Lesson 1.2 Farmer Level Data and Farmer Profiling
Learning outcomes
At the end of this lesson, you will (be able to):
Understand the importance for farmer groups and cooperatives to profile their members
Identify the core components and the partners required to organize a profiling project
Understand the challenges related to farmer profiling and the strategy to address them
Design a business model to sustain the profiling of the members
1. Introduction
In order to increase their production and increase their income, smallholder farmers need various types of services such as extension services, financial services or trade services. As presented in Lesson 1.1, the design, deployment and delivery of these services in their physical or ICT format require the mash-up of data at the farm level with data from the global context and available through open datasets released by national, regional or international organizations. This lesson presents a practical and operational recipe for actors such as farmers’ organizations, cooperatives, agribusinesses or governments to build a digital farmer profiling platform with the aim of aggregating a series of farm-level and farmers’ data to ease the delivery of added-value services.
2. Digital Farmer Profiling: An introduction
As highlighted in Lesson 1.1, the delivery of tailored, actionable services to farmers require the mash-up of farm-level data with global information (e.g. weather forecasts, etc.). These services are usually designated under the term ‘precision agriculture’ and they have been developed and deployed all over the world for more than a decade. These services take different forms. In countries, usually developed countries, where farmers have capacity, infrastructure and equipment, they can themselves manage their own farm data that they collect either directly or via sensors. They are then able to select the services they need and share their own data with third parties who are providing those services. In most developing countries, the situation is very different. Farmers lack capacity, infrastructure, and equipment. They are not yet in a position to collect and manage their own data and then interact directly with service providers. As a result, different service providers directly collect different types of data on the farm. This situation presents a series of challenges such as:
Lack of sustainability for service-providers: At the moment each and every service provider has to put in place a data collection and update process. Such processes are extremely costly. As a result, the services provided are expensive for farmers and hardly sustainable.
Farmers’ rights are not respected: At the moment, farmers are not really aware of the use of their data and individually cannot really protect their interest. In practice, they lose ownership of their own data, and at the same time, this data is in the hands of service providers and not necessarily used to maximize farmers’ benefits.
Service provider lock-in: As service providers are putting in place an end-to-end service, it is difficult for farmers to switch to another provider. For example, if a financial service institution collects and builds a database of information about farmers to compute their credit scoring and their eligibility, this information (e.g. repayment rates or credit values) is stored at the institution and will not support the farmer who applies to other financial institutions for credit.
To address these issues, a new approach is now emerging: The Farmer Digital Profiling (FDP) platform. The concept is for an organization to aggregate all the profile information of the farmers under its umbrella and then leverage this information to support the development of services. Figure 1 illustrates the concept:
Organizations engaged in this process include farmers’ organizations (e.g. FEPAB in Burkina Faso), cooperatives (Igara Tea Factory in Uganda, Nucafé in Uganda), agribusinesses managing a series of farmers under contract (e.g. Meridian in Malawi), or even a government agency building a national repository (e.g. Fiji Crop Livestock Council in Fiji, Rwanda Ministry of Agriculture).
These different types of organizations can take advantage of a deep understanding and knowledge of their members and in particular who they are, what they do, where they live, what they produce, etc. This information is essential for many reasons:
Planning & Strategy: Organizations can plan their services, their intervention and their areas of investment based on real data. They will be able to identify areas where they could expand, or places where there are specific opportunities in terms of production or selling. A deep understanding of their member allows organizations to define their roadmap and identify new services. It also allows them to make financial forecasts and evaluate with high precision potential markets for various services. In short, this information will allow organizations to plan their activities based on real facts and data.
Easier membership management: For farmer organizations, cooperatives or similar organizations, the use of a Farmer Digital Profiling (FDP) platform helps in the management of membership for all the internal activities such as payments, elections and votes. The use of such platforms helps them to save both time and money in those tasks.
Easier communication: The capture of communication details, in particular phone numbers, allows the use of communication platforms that automate the sending of information in various formats (voice, SMS, etc.). The use of such platforms supports a better and more regular communication between the central organization and its members. The use of new communication channels enables organizations to
better understand the needs and demands of their members
better understand constraints and pain points
query their members on specific topics and get their feedback.
