Review Delivering nutrient management impact through farmer-centric research: a systematic review of innovation systems in African smallholder systems Ivan S. Adolwa a,*, Shamie Zingore b, James Mutegi a, Matthew McNee c, Bolaji A. Akorede d, Denver Masidza a, T. Scott Murrell b, Samuel Mathu Ndungu e,h, Eileen Nchanji f, Simon Cook g, Thomas Oberthür b a African Plant Nutrition Institute (APNI), c/o ICIPE Compound, PO Box 30772-00100, Nairobi, Kenya b African Plant Nutrition Institute (APNI), UM6P Experimental Farm, Ben Guerir 41350, Morocco c Independent Consultant, Brisbane, Australia d School of Collective Intelligence, Mohammed VI Polytechnic University, Lot 660, Ben Guerir 43150, Morocco e International Institute of Tropical Agriculture (IITA), c/o ICIPE Compound, PO Box 30772-00100, Nairobi, Kenya f International Center for Tropical Agriculture (CIAT), c/o ICIPE Compound, PO Box 30772-00100, Nairobi, Kenya g Centre for Digital Agriculture, Murdoch University, Perth, Western Australia, Australia h World Vegetable Center, Shanhua, Tainana, Taiwan H I G H L I G H T S G R A P H I C A L A B S T R A C T • Current innovation systems contribute to enhanced co-learning, but do not effectively engage farmers. • The value generated out of an innova tion process determines its efficacy. • On-farm experimentation helps to spur farmer-driven innovation. • Agricultural innovation systems should be aligned with farmers’ processes for sustained impacts. • Actionable and contextualized agronomy a key tenet for on-farm experimentation operationalization. A R T I C L E I N F O Editor: Mark van Wijk Keywords: Innovation processes Participatory research Agricultural knowledge and innovation systems On-farm experimentation Nutrient management Value creation framework African smallholder farming A B S T R A C T CONTEXT: The performance of the agricultural sector in Africa still lags behind other regions. The current average maize productivity of 2 t ha-1 is below the global average of about 6 t ha-1. This low productivity threatens the livelihoods of a majority of the population. Despite decades of research and development in vestments, current agricultural innovation systems remain ineffective in supporting sustainable agricultural transformation. OBJECTIVE: This study traces the evolution of innovation systems in Africa as a backdrop to the adaptation of on- farm experimentation (OFE), which is a novel framework for accelerating research and development (R&D) impact. * Corresponding author. E-mail address: i.adolwa@apni.net (I.S. Adolwa). Contents lists available at ScienceDirect Agricultural Systems journal homepage: www.elsevier.com/locate/agsy https://doi.org/10.1016/j.agsy.2025.104416 Received 21 November 2024; Received in revised form 21 April 2025; Accepted 2 June 2025 Agricultural Systems 229 (2025) 104416 Available online 13 June 2025 0308-521X/© 2025 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ). mailto:i.adolwa@apni.net www.sciencedirect.com/science/journal/0308521X https://www.elsevier.com/locate/agsy https://doi.org/10.1016/j.agsy.2025.104416 https://doi.org/10.1016/j.agsy.2025.104416 http://creativecommons.org/licenses/by/4.0/ METHODS: A systematic review approach is augmented with social network analysis methods and primary and secondary data. RESULTS AND CONCLUSIONS: We find that although current innovation systems have contributed to enhancing co-learning processes and have enabled the partial adoption of improved agronomic practices, resulting in increased nutrient uptake efficiency and crop productivity, several shortcomings have limited their impact. Despite their core focus on participatory and systemic R&D processes, our review points to their inability to effectively engage farmers. Hence, failing to generate scalable learning, and demonstrate sufficient value to farmers and other stakeholders. The OFE initiatives demonstrate how farmer-relevant insights integrated with field-based and digital evidence help spur a farmer-driven innovation development and decision support framework. SIGNIFICANCE: OFE is potentially a powerful enabler of current innovation systems performance as it provides the platform for a transformative farmer-led innovation process. 1. Introduction African agriculture remains gridlocked in a vicious cycle of nutrient mining and low crop productivity, causing widespread hunger, poverty and land degradation (Bationo and Waswa, 2011; Vanlauwe et al., 2017). This situation has been exacerbated by weak institutions that constrain agricultural investments and innovation, leading to low adoption of yield-enhancing technologies, and inappropriate nutrient management (Spielman et al., 2009). From the agronomic point of view, crop yields are determined by genotype (G) x environment (E) x man agement (M) interactions (Zingore et al., 2022). Hence, a proper nutrient management strategy must consider seed germplasm, climatic and soil fertility conditions, and management practices, including fer tilizer inputs. However, there are other factors transcending agronomic considerations that affect actual crop yield on farmers’ fields. These exogenous factors to the farm, such as markets, labor, and institutional services (e.g., extension and access to affordable credit), are also critical for raising crop productivity as part of innovation systems (Ojiem et al., 2006; Sumberg, 2012). Due to the complex and systemic nature of challenges present in African farming systems, e.g., poor infrastructure, unfavorable socioeconomic circumstances, non-supportive policy envi ronments, etc., holistic solutions are paramount. The complex interface between best management plant nutrition and exogenous influences can be addressed only by multiple actors working synergistically within innovation systems. An innovation system is a set of actors involved in innovation pro cesses that integrate the mechanisms underpinning the complex in teractions among disparate actors, organizational cultures and practices, learning behaviors and cycles, and rules and norms (Klerkx et al., 2012). Inherent in innovation systems is the value being created and shared among different actors. Value, in this context, is described as “... increasing ‘gains’ or reducing ‘pains’...” (Cook et al., 2024; Osterwalder et al., 2015). Value is captured in various forms: financial, human, so cial, institutional, physical or natural capital (Dfid., 1999; Malherbe et al., 2024). Financial and physical capital help farmers improve their farming enterprise. Indicators include yield gain, return on investment or labour, and mechanization. Human capital improves the capacity to solve problems exogenous or endogenous to the farm. A key indicator is increased learning. Social capital enhances participation in collective action activities, whereas institutional or political capital helps to backstop the farm enterprise by exerting control over institutional or political processes (Malherbe et al., 2024), e.g. improved capacity to negotiate with financiers or suppliers. Current innovation systems in Africa exist as: transfer of technology (ToT); innovation platforms (IP; Schut et al., 2018); agricultural research for development (AR4D) partnerships (Vanlauwe et al., 2017); or Agricultural Knowledge and Innovation Systems (AKIS; Pascucci and de Magistris, 2011). ToT is based on Rogers (2003) diffusion of in novations, defined as, “...the process by which (1) an innovation (2) is communicated through certain channels (3) over time (4) among members of a social system.” It describes how innovation (supposedly emanating from scientists) spreads across a population (of end users e.g. farmers) through time using various communication channels. An IP is, “...a space for learning and change. It is a group of individuals…with different backgrounds and interests…[who]...come together to diagnose problems, identify opportunities and find ways to achieve their goals” (Homann-Kee Tui et al., 2013). The AR4D is, “...a set of applied research approaches that aim to contribute directly to the achievement of inter national development targets…through growth of and innovation in the agricultural sector” (Thornton et al., 2017). An AKIS, which is a network of agricultural institutions, organizations, individuals engaged in the generation of new products, processes, and/or innovation, combines concepts of agricultural innovation systems (Klerkx et al., 2012) and agricultural knowledge and information systems (Roling, 1989). The national agricultural research and extension systems (NARES), which is the fulcrum for agricultural innovation development in most African countries, have mainstreamed IP, AR4D and ToT approaches to varying extents. Innovation systems on the continent are characterized by transformation processes that are driven by what Conti et al. (2024) refer to as technology mission-oriented scenarios, where agricultural innovation development is mediated and directed by public sector investments. A major focus of such investments has been sustainable agricultural intensification. Pretty (1997) described it as, “...the integrated use of a wide range of methods and technologies to manage pests, nutrients, soil and water…to foster increased diversity of enterprises within farms combined with increased linkages and flows between them.” It has been lauded as one of the ways to bolster crop productivity and meet the food security needs of the continent (Schut et al., 2016; Styger et al., 2011). Coupled with this is the need to build the resilience of current farming systems to cope with climate change and unstable input/output com modity prices. For the last fifty years, significant investments have been made into AR4D organizations to support sustainable agricultural transformation in Africa (Vanlauwe et al., 2017). Within this time, agricultural innovation system development has gone through various paradigm shifts; from top-down or linear to systemic approaches. However, African innovation systems have not been effective in deliv ering viable innovations tailored for smallholder farming systems (Dawson et al., 2016). The success of the green revolution (GR) in Asia and Latin America failed to materialize in SSA as little consideration was given to the relevance of the GR technologies in complex African crop ping systems, resulting in the limited adoption of improved crop vari eties and fertilizer (Evenson and Gollin, 2003). Furthermore, agricultural productivity growth was hampered by infrastructural lim itations and inadequate institutional support (Denning et al., 2009). Subsequently, there is a growing demand for AR4D organizations to demonstrate the impact of their interventions at scale (Schut et al., 2020). Given the disproportionate focus on myopic goals rather than critical underlying processes that underpin successful societal outcomes, it is not surprising that the success of such efforts is limited (Hall and Dijkman, 2019). This study reviews the progress of past and current innovations I.S. Adolwa et al. Agricultural Systems 229 (2025) 104416 2 systems as applied to nutrient management in African farming systems and identifies shortcomings and limitations that can be addressed by a novel systemic on-farm experimentation (OFE) innovation process, which brings agricultural actors together around mutually beneficial experimentation to support farmers’ management decisions (Lacoste et al., 2022). Therefore, the first objective of the study is to synthesize and compare past and current innovation systems by assessing their efficacy in generating different forms of value for diverse sets of agri cultural stakeholders. Specific questions around this objective revolve around the efficacy of current innovation systems and their strengths and shortcomings. The second objective is to demonstrate the concep tual and methodological difference between OFE and preceding