Proposal by Busara for Vanilla Value Chain: Increasing Smallholder Farmer Productivity – Tanzania

Busara

Proposal

Vanilla Value Chain: Increasing Smallholder Farmer Productivity in Tanzania

This proposal has been created by Busara Centre for Behavioral Economics.

SECTION 1: BACKGROUND

Project name: Vanilla Value Chain: Increasing Smallholder Farmer Productivity in Tanzania

Name of submitting organization: Busara Center for Behavioral Economics

Proposed skill contributions:
• Legal and Institutional Due Diligence, Risk Analysis/Policy Analysis

• Stakeholder Mapping and Stakeholder/Community Consultation,

• Community Capacity Building

• Environmental Impact Assessment and Mitigation

• Social Impact Assessment and Mitigation

• Gender Impact Assessment / Gender Analysis

• Outgrower Support and Engagement,

• Monitoring and Evaluation

• Behavioral Economics Research

 

Background:

Who we are 

Busara Center for Behavioral Economics (Busara) is an advisory and research organization focused on advancing and applying behavioral science in the Global South. We work with organizations to build behaviorally-informed solutions to help scale their products, programs, and policies. We work to understand the context of decision-making in new markets, with a strong focus in Sub-Saharan Africa. 

We operate from our permanent offices in Kenya, Ethiopia, Uganda and Tanzania, and work on projects with partners across the continent. Our mission Busara’s mission is to work with researchers and organizations to advance and apply behavioral science in pursuit of poverty alleviation. Our vision Behavioral science is a highly interdisciplinary field that draws on methods and insights from psychology, anthropology, economics, data science and more. 

Our team combines academic rigor with strong private sector experience to build evidence-driven yet pragmatic analyses and solutions. Our 40+ engagement teams are decidedly global to ensure our interventions are leveraging cutting edge behavioral science but also grounded in the context and cultures we work in. Our staff comes to Busara with a diverse set of expertise in health, education, economics, political science, finance, gender, development studies, behavioral science, mathematics, environmental studies, anthropology, and nutrition. Our vision is to build behaviorally-informed solutions to help our partners avoid risks, save money, and drive impact. 

We believe in the importance of behavioral science because while people may feel they have clear intentions, their actions do not always reflect these goals. Behavioral science focuses on understanding these inconsistencies and building models that are more reflective of the human experience. What we do Behavioral economics, though recently popularized, is still a relatively nascent field. Busara is committed to advancing behavioral science in emerging markets, and focuses on three primary barriers to overcome: data, capacity, systems. Our work always starts with removing assumptions and pre-conceived notions about culture and context to look at the world through a fresh perspective. 

We spend a long time listening, learning, designing and discarding potential solutions before we find the ones that may work. Then we test, scale, break, and iterate, again and again until we have it right. Our challenge is that survey data largely focuses on attitudes, preferences, and beliefs, but rarely actions.

New data collection methods and measurement tools are needed. At Busara, we build new data collection tools through incentivized games and vignettes in lab simulations or through embedded tools in ongoing operations: 

• Decision labs: Standing, mobile, and embedded labs are available to accommodate projects of all sizes and scope. We currently have deployed labs in 5 countries: Kenya, Uganda, Nigeria, India and Tanzania. 

• Decision data panels: We use an in-house tool that leverages phone sensors and experience sampling to aggregate behavioral patterns.

• Passive sensing: We utilize mobile phone remote sensing to provide objective, granular data on user digital behaviors to try to understand behavioral trends and patterns. 

• Active sensing: Wearable and environmental sensors detect events in a physical location and provide continuous data to support environmental design and provide broader metrics. 

We develop software, data panels, and models that can be built into existing systems to enable continuous testing and application of behavioral interventions. We build predictive models that help organizations to both predict when certain behaviors will happen, as well as make decisions to intervene and prescribe solutions when triggered. 

Another key service at Busara is that we not only help organizations solve problems, but we also cultivate behavioral expertise within our partners to build a more behaviorally-informed market. We do so through workshops and executive education. 