Greater opportunities to identify and put in place new services: A better knowledge and understanding of its membership enables an organization to identify new targeted value-added services. These services not only provide new benefits to the members, but also increase the value of membership and enable farmers’ organizations (FOs) & cooperatives to recruit new members. Among potential new services, ICT services developed by third-party suppliers have a specific place. The availability of a maintained FDP platform is critical for ICT service providers. It saves them huge costs, making services more affordable and at the same time sustainable. To reach this goal, those services must be designed jointly between cooperatives & FOs, service providers and farmers
Greater power in advocacy: A deep knowledge of its membership allows organizations to have a stronger voice in advocacy. At a basic level, an organization with an FDP platform can prove their membership and demonstrate the number of people it is representing, who they are and where they are. This gives power to their voice based on their representativeness. At a more advanced level, an organization can exploit its membership to inform policy makers in various ways:
Simulating the impact of proposed measures: based on farmers’ information, it will be possible for an organization to measure or simulate the impact of new measures (e.g. a new subsidy scheme) and define its position with regards to proposed measures based on real data.
Quickly executing surveys to get members’ positions on specific topics: Organizations can mobilize their members and collect their opinions on specific topics. Here again, such a process will help an organization to define its position and defend it based on real data and based on a clear mandate from their members.
New sources of revenue: the farmers’ profiles are potential sources of revenue for third-party activities such as
research
market surveys
advertisement.
However, it is important to note, that farmers must agree with the use of their data and their participation in such activities. These revenue streams are presented in detail in the last section of this lesson.
In short, an FDP platform is both an opportunity for FOs, cooperatives and similar organizations managing a group of farmers, and an enabler for new farmer-centric services that are critical for increasing production, decreasing loss, and maximizing income for smallholder farmers. It is also a way to protect farmers’ rights and ensure that they benefit from the disclosure of their farm-level data.
The following video, https://vimeo.com/263101646, about a farmer profiling project in a tea factory in Uganda illustrates the value of an FDP platform for a cooperative.
While an FDP platform is an important enabler for the design and delivery of services, it fits in a larger context, where other enablers could also provide important impact. Figure 2 below provides a global view of the different components.
As the diagram shows, other enablers include an identification service to uniquely identify individuals (e.g. India’s Aadhaar system), a communication service to ease the collection and delivery of information (via various means such as USSD services, SMS services or voice-based services) or a payment service (e.g. mobile money) to ease transfer of money between parties (lending, payment for information service, etc.). However, these enablers are not required to exploit the full potential of a farmer profiling platform.
Finally, it is important to note that an FDP platform is an enabler for both public services (e.g. planning, subsidiary scheme design, extension services) and private services (e.g. financial loans). Apart from information services, policymakers are also potential consumers of data stored in profile information. The profiling platform can provide raw content to compute key policy indicators (e.g. land planted, size of land under irrigation, etc.). In the same way, the profiling platform can be used to forecast the impact of policy interventions.
However, while all these applications can potentially be enabled by a profiling platform, the success of such platform and its ability to deliver expected results depends on several elements including the implementation context; the content of the profiles; the quality, timeliness and completeness of data stored; and the usability, reliability and effectiveness of the platform from a technical point of view. This lesson will introduce these different aspects below and present some best practices to build an FDP platform.
3. Components of a farmer profiling project
This section includes five subsections that introduce the different components or dimensions that must be considered to ensure the development of a farmer profiling project in order to maximize the chances of its success.
3.1. National context
In the case of a national-level profiling project, before engaging with the design and development of the FDP platform, it is important to explore and review other initiatives taking place in the country in order to identify existing public or private databases that could be linked to the profiling platform. This includes identity databases (e.g. Aadhaar system in India), land databases, subsidiaries databases, etc.
