The World’s Largest Open Access Agricultural & Applied Economics Digital Library This document is discoverable and free to researchers across the globe due to the work of AgEcon Search. Help ensure our sustainability. Give to AgE con Search AgEcon Search http://ageconsearch.umn.edu [email protected] Papers downloaded from AgEcon Search may be used for non-commercial purposes and personal study only. No other use, including posting to another Internet site, is permitted without permission from the copyright owner (not AgEcon Search), or as allowed under the provisions of Fair Use, U.S. Copyright Act, Title 17 U.S.C. Bites into the Bits: Governance of Data Harvesting Initiatives in Agrifood Chains Lan Ge1 and Marc-Jeroen Bogaardt2 1,2LEI Wageningen UR 1E-mail: [email protected] 2E-mail: [email protected] Paper prepared for presentation at the 148th seminar of the EAAE, ‘’Does Europe need a Food Policy?”, Brussels, Belgium, 30 November – 1 December, 2015 Copyright 2015 by Lan Ge and Marc-Jeroen Bogaardt. All rights reserved. Readers may make verbatim copies of this document for non-commercial purposes by any means, provided that this copyright notice appears on all such copies. 1 Bites into the bits: Governance of data harvesting initiatives in agrifood chains Lan Ge and Marc-Jeroen Bogaardt Abstract Data harvesting is becoming a booming business in agrifood chains where many players are taking bites into data generated by farming. While data technologies such as mobile apps, cloud computing and big data analytics rapidly develop and mature, business models and governance arrangements are still evolving. From a New Institutional Economics perspective and using the theory of multiple rationalities, this paper studies a number of data harvesting initiatives in agrifood chains to identify the key governance issues to be addressed. Implications for ongoing data harvesting initiatives such as the FarmDigital programme are discussed. 1. Introduction As digitalisation of farming processes continues to expand and intensify, the supply and demand of farming data is rapidly growing (Sonka, 2014; Zhang & Shen, 2011). Demand for farming data is on the one hand driven by the need to make informed decisions and on the other hand pulled by informational institutions for governance purposes like transparency and sustainability (Ge & Brewster, 2016; Verbeke, 2005). On the supply side, vast amount of farming data are being generated or automatically collected by smart machines. The internet, mobile technologies and cloud computing have accelerated the transfer, processing and sharing of data. There is a surge of data-tools in the market and even more are in the making1. Spurred by the supply of data and data technologies, data-driven business initiatives are steadily increasing in agrifood chains. Farming is no long just about harvesting food, but becoming a booming data harvesting business where many players are taking bites into data generated by farming (Orts & Spigonardo, 2014). While data technologies like cloud computing and big data analytics rapidly develop and mature, many data harvesting initiatives are still exploring viable business models and 1 See e.g. http://guides.library.cornell.edu/ag-food-data-guide/ag-food-data-tools 2 governance structures to capture the value of data. A variety of business models are being used and developed with different value propositions to different stakeholders, ranging from established ones like those in precision farming aiming at optimizing production2 to nascent ones such as big data analytics and digital compliance. As a result, data harvesting initiatives, in conjunction with the advent of big data in agriculture, feature prominently in agribusiness research, especially in the context of industrialised agriculture in the US, Canada and the EU. Data-drive innovations and data harvesting initiatives are flourishing in the Netherlands. Being part of Public-Private Partnership (PPP) research programme on the transparency and traceability of agrifood chains, the authors have been involved in a number of research projects concerning data harvesting initiatives in agriculture. The recent and ongoing one, FarmDigital, is on the development of a digital compliance platform to reduce the administrative burden for growers and auditors of sustainability standards (www.farmdigital.nl). While reviewing relevant literature and observing the development process of FarmDigital to unfold, our attention was drawn to the lack of a consistent methodology to choose the most suitable governance arrangements. The lack of methodology to analyse governance arrangements is in stark contrast to the abundance of governance issues discussed in academic and grey literature. Parallel to rapid developments in information and communication technologies (ICT) and increased data and information flows, power and economic relations among different stakeholders are undergoing visible changes. Data-driven initiatives have created new social relations characterized by old players taking on new roles and new players taking over roles traditionally played by others. These changes require new governance arrangements to be made for business models to harvest value from farm data. In this process, many have recognized that governance is a thorny issue, but few have systematically analysed governance issues. This has prompted us to develop our own analytical framework for studying governance of data harvesting initiatives in general, and digital compliance platform in particular. In what follows, we will first present our framework based on new institutional economics (NIE) and the multiple rationalities theory to identify the key aspects and issues. 