inno vation systems. Specific questions around this objective are concerned with how OFE could be used to develop a farmer-centric knowledge discovery system within existing African innovation systems, increase their performance, and adapt them to support or deploy OFE. 2. Theoretical and conceptual framing A theoretical framework that describes the systematic to systemic progression of innovation systems in African agriculture (Fig. 1), informed the systematic review. Systematic approaches stress objective measurements, reductionism, mechanistic synthesis and quantification. Systemic approaches are directed by the complex systems methodology (CSM) entailing non-linear system behavior, systems relationships and holistic effects, and uncertainty in its different dimensions (Cook et al., 2013; Schiere et al., 2004). The innovation systems progress from the Transfer of Technology (ToT) paradigm to AKIS against the backdrop of the novel OFE that has to be adapted to the African context. The ToT paradigm has served its purpose well in the global North and tropical regions such as Latin America and Asia (Dawson et al., 2016); however, results in Africa have been disappointing due to complex and unique socio-organizational conditions and production environments within the continent (Schut et al., 2020). The top-down approaches in African innovation systems (AIS) have been complemented by participatory approaches and inter-disciplinarity, where farmers and even non- governmental (NGO) actors have been involved in the innovation pro cess (Bakker et al., 2021; Chambers, 1994; Hauser et al., 2016). More recently, there have been attempts to embrace systemic thinking in the innovation process, with adaptive management, reflexivity, in terrelationships, and non-linear system behavior becoming important considerations (Klerkx et al., 2012; Schiere et al., 2004). With the emergence of AKIS, agricultural systems are increasingly viewed in terms of complex systems methodologies (CSM), where rapid trans formation emanates from multiple coinciding influences, events, trends or even shocks (Schut et al., 2020). The scope of OFE, which distinguishes it from preceding innovation systems paradigms, focuses on the value it generates for its end users. Hence, we utilize a value creation framework based on the value proposition approach (Osterwalder et al., 2015) that defines how value is generated for diverse actors i.e., farmers, researchers, extension agents and others complemented with the capitals approach (Malherbe et al., 2024). The value emanating from a process, which equates to its efficacy, is captured as financial capital (increase in crop yield and in come), human capital (accelerated learning, increased confidence or status), social capital (improved actor relations), institutional capital (improved negotiation with financiers or suppliers), natural capital (improved soil organic matter, increased nutrient uptake), and physical capital (technological developments, infrastructure; Dfid., 1999; Mal herbe et al., 2024). Value creation entails processes such as co-learning, co-creation and co-development, which encompass joint experimenta tion and exploration between disparate actors (Falconnier et al., 2017; Lacoste et al., 2022). Key to knowledge co-creation is the saliency, credibility, and legitimacy of the knowledge generated (Cash and Belloy, 2020). Salience refers to the connection between the generated knowl edge and daily practices or ‘jobs’ of the end users (Eyers and Howarth, 2024). Similarly, end users should trust the scientific aspects of knowl edge (credibility) and perceive that their vision has been considered (legitimacy; Akimowicz et al., 2022). The scalability of such processes could be enhanced by digitalization (Cook et al., 2013), or through networks (Klerkx et al., 2012). Furthermore, learning occurs through environmental (experimentation, experiential or agency), social (emulation or confirmation) or didactic (externally sourced) modes (Henrich and Gil-White, 2001; Stone, 2016; Sumane et al., 2018). Aki mowicz et al. (2022) contend that the braiding or integration of farmers’ experiential knowledge with technical knowledge derived from external sources may enhance scalable learning that is legitimized (i.e. includes farmers’ input). The OFE paradigm seeks to align these three modes of knowledge acquisition and learning to support farmers’ own decision making while legitimizing technical knowledge drawn from external agents. Underpinning the value creation process are enablers, such as farmer-centric engagement, experimentation at landscape scale, and evidence-driven and specialist-enabled approaches (Lacoste et al., 2022). A farmer-centric approach implies strong roles for farmers in research and development. Operating at the farm or landscape scale ensures experimental designs fit into farmers’ specific production sys tem (Conti et al., 2024; Laurent et al., 2019). A data-driven approach supported by specialists helps to spur endogenous productivity growth (Bramley et al., 2022; Cook et al., 2013; de Lara et al., 2023). Unlike the Fig. 1. The systematic (TT) to systemic (AKIS) evolution of innovation systems in African agriculture (adapted from Klerkx et al., 2012; Lacoste et al., 2022; Norman, 2002; Rogers, 2003; Schiere et al., 2004; Stone, 2016; Van Rooyen et al., 2021) I.S. Adolwa et al. Agricultural Systems 229 (2025) 104416 3 other approaches, OFE’s methodological approach entails a synthesis of hard systems methodology (HSM; mechanistic), soft systems method ology (SSM; qualitative) and CSM (holistic), marking a shift from uni disciplinarity to transdisciplinarity (Fig. 1). 3. Methods 3.1. Literature search strategy This systematic review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA; Moher et al., 2010). The general format for searches was premised on the Population Inter vention Comparator Outcome (PICO) categories complemented by au thors’ expert knowledge in the subject matter (Moher et al., 2010; Zingore et al., 2022). The PICO categories for the innovation processes in African agriculture reviewed and the inclusion/exclusion criteria are tabulated in Table 1. After establishing the inclusion/exclusion criteria, we identified a limited set of papers to serve as a “gold standard’ of the types of papers that would be appropriate for this analysis. Some of these publications include those authored by some of the co-authors of this manuscript. The R Language and Environment for Statistical Computing (R Core Team, 2024; packages: litsearchr, synthesisr, igraph) was used to generate a Boolean search phrase using a methodology described in Zingore et al. (2022). The final printed expression (search string 1) was used in the Web of Science (WoS) and Google Scholar search engines (Table 2). Using expert knowledge, a new and improved search string (search string 2) was generated (Table 2). When this Boolean search string was used in WoS, about 25,000 records were returned. After duplicates were removed and several screenings implemented, 101 articles remained for detailed analysis (Fig. 2). 3.2. Data abstraction and analysis The first objective was addressed with a systematic review. Three reviewers that form part of the co-authorship were involved in screening and reviewing the articles. Data were abstracted using Covidence (Veritas Health Innovation, 2024), then exported to spreadsheets for further synthesis. Microsoft Excel was used to categorize the number of studies according to the regions of focus in Africa. The abstracted data was used to identify and extract key principal terms, which were in turn used to generate a network diagram. We used the NetDraw software package within UCINET 6.0 (Borgatti et al., 2002), to indicate the fre quency or intensity with which key principal terms related to processes (i.e., co-learning, co-creation, co-development, scalable) and enablers (i. e., farmer-centric, specialist-enabled, evidence-driven) within innova tion systems had been mentioned or studied in the literature. Apart from mentions or inference, we also established whether the principles behind the terms had been applied in the studies. A network diagram is used to analyze interactions or relationships of any nature, ranging from biophysical, e.g. protein interactions, to social or commercial, e.g. international trade (Borgatti et al., 2009). Networks consist of nodes (or actors) linked together by dyadic ties and are stored as matrices in spreadsheets for subsequent analysis by an application e. g., the UCINET program (Adolwa et al., 2017). Hence, networks are synonymous with mathematical graphs and have properties such as connectedness, centrality or embeddedness, positioning and tie strength as denoted by frequency of interaction (Borgatti et al., 2009). In our case, the embeddedness of a principal term suggests to what extent it is prioritized in the interventions undertaken by research and develop ment actors. Therefore, the embeddedness and tie strength of principal terms within the network structure could reflect how innovation systems are working as reported in the literature. The network analysis approach thus helps us understand the extent to which various processes and enablers in innovation systems have been successful in generating pos itive outcomes (i.e., behavioral change, adaptive change, institutional change) for nutrient management in African agriculture. 4. Results In this section, the synthesis of the literature on African agricultural innovation systems is presented in three parts. First, we provide an assessment of: a) the operationalization of different innovation system paradigms in Africa b) the advance of AIS along the systematic-systemic approach continuum towards positive nutrient management outcomes, and c) the efficacy from implementation of the innovation systems, in relation to how value was created in each of the six capitals discussed above i.e. financial capital, human capital social capital, institutional capital, natural capital, and physical capital. 4.1. Operationalization of different innovation system paradigms in Africa There exists a rich tapestry of information on African innovation systems transcending nutrient management, seed systems and technol ogies, value chains, and institutional arrangements and processes (Table 3). Nutrient management interventions that were key strategies across innovation stages included fertilizer trees, use of mineral and organic fertilizers, organic agriculture and sustainable intensification. Less than half of the relevant literature touches on the systemic AKIS and OFE approaches (Fig. 3). About 55% of the studies on innovation systems were from Eastern Africa, 38% from Western Africa and 25% from Southern Africa (Fig. 3). Most of the studies on AKIS were concentrated in the Eastern African region, which is not surprising given the large investment in agricultural research and development initia tives in that region. The Eastern African region is host to the head quarters of two Consultative Group on International Agricultural Table 1 Inclusion and exclusion criteria based on PICO PICO Inclusion Exclusion Population of interest Smallholder farmers in Africa Farmers not in Africa Intervention Innovation processes including participatory research implemented in various African agricultural innovation system domains Innovation processes not documented in English Comparator Reference to how the OFE process is implemented to improve agricultural innovation systems OFE process not implemented in Africa Outcome Explicit reference to impacts of innovation processes to livelihoods, income, crop yield, food security, learning and environmental outcomes e.g., nutrient use efficiency Impacts are not linked to crop production Table 2 Final search strings used for identifying relevant articles Web of Science database