Video-based support for small-scale farmers around the world

Governments invest heavily in agricultural outreach programs to help small-scale farmers improve their farming practices, but these programs often struggle to make information customized, relevant, and timely. To address this challenge, the organization Digital Green develops short how-to videos, featuring local farmers executing complex agricultural practices, to incorporate into government-run outreach programs. A randomized evaluation in India found that the approach increased women farmers’ agricultural productivity and profits at low cost. Digital Green, in collaboration with India’s Ministry of Rural Development and NGO partners, has since scaled this model throughout India and adapted it to twelve countries, reaching 7.2 million farmers. 

The Problem

Traditional outreach models, known as agricultural extension, are costly and do not effectively help small-scale farmers adopt new practices that could increase their yields and profits. 

Small-scale farmers manage over three-quarters of the world’s farms, yet they only produce 35 percent of the world’s food.¹ Many of these farmers experience poverty and are not aware of or don’t have access to new technologies and practices that could potentially increase their productivity. Agricultural extension programs, in which public agents visit farmers individually or in groups to share information, are designed to help bridge this gap. 

However, research shows traditional extension often does not effectively deliver customized, relevant, and timely information. Providing in-person agricultural advice at scale, across the remote areas where farmers live, poses operational challenges. In India, for example, the ratio of public extension workers to farming families has consistently been over one to 1,000.² Furthermore, traditional extension often provides broad recommendations, which might be too generic to help farmers tackle specific issues on their farms. In addition, extension is rarely tailored to a farmer’s level of education, technical understanding, or local language.

There have been many efforts to innovate extension models by harnessing information and communication technology (ICT). Mobile phones, for example, offer a way to reach more remote farmers at low cost. However, these approaches aren’t suitable for all farmers since SMS requires farmers to be literate, which is not common in many rural and remote areas, often particularly among women. The level of detail SMS interventions can convey is limited, which can prevent farmers from being able to adopt more complex practices.

This challenge of effectively communicating complex practices is particularly salient as climate change alters the landscape of inputs and techniques beneficial to farmers. As droughts become more frequent, new practices like system rice intensification (SRI), a multi-step technique for cultivating rice that reduces water use while increasing yields, have the potential to help farmers adapt. Improving extension to help farmers adopt complex practices is a key policy priority for many governments seeking to improve farmers’ incomes and local food systems.

The Research

Researchers found Digital Green’s community video advisory model to be a cost-effective approach to increasing farmers’ agricultural productivity and profits. 

Digital Green is a global development organization founded in India that harnesses digital tools to improve traditional extension systems. In 2012, Digital Green began collaborating with Jeevika, a state-funded organization under India’s National Rural Livelihood Mission (NRLM). NRLM is one of the largest poverty reduction programs in the world, run by the Indian Ministry of Rural Development and supported by the World Bank. 

In 2014, J-PAL affiliated professor Dean Karlan (Northwestern University), together with Tushi Baul, Kentaro Toyama, and Kathryn Vasilaky, partnered with Digital Green and Jeevika to evaluate the impact of incorporating short how-to videos into group extension sessions run by Jeevika’s network of self-help groups. The videos, which featured women farmers from local communities, showed how to execute the many steps in the system rice intensification (SRI) practice. Videos included information on additional labor needed during pre-planting and planting phases, as well as messages designed to build farmers’ confidence in their ability to independently carry out the steps in SRI. 

To measure the impact of the video model compared to traditional face-to-face extension, researchers randomly assigned 280 villages in Bihar to watch the videos during extension sessions. Another 120 villages continued to participate in extension sessions without the videos. In the video villages, some also received either a labor cost message, a self-confidence message, or both to address the specific barriers identified for using SRI.

Researchers found that the video increased farmers’ agricultural productivity and profits.³ After one year, farmers who were offered the videos increased yields by 12-18 percent and estimated profits by 9-24 percent, relative to farmers who received conventional extension, with smaller effects after two years. The video was most effective at increasing agricultural output and profits when complementary messages on both labor costs and self-confidence were included in the videos, or when neither of these additional messages was added.⁴

The study did not find increased adoption rates of SRI, but points to two possible explanations for this. First, since all farmers received the same information about SRI through group extension with or without the videos, the video may not have increased adoption rates beyond those who received conventional extension. Second, the study did not measure partial adoption of SRI nor did it measure farmers’ execution of each task, both of which may have been affected by the videos. 