3.2. Legal and policy context
A farmer profiling platform collects and stores data about farmers and farms that are, by their nature, classified as personal data. In many countries, the collection, storage and management of personal data is regulated by specific legislation. It is therefore essential to review the legislation and to capture key requirements such as official declarations, data sharing rights, obtaining farmer consent, etc., that have to be considered for the building of the platform and to ensure that the project does not breach any regulations. See Lesson 2.4 for details about this aspect.
In the same way, other legislation related to open data or official statistics may also have an impact on the ability to publish anonymized aggregated datasets, particularly if the platform is under the authority of a public agency. It is therefore essential to identify the legal constraints under which the technical platform is to be developed if it is to be compliant with national legislation. International and continental treaties, agreements and policies also need to be considered.
3.3. Operational dimensions
The success of a farmer profiling project largely depends on the operationalization of the data collection tasks. There are two phases to consider:
The setup phase: the profile information will be collected for the first time.
The operational phase: the post-setup phase when the profiles are updated regularly.
Setup phase
The success of the setup phase relies on a series of factors:
Engagement with targeted farmers: The collection of personal details of farmer and farm is something that is not as easy as it first seems. The literature shows that farmers are reluctant, if not opposed, to provide their details if someone shows up at their place. In order to support this task, and to ensure a faithful contribution by farmers, a series of activities have to be organized in advance.
Awareness raising campaign: Meetings with farmers as well as radio spots have to be organized to explain the concept, the process, the rationale, and the potential benefits for farmers. These campaigns should also explain in detail the information that will be collected and why.
Data sharing agreement: Farmers are usually not willing to share their data if they are not clear about who is going to use the data. From an ethical perspective, sometimes a legal perspective, and an operational perspective, it is important to present a data sharing agreement to farmers. Note that the data sharing agreement is the result of a careful analysis of the personal data protection regulatory context analysis. See Lesson 2.4 for more details on that matter and on the tasks that have to be implemented at this stage.
Increase trust in data collectors: Farmers need to be sure that those coming to their farms asking for data are authorized data collectors. It is recommended that enumerators are equipped with a professional card and an easily recognizable item of clothing such as a jacket or hat. These should be promoted during awareness raising campaigns. Other elements that increase trust and ease the work of the data collector include an announcement of the timing of the visit or the introduction of the collector to the farmers by a trusted person (extension agent, cooperative/farmer group leader, etc.).
Training of enumerators: The training of enumerators, not only on the technical platform, but more importantly on the data collection process is essential for the success of the task. In particular, the training must include awareness raising on data security and the protection of their equipment, confidentiality of data collected and presentation of the data sharing agreement and capture of farmers’ agreement. It should also include best practices on the sending of profiles to the central platform for review. It is recommended to design a charter that data collectors must sign that includes all elements of their tasks, as well as potential legal risks if they breach some of the legal requirements (confidentiality, etc.).
Provision of robust equipment: the data collector’s equipment is a critical element for the process. This equipment has therefore to be reliable. Different elements should be considered:
Power: equipping a data collector with one or two power banks to ensure that they can conduct a full day of collection without energy issues. Depending on the location, organizing how data collectors can recharge their power banks and equipment should also be organized
Tablet: The robustness of the tablets and the quality of their components, in particular the onboard GPS chip, are essential for the quality of the information collected.
Memory: depending on the profile content, it is safe to plan extra memory on the tablet to ensure that all content can fit.
Funding of data collectors: it is essential that data collectors have enough funding for travel, and communication (calling for support, transmitting profile data). An inadequate support can have a major effect in timing.
Monitoring: It is essential to monitor the quality of the profile collected in almost real time, particularly at the beginning of the process, to ensure that each and every data collector has well understood the tasks. When problems are detected on profiles, the information must be communicated to the collector so that he completes the problematic profiles. One efficient way to ensure the quality of the profile is to put in place a payment per profile collected and validated.
Support: Data collectors will always encounter problems they are unable to fix or will have questions that need to be answered for them to execute their tasks. It is critical to put in place a support mechanism (e.g. a hotline) to allow them to access the necessary support in the field when they need it.
