EIP-AGRI Seminar

‘Data revolution: emerging new data-driven

business models in the agri-food sector’
SEMINAR REPORT
22-23 JUNE 2016

funded by
 Table of contents

1.Introduction 3
2. Summary 4
3. Participants 5
4. Data innovations in agriculture and the supply chain 7
5. Data-driven business cases 8
6. Limiting factors 13
7. Encouraging factors and available support 15
8. High potential actions to trigger data-driven business models 17
References 19

2 funded by
1. Introduction
The development of Information and Communication Technologies (ICT) and sensor-based technologies,
procedures and software has enormously increased the amount of data collected and data available in
agricultural sectors and throughout the whole supply chain (from farm to fork). The appearance and adoption
of biosensors, nanotechnology, low-cost electronics, the Internet of Things or remote sensing devices, among
others, will further boost the data domain. This development offers significant potential for new data-driven
business models.

Several initiatives are already in place, many are quickly developing, and even more initiatives are just ideas.
This state of infancy suggests that it makes sense to connect initiatives by bringing the community together
and by giving people who are launching or conceiving new data-driven businesses the opportunity to discuss,
test and enrich their ideas while meeting other members of the community. This was the idea that triggered
the organisation of a seminar within the framework of the European Innovation Partnership for Agricultural
Productivity and Sustainability (EIP-AGRI) by the European Commission’s Directorate-General for Agriculture
and Rural Development in cooperation with the Institute for Agrostrategies and Innovations in Sofia, Bulgaria,
on 22-23 June 2016.

The seminar also reflected on how agricultural and rural development policy can support the data revolution
for an enhanced productivity and sustainability in the wide agri-food chain, covering different sectors, farm
types and production systems. The concrete objectives of the seminar were:

• To identify and discuss existing and potential data-driven business models in the agricultural sector,
including the whole supply chain
• To identify enabling conditions for these business models, and strategies to support their development
• To bring people together to further develop data-driven business models

Experiences and ideas of participants were at the core of the seminar by using participative methods.
Many participants presented their business models to other colleagues and innovators.
This document is the report of the seminar. It is partly based on keynotes given at the seminar by Krijn
Poppe (LEI Wageningen UR) on data-driven business models and by Iman Boot (DG AGRI) on available
support measures. Many of the insights in the report come from the interactive workshops in the seminar.

More information on the seminar, including the agenda and presentations, can be found on the
EIP-AGRI event webpage.

3 funded by
2. Summary

ICT and sensor-based technologies increase the amount of data collected in agriculture and throughout the
whole supply chain. This development offers significant potential for new data-driven business models. In
June 2016, the European Commission’s Directorate-General for Agriculture and Rural Development organised a
seminar in Sofia, Bulgaria, in the framework of the European Innovation Partnership for Agricultural Productivity
and Sustainability (EIP-AGRI). The goal was to reflect on how agricultural and rural development policy can
support the data revolution for an enhanced productivity and sustainability in the wide agri-food chain. Just
over 100 persons from all over Europe participated in the seminar, which was organised in a very interactive,
bottom-up way. The seminar was designed according to its three main objectives: To identify and discuss
existing and potential data-driven business models; to identify enabling conditions; to bring people together
to further develop data-driven business models.

ICT can be very disruptive to current business models, as for instance Uber shows in the taxi world. A very
dramatic example is Wikipedia, where an arrangement on a voluntary basis wiped out the commercial market
for encyclopedias. It also raises the question how money can be earned with data. Five types of business
models are typical (Van’t Spijker, 2014):
1. basic data sales
2. product innovation
3. commodity swap: data for data
4. value chain integration
5. value net creation

Examples of all five types were discussed at the seminar, with an emphasis on data (or software) sales, value
chain integration and value net creation. Small or medium-sized enterprises (SMEs) dealing with ICT were the
main types of participants at the seminar. For these participants, product innovation was more associated with
manufacturing agricultural equipment, and commodity swaps rather with data exchange between farmers and
food processors.
Discussions focused on limiting and encouraging factors for data-driven business models. The following groups
of limiting factors were identified:

• A lack of awareness of the possibilities and benefits of data-driven applications for agriculture. For ICT
developers this implies a necessity to focus on farmers’ needs.
• The lack of standardisation and interoperability.
• Incentives for letting farmers share data are needed to add value to data.
• Discussions on the governance of data and the potential effects of ICT on the food chain.
• Financing the investments in data-driven solutions.