We offer staff training and capacity building to support organizations in developing in-house capacity to maximum comprehension, implementation speed, and data quality. For examples of projects we worked on, please refer to our case studies on our website:

https://work.busaracenter.org/case-studies.php 

 

SECTION 2: PROJECT APPROACH

Context for the project:

As part of its role as a purchaser and producer of vanilla products, Natural Extracts Industries Ltd (NEI) supports smallholder farmers in the vanilla value chain with trainings and capacity building. Through doing so, NEI aims to drive farmer productivity and output. NEI collects extensive data on farmer productivity and output metrics, and through this has produced preliminary findings that NEI’s trainings positively correlate with farmer performance (measured through the constructed Farmer Health Index), and through this productivity.

This tentatively suggests that NEI’s trainings and interactions with farmers improve farmer performance to the benefit of both the farmers and NEI. However, this cannot be concluded without a more rigorous evaluation of the training. Secondly, there is high heterogeneity in results: the correlations in performance metrics are relatively weak, suggesting a strong effect in some locations, but not in others.

Engagement objectives:

The primary aim of this engagement is to improve the productivity of vanilla farmers through more robust and precise communication of farming methods to them. This aim breaks down into three main objectives:

    • To understand the needs of smallholder vanilla farmers, how they respond to NEI’s trainings, and what they require from these trainings. The first objective in this engagement is to conduct a deep dive into the context of the smallholder vanilla farmers, including modelling their decision processes (i.e. their barriers and drivers), and mapping how NEI’s trainings can and do fit into these. This objective includes segmenting vanilla farmers into groups, based on heterogeneous needs.
    • To design improvements to NEI’s capacity building trainings: Building on a thorough understanding of the smallholder farmers, this engagement will seek to develop effective improvements to NEI’s trainings, focusing on addressing barriers and leveraging drivers in order to promote effective behavioral change.
    • To estimate the effectiveness of NEI’s training programmes. Before rolling out suggested adjustments to NEI’s trainings, this engagement will seek to quantify their effectiveness through a lab in the field approach, by segment.

Busara will seek to meet these engagement objectives through a four-phased approach.

Our approach:

There is a growing body of literature that highlights the deviations of human behavior from rationality, especially when it comes to adoption and sustained usage of new technologies and practices.

The interplay of behavioral biases and heuristics, and environmental and structural factors, such as access to finance, markets, and logistics, can give us novel insights into the problems adoption and ownership of new technologies. Busara adopts this integrated approach of behavioral economics and human centered design to help identify effective ways of promoting sustainable ownership of interventions aimed at increasing productivity in the vanilla value chain.

In the following subsections, we describe the four stakes we propose to inform an effective and evidence-based strategy for increasing productivity among vanilla farmers. In summary, these are:

    1. Initial qualitative research with vanilla farmers to build contextual understanding.
    2. Large sample quantitative research with vanilla farmers to quantify this understanding, and to build segments based on needs and required trainings.
    3. Designing interventions for each segment, focused on the capacity building support provided by NEI.
    4. Lab in the field testing of interventions to provide an evaluation of their effectiveness.

SECTION 3: ACTIVITIES AND DELIVERIES

Activity 1: Phase 1: Foundational Qualitative Research and Decision Process Modelling
Description

Objective: Obtain a high-level understanding of the current beliefs, barriers and behaviors of farmers and key stakeholders around the acceptability and potential uptake of improved vanilla management management technologies methods, and general farmer needs, to inform key areas of exploration for the quantitative survey in phase 2.

Busara will design and run a foundational qualitative study to more deeply understand the nuances of everyday life among the target population – i.e. what they care about, what they are motivated by, what their aspirations are, who they are influenced by, and what discourages them from seeking out interventions such as improved vanilla management technologies and financing options – and the context in which they make farming decisions. This will lead us to key determinants of the acceptability and uptake of improved vanilla management technologies, awareness of the existence of such technologies, accessibility, learning curves, prime decision-makers along the user journey, and other structural and behavioral barriers/levers. 