Search string 1 "((africa* OR farmer* OR farm* OR local* OR smallhold* OR (extens* worker*) OR (medium* resourc*) OR (resourc* endow*)) AND (agricultur* OR knowledg* OR participatori* OR practic* OR system* OR (cowpea* intercrop*) OR (target* option*) OR (technic* option*)) AND (on-farm* OR process*) AND (enhanc* OR fertil* OR perform* OR resourc* OR yield*))" Search string 2 "((africa* OR farmer* OR farm* OR smallhold* OR (extens* worker*) OR (medium* resourc*) OR (resourc* endow*)) AND (agricultur* OR (local*knowledg*) OR participator* OR (act* research) OR system* OR (rural appr*) OR (innovati* syst*) OR (cowpea* intercrop*) OR (target* option*) OR (actor tie*) OR (technic* option*)) AND ((on- farm experiment*) OR Scal* OR (innovat* process)* OR (farm* experiment*) OR (business model*)) AND (enhanc* OR fertil* OR (food self*) OR perform* OR (local adapt*) OR resourc* OR yield*))" I.S. Adolwa et al. Agricultural Systems 229 (2025) 104416 4 Research (CGIAR) centers and numerous regional offices affiliated to the CGIAR and other international AR4D organizations. Also, the maize- mixed farming systems that dominate the region are considered to have the greatest potential to generate high returns-on-investment (ROI) due to the highly dense rural populations and relatively high poverty levels (Garrity et al., 2012). Studies that adopt the participatory research paradigm focusing on mineral and organic fertilizer interventions are spread evenly across the regions except for Central and northern African. The latter, particularly Central Africa, which is dominated by forest-based farming systems, has long been neglected by the donor and research community, probably due to the low to moderate growth potential (Dixon et al., 2001). Participatory research has been conducted for a longer period relative to AKIS, hence more AR4D organizations and actors across the board are familiar with this approach. Farming systems research (FSR) and transfer of technology studies are scarce due to the shift to participatory and systemic approaches in agricultural research and development. The FSR approach coincided with the period when a strong emphasis was placed on low input sustainable agriculture (LISA) technologies (Bationo et al., 2008). These included agroforestry technologies, such as improved fallow, that were widely disseminated in Kenya, Zambia, and Zimbabwe using FSR approaches (Bationo et al., 2008; Hildebrand et al., 1993). Given that OFE is a relatively new perspective globally, it is not surprising that it is the least applied and explored in African agricultural research and development discourses. A continent-wide OFE initiative that co-designed, co-developed and delivered relevant precision nutrient management innovations to farmers in cereal cropping systems across Africa is one of the few examples (Adolwa et al., 2022; Fig. 4). 4.2. The advance of African innovation systems along the systematic- systemic continuum There is a general tendency for the nutrient management strategy to shift from single technologies (e.g. fertilizers, improved seed) to com plex system technologies (e.g. sustainable intensification strategies such as integrated soil fertility management) as innovation systems move from top-down ToT to systemic AKIS (Table 3). It is worth noting, however, that several African AR4D actors and institutions have become more amenable to hybridizing innovation systems by combining top- down, participatory and systemic approaches (Conti et al., 2025). Several African AR4Ds and project initiatives have incorporated ele ments of participatory research, conceptualized in the 1990s to move farmers towards the center of research processes (e.g. Defoer, 1998; Falconnier et al., 2016; Misiko et al., 2008; Steiner et al., 2018). Furthermore, some researchers and development actors are now shifting to systemic approaches, associated with AKIS, which are more inclusive and recognize the roles of diverse actors in agricultural value chains, examine issues holistically and are more responsive to rapidly changing contexts, e.g. climate change, input price shocks (McIntire and Dober mann, 2003; Schiere et al., 2004; Van Rooyen et al., 2021). The systemic attributes of AKIS were conceptualized in the 2000s to tackle the problems of limited scalability and low adaptive capacity to rapidly changing contexts, e.g. climate change mitigation. Also, a shift from LISA to system innovations such as integrated nutrient manage ment (INM), integrated soil fertility management (ISFM) and conser vation agriculture (CA; Bationo et al., 2008) necessitated the exploration of systemic approaches in AIS. Therefore, ‘systems thinking’ is at the core of studies reviewed under this category. Methodological tools such as fuzzy cognitive mapping (Lalani et al., 2021) or combinations of Fig. 2. Flow diagram representing the systematic search approach I.S. Adolwa et al. Agricultural Systems 229 (2025) 104416 5 Table 3 A synthesis of papers reporting current agricultural innovation systems in Africa Innovation system paradigm Innovation or process in focus Main findings, including weaknesses and strengths Major reported outcomes Value form Author/s Transfer of technology Agricultural extension services; input subsidies; push-pull technology Studies report on processes that rely primarily on linear approaches where there is a lack of context within which farmers’ decisions on continued adoption or dis-adoption are made. Higher farm and labor productivity Financial Adesina et al. (2023); Malimi (2023); Morgan et al. (2020) Improved soybean varieties; fertilizer trees; agronomic practices Delineation of linear approaches that are largely driven by a single actor (mainly scientists) and constrained by the lifecycle of projects. An economic model, in one case, is used to justify top-down extension approaches. Increased yields and incomes; increased welfare gains Financial Ajayi et al. (2011); Mohammed and Abdulai (2022) Early farming systems research Fertilizers; pesticides; hybrid seeds Extension of the linear approach but with the added dimension of a whole- farm approach. Issues revolve around constraints to farmer adoption of technologies, economic returns to technology use, and farm and farmer characterization. Increased maize yield; improved food self-sufficiency and income Financial Branca et al. (2022); Freeman and Qin (2020); Pandey et al. (2001); Posner and Crawford (1992); Ronner et al. (2018); Vanlauwe et al. (2019) Organic agriculture, agroforestry, weed control Constraints to the adoption of low input sustainable agricultural technologies are examined at different scales e.g., household, farm Increased technology adoption Financial Bottazzi et al. (2023); Hildebrand et al. (1993); Posner & Crawford, (1991) Participatory research Mineral fertilizer, organic fertilizer (e.g. biochar, nutrified urine fertilizer) Participatory approaches ranging from functional/consultative (researcher- dominated) to participatory trial management, monitoring and evaluation (with some farmer involvement and agency). Participatory action research facilitated by semi-quantitative techniques, e.g. resource flow models where farmers are involved in data collection and analysis. Increased grain and biomass yields; increased nutrient uptake; more farmers establishing their own experiments; enhanced farmer learning Financial, Human, Natural Astatke et al. (2004); Defoer (1998); Eldon et al. (2020); Falconnier et al. (2016); Kamanga et al. (2010); Misiko et al. (2008); Odendo et al. (2006); Steiner et al. (2018); Wilde et al. (2022); Vanlauwe et al. (2017) Hybrid seeds A range of participatory methods are utilized to increase the efficacy of seed technologies and their uptake. These range from paired-plot participatory trials to mother-baby trials, farmer group experimentation, on-farm demonstrations, and farmer varietal evaluations and selections. However, the scalability of these approaches often came into question. Crowd sourcing techniques utilize digital tools to improve the data collection capacity of farmers, though this could simply be using farmers to harvest data. Increased adoption rates; higher yields Financial Amoako et al. (2023); Dibi et al. (2017); Etten et al. (2019); Harris et al. (2001); Kerr et al. (2007); Mekbib (1997); Mukanga et al. (2019); Mulatu and Belete (2001); Ssali et al. (2010); Waldman et al. (2014); Worku et al. (2020) Agroforestry, good agricultural practices Focus on system, complex technologies such as agroforestry, intercropping/ rotation systems, weed management and plant arrangement, which involve substantial internalization and adaptation processes. Participatory approaches incorporate farmer indigenous knowledge and semi- quantitative tools, e.g. board games. It emerged that whole-farm perspectives rather than individual technologies were of critical value to farmers. Impacts of joint experimentation or use of simple paired plots were assessed; however, the problems persist in how to scale the generated learnings and insights and how to determine ‘true’ farmer-centricity, given the considerable control researchers maintain. Narrowed yield gaps; integration of tacit knowledge with technical solutions; increased productivity Financial, Human Ajeigbe et al. (2010); Awio et al. (2022); Barrios et al. (2006); Franzel et al. (2001); Mafongoya and Kuntashula (2005); Senthilkumar et al. (2018); Tippe et al. (2017) Farmer field schools, Farmer business schools, co-designing, co- learning, systems thinking Participatory approaches advance to the point of re-designing farm systems, e.g. using DEED (Describe Explain Explore Design) and participatory experimentation. Farmers and Crop income increase; higher crop yields; moderate increase in cropping diversity; improvement of agricultural skills Financial, Human, Natural De Trincheria et al. (2015); Chilemba and Ragasa (2020); Douthwaite et al. (2009); Falconnier et al. (2017); Jaouadi et al. (2022); Husson et al. (continued on next page) I.S. Adolwa et al. Agricultural Systems 229 (2025) 104416 6 surveys, modelling and databases (Giller et al., 2011) further entrench a ‘systems thinking’ approach in examining farmer decision-making in complex environments. Such perspectives encompass co-learning among multi-stakeholders (Abate et al., 2011), re-design of cropping and farming systems, albeit with minimal farmer involvement (Bakker et al., 2021), demand-driven actor interactions (Clarkson et al., 2018), consideration of value chains (Horton et al., 2022), and application of process-based intervention strategies (Pircher et al., 2022). Typically, AR4D organizations consider the scaling components of their intervention strategies as beyond their purview. Partnerships with value-chain actors in multi-stakeholder arrangements may result in success, but be hardly scalable, as was the case with Low and Thiele (2020) in their orange-fleshed sweet potato scaling African program. Scaling efforts of many AKIS interventions have often been hampered by the short-term nature of donor-funded projects (Shilomboleni et al., 2023), or dependency on external actors rather than the more sustain able endogenous growth processes (e.g. Roques et al., 2022). There is also scant evidence for trans-disciplinarity and value chain approaches, key tenets of AKIS, based on the results of the review. The only studies to demonstrate transdisciplinary research approaches that traversed knowledge beyond academia into other domains were not directly linked to nutrient management (Restrepo et al., 2018; Restrepo et al., 2020). These are pointers to the impediments of AIS advancing along the systematic-systemic continuum, which could potentially be overcome through implementation of OFE processes. Table 3 (continued ) Innovation system paradigm Innovation or process in focus Main findings, including weaknesses and strengths Major