By far the most encouraging factor for data-driven business models in agriculture and the food chains is of
course the insight of innovators that problems (whether they are operational problems at farm level or societal
challenges at large) can be solved by a clever application of ICT. The problems of today are the business of
tomorrow. Governments offer support to innovators in many ways, from (financial) support for starters to
research projects that provide results that can be used in commercial applications. In relation to data-driven
business models in agriculture, the European Commission – DG AGRI in this respect focuses on the innovation-
driven research approach and on the interactive innovation model through the European Innovation Partnership
“Agricultural Productivity and Sustainability” (EIP-AGRI). Operational Groups, Thematic Networks and Multi-
actor research projects are key activities, with financial support from the Rural Development Programme

4 funded by
and Horizon 2020. The EIP-AGRI Service Point organises relevant seminars and focus groups, and enables
networking.

The EIP-AGRI seminar has helped to create awareness that ICT and sensor-based technologies, procedures
and software are being implemented in agriculture and the food chain. However, it is also clear that we are still
in a state of infancy concerning data-driven business models. Seminar participants welcomed the opportunity to
connect initiatives by bringing the community together. In line with practices in the ICT community, participants
decided to create an online platform to stay connected and exchange information on Slack: https://agridata-
eu.slack.com/

In addition to what innovators can do themselves, the European Commission could take actions that overcome
the limiting factors identified at the EIP-AGRI seminar. Five actions were suggested that have a high potential
to promote data-driven business models:

• Develop an EU ICT Architecture Strategy for AGRI-FOOD to provide guidance on the desired information
architecture, for a food sector based on family farms and SMEs in food processing
• Active provision of digital data by government organisations with open data, and integration of government
data capturing with systems in the food chain
• Solve the discussions on data ownership and data governance
• Make a more in-depth analysis of data-driven business models to learn what works and what doesn’t
• Create a research agenda for real-time agronomic models, to transform data into information that can be
used for making farm decisions.

3. Participants
108 persons participated in the seminar, as a result of two invitation rounds. The first one followed a public
call for participants. That call resulted in more than 300 applications from many countries, including several
outside of Europe. From this round, the participants that are the most relevant and interesting for Europe
were selected, based on their cases and motivation. The second invitation round was based on a selected pull
of organisations which were kindly requested to propose participants. These included among others some
standard organisations, industry organisations like CEMA, IFOAM, CopaCogeca, CEJA, EUFRAS, etc. By far the
majority of the participants came from the first group and more than 40 participants actively wanted to present
a case in which they were involved.

The group of participants was relevant to the topic, well balanced geographically (see table) and in terms
of profile. There was a significant participation of the private sector through many ICT-SMEs and not much
participation from big players (big machinery companies, food industry, other ICT-providers, etc.).

The seminar was designed according to its three main objectives (Identify and discuss existing and potential
data-driven business models; identify enabling conditions; bring people together to further develop data-
driven business models). It put the participants and their contributions at the core of the event with only
two classical presentations. The rest of the time was devoted to different types of group work and interactive
discussions. According to the evaluation by the participants, the third objective was clearly achieved.

5 funded by
 Participants very much stressed the opportunities to meet
potential partners and extend their network. The first
objective was also quite well covered. The second one, the
identification of enabling conditions for data-driven business
models, was only identified as being reached by around
one third of respondent participants. Nevertheless, 65% of
respondents improved their knowledge on the EIP-AGRI and
on Operational Groups (OGs).

