Precision Agriculture for Development

I grew up in a small town in Western Kenya. I saw how hard my parents worked on our farm. Despite many challenges, they ensured that I could access education. The example of my parents, and their values, shaped the person I am today.

While working as a data entry clerk in 1996, I met a young economist named Michael Kremer. I was Intrigued by his research on isolating causal impacts to scale actions to address poverty. This led me to enroll in the MPA program at Harvard Kennedy School. 

I have long believed in the power of technology to create impact, and was thrilled when Prof. Kremer co-founded PAD in 2016. I am particularly motivated by PAD’s drive to empower farmers with information using technology. Through their mobile phones, farmers can make informed productive decisions and no longer need to repeat mistakes that were common among my parents' generation.

Most of the world’s 600 million farmers live in a village setting in a developing country and typically harvest just 30 to 50% of what their land is capable of producing. Despite significant public spending in support of over one million agricultural extension workers worldwide, farmers continue to lack the information and advice they need to close the “yield gap” and maximize their net incomes. PAD provides a two-way flow of digital information to farmers using SMS, interactive voice response (IVR) and other digital tools to deliver low cost, customized advice to improve on-farm practices, input utilization, pest and disease management, climate and weather resilience, environmental sustainability, and access to markets. We propose to expand our services to 20 million farmers in 10 countries in Africa and Asia. These services will increase farmer knowledge about and adoption of improved technologies and practices, as well as improving their yields and income.

The average smallholder farmer cultivates a small area (85% of the world’s farms are 5 acres of land or less), and harvests only one third to one half of what they could. These “yield gaps” affect hundreds of millions of farmers worldwide and can be attributed to suboptimal farming practices caused by lack of information about and/or low availability of inputs or technologies. Farmers typically obtain agronomic information by asking their peers, or, if available, from government agricultural extension workers (EWs). The challenge, however, is that peers have only limited information which may not be accurate, and government EWs cannot meet the large demand for their services. In one study in India, for example, only 10% of farmers had interacted with an EW in the past year. We help governments and other partners to develop impactful digital services that can be accessed by farmers for free, using their own mobile phones.

Our approach relates directly to the lack of information experienced by farmers; we provide real-time, highly localized information to farmers that expands their knowledge and leverages their own mobile phones, which allows 2-way communication to further customize information as farmers share their information with us.

We have developed a set of 2-way, interactive digital information services that allow farmers to send and receive information about their farms using their simple mobile phones. These services are customized to the specific needs of farmers, and the communication channel varies by context. For example, in Kenya, where text messaging is very common, we have developed an interactive SMS service. In India, where SMS is less common, we’ve developed an Interactive Voice Response (IVR) service that allows farmers to call in and receive pre-recorded messages or record their own agronomic questions.

All services have the following core features:       

• Free for farmers

• 2-way, which enables us to gather information from users and collect real-time contextual information (e.g. pests, input availability, market conditions)

• Scalable and highly cost-effective services, with very low marginal costs, enabling rapid scale-up with very high benefit-cost ratios at scale.

• Evidence-led 

• Trusted, working with partners in government and research institutions. 

• Customized, to the extent possible, so that farmers receive the most relevant information for them. 

• User-centred, informed by behavioural economics and A/B testing, so that information is provided in the most actionable and effective form.

• Data-led, enabling rapid iteration, learning and adaptation.

Our project targets smallholder farmers living in rural areas of developing countries, often living at or below the poverty line. As subsistence farmers, agricultural production constitutes their primary source of food security, nutrition and income. We partner with local governments, research organizations and other stakeholders to assess the agronomic and other information needs of our users, and use our expertise in behavioral economics and A/B testing of messaging and service features to continually improve our services. Our engagement typically involves face-to-face meetings with representative farmer focus groups during which we engage with farmers and hear feedback from them about their own needs and challenges. We also consult with local experts and partners including, in most cases, the national Ministry of Agriculture, who provide us with their own agricultural extension content and recommendations. 