reported outcomes Value form Author/s researchers are involved in co- designing innovations. Participatory prototyping trials help in fostering co- learning between scientists and farmers. Nevertheless, in many of the studies, farmers do not take a central role even when participatory elements are present. (2016); Kaaria et al. (2008); Hauser et al. (2016); Hounkonnou et al. (2006); Kraaijvanger et al. (2016); Kraaijvanger and Veldkamp (2017); Marinus et al. (2021); Périnelle et al. (2021); Traore et al. (2015); Traore et al. (2021);Van Mele et al. (2010); Van Vugt et al. (2017) Agricultural knowledge and innovation systems Seed technologies Resilience of seed systems is built through inclusive business models, networked scaling and institutionalization; however, since the push is exogenous, i.e. donor funded, impacts are often limited. Improved relations between actors; increased yield, improved adaptive capacity and competencies Financial, Human, Social Low and Thiele (2020); Shilomboleni et al. (2023) Sustainable intensification, system innovations Studies examine system innovations that integrate component technologies within a prism of innovation systems. Co-learning is crucial for buy-in of technologies. Others explore the complexity of farming systems by adopting the socio-ecological framework to innovation development. Enhanced understanding of market demands and opportunities Human Adolwa et al. (2017); Giller et al. (2011); Lalani et al. (2021); Ojiem et al. (2006); Schut et al. (2016); Styger et al. (2011); Horton et al. (2022) Innovation platforms, innovation processes, inclusive innovation, institutional change, value chains Innovation platforms is a widely used strategy for various agricultural projects and interventions. In a few reported cases there is an emphasis of transdisciplinary research where farmers co-develop research outputs with others resulting in co-created contextual innovation. The value farmers envisage out of the whole process marks out a unique feature of such transdisciplinary studies. One shortcoming of these studies is that being mostly qualitative in nature, they do not sufficiently account for exogenous factors when attributing impacts of the innovation process. Also, innovation platforms examined are too limited in scope to have an impact. Ultimately, for these processes to be successful, farmers must be active participants. Collaborative learning; increased useability of research outputs; skill development and capacity building; higher crop yields; increased generation of on-farm inputs from re-designed systems; increased crop earnings; improved inter-actor relationships and interactions. Financial, Human, Social, Natural Abate et al. (2011); Ahoa et al. (2020); Annette et al. (2023); Bakker et al. (2021); Bakker et al. (2022); Bertin et al. (2014); Bisseleua et al. (2018); Clarkson et al. (2018); Davies et al. (2018); Fofana et al. (2020); Horton et al. (2022); Iza and Dentoni (2020); Kingiri (2021); Kirina et al. (2022); Kuntosch and König (2018); Kwapong and Ankrah (2023); Magala et al. (2019); Makate (2019); Ochago et al. (2021); Opola et al. (2023); Pircher et al. (2022); Restrepo et al. (2020); Restrepo et al. (2018); Richardson-Ngwenya et al. (2019); Sanyang et al. (2016); Schut et al. (2018); Spielman et al. (2009); Totin et al. (2020) Fig. 3. Number of studies assessed by region and agricultural innova tion paradigm I.S. Adolwa et al. Agricultural Systems 229 (2025) 104416 7 4.3. Analysis of the efficacy from the implementation of the innovation systems 4.3.1. Value Creation A major goal of African AR4D organizations, particularly those that are biased towards top-down approaches, has been solely to increase crop productivity and incomes. The value creation, hence, is limited to financial capital gains, and is mainly assessed through biophysical and econometric modelling analysis on limited datasets (Ajayi et al., 2011; Mohammed and Abdulai, 2022; Tufa et al., 2019). However, to assess how effectively innovation systems are implemented there is need to consider aspects of the saliency, credibility and legitimacy of the inno vation process (Cash and Belloy, 2020). A trend analysis of global cereal data over the last sixty years suggests that cereal yields in Africa, especially for maize and rice, have increased at a slower pace compared with global trends (Fig. 5). Clearly, the focus on financial value alone has not been as successful in Africa as it has been in other continents and regions with highly controlled production systems (Schiere et al., 2004). Participatory approaches, however, had the effect of building addi tional human capital for farmers in various AIS. This was achieved by building the human capital of farmers (e.g. increased confidence) and scientists (e.g. application of new methods). With the participatory paradigm, farmers were part of the research process (Kraaijvanger et al., 2016; Kraaijvanger and Veldkamp, 2017; Mekbib, 1997; Tippe et al., 2017). In Burkina Faso and Mali, for instance, farmers were introduced to new knowledge on legume diversification and scope was provided for knowledge sharing and building on local knowledge (Kerr et al., 2007; Périnelle et al., 2021). Nevertheless, their efforts were continually (re-) directed towards improving the scientific process and its outputs rather than entrenching co-learning processes, hence constraining the learning, insights and behavioral development needed to drive change. In the literature we reviewed, co-learning is the value-creation process most frequently mentioned in AKIS, participatory and OFE initiatives (Fig. 6). There are also cases where communication broke down between stakeholders because power imbalances hindered farmers’ thinking and involvement in the research process (Jaouadi et al., 2022). Ultimately, the major undoing of all the examined participatory approaches was the limited scalability of the generated learnings, insights, and behavioral changes. Scaling as a process for value creation had minimal bearing for the AIS engaged in nutrient management (Fig. 6). The potential power of data analytics to inform scaling approaches of improved processes and insights (e.g. from farming communities to agribusinesses) is not Fig. 4. A representation of OFE pilots in Kenya and Cote d’Ivoire where co-learning, and nutrient management innovation co-development and co-design has been taking place Fig. 5. Yield trends of major cereals from 1961-2022. Source: https://www.fao.org/faostat/en/#data/QCL I.S. Adolwa et al. Agricultural Systems 229 (2025) 104416 8 https://www.fao.org/faostat/en/#data/QCL harnessed. There is also no compelling evidence of hybridization of scientific and local knowledge where co-learning has occurred, nor is there evidence of a shift in mindsets of scientists and farmers. Using an AKIS approach, Lalani et al. (2021) in Mozambique demonstrated the build-up of natural capital, i.e., soil quality improve ment through best nutrient management practices, with ramifications on climate change adaptation. Other AKIS approaches explored the role of social capital in improving farmers’ ability to invest in technologies (Adolwa et al., 2017; Kwapong and Ankrah, 2023). However, aspects of institutional value creation, i.e., improved farmer negotiation with fi nanciers, even when addressed were not explicitly described, e.g. the case of Horton et al. (2022) who sought to enhance the understanding of market demands. Nevertheless, the AKIS are better positioned than other paradigms in capturing diverse forms of value, as informed by the capitals approach. What remains uncertain is whether AKIS has built up the institutional capital of actors, particularly farmers, required to exert influence on technological processes. Judging by the small size of the nodes and lack of embeddedness in the network (Fig. 6), adaptive change, behavioral change and institutional change as change outcomes of inno vation systems geared towards nutrient management are not commonly reported implying a lack of a thorough introspection of these aspects in the current literature on African innovation systems. This seems to suggest that whilst we may criticize over-emphasis on value creation for financial capital, the components of institutional capital and to a lesser extent human, social and natural capitals, are insufficiently understood in the context of AIS. How to set up these innovation systems for diverse value creation to be exploited by farmers and other actors for better outcomes is an issue of concern. 4.3.2. Enablers From the network analysis, we find that co-learning processes are specialist-enabled (i.e., enabled by ‘specialists’) and are evidence-driven (informed and enabled by data), which results in co-created knowledge hybridization (Bramley et al., 2022; Bullock et al., 2019). Co-learning is also enabled when a farmer-centric approach is implemented (Fig. 6). For instance, Bertin et al. (2014) reported how new technologies and trainings on agroforestry shared with Cameroonian farmers through a co-learning process and evidence-based demos successfully imparted new skills in them. The buy-in from farmers was ultimately achieved because they were convinced the resulting knowledge was relevant to solving their problems. Maize yield improvement on farmer experi mental units corresponded with continuous co-learning between re searchers, farmers and other actors initiated under the African OFE initiative (Fig. 7). All the enablers to co-learning will determine the level of saliency, credibility and legitimacy in any given intervention, but the farmer-centric approach is the most variable of the enablers. We found that the extent of farmer-centricity varied with the innovation system paradigm. Farmer-centricity as an enabler tended to be more visible in AKIS and participatory approaches. Even so, participatory approaches ranged from those that were researcher-directed (Falconnier et al., 2016; Steiner et al., 2018), to those where farmers were involved in participatory trial management designing and conducting their own experiments (Misiko et al., 2008), and to action research using a range of semi-quantitative tools, including resource flow models (Defoer, 1998) and board games (Mafongoya and Kuntashula, 2005). Some of this work progressed to the extent where, for example, scientists and farmers in southern Mali were engaged in co-designing innovations using process- based frameworks such as DEED (Describe Explain Explore Design) and participatory experimentation (Falconnier et al., 2017). Fig. 6. A network graph showing the number of times principal terms were mentioned or inferred (denoted by size of circle/node) and the number of times the terms were mentioned (or referred to) together (denoted by thickness of the connections). Principle terms categorized as either value creation processes (blue), enablers (green) or change outcomes (red). Fig. 7. Trend of maize yields across seasons in the optimized treatment (OT) and farmer practice (FP) plots under an OFE initiative in Kenya and Cote d’Ivoire. Bands represent co- efficient of variation. (Source: Authors) I.S. Adolwa et al. Agricultural Systems 229 (2025) 104416 9 Enablers to the decentralization of innovation that can rectify power imbalances such as trans-disciplinarity are rarely actualized in AIS that center their activities around nutrient management (Fig. 6), given the asymmetrical power arrangements between actors. It is only in the Ethiopian innovation systems where such an attempt has been made through skill development and capacity building among NARES staff, farmers and other actors (Abate et al., 2011). But even in such a case, there was an absence of data-driven enablers to furnish the evidence base that can attribute impacts generated to innovation systems improvement. 