscape and the key drivers that may support or hinder innovation [97].

In the specific-context perspective, market barriers cannot be ignored, as they represent a factor that hinders innovation [53,98].

2.4. Willingness to Innovate (WI) and Innovation Capability (IC)

The WI, which preceded PI [35,36], is rooted in human dispositions and refers to a firm’s favourable disposition to engage in distinctive behaviour [99,100]. It reflects the organisation’s openness to growth, its readiness to engage in innovation-oriented activities [99,101], and its broader innovation orientation, which underpins organisational ambidexterity—the ability to balance the exploitation of existing capabilities with the exploration of new ones [34]. Therefore, WI is crucial for long-term growth, as it drives new technologies, processes, and product adoption [102]. However, it represents a necessary but not sufficient condition for IC, since its translation into outcomes depends on the strategic orientation and the ability to mobilise resources based on governance and network [103].

In SMEs, WI plays a central role in stimulating organisational processes that may lead to both internal and external resource development. Specifically, WI operates as an antecedent that activates investment and reconfiguration processes within the firm, thereby transforming strategic intent into tangible resource commitments. Entrepreneurial orientation—closely related to WI—is positively associated with IC and supports the flexible allocation of internal resources in response to dynamic environmental conditions, while also fostering organisational learning and knowledge creation processes [104,105]. Firms characterised by a strong innovation orientation are more likely to invest in human capital and organisational capabilities—such as employee training, R&D activities, and internal knowledge sharing—thereby strengthening their internal IA base. Through these processes of internal resource development and capability enhancement, WI may contribute to the gradual strengthening of IC. In this sense, IC can be understood as the outcome of sustained investments in organisational learning, human capital, and knowledge integration, which are facilitated—but not automatically determined—by a firm’s willingness to innovate. Simultaneously, this orientation encourages openness to external knowledge and support, including collaborations with universities, partnerships with other firms, and access to institutional or financial resources [51,106].

In the agri-food sector, innovation is typically incremental, encompassing not only products but also processes, firm organisation, and even marketing activities, extending along the supply chain [29]. Therefore, starting from the various innovation forms identified by Schumpeter [39], WI can be defined as openness to change towards new production processes, marketing practices, organisation—referring to new forms of partnership—and innovation investment.

Precisely, the willingness towards marketing practices can be declined in the two Ps of product and place. The growing centrality of both physical and intangible product attributes as key drivers of business competitiveness underscores the strategic importance of a firm’s willingness to introduce new products and develop new brands [107]. In parallel, openness to exploring new distribution channels and entering new markets proves equally critical, since it plays a crucial economic and social role in achieving business and marketing objectives and creating additional access to resources, enabling businesses to scale up quickly when new opportunities arise [108].

An important form of innovation for SMEs is organisational innovation, understood by Geldes et al. [109] as networking with other organisations and focusing on new methods of organising external relations with other firms or public institutions. For agri-food SMEs, collaborative relationships are a decisive lever for innovation and long-term competitiveness. In a sector shaped by rapid structural and market changes, collaboration is no longer optional but essential: through network-building and knowledge exchange, these firms can strengthen their organisational structures, enhance product valorisation, access critical resources, and generate sustained value creation [50,110,111,112]. Research institutions, in particular, contribute essential resources—such as human capital and infrastructure—across various innovation stages [113]. Therefore, the willingness to develop collaborative relationships with entities, such as research institutes, universities, public organisations, and various stakeholders along the supply chain—both vertically and horizontally—is crucial in understanding a business’s IC [33].

Furthermore, innovation investment in terms of both human and financial resources in all stages of innovation development is a prerequisite to generate innovation outcomes [6,114]. SMEs often encounter extra challenges like limited technology, technical skills, and financial resources, which can hinder them from investing in innovation [115]. Consequently, in the context of SMEs, the favourable disposition to engage in innovation investment is essential.

2.5. Innovation Ability (IA) as Resource Mobilisation Capacity

According to Teece [116], Lisowska [117] and Chrisman et al. [101], IA is a broad concept concerning the organisation’s ability to allocate, dispose of, and reconfigure its resources in the area of innovation to action and adapt to an ever-changing context. From this perspective, competitive advantage does not rest primarily on specific products, but on the pool of resources firms possess and, above all, on their ability to reconfigure them in response to change [118]. In this sense, IA refers to the firm’s capability to transform organisational, relational and institutional resources into new products, processes or technological adaptations. It relies on firms’ ability to acquire, assimilate and exploit knowledge, which in agri-food SMEs is largely accessed through supply-chain relationships, networks and institutional support mechanisms rather than in-house R&D. Therefore, the nature and quality of firms’ resources play a crucial role in shaping their innovation capacity and, consequently, their ability to generate and implement innovations [33].

Within this field, starting from Grunert et al.’s [47] contribution to IC in the agri-food sector, the previous literature sometimes focused on internal resources and sometimes on external resources innovation processes activators within the firm and along the supply chain [33,119,120,121,122]. Internal resources, including tangible assets and employee skills, are crucial for fostering innovation, while external knowledge and resources acquired through a firm’s relationships can complement internal capabilities and enhance IA [123]; conversely, the absence of internal resources may slow IC development, whereas external factors such as public infrastructure, policies, market conditions, and competition shape firms’ innovation strategies [103]. According to previous studies [33,65,124], IA is here understood as a construct built upon multiple internal and external variables that interact with each other to drive innovation within agri-food businesses and along their supply chain.

Based on the studies by De Martino and Magnotti [33] and Castillo-Valero and García-Cortijo [5], six internal (size, qualified staff, export orientation, R&D investment and legal form) and three external resources (market and science-based information sources, collaborative relationships both vertically and horizontally across the supply chain, including the research institute, university, and public entity, and public financing) influencing agri-food SMEs’ IA can be identified.

Business size can be seen as a measure of available resources [125] and is a pivotal internal factor affecting agri-food innovation due to the high operation and maintenance costs involved, which only profitable firms can cover [5]. Business size is defined by its revenue and number of employees [33]; however, the training profile of the business workforce (technical, professional and managerial) can contribute to or limit its IA [5]. Orientation to the export is crucial for SMEs aiming to grow and progress, and for product innovation along the supply chain [33,126]. Although the agri-food sector’s reputation is that of a low-technology and research-scarce industry [11], investing in R&D remains a crucial factor for fostering innovation and supporting long-term business success [33].

A firm’s access to external knowledge significantly shapes its IC. First, given the financial constraints commonly faced by SMEs, which can significantly limit their ability to innovate, public funding becomes a crucial resource in supporting their R&D efforts [33,127].

Second, for SMEs, leveraging external market and scientific information enhances their ability to identify innovation opportunities, understand competitive dynamics, formulate effective strategies, and monitor their progress [128,129,130]. External actors—such as suppliers, competitors, customers, universities and research institutions—provide critical insights into market needs, emerging technologies and technical solutions [131]. Interorganizational collaboration—whether vertical, horizontal, or institutional—serves as a vital mechanism for knowledge exchange that fosters innovation while reducing costs and risks [132]. In the agri-food sector, such collaboration operates as a crucial external resource by facilitating co-creation, risk sharing, integrating dynamic capabilities and access to complementary competencies, thereby strengthening firms’ innovative performance, promoting technology adoption, and supporting systemic transformation, particularly where internal R&D capacities are limited [51,133,134,135].