Once initial content is developed and services are launched, we continue to test and iterate on them, measuring the impact of the information on farmers’ knowledge, behavior and, where possible, their agricultural yields and income.

Over 80% of the world’s farms are 5 acres (2 hectares) in size or less. Than almost any other profession in the world, smallholder farmers are more likely to live in poverty and lag behind on human development. Compared to larger farmers, they lag behind on access to agronomic support services, information, and appropriate technology. As a result, they typically harvest only 30-50% of what their land could produce. Our approach empowers these farmers - most of whom now own a mobile phone - to gain access to reliable, accurate and customized information to expand their production and income.

Mobile phone penetration is the underlying factor that drove Prof. Michael Kremer, a professor at Harvard and recent Nobel Laureate for Economics, into investigating the impact of SMS-based agricultural advice among sugar cane farmers in Kenya in 2011. The results were encouraging; farmers experienced yield gains of 11.5 percent, while interruptions in fertilizer distribution were reduced by 21.6 percent.  Since then, mobile phones have become even more affordable, lowering the cost of digital communication and potential for scale. 

I was involved in the earlier project in 2011 and when Prof. Kremer co-founded PAD in 2016, I was excited to join the project to share lessons that I had learnt previously.  Our initial project in Kenya was initiated by an outcry from the government. A nation-wide infestation of a new, destructive pest - the Fall Armyworm - threatened the country’s main staple crop and food security. 

To respond to the emergency, PAD set up a two-way SMS platform and worked together with the government, experts and Kenya’s mobile networks to send out over 20 million SMS inviting farmers to join the platform. Since then, we have reached millions of users and expanded our approach to seven other countries.

Growing up in a farming community where livelihoods depended predominantly on traditional practices, intuition, personal experience and handed-down beliefs, I am particularly motivated to implement data-driven solutions to optimize inputs and maximize agricultural productivity.  

As a child I was engaged in farming practices without knowledge of what I was doing. Every day before cooking, for example, I had to take cold ash from the fireplace to the farm. Years later, I learned  that ash improves soil pH, making nutrients more available to plants, which can greatly improve productivity. 

I was privileged to attend one of the very top universities in the world, where information is abundant. But there are many farmers in my community even today who still don’t have access to basic agronomic information that can drastically improve their lives. 

 My hope for Kenya is that information won’t remain locked up inside of ivory towers, but will be freely available for those who need it the most - the millions of farmers who together can reach the nation’s huge agricultural potential. This is why I am so passionate about my work at PAD.

I am a seasoned development practitioner with over two decades of experience. I have a degree in management from Africa Nazarene University and an MPA from the Kennedy School at Harvard University.  I believe that the knowledge that I have acquired through the course of my studies, and the skills I have developed through my experience in project evaluation,  leadership and management, are tremendous assets that can contribute towards the success of this project.  My decade of experience in managing and leading high-level government engagements will assist me in navigating unforeseen challenges that will inevitably arise as we expand and deepen our platform (new crops, content, etc), develop new partnerships, and expand into new countries.

For example, I have organized and led high level discussions between PAD Kenya and the government regarding the expansion model for the MOA-INFO platform, I have also participated in numerous County Agricultural Strategic meetings. Therefore my ability to navigate both the National and County governments is an asset to PAD Kenya’s expansion plan.

My personal engagement in the community, my understanding of culture and local context - including my own first-hand experience growing up in a Kenyan farming family  - are relevant to the success of this project.  My ability to assess, predict and mitigate risks will also play an important role during implementation of the project.  I have very strong leadership skills, am highly motivated, with a passion to use evidence to alleviate poverty among the underserved communities.

Our project relies on support through the Ministry of Agriculture. All content developed and shared through our platform must be reviewed and approved by government experts.  Government experts have their own work assignments and contrarily work with several partners with competing needs, as a result navigating their systems is an uphill task that has sometimes delayed our ability to provide farmers with timely content.