5. Discussion 5.1. Using OFE to develop a farmer-centric knowledge discovery within existing AIS The results highlight the progressive evolution of AIS in tandem with the shift from systematic to systemic thinking among AR4D organiza tions and actors. In many ways the OFE paradigm builds on AKIS thinking as well as the participatory, FSR and linear (ToT) approaches. Several aspects of OFE, including farmer experimentation, are not new. Experimentation, a process of discovery and hypothesis testing, is the domain of both scientists and farmers (Akimowicz et al., 2022; Cook et al., 2013). A key question, therefore, hinges on how farmer-centric knowledge discovery could be developed within existing innovation systems. A major pillar of OFE is to enable and support co-learning be tween stakeholders (farmers, scientists, extension workers and others) that is scalable through community networks to share insights (Sumane et al., 2018; Thompson et al., 2019), or data analytics to inform formal research processes (Laurent et al., 2019). The attributes of farmer- centric OFE outlined earlier e.g. experimentation at scale, data-driven approaches etc., align with the ways farmers learn. Our literature syn thesis shows that the co-learning concept is fairly well entrenched in the relevant literature (Table 3; Fig. 6). The results also point to how cred ibility and legitimacy are entrenched through co-learning, which is in turn enabled when the process is farmer-centric and evidence-driven. A lingering question, however, is on mechanisms supporting the scaling of processes that underpin learning. Limited attention is given to empow ering farmers to share their data, ideas, and insights with the wider community. These concerns have been addressed to some extent through a few African OFE initiatives. Adolwa et al. (2022) helped to reinforce and accelerate learning on best nutrient management practices among Kenyan and Cote d’Ivoirian farmers by adopting a mixed methodology that integrated HSM, SSM and CSM. This approach entailed analysis of agronomic (crop cuts) and socio-economic data complemented by en gagements with farmers and digital (spectral data) solutions. Fairly large experimental units (1 ha) with simple designs in line with the ‘real systems’ concept of farmer-centric OFE (Lacoste et al., 2022) were adopted. Hence, the value of OFE was unpacked by collecting agronomic information to alleviate metrical uncertainty, spectral data to address variability in measurement, and socio-economic data to remove trans lational uncertainty in explaining results that were not clear (Cook et al., 2018). For instance, spectral information was converted to yield maps that were in turn shared with farmers to garner their insights on why there was high variability in yield response on their farms or within the farmer practice (FP). Some of the reasons cited for this variability included tree shading, management aspects e.g. differential manure application, and edaphic factors (Adolwa et al., 2022). In West Africa, Alexandre et al. (2023) focused on the co-development of a decision- support system to help farmers estimate, analyze and share data on mango yields. Thus, these OFE examples report on accumulated changes across several domains that individually may not be spectacular but collectively realize a change significant enough to address the limita tions of AIS. 5.2. Potential role of OFE to increase performance of current agricultural innovation systems The performance of innovation systems can be viewed from the prism of dynamic processes of actor interactions in complex adaptive systems (Klerkx et al., 2012), or from functionality in terms of enhanced information and knowledge exchange, learning, technological produc tivity, economic efficiency, and entrepreneurial activities (Hekkert et al., 2007; Kraaijvanger et al., 2016). The value creation framework proceeds to unpack this functionality in terms of the value generated (Malherbe et al., 2024) to implement functional, personal, social and supporting jobs (Cook et al., 2024). Functional jobs contribute to the building of financial and/or physical capital. Personal jobs contribute to the build-up of human and natural capitals. Social jobs contribute to the social capital build-up. Supporting jobs contribute to the build-up of institutional capital. The studies reviewed suggest that innovation sys tems have performed well, for instance by imparting financial benefits through increased crop yields and incomes (Mohammed and Abdulai, 2022; Odendo et al., 2006; Posner and Crawford, 1992; Steiner et al., 2018; Tufa et al., 2019). However, the emergence of ‘food systems agronomy’ more in line with ‘systems thinking’ calls for consideration of value creation over the entire food system (Conti et al., 2025). Thus, there is need to move beyond a sole focus on yield improvement in Af rica. Processes such as OFE that apply business model development to support crop nutrition agronomy, have a role in addressing the weak points of the scaling components in AIS. Although human capital has been built in select AIS (Kraaijvanger et al., 2016; Low and Thiele, 2020), there is a need to further entrench co-learning processes to bolster positive nutrient management outcomes. The interventions report on the build-up of natural capital through improvement of soil organic matter (Eldon et al., 2020), and increased nutrient uptake effi ciency (Burke et al., 2022; Wilde et al., 2022). The scope for natural capital through, for instance, build-up of soil carbon stocks is wider than suggested in the AIS literature. The 4R nutrient stewardship practices (Fixen, 2020) complemented with ISFM/CA are viable strategies for building natural capital on farms, which through carbon farming (Schilling et al., 2023) or payment for ecosystem services schemes (Engel and Muller, 2016) can be leveraged for the benefit of African farmers. As our results show, it will be difficult to achieve desired out comes in the absence of viable OFE networks that are sufficiently incorporated with value chain approaches. In assessing innovation systems, it is important not to focus on iso lated performance metrics at the expense of the legitimacy or credibility of the process. Akimowicz et al. (2022) corroborate this assertion where they report a lack of legitimacy in the French varietal innovation system despite high adoption rates. Innovations actors prioritize information uncertainties variably due to their diverse interests and objectives. For instance, agronomic and soil fertility indicators may be of interest to scientists to help them in tackling metric uncertainty, but a farmer may be more concerned with translational issues such as the price and availability of quality fertilizer inputs (Cook et al., 2013). There are scarcely any studies that report on institutional benefits, i. e. increased capacity for entrepreneurial activities, and the build-up of physical capital as a pointer towards increased digitalization in agri culture. This suggests that current innovation systems have failed to sufficiently demonstrate these forms of value to end users. Thus, a pertinent question would be on the value an OFE process creates for different actors in the AKIS, particularly smallholder farmers. Addi tionally, what is the potential role of OFE to increase performance of current agricultural innovation systems? Addressing these questions calls for adopting a nuanced approach, which systemic approaches like OFE are well attuned to. Reflexive methodologies, e.g. knowledge mapping and value proposition canvas tools, can be applied to assess the value of an innovation system and associated processes. Using such techniques, Adolwa et al. (2022) found that the OFE process met some of the value propositions of smallholder farmers. Farmers were able to I.S. Adolwa et al. Agricultural Systems 229 (2025) 104416 10 perform functional jobs that led to high yields and income, deriving financial value from OFE. Human capital benefits were also realized through learning and internalization of agronomic concepts (Fig. 4). Farmers adopted some of the 4R Nutrient Stewardship principles (Fixen, 2020), entailing the use of the right fertilizer source applied, at the right time, right rate, and right place. Natural capital was enhanced through their re-design of farming systems using soil and water conservation technologies to help conserve soil moisture for resilience to climate change impacts. An increase in knowledge and its application translated into improved livelihood outcomes, e.g. availability of funds to educate children, and ultimately an improvement in social standing. It was observed that farmers valued change processes that are holistic and complementary to agronomic knowledge. There are several other successful OFE initiatives around the globe including: a pilot farmer-based research program in the Falklands to assess the utility of digital soil and vegetation maps in an engagement process (McNee and Lacoste, 2021); an on-farm research network in the United Kingdom to compare fungicide programs (Roques et al., 2022); and an on-farm research network in the United States that conducts agronomic experiments under local conditions (Laurent et al., 2019). Several factors, e.g. large landholdings, access to finances, institutional support, highly developed human capital etc., have seemingly favored the success of OFE. Robust and fruitful farmer-centric OFE imple mentation in Africa will have to contend with challenges and opportu nities that are unique to the African agricultural production environment. Key barriers for successful OFE implementation include rampant fragmentation of landholdings, low income, high spatial and temporal variability within and across landscapes, heterogeneity in socioeconomic conditions, and low extension-to-farmer ratios (Dawson et al., 2016; Schut et al., 2020; Vanlauwe et al., 2017). Contra wise, OFE provides the means to, for instance, tackle inherent variability through the deployment of contextual and data-driven solutions. OFE could fill the gap occasioned by a low extension-to-farmer ratio, hence limited interactions with extension, by facilitating demand-driven and farmer- led processes. Against a backdrop of the Africa Fertilizer and Soil Health (AFSH) action plan aimed at building sustainable soil health and increasing access to context-relevant inputs (African Union Develop ment Agency- NEPAD, 2024), increased farmer agency enabled by OFE will be a key driver for success. Therefore, OFE has the potential to be a major enabler to the performance of AR4D, IPs and AKIS (Fig. 8). Given the stated intentions of IPs, for instance, to identify problems, devise solutions, implement the solutions, and evaluate the cycle, OFE is best placed to offer the approaches and tools to enable such processes if well operationalized. Key facilitators for farmer-centric OFE include: 1) the provision of an enabling policy environment and favorable institutional arrangements, including market access, that spur entrepreneurship; 2) recognition of farmers’ aspirations, objectives, motivations and strate gies for coping with ‘real’ world problems, e.g. input price shocks; and 3) access to networks and platforms that enhance continuous learning and foster contextualized and actionable agronomy (Burke et al., 2022; Glover et al., 2019; Van Rooyen et al., 2021). 