2 See e.g. https://www.foreignaffairs.com/articles/united-states/2015-04-20/precision-agriculture-revolution) 3 Following this framework, we studied seven data harvesting initiatives in agrifood chains. The findings of the study are presented and discussed in Section 3. 2. Conceptual framework In developing our conceptual framework, we draw insights from two streams of economic theories: the New Institutional Economics (NIE) and the multiple rationalities theory. This choice is based on our conceptualisation of data-driven initiatives as economic organizations that, instead of producing agricultural goods like other organizations, produce information that is of value to the users through the collection, storage, transfer and analysis of farm data. Since the data processes inevitably involve a network of actors, network features are likely to play an important role in the governance of these organizations. Governance in this respect is the use of institutions and structures of authority and collaborations, i.e., governance arrangements, to allocate resources and to coordinate and control joint action across a group or network of organizations that work together to achieve a collective goal (Provan & Kenis, 2007). According to NIE, governance arrangements are made to align transactions or economic exchanges to their institutional environment (Williamson, 2000). A key insight from NIE is that efficient governance arrangements should reduce transaction costs, i.e., resources used to establish and maintain human exchanges. Uncertainty is an important determinant of transaction costs as governance arrangements are made essentially to cope with uncertainties in human exchanges. Uncertainties arise due to on the one hand the lack of information or asymmetric distribution of information among different stakeholders about the value of the goods to be exchanged and on the other hand the lack of control over transactions among stakeholders. Transaction costs consist of ‘mundane’ coordination costs (Baldwin & Clark, 2000), information costs (to obtain and process information) and negotiation costs (to make sure the same information is accepted by different parties) (Lv et al., 2012). Archetypical governance arrangements known in the literature are firms (or using hierarchy for coordination), market (using the price mechanism for coordination), and various hybrid forms (Slangen et al., 2008). The multiple rationalities theory views rationalities as ‘frameworks of giving meaning’ incorporating ethical norms and values (Edwards, 1998; Snellen, 1987). Four rationalities have been identified: the political, legal, economic and scientific rationality. Each form of 4 rationality leads to a specific way of acting. The four rationalities can be seen as systems of criteria for responsible governance. We apply the rationalities theory to analyse the governance of the data-driven business initiatives in conjunction with NIE. In this paper economic rationality concerns whether the goal of a farm data-driven initiative is obtained in an efficient way from a micro-economic perspective. The legal rationality concerns governance according to and in compliance with the statutory laws. For example problems are tackled by applying new rules or more control and enforcing measures. Political rationality relates to issues of support and legitimacy of the initiative such as access and decision making. Whether farmers can join easily and for free is for example an issue of political rationality. Finally, scientific rationality refers to the scientific knowledge and information significant for the functioning of the initiative in relation to realizing the common outcome. We posit that when designing the governance arrangements for data harvesting initiatives, besides their impact on transaction costs, the four rationalities should be fully reflected on to ensure its viability. More specifically, one should evaluate the extent to which governance arrangements are congruent to those rationalities. Through the lenses of transaction costs and multiple rationalities, we examine a number of known data harvesting initiatives and draw on our own extensive experience of participating in data harvesting initiatives. We use publicly available information to derive the rationalities behind these initiatives and the governance arrangements. The sources of information include company websites, news items, and a wide range of business analysis reports. 