Ultimately, one aspect that has not been widely addressed in the literature is the legal form of firms, as distinctions between different types of firms and cooperatives are usually overlooked [5]. Cooperatives—distinct from private firms in their statutory objectives, governance structures, investment patterns, and shared member responsibilities—play a crucial role in the development of innovation systems from both evolutionary and territorial perspectives. Their share capital constitutes a strategic intangible asset that strengthens innovation processes and enhances the capabilities required for generating and disseminating knowledge and information [5,136]. Participation in cooperatives positively influences technology adoption, fosters incremental innovation, supports social innovation processes, and enables members to benefit as co-innovators [137,138,139,140]. Within the European Union agri-food system, agricultural cooperatives are pivotal in improving members’ market access, reducing opportunistic behaviour along supply chains, and promoting sustainable agricultural practices. Empirical evidence indicates that cooperative membership mitigates supplier and buyer opportunism, thereby facilitating compliance with sustainability standards and reinforcing competitiveness [141]. Bibliometric analyses further reveal a growing scholarly focus on cooperative governance, social capital, sustainability, innovation, economic performance, and digitalisation in the EU context [142,143]. Survey-based research also confirms that cooperatives make a substantial contribution to sustainable agriculture and food security by encouraging resource sharing, collective bargaining, and information exchange, ultimately enhancing farm productivity and resilience [144].

2.6. Interplay Among WI–PI–IA

The WI–PI–IA framework is based on the idea that IC is not simply an aggregation of isolated dimensions but an interdependent capability-building process, particularly relevant for agri-food SMEs. In this model, WI represents the dispositional prerequisite for innovation, PI constitutes the strategic orientation through which disposition is translated into concrete actions, and IA reflects the operational capacity to realize innovation by mobilizing available resources. IC thus emerges from the sequential and dynamic interplay of these three elements, in contexts characterized by resource constraints and predominantly incremental innovation patterns. Building on the RBV and outcome-focused approaches to innovation, the framework focuses on SMEs in the agri-food sector and integrates established perspectives [35,36] that consider WI as preceding PI, while IA depends on the ability to act on PI to transform resources into tangible innovations [34]. This process-oriented logic distinguishes the model from previous multidimensional approaches [55,61,70,74], which often treat IC as a relatively static set of internal or functional competences, emphasizing organizational processes, technological upgrading, or leadership. The innovative contribution of the WI–PI–IA framework lies in its layered and context-sensitive architecture: it clearly differentiates disposition, strategic orientation, and operational capacity, situating IC within governance arrangements and institutional contexts, which are particularly relevant in traditional agri-food SMEs where innovation is driven more by cooperation, incremental adaptation, and territorially embedded networks than by science-based technological intensity. In this way, the model captures IC as a relational phenomenon, sensitive to context, and dependent on how resources and strategies interact to generate innovation. Figure 1 provides a schematic representation of the proposed framework.

Unlike traditional multidimensional models of innovation capability, which conceptualise IC as a set of coexisting resource-based dimensions, the proposed framework adopts a processual perspective that disentangles its dispositional, strategic, and operational components. By distinguishing WI, PI, and IA, the model captures the sequential transformation of cultural orientation into strategic commitment and, ultimately, into resource mobilisation. This distinction is particularly relevant in low-tech agri-food SMEs, where innovation does not primarily emerge from formal R&D structures but from incremental, path-dependent, and territorially embedded processes. Accordingly, the framework contributes to the literature by offering a context-sensitive interpretation of innovation capability formation rather than a purely dimensional classification.

2.7. Hypothesis Development

The proposed framework conceptualises innovation capability as emerging from the structural relationships among WI, PI, and IA. In line with Figure 1, the hypotheses correspond to the directional paths represented in the model.

The first path (WI → PI) reflects the transition from dispositional openness to strategic commitment. Firms characterised by stronger WI are more likely to formalise innovation-oriented strategies and exhibit a higher PI.

H1. 

(WI → PI). WI is positively associated with PI.

The second path (PI → IA) captures the translation of strategic orientation into operational capability. A stronger propensity to innovate enhances the firm’s ability to mobilise resources, coordinate processes, and implement innovative solutions. Moreover, in low-tech agri-food SMEs, cooperative embeddedness, territorial networks and public financing may strengthen the extent to which strategic orientation translates into internal resources. Given the multidimensional nature of IA, this relationship is examined with respect to both internal and external dimensions of innovation ability.

H2a. 

(PI → IA). PI positively influences firms’ Internal IA.

H2b. 

(PI → IA). PI is positively associated with External IA.

H2c. 

(Moderation on Path b). Sector-specific embeddedness positively moderates the relationship between PI and Internal IA, such that the relationship is stronger in firms with higher levels of embeddedness.

The third path (WI → IA) captures both the direct and the mediated relationship between WI and IA. At the organisational level, WI reflects a cultural orientation toward innovation characterised by openness to change, tolerance for experimentation, and managerial support. Such cultural dispositions shape behavioural routines and learning patterns over time, thereby fostering an internal climate conducive to capability accumulation. This direct influence is particularly relevant in low-tech agri-food SMEs, where innovation frequently emerges through incremental learning-by-doing and learning-by-interacting. In such contexts, a favourable disposition toward innovation stimulates informal experimentation, relational learning, and gradual skill enhancement, strengthening both internal and externally embedded components of IA. Moreover, from a path-dependent perspective, capability building is cumulative and historically embedded; sustained willingness to engage in innovation activities contributes to the progressive development of routines, relational ties, and absorptive capacity that consolidate into structured innovation abilities. Therefore, WI may also exert a direct—though partial—effect on IA. Nevertheless, dispositional openness alone does not automatically generate operational outcomes. Its impact is largely channelled through PI, which reflects the firm’s deliberate strategic commitment to innovation-oriented decisions, resource allocation, and competitive positioning. In this sequential logic, WI operates as the upstream attitudinal driver, PI as the strategic conduit, and IA as the downstream operational manifestation of innovation processes. The indirect path (WI → PI → IA) reflects the sequential logic of the framework. Accordingly, the framework assumes a partial mediation structure, in which WI influences IA both directly and indirectly through PI.

H3a. 

(Direct effect WI → IA). WI is positively associated with IA.

H3b. 

(Indirect effect WI → PI → IA). The relationship between WI and IA is partially mediated by PI.

3. Methodology

3.1. The Sector

The European sheep dairy industry has gained increasing importance globally over the years [145] as it greatly improves food security and supports economic growth and development in both developed and developing regions [146]. Dairy processing, a vital subsector of the European Union’s food processing industry [147], to be sustainable, it must be both efficient and competitive [148], which in turn depends on innovation [7,10].

This research focuses on the sheep dairy industry in Sardinia (the second largest island in the Mediterranean) as it serves as a valuable case study for several reasons. Sardinia’s sheep milk industry dominates the country’s entire sheep milk landscape and takes a leading role in the regional economy [149]. In 2017, sheep and goat milk and meat produced in Sardinia accounted for 58% and 41% of national production and 13% and 4% of agricultural regional output [150]. Sardinia also houses 40% of the national sheep livestock, contributing over 68% of Italy’s sheep milk production [151]. The “Pecorino Romano PDO” (PR) is Sardinia’s leading cheese, representing over 60% of the island’s total cheese production value. It is also a significant export product, especially to the US market, ranking third in Italy behind Parmigiano Reggiano and Grana Padano [152,153].

Italy, along with France, Greece, and Spain, which together produce approximately 46% of the sheep milk in the Mediterranean [154], has adopted a modern transformation process aided by advanced commercial systems [145]. Since 2000, the Sardinia sheep milk industry has undergone significant restructuring due to a crisis in the traditional PR market and the cessation of export aid. The Sardinian private and cooperative sheep dairy firms pursued innovation and modernisation strategies and strengthened their commercial and distribution in the past. For example, they implemented (i) diversification of cheeses, moving away from a single product to offer various types of pasteurised pecorino cheeses, giving rise to actions to reevaluate and improve the quality of PR, (ii) organisational innovation processes by introducing new “quality” and “commercial” departments; (iii) actions aimed at reducing the weight of intermediaries along the supply chain [155].

Recent research adopting a systems perspective highlights the complexity of the Sardinian dairy sheep sector, where innovation must integrate environmental, economic, and social objectives to enhance sustainability performance and policy coherence [156]. Locally produced sheep milk systems also provide a broad range of ecosystem and social services, whose public recognition influences the perceived value and innovation potential of extensive dairy sheep farming [157]. Moreover, recent innovation projects in the Sardinian sheep dairy sector, such as the interactive innovation activities of the PASCHINRES Operational Group [158] and the regionally supported digital platform APPàre [159], illustrate ongoing efforts to introduce participatory methods, knowledge-sharing tools, and data infrastructure to enhance resilience and business performance in traditional livelihood systems.