5.3. Adapting current agricultural innovation systems to support or deploy OFE Farmer-centric OFE can be implemented as a stand-alone project or Fig. 8. Framework for an OFE-driven innovation process embedded within existing innovation systems. Source: Adolwa et al., 2022 I.S. Adolwa et al. Agricultural Systems 229 (2025) 104416 11 as a complement to AKIS, IPs and other multi-stakeholder partnerships (Fig. 8). For the latter case, the innovation systems need to be well adapted to benefit from OFE’s principles as espoused by Lacoste et al. (2022). Indeed, Conti et al. (2024) envisage a scenario where agri-food system transformation occurs through multiple innovation pathways engendering both bottom-up (OFE) and top-down approaches (e.g. multistakeholder arrangements led by IARC or NARES). For these innovation systems to deploy OFE, they should have built- in agility that allows a re-configuration or re-design of technological processes based on field-based evidence (Glover et al., 2019). Post- harvest dialogue platforms used by a recent initiative to deploy OFE provided the opportunity to link OFE processes with the wider AKIS (Adolwa et al., 2022). External experts learned from farmer experts on some of the emerging, contextual issues that were affecting maize pro duction in northern Cote d’Ivoire and Kenya, e.g. Spodoptera frugiperda (Fall armyworm) infestation, inefficiencies in the fertilizer supply chain, soil fertility constraints, and fluctuating weather patterns among other issues. Consequently, these CITs adjusted the treatment options of the researcher-led trials to address some of these issues. For instance, the agronomic protocol for Kenya CIT was revised to include organic amendments needed for water and nutrient retention under erratic rainfall conditions, and the yield targets were all revised downwards (Adolwa et al., 2022). Therefore, a careful balance between farmer- centred and experimental-centred (that accounts for complexity and uncertainty) approaches may be required for a successful transformative process for food systems (Conti et al., 2024; Conti et al., 2025). A successful implementation of any innovation process requires a careful consideration of practical, capacity and cost implications. Past (and current) innovation paradigms always required a huge injection of resources to get results. Case in point, is the training-and-visit (T&V) extension program funded by the World Bank, conceived to transfer fertilizer and seed technologies to African smallholder farmers (Anderson et al., 2006). At the cessation of funding that coincided with onset of structural adjustment program conditionalities, the T&V pro grams were discontinued to the detriment of smallholders (Dawson et al., 2016). The participatory and AKIS initiatives have also encoun tered the same fate, with efforts halting as soon as funding ceased (Shilomboleni et al., 2023). Given that OFE promotes an endogenous growth process supported by viable business models (Cook et al., 2018), it is envisaged that constraints related to funding, capacity and practi cality will be minimized. This is because the end users, interacting with innovations within their context and having realized different forms of value from OFE, will inject their own resources (e.g. time, capital, knowledge) to sustain the benefits. Similarly, other actors, e.g. Agri preneuers, will contribute to building and sustaining the innovation process given that they are also obtaining value from it. The budding OFE initiatives in Africa, if built upon, can help reduce initial costs of setting up subsequent OFE processes. 6. Conclusion This study employs a systematic review approach to track the evo lution of African innovation systems and to contextualize the novel OFE paradigm. Current innovation systems, particularly the later-stage ones, such as participatory research and AKIS, have been instrumental in engendering co-learning processes thereby enabling partial adoption of improved agronomic practices. This has in turn led to an improvement of soil organic matter, increased nutrient uptake efficiency, and higher crop productivity, albeit to a limited extent. The shortcomings of linear processes are widely accepted given their top-down approaches that focus exclusively on crop productivity increase without due consider ation to prevailing socio-organizational conditions and production en vironments. Participatory and AKIS processes address some of these concerns; however, they fall short in terms of generating scalable learning or solutions, demonstrating sufficient institutional value to end- users, and re-designing of technological processes based on the available evidence. In the few cases where re-design is done, farmers are often not central to the process. The implications are low understanding and ownership of developed technologies leading to low technology adop tion. This has led to continued low productivity, minimal livelihood improvement and high environmental degradation, even though tech nologies exist to address them. Farmer-centric on-farm experimentation builds on participatory research and the systemic approaches of AKIS to potentially deliver an approach that will address these challenges and gaps. The integration of tools, methodologies and concepts drawn from agronomy, digital sci ences and social sciences allows OFE to unpack farmer-relevant insights from the field to support innovation development and the decision- making of a range of actors, including smallholder farmers, policy makers, fertilizer and seed market players. By aligning innovation sys tems with experimental and decision-making processes of farmers, we do not anticipate a reversal of gains made as has often been the case with AKIS and participatory approaches. Therefore, OFE can be a powerful enabler to the performance of current innovation systems by unlocking context-specific and smallholder farmer-relevant nutrient management solutions. The highlighted African OFE examples only provide early insights into making OFE operational in a smallholder context. Thus, it is envisioned that more concrete insights will be generated as more OFE pilots are prototyped on the continent. This will determine whether innovation systems need to be built from the ground up through an OFE process or be adjusted to be responsive to emerging demands, chal lenges, and opportunities. In any case, on-farm experimentation is an innovation process that is uniquely placed to drive the transformation of African agriculture through sustainable food production by smallholder farmers. CRediT authorship contribution statement Ivan S. Adolwa: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptu alization. Shamie Zingore: Writing – review & editing, Supervision, Conceptualization. James Mutegi: Writing – review & editing. Matthew McNee: Writing – review & editing, Methodology, Concep tualization. Bolaji A. Akorede: Data curation. Denver Masidza: Formal analysis, Data curation. T. Scott Murrell: Methodology, Writing – re view & editing. Samuel Mathu Ndungu: Methodology. Eileen Nchanji: Writing – review & editing. Simon Cook: Conceptualization. Thomas Oberthür: Writing – review & editing, Supervision, Method ology, Conceptualization. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgement We sincerely thank Dr. Cai Chao (Purdue University) for his assis tance with providing systematic review tools and helping with the literature search and screening process. We also thank the anonymous reviewers for their comments and suggestions that considerably improved the manuscript. Data availability Data will be made available on request. References Abate, T., Shiferaw, B., Gebeyehu, S., Amsalu, B., Negash, K., Assefa, K., Eshete, M., Aliye, S., Hagmann, J., 2011. A systems and partnership approach to agricultural I.S. Adolwa et al. Agricultural Systems 229 (2025) 104416 12 research for development Lessons from Ethiopia. Outlook Agric. 40 (3), 213–220. https://doi.org/10.5367/oa.2011.0048. Adesina, O.S., Whitfield, S., Sallu, S.M., Sait, S.M., Pittchar, J., 2023. Bridging the gap in agricultural innovation research: A systematic review of push-pull biocontrol technology in sub-Saharan Africa. Int. J. Agric. Sustain. 21 (1). https://doi.org/ 10.1080/14735903.2023.2232696. Adolwa, I.S., Schwarze, S., Bellwood-Howard, I., Schareika, N., Buerkert, A., 2017. A comparative analysis of agricultural knowledge and innovation systems in Kenya and Ghana: sustainable agricultural intensification in the rural–urban interface. Agric. Hum. Values. 34 (2), 453–472. https://doi.org/10.1007/s10460-016-9725-0. Adolwa, I.S., Cook, S., Oberthür, T., Phillips, S., Agneroh, T., Amouzou, K., 2022. Post- harvest assessments of on-farm maize experimentation provide key checkpoints for farmers and stakeholders. Growing Africa 1 (2), 10–14. https://doi.org/10.55693/ ga12.UATZ3489. African Union Development Agency- NEPAD, 2024. AICCRA Brief. Accelerating Impacts of CGIAR Climate Research for Africa (AICCRA). Accessed 24.02.25. https://hdl. handle.net/10568/169170. Ahoa, E., Kassahun, A., Tekinerdogan, B., 2020. Business processes and information systems in the Ghana cocoa supply chain: a survey study. NJAS - Wagening. J. Life Sci. 92. https://doi.org/10.1016/j.njas.2020.100323. Ajayi, O.C., Place, F., Akinnifesi, F.K., Sileshi, G.W., 2011. Agricultural success from Africa: the case of fertilizer tree systems in southern Africa (Malawi, Tanzania, Mozambique, Zambia and Zimbabwe). Int. J. Agric. Sustain. 9 (1), 129–136. https:// doi.org/10.3763/ijas.2010.0554. Ajeigbe, H.A., Singh, B.B., Adeosun, J.O., Ezeaku, I.E., 2010. Participatory on-farm evaluation of improved legumecereals cropping systems for crop-livestock farmers: Maize-double cowpea in Northern Guinea Savanna Zone of Nigeria. Afr. J. Agric. Res. 5 (16), 2080–2088. Akimowicz, M., Del Corso, J.-P., Gallai, N., Kephaliacos, C., 2022. The leader, the keeper, and the follower? A legitimacy perspective on the governance of varietal innovation systems for climate changes adaptation. The case of sunflower hybrids in France. Agric. Syst. 203, 103498. https://doi.org/10.1016/j.agsy.2022.103498. Alexandre, C., Tresch, L., Sarron, J., Lavarenne, J., Bringer, G., Chaham, H.R., Bendahou, H., Carmeni, S., Borianne, P., Koffi, J.M., Faye, E., 2023. Creating shared value(s) from On-Farm Experimentation: Ten key lessons learned from the development of the SoYield® digital solution in Africa. Agron. Sustain. Dev. 43 (3). https://doi.org/10.1007/s13593-023-00888-7. Amoako, O.A., Oteng-Frimpong, R., Yirzagla, J., Kassim, Y.B., Tengey, T.K., Adogoba, D. S., Mingle, M., Alhassan, R., Ibrahim, A.A., 2023. Participatory on-farm evaluation of improved groundnut genotypes in the Guinea Savannah agro-ecological zone of Ghana. Agriculture-Basel 13 (12). https://doi.org/10.3390/agriculture13122249. Anderson, J.R., Feder, G., Ganguly, S., 2006. The rise and fall of training and visit extension: an Asian mini-drama with an African epilogue (World Bank Policy Research Working Paper 3928). Annette, N.M., Makeda, T., Mukani, M., Tawanda, M., 2023. Tailoring business models for small-medium food enterprises in Eastern Africa can drive the commercialization and utilization of vitamin A rich orange-fleshed sweet potato puree. Open Agric. 8 (1). https://doi.org/10.1515/opag-2022-0168. Astatke, A., Mamo, T., Peden, D., Diedhiou, M., 2004. Participatory on-farm conservation tillage trial in the Ethiopian highland vertisols: the impact of potassium application on crop yields. Exp. Agric. 