Country Participants
Austria 2
Belgium 7
Bosnia and Herzegovina 1
Bulgaria 13
Croatia 1
Cyprus 1
Estonia 2
Finland 3
France 5
Germany 5
Greece 8
Hungary 3
Ireland 4
Italy 10
Lithuania 3
Montenegro 1
Netherlands 9
Poland 2
Portugal 3
Romania 1
Serbia 1
Slovakia 1
Slovenia 2
Spain 12
Sweden 4
United Kingdom 4

6 funded by
4. Data innovations in agriculture and the supply chain
Information and Communication Technologies (ICT) change the world we are living in. Cloud computing, Inter-
net of Things, location-based monitoring, social media, block chain technology and big data are technologies
that have a high potential for unprecedented innovations and that have led to disruptive business models in
several industries. These technologies are quickly invading agriculture and the supply chain, from input indus-
tries (like the machinery industry) via farming and food processing, to logistics, retail and even the consumer
when it comes to food and health apps.

The state of the macro-economy is characterised by a major economic crisis. Innovation is needed to create
a smart, sustainable and inclusive society. This situation suggests that we need institutional innovations that
change the way we organise our economy. These could come from a focus on using ICT as driving technology
to solve societal challenges like food and nutrition security, climate changes, environmental issues and healt-
hy diets for a healthy life (Perez, 2002; Poppe et al, 2013). These innovations include new business models
and they will probably bring further changes in the organisation of the supply chain. Large organisations in
the supply chain have already digitalised most of their data, but data exchange and interoperability between
organisations is still poor. There is room for productivity gains (fewer administrative burdens) and there is an
under-exploitation of ICT in monitoring and managing production and consumption processes, to make them
more sustainable. The increased interoperability that is needed can be created by software ecosystems for
ABCDEFs: Agri-Business Collaboration & Data Exchange Facilities (Poppe et al, 2015a).

Depending on how such software ecosystems are developed, two scenarios have been sketched (EU SCAR,
2015). One is that of a Captive Prescriptive Model: the farmer becomes part of one integrated supply chain
as a franchiser / contractor with limited freedom, and is supported by one platform for e.g. the potato breeder,
machinery company, chemical company, and the french fries processor. In this case the integration with ser-
vice providers and the government is likely to be weak. A second scenario is that of an Open Collaboration
Model: markets for services, apps and data are developed in common, open platforms with fewer lock–in
effects. The business model of such a platform is probably more difficult, because it requires a higher common
investment, to be paid up front. But probably it would also mean more empowerment for farmers and coope-
ratives. Against this background, it makes sense to look to new data-driven business models in more detail.

7 funded by
5. Data-driven business cases
A business model describes the rationale of how an organisation creates, delivers and captures value in
economic, social, cultural or other contexts. The process of business model construction is part of a company’s
business strategy. A business model has several dimensions. The value proposition is the main element.
It describes what service is delivered to a specific group of clients. The value architecture describes the
organisational infrastructure and the technological architecture which allows the movement of products,
services, and information. The value finance provides details on information related to how income is generated,
with elements such as the total cost of ownership, pricing methods, and revenue structure. The value network
articulates how the company collaborates with other organisations. A very popular tool to create and visualise
a business model is Osterwalder’s Canvas Model. Several interesting examples can be found in the literature
and on the worldwide web. The attractiveness of the tool is that it visualises the different aspects of a business
model, as described above, on a large surface (at least an A4 landscape piece of paper) so that groups of
people can jointly sketch and discuss business model elements with post-it notes or board markers. It is a
hands-on tool that fosters understanding, discussion, creativity, and analysis, and that supports all types of
business models (see Osterwalder et al, 2010).

ICT has an effect on business models. It can be very disruptive to current business models, as for instance
Uber shows in the taxi world. A very dramatic example is Wikipedia where an arrangement on a voluntary basis
(in the ‘commons’) wiped out the commercial market for encyclopedias. It also raises the question how money
can be earned with data. In his book “The New Oil - using innovative business models to turn data into profit”,
Van’t Spijker (2014) has suggested five types of business models:
1. basic data sales
2. product innovation
3. commodity swap: data for data
4. value chain integration
5. value net creation

Below we discuss these types in more detail, based on the literature (Poppe et al, 2016, some of the text below
is taken from that report; and Ge and Bogaardt, 2015) and illustrated with some inspiring examples presented
at the EIP-AGRI seminar reported on here.