That evidence highlights the vital role of the Sardinian sheep and dairy sector in driving regional economic development and how any issue related to sheep milk production in Sardinia resonates at regional and national levels in economic and social terms. Therefore, as improving firm productivity through innovation is crucial for increasing profitability, it is particularly suitable to investigate the IC of the Sardinian sheep dairy industry.

Moreover, as many firms operate as cooperatives, investigating the IC of the Sardinian sheep dairy industry can provide valuable insights into how different legal forms may influence the IA of businesses, as well as the role of cooperatives in the European agri-food landscape [160].

3.2. Research Design

The study adopts a quantitative explanatory research design grounded in the hypothesis-guided framework developed in Section 2. The empirical analysis examines the relationships among WI, PI, and IA in agri-food SMEs, with particular attention to both direct and mediated effects.

Given the sector-specific population and the case-oriented nature of the study, the empirical analysis relies on composite indicators and regression-based mediation analysis rather than full latent-variable structural modelling.

3.3. Data Collection and Sample

Using the AIDA database, 58 firms (both private and cooperatives) operating in the Sardinian sheep milk processing sector in 2023 were identified. Of these, 15 were eliminated because (i) they were found to be inactive; (ii) the firm’s activity is purely the commercialisation of products branded by other firms; (iii) there were duplicates of firms operating under multiple firm names for fiscal and/or organisational reasons. Additionally, since the survey targets SMEs—according to the Commission Delegated Directive (EU) 2023/2775 dated 17 October 2023—two large businesses, which together account for 23.90% of the sector’s total turnover in 2023, and four micro enterprises, representing only 0.23% of the sector’s turnover, have been excluded from consideration.

Data were collected through a structured questionnaire administered to owner-managers and senior decision-makers of agri-food SMEs. This choice was considered appropriate for two main reasons: (i) it allowed the operationalisation of the theoretically defined constructs (WI, PI and IA) and ensured comparability across firms within the sectoral case study; (ii) in SMEs innovation-related decisions are typically concentrated at the owner-manager level; therefore, a structured instrument enables the systematic collection of self-reported managerial perceptions regarding innovation strategies, resources and organisational behaviour. This approach combines the contextual depth of a sector-specific case study with the analytical consistency required for hypothesis testing.

Research occurred in the field for two months (from 1 June to 31 July 2024). To get the maximum number of answers, firstly, all firms were contacted by telephone to present the purpose and significance of the research and to request their participation. Subsequently, a first email has been sent requesting them to complete the online questionnaire. Subsequently, two more reminder emails were sent. As of 1 August, 20 firms (11 private and 9 cooperatives) responded fully to the questionnaire, equalling 49% of the sector’s total number of firms, 38.72% of its total turnover, and 51.43% of the sector’s total turnover generated by SMEs in this sector.

3.4. Operationalisation of Constructs

The three constructs—WI, PI, and IA—are operationalised as composite indices derived from theoretically coherent groups of variables. In particular, PI and IA are computed by combining their two complementary dimensions, reflecting their multidimensional nature.

Composite indicators were constructed through a structured three-step procedure.

First, all selected variables were standardised using z-score normalisation based on the sample standard deviation. This transformation ensures comparability across different measurement scales, centres each variable around zero, and preserves relative dispersion. The use of z-score standardisation is particularly appropriate given the heterogeneous scales of the original variables and the subsequent inferential analyses.

Second, internal consistency was assessed using Cronbach’s alpha for constructs conceptualised as reflective dimensions. Reliability analysis was conducted on the standardised variables to avoid distortions arising from scale heterogeneity.

Third, the composite indices were calculated as the arithmetic mean of the standardised variables associated with each construct. Equal weighting was adopted to preserve theoretical coherence and transparency, consistent with the conceptualisation of WI, PI, and IA as configurational constructs emerging from the joint contribution of complementary components. Moreover, no strong theoretical or empirical justification exists for assigning different weights to the variables. This approach is commonly used in exploratory sectoral studies where the objective is to capture the overall configuration of related factors rather than to estimate latent-variable structures, e.g., [161]. For multidimensional constructs (PI and IA), each sub-dimension was first computed separately and subsequently aggregated to obtain the overall index.

Finally, IC is conceptualised as the systemic configuration emerging from the interaction among WI, PI, and IA, rather than as a directly measured variable.

3.5. Analytical Strategy

The empirical analysis proceeds in two stages. First, descriptive statistics are presented to provide an overview of the sample characteristics and the distribution of the constructed indices.

Second, the hypotheses are tested using linear regression analysis. To examine the mediating role of PI in the relationship between WI and IA, as well as the moderating role of External IA in the relationship between PI and Internal IA, bootstrap-based procedures (5000 resamples) are employed. Multicollinearity diagnostics (VIF) were also conducted to assess potential multicollinearity among the predictors, and the results indicated that multicollinearity was not a concern.

Given the relatively small population of firms operating in the Sardinian sheep dairy processing sector, the empirical analysis adopts a sectoral case-study perspective rather than aiming at statistical generalisation. The sample of 20 firms represents a substantial share of the population in terms of both the number of firms and sector turnover. To further strengthen the reliability of the findings, the analysis relies on standardized coefficients, effect size measures (Cohen’s f²), and bootstrap resampling procedures, which provide robust estimates even in small samples. All statistical analyses were performed using IBM SPSS Statistics (version 29) (IBM Corp., Armonk, NY, USA).

3.6. Variables and Questionnaire

Before being administered to companies, the questionnaire was pre-tested with a small group of sector experts and academics to ensure the clarity, relevance, and consistency of the questions. The structured questionnaire comprising four parts was created in Jotform in Italian. The English translation of the questionnaire can be viewed at the link: https://form.jotform.com/251945597373370, accessed on 29 January 2026. Participants were first introduced to the study through a brief explanation of the survey’s objectives. They were informed about ethical considerations and how their data would be processed anonymously in accordance with EU Regulation No. 2016/679 (GDPR). After this introduction, participants were invited to complete four sections of the questionnaire. Respondents were free to skip questions; however, the questionnaire design encouraged complete responses, and the final dataset contains no missing values for the variables used in the empirical analysis. Only the business name was mandatory to link the responses with financial information obtained from the AIDA database.

The first section included questions regarding businesses’ IA. Concerning the internal resources, while data on revenue and legal forms was gathered by Aida database, firms were asked to indicate (i) the number of temporary and permanent employees, (ii) the percentage of revenue invested in R&D activities, (iii) the percentage of production sold through various channels, and (iv) the type of qualified staff the business can rely on.

Concerning the external resources, questions were asked regarding the subjects and/or institutions from which the firm obtains useful information for defining its strategies, external entities with which relationships have been established, and the sources of public funding from which businesses benefit. The second section focuses on the firm’s WI. It includes specific questions rated on a scale of 1 to 5 about the firm’s openness to engaging in new endeavours, such as production processes, products, brands, distribution channels, markets, partnerships, and making new investments to foster innovation.

The third part investigated the PI of businesses. It begins with a question about how necessary the firm believes it is to introduce innovations into its organisation, using a scale of 1 to 5. Following this, first, specific questions are posed regarding the firm’s previous strategies adopted. We were specifically asked about the main strategies previously pursued from a list of nine possible options, and for each pursued strategy, which specific actions were taken. Subsequently, close questions were asked regarding perceived sector barriers, strengths and weaknesses, and broader opportunities and threats identified. The fourth section allows the interviewee to provide additional comments.

4. Results

4.1. Descriptive Statistics

The first step in applying the elaborated framework was to examine the firms’ profiles and available resources, which act as antecedents to IA. The legal structure of a firm, considered as an external resource, can influence its dynamics regarding IC. Therefore, when possible, the results will be categorised to distinguish between those derived from private firms and those from cooperatives.

