A Controversy Born
In February of 2020, farmers took to Twitter to express their concerns about the partnership between The Climate Corporation FieldView platform and Tillable.(1)
FieldView is a digital software platform which ‘“listens” to the massive amounts of data generated on farms to help producers “optimize yield” and “maximize profit”. Signing up to their service gives the Climate Corporation access to information as disparate as volume of water utilized, to units of fertilizer used, to anticipated yield.
Tillable is a agtech “disruptor” that bills itself as a new platform that simplifies the relationship management between landlords and renters. In essence, they have created a new industry: they are a data-driven farm management and subletting company that severs the direct connection between landlord and lessee.
Tillable has been sending out letters to landlords that contained granular information about farms and input transactions which producers believed went beyond what could be publicly available. The concern of farmers was that the corporation was leveraging their partnership with FieldView to gather data and make offers. This was hotly denied by Tillable and Climate, but farmers remained unconvinced and the relationship was quickly severed.
Although there were many reasons for this specific suspicion (a dense network of connections exist between Tillable and Climate (2)), this incident is emblematic of a much larger set of questions about the data ecosystem in agriculture. What are the types of data collected on farms? Could Tillable have found such detailed information without gathering it directly from the Climate Corporation?
Three Types of Data
According to recent research (Ellixson et al, 2019), farm data can be divided into three categories: data collected by the farmer on the farm; data collected on the farm by a secondary party at the request of the farmer; and data collected about the farm by an outside party. A brief outline of the differences of each is discussed below.
Note that this is a very superficial view of data collected, and is largely limited to to the United States, although is generally applicable globally.
Data Collected by the Farmer on the Farm
Farmers have been diligent collectors of information about their farms for centuries, often famously carrying dust and sweat-caked notebooks in their breast or pants pockets to jot down specifics of planting or economic decisions. Much of this was done to adhere to rigorous standards and regulatory requirements, although farmers have also been early adopters of new technologies such as yield monitors and GIS-driven soil sampling, driven by economic need as much as by incentive (governments providing financial stimulus, for example). Software programs which provide direct access to information without second-party intervention is included in this category.
Data Collected on the Farm by a Secondary Party
This same drive has producers investing in new types of precision and data technologies. Many software programs such as Climate FieldView would be included in this category, as would newer models of tractors and combines. This may seem counter-intuitive, as farmers themselves are “collecting” the data — but it is being transmitted and analyzed independently by secondary parties. Farmers themselves may ostensibly have “ownership” of the data, but would find it difficult (if not impossible), to make use of it independently. Additionally, this data is utilized by the secondary provider in an aggregated manner to generate insight into the larger picture of the state of agriculture, not just the individual farm.
Most, if not all, modern farming equipment is sold with sensors which transmit information of interest not only of the farm but of the machines themselves. This allows the equipment provider to monitor farmers for “off-brand” use — allowing them to shut down, or “brick” a device if they sense it’s being used for an undesignated purpose (this is at the heart of the Right to Repair debate in Ag).
But farmers also work with any number of secondary individuals, from insurance providers, agronomists, manufacturers, and consultants, who can use a myriad of tools (drones being primary among them) to gather data consensually from farmers).
Data Collected about the Farm by an Outside Party
Large amounts of data are also collected independently about individual farms which are readily available to the public — although often, you have to be a) quite sophisticated to interpret it, or b) rich enough to buy it.
Public entities such as the United States Department of Agriculture regularly publishes large data sets of granular information about the agricultural sector. This includes weather information, yield data, financial conditions, production practices, and more. These data sets have historically been provided on the basis that they benefit public institutions and provide farmers insights they could not individually access, but arguments have been made recently that this information more benefits large corporations and speculators.(3)
Challenges to the release of this data have often been ignored. In 2008, the D.C. Circuit Court (515 F. 3d. 1224, D.C. Cir.) ruled that the USDA could not withhold information on “irrigation practices, farm acreage, and the number and width of rows of tobacco and cotton” or GIS information on “farm, tract, and boundary identification, calculated acreage, and characteristics of the land such as whether it is erodible, barren, or has water or perennial snow cover” (Rasmussen, 2016). While the court recognized that there was a potential privacy interest, they were not persuaded that it outweighed the public interest in such disclosures (this may be countered by the specific prohibition in Section 1619 of the 2008 Farm Bill which prohibits the publication of public geospatial data).
But even a simple Google Earth search could provide surprising amounts of information to interested parties. A trained individual could identify crop or cattle-type simply with a search.
Looking beyond data about the physical aspects of farms, it is possible to obtain info on the amount of public monies a farm receives by filing a FOIA request. Firms like FarmMarket ID pulls from publicly available information, credit reports, mortgage information — even magazine subscription data, to build robust profiles on individuals and farms.
Discussion
At the start of this post, I asked two questions: what are the types of data collected on farms? And could Tillable have found detailed information about individual properties without gathering it directly from the Climate Corporation?
The answer to the latter is almost decidedly yes — as long as they could pay for the information, or had the capabilities to aggregate and analyze it.
Data of all kinds — macro and micro, is increasingly being collected about individual farms and agriculture in general. But the fury generated by the Tillable controversy was about potential use of individual data without consent. This is emblematic of a much larger issue: the growing data economy in farming, and the way that data is extracted, exploited, and capitalized upon without providing substantial benefits to the ones generating that data. This will become increasingly important and controversial as the technology proliferates.
(1) The Climate Corporation is a subsidiary of Bayer, but was originally purchased by Monsanto (which was itself purchased by Bayer).
(2) The current CEO of Tillable worked on the FieldView product at one point, for example.
(3) This argument has gained more currency with the advent of ML/AI
Coble, K., Mishra, A., Ferrell, S., & Griffin, T. (2018). Big Data in Agriculture: A Challenge for the Future. Applied Economic Perspectives and Policy. Vol 40(1). March 79–96.
Ellixson, A., Griffin, T., Ferrell, S., & Goeringer, P. (2019). Legal and Economic Implications of Farm Data: Ownership and Possible Protections. Drake Journal of Agricultural Law. Vol 24(1). 49–66.
Rasmunssen, N. (2016). From Precision Agriculture to Market Manipulation: A New Frontier in the Legal Community. Minnesota Journal of Law, Science, and Technology. Vol 17. 489.