Basic data sales

Most examples on data-driven business models, also those presented at the EIP-AGRI seminar, can be
classified as a form of basic data sales: software is created to help farmers or others to collect data (by manual
registration or a relatively simple sensor), data is in some cases linked to other, open data and then information
for the decision maker is generated. Essentially, the buyer pays for the software or data, either by subscription
or by paying up front for the software package or data set. Two examples from the EIP-AGRI seminar deal
with the cost of production data in Italian dairy farming, and tracing and tracking in honey production in
Montenegro (see box 1).This category includes examples where the software is sponsored by the government
or a research / advisory organisation, and where it is given away for free. Food processors sometimes offer a
similar support, but often these are examples of commodity swaps (see below), as the farm data is also useful
for the food processor.

In the case of basic data sales in agriculture, we often think of the farmer as being the client of the data-
driven business. It would be interesting to test the idea of farmers sharing or selling their data to others, for
instance via a cooperative. An example from the USA is the Farmers Business Network (FBN). Its mission is to

8 funded by
 Milk and honey

MILK MONEY is the first online internet-based system to benchmark milk production costs in Italy. The tool was developed by
CRPA with the financial support of the Emilia Romagna Region. Any technical adviser or dairy farmer who pays an annual fee
can easily get access to the system, calculate his milk production costs, revenues and profitability. They can then compare
the result with a range of farms selected according to specific criteria. The possibility to benchmark the farmer’s own results
with groups of farms located in the same area gives the farmer the opportunity to understand which cost or revenue items
are out of range and need to be investigated to understand the reasoning behind them. This knowledge has in most of the
cases been an important input to adapt the farm strategy and improve profitability.

The honey traceability programme in Montenegro implements value-added services through food safety and quality assuran-
ce. According to the users, low quality honey from unknown sources is a significant concern for the growing industry. This
creates food safety issues, undercuts fair market prices and damages the industry’s reputation for quality and safety. The
goal is to create an online database of all natural honey producers in Montenegro with data about the beekeeper, bee hive,
queen bee, GPS position of bee hives, weather conditions at the bee hive location, and honey variety produced. Collected
data will allow the creation of a central website with all data for all beekeepers in Montenegro. The website would also offer
QR codes for each beekeeper, which should be printed to be put on honey jars. The QR code, scanned with a smartphone,
will allow users (customers, consumers, but also beekeepers, importers and exporters) to access the respective website
with all data. This will bring transparency and will clearly demonstrate the value of the honey that is provided – starting at
the hive. Feedback from customers and consumers will be encouraged, and it will be open and displayed, ensuring healthy
business competition. Beekeepers will pay a monthly subscription fee for website data updates.

Box 1

make data useful for farmers to select the optimal seeding grade for their variety and their field, in order to
reach a maximum potential. In 2015 FBN aggregated data from 7 million acres of farm land across 17 states
in the USA. FBN is able to assess the performance of 500 seeds and 16 different crops. No data is shared with
other parties. Access is by payment: 500 USD per year. FBN is not linked to any agri-business company, but it
received investment by Google Ventures. They are a community of farmers and independent persons (Ge and
Bogaardt, 2015).

Another example is Farmobile, which sells a simple data collection tool that centralises growers’ agronomic
data from multiple systems in one electronic farm record. Farmobile standardises the data and makes it
easily searchable for customers who want to purchase data. Farmobile’s data management system starts
with a $1,250 annual subscription fee. If farmers opt to share their data through Farmobile, they will get 50
percent of the revenue derived from selling the data. At Farmobile the electronic farm record (EFR) is owned
by the farmer. Farmers have the power to authorise or deny access. Data is stored as long as the subscription
remains active. The farmer’s data is housed on cloud servers of Farmobile. In its marketing, Farmobile claims
that “farmers believe their trust has been violated”: their data go to multinationals, which announce a large
future income from big data, while the farmers have to pay for everything. By the end of 2015 Farmobile had
raised $5.5 million in equity financing from Anterra Capital (a growth capital fund, jointly funded by proprietary
investment of FIL and Rabo Private Equity).