There is a significant difference in turnover percentages both in the total sample and in the partial samples. Among the firms that have participated in the survey, private firms are in the first position in terms of turnover. The number of employees tends to vary depending on the firm’s size, suggesting the adequacy of this data (

Table 1. Firms’ profile.

Firm	Years of Activity	% Revenue on Total SMEs in the Sector	Temporary and Permanent Employees
Cooperative 1	100	1.33	17
Cooperative 2	70	1.14	20
Cooperative 3	44	1.52	17
Cooperative 4	54	0.50	15
Cooperative 5	76	0.49	12
Cooperative 6	60	0.25	32
Cooperative 7	117	5.94	65
Cooperative 8	73	0.90	29
Cooperative 9	65	1.27	20
Private 1	34	8.49	65
Private 2	61	9.00	75
Private 3	26	2.84	23
Private 4	42	0.64	8
Private 5	32	1.63	21
Private 6	38	0.85	15
Private 7	74	0.97	17
Private 8	27	6.02	85
Private 9	7	0.28	8
Private 10	51	6.74	100
Private 11	50	0.60	18

).

Data showed no significant differences in the choice of outlet market, and both private and cooperative firms favour selling in the national market (77% and 63%, respectively). In terms of export orientation, only three cooperatives are greatly engaged in this activity, with their export sales percentages ranging from 20% to 32%. Among private firms, one participates in international markets with 20% of its sales, while two others have a much stronger presence, with 70% and 90% of their sales coming from exports, respectively (all details about the questionnaire responses are available in the Appendix A).

When we examine the sales channel, we find that the firm store is the primary channel for only one cooperative, and local markets have a residual impact on sales. The Ho.Re.Ca. channel is significant for one cooperative and three private firms. Sales to large-scale retailers and wholesalers present a different situation. While private firms adopt both channels, cooperatives prefer wholesalers.

Turning to the analysis of the other internal resource antecedents of IA (see

Table A2. Internal resources: The number of companies that employ specific professional roles among their staff.

Firm	Cheesemakers	Food Technologists	Logistics Managers	Highly Skilled Administrative Staff	None
Cooperatives	6	2	3	6	0
Privates	7	5	3	7	0
Total	13	7	6	13	0

in the Appendix A), the distribution of qualified personnel highlights a greater concentration of specialised staff in private firms, which employ a greater number of food technologists and logistics managers than cooperatives, suggesting a slightly higher level of investment in technical and managerial skills. Concerning the cheesemakers and highly skilled administrative personnel, 65% of the firms—with minimal differences between private and cooperative firms—report the presence of these figures among their staff.

Five cooperative and seven private firms reported a standard R&D investment level of 10%; the second-largest private firm by turnover reports an investment of 20% (see

Table A3. Internal resources: Percentage of revenue invested in research and development activities.

Firm	%Revenue Invested in R&D Activities
Cooperative 1	10
Cooperative 2	10
Cooperative 3	10
Cooperative 4	10
Cooperative 5	10
Cooperative 6	0
Cooperative 7	0
Cooperative 8	0
Cooperative 9	10
Private 1	20
Private 2	10
Private 3	10
Private 4	10
Private 5	10
Private 6	0
Private 7	10
Private 8	0
Private 9	10
Private 10	0
Private 11	10

in the Appendix A). In analysing the external resource antecedents of IA (see

Table A4. External resources: public financing.

Firm	Rural Development Programmes	European Programming (i.e., EFSL; FSC)	European Planning (i.e., Horizon, Interreg, Life)	Other Public Financing	None
Cooperatives	4	0	0	5	0
Privates	1	2	0	6	2
Total	5	2	0	11	2

,

Table A5. External resources: information sources.

Firm	Universities	Consortia	Agronomists	Milk Suppliers	LSD
Cooperatives	2	3	2	3	3
Privates	1	2	1	1	3
Total	3	5	3	4	6

and

Table A6. External resources: information sources (continued).

Firm	Wholesalers	Fairs	Word of Mouth from Other Processors	None
Cooperatives	6	4	6	3
Privates	3	3	3	1
Total	9	7	9	4

in the Appendix A), it is noted that only one private firm and four cooperatives participated in rural development programs. Additionally, only two private firms benefited from European funding, while 55% of private firms and cooperatives received support from other public financing sources.

The analysis of the external sources utilised reveals a general reluctance to engage with such resources, as evidenced by the fact that 20% of the firms do not rely on external information. Notably, one private firm and two cooperatives report using scientific information obtained from universities. Regarding external market information, 67% of cooperatives rely on wholesalers and word-of-mouth from other processors, in contrast to only 27% of private firms.

Finally, 30% of the firms reported having no form of collaboration with external entities. Cooperatives, in particular, are the most likely to establish such collaborations, especially through affiliation with production consortia and other collective organisations or partners, either under formal agreements or informal arrangements.

When we delve into the WI construct, the data reveal distinct patterns in innovation priorities and approaches between the two organisational forms. The findings suggest that while both cooperatives and private firms are engaged in innovation, their strategies differ in orientation: cooperatives tend to follow a more cautious and incremental approach, and private firms appear more dynamic and proactive in adopting innovations that require investment, openness to new markets, and strategic reorientations. These differences suggest that organisational form may influence how firms translate innovation willingness into innovation capability within the WI–PI–IA framework. In particular, private firms appear more able to convert innovation willingness into strategic and organisational capabilities. Overall, private firms show higher levels of WI in almost all categories considered. The most marked differences emerge for the introduction of new products (45% versus 31%) and new commercial channels (44% versus 33%), indicating a greater openness of private firms toward externally visible innovations that drive market competitiveness. Private firms also show higher values for new markets, new forms of collaboration, and new investments, suggesting a greater capacity or willingness to take risks and explore external opportunities. Cooperatives, while showing lower levels, maintain a certain stability across the various areas of innovation, with relatively homogeneous values.

Investigating the PI revealed that four private firms (with a turnover percentage of 12.5% in the total sector and 16.7% of total SMEs, respectively) responded negatively when asked, “Do you think it is necessary for your firm to introduce innovations?”.

Looking at the business-specific antecedents of PI, we first investigated the previous strategies adopted. The data reveal that private firms have adopted a broader and more proactive range of strategies compared to cooperatives. They show a clear lead in key areas such as technological innovation, market orientation, pricing, and diversification, indicating a stronger focus on competitiveness and growth. Both firm types engage similarly in financialization, while cooperatives demonstrate a more cautious approach overall, particularly avoiding strategies like externalisation. Going into the details of the single macro-strategies (detailed in

Table A12. Antecedents to propensity to innovate: detail of main strategies pursued.

	Intensification and Upscaling
Firms	Increase in Workers	Integration with Upstream or Downstream Phases of the Transformation	Increase in Production Capacity	Increase in International Relations	For Cooperatives, Increase in Members	None
Cooperatives	0	0	3	0	4	3
Privates	1	2	4	2	0	3
Total	1	2	7	2	4	6

,

Table A13. Antecedents to propensity to innovate: detail of main strategies pursued (continued).

	Technological Innovation				Market Orientation
Firms	Mechanization	Robotization	Informatization	None	Tools for Recognising the Needs Expressed by Consumers	Communication Tools (Website, Social Media, etc.)	None
Cooperative	5	4	3	2	3	2	1
Private	6	2	6	0	5	4	1
Total	11	6	9	2	8	6	2

,

Table A14. Antecedents to propensity to innovate: detail of main strategies pursued (continued).

	Price
Firms	Price Reduction Due to Reduction in Production Costs	Price Reduction to Gain Market Share	Price Increase to Enhance a Quality Product	Application of a Low Price to Launch a New Product on the Market	None
Cooperative	0	1	2	0	6
Private	1	1	5	3	1
Total	1	2	7	3	7

,

Table A15. Antecedents to propensity to innovate: detail of main strategies pursued (continued).