The gender dimension is an additional factor to consider in a farmer profiling project. The gender aspect is sometimes critical and may dramatically influence the quality of the data collection and the level of contribution by the farmer. For example, male data collectors interviewing female farmers or vice versa may impact of the level of engagement and the quality of the data provided. In some countries, the issue of gender is not critical, but in others it poses a major risk to data collection. Gender aspects have to be integrated at different levels. This includes, in particular:
The gender of the data collector (ensuring that female enumerators are used to collect data from women farmers) and the language enumerators speak
The time and language of the awareness raising campaign at radio station. The person (e.g. the voice used in radio spot) is also important.
The time (day and time) at which data collection is organized.
Operational phase
Organizing one-off collection of data (for example, by means of a census) is relatively easy, but putting in place a system that will ensure that profile data are updated regularly, is more challenging. There are different models that could be explored, and different potential organizational and institutional arrangements that should be considered. The options are presented below.
Centralized versus decentralized model: There are different ways of organizing an FDP platform. One way to do it is by using a centralized model, where a given organization decides to build the platform and map all farmers. This model fits well for any local organization in direct contact with their farmers (a local farmers organization, an agribusiness, a cooperative, etc.). Another way to do it is through a decentralized model, where the profiles are provided by a series of organizations. This model fits well for nationwide FDP platforms or for decentralized organizations (e.g. a group of cooperatives). In such a setup, it is then more efficient to rely on an individual cooperative or farmer group to map their own members. The decentralized model is usually more efficient, for two main reasons: (1) the mapping organization is in direct contact with farmers, and a trust relationship already exists; (2) the mapping organization can use profile information for their activity, creating an important incentive to do a good job. However, such a model is possible only when farmers are well organized, and each entity has enough capacity to conduct the tasks and exploit the profile.
Specific task versus part of other tasks: Usually the setup phase, as presented above, is run like a census. It is possible to organize updates of profiles in the same way with annual or seasonal profile update tasks. However, this model presents two challenges: (1) such a task requires a similar mobilization of funds as the launch phase, and it is therefore costly; (2) the data are updated once a year at best, usually every 3 to 5 years and this may not be appropriate for specific tasks or value chains. For example, for tea growers, coffee growers or coconut producers the profile information will not vary much over the years; for rain-fed crops, information must be updated at least once a year, and usually twice or three times. The second option is to give the task of updating profiles to people who are regularly visiting farmers. Such a model allows a far more flexible data update and does not require major funding. Such a model is easily implementable in countries where there is a strong extension agents’ network and where the tasks can be added to their job description, or in the case of a decentralized model presented above where mapping organizations are closely linked to farmers. The best approach is usually a mixed approach where a full census is run every 3 to 5 years, and a seasonal update is in place but focusing on only a limited number of types of information and using more basic technologies.
Farmer-led process versus organization-led process: Updating profile information can be conducted as a top-down task with a mapping organization collecting information as presented above. Alternatively, farmers can update their own profiles[1]. This model requires two main elements to be successful. The first is that different channels must be put in place to enable all farmers to update their profiles. This may include a call centre as many farmers in developing countries will not be able to use smartphone applications, SMS/USSD services or even voice-based (also known as Interactive Voice Response or IVR) services. A farmer-led model also requires regular communication campaigns to remind farmers to update their profiles. The second and more important element is the incentive that rewards farmers for updating their profiles. Farmers will spend time updating their profiles if there is a direct benefit. Incentives can assume various forms, including free extension services, financial rewards, access to a subsidy scheme, etc. The incentive plan is the cornerstone of the farmer-led model.
It is important to note that the selection of specific options for data collection has a major impact on the technical platform, its functionalities and the infrastructure to put in place (setting up a call centre, IVR services, etc.). It also impacts on the overall organization of the project (who to train, how to train them, etc.) and on the overall budget. It is therefore essential that these elements are selected before technical choices are made.