We classify such services as data sales, as it is software that helps farmers to manage their operations and/or
to sell their data. But they could easily develop into a commodity swap or platform solution (see below) when
food processors or others add the data they have that is relevant for farmers.

9 funded by
Product innovation
In the category product innovation, existing products (often machinery) become much more data-intensive.
It could even be that the hardware or product turns into a service. An example of the latter is a breeding
company that sells sperm for cows and adds services on bull selection and inventory management (of frozen
sperm in the farm), or that sells pedometers to track if cows are in heat (to trigger this service delivery).
Especially the development of Internet of Things stimulates original equipment manufacturers like tractor and
machinery companies, but also developers of milking robots or stable and glasshouse equipment to be very
active in adopting this type of business model.

Examples are machine companies like the American corporation John Deere, which collects data from the
agricultural machinery of the farmer. This includes the location of the machinery, engine hours, operational
data (e.g. amount of fuel used) and diagnostic data of the machinery. All the data is collected in the web portal
MyJohnDeere.com. Another example is the Dutch company Lely Industries, which manufactures milking robots
and collects data on the performance of the individual cows.

Commodity swap
In a commodity swap, data is exchanged between (for instance) farmers and food manufacturers to increase
the service component of the transaction. Examples show that software can be made available by processors
of agricultural products to support the management of the farmer and at the same time improve the
production or marketing process of the food company. The UniTip software of the Dutch sugar cooperative

Farmers meet consumers on Facebook

Short supply chains that bypass several levels in the food chain and directly link consumers to farmers have many attractive
characteristics for both sides. It can mean lower prices, better traceability, a better understanding of how the product is
made, more sustainability, and it contributes to local communities, to name a few of the advantages. Farm shops and farm
markets are classical ways to organise short supply chains. Like AirBnb and Uber, several initiatives have turned to ICT to
organise short supply chains. Several examples were discussed at the EIP-AGRI seminar.

An extremely successful and very interesting example comes from Finland. REKO is a direct selling system from producers to
consumers. It is based on the use of social media as a market channel. It is embedded in Facebook, where volunteer admi-
nistrators run closed Facebook groups with producers and consumers. Such groups are extremely easy to establish (from an
ICT point of view): a group is formed, farmers post their offers on a weekly basis, and consumers use the comment option
in Facebook to order (and inspire others in the group to place an order too). The orders are delivered at a central point and
at a certain time slot for pick-up and payment (e.g. Saturday morning between 10.00 and 11.00 hours at the parking lot of
the local school). Existing retailers are often happy to allow the hand-over of products at their own parking lot, hoping that
this will lead to visits of consumers to their shop for other products.

The characteristics of this solution are that it is totally free for all (both for producers and consumers), and it requires very
little administration and no middlemen. By using an established social media tool there are no costs for developing and main-
taining ICT. Three years ago REKO started with two projects in the western part of Finland. It went well and now there are
more than 130 projects (REKO-circles) running, with a total number of more than 180 000 members involved. The estimated
turnover for 2016 is 30 million euros. Initiator Thomas Snellman received several awards for his initiative.
This example shows that it is not technology as such but the aspect of social innovation that creates new opportunities with
a data-driven business model. Data-driven, as ultimately it is the exchange of data in Facebook that generates the business
in REKO.
10 funded by Box 2
Cosun is an example: farmers can register their field data in the software of the cooperative. They then receive
management tips and benchmarking data. At the same time the cooperative uses this data to organise its
logistics, production planning and its marketing (as it can provide its clients with sustainability data). In this
example the pressure of the food and drink industry for sustainable production methods at farm level has
brought the cooperative to the decision to make the software mandatory for its members by 2018.