Financialisation
Firms	Ordinary Credit	Credit Facility	Secured Credit (e.g., from the Warehouse)	Other Financing	None
Cooperative	0	2	1	1	6
Private	5	2	0	0	4
Total	5	4	1	1	10

,

Table A16. Antecedents to propensity to innovate: detail of main strategies pursued (continued).

	Externalisation	Diversification
Firms	Externalisation	Deepening	Broadening	Re-Grounding	None
Cooperative	0	2	2	0	4
Private	2	7	1	0	3
Total	2	9	3	0	7

and

Table A17. Antecedents to propensity to innovate: detail of main strategies pursued (continued).

	Risk Management	Coping with Firming Decline
Firms	Insurance Contracts	Abandonment (Search for Other Employment, Withdrawal from the World of Work)
Cooperative	3	0
Private	3	0
Total	6	0

of the Appendix A), it should be noted that private firms demonstrate a slightly higher incidence of integration with upstream or downstream phases and engagement in international relations. Technological innovation represents another key divergence: while both organisational forms adopt mechanisation, private firms exhibit greater informatisation, but less investment in robotisation. Private firms exhibited more engagement in tools for recognising the needs expressed by consumers and the use of a price increase strategy to enhance the quality of a product. Moreover, diversification strategies—particularly deepening—are more prevalent among private firms (64%) than cooperatives (22%), reflecting a broader commitment to adaptive resilience.

Among perceived strengths (see

Table A18. Antecedents to propensity to innovate: perceived strengths.

Firms	Suitability of Installations and Premises	Supplier Reliability	Good Relationship with LSD, Wholesalers	Good Administrative Management	Market Channels	Certification of Origin
Cooperative	1	2	0	5	0	0
Private	1	1	2	5	0	0
Total	2	3	2	10	0	0

,

Table A19. Antecedents to propensity to innovate: perceived strengths (continued).

Firms	Quality Certifications	Collaborations with Other Companies	Collaborations with Research Institutes	Company Size	Skilled Labour Level	Participation in the Life of the Cooperative
Cooperative	1	0	0	1	0	2
Private	0	1	0	3	1	0
Total	1	1	0	4	1	2

and

Table A20. Antecedents to propensity to innovate: perceived strengths (continued).

Firms	Quality of the Raw Material	Generational Change	Production Variety	None
Cooperative	6	0	1	0
Private	2	0	4	0
Total	8	0	5	0

in the Appendix A), administrative management is noted by five cooperatives and five private firms. Few firms highlight relationships with retailers, firm size, or workforce specialisation. Market channels and certifications are not considered strengths. Notably, raw material quality is valued by six cooperatives and two private firms, while four private firms emphasise product variety.

Regarding perceived weaknesses (see

Table A21. Antecedents to propensity to innovate: perceived weaknesses.

Firms	Suitability of Installations and Premises	Supplier Reliability	Good Relationship with LSD, Wholesalers	Good Administrative Management	Market Channels	Certification of Origin
Cooperative	2	0	1	1	2	0
Private	1	2	0	0	2	0
Total	3	2	1	1	4	0

,

Table A22. Antecedents to propensity to innovate: perceived weaknesses (continued).

Firms	Quality Certifications	Collaborations with Other Companies	Collaborations with Research Institutes	Company Size	Skilled Labour Level	Participation in the Life of the Cooperative
Cooperative	0	1	1	4	1	3
Private	0	0	1	2	3	0
Total	0	1	2	6	4	3

and

Table A23. Antecedents to propensity to innovate: perceived weaknesses (continued).

Firms	Quality of the Raw Material	Generational Change	Production Variety	None
Cooperative	0	3	0	0
Private	0	0	0	2
Total	0	3	0	2

in the Appendix A), no firms identify quality or origin certifications, nor product variety, as problematic. Among cooperatives, firm size is the most cited issue (four cases), followed by low participation in cooperative life and generational turnover (three each), and inadequate administrative management (one). Notably, raw material quality is not considered a weakness by any respondents, and two private firms report no significant criticalities.

In terms of context-specific factors influencing the PI, the perceived barriers (see

Table A24. Antecedents to propensity to innovate: perceived barriers to entry in the industry.

Firms	High Business Startup Costs	Presence of Numerous Competitors	Know-How Needed for Starting the Business	High Bureaucracy	None
Cooperative	7	2	2	4	1
Private	6	6	2	4	0
Total	13	8	4	8	1

in the Appendix A) to entering the market are primarily related to high start-up and sunk costs, followed closely by bureaucratic hurdles. Approximately 55% of private firms experience challenges due to the presence of numerous competitors. Conversely, the barriers to exiting the sector are relatively minor (see

Table A25. Antecedents to propensity to innovate: perception of barriers to exit from the sector.

Firms	Difficulty in Liquidating the Business	Ensuring Employment of Employees	Emotional Attachment to the Business	Hope of Potential Recovery	Specialized Skills Not Usable in Other Types of Businesses
Cooperative	1	2	2	1	0
Private	5	3	2	0	1
Total	6	5	4	1	1

in the Appendix A). Specifically, five interviewees mentioned challenges in securing employment after leaving, and four mentioned an emotional attachment to their current firm. Additionally, five private firms reported adversity in business liquidation.

In terms of perceived market opportunities and threats (see

Table A26. Antecedents to propensity to innovate: perceived opportunities.

Firms	Satisfactory Milk Price	Rural Development Policies to Support Income Diversification	Consumer Attention vs. Sustainable and Local Production	Growth of the “Lactose-Free” Segment	Expansion of International Demand
Cooperative	2	0	3	1	2
Private	1	2	4	0	1
Total	3	2	7	1	3

,

Table A27. Antecedents to propensity to innovate: perceived opportunities (continued).

Firms	Collaboration with Producers	New Consumer Targets	Value of Certification of Origin	Fluctuation of Wholesale Prices	Instability of Production Factor Prices
Cooperative	0	0	1	0	0
Private	2	1	2	1	1
Total	2	1	3	1	1

,

Table A28. Antecedents to propensity to innovate: perceived opportunities (continued).

Firms	LSD Pricing Policies	Defense of PDOs on International Markets	EU and Non-EU Competition	Concentration of Import Activity	None
Cooperative	0	2	0	0	3
Private	0	2	0	0	1
Total	0	4	0	0	4

,

Table A29. Antecedents to propensity to innovate: perceived treats.

Firms	Satisfactory Milk Price	Rural Development Policies to Support Income Diversification	Consumer Attention vs. Sustainable and Local Production	Growth of the “Lactose-Free” Segment	Expansion of International Demand
Cooperative	2	0	0	0	0
Private	1	0	0	1	1
Total	3	0	0	1	1

,

Table A30. Antecedents to propensity to innovate: perceived treats (continued).

Firms	Collaboration with Producers	New Consumer Targets	Value of Certification of Origin	Fluctuation of Wholesale Prices	Instability of Production Factor Prices
Cooperative	0	1	0	4	4
Private	0	0	0	4	6
Total	0	1	0	8	10

and

Table A31. Antecedents to propensity to innovate: perceived treats (continued).

Firms	LSD Pricing Policies	Defense of PDOs on International Markets	EU and Non-EU Competition	Concentration of Import Activity	None
Cooperative	1	1	1	2	2
Private	2	0	2	2	1
Total	3	1	3	4	3

in the Appendix A), firms do not perceive many opportunities in the market. Just over 30% of cooperatives and private firms report consumers’ attention towards sustainable and local production as opportunities. Five firms reported the defence of DOP on international markets (counterfeiting, Italian sounding) and seven (three cooperatives and four private firms) the consumer’s attention towards sustainable and territorially linked production as opportunities. Only two cooperatives and one private firm perceive the satisfactory milk price and the expansion of international demand for cheese in “new consumer” countries (e.g., Asian countries) as market opportunities. The most perceived threat (four cooperatives and six private firms) is the instability of the prices of production factors, followed by the fluctuation of wholesale prices (four each).