Through understanding this, Busara can model the farmers decision processes, including exploring barriers and drivers to the uptake of new technologies along the customer journey. Using a qualitative lens means we will be better able to assess how farmers’ articulated preferences and values (acceptability) align with observed practices and behaviors (uptake). This evidence will ultimately inform the design of the quantitative study. 

During this phase, we will also work with the key stakeholders who will serve both as expert informants for the qualitative research and as invaluable partners in developing intervention ideas. The design of the qualitative study will be informed by an initial literature review, which will synthesize the behavioral and structural problems facing the uptake of technologies, as well as global best practices that have been adopted to overcome barriers and drive demand for this kind of intervention. 

Phase 1 deliverables:
    • Findings from literature review, qualitative research and high level-environmental scan.

    • Barrier map with prioritization themes to inform phase 2 research

Duration: 2 – <4 weeks
Deliverable: Report, Presentation

Activity 2: Phase 2: Quantitative Research and Segmentation
Description

Objective: Explore emerging themes from phase 1 to identify and differentiate the end user groups that will be targeted for improved capacity building trainings according to the end user demographics and behavior. Design and prototype interventions targeted to the specific user profiles. 

Busara will conduct a quantitative survey, that will look to understand and quantify themes and behaviors arising from the Phase 1 research. Variables of interest here include but are not limited to demographic variables (i.e. age, education, income, location), behaviors (i.e. awareness of improved vanilla management techniques, self-efficacy, openness to using new vanilla management technologies) and environmental factors (stakeholders and decision-makers, openness to communication, information asymmetry, social norms, social preferences/trust). 

Having this breadth of data will allow us to have a richer understanding of the user profiles across all geographic areas. This survey will be used as the basis of a segmentation exercise, aimed at defining segments of vanilla farmers who would benefit from a different capacity building approach. This segmentation will depend on the findings from phase 1 and the phase 2 quantitative study, but we expect that farmers with different needs, barriers and drivers, and different access to information sources, will benefit from different approaches. 

Busara will explore a variety of methods for segmentation, including manual segmentation (i.e. division of the population by hand on key variables) and unsupervised machine learning approaches (i.e. feeding the data into an algorithm to divide the population into groups based on multiple variables).

Phase 2 deliverables: – Quantitative findings and segmentation report

Duration: 4 weeks – <3 months
Deliverables: Report

Activity 3: Phase 3: Intervention Design and Rapid Prototyping
Description

Objective: Collaboratively design and prototype interventions to improve the capacity-building trainings offered by NEI, targeted to the specific farmer segments.

Having built and understood farmer segments in phase 2, Busara will (with NEI) design effective interventions that can be targeted to each specific profile. These interventions will be focused on improving NEI’s capacity-building trainings, focusing on:

    • Improving and tailoring the content of the trainings, for example embedding behavioral language focused on promoting behavioral change.
    • Improving and tailoring the timing of the trainings.
    • Developing new pathways for training, including the development of ongoing behavioral ‘nudges’
      • SMS messaging
      • video content, like those offered by Digital Green or Mediae (e.g. Shamba Shape-up)
      • physical tools to improve adherence to schedules
      • changes in incentivisation techniques

These potential interventions are left deliberately broad, as they will depend on the findings from phases 1 and 2. Busara proposes to develop these interventions through a collaborative design workshop with NEI, bringing together Busara’s behavioral knowledge and phase 1&2 research, and NEI’s deep contextual and technical expertise.

As part of their development, these interventions will be prototyped and taken to farmers for their feedback. Overall, we will adopt a human centered design approach, consisting of iterated prototype development and feedback loops through focus group discussions with a sample of user groups.

Busara expects to conduct 2-3 rounds of iterations to develop final prototypes, which will be low to medium fidelity interventions for NEI to put into practice.

Phase 3 deliverables:

    • Design workshop.
    • Final, prototyped interventions for testing.

Duration: 4 weeks – <3 months
Deliverables: Presentation, Workshop or meeting

  •  

Activity 4: Phase 4: Testing and Recommendations for Scale-up
Description

Objective: Test the final set of interventions using Busara’s lab in the field approach to arrive at tangible, successful strategies that can be used by NEI and to drive uptake and sustained ownership of improved vanilla management techniques among target populations.