One of the major challenges that we have faced is coordinating the government team and ensuring that the content is reviewed in time for the crop cycle.

To manage this challenge we  applied multiple strategies; 1) volunteered our agronomist to work with the government experts at their offices 2) facilitated meeting requirements such as transport 3)developed crop calendars for major crop and shared them with the relevant experts to insinuate urgency 4) devoted more time towards networking with the government experts by attending meeting invited by both the government and other partners 5) set up update meetings with top officials in the ministry to keep our services on their radar.   This multifaceted approach has over time improved our relationship with the government and improved our engagement processes leading to timely content development for our farmers.

Prior to joining PAD, I worked as a Deputy Country Director (DCD) at Innovations for Poverty Action Kenya (IPAK).  IPAK had over thirty researchers running multiple projects annually with over 400 employees. Even though all the projects were under one umbrella,  operations such as staff contracts were guided by individual projects, a process that was perceived as unfair.  While researchers were keen at achieving high-quality, low-cost operations, staff demanded pay raises and job security. As DCD, one of my roles was to balance the needs of both parties equitably.  The underlying problem was caused by researchers who did not want to publicly share hiring information such as contract terms,  a practice perceived as undesirable among high performing staff who would benefit from internal transfers.

 To mitigate this problem, I created a collaborative environment that promoted individual communication among researchers and an understanding of standardizing salaries across projects. I clarified the importance of publicly declaring staff contractual terms and making known available positions ahead of time to guarantee our top performers’ job security.  After repeatedly engaging them in this process, I received positive feedback from both researchers and staff expressing their satisfaction with the new placement system.

Beyond simply creating a digital version of agricultural extension services, our approach relies heavily on human behavioral economics, which is grounded in the motivations, constraints and nuances that shape behavior in various settings. This is very important for digital advisory services, where small tweaks to content, tone or messaging can have significant impacts on user experience and comprehension. We are constantly measuring the impact of these tweaks, or A/B tests, on those outcomes.

 A good example of this is from our 2-way SMS platform built in cooperation with the Kenyan Ministry of Agriculture. We conducted A/B tests to measure the impact of messages on self-reported behavior change. In two trials, SMS messages increased the self-reported likelihood of adopting the recommended practices by 5-23%. But those gains would never have been made had we not tested out the messages, and systematically iterated on our communication approach.

A wealth of existing evidence suggests that mobile phone-based advisory services positively impact outcomes, such as agricultural information dissemination, knowledge and comprehension, adoption behavior, and agricultural yields. A recent article published in the journal Science takes stock of existing evidence on digital agricultural extension and reports the results of a meta-analysis of available impact estimates (Fabregas, et al. 2019a). Meta-analysis helps us to systematically identify the magnitude of the potential impact of digital agricultural extension across varying contexts. These meta-analyses find evidence that PAD’s model can change farmer behaviour and improve yields. 

 In their review of six studies, Fabregas, et al. found that farmers who received advisory text messages promoting the use of lime, on average, adopted lime at a rate 11.28% higher than farmers who did not (Fabregas, et al., 2019b). For context, the proportion of farmers in the control groups that adopted the recommended inputs ranged from 5 to 32%. While the underlying studies did not collect data on yields, evidence from agronomic trials suggests that the application of lime increases maize yields by 5-75% (Fabregas, et al., 2019b). This suggests that the profit increase per treated farmer is between 10 and 100 times the per farmer cost of implementing the programs (Fabregas, et al., 2019a). 

 In a 2018 pilot involving 1,585 cotton farmers in Gujarat, we tested the effect of customized recommendations on soil fertility delivered to farmers through the use of visual aids and weekly push calls. Preliminary analysis suggests a more than two-, four-, and five-fold increase in the likelihood of treated farmers using UREA, MOP, and Zinc, respectively, when compared to farmers in the control group (Shawn Cole, et al. 2018). 

We currently reach 3.6 million farmers in 8 countries. These numbers represent only “active” users of our various advisory platforms who engage with our services - and don’t include the millions more that have been made aware of or invited to use our services. 