To facilitate interpretation of the comparative results,

Table 2. Comparative innovation patterns of cooperatives and private firms within the WI–PI–IA framework.

Dimension (Framework)	Cooperatives	Private Firms	Empirical Evidence	Implication for the WI–PI–IA Framework
Willingness to Innovate (WI)	More cautious and stable innovation orientation across activities	Higher willingness to introduce new products, channels, and investments	Higher WI scores for new products (4.09 vs. 3.44), new channels (4.00 vs. 3.67) and investments (4.18 vs. 3.89)	Private firms show stronger willingness for product, channel and investment innovation, while cooperatives report slightly higher openness toward new markets
Propensity to Innovate (PI)	More conservative strategic behaviour and incremental adjustments	Broader strategic portfolio including technological innovation, diversification, and market strategies	Diversification strategies more frequent in private firms (64% vs. 22%)	Private firms demonstrate greater strategic flexibility in translating innovation willingness into action
Human capital resources (Internal IA)	Similar presence of cheesemakers and administrative staff	Higher presence of specialised technical roles (e.g., food technologists)	Food technologists: 5 private vs. 2 cooperatives; logistics managers similar (3 vs. 3)	Specialised competences appear to support stronger internal innovation capability
R&D investment (Internal IA resources)	Mostly moderate or zero R&D investment	Similar pattern but with occasional higher investments	One private firm reports 20% R&D investment	R&D investments supporting internal innovation capability remain limited across firms but are slightly higher among private firms
External knowledge sources (External IA)	Greater reliance on traditional market-based information sources and slightly higher overall use of external market knowledge	More diversified sources of information	Wholesalers: 6 cooperatives vs. 3 private firms	Different external knowledge configurations may shape firms’ ability to mobilise external innovation resources (External IA)
Collaborative networks (External IA)	Stronger involvement in consortia and collective organisations	More heterogeneous patterns, including firms with no collaborations	Consortia affiliation: 4 cooperatives vs. 1 private firm; firms with no collaborations: 2 cooperatives vs. 4 private firms	Cooperative structures favour institutional collaboration but not necessarily strategic diversification
Market exposure	Lower export orientation and stronger reliance on wholesalers	Higher engagement in export markets and diversified sales channels	Two private firms export 70–90% of sales	Market exposure may reinforce the translation of innovation propensity into innovation ability

summarises the main differences between cooperatives and private firms across the key dimensions of the WI–PI–IA framework emerging from the descriptive analysis.

4.2. Hypothesis Testing

Given the relatively small sample size (although it represents nearly half of the population of firms operating in the sector), the analysis emphasizes standardized coefficients, effect size measures (Cohen’s f²), and bootstrap resampling procedures, which provide robust estimates without relying on strict distributional assumptions.

Hypothesis 1 (Path a) proposes that WI positively influences PI. The results of the linear regression analysis support this hypothesis (see

Table 3. Regression results for Hypothesis 1.

Linear Regression Results
Predictor	β (Std.)	b	SE	t	p-Value
WI	0.38 *	0.84	0.38	2.21	0.040
Constant	—	−0.23	0.32	−0.71	0.486
Model fit
Statistic		Value
R2		0.21
Adjusted R2		0.17
F(1,18)		4.88
p-value (model)		0.040
Cohen’s f2		0.27

* Bold values indicate statistically significant coefficients in the regression results and statistically significant model tests in the model fit section (p < 0.05).

). WI shows a positive and statistically significant effect on PI, explaining 21% of the variance in the dependent variable, indicating moderate explanatory power. The overall model is statistically significant, and the effect size is moderate (Cohen’s f² = 0.27). These results suggest that higher levels of WI are associated with higher levels of PI among the surveyed firms, confirming the expected positive relationship. The reliability analysis also supports the robustness of the constructs used in the model. The WI shows high internal consistency (Cronbach’s α = 0.881), while the PI demonstrates acceptable reliability (Cronbach’s α = 0.72).

Building on the evidence supporting the positive relatio nship between WI and PI, the analysis next investigates whether this propensity contributes to firms’ IA by estimating two regression models: the effect of PI on Internal IA and the effect of PI on External IA (H2, Path b). Reliability analysis shows acceptable internal consistency for the Internal IA (Cronbach’s α = 0.78) and External IA (Cronbach’s α = 0.69) dimensions.

The results indicate that PI has a significant positive effect on Internal IA, consistent with H2a, suggesting that firms with a stronger PI are more likely to develop Internal IA. In contrast, the direct effect of PI on External IA is not statistically significant, indicating that a stronger innovation orientation does not necessarily translate into higher levels of external innovation engagement (see

Table 4. Regression results for Hypothesis 2 (direct effects).

Linear Regression Results
	Predictor	β (Std.)	b	SE	t	p-Value
Model 1: PI influences Internal IA	PI	0.52 *	2.64	0.93	2.84	0.011
	Constant	—	−1.18	0.71	−1.66	0.114
Model 2: PI influences External IA	PI	0.18	0.89	1.12	0.79	0.441
	Constant	—	−0.07	0.83	−0.08	0.936
Models fit
Statistic	Value of model 1: PI influences Internal IA			Value of model 2: PI influences External IA
R2	0.27			0.03
Adjusted R2	0.23			−0.02
F(1,18)	8.07			0.63
p-value (model)	0.011			0.441
Cohen’s f2	0.37			0.03

* Bold values indicate statistically significant coefficients in the regression results and statistically significant model tests in the model fit section (p < 0.05).

).

To further explore the role of external innovation engagement, a bootstrap moderation analysis (5000 resamples) was conducted to test whether External IA moderates the relationship between PI and Internal IA (see

Table 5. Moderation analysis for Hypothesis 2.

Bootstrap Moderation Analysis (Dependent Variable: Internal IA)
Predictor	β	SE	t	p-Value	95% Bootstrap CI
PI	2.47	0.92	2.69	0.015	[0.55, 4.36]
External IA	0.88	0.64	1.38	0.185	[−0.44, 2.18]
PI × External IA	1.12 *	0.46	2.43	0.026	[0.19, 2.08]
Model fit
Statistic		Value
R2		0.41
Adjusted R2		0.32
F(3,16)		3.71
p-value		0.033
Bootstraps		5000
Simple Slopes (Conditional Effects)
External IA Level		Effect of PI on Internal IA			p-value
Low (−1 SD)		1.35			0.087
Mean		2.47			0.015
High (+1 SD)		3.59			0.006

* Bold values indicate statistically significant coefficients in the regression results and statistically significant model tests in the model fit section (p < 0.05).

).

The interaction term (PI × External IA) is statistically significant, indicating that External IA strengthens the positive relationship between PI and Internal IA. This finding suggests that external relational embeddedness does not directly result from PI but instead enhances firms’ ability to transform such propensity into Internal IA. The conditional effects analysis further shows that the positive effect of PI on Internal IA increases as External IA rises. Thus, external relational and institutional embeddedness acts as an enabling context that amplifies firms’ ability to translate PI into Internal IA, supporting the configurational nature of the proposed WI–PI–IA framework.

Finally, hypothesis 3 examines whether WI influences IA (Path c). Reliability analysis shows that the overall IA index has lower internal consistency (Cronbach’s α = 0.42), which reflects the configurational nature of the construct, combining heterogeneous internal and external resources rather than a reflective latent dimension.

The regression results indicate that WI has a positive but non-significant effect on overall IA. However, when IA is decomposed into its internal and external components, WI shows a strong and significant positive effect on Internal IA, whereas no significant relationship emerges with External IA. Overall, these findings suggest that WI mainly translates into the development of Internal IA rather than external relational resources (see

Table 6. Regression results for Hypothesis 3 (direct effect).