3. Results and discussion 3.1. Key features of data harvesting initiatives Table 1 presents an overview of the data harvesting initiatives we studied. Among numerous data-harvesting initiatives that are known to us, we consider these initiatives as landmarks in the world of data business due to their conspicuous web-presence and marked differences in origin, business model, development path, and governance arrangement. Besides Farm Digital, we have chosen to study six other initiatives in order to see to what extent their governance arrangements and rationalities differ and what factors would explain the difference. These insights, we hope, would enable us to draw design implications for FarmDigital. As shown in Table 1, besides countries of origin, the initiatives also differ in their business status, data technology used or offered, and their use of data. In terms of the ‘initiator’ 5 of the initiatives, we observe that both large agricultural firms and start-ups are active in the data harvesting business. While large firms like Monsanto and John Deere typically adopt the strategy of acquiring start-ups to strengthen their existing position (e.g., Monsanto acquired the company Climate Corporation to provide planting advice), new start-ups continuously seek investors and funding from venture capital and large technology firms to expand their service and influence (e.g., Farmer Business Network raised funds from Google Ventures). Compared to other initiatives, FarmDigital has a rather unique niche of data harvesting business that is intimately linked to the certification industry. To provide a common background for comparison, we have pictured the current landscape of data harvesting business in farming as a ‘battle field’ fought by major players in venture capital, agribusiness (like Monsanto and DuPont), ag-tech (like John Deere and Trimble), and other technology companies (like IBM and Oracle). The landscape is visualised in Figure 1, where we also show the driving force of data flows within the farming business and the main sphere of influences aiming at farming data. As in all businesses, developments in ICTs have resulted in an explosive increase in data flows in agriculture and the trend is likely to be self-reinforcing. NGOs are considered a major driving force for the increase of data flows due to their demand for transparency and evidence for sustainability (Vellema & van Wijk, 2014). Standards refer to requirements on products and production processes which necessitate measurements, data transfer and data analysis. While all initiatives perform data analysis to get information or insights, the content and natures of the insights vary from whole farm management advice to a specific aspect of verifying compliance (certification). These differences imply different exchanges of data and the value created for the parties involved. Understanding these differences is critical for understanding the different choices of business models and governance arrangements. 3.2. Governance arrangements Table 2 summarizes the main features of the governance arrangements. In several aspects, governance arrangements of the data-harvesting initiatives exhibit a considerable degree of diversity in organisational forms. With regard to the coordination mechanism, although the price mechanism prevails, it is almost always used in combination with mutual agreements or relational arrangements. 6 A feature of the data harvesting initiatives originated as start-ups is the joint creation of value with a network of actors providing data, knowledge (know-how) and tools. This could be a network of farmers sharing farm data with the platform or a network of knowledge providers sharing different data tools. For example, 365FarmNet is a consortium whose members include e.g. Allianz, Bayer, KWS, farm-equipment makers CLAAS and Amazone- Werke. The consortium has established a marketplace for agricultural information where growers can buy GPS, diagnostic, crop, fertilizer and other data from any consortium member; download it to their computers and farm equipment; and use it to take action, such as drawing up crop plans for the coming planting seasons. This networked nature of value creation implies that price alone is unlikely to be sufficient in coordinating the value creation from data. Mutual agreements must be made in addition to price mechanism. The diversity in governance arrangements as shown in Table 2 can be explained by different features of the transactions and interactions between the suppliers of data, the supplies of data tools, knowledge and know-how, and the user/buyer of information. Price mechanisms seem to prevail when ownership of data and the value of the information are both clearly defined with regard to the actors concerned and the transactions are standard (recurring). Although there is a general consensus that farmers own their data and should have control over the access to and use of their data, it is much more complicated to assess the value of information as it is intricately linked to the value of data services and its impact on decisions. This implies that there is a high level of uncertainty with regard to the value of the ‘product’ (i.e. information) in these transactions. High transaction costs may arise when parties disagree. Contractual relations are likely to be influenced by uncertainties about the value of information. Data harvesting business derive the value from the information they provide that is meant to improve decision-making. This value can hover be pervasive to the user/buyer as the outcome of his decision depends on numerous factors beyond the control of the decision- maker. Since the circumstances of the growers differ, customers' results may differ materially from those stated by the data harvesting initiatives. 