Lessons Learned from the YieldWise Experience in the Kenyan Mango Sector: Initial Findings

Sonka, Shah, Wormald, and Agarwal

Ed Snider Center for Enterprise and Markets: University of Maryland’s Robert H. Smith School of Business

Consortium for Innovation in Post-Harvest Loss & Food Waste Reduction


This report identifies ten “lessons learned” arising from the YieldWise intervention conducted in the Kenyan mango sector. These initial findings are based upon intensive qualitative interviews, held early in 2019, with stakeholders from throughout the sector. The findings emphasize the value of a more comprehensive approach when attempting to understand the contribution of development interventions. A key premise of the YieldWise effort is that effective interventions need to extend beyond technology provision to be effective in reducing post-harvest loss. The lessons learned which are described here provide strong support for that perspective.


Managing post-harvest loss (PHL) is a challenge endemic to agricultural and food systems. Time and resources are devoted to producing an orchard of trees brimming with fruit. However, before that produce becomes available to consumers it must be harvested, stored, transported, and/or processed within systems that may operate over several months and considerable distance. Without appropriate care and attention, loss can occur at distressingly high levels.

In developing countries, PHL routinely exists at excessive levels resulting in reduced well-being for smallholder farmers, higher prices for consumers, diminished food security, and unnecessary environmental degradation. Natural and human resources employed to produce agricultural products, which are then not consumed, enlarge the environmental footprint of agriculture[1].

Recognizing this challenge, in 2016 the Rockefeller Foundation launched YieldWise Food Loss—focused on reducing food loss by focusing on fruits, vegetables, and staple crops in Kenya, Nigeria, and Tanzania—countries where up to half of all food grown is lost (https://www.rockefellerfoundation.org/our-work/initiatives/yieldwise/).

The YieldWise Initiative incorporates the following four action components:

1. Access to technologies: promoting the adoption of appropriate loss-reducing technologies.

2. Access to finance: collaborating with financial institutions to develop credit products that can be accessed by farmers and farmer-based organizations.

3. Aggregation and training: training farmers and other supply chain actors in post-harvest management and facilitating development of local aggregation centers.

4. Access to markets: stimulating demand by engaging actors across the diverse ecosystem of buyers.

Preparation and Conduct of In-field Interviews

The following discussion describes the interview process and then briefly describes the ten lessons initially identified.

Employing sophisticated qualitative methods, a team of researchers from the University of Maryland and the University of Illinois examined effects of the YieldWise initiative to introduce system-wide change to reduce post-harvest loss (PHL) in Kenya. The research goal was to document lessons learned as a result of the YieldWise efforts to reduce post-harvest loss of mangos. In doing so, the research strives to collect and better understand how and why this process (sometimes) worked and (sometimes) did not. We studied how the process of system-wide change unfolded over time. In addition, attention was devoted to how other actors’ choices and behaviors with respect to implementing changes influenced outcomes.

The interviews were carefully designed to provide broad and deep data about the YieldWise initiative, with the goal of yielding in-depth insights across the value chain. Interviews ranged in length from 30 to 90 minutes. Conducted by the same two researchers, the majority of interviews were in-person. All other interviews were conducted via video or telephone calls. After receiving verbal consent of the interviewee, each interview was recorded and transcribed to facilitate data analysis. A local translator, familiar with PHL practices, was employed in situations where the interviewee did not speak English or spoke limited English.

The majority of the participants were interviewed once. A few, highly knowledgeable employees of the Rockefeller Foundation and its implementation partners were interviewed (or asked questions) more than once; some of these questions were about their own activities, others clarifying factual questions in areas of their expertise (e.g., regarding farming, storage and transportation practices, geographic differences in surrounding areas, etc.). These individuals generally were college-educated, with considerable expertise in their fields. The interviewer(s) took detailed notes during all interviews, which were summarized and circulated to the entire research team for review and to facilitate identification of patterns.

Twelve interviews (12.57 hours of transcribed material) took place prior to the research team arriving in Kenya. These pre-trip interviews provided important contextual information and enabled the research team to validate and refine the initial interview protocol. While in Kenya, 60 interviews (53.78 hours) took place over the course of two weeks in Nairobi and in Embu, Murang’a, and Tharaka counties. An additional 4 interviews (2.32 hours) took place after the two- week trip in Kenya. In total, we conducted 76 interviews (68.67 interview hours).

As this effort is ongoing, additional in-field interviews will be conducted to gain information from individuals not available during the time of the prior fieldwork and/or to seek clarification and a deeper understanding of topics identified previously.

Incomplete List of Lessons Learned

This section identifies and briefly describes ten important lessons learned from the YieldWise intervention efforts in Kenya. It is important to note the phrase, an incomplete list, in the title of the subsection. The lessons noted below are compiled primarily from the research teams initial in-field interviews. Future efforts will integrate these findings with information from additional interviews and from other published sources.