In the final work stream, we will test the interventions developed in phase 3 in order to provide evidence of their effectiveness prior to rolling out.

Testing is crucial to ensuring any intervention has been validated with the target audience. To provide an accurate estimate of the effect of the interventions on farmers’ behavioral mechanisms, and to avoid a large scale and expensive field experiment, Busara proposes testing the interventions in a controlled lab environment using Busara’s lab in the field approach.

This approach is a useful, light way to test communications in a controlled setting with the target audience. The goal will be to rapidly identify low-cost yet high-impact interventions that overcome behavioral and/or structural barriers, driving the sustained uptake of the improved vanilla management techniques.

Phase 4 deliverables:

    • Pre-analysis plan to measure outcomes of interest from the testing phase
    • Preliminary findings from the lab in the field testing – Recommendations report with strategy for scaling interventions
    • Delivery of recommendations through VCoEs

Duration: 4 weeks – <3 months

Deliverables: Report, Presentation, Workshop or meeting

  •  
  • Project measureables:

We design a very customized approach to monitoring, evaluation and reporting for each of our project. We therefore use different tools along the process based on the activities undertaken. Please see below suggested monitoring and evaluation tools we envision for each phase of the project, but we are very open to change or add elements of our approach.

Phase 1: During the qualitative phase, we will conduct interviews with key stakeholders. We will monitor and report on number of interviews, evaluate the quality of interviews through audio recording and checks on transcripts.

Phase 2: During the quantitative phase, we will monitor survey progress through the number of interviews completed. We will ensure quality of the interviews using several data collected through the tablets used for the interviews and check key variables at the level of interviewers. Progress can be reported on a weekly basis.

Phase 3: During the co-design phase, the main aspect to monitor is the number of interventions we will be designing and which will be approved by NEI. Focus group will be closely monitored with number of participants, transcripts verification and audio recording.

Phase 4: Once interventions have been designed and before testing these, we will prioritize them. In order to move from the list of intervention ideas and low-fidelity prototypes to a fully realized intervention design, Busara utilizes a comprehensive selection process.

Each low-fidelity prototype will be scored according to:

    • Impact: share of the target audience covered, change in behavior, likelihood of driving uptake, end-user interaction, and relevance of the barrier being targeted
    • Feasibility: resources, scalability, cost of implementation, cost of target group, timeline (internal), testable in lab (internal), replicability, practicality, and accessability
    • Social Acceptability: buy-in from other stakeholders, expressed appeal/desirability from target group, and appropriateness

We will also consider how the intervention might be integrated into existing government programs and/or have demonstrated efficacy and results at scale to make a case for government adoption. This classification and indicators will be monitored and adjusted during the last testing phase.

Finally, an output of the project will be this landscape analysis with barrier/levers of contextual factors mediating farmers’ awareness, access, and uptake of vanilla management techniques/services. This will be helpful tools to continue monitoring the success of the adopted interventions beyond the scope of the project.

Methods of data collection:

Interviews

Focus Group Discussions

Technology

Desk Review

 

Project Risks:

During the project cycle we shall use the following project management process to identify and manage risks in the proposed project:

    • Communicate and consult with partners throughout the project.
    • Establish the context for project risk management e.g. policies, roles
    • Identify risks events and their causes
    • Analyze risks – i.e. consequence and likelihood of each risk event.
    • Evaluate risks – prioritization of risk events for management.
    • Treat risks – i.e. implementation of strategies to manage risk events

Monitor and review the effectiveness of the project risk management process.