3.3. Rationality behind the data harvesting initiatives 7 To understand the rationality behind the data harvesting initiatives, we examined the mission statements on the websites of the initiatives as well as announcements regarding governance issues such as the rights and liability. The results are summarized in Table 3 and elaborated below. The urge to improve efficiency has evidently motivated all the initiatives studied. For most of them, this further translates into the improvement in productivity and profitability. For example, Farmers Edge began in 2005 with the vision to use technology to help growers become more efficient3. According to Farmers Edge their service assists growers to make advanced management decisions which results in higher yields and higher returns based on farm data. In 2015, with a price of $3.95 per acre for a full service package (variable rate fertilizer programs, telematics packages to manage fleets of machinery, updated satellite images every 7 to 10 days during the growing season, localized weather information) to farmers on over four million acres4, FE with around 160 employees (GIS experts, technology specialists, data scientists, research and development team, precision agronomists, sustainability analysts, carbon specialists, laboratory technicians, and soil scientists among others5), could account for a turnover of almost 16 million a year. With the average farm size for 2011 in Canada of 778 acres6, a farmer would pay around $3073 for the full service package of Farmers EdgeTM. That is more than 5 percent of the average net operating farm income of $59,402 for 2009-2013 in Canada7. Similarly to Farmers Edge, the aim of Farmers Business Network (FBN) is to help farmers to select the optimal seeding grade for their variety and their field in order to reach the maximum potential. John Deere states that its data platform will increase the productivity and efficiency of the crops and lead in the end to higher production and revenue8. According to Monsanto, FieldScripts help farmers to get the most out of every acre. FieldScripts 3 Source: http://www.nutrientstewardship.com/partners/products-and-services/farmers-edge-precision- consulting-inc (Consulted November 14th 2015). 4 Source: http://www.grainews.ca/2014/12/10/a-new-business-model-for-precision-ag-data-packages/ (consulted November 8th 2015). 5 Source: http://usbusinessexecutive.com/agriculture/case-studies/farmers-edge-growing-more-precisely- through-manitoba-based-farm-data (Consulted November 8th 2015). 6 Source: http://www.statcan.gc.ca/pub/96-325-x/2014001/article/11905-eng.htm#a4 (Consulted November 14th 2015). 7 Source: Agriculture and Agri-Food Canada. 2015. Canadian Agriculture Outlook. See: http://www.cahrc- ccrha.ca/sites/default/files/AAFC%202015%20Canadian%20Agriculturual%20Outlook%20%20AAFCAAC- %23101147675-v1-12325E_-_2015_Canadian_Agricultural_Outlook_0.pdf 8 Source: https://datafloq.com/read/john-deere-revolutionizing-farming-big-data/511 (consulted October 15th 2015). 8 maximize productivity, allow farmers to minimize risks and realize higher yields9. As stated by 365FarmNet, its responsibility is to provide and cleverly cross-link agricultural know how in the form of a single software. Farmers and contractors are able to combine partner applications on a modular basis depending on their needs. Besides improving the efficiency in agricultural production, the initiatives also aim to improve the efficiency in using data for producing information and insights-their core business-by providing tools for easy collection, storage, integration and analysis of data. For example, Farmobile sells a simple data collection tool that centralises grower’s agronomic data from multiple systems in one electronic farm record. Farmobile standardize the data and make it easily searchable for customers who want to purchase data10. In the case of John Deere, all data coming from sensors at farming equipment are collected in its web portal MyJohnDeere.com and combined with historical and real-time data regarding weather prediction, soil conditions, crop features etc. in order to help farmers to run and manage all their operations. The company 365FarmNet enables the farmer to manage his entire agricultural holding for 365 days a year with one single software that is independent of manufacturers. 365FarmNet states that it saves time and makes data usable for the farmer. The political rationality is assessed from the perspective of access and decision- making. As also shown in Table 2, the access control is generally market-based with mutual agreements on value-sharing, privacy and liability issues. FBN offers benchmarking information to member farmers by collecting and analysing data from thousands of fields around the US for a membership fee of $500 per year. The data management system of Farmobile originates with a $1,250 annual subscription fee11. If farmers opt to share their data through Farmobile, they will get 50 percent of revenue derived from selling the data12. Monsanto charges $10 per acre13. Legal rationality is mostly reflected in the ‘terms of use’ articles and subscription agreements concerning ownership, privacy and liability. Farmers Edge (FE) states that the agronomic and financial data of the farmer belongs to the farmer. In its contract with a farmer 9 Source: https://www.youtube.com/watch?v=m-pn9ytxihQ 10 Source: http://www.croplife.com/equipment/precision-ag/farmobile-coming-online-in-2015/ (Consulted November 8th 2015). 11 Source: Farmobile in the Press: https://www.farmobile.com/blog/news (consulted November 8th 2015). 12 Source: https://www.fcc-fac.ca/fcc/agKnowledge/publications/agrisuccess/pdfs/agrisuccess-mar-apr-2015.pdf (consulted November 8th 2015). 13 Source: Franklin, D. (2014). Monsanto’s FieldScripts an Early Move on the Next Phase of Farming. Blueshift Research. San Francisco. US (see: http://blueshiftideas.com/reports/021405MonsantosFieldScriptsanEarlyMoveontheNextPhaseofFarming.pdf) 9