Linear Regression Results
	Predictor	β (Std.)	b	SE	t	p-Value
Model 1: WI influences IA	WI	0.33 *	1.21	0.79	1.53	0.143
	Constant	—	−0.19	0.68	−0.28	0.784
Model 2: WI influences Internal IA	WI	0.59	3.67	1.13	3.25	0.004
	Constant	—	−0.43	0.84	−0.51	0.615
Model 3: WI influences External IA	WI	0.14	0.64	1.08	0.59	0.563
	Constant	—	0.17	0.80	0.21	0.836
Models fit
Statistic	Value of model 1: WI influences IA		Value of model 2: WI influences Internal IA		Value of model 3: WI influences External IA
R2	0.12 *		0.35		0.02
Adjusted R2	0.07		0.31		−0.03
F(1,18)	2.35		10.58		0.35
p-value (model)	0.143		0.004		0.563
Cohen’s f2	0.14		0.54		0.02

* Bold values indicate statistically significant coefficients in the regression results and statistically significant model tests in the model fit section (p < 0.05).

).

A bootstrap mediation analysis (5000 resamples) was conducted to assess whether PI mediates the relationship between WI and IA (see

Table 7. Mediation analysis for Hypothesis 3.

Bootstrap Mediation Analysis
Effect	Coefficient	SE	95% Bootstrap CI	p-Value
Total effect (c) WI → IA	1.20	0.78	[−0.41, 2.78]	0.141
Path a WI → PI	0.72	0.28	[0.14, 1.26]	0.018 *
Path b PI → IA	1.56	0.47	[0.64, 2.48]	0.004
Direct effect (c′) WI → IA	0.32	0.74	[−1.11, 1.79]	0.664
Indirect effect (a × b)	1.12	0.41	[0.38, 2.04]	0.007
Model Fit
Statistic		Value
R2 (full mediation model)		0.36
R2 mediator model		0.21
Bootstraps		5000

* Bold values indicate statistically significant coefficients in the regression results and statistically significant model tests in the model fit section (p < 0.05).

). The results reveal a significant indirect effect (β = 1.12, 95% CI [0.38, 2.04]), as the confidence interval does not include zero. After including the mediator, the direct effect of WI on IA becomes non-significant (β = 0.32, p = 0.664), indicating full mediation. These results indicate that WI influences IA primarily through PI, supporting the sequential structure of the proposed framework. Although the theoretical framework allows for partial mediation, the empirical results suggest that the relationship operates predominantly through PI in this sectoral context.

Overall, these findings provide empirical support for the proposed WI–PI–IA framework and highlight the sequential relationship between WI, PI, and firms’ ability to mobilise Internal and External IA.

Table 8. Summary of hypothesis testing results.

Hypothesis	Relationship	Method	Result
H1	WI → PI	Linear regression	Supported
H2a	PI → Internal IA	Linear regression	Supported
H2b	PI → External IA	Linear regression	Not supported
H2c	PI × External IA → Internal IA	Moderation analysis (bootstrap)	Supported
H3a	WI → IA	Linear regression	Not supported
H3b	WI → PI → IA	Mediation analysis (bootstrap)	Supported (indirect effect significant; full mediation)

summarises the empirical results of the hypothesis testing and highlights the sequential nature of the WI–PI–IA framework.

5. Discussion

Grounded in the RBV, DC and KBV, this research advances theoretical understanding of innovation by addressing existing conceptual gaps in the literature, particularly regarding IC in agri-food SMEs, which operate within a sector that is often underrepresented in innovation studies [14,30].

This study contributes to the literature by conceptualising IC in agri-food SMEs as a sequential and context-dependent capability-building process linking WI, PI, and IA. In contrast to many multidimensional models of IC—which often conceptualise IC as a set of parallel organisational resources or competences [29,34,70]—the WI–PI–IA framework interprets IC as a dynamic process linking innovation orientation, strategic enactment, and resource mobilisation. The empirical results provide support for the proposed framework, suggesting that IC emerges from the sequential interaction among WI, PI, and IA and highlighting the sequential nature of capability formation.

Moreover, focused on the Sardinian sheep dairy sector—an important and representative context in Mediterranean agri-food systems—this framework facilitates a detailed analysis of both the internal and external factors influencing innovation in these firms. The results suggest that IC in agri-food SMEs is shaped not only by internal technological resources but also by governance arrangements, territorial embeddedness, and the ability to mobilise institutional resources. In this context, WI alone cannot explain innovation in agri-food SMEs as it also depends on PI, which is sensitive to context and history, and support received from relational and institutional capabilities. These findings also resonate with the literature on regional innovation systems (RISs) [162,163], which emphasises how innovation processes in traditional sectors are shaped by territorially embedded institutions, networks, and knowledge infrastructures. In this perspective, IC does not depend solely on firm-level resources but also on the interactions among firms, research institutions, public organisations, and sectoral intermediaries operating within the regional innovation environment. In the Sardinian sheep dairy sector, cooperative structures, production consortia, and institutional support mechanisms contribute to shaping the relational context in which firms develop their innovation propensity and mobilise internal and external innovation resources.

The findings support the framework foundation in the RBV by demonstrating that internal characteristics of firms—such as qualified staff, R&D investment, and legal structure— are key enablers of innovation [26,65]. Indeed, sixty-five per cent of firms have cheesemakers and qualified administrative staff, highlighting the importance of internal human resources in supporting innovation. Many firms, both private and cooperative, report an investment, albeit low, in R&D, which is a key internal tangible resource within the RBV model. Notably, firm legal form also emerged as a decisive structural factor shaping innovation patterns, with cooperatives and private firms demonstrating distinct behaviours, consistent with Castillo-Valero and García-Cort [5]. These differences highlight the importance of organisational form in shaping innovation behaviour, confirming that cooperatives and private firms follow partially distinct innovation trajectories within the sector.

Additionally, firms’ active identification and response to opportunities, particularly through organisational restructuring and the strategic reallocation of resources, show that the new framework is firmly grounded in the DS perspective. Private firms appear to exhibit a more proactive and market-driven innovation orientation—through diversification, digitalisation, and product development—while cooperatives focus on gradual transformation, structural refinement, and collaborative partnerships. Such actions, including diversification, exploring new sales channels, forming strategic partnerships, and participating in consortia, exemplify the firms’ ability to “sense,” “seize,” and “transform”, as per the DC framework [68].

Absorptive Capacity validates the grounding of our framework in the KBV. Within this context, cooperative firms showed greater engagement with external scientific sources, thus displaying greater innovation openness and adaptability. In particular, cooperatives maintained frequent formal collaborations with consortia and public institutions and strengthened their knowledge integration processes [71]. Private firms adopted broader strategic approaches (e.g., technological innovation, diversification), while cooperatives were more cautious. SWOT data revealed that lower innovation propensity correlated with reduced opportunity perception and heightened threat awareness. Moreover, firms seem to have difficulty indicating their strengths and weaknesses, but this is essential for defining an effective business strategy and achieving the set objectives.

Lastly, IA is portrayed to be a dynamic function of both internal and external resources. In fact, workforce skills, export orientation and firm size (indicators of production capacity and openness), limited participation in rural development programmes, but significant use of public funding (55% of firms), demonstrate that IA is not simply given by the static availability of resources, but by the firm’s ability to activate and integrate these resources, consistently with the dynamic definition provided by Daronco et al. [34]. It is important that many firms—especially private ones—show limited engagement with external networks, highlighting untapped potential [115].

This study enhances Dobni’s [35] conceptualisation of WI by emphasising structural dependence on the type of firm. While Dobni defines WI as an internal organisational position rooted in culture and openness to innovation, our findings reveal that this disposition varies systematically across different organisations. Private firms demonstrate stronger market-oriented innovation behaviours, such as developing new products and investing in new channels, while cooperatives focus more on process improvements and collaborative approaches. These differences indicate that WI is not only embedded in culture or merely firm size but also influenced by governance models and institutional contexts. Therefore, this study broadens the theoretical understanding of WI by highlighting it as a context-sensitive construct shaped by both firm-level structures and cultural orientations.