 In the next year, we have already secured funding and/or partnerships to expand into Nigeria, with potential new partnerships in the DRC, Brazil and Colombia. We also expect to continue to expand our reach within Kenya, Pakistan, and India. Thus, by the end of 2021, we expect to serve at least 5 million farmers in 10 countries. 

 In five years, we expect to be active in another 3 to 5 countries. Thus, by the end of 2025, we expect to serve 20 million active users in 12 countries.

 As an example of how we will obtain these ambitious goals, we have recently entered into a partnership with the International Fund for Agricultural Development (IFAD) - a specialized agency of the UN active in 177 member states. We have already begun partnering with IFAD in Kenya, Nigeria and Pakistan, and we expect to expand this partnership in at least 5 more countries in the coming years. Our co-founder and recent Nobel Prize Laureate, Prof. Michael Kremer, recently conducted a webinar with IFAD’s president, Gilbert Houngbo, entitled “Farmers on the Frontline: Making Digital Agriculture work for Small-scale Farmers in a Post-Coronavirus World” - evidence of the strategic alignment of our two organizations.

We measure our impact success based on 2 main metrics: reach and benefit cost ratio (BCR).

 Reach:

 We currently reach 3.6 million farmers with digital agricultural services in eight countries. Within the next year, we hope to reach 5 million farmers in ten countries, and within the next five years, we hope to reach 20 million farmers in twelve countries. 

 BCR:

 While it is not feasible to measure the economic value of our service to farmers and compare it to the costs in every initiative, we expect to be able to measure the BCR in a large enough sample to make data-driven insights about our impact and cost effectiveness. Our goal is to maintain at least a BCR of 5:1 in our initiatives. 

As any non-profit organization, we face financial challenges as we aim high to maximize our reach and impact among our constituents. This involves primarily the challenges associated with fundraising, especially during the COVID pandemic with its environment of heightened uncertainty.

Operationally, our organization must find the right balance between rapid growth and human and financial capacity to support operations. Should we grow too slowly, we may fail to optimize our impact and cost effectiveness as many fixed costs are divided across a smaller pool of farmers. But if we grow too quickly, our capacity as an organization to support each farmer, and the ongoing process of iteration and learning that we prize, may be diluted across too many initiatives, leaving us unable to fully support our existing services.  

 We have learned that each initiative has its own growth cycle, which must be fully supported at each stage for it to succeed. The initial stages usually last a year or two, and involve a high amount of relationship building among partners. This stage takes time as we develop trust and rapport with local government and institutions, and hire and support local teams. This often results in an initial pilot, which involves reaching a smaller number of farmers (i.e. 100,000 farmers). Most often, it is not until we successfully roll out this stage that larger scale expansion is possible (i.e. reaching 1,000,000 farmers or more). Prioritizing and allocating resources to various stages of this process is a challenge in itself. 

To address our financial barriers, we plan to strategically approach new types of donors and organizations that would be potentially interested in partnering with us in reaching smallholders at scale. Specifically, multilateral organizations are beginning to take interest in our work. The most recent example is our upcoming partnership with IFAD, which is beginning as a grant to address COVID-related issues in agricultural value chains in India, Kenya and Nigeria, but may be expanded to a broader partnership across many other geographies. 

 We have a strong team of development practitioners within our organization, but we are also seeking to structure our organization’s development for more ambitious growth by hiring for new positions to support this work as well as by using software to track and identify new opportunities. 

 Once funding is secured, allocating it in an optimal manner to new and existing initiatives requires us to maintain a delicate balance across our teams. Our priority is to focus first on finding the right people for any new team or initiative, and then supporting them throughout their development. Our greatest resource here is our past experience - including our mistakes - that has shaped our strategy moving forward. 