Overall, the study found that the video model was cost-effective relative to in-person-only extension, indicating a viable pathway for scale. Researchers estimated that a farmer’s average additional profit was about 16 times greater than the additional cost of implementing the video approach. Another team of researchers estimated that the cost of a farmer adopting a new practice as a result of the video-based model was roughly US$3.50, compared to US$35 through traditional extension.⁵

To learn more, read the evaluation summary. 

From Research to Action

Backed by evidence from the randomized evaluation in Bihar, Digital Green and its government and NGO partners scaled video-based models in India and around the globe, reaching over 7.2 million farmers.

From its inception in 2008, research, testing, and iteration have been at the heart of Digital Green’s approach. In 2009, qualitative fieldwork showed that farmers prefer to see and hear information coming from peers, rather than trainers or government officials. This became a key pillar in the video-based approach. 

The randomized evaluation in Bihar convinced Digital Green’s government partners of the video model’s cost-effectiveness. Within Bihar, the results encouraged Jeevika to expand the model to 3,000 additional villages in Bihar. The evaluation also led Digital Green to incorporate farmer agency and self-confidence as a north star in their theory of change. 

The study led other state rural livelihood missions, including in Andhra Pradesh, Jharkhand, and Odisha, to incorporate the video model into self-help groups. In Andhra Pradesh, for instance, the Department of Agriculture and Cooperation worked with Digital Green to add the video to its existing extension services across all 13 districts.⁶  In recent years, the Government of India, the World Bank, and state governments of Bihar and Andhra Pradesh provided co-funding of US$23.2 million, covering the costs of cameras, projectors, and staffing for scaling.⁷

The rigorous evidence supporting Digital Green’s model in India caught the attention of other governments with large rural populations. In 2014, Ethiopia’s Ministry of Agriculture approached Jeevika and Digital Green to learn about the video model. The ministry sent a delegation to India to see the model in action in Andhra Pradesh and Bihar. Researchers from the evaluation in Bihar participated in these exchanges and provided input on adapting the model to Ethiopia. 

Subsequently, in 2017, J-PAL invited researcher Tanguy Bernard (University of Bordeaux) and others began evaluating the video model in Ethiopia, which showcased agricultural practices recommended by the Ministry of Agriculture (MoA). Researchers found that video advisory increased farmers’ knowledge and adoption of techniques, though it did not increase their yields in the short term compared to conventional extension.⁸ Researchers estimated that, when scaled to all villages in the study, the program would cost US$3-6 per adoption, making it potentially cost-effective at scale. 

Considering the lessons from these two randomized evaluations, the Ethiopian government integrated the video model into its Second Growth and Transformation Plan and the Second Agricultural Growth Program with a US$12 million investment in 2015. The MoA and the Regional Bureaus of Agriculture have since scaled the program to more than 600,000 farmers.⁹

Digital Green’s community video model now operates in twelve countries and has reached 7.2 million small-scale farmers to date, of which 57 percent are women. In all geographies, Digital Green collaborates with government ministries and NGOs to institutionalize use of the videos. 

Digital Green continues to test and evaluate innovations to this model. Fueled by the rise of mobile connectivity and the global pandemic, when in-person meetings were not possible, Digital Green began sharing videos with farmers through WhatsApp. The organization is now exploring—and rigorously evaluating—ways to harness generative AI to increase customization, such as using chatbots in local languages. In Kenya, for example, Digital Green is working with IFPRI to evaluate Farmer.Chat Kenya, an AI-powered digital assistant. 

As they explore the use of AI, the organization is maintaining the core components of the video model that have been rigorously evaluated, like the use of visuals and focus on building self-confidence, while considering how extension agents can backstop when AI reaches its limits. Digital Green aims to reach five million additional farmers with its evidence-informed programs by 2029.