The format of this section is as follows. A one sentence summary of the lesson is first provided, followed by one-word phrases that link the lesson to one or more of the four key components of the YieldWise initiative. Then a brief explanatory paragraph is provided. The 10 lessons are grouped into the following three categories:

· technology implementation,

· technology interaction with other YieldWise components, and

· market interactions with other components.

Technology Implementation

1. Effective implementation of improved technologies at the farmer/first-handler level must provide supporting information, knowledge and training. <Technology>

PHL-reducing technologies must be both technically efficient and cost effective for their use and adoption by small-holder farmers and others in the mango supply chain. However, especially at the farmer and first-handler level, sophisticated and extensive support is needed to speed the adoption process. In the YieldWise project, support personnel provided deep knowledge, not just of the technology, but also of the context and constraints that characterize mango farming. On-farm demonstrations, initiation of farmer groups, and coordination with input suppliers proved effective, in large part because of the implementing team’s ability to generate credibility and trust. Over time, these extensive interactions fueled innovation in terms of cost-reducing modification of technologies and their application to other crops.

2. Although targeting a specific commodity can provide necessary focus for the development intervention, farmers are adept at expanding the scope of the technology’s utilization, if net benefits are available. <Technology>

Throughout the history of agricultural innovation, focus has been an important principal to achieve adoption. Typically done in terms of commodity and region, focused application allows for rapid identification and resolution of issues that impede adoption. Once initial impediments are resolved, however, more innovative farmers are particularly adept at expanding the scope of the technology’s application to similar commodities and circumstances. This process appears to have occurred in a robust fashion in the context of the YieldWise mango intervention. Farmers describe application of YieldWise technologies to crops such as avocados, bananas, passion fruit and tomatoes. In addition, some farmers fabricated less expensive versions of the technologies. It is interesting to note that measurement and evaluation methods typically aren’t designed to capture the effects of this informal expansion of scope.

3. Reliance on recall-based estimates of loss is problematic for large scale interventions <Technology>

Careful analysis of recall-based estimates of post-harvest loss have consistently shown those estimates to be subject to considerable error. When interventions target thousands of SHFs, surveys can be expensive to conduct. Further, analysis of their findings can be limited when survey participants change from year-to-year which often is the case when participants live in rural areas. Because of the time and cost needed to personally survey participants, there is a natural tendency to expand the range of topics for which responses are sought. However, doing so heightens the potential for respondent fatigue to reduce the amount of useful information obtained. Approaches are needed which combine smaller-scale research efforts to validate the linkages between technology use and reduced PHL with large-scale surveys that focus only/primarily on actual use of PHL-reducing technologies.

Technology’s interaction with other components

4. Effective interventions apporach the mango supply chain as an interdependent combination of “market” segments <Technology> <Market linkages> <Aggregation>

Often the issue of excess post-harvest loss is framed as a problem affecting a general commodity sector, such as the “Kenyan mango sector”. In reality, implementation of the YieldWise interventions took place in the context of a variety of market segments. Those segments exist across diverse dimensions; for example whether the eventual use is as processed or fresh product and/or whether the use is for export or domestic consumption. Further, individual varieties can better serve differing needs and the time of harvest varies across production regions. As demonstrated in YieldWise, the use and effectiveness of specific technologies can be differentially affected by these segmentation characteristics. For example, reduction of the impact of fruit flies does have general applicability. However, practices to reduce harvest damage such as use of tarps and harvesting tools have differing value within alternative market segments.

5. Sustained use and expansion of PHL-reducing technologies and practices requires on-going support which can generate entrepreneurial action and economic development in rural areas. <Technology> <Market linkages> <Financing>

Improved technology often provides only a portion of the capabilities needed by farmer-based-organizations and entrepreneurial firms to establish new market outlets. Enhancements to management capability, market intelligence and access to finance also can speed successful implementation. Further, access to such capabilities can identify organizational shortcomings before resources are invested in activities with limited potential for success.

Inspiring entrepreneurial initiatives have been catalyzed through the YieldWise innovation. It is important to note that these efforts have flourished at multiple levels surrounding the farmer segment of the supply chain. Provision of services to support continued farmer use of YieldWise technologies is one example. Aggregation capabilities are being tested both by farmer groups and by individual entrepreneurs. In Kenya, both new and established SMEs are striving to expand mango use and markets.