40 (3), 369–379. https://doi.org/10.1017/ S0014479704002029. Awio, T., Struik, P.C., Stomph, T.J., 2022. Management practices and rice grain yield of farmers after participation in a joint experimentation. Front. Sustain. Food Syst. 6. https://doi.org/10.3389/fsufs.2022.1009469. Bakker, T., Dugué, P., de Tourdonnet, S., 2021. Assessing the effects of Farmer Field Schools on farmers’ trajectories of change in practices. Agron. Sustain. Dev. 41 (2). https://doi.org/10.1007/s13593-021-00667-2. Bakker, T., Dugue, P., Tourdonnet, S., 2022. How do farmers change their practices at the farm level after co-design processes in Farmer Field Schools? Agric. Syst. 201. https://doi.org/10.1016/j.agsy.2022.103457. Barrios, E., Delve, R.J., Bekunda, A., Mowo, J., Agunda, J., Ramisch, J., Trejo, M.T., Thomas, R.J., 2006. Indicators of soil quality: A South-South development of a methodological guide for linking local and technical knowledge. Geoderma 135, 248–259. https://doi.org/10.1016/j.geoderma.2005.12.007. Bationo, A., Waswa, B., 2011. New challenges and opportunities for Integrated Soil Fertility Management in Africa. In: Innovations as key to the Green Revolution in Africa, 1. Springer, pp. 3–18. Bationo, A., Kihara, J., Vanlauwe, B., Kimetu, J., Waswa, B.S., Sahrawat, K.L., 2008. Integrated nutrient management: concepts and experience from Sub-Saharan Africa. In: Integrated Nutrient Management for Sustainable Crop Production. Taylor & Francis Group, pp. 467–521. Bertin, T., Zacharie, T., Ann, D., Ebenezar, A., Alain, T., 2014. Scaling-up sustainable land management practices through the concept of the rural resource centre: reconciling farmers’ interests with research agendas. J. Agric. Educ. Ext. 20 (5), 463–483. https://doi.org/10.1080/1389224X.2014.913984. Bisseleua, D.H.B., Idrissou, L., Olurotimi, P., Ogunniyi, A., Mignouna, D., Bamire, S.A., 2018. Multi-stakeholder process strengthens agricultural innovations and sustainable livelihoods of farmers in Southern Nigeria. J. Agric. Educ. Ext. 24 (1), 29–49. https://doi.org/10.1080/1389224X.2017.1392992. Borgatti, S.P., Everett, M.G., Freeman, L.C., 2002. Ucinet 6 for Windows. Software for social network analysis. Analytic Technologies, Harvard. Borgatti, S.P., Mehra, A., Brass, D.J., Labianca, G., 2009. Network analysis in the social sciences. Science 323 (5916), 892–895. Bottazzi, P., Seck, S.M., Niang, M., Moser, S., 2023. Beyond motivations: a framework unraveling the systemic barriers to organic farming adoption in northern Senegal. J. Rural Stud. 104. https://doi.org/10.1016/j.jrurstud.2023.103158. Bramley, R.G.V., Song, X.X., Colaço, A.F., Evans, K.J., Cook, S.E., 2022. Did someone say “farmer-centric”? Digital tools for spatially distributed on-farm experimentation. Agron. Sustain. Dev. 42 (6). https://doi.org/10.1007/s13593-022-00836-x. Branca, G., Cacchiarelli, L., Haug, R., Sorrentino, A., 2022. Promoting sustainable change of smallholders’ agriculture in Africa: policy and institutional implications from a socio-economic cross-country comparative analysis. J. Clean. Prod. 358. https://doi. org/10.1016/j.jclepro.2022.131949. Bullock, D.S., Boerngen, M., Tao, H.Y., Maxwell, B., Luck, J.D., Shiratsuchi, L., Puntel, L., Martin, N.F., 2019. The data-intensive farm management project: changing agronomic research through on-farm precision experimentation. Agron. J. 111 (6), 2736–2746. https://doi.org/10.2134/agronj2019.03.0165. Burke, W.J., Jayne, T.S., Snapp, S.S., 2022. Nitrogen efficiency by soil quality and management regimes on Malawi farms: can fertilizer use remain profitable? World Dev. 152, 105792. https://doi.org/10.1016/j.worlddev.2021.105792. Cash, D.W., Belloy, P.G., 2020. Salience, credibility and legitimacy in a rapidly shifting world of knowledge and action. Sustainability 12, 7376. https://doi.org/10.3390/ su12187376. Chambers, R., 1994. Participatory rural appraisal (PRA): Analysis of experience. World Dev. 22 (9), 1253–1268. https://doi.org/10.1016/0305-750X(94)90003-5. Chilemba, J., Ragasa, C., 2020. The impact on farmer incomes of a nationwide scaling up of the farmer business school program: lessons and insights from central Malawi. Eur. J. Dev. Res. 32 (4), 906–938. https://doi.org/10.1057/s41287-019-00246-y. Clarkson, G., Garforth, C., Dorward, P., Mose, G., Barahona, C., Areal, F., Dove, M., 2018. Can the TV makeover format of edutainment lead to widespread changes in farmer behaviour and influence innovation systems? Shamba Shape Up in Kenya. Land Use Policy. 76, 338–351. https://doi.org/10.1016/j.landusepol.2018.05.011. Conti, C., Hall, A., Percy, H., Stone-Jovicich, S., Turner, J., McMillan, L., 2024. What does the agri-food systems transformation agenda mean for agricultural research organisations? Exploring organisational prototypes for uncertain futures. Glob. Food Secur. 40 (100733). https://doi.org/10.1016/j.gfs.2023.100733. Conti, C., Hall, A., Moallemi, E.A., Laila, A., Bene, C., Fanzo, J., Gibson, M.F., Gordon, L., Hicks, C., Kok, K., Rao, N., Laxminarayan, R., Mason-D’Croz, D., 2025. Top-down vs bottom-up processes: A systematic review clarifying roles and patterns of interactions in food system transformation. Glob. Food Secur. 44 (100833). https:// doi.org/10.1016/j.gfs.2025.100833. Cook, S., Cock, J., Oberthür, T., Myles, F., 2013. On-farm experimentation. Better Crops 97 (4), 17–20. Cook, S., Lacoste, M., Evans, F., Ridout, M., Gibberd, M., Oberthür, T., 2018. An on-farm experimental philosophy for farmer-centric digital innovation. In: Proceedings of the 14th International Conference on Precision Agriculture. Cook, S., McNee, M., Adolwa, I., Lacoste, M., Longchamps, L., Oberthür, T., 2024. What is the value of OFE? Who gets it, and how is it created?. In: Proceedings of the 2nd International Conference on Farmer-centric On-FArm Experimentation ’Exploring Shared Value Propositions’- 5-7th December 2023 https://doi.org/10.7298/tjcv- ht23. Davies, J., Maru, Y., Hall, A., Abdourhamane, I.K., Adegbidi, A., Carberry, P., Dorai, K., Ennin, S.A., Etwire, P.M., McMillan, L., Njoya, A., Ouedraogo, S., Traoré, A., Traoré- Gué, N.J., Watson, I., 2018. Understanding innovation platform effectiveness through experiences from west and central Africa. Agric. Syst. 165, 321–334. https://doi.org/10.1016/j.agsy.2016.12.014. Dawson, N., Martin, A., Sikor, T., 2016. Green Revolution in Sub-Saharan Africa: Implications of Imposed Innovation for the Wellbeing of Rural Smallholders. World Dev. 78, 204–218. de Lara, A., Mieno, T., Luck, J.D., et al., 2023. Predicting site-specific economic optimal nitrogen rate using machine learning methods and on-farm precision experimentation. Precis. Agric. 24,, 1792–1812. https://doi.org/10.1007/s11119- 023-10018-8. De Trincheria, J., Craufurd, P., Harris, D., Mannke, F., Nyamangara, J., Rao, K.P.C., Leal, W., 2015. Adapting agriculture to climate change by developing promising strategies using analogue locations in Eastern and Southern Africa: A Systematic approach to develop practical solutions. In: Adapting African Agriculture to Climate Change: Transforming Rural Livelihoods. Defoer, T., 1998. Participatory action research and quantitative analysis for nutrient management in southern Mali: A fruitful marriage? Agric. Ecosyst. Environ. 71, 215–228. Denning, G., Kabambe, P., Sanchez, P., Malik, A., Flor, R., Harawa, P., Nkhoma, P., Zamba, C., Banda, C., Magombo, C., Keating, M., Wangila, J., Sachs, J., 2009. Input subsidies to improve smallholder maize productivity in Malawi: toward an African green revolution, 7 (1), e1000023. https://doi.org/10.1371/journal.pbio.1000023. Dfid., 1999. Key sheets for sustainable development: Overview. Department for International Development. Dibi, K.E.B., Essis, B.S., N’Zué, B., Kouakou, A.M., Zohouri, G.P., Assouan, A.B., Van Mourik, T., 2017. Participatory selection of orange-fleshed sweetpotato varieties in north and north-east Cote d’Ivoire. Open Agric. 2 (1), 83–90. https://doi.org/ 10.1515/opag-2017-0009. Dixon, J., Gulliver, A., Gibbon, D., Hall, M., 2001. Farming Systems and Poverty: Improving Farmers’ Livelihoods In A Changing World. Food and Agriculture Organization of the United Nations. http://www.fao.org/3/Y1860E/y1860e04. htm#P1_2. Douthwaite, B., Beaulieu, N., Lundy, M., Peters, D., 2009. Understanding how participatory approaches foster innovation. Int. J. Agric. Sustain. 7 (1), 42–60. https://doi.org/10.3763/ijas.2009.0339. I.S. Adolwa et al. Agricultural Systems 229 (2025) 104416 13 https://doi.org/10.5367/oa.2011.0048 https://doi.org/10.1080/14735903.2023.2232696 https://doi.org/10.1080/14735903.2023.2232696 https://doi.org/10.1007/s10460-016-9725-0 https://doi.org/10.55693/ga12.UATZ3489 https://doi.org/10.55693/ga12.UATZ3489 https://hdl.handle.net/10568/169170 https://hdl.handle.net/10568/169170 https://doi.org/10.1016/j.njas.2020.100323 https://doi.org/10.3763/ijas.2010.0554 https://doi.org/10.3763/ijas.2010.0554 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0040 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0040 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0040 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0040 https://doi.org/10.1016/j.agsy.2022.103498 https://doi.org/10.1007/s13593-023-00888-7 https://doi.org/10.3390/agriculture13122249 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0060 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0060 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0060 https://doi.org/10.1515/opag-2022-0168 https://doi.org/10.1017/S0014479704002029 https://doi.org/10.1017/S0014479704002029 https://doi.org/10.3389/fsufs.2022.1009469 https://doi.org/10.1007/s13593-021-00667-2 https://doi.org/10.1016/j.agsy.2022.103457 https://doi.org/10.1016/j.geoderma.2005.12.007 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0095 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0095 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0095 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0100 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0100 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0100 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0100 https://doi.org/10.1080/1389224X.2014.913984 https://doi.org/10.1080/1389224X.2017.1392992 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0115 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0115 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0120 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0120 https://doi.org/10.1016/j.jrurstud.2023.103158 https://doi.org/10.1007/s13593-022-00836-x https://doi.org/10.1016/j.jclepro.2022.131949 https://doi.org/10.1016/j.jclepro.2022.131949 https://doi.org/10.2134/agronj2019.03.0165 https://doi.org/10.1016/j.worlddev.2021.105792 https://doi.org/10.3390/su12187376 https://doi.org/10.3390/su12187376 https://doi.org/10.1016/0305-750X(94)90003-5 https://doi.org/10.1057/s41287-019-00246-y https://doi.org/10.1016/j.landusepol.2018.05.011 https://doi.org/10.1016/j.gfs.2023.100733 https://doi.org/10.1016/j.gfs.2025.100833 https://doi.org/10.1016/j.gfs.2025.100833 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0175 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0175 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0180 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0180 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0180 https://doi.org/10.7298/tjcv-ht23 https://doi.org/10.7298/tjcv-ht23 https://doi.org/10.1016/j.agsy.2016.12.014 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0195 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0195 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0195 https://doi.org/10.1007/s11119-023-10018-8 https://doi.org/10.1007/s11119-023-10018-8 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0200 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0200 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0200 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0200 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0200 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0205 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0205 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0205 https://doi.org/10.1371/journal.pbio.1000023 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0215 http://refhub.elsevier.com/S0308-521X(25)00156-8/rf0215 https://doi.org/10.1515/opag-2017-0009 https://doi.org/10.1515/opag-2017-0009 http://www.fao.org/3/Y1860E/y1860e04.htm#P1_2 http://www.fao.org/3/Y1860E/y1860e04.htm#P1_2 https://doi.org/10.3763/ijas.2009.0339 Eldon, J., Baird, G., Sidibeh, S., Dobasin, D., Rapaport, P., Cheng, W.X., Shennan, C., 2020. On-farm trials identify adaptive management options for rainfed agriculture in West Africa. Agric. Syst. 182. https://doi.org/10.1016/j.agsy.2020.102819. Engel, S., Muller, A., 2016. Payments for environmental services to promote ‘climate- smart agriculture’? Potential and challenges. Agric. Econ. 47 (S1), 173–184. https:// doi.org/10.1111/agec.12307. Etten, J.V., Beza, E., Calderer, L., Duijvendijk, K.V., Fadda, C., Fantahun, B., Kidane, Y. G., Gupta, A., Mengistu, D.K., Kiambi, D., Mathur, P.N., Mercado, L., Mittra, S., Mollel, M.J., Rosas, J.C., Steinke, J., Suchini, J.G., Zimmerer, K.S., 2019. First experiences with a novel farmer citizen science approach: crowdsourcing participatory variety selection through on-farm triadic comparisons of technologies (TRICOT). Exp. Agric. 