3.4 Technical dimension
A farmer profiling platform is by definition an ICT platform and the technical and technological dimension is therefore an important component to consider. A profiling platform is made of three main elements:
The data collection module: the data collection module is usually in the form of a tablet application. In most countries, internet connectivity in rural areas is not reliable enough to opt for a connected application, and an offline application should be preferred. The application has to have a series of functionalities that include (but are not limited to):
Ability to synchronize the profile information when connectivity is available.
Security: the tablet and/or the application have to be secured to protect data:
the tablet should have antivirus protection
the tablet should be localizable and erasable remotely when it connects to the internet
the tablet should be protected and dedicated to data collection to avoid collectors sharing with others for e.g. pictures, games, etc. that risk allowing others to access confidential data
the application has to have an authentication feature to ensure that data is accessible to authenticated data collectors only.
Flexibility to adapt to profile changes: the content of the profile will likely evolve over time and the application should be offer such flexibility
Geolocation: the application will most likely have to capture GPS coordinates (farms, fields, etc.) and it should therefore be able to handle an onboard GPS chip.
Validation: the application should support the data collector and detect at the time of entry any typos or errors by conducting a data check.
Update: the application has to offer a way to review and update existing profiles (operational phase). The application and the central profile repository have to synchronize their profiles, and specific data collectors have to have access to specific profiles assigned to them.
The central profile repository: the central repository is the heart of the platform. Key functionalities or elements to consider include:
Agility: The platform should enable an easy update of profile template as the profile content will likely evolve over time as new applications emerge that require specific data to be collected.
Robustness: The platform and underlying infrastructure should be designed based on the targeted number of profiles, the targeted number of data collectors and the likely number of simultaneous connections from third-party applications.
Security: The security of the central platform relies on different elements:
Software security: the platform must be up to date with regards to operating system, and have an updated antivirus
Protection: the platform must be hosted in a data centre that provides protection related to power, physical intrusion, temperature, water, etc.
Backups: backups must be conducted at least on a daily basis and content has to be stored securely at different physical sites
Access: the platform should offer the latest technology for secured access (in particular the use of encrypted password from end-to-end) and should have a fine-grain authorization model that will offer a flexible framework for different categories of users. In particular, the authorization model should at least work at the individual profile level (providing access to specific profiles) and at the profile content level (providing access to specific parts of profiles).
Validation: The platform should have a series of built-in checks that will highlight potential inconsistencies on stored profiles.
Analytics: The platform should offer a series of analytics to visualize different elements such as the activities of data collectors, the content of the database (e.g. to drive the extension or prioritize tasks) or the age of the information. These analytics are important to support and monitor the data collection tasks, and to evaluate the quality and completeness of the information stored.
Technologies and standards: The evaluation of the technical platform should also cover the technologies and standards selected for the platform to ensure that the code could be easily maintain and updated over time, and to avoid any vendor lock-in.
The data access/publication module: As the role of the profiling platform is to enable other ICT services, the platform should offer a series of access channels. This should include at least a web interface for human access, an API for software access, and potentially an export of open data to provide anonymized statistics and content for policymakers and other stakeholders. Depending on the existence of a national open data policy, the platform may include a functionality to publish datasets automatically on a national open data portal.
The exact set of functionalities largely depends on the operational setup and, in particular, the model selected for the running phase. The most common approach includes the following elements:
A smartphone-based data collection tool for the first data collection. The reference free and open source tool for that purpose is Open Data Kit (ODK).
A central repository developed as a web platform: For this part, there isn’t a reference free and open source package. There are different modules that could be used but would need to be integrated. If there is no geospatial data, packages such as ODK collect (part of ODK solution), ONA or kobo toolbox are potentially interesting options. A comparison of the different tools used in August 2017 is available. If geospatial data are included, the reference free and open source geographic information system is QGIS.
A service for information update: As presented before, such a module can be implemented in different ways such as a USSD service, a voice-based service or a call centre. There is no reference package for such services, and they require ad hoc development. If there is a large agent network on the ground and if they are equipped with smartphones, the same tool as the one for first collection should be used.
An example of a complete architecture is presented in an article entitled “Farmer Registration and Profiling: How Did it Go?” published by CTA. This architecture has been used in a series of projects in Africa.