Value chain integration

In a value chain integration business model the activities in an existing chain are organised by ICT in an
alternative way, as the availability of data makes decision making at another point in the chain more efficient.
An example is prescriptive farming, where some of the decision making moves from the farm (based on local
knowledge or ‘green fingers’) to software at another level in the value chain.

Multinational agricultural biotechnology corporations like Monsanto and Dupont have adopted the strategy of
acquiring start-ups, to adopt this business model and strengthen their existing position. For example in 2013
Monsanto acquired the weather and agronomic data modelling start-up Climate Corporation, which provides
planting advice to farmers based on data science. Monsanto’s primary software product FieldScripts also helps
farmers to maximise productivity, minimise risks and realise higher yields. This move into prescriptive farming
originally also included an investment in a machine manufacturer, but this business has now been sold to John
Deere. Monsanto charges $10 per acre for FieldScripts (Ge and Bogaardt, 2015).

A European example in services is the Dutch-Flemish breeding cooperative CRV, which supports dairy farmers
in insemination decisions for cows. Where traditionally the farmer looked in the meadow if a cow was in heat
and ready for insemination, pedometers have taken over this detection. CRV developed an app that not only
signals this status but also suggests sperm from three preferred bulls. As most farmers always choose option
A of the list of three, they can now subscribe to a service where CRV automatically delivers the sperm, if not
in stock with the farmer already.

Value net creation

Business models that create value out of data by creating new value nets are in essence platforms that link
different groups of clients and support their interaction. Often there is an element of co-creation: the data of
one group triggers activities by the other group and vice versa. Sometimes such platforms have strong network
effects: for users it is attractive to join a platform that other customers have already subscribed to, as in the
case of LinkedIn or Facebook. Some of these platforms create new markets, like AirBnb has done by linking
house owners and travellers.

Within such platforms a market of specialised apps can be created, like in Facebook. In agriculture several
platforms have been set up that create an ecosystem of apps. European examples are 365Farmnet and the
Dutch Akkerweb by the international cooperative Agrifirm. Another example is the EU project FISpace (Future
Internet Business Collaboration Network) which is now available for commercial exploitation by offering
a business-to-business collaboration platform that could link platforms like MyDeere.com, 365Farmnet,
Akkerweb, Agriplace and others via a Linux-like Open Source model. Several EU FI-PPP accelerator projects
like SmartAgrifood, FINISH and Fractals are using the platform.
At the EIP-AGRI seminar, several interesting initiatives that were presented fall into this category of data-driven
business models. Some of these linked farmers and consumers, as in the very successful Finnish example
described in Box 2. Others had a business-to-business focus. One of these is the Dutch example of AgriTrust
(Box 3), which is of special interest because it helps farmers to execute the ownership of their data and reduce
administrative costs.

11 funded by
 Farmers as data owners
Paperwork is important in farming, especially if (tax) accounts are obliged and compliance with
certification schemes (like Global Gap) has to be demonstrated. Much of that data is available on
computers of food processors or input suppliers, in the form of invoices, delivery notes etc.

In the Netherlands, a huge datahub (EDI-Circle: http://www.edi-circle.nl/) is used for exchanging invoices and several other
messages from feed suppliers, dairy companies and the government, to accounting firms and farm management software.
In dairy farming it is also the basis for a farm mineral accounting system (ANCA - annual nutrient cycle assessment). Farmers
control the flow of data in EDI-Circle with authorisations. They authorise companies to provide their documents to other
business partners (e.g. they authorise the feed company to send their invoices in EDI-format to their accountant, their farm
management software and their veterinary). Currently this authorisation is done on paper, but a special website (AgriTrust)
is under development now that a standard EDI message for authorisations is becoming available. This authorisation system
builds trust in (digital) data exchange because it allows farmers to execute the ownership of their data.

EDI-Circle is also used by Agriplace (https://www.agriplace.com/en/), a start-up by a Dutch NGO with a sustainability com-
pliance objective. Agriplace offers farmers a platform to share their certification data with the auditing organisations (like
SGS or Control Union which audit GlobalGap and other standards), the certification bodies (like the Dutch SKAL which certi-
fies organic production) and the food processors (or, in the case of farmers in developing countries, with Dutch importers).