Risk management strategies:

Some of the risks anticipated and how to manage them:

    1. Research Achievability & Integrity:
      • Research has low chance of achieving research outcomes. Mitigation strategy: Proposed research has been carefully designed to meet its objectives.
      • Misconduct or fraudulent behavior. Mitigation strategy: The project staff are highly trained and professional. Staff are reminded to sign NDA for every project they handle
      • Researchers make unjustified claims not supported by data. Mitigation strategy: Staff have undergone Ethics training and trained on data and research misconduct.
    1. Research Methods & Process
      • Research undertaken without approval of the research design. Mitigation Strategy: All study protocols and designs are submitted to ethics review committee which approve the study
      • Research methods outside of approved research design. Mitigation Strategy: All changes are submitted to Ethics review committees for approval before study continuation
      • Data is not secured, lost or unverified. Mitigation Strategy: Staff have undergone data training and back process has been in place to retrieve all data that may get lost. All devices, laptops and servers are password protected and only accessed by authorised senior staff.
      • Research methods are not adequately documented. Mitigation Strategy: Ethics Review Board do monitoring visits and Research Compliance Officer has been mandated to be in charge of project documentation.
    1. 3. Ethics
      • Physical or psychological harm to research subjects. Mitigation Strategy: These are reported as social harms to the Institutional Review Board
      • Conflict of interests are undeclared. Mitigation Strategy: Staff have undergone Conflict of interest online course and sign forms for future reference
      • Researchers fail to obtain consent. Mitigation Strategy: Staff on the project will be trained and importance of consenting process and auditing process will be done for verification that all study participants were consented.
      • Researched sample is disadvantaged by the research. Mitigation Strategy: All adverse events during study implementation is reported to Ethics board
      • Research methods are not adequately documented. Mitigation Strategy: Dedicated staff members will be mandated to be in charge of project documentation and reporting.
      • Research sample change their behaviors as they know they are being researched. Mitigation Strategy: All adverse events during study implementation is reported to Ethics board.

SECTION 4: SKILLS AND TEAM

Relevant past experience:

Busara has both extensive experience applying behavioral economics to the agricultural field in Tanzania and using behavioral insights for designing customized learning tools. Please see below a few examples of relevant past projects:

    • Mercy Corps AFA, a mastercard funded program that sought to offer financial and technical support to innovative organisations that created and were implementing cutting edge products and service at the intersection of agriculture and finance. AFA engaged Busara as their primary learning partner for 2 years to work directly with their partners while also producing learning outputs that ultimately would be shared with other players in the ecosystem. Engagements varied from running data analytics to inform strategic recommendations, executing qualitative exercises to tease out behavioral elements that could fit into to capturing learnings and supporting the development of learning agenda during their yearly learning partner events.
    • Agricultural Transformation Agency (ATA) / Precision Agriculture Development (PAD): PAD partnered with Busara in Ethiopia in order to conduct a two part diagnostic study consisting of:
    1. international benchmarking of the ATA 8028 hotline to generate actionable insights to aid expansion;
    2. In-depth interviews and focus group discussions with farmers to identify additional services to be developed on the ATA System.

Similar to this project, Busara led both qualitative and quantitative surveys with smallholder farmers around the country and could extract meaningful conclusions and recommendations thanks to the integration of behavioral insights.

    • Financial Sector Deepening Trust in Tanzania: Busara was tasked to analyze and make recommendations on designing a target operating model to drive digital financial inclusion at scale in several value systems: maize, transportation and storage and wholesale & retail.To do this, Busara conducted:
    1. a literature review and key expert interviews aimed at understanding the context of each value system
    2. a qualitative deep dive into these value chains to understand the lives of the primary actors in each and
    3. a quantitative deep dive to identify the prevalence of different behaviors and opportunities.

Together, this stage details the lives, needs, barriers and connections of the primary actors in each value system. We also have a strong experience in developing and conducting trainings, here are a few examples of past projects for your references that we can further discuss:

    1. Trained MFI leadership, managers, and agents on improved service provision and customer outreach with ICCO in Ethiopia.
    2. Trained both government veterinarian and commercial sales agents on the best ways to communicate the benefit of improved breed chickens in Ethiopia with EthioChicken, with Acumen, funded by Gates Foundation
    3. Trained sales agents on methods to improve uptake of financial services in rural Tanzania with Access Bank Tanzania in 2018.

Key staff experience:

Our core team for this project is composed of one Project Lead, 3 Associates with various expertise needed for the project and one Consortium Coordinator.

