Could a Data Sharing Protocol be Agriculture's Missing Link?

A Case for Data Sharing

Sharing of data across organizations, irrespective of if it is governments, NGOs or private sector, is hard. Agricultural data in the Global North are aggregated by relatively few big agricultural companies, whereas data in the Global South are fragmented, resting with thousands of companies, governments and NGOs. In countries like Ethiopia where farm & farmer data are not readily available in digital form, we still need to make significant parallel investments in the development of digital registries. It’s encouraging to see governments recognizing this need with the Ministry of Agriculture in Ethiopia investing in developing a registry of its 60,000 extension agents and over 15 million farming households. Even though we increasingly have open public datasets—for example on soil types and weather—led by organizations such as Global Open Data for Agriculture and Nutrition (GODAN), more investment is needed to address the persistent issues of reliability, lack of digitization, limited resolution and low frequency of updation.

But what really hinders data sharing is the need for privacy, especially for personally identifiable data, and lack of incentive to share proprietary & monetizable data. For instance, we recently did a study of over 60 potential data providers and users in Ethiopia, which outlined that the need for trust, security and ownership remains organizations’ main barrier to sharing their data. If organizations could securely share data on their own terms, we could find win-win solutions especially through vertical collaboration across the value-chain. For example, a farmer network could share input demand data with suppliers to ensure adequate & timely availability of inputs. 

If data sharing is challenging organizations, you can only imagine the difficulty farmers face, especially since most don’t control their data. If farmers controlled their data, they could choose to share it with different service providers thereby unlocking new services, products and markets; monetize the data for themselves; and mitigate being exploited by the increase in transparency & auditability of data. This is where a need for a data sharing protocol comes in.

Need for Protocol or Platform?

A protocol is a standard that anyone can use and build off of—for instance Hyper Text Transfer Protocol (HTTP) which is used to access the web or Simple Mail Transfer Protocol (SMTP) for emails. In contrast, platforms such as those run by Amazon and Facebook are networks that are run by central authorities dominated by their pre-defined sets of policies and rules.

Proprietary platforms often require data aggregation which ends up giving disproportionate control to the central authorities. As more customers use and contribute their data, the more the platform generates additional usage data resulting in a sticky self-perpetuating cycle that can give rise to rent seeking & “winner takes all” markets. In contrast, protocols too can enable network effects but can be used as a public good without the pitfalls of proprietary platforms.

What is needed therefore is a collaborative model in which organizations can align on a pre-competitive basis on the value, standard, and governance of a protocol. For instance, imagine the confusion if there had been no standardization in the early internet days resulting in multiple competing protocols for HTTPS or SMTP. Solutions such as the OpenCollective are already showing the way for organizations to coordinate & co-invest to develop shared infrastructure with transparency.

FarmStack: A Data Sharing Protocol

FarmStack, is an example of an open-source protocol that enables the secure transfer of data between organizations and farmers.

Rather than requiring organizations to share their data with a third party, data providers and data consumers can use FarmStack’s self-service peer-to-peer connectors by installing a certified software package on their existing infrastructure. FarmStack enables a trusted network of certified users who can assert control over the usage of data by configuring usage policies. Whereas most data sharing solutions provide ways to restrict access, FarmStack goes a step beyond and helps data owners define how their data can be used. It also helps organizations adhere to local data policies such as Europe’s General Data Protection Regulation (GDPR) & India’s Data Empowerment and Protection Architecture (DEPA). By utilizing an open architecture, organizations can use the protocol irrespective of where & how they store their data. Beyond organizations, farmers themselves can control their data through open data wallets that can help farmers consent to what data they want to be shared with whom.

Digital Green, with the support from Bill & Melinda Gates Foundation and the United Kingdom’s Foreign, Commonwealth, and Development Office, is currently implementing an instance of FarmStack in Ethiopia where the Ministry of Agriculture will serve as a data steward responsible for operating the gateways for participants while allowing for the decentralized sharing of data between member organizations. A free version of FarmStack is available on Github.

Some sectors have started to crack the code of interoperability and data sharing as a collective public good. Banks use SWIFT codes, health records are increasingly being shared across providers and the transport sector has standardized the way traffic data is collected & shared. The question remains: can the food and agriculture system be the next frontier?