From Pilot to Practice: Why Agricultural Innovations Struggle to Scale

By Aditya K S[1], Bhuvana N[2] and Ravi Nandi[3]

The phrase ‘scaling for impact’ has become a common term in agricultural development. While many scientifically sound technologies demonstrate promise in trials, few are widely adopted to make a real-world impact. The blame often falls on extension systems, but this is an oversimplification of the issue. In this blog, we argue that successful scaling requires rethinking how technologies are piloted, framed, and supported, recognizing the broader network of stakeholders and incentives involved.

In his book ‘The Voltage Effect’, John List explores the idea of why some technologies scale, while many others fail to do so (even when most pilots report positive impacts). The book identifies five key pitfalls that prevent successful scaling, including false positives, overestimating growth, unique circumstances, unintended consequences, and excessive costs. It clearly highlights that ‘success in piloting’ is not a sufficient condition to determine whether the technology or idea has the potential to scale. So, the first point to understand is that ‘not every technology or an idea’ is suitable to be scaled. Even when technology developers highlight the benefits of their technologies, compared to checks, those benefits may not correspond to what farmers value or want. The technology might produce some negative externalities or require resources or context-specific considerations that make it not worth scaling. Under such situations, social scientists will have a major role to assess and identify technologies that can be scaled for a specific context and agro-climatic zone.

Technology or idea piloting: One crucial strategy for determining which ideas or technologies can potentially scale is to implement ‘piloting’. Ideally, pilots are supposed to test the effectiveness of the intervention in a real-world setting. However, that is rarely the case. ‘India is a burial ground for pilots’ is a famous anecdote among developmental researchers working in India, reflecting a large number of ‘successful’ pilots who often fail to scale. The problem is, how the pilots are implemented rather than testing the feasibility of the idea as a proof of concept, everyone involved in implementation tries to make it a success and show impact even when there is none. Often, pilots are implemented in places with good infrastructure (often in some large farmer’s fields). Then, assuming the limelight is on them, all the institutions involved assume that it is their job to make it a success, to show the world that it is an idea worth implementing. In all these, the core concept of testing the idea in a real-world setting goes by the wayside. Even at farmer field days, it is often the farmer who is better connected to the extension system and participates regularly in training programs who is chosen. The yield and benefits they obtain in their field cannot be compared to those of an average farmer. Therefore, the promise of a technology, even in a pilot or field demonstration, does not necessarily reflect its scalability.

Stakeholder in technology change: Technology change is a complex process, involving multiple stakeholders who coordinate with each other, exchange ideas, and co-innovate as reflected in the Agriculture Innovation System (AIS) concept, challenging the traditional notion that technology adoption is a linear process, where farmers receive the technology as a package, and implement it without any changes. In spite of the recent recognition of the need to bring in the AIS lens to technology adoption, most studies focus on farmers alone. Any technology or innovation, for that matter, will benefit some stakeholders through its widespread adoption, while others may lose. This different distribution of ‘gains’ and ‘losses’ of the transition across stakeholders needs explicit attention in policy making. If a sufficient number of stakeholders do not have an incentive to push for scaling, it will be challenging to scale, despite the policy push and technological promise (Barrett 2023).

Institutional silo’s: Another important dimension of this problem is ‘path dependency’ and ‘disciplinary/institutional silo’. Most agricultural technologies or innovations span multiple departments and may require collaboration among various institutions for implementation. However, there is an ‘inertia’ or path dependency that the department traditionally involved with a given technology will continue to implement it with minimal convergence of relevant departments/institutions. A classic case is Solar Irrigation Pumps in India under PM KUSUM, where the Ministry of New and Renewable Energy is the Implementing agency.

Wrong framing of the technologies: Many technologies are typically framed as ‘climate smart’ / ‘resource conservation’ technologies and often fail to scale. Many of these technologies have large social benefits of adoption but little to no private benefits. Even if there are private benefits, they are not highlighted in the framing of the technologies to reflect this. The result is that they are not adopted at scale, thereby limiting the benefits. For example, if Alternate Wetting and Drying (AWD) is promoted as a water conservation technology without any direct financial gain from adoption, it may not be attractive to farmers. Though from a rationalistic perspective, the gains from adoption are positive, as without a yield penalty, farmers are conserving water. However, farmers often prioritize immediate economic benefits over abstract environmental gains, and rightly so. Any new technology will come with substantial learning costs and uncertainty (whether perceived or real), and the incentive must outweigh these costs for the farmer to adopt it. Similar is the case with Solar Irrigation Pumps. If the farmer already has an electric pump, electricity for pumping being free, ‘clean energy’ cannot be an incentive for farmers to adopt. In fact, as the research highlights, main incentive for farmers to adopt is ‘day time power’ as it reduces the drudgery of going to the field at night, as the regular grid-based electricity supply is rationed and provided only during night (Korekallu & Mithöfer, 2025). Ultimately, the point is, the framing of technology has to be ‘farmer-centric’- it should answer the question ‘what is in it for a farmer to adopt?’.