4. Profile content
The previous sub-section presented the core elements to successfully conduct farmer profiling tasks. However, as presented in the introduction, a farmer profiling platform is not a goal in itself but is an enabler that supports and eases the delivery of targeted information and other services to farmers. An FDP platform can support those services depending on the information stored in each profile. In that regard, it is important to note two points:
A profile cannot be exhaustive and capture everything about a farmer and his or her farm. Some crops require specific data; livestock and crop farming are, for example, very different both in terms of the data available and in terms of the interests of service providers. The content of a profile is therefore usually guided by how the profile data will be used.
There is a trade-off between, on the one hand, collecting a vast amount of data that makes the collection process longer, more expensive, more difficult and will likely create more resistance from the farmers, and, on the other hand, the collection of basic data that are easy to capture, may not vary much over time, but may not allow the development of advanced information services.
Based on these elements, it is critical to define the content of profiles based on stakeholders’ perspectives and plans. It is not possible to be exhaustive, nor it is possible to identify all possible applications[2], but the success of the project will rely on a bootstrapping phase that will demonstrate how such a platform can enable useful services for farmers, who, in turn, become more open to providing additional data to access more services. The best practice to define the profile content can be summarized in the following steps:
The organization in charge of the FDP platform should identify the set of services they want to implement and that represent the underlying rationale for setting it up.
For each service (e.g. access to credit, trade, etc.), the organization should organize a workshop with other parties involved in the service (e.g. micro-finance institutions) to identify the set of information that would need to be collected.
This process will ease the design of the profile content.
Apart from the content of a profile, another critical element is the identification of the farmer. An FDP platform stores a large set of profiles and it is therefore important to know how to retrieve the specific record attached to a farmer. Some countries implement a national scheme for personal identification (e.g. ID cards in European countries and the Aadhaar system in India). Such schemes can easily be used as the index in the database of profiles. However, in most developing countries, there is no reliable ID number or other unique identifier to identify a specific farmer. In such a case, it is critical to understand the element of information that could uniquely identify a person. In some countries, first and last names are unreliable identifiers, so too are farm addresses. The use of biometrics such as fingerprints is relatively difficult to put in place and presents several challenges, outside the cost dimension. It usually requires a series of elements such as name, address related to a specific point of interest (school, health centre...), phone number, etc. that are useful to identify the person. Note that a picture is a potentially useful element to verify the profile information, but it is useless as a search criterion. The design of the identification process is an extremely important element that should be explored at the early days of the setup to avoid duplicate records and splits of information between multiple records. In the absence of a robust identification scheme, different use-cases have to be considered:
Identification of the person for update of his/her profile in a face-to-face interaction.
Identification of the person without face-to-face interaction (via a call, or via a USSD service). In that case, the technology may also have a role to play (e.g. whether the technology used allows the capture of GPS coordinates or not).
Finally, the profile is very valuable information for farmers, and each profile records individual farmer data. Outside the services provided by the organization setting up the FDP platform, a farmer may be interested in using his profile information for other purposes. At the same time, the ability to see the profile information depends on an element of trust between the organization and the farmer. It is therefore highly recommended to include the design and delivery of a paper-based profile to each farmer.
An example of such dossier provided to farmers is implemented by Igara Tea Factory (Uganda) in their profiling initiative.
5. Farmer profiling platform business model
The sustainability of an FDP platform is a critical element to consider at the early days of the project. There are three elements to consider as part of the business model:
The platform development and operational business model: whether the organization who want to setup the FDP platform will cover the costs of development and operation of the platform and collect all revenue, or whether a PPP model should be put in place?
The revenue streams: who are potential customers of the FDP platform?
The operational costs: what are recurring costs?