In Farm Digital, a Dutch multi-actor public-private partnership project, 1.000 farmers from the Netherlands, South Africa and
Costa Rica have helped to create this platform as testers and first users.

Box 3

All in all, the different types of data-driven business models lead to a very dynamic landscape in which the
traditional actors of farms, agricultural technology companies (machinery manufacturers) and agri-business
companies (chemicals, seeds, food processors) now also see new players like the classical ICT companies
(large and small) and the venture capitalists that support start-ups in ICT and agricultural technology (as in
urban farming).

Among farmers, the development of new data-driven business models in agriculture has also raised issues
about data governance and the companies that exploit such models. These include questions of data ownership
(legally a difficult concept), data access rights (including the question of farmers if government agencies have
access), but also intellectual property rights on machinery (“Do I still own my tractor if so much of it is a
software licence?”). In some cases there are worries about lock-in effects (can farmers take the historical data
with them if they move to another supplier?). Some of the business models have also raised the discussion
whether big corporations do gain market power on future markets by having access to all farmers’ data, or to
the role of the farmer in the future (does he become a franchiser with the risks but not the returns?). There
are signs that this picture is becoming more clear: companies tend to handle data of farmers as personal
information, and systems are being developed (like box 3) where farmers can influence where the data ends
up. Nevertheless, there is also more work to do, especially in a European context where privacy issues differ
between member states. The privacy aspect of data is only one aspect. It helps farmers to prevent data from
travelling to the public, the government and to competitors that farmers do not want to give data access
to. The other aspect is being able to use the data as a farmer yourself, and combine it with data from other
applications. This can be at this very moment on your management dashboard, or in the future in case you
switch to another equipment supplier.

12 funded by
6. Limiting factors
The EIP-AGRI seminar used several interactive methods to discuss the state of data-driven business models
in agriculture and the food chain. Discussions focused on limiting factors, but also encouraging factors were
discussed (chapter 7).

First of all, there is a lack of awareness of the possibilities and benefits of data-driven applications for agriculture.
If farmers and other operators do not know about the new possibilities, they will not use them. Linked to this
lack of awareness is the need for any new data-driven application to show tangible results for the end user.
There needs to be a clear return on investment, and the potential buyer of an application should have clear
information about its costs and benefits to be able to decide to buy the application or not.

For ICT developers this implies a necessity to focus on farmers’ needs: many current applications do not provide
real solutions in practice, perhaps because the developers do not understand the reality of farmers’ problems
very well. Cooperation with clients in developing data-driven business products is essential: speaking with,
and understanding farmers and other people in the farming and agri-food sector is essential to develop user-
friendly products that address real questions. Especially small farms, where the use of formal management
systems to keep oversight is less needed and where farmers are on average much older, are far from convinced
that they benefit from applying ICT in their business.

Data as such does not always have value, up-to-date agricultural models are needed to turn data into
information. If a farmer knows - thanks to precision farming - that at a certain square meter wheat is not
growing very well, he or she can still be unsure what the cause is (weather can be very local, damage by wild
animals even more) and if more fertiliser (to promote growth) or less (as it is a waste of money on this poor
soil) has to be given.

13 funded by
Another issue related to the awareness and cost / benefit considerations is the speed of development in itself.
This makes farmers and other end users hesitant to buy applications as it is not clear when the best moment
is to buy a system instead of waiting for another one that may be even better.

A second limiting factor is the lack of standardisation and interoperability. This also affects the data that
is produced and (not always) delivered by public organisations (i.e. several data in the realm of IACS or
other weather, agronomic or socioeconomic data). This also hampers the development of new applications.
Integrated solutions are asked for by the sector, with flexibility to move between different systems / platforms.
This is currently not guaranteed.