6. Effective interactions with appropriate government entities can aid implementation of post-harvest loss reducing interventions <Technology> <Market linkages> <Aggregation>

While implementation of post-harvest reducing technologies requires extensive involvement with farmers and related private-sector firms, interaction with and support of government entities also is important. YieldWise efforts have effectively engaged with the government sector. Even though local extension offices often lacked resources, YieldWise implementors established linkages to those entities to assist in reaching farmers and distributing information even if that required use of YieldWise resources. County governments in Kenya often are interested in local economic development. Where those interests were synergistic with actions that could reduce post-harvest loss, early interaction by YieldWise spurred activities to achieve mutual benefit.

Market interactions with other components

7. Fragmentation at the trader/first handler level contributes to supply chain inefficiencies that result in lower returns to producers <Market linkages> <Aggregation>

In developing country agriculture, large price differences between the farm gate and the consumer marketplace have tended to be explained by demonizing the traders and brokers who link those segments of the supply chain. The phrase, “evil middle-man”, is often employed to explain price differences. However, even casual exploration of the costs and risks that traders and brokers indicate that a substantial portion of that price gap can arise from real economic factors. Often these factors can be linked to inadequate physical infrastructure and weak government institutions. While the direct cost of these inadequacies occurs at the trader and broker level, a substantial portion of the resulting economic penalty is borne at the farmer level.

Efforts in YieldWise have made important initial strides to address this issue. Providing catalytic support for the Association of Kenya Mango Traders (AKMT) is illustrative. The AKMT could act to reduce inefficiencies and their impact at this important stage of the supply stage. More work to rigorously determine and document sources of inefficiency at this stage would seem to have strong potential to strengthen the mango supply chain. In addition to identifying means to reduce cost in the current channels, a more efficient supply chain could attract entrepreneurial behavior focused on expanded uses for mangos.

8. Aggregation at the first-sale level offers considerable promise, especially with a primary focus on enhanced supply chain performance <Market linkages> <Aggregation> <Financing>

Supply chains for perishable crops in developing countries, such as for mangos in Kenya, tend to exhibit large price differentials between the farm gate and the consumer. While infrastructure deficiencies contribute to that gap, fragmentation of the actors and lack of reliable information contribute as well because of the effects of risk and uncertainty. Although in early stages, YieldWise efforts indicate that catalytic steps can foster development of distributed aggregation capabilities relatively close to local production. As noted previously, entities which provide effective aggregation can be instituted by farmer groups, entrepreneurs or established firms. While some amount of market power can be obtained, improved outcomes are more likely to be earned as a result of enhanced operational and logistical efficiencies.

9. Especially for perishable commodities, improving the capacity and effectiveness of downstream market channels is an important means to reduce PHL. <Market linkages> <Aggregation> <Financing>

Rejection of mangos that have been damaged from fruit flies or bruising during harvest is directly identifiable as a source of post-harvest loss. However, at a more systemic level, the mismatch between supply and demand (both within and between seasons) may contribute as significantly to excessive loss and reduced economic performance. The absence of information as to likely in-season production levels, as well as insights for future seasons, constrains farmers and first handlers in their ability to make effective operational and tactical decisions. Although commodity markets seldom are in “perfect” balance, the lack of reliable information likely exacerbates boom and bust scenarios. YieldWise efforts to foster farmer-based organizations and aggregation units and to provide information via social media have increased the capacity of the system to disseminate needed market information. However, the overall capability to produce and disseminate such information, which typically is a public-sector activity, seems to be lacking.

10. Expansion of market uses for mangos can act to reduce post-harvest loss, however, access to applied research and development (R and D) capabilities is needed to facilitate the learning and experimentation that supports initiation of novel ventures. <Market linkages><Financing>

Expanding the market uses for mangos also is an important means to create a more effective market channel, with reduced levels of post-harvest loss. This is particularly important in the Kenyan mangos setting where expanded production has occurred and likely could continue. At the processing level, market expansion can occur through additional use of puree in juice and/or development of products which are relatively new to the market (dried mango chips). Export markets offer opportunities, for example, through innovations which lengthen the time between when mangos are harvested and when they no longer can be consumed. For export, application of a wax-type coating to mangos might allow use of ocean freight for transport to the Middle East rather than using air freight. Resulting cost reductions would enhance the competitiveness of Kenyan mangos. Further, creation of “pest-free zones” could open access to European markets.

In addition to entrepreneurial leadership, the capability to conduct applied research targeted to timely market opportunities is urgently needed. YieldWise has fostered important efforts to create such capabilities and to foster linkages between private sector firms with R and D needs and entities which can produce credible results. Private sector entities engaged in the YieldWise- fostered processes range from small farmer-based organizations to established Kenyan SMEs striving to expand their mango-related capabilities.

[1] In developed nations food waste (loss at the retail and consumer level) comprises the dominant component of total food waste and loss. This discussion, however, is focused on loss at the producer and supply chain level in developing nations.