Concerning the operational costs, they depend on the data collection and update model that is selected. The different elements of the costs include at least:
the hosting of the FDP platform
the hosting of the update services
the incentives for the different stakeholders (farmers, intermediaries in charge of updating profile information, etc.)
the depreciation/renewal of equipment
the communication costs between various stakeholders
the staff in charge of monitoring the platform
5.1 Public–Private Partnership vs. organization-owned platform
The term Public–Private partnership (PPP) is used in a broad sense where a (public) entity needing an investment doesn’t pay for the investment, but then either pays a fee annually to the other (private) entity that make the investment, or allows it to make revenue from the investment. The choice of a PPP model versus an organization-owned model should be driven by a careful evaluation of each option. The factors that should drive the selection of an appropriate model are:
Funding: One of the main factors for the choice of the model is usually driven by the funding available. If specific funding from a development partner is available, such investment does not fit well with a PPP model. In contrast, in the absence of specific funding for the platform, a PPP model is often the only option given the size of the required investment.
Operational costs vs investment costs vs. potential revenue streams: The choice of the model should be influenced by the evaluation of the operational costs, the investment costs and the potential revenue streams. All these elements must be balanced to evaluate the best option. This factor should be evaluated only at the end of the requirement phase, depending on the identification of the real costs versus non-financial incentives.
Feature evolutions: In a PPP model, the software is property of the private sector partner and therefore the evolution of functionalities is not under the authority of the organization in charge of the FDP platform. Each extension will require a specific discussion. In both cases, for any extensions, the organization will likely have to pay for it, but in the case of the PPP option, there will be only one possible provider. In that regard, the PPP model offers slightly less freedom and is likely to be more expensive.
Data governance: The governance issue includes the data ownership and access. In the case of an FDP platform, the collected data can be monetized. At the same time, the protection of personal data should also be considered (see Lesson 2.4).
Technical capacity: One of the key advantages of the PPP model in technology platforms is that the platform is operated and maintained by the private sector partner and does not require technical expertise or transfer of ownership to an IT team that may or may not have the required capacities.
The choice is not really a binary choice of a PPP versus an organization-owned model. Mixed models should also be considered. A typical mixed model for technical platforms is a model where the organization pays for the development, owns the platform, owns the data and owns the revenue, but outsources the operations. If the organization in charge does not have the staff or the capacity to manage the platform, this is a suitable option.
5.2 Revenue Stream
Different products can be designed from an FDP platform and can generate revenue. The table below presents an overview of the needs, requirements, and potential uses of main actors in the agriculture sector.
Table 1 Overview of actors and requirements in the agricultural sector
Potential clients
Examples of data needs
Data available in the profiling platform
Client likely to access the profiling platform
Potential customer’s requirements for the profiling platform
Potential customer’s activities based on profiles data
NGOs (and donor partners)
Performance data to support program or intervention designs, contact data to support advocacy and outreach campaigns
Yes
Possible
Data quality, timeliness and completeness; *
neutrality in data access; # personal data in compliance with open data and privacy principles;
affordability; $
real-time web and API access
Define program interventions (community, geography, targets, etc.); support program activities (communication, support of beneficiaries; M&E)
Financial institutions
Data on unbanked farmers; data on farmers unable to purchase inputs due to financial constraints; financial (assets) and performance (yield) data to design agri-specific financial products, and to issue loans and credit facilities
Mostly
Possible
Data quality, timeliness and completeness;
data provided by neutral party; personal data; affordability;
real-time web and API access
Feed credit scoring engine with data
design new financial product;
Conduct market analysis
targeted advertisements
Insurance companies
Performance (yields) and environmental data to design agri-specific insurance products
Yes
Possible
Data quality, timeliness and completeness;
geolocalized data;%
neutrality in data access; affordability;
real-time web and API access
Implementing new insurance products at lower cost; market research;
product design; targeted advertising
Food processing and exporters
Data on certification, yields, prices for planning and to ensure quality
In part
Unlikely
Data quality, timeliness and completeness; neutrality in data access; affordability;
real-time web and API access; specific data needs&
Identifying new production opportunities;
developing agri-food supply chains
Agricultural products dealers
Data on crops, land, pests, inputs, etc. to provide farmers with targeted products and services
Mostly
Likely
Data quality, timeliness and completeness; neutrality in data access; affordability;
real-time web and API access
Market-research
design new product
targeted advertisement
Researchers
Empirical data to support scientific studies; data to design innovative agri-products to support transformation to the knowledge economy
Mostly
Likely
Data quality, timeliness and completeness; open data@
Research planning and analysis; replication; quantitative surveys; modelling; design of R&D; prototypes
Agritech startup
Profile information to complement specific activities (e.g. drone mapping, IoT/sensor, tools, lending, etc.)