55 (S1), 275–296. https://doi.org/10.1017/ S0014479716000739. Evenson, R.E., Gollin, D., 2003. Assessing the impact of the green revolution, 1960 to 2000. Science 300, 758. https://doi.org/10.1126/science.1078710. Eyers, A., Howarth, C., 2024. The role and effectiveness of climate commissions in engaging the public on climate change in the UK. Environ. Sci. Policy. 157, 103760. https://doi.org/10.1016/j.envsci.2024.103760. Falconnier, G.N., Descheemaeker, K., Van Mourik, Adam, M., Sogoba, B., Giller, K.E., 2017. Co-learning cycles to support the design of innovative farm systems in southern Mali. Eur. J. Agron. 89, 61–74. https://doi.org/10.1016/j.eja.2017.06.008. Falconnier, G.N., Descheemaeker, K., Van Mourik, T.A., Giller, K.E., 2016. Unravelling the causes of variability in crop yields and treatment responses for better tailoring of options for sustainable intensification in southern Mali. Field Crops Res. 187, 113–126. https://doi.org/10.1016/j.fcr.2015.12.015. Fixen, P.E., 2020. A brief account of the genesis of 4R nutrient stewardship. Agron. J. 1–8. https://doi.org/10.1002/agj2.20315. Fofana, B., Halos-Kim, L., Akeredolu, M., Okiror, A., Sima, K., Naibakelao, D., Oluoch, M., Iseki, F., 2020. Innovative agricultural extension value chain-based models for smallholder African farmers. Front. Agric. Sci. Eng. 7 (4), 418–426. https://doi.org/10.15302/J-FASE-2020358. Franzel, S., Coe, R., Cooper, P., Place, F., Scherr, S.J., 2001. Assessing the adoption potential of agroforestry practices in sub-Saharan Africa. Agric. Syst. 69 (1–2), 37–62. https://doi.org/10.1016/S0308-521X(01)00017-8. Freeman, K., Qin, H., 2020. The role of information and interaction processes in the adoption of agriculture inputs in Uganda. Agronomy-Basel 10 (2). https://doi.org/ 10.3390/agronomy10020202. Garrity, D., Dixon, J., Boffa, J.-M., 2012. Understanding African Farming Systems: Science and Policy Implications. Australian Centre for International Agricultural Research and Australian International Food Security Centre. Giller, K.E., Tittonell, P., Rufino, M.C., van Wijk, M.T., Zingore, S., Mapfumo, P., Adjei- Nsiah, S., Herrero, M., Chikowo, R., Corbeels, M., Rowe, E.C., Baijukya, F., Mwijage, A., Smith, J., Yeboah, E., van der Burg, W.J., Sanogo, O.M., Misiko, M., de Ridder, N., et al., 2011. Communicating complexity: Integrated assessment of trade- offs concerning soil fertility management within African farming systems to support innovation and development. Agric. Syst. 104 (2), 191–203. https://doi.org/ 10.1016/j.agsy.2010.07.002. Glover, D., Sumberg, J., Ton, G., Andersson, J., Badstue, L., 2019. Rethinking technological change in smallholder agriculture. Outlook Agric. 48 (3), 169–180. https://doi.org/10.1177/0030727019864978. Hall, A., Dijkman, J., 2019. Public agricultural research in an era of transformation: The challenge of agri-food system innovation. (2109; p. 67). In: CGIAR Independent Science and Partnership Council (ISPC) Secretariat and Commonwealth Scientific and Industrial Research Organisation (CSIRO). Harris, D., Pathan, A.K., Gothkar, P., Joshi, A., Chivasa, W., Nyamudeza, P., 2001. On- farm seed priming: Using participatory methods to revive and refine a key technology. Agric. Syst. 69 (1–2), 151–164. https://doi.org/10.1016/S0308-521X (01)00023-3. Hauser, M., Lindtner, M., Prehsler, S., Probst, L., 2016. Farmer participatory research: Why extension workers should understand and facilitate farmers’ role transitions. J. Rural Stud. 47, 52–61. https://doi.org/10.1016/j.jrurstud.2016.07.007. Hekkert, M., Suurs, R., Negro, S., Kuhlmann, S., Smits, R., 2007. Functions of innovation systems: A new approach for analysing technological change. Technol. Forecasting Social Change 74 (4), 413–432. https://doi.org/10.1016/j.techfore.2006.03.002. Henrich, J., Gil-White, F.J., 2001. The evolution of prestige: Freely conferred deference as a mechanism for enhancing the benefits of cultural transmission. Evolution and Human Behavior 22 (3), 165–196. https://doi.org/10.1016/S1090-5138(00)00071- 4. Hildebrand, P.E., Singh, B.K., Bellows, B.C., Campbell, E.P., Jama, B.A., 1993. Farming systems research for agroforestry extension. Agrofor. Syst. 23 (2–3), 219–237. https://doi.org/10.1007/BF00704917. Homann-Kee Tui, S., Adekunle, A., Lundy, M., Tucker, J., Birachi, E., Schut, M., Klerkx, L., Ballantyne, P., Duncan, A., Cadilhon, J., Mundy, P., Ombati, A., 2013. What are innovation platforms? Innovation platforms practice brief 1. http://oar. icrisat.org/id/eprint/7343. Horton, D., Devaux, A., Bernet, T., Mayanja, S., Ordinola, M., Thiele, G., 2022. Inclusive innovation in agricultural value chains: Lessons from use of a systems approach in diverse settings. Innov. Dev. https://doi.org/10.1080/2157930X.2022.2070587. Hounkonnou, D., Kossou, D.K., Kuyper, T.W., Leeuwis, C., Richards, P., Röling, N.G., Sakyi-Dawson, O., Van Huis, A., 2006. Convergence of sciences: The management of agricultural research for small-scale farmers in Benin and Ghana. NJAS - Wagening. J. Life Sci. 53 (3–4), 343–367. https://doi.org/10.1016/S1573-5214(06)80013-8. Husson, O., Quoc, H.T., Boulakia, S., Chabanne, A., Tivet, F., Bouzinac, S., Lienhard, P., Michellon, R., Chabierski, S., Boyer, J., Enjalric, F., Rakotondramanana Moussa, N., Jullien, F., Balarabe, O., Rattanatray, B., Castella, J.C., Charpentier, H., Séguy, L., 2016. Co-designing innovative cropping systems that match biophysical and socio- economic diversity: The DATE approach to Conservation Agriculture in Madagascar, Lao PDR and Cambodia. Renew. Agric. Food Syst. 31 (5), 452–470. https://doi.org/ 10.1017/S174217051500037X. Iza, C.L.B., Dentoni, D., 2020. How entrepreneurial orientation drives farmers’ innovation differential in Ugandan coffee multi-stakeholder platforms. J. Agribus. Dev. Emerg. Econ. 10 (5), 629–650. https://doi.org/10.1108/JADEE-01-2020-0007. Jaouadi, R., Lestrelin, G., Benaissa, N., Scopel, E., Ben Mimoun, M., Bouarfa, S., 2022. Assessing farmer field schools process to improve the participation of local actors: A case study in Tunisia. XXXI International horticultural congress, ihc2022: International symposium on agroecology and system approach for sustainable and resilient horticultural production. Kaaria, S., Njuki, J., Abenakyo, A., Delve, R., Sanginga, P., 2008. Assessment of the Enabling Rural Innovation (ERI) approach: Case studies from Malawi and Uganda. Nat. Resour. Forum. 32 (1), 53–63. https://doi.org/10.1111/j.1477- 8947.2008.00174.x. Kamanga, B.C.G., Whitbread, A., Wall, P., Waddington, S.R., Almekinders, C., Giller, K. E., 2010. Farmer evaluation of phosphorus fertilizer application to annual legumes in Chisepo, Central Malawi. Afr. J. Agric. Res. 5 (8), 668–680. Kerr, R.B., Snapp, S., Chirwa, M., Shumba, L., Msachi, R., 2007. Participatory research on legume diversification with malawian smallholder farmers for improved human nutrition and soil fertility. Exp. Agric. 43 (4), 437–453. https://doi.org/10.1017/ S0014479707005339. Kingiri, A.N., 2021. Agricultural advisory and extension service approaches and inclusion in reaching out to Kenyan rural farmers. AJSTID 13 (7), 797–806. https:// doi.org/10.1080/20421338.2020.1823098. Kirina, T., Groot, A., Shilomboleni, H., Ludwig, F., Demissie, T., 2022. Scaling climate smart agriculture in East Africa: Experiences and Lessons. Agronomy-Basel 12 (4). https://doi.org/10.3390/agronomy12040820. Klerkx, L., Mierlo, V.M., Leeuwis, C., 2012. Evolution of systems approaches to agricultural innovation: Concepts, analysis and interventions. In: Farming systems research into the 21st century: The new dynamic. Springer, pp. 457–483. Kraaijvanger, R.G., Veldkamp, T., 2017. Four years of farmer experimentation on soil fertility in Tigray, northern Ethiopia: Trends in research strategies. J. Agric. Educ. Ext. 23 (4), 373–391. https://doi.org/10.1080/1389224X.2017.1289962. Kraaijvanger, R., Veldkamp, T., Almekinders, C., 2016. Considering change: evaluating four years of participatory experimentation with farmers in Tigray (Ethiopia) highlighting both functional and human-social aspects. Agric. Syst. 147, 38–50. https://doi.org/10.1016/j.agsy.2016.05.001. Kuntosch, A., König, B., 2018. Linking system perspectives with user perspectives to identify adoption barriers to food security innovations for smallholder farmers—Evidence from rural Tanzania. Glob. Food Secur. 10 (4), 881–896. https:// doi.org/10.1007/s12571-018-0821-4. Kwapong, N.A., Ankrah, D.A., 2023. Understanding innovation process within an interactive social network: empirical insights from maize innovations in southern Ghana. Cogent Soc. Sci. ll9(1). https://doi.org/10.1080/23311886.2023.2167390. Lacoste, M., Cook, S., McNee, M., Gale, D., Ingram, J., Bellon-Maurel, V., MacMillan, T., Sylvester-Bradley, R., Kindred, D., Bramley, R., Tremblay, N., Longchamps, L., Thompson, L., Ruiz, J., García, F.O., Maxwell, B., Griffin, T., Oberthur, T., Huyghe, C., Hall, A., 2022. On-Farm Experimentation to transform global agriculture. Nat. Food. 3 (1), 11–18. https://doi.org/10.1038/s43016-021-00424-4. Lalani, B., Aminpour, P., Gray, S., Williams, M., Büchi, L., Haggar, J., Grabowski, P., Dambiro, J., 2021. Mapping farmer perceptions, Conservation Agriculture practices and on-farm measurements: The role of systems thinking in the process of adoption. Agric. Syst. 191. https://doi.org/10.1016/j.agsy.2021.103171. Laurent, A., Kyveryga, P., Makowski, D., Miguez, F., 2019. A Framework for visualization and analysis of agronomic field trials from on-farm research networks. Agron. J. 111 (6), 2712–2723. https://doi.org/10.2134/agronj2019.02.0135. Low, J.W., Thiele, G., 2020. Understanding innovation: the development and scaling of orange-fleshed sweetpotato in major African food systems. Agric. Syst. 179. https:// doi.org/10.1016/j.agsy.2019.102770. Mafongoya, P.L., Kuntashula, E., 2005. Participatory evaluation of Tephrosia species and provenances for soil fertility improvement and other uses using farmer criteria in eastern Zambia. Exp. Agric. 41 (1), 69–80. https://doi.org/10.1017/ S0014479704002339. Magala, D.B., Mangheni, M.N., Miiro, R.F., 2019. Actor social networks as knowledge sharing mechanisms in multi-stakeholder processes: a case of coffee innovation platforms of Uganda. J. Agric. Educ. Ext. 25 (4), 323–336. https://doi.org/10.1080/ 1389224X.2019.1629971. Makate, C., 2019. Local institutions and indigenous knowledge in adoption and scaling of climate-smart agricultural innovations among sub-Saharan smallholder farmers. Int. J. Clim. Change Strateg. Manag. 12 (2), 270–287. https://doi.org/10.1108/IJCCSM- 07-2018-0055. Malherbe, W., Biggs, R., Sitas, N., 2024. Comparing apples and pears: linking capitals and capacities to assess the resilience of commercial farming operations. Agric. Syst. 217 (103934). https://doi.org/10.1016/j.agsy.2024.103934. Malimi, K., 2023. Agricultural input subsidies, extension services, and farm labour productivity nexus: Evidence from maize farmers in Tanzania. Am. J. Agric. Econ. 74 (3), 874–898. https://doi.org/10.1111/1477-9552.12537. Marinus, W., Descheemaeker, K., van de Ven, G., Waswa, W., Mukalama, J., Vanlauwe, B., Giller, K.E., 2021. “That is my farm” – An integrated co-learning approach