Feed the Future
This project is part of the U.S. Government's global hunger and food security initiative.

Yelp for Smallholder Farmers: Does It Work and How?

Where can you find the best burrito in San Francisco’s Mission District? What brand of neck pillow works best for a long trans-Pacific flight in coach class?

The way modern consumers would quickly answer these questions by turning to Yelp or Amazon’s customer ratings demonstrates how dramatically crowdsourced information and opinions have transformed commerce in the United States. Not only have these platforms changed how we make purchasing decisions, they have also affected businesses’ bottom lines. Research from UC-Berkeley professors Jeremy Magruder and Michael Anderson indicates that just an extra half-star rating on Yelp led to a 19 percent increase in restaurant reservations in San Francisco. While impacts of review and ranking sites on the consumer space are clear, perhaps a much more important question is what crowdsourced information might mean for achieving global public goods like advancing the plight of smallholder farmers in sub-Saharan Africa and South Asia.

This issue is of growing interest to many influential stakeholders and, fortunately, there is evidence that these platforms have potential to make a difference. In a study funded by the Agricultural Technology Adoption Initiative, which is generously supported by the Bill & Melinda Gates Foundation and UK Aid, researchers found that a Yelp-like platform significantly increased the success rate by which livestock farmers in Punjab, Pakistan received artificial insemination (AI) services from local veterinarians. The study was led by Arman Razaee, a researcher affiliated with the Center for Effective Global Action (CEGA). Razaee and his collaborators specifically created an information clearing house through which basic information on vets providing AI services in one district of Punjab was collected from rural farmers, aggregated across many observations and transmitted back to those same farmers. Farmers who received information on the average price, success rates and the number of observations these figures were based on demanded more from local AI providers and, on average, experienced a 27 percent higher success rates than those without this information. This success is pretty astounding, especially so when considering that constructing the platform and transmitting the information on average costs around $8 per farmer (including the fixed cost of establishing the service), while farmer returns are estimated at $32.

Full details on this three-year study are found in Rezaee’s working paper, project summary and blog. However, perhaps the most interesting aspect of this story for Agrilinks readers is what drove the success of the platform and what that might mean for similar efforts in the agricultural sector. Below are three quick insights adapted from a presentation Rezaee gave at CEGA’s annual Evidence to Action symposium last May in San Francisco.

  1. Keeping messaging simple. Rezaee identifies the simplicity of the “collection and transmission” of information as one of the key features that drove his success. The only information collected and relayed to farmers during the experiment was the average price, average success rate of the service performed and the number of observations. During Evidence to Action, Rezaee encouraged others working on monitoring platforms to keep their messaging simple, so as not to overwhelm people, especially when targeting busy farming households in rural areas.
  2. Incentives to “care." A key success factor for Rezaee was also that targeted veterinaries had a clear incentive to care about their customers’ opinions. While unlikely to lose their jobs or significant amounts of salary as a result of poor performance, the vets did receive a payment each time they offered their services, making citizen perceptions of their performance more salient. Finding settings where service providers have incentives to respond to citizen opinions is important for the outcomes of monitoring platforms. An increasing number of developing countries are also experimenting with creating new monetary incentives for public sector providers, including pay-for-performance schemes in Rwanda, India, China, Pakistan and others. Combining monitoring efforts with salary incentives may increase the impacts of citizen monitoring efforts.   
  3. The technology can shape the crowd. Rather than a high-tech platform with sophisticated features, Rezaee’s work in Pakistan operated over feature phones, with information collected through actual person-to-person phone calls. Users of the platform similarly received one simple SMS reminder on service rating. Lower-tech channels for the platform were appropriate for the setting in Pakistan, and those that want to build on this success should think carefully about the technology they use. Research on technology for citizen mobilization in South Africa finds enormous differences in the composition of participants across different technologies. While more sophisticated platforms are easier to use and offer more features, in the South Africa study, participants identified through lower-tech options were more motivated and engaged in the program over a longer period.

The Agricultural Technology Adoption Initiative (ATAI) is committed to building rigorous evidence on ways to unlock the potential of smallholder farming in South Asia and sub-Saharan Africa. The success of Rezaee’s work in Pakistan warrants more dialogue and consideration from the donor and implementation community. While the three insights above demonstrate in part how this work might be built upon, a much more robust dialogue is needed to explore the potential of Yelp-like platforms to support farmers in developing countries. ATAI and its network of researchers are eager to engage in this conversation, so please reach out at awestbury@berkeley.edu