Mostly
Likely
Data quality, timeliness and completeness; neutrality in data access; affordability;
real-time web and API access
market research;
product design & deployment;
targeted advertising
* The data collection and validation chain should be transparent with various elements of trust added to the profile, such as the time of update, the evolution of profile over time, the name of operator updating the profile, the GPS coordinates of where the profile update was made.
# The provider of the data should be a neutral party without any bias and not providing preferential access to specific parties.
$ The concept of affordability is hard to define because affordability is not only a function of a potential customer’s ability to pay but of the value they attach to the product or service being offered to them. During the interviews conducted, interviewees were not in a position to provide indications of what they are prepared to pay for data because they were uncertain which data would be available, and whether this is accurate and reliable. Most organizations have access to some data via their own or other existing systems (e.g. access to credit information from the Credit Reference Bureau in the case of financial institutions) and they can’t yet evaluate how a new approach could ease their work and increase their opportunities.
% The trend in agriculture at the moment is on weather-based insurance which requires information about the exact location of fields owned by the farmer or cooperative applying for insurance.
& Each value chain has specific data needs depending on the commodity and a farmer profile may therefore to accommodate customized value-chain specific data. For example, the altitude of a wash station is an important data point in the coffee value chain but does make sense in other value chains.
@ Researchers in agriculture are not usually interested in personal data, but more in anonymized open data for analysis and publication.
6. Summary
In this lesson, you were presented in detail with the rationale for the setup of a digital farmer profiling platform by different types of organizations. The goal was also to provide a practical, implementation-oriented approach to support the execution of a profiling initiative. Particular attention was paid to the sustainability of the profiling platform, and possible business models.
There isn’t one solution that fits all context, but depending on the size, the capacities, the structure and the objective of a given organization, there is a series of options to select to make an efficient, sustainable, up-to-date FDP platform. This lesson is a guide to help organizations understand the benefits they can expect for the setup of such platforms, and it is also a guide for them to select the best approach for their needs and to adopt a sustainable model.
Bibliography
Boyera, S., Addison, C. and Msengezi, C. 2017. Farmer profiling: making data work for smallholder farmers. CTA Working Paper 17/09, CTA, Wageningen, Netherlands. https://cgspace.cgiar.org/handle/10568/89763
CTA, 2017. Farmer Registration and Profiling: How did it go? CTA, Wageningen, Netherlands. http://experience-capitalization.cta.int/farmer-registration-and-profiling-how-did-it-go/
Gray, B., Babcock, L., Tobias, L., McCord, M., Herrera, A., Osei, C. and Cadavid, R. 2018. Digital Farmer Profiles: Reimagining Smallholder Agriculture, USAID, Washington DC, USA. https://www.usaid.gov/sites/default/files/documents/15396/Data_Driven_Agriculture_Farmer_Profile.pdf
ICT Update 2018. Data4Ag: New opportunities for organised smallholder farmers, ICT Update, issue 89, December 2018, CTA, Wageningen, Netherlands. Available online at: http://ictupdate.cta.int/category/issues/89-data4ag/
Owoyesiga H., 2018. The Impact of Improved Technologies at Igara Growers Tea Factory Ltd, Uganda, CTA Brussels Briefing, https://www.slideshare.net/brusselsbriefings/the-impact-of-of-improved-technologies-at-igara-growers-tea-factory-ltd-uganda
Footnotes
[1] We mean here that farmers would proactively get in touch with the organization running the FDP platform to update their profile
[2] See lesson 1.1 for more details on categories of applications
Last updated