Under the PM-KUSUM scheme in India, the Ministry of New and Renewable Energy (MNRE) is responsible for overseeing the program at the national level. However, since agriculture is a state subject, each state handles implementation differently. For example, in Karnataka, the Karnataka Renewable Energy Development Limited (KREDL) is the agency in charge. What’s striking is that the agriculture department is not involved at all in a few of the states, including Karnataka. This creates a problem because farmers are more familiar with the agriculture department, which usually handles subsidies and farm necessaries. The department also has better reach and experience working with farmers. In contrast, other institutions may not always have the experience of interacting with individual farmers, making it harder for them to manage the process effectively. (Korekallu & Mithöfer, 2024).

Innovation bundling: One key benefit of farmer-centric thinking is that it is easy to understand why many technologies have failed to scale. The benefits are often indirect and hard to measure, while the direct benefits are uncertain. Farmers tend to have a strong status quo bias, resisting change unless there is a significant incentive. Recent studies also show that farmers exhibit hyperbolic discounting; a high preference for immediate benefits, even when the real value of future benefits outweighs the present benefits (Begho & Anik, 2022). Broadening this thinking, as mentioned earlier, even among different stakeholders, a given technology can have varying costs and benefits. Therefore, stakeholders who will be disadvantaged by the adoption of a new technology may not cooperate and can create hurdles at various levels based on their power and control over resources. Barrette (2023) highlighted that one strategy in these situations is to bundle innovations in such a way that, as a technology bundle, the innovation or technology has sufficient incentive not only for farmers but also for a sufficiently large group of stakeholders, thereby accelerating the transition. Another significant side benefit of bundling is ‘political economy’. Only those policies that can be seen as ‘popular’ and get votes are the ones to get implemented with political will (dangle the carrot, hide the stick). Innovation bundling (either tech-tech bundling or socio-technological bundling) can also make it more attractive to policymakers.

Barette (2023) provides clear examples of innovation bundling in China’s science and technology sector. The main challenge was that farmers were not adopting most soil and water conservation practices, often because these involved unfavorable trade-offs like increased labor. Farmers sought technologies and products that would boost their income. Conversely, scientists aimed to develop new technologies that could be published. Input suppliers needed new markets to establish new channels for their products. China’s Science and Technology Backyard (STB) program, launched in 2009 by China Agricultural University, is an innovative effort to connect scientific research with the practical needs of rural farmers. In this model, professors and graduate students live and work alongside farmers in rural villages, often renting farmers’ backyards to conduct research, provide technical training, and co-create solutions for local agricultural problems. As farmers saw benefits from backyard research farms, these sites became technology hubs, attracting more farmers, with local governments supporting close collaboration between researchers and farmers. Input suppliers and companies also gained by testing and promoting their products, turning Science and Technology Backyards into multi-actor platforms where all parties advanced their goals while contributing to broader agricultural development. This exemplifies socio-technological bundling, where technology is combined with institutional innovation.

Conclusion: Scaling agricultural technologies is not merely a matter of technical efficacy or policy endorsement. It requires thoughtful piloting, inclusive stakeholder engagement, appropriate framing, and institutional coordination. Innovations must offer tangible incentives to farmers and other actors in the ecosystem. Approaches like innovation bundling, as seen in China’s STB model, offer promising directions. If we wish to move from ‘pilot projects’ to transformative impact, we must design for scale from the start with the farmer at the centre. To address the scaling challenges, CGIAR has launched its unique Scaling for Impact (S4I) program to better align research with real-world needs (CGIAR, 2024). Further, to accelerate the adoption and scaling of sustainable, inclusive innovations across agrifood systems to tackle pressing global challenges. It does so by combining scientific rigor, strategic partnerships, financial mechanisms, system-level learning, and a commitment to inclusivity.

Reference

Barrett, C. B., Benton, T., Fanzo, J., Herrero, M., Nelson, R. J., Bageant, E., … & Wood, S. (2022). Socio-technical innovation bundles for agri-food systems transformation (p. 195). Springer Nature.

Begho, T., & Anik, A. R. (2022). Examining the effect of climate vulnerabilities on the discounting behaviour of farmers. Climate Resilience and Sustainability, 1(4), e46.

CGIAR. 2024. Scaling for Impact Program: Full design document. Agenda item SC21-05a, 21st CGIAR System Council meeting, Berlin, Germany, 11-12 December 2024. Montpellier: CGIAR

Korekallu Srinivasa, A., & Mithöfer, D. (2024). Unpacking stakeholder perceptions on challenges for increasing adoption of solar-powered irrigation systems in India: AQ methodology study. Q Open, 4(2), qoae020.

Korekallu Srinivasa, A., & Mithöfer, D. (2025). Understanding Farmers’ Policy Preferences for Solar‐Powered Irrigation Systems in Karnataka, India: A Choice Experiment Approach. Australian Journal of Agricultural and Resource Economics.