Our current partners include private sector, government and non-profit organizations. In most cases, we work directly with national or regional governments to provide digital supplements to their face-to-face extension work. These and other partners provide us with local knowledge, content for our advisory services, and access to our end users. In some cases we develop services for our partners (such as Kenya’s Ministry of Agriculture MoA-INFO Service), and in other cases, our partners invite us to improve their own services (such as Ethiopia’s 8028 Farmer’s Hotline). 

 Global: International Fund for Agricultural Development (IFAD) (subject to final signatures)

 India 

Coffee Board of Karnataka

State Government of Odisha  

State Government of West Bengal

The Nature Conservancy

Pakistan 

Prime Minister’s Reform Team

Provincial government in Punjab 

HarvestPlus

Bangladesh 

mPower

Kenya 

Ministry of Agriculture and Livestock

Kenya Agricultural and Livestock Research Organization (KALRO)

One Acre Fund

CABI

Tulaa

World Bank

Ethiopia 

Ministry of Agriculture

Agricultural Transformation Agency

Rwanda 

One Acre Fund

Root Capital

Ministry of Agriculture

Rwanda Agricultural Board

Zambia

Ministry of Agriculture

Smart Zambia Institute

CABI

FAO

Uganda

TechnoServe 

HRNS

As a non-profit organization, we support our work primarily through philanthropic donations, but also pursue partnerships with governments or public organizations to offset some of our costs. For example, through our partnership with the Ministry of Agriculture in Kenya, we have been able to cooperate with local mobile networks and invite farmers onto the platform via SMS at reduced rates. We are also exploring revenue generating activities that complement our work with private sector players. 

Due to our success at raising financial support through donations and grants, we believe that a significant portion of our operation costs will continue to be covered through philanthropic means. Since our founding, we have raised over $19 million in the past 4 years to cover our costs and rapidly expand new initiatives. 

 However, we are also mindful that relying entirely on donor funding may create perverse incentives and may distract our focus from the smallholder farmers whom we are seeking to serve. As such, we are actively pursuing partnerships with private sector businesses that may generate additional value to our users and revenue streams to our organization. 

 One example of this includes our current partnership with BigHaat, an online marketplace in India for agricultural inputs. Since their platform requires internet access, we are offering to provide them with our call center and hotline service to inform our users about the inputs that they offer, and enable them to place orders without needing to first gain access to the internet. We will be receiving a 3% commission on all orders placed by users of our platform. 

 We expect there to be other similar partnerships possible as we expand our work in the years ahead. 

Since 2017, we have raised over $19 million.

 Major sources of funding include:

Unorthodox Philanthropy - $1,193,500 - grant

Bill & Melinda Gates Foundation - $4,300,000  - grant 

Anonymous - $6,241,100 - grant

Anonymous - $4,500,000 - grant

International Fund for Agricultural Development (pending signatures) - $2,000,000 - grant

Goldsmith Foundation - $400,000 - grant

Mulago Foundation - $100,000 - grant

Dioraphte Foundation - $191,440 - grant

Jasmine Foundation - $200,000 - grant

Yajilarra Foundation - $300,000 - grant

Vitol Foundation - $200,000 - grant

PAD is seeking to raise $20 million to expand our work over the next 5 years to reach at least 20 million farmers in 12 countries. Our current cost per farmer is less than $1.50 per year, and we expect at that scale to be able to reach farmers with our services for less than $1 per year.

Our total projected expenses for 2020 are $4.4 million. This includes “all-in” fixed and operational expenses for reaching roughly 6.4 million farmers across our 8 countries of operations by the end of the year. 

Regarding fundraising, we believe The Elevate Prize, by raising awareness of our work among your network, will greatly help us achieve our near term fundraising goals as well as develop lasting, strategic partnerships. 

Regarding our operational challenges and finding the right balance of growth, we believe that the Prize's consulting team may help us with some of those strategic questions that we face as we move forward into the scaling phase of our organization.

We wish to align our organization with like-minded individuals, philanthropic organizations, and impact-driven actors in our next phase of development. We are particularly interested in partnering with multilateral organizations and governments in developing countries seeking to provide a scalable solution for their farmers.