The team speaks both English and Swahili, has strong experience conducting both qualitative and quantitative research and has extensive expertise in the local context:

Nathanial Peterson, VP Partnerships, will be the Team Leader for the project. He has extensive experience leading projects in the agriculture sector. His primary interest is in how people perceive and manage risk, so he is especially drawn to projects involving insurance uptake and credit for farm and business investment.

His current portfolio includes behavioral and psychometric segmentation to promote financial inclusion, understanding how SMEs can overcome behavioral supply chain issues, and communicating improved agricultural practices. He has also led several projects developing and conducting trainings for various types of public and in various contexts in Tanzania and abroad.

Gideon Too, Senior Associate, has been at Busara since 2016. During this time, he has led the design and execution of multiple large-scale behavioral research and advisory projects across a range of sectors and countries in Africa. As a Project Manager for this work, he will be in charge of leading the development and delivery of project deliverables, including client presentations and stakeholder workshops. He will coordinate the different phases activities and research teams associated.

Salim Kombo, Research Associate, joined Busara in 2017 and has worked on various research projects applying Behavioral Economics in the energy, agriculture, environmental conservation, and governance fields. Salim will bring his extensive experience in experimental design, qualitative and quantitative survey design, alternative data collection tools and data analysis. A recent project he worked on was for Financial System Deepening Trust about ecosystem design and development approach research for financial inclusion in the maize transportation & storage and wholesale & retail value systems in Tanzania.

Rosa Tesfay, Associate, has worked for Busara since 2016. Prior to joining Busara, she worked at the International Trade Center, Geneva, conducting assessments on Aid for Trade initiatives and developing project ideas to enable SMEs to participate in and benefit from global value chains, and supported in managing projects in Tanzania and Lesotho. Prior to her studies, she led community development, women empowerment, SMEs’ capacity building and entrepreneurial skill development projects in Ethiopia. Since working at Busara, she has led several projects in the agricultural sector, working on behavioral insights from smallholder farmers. She currently leads the designing and testing of market linkages and technical support systems for smallholder farmers in Ethiopia, and leads a project with an overall aim of identifying support systems for farmers via input linkage, technical support and market linkage services for smallholder farmers in Kenya.

Gladys Muange, Consortium Coordinator, joined Busara in 2013 as a Field Officer and has risen through the ranks of Senior Field Officer to Project Lead and currently Senior Project Lead as from November 2016. She has led several advisory studies on Humanitarian, Health Behaviors, Financial Inclusion and Agriculture. She has strong experience in managing large-scale field and lab-based research.

Ann Wanjiku, Senior Research Analyst, has worked at Busara on a wide range of research projects, especially in Tanzania, since she joined the team in 2016. She recently worked on a project for CGAP and TIRA where she developed research tools, led data collection and analyzed qualitative data. She is an expert in all the research tools presented in this proposal, especially in the Tanzanian context.

SECTION 5: FEES, REPORTING AND FOLLOW ON

Currency Selection: USD

Upper fee limit: $99,108

Payment structure: Milestone/deliverable based

Reporting processes: We can customize our monitoring and evaluation tools to the needs of the project. Based on further discussions on the details of the projects, we can offer further monitoring and evaluation tools and adjust the reporting frequency to the need of NEI.

Desired frequency of reporting: Monthly

Interest in follow on work: Yes. We would be very keen to understand farmer decision-making processes and reaction to longer term trends in adaptation or usage of NEI’s recommendations. When there are new value chain supports put into place, much interest can be generated in taking part after which, perhaps, a few high yielding farmers emerge as more successful which may create new, possibly unintended beliefs about the causal mechanisms behind production. Producing mental models of causality of yield over time could be highly instructive. Additionally, long-term perceptions of fairness of NEI practices, preferred farmers, those included in pilots, etc. would be very useful to understand. Finally, we may find that different types of learning are necessary for farmers or that a particular tool may be surprisingly useful to them and worth training with others. We’ve seen phenomena with some simple inventory management tools used by MSMEs and in other contexts as well. It would also be great to know if farmers are demanding some specific tool.

SECTION 6: ADDITIONAL FILES

CVs
1. Nathanial Peterson CV

Workplan and budget
1. Workplan

2. Budget