This study emphasises the understanding of PI as a path-dependent concept, consistent with the findings of Iranmanesh et al. [37]. Analysing PI antecedents shows that a firm’s propensity for innovation is influenced by its historical experiences and contextual understandings. For instance, private firms exhibited a broader strategic engagement—such as technological innovation, market expansion, and diversification—while some questioned the necessity of innovation altogether. This underscores how strategic orientation, accumulated experience, and environmental perceptions come together to create diverse innovation propensities, thereby confirming that PI is sensitive to context and history.

Lastly, the findings support Daronco et al. [34] in viewing IA as a dynamic construct shaped by internal and external resources. While factors like staff qualifications, R&D investment, and export orientation influence IA, they must be combined with external enablers—such as public funding and collaborations—to achieve innovation. This highlights that IA is an ability that evolves through relational, financial, and knowledge linkages, especially in resource-constrained SMEs. Notably, many firms—especially private ones—reported limited collaboration with external entities, highlighting an underutilised potential for knowledge acquisition. This confirms the structural innovation constraints identified in previous studies [115], especially among SMEs that lack the institutional support networks typical of larger firms or cooperatives.

Collectively, these findings confirm the analytical value of conceptualising IC through the interplay of WI, PI, and IA. By operationalising IC as a process influenced by context, strategic behaviour, and resources, this study addresses the fragmentation observed in earlier models and offers a robust foundation for future theoretical advancements in agri-food innovation research. This study challenges the implicit assumption in part of the innovation literature according to which IC is mainly driven by internal technological resources and R&D investments. In the agri-food context, IC occurs instead as a governance- and network-based phenomenon, where legal form, previous path, and access to institutional resources may play a decisive role.

5.1. Theoretical Implications

This study enhances the theoretical discussion on innovation in SMEs by integrating multiple theoretical perspectives, including the RBV, DC, and KBV, into a comprehensive model. It moves beyond traditional linear models of innovation by conceptualising IC as a composite construct influenced by organisational willingness, contextual factors, and resource-based ability, identifying antecedents that can explain the firm’s innovation, understood as an outcome that contributes to an organisation’s success, growth, and competitiveness.

The proposed framework responds to the call to explore IC as a multidimensional concept and addresses the limitations identified in the existing literature, particularly the lack of empirical applications of theoretical models and operational clarity regarding IC in agri-food SMEs [14,26,30,32]. Furthermore, it addresses the call for case-oriented empirical research to fill the gaps in comprehending how IC is understood and implemented within SMEs, and to highlight the role of mixed contextual factors in IC SMEs’ outcomes [29]. It contributes to ongoing academic discussions by emphasising the significance of context-specific and firm-type-specific pathways to innovation (largely overlooked in prevailing theoretical models), including in the framework and test external aspects, such as barriers to innovation, and focused not only on product innovation but also on other types of capabilities that SMEs should develop [29,33,34].

Moreover, it should be highlighted that the key factors explaining the business and context-specific antecedents of PI, as well as the sub-categories of internal and external resources that elucidate IA, can be adapted to the industry under analysis. Therefore, the proposed framework is well-suited for applying the research to various agri-food sectors and even non-agri-food industries.

Although the empirical analysis focuses on the Sardinian sheep dairy sector, some of the findings appear to reflect broader dynamics observed in agri-food SMEs. In particular, the sequential relationship between WI, PI, and IA is likely to characterise innovation processes in other traditional agri-food systems characterised by small firm size, limited internal R&D capacity, and strong reliance on relational and institutional resources. At the same time, some results are more context-specific. The prominent role of cooperatives, production consortia, and territorially embedded networks reflects institutional characteristics that are particularly strong in the Sardinian sheep dairy sector and may vary across other agri-food regions. Therefore, while the WI–PI–IA framework may provide a useful analytical lens for studying innovation capability in similar agri-food SME contexts, the relative importance of specific resources and governance arrangements should be interpreted in relation to the institutional and organisational configuration of each regional agri-food system. In this sense, the framework may provide a useful analytical lens for interpreting innovation dynamics in other traditional agri-food SME systems.

5.2. Managerial and Policy Implications

The study offers valuable insights for both managers and policymakers. For managers, the framework assists in identifying internal bottlenecks and strengths within the three constructs of intellectual capital, allowing for more focused strategic planning and resource allocation. Specifically for cooperatives, enhancing external connections and improving access to scientific knowledge can lead to better innovation outcomes. From a practical standpoint, it provides policymakers with a diagnostic tool to customise innovation support programs and define effective policy actions that account for structural differences, local context, and sector-specific challenges [33]. The study suggests that support measures be tailored to different organisational forms. This includes fostering collaborative ecosystems, supporting collaborative innovation and knowledge exchange rather than relying solely on formal R&D investments, and simplifying access to public funding programs to overcome the structural and cultural barriers that hinder innovation in traditional agri-food SMEs.

More specifically, the empirical results suggest several concrete policy directions for regional and sectoral innovation policies. First, public innovation programmes should strengthen intermediary organisations and knowledge brokers that facilitate interactions between SMEs and research institutions, given the limited use of scientific information sources observed among many firms. Second, targeted support schemes could encourage collaborative innovation projects within production consortia and cooperative networks, which appear to play a key role in mobilising external knowledge and institutional resources in the sector. Third, policy instruments aimed at strengthening managerial and technical competencies—such as training programmes, innovation advisory services, and technology transfer initiatives—could enhance firms’ internal IA and facilitate the mobilisation of resources required for innovation in agri-food SMEs.

5.3. Limitations and Future Research

This study provides novel insights into the determinants and formation mechanisms of IC in agri-food SMEs through a sector-specific quantitative case study. However, some limitations should be acknowledged.

First, the empirical analysis focuses on a single regional sector—the Sardinian sheep dairy processing industry—and relies on a relatively small sample of firms. Although the sample represents a substantial share of the population in terms of both the number of firms and sector turnover, the regional focus may limit the generalisability of the findings. Future research could therefore extend the empirical validation of the proposed WI–PI–IA framework to other agri-food sectors and geographical contexts, including cross-regional or longitudinal analyses. Furthermore, examining the roles of cooperative governance and inter-organisational networks holds the potential to reveal important strategies for fostering collective innovation.

Second, part of the empirical evidence relies on respondents’ self-perception data collected through a structured questionnaire. While this approach is common in studies on innovation behaviour and strategic orientation in SMEs, self-reported data may involve potential perception biases or subjective evaluations of firms’ innovation practices. This limitation is partially mitigated by the fact that the questionnaire targeted owner-managers and senior decision-makers who are directly responsible for strategic and innovation-related decisions in SMEs. Moreover, several variables used to operationalise IA (e.g., firm size) were complemented with objective data obtained from the AIDA database. Future research could combine perceptual measures with additional objective indicators of innovation performance and expand the number of firms analysed. In addition, further studies could investigate more deeply the relationships among the WI–PI–IA constructs and explore the mechanisms through which propensity and organisational resources translate into innovation ability.

6. Conclusions

This paper demonstrates that innovation capacity in agri-food SMEs is not primarily technology-driven, but stems from governance structures, territorially embedded networks and the ability to mobilise institutional resources, calling for sector-specific approaches to innovation analysis and policy design. This study makes a significant theoretical and empirical contribution to the understanding of innovation in agri-food SMEs by introducing and validating an integrated framework for IC. Conceptualising IC as the result of three interacting dimensions (WI, PI and IA), the model captures the complexity and context dependence of innovation processes in resource-constrained environments.

To the best of our knowledge, this is the first study on agri-food SMEs to investigate IC in the agri-food sector through the lens of these three interconnected constructs.

By applying the framework to the Sardinian sheep dairy industry, the study illustrates how firm type, strategic behaviour, and access to resources collectively influence innovation outcomes. These insights provide a strong foundation for future research and serve as a valuable tool for designing innovation policies tailored to the unique characteristics of agri-food SMEs.