Supporting Digital Green: bringing tech-enabled solutions to agricultural extension

Decision-maker’s challenge

Boosting farmer productivity with the most up-to-date information can be labor-intensive and costly. Digital Green is working in India to bring tech-enabled solutions to agriculture extension for large-scale impact. The NGO supports state agencies to deliver low-cost, interactive, and relevant content through a combination of videos, mobile phones, and participatory workshops. Digital Green’s aim is to boost farmers’ adoption of new practices and, as a result, increase their incomes. 

Impact opportunity

In India, small-scale farmers, who account for 80% of the country’s farmers, learn about better cultivation methods through government-run agricultural extension services. However, traditional extension is labor-intensive, and relies on large-group lecture formats that are costly to scale-up and difficult to tailor to farm communities. To date, this approach has not resulted in high adoption rates by farmers. 

DG has been innovating to create less costly, higher impact services. Over the last decade, the organization has created a video-based initiative that disseminates information on best-available cultivation methods. The video program, which has reached more than 2 million farmers around the world, is focused on meeting needs and building on assets in farm communities, 

In 2020, IDinsight completed a study with DG to assess the effectiveness of their new  “FarmStack” system. In the project, we compiled evidence on the program’s impact on farmers and its institutionalization within the government delivery system.

Our approach

IDinsight worked with DG on three workstreams:

First, we conducted an evidence/landscape review, collating findings from internal and external studies conducted on DG to develop a narrative about the organization’s impact on farmer reach, technique adoption, cost-effectiveness, and farm yields and income. 

Second, given that DG’s model has been adopted by government partners in some Indian states, we conducted an in-depth process evaluation of the organization’s efforts to integrate its video-based approach with public sector extension systems in Andhra Pradesh. Data collection involved interviews with 57 DG staff members and government officials, plus 20 beneficiary farmers. The work led to key recommendations from IDinsight on how to advance government institutionalization.

Third, we designed and ran a randomized controlled trial (RCT) to estimate the impact of DG’s newest tech platform called FarmStack, which delivers customized and time-relevant cultivation, crop, and weather information directly to farmers’ phones. The RCT was conducted with more than 2,000 small-scale cashew farmers across 81 rural communities in Andhra Pradesh.

The results

In our evidence review, we found that:

1. A wide array for research has been conducted on Digital Green’s video-mediated approach, including four rigorous RCTs.
2. Studies generally show that Digital Green’s programs increase access to extension, knowledge, and adoption of practices. Results on production and yield are mixed. 
3. More research is needed on long-term effects, how to maximize impact on women, and how to best leverage new technologies (such as mobile phone).

In our process evaluation on government institutionalization, we identified and described:

• Current versus ideal distributions of responsibilities between DG and the Department of Agriculture (DoA)
• Key explanatory drivers of uneven adoption across state districts, including divergent levels of buy-in by DoA leadership 
• Strategies for accelerating institutionalization, such as improved payments systems for field staff to boost motivation, and better data collection infrastructure to strengthen performance management.

In our RCT analyses, we found highly promising results when comparing farmers who were randomly assigned to participate in the program with similar farmers who were randomly selected not to participate: 

• Farmers assigned to participate in the program showed a significant increase (0.232 standard deviations) in their knowledge of novel practices relative to farmers who did not participate.
• Farmers assigned to participate in the program reported higher adoption of novel practices by 5-8 percentage points relative farmers who did not participate (where adoption ranged from 13-30 percent).
• In an “at-scale” scenario, we estimated it would cost DG approximately $16 to influence a farmer to adopt a high-productivity practice.
• FarmStack effectively reached under-represented populations. Among the participants, 81% of farmers accessed at least one FarmStack advisory — a strong sign of the potential for mobile-based extension to perform in hard-to-reach geographies. Additionally, across several knowledge and adoption indicators, participating female farmers showed greater gains relative to male farmers.

DG will use these and other findings to bolster its scale-up efforts with the Department of Agriculture. They expect to be able to exceed a target of providing high quality extension to 300,000 farmers across the state (currently the recorded reach is 240,000 farmers). We anticipate IDinsight’s decision support can accelerate the participation of at least 60,000 more farmers receiving training services. DG used our initial RCT results to help secure additional funding from the Walmart Foundation to fuel FarmStack’s expansion in Andhra Pradesh in 2021.