A third group of limiting factors deals with incentives for sharing data by farmers. Why should farmers share
their data? There should be a clear and ‘real’ benefit for them to do so. Farmers’ trust in how collected data
will be used (also by NGOs and government agencies) is important. To generate value out of data, data often
has to be aggregated, and data from different sources has to be combined. Data sharing is difficult to achieve,
especially when businesses involved think that they can earn money with their own data, and when clear
methods to value the contribution of each other’s data to the common product or service are not available. A
similar reluctance to share data can be observed with public agencies that experience institutional, legal and
technical limitations.

More general is a fourth group of limiting factors which deals with discussions on the governance of data
and the potential effects of ICT on the food chain. Some worry that the increasing use of ICT will change
the division of the value added in the chain and that it will influence the distribution of market power within
the sector. It raises the question what type of agriculture we want for the future (family farms or another
structure). Some participants at the EIP-AGRI seminar stated that the general framework of where we want to
go is not clear. It would be easier to invest with lower risks if there was a clear strategy on ICT in agriculture
and the food chain, with some common infrastructure (e.g. data standards) as a result. If this strategy would
support a structure of family farms, and if agriculture becomes very much based on ICT, the question arises
whether cyber security is sufficiently guaranteed in a system with so many small players.

Appropriate legal frameworks for data ownership are important. Data ownership is in itself a difficult legal
concept, as it only exists in privacy laws for natural persons and intellectual property rights including business
secrets. This issue is certainly not specific for agriculture, although family farms are an interesting entity
between natural persons and (larger) businesses.

A fifth group of limiting factors discussed at the EIP-AGRI seminar relates to financing the investments in data-
driven solutions as costs tend to precede benefits. Pricing structures to spread the payment for data services
over those who benefit, including farmers, consumers, and society in general could be difficult, especially if
environmental or climate benefits are an important societal benefit. Some wonder what the added value is
for the end consumer (the citizens) and the farmer (and the operators in between). Another question is who
will in those cases pay to implement such technologies. The discussion addressed the issue of food prices:
if consumers are not willing ‘to pay the real food price’, how could new costs be charged or how could new
revenues be obtained from the use of these new opportunities? A solution to this problem could be to support
the development of such ICT tools, to lower costs and support innovation.

Public-private cooperation / investment to stimulate data-driven business development could be one way to
go. Finding the right balance between the requirement to share new findings and results paid by public money,
and a market approach would then be important. Discussions at the EIP-AGRI seminar showed diverging
opinions on the need for public investment to stimulate data-driven business, and on the support schemes
that would be most effective.

14 funded by
 7. Encouraging factors and
available support
By far the most encouraging factor for data-driven business
models in agriculture and the food chains is of course the
insight of innovators that problems (whether they are
operational problems at farm level or societal challenges
at large) can be solved by a clever application of ICT. The
problems of today are the business of tomorrow. Several
business models presented by participants at the EIP-AGRI
seminar proved this point. As several aspects of the food
chain are effected by ICT (see for instance the retail and
the growth of webshops) and some consumer segments
are unsatisfied with the current food system (for instance
making them look for more sustainable or regional food),
there are plenty of chances for new value propositions based
on a data-driven business model. However, it is also clear
that innovation is a risky business and that in this confusing
environment many projects do not deliver what their initiators
had hoped for.

Governments offer support for innovators in many ways,

from (financial) support for starters to research projects that
provide results that can be used in commercial applications.
In relation to data-driven business models in agriculture, the
following information of the European Commission – DG AGRI
is relevant (text taken and adapted from EU SCAR, 2015):

At European level, the innovation-driven research approach

and the interactive innovation model are promoted through
the European Innovation Partnership “Agricultural Productivity
and Sustainability” (EIP-AGRI). The focus is on bottom-up
approaches and cooperation between farmers, advisers,
researchers, businesses and other actors in Operational
Groups (OGs) to realise innovations. It is expected that
this knowledge “exchange” will generate new insights and
ideas and will mould existing tacit knowledge into focused
solutions. Such an approach should stimulate innovation
from all sides and should help to target the research agenda.

The EIP-AGRI is challenge-driven, focusing on societal benefits

and rapid modernisation. Like other EIPs, it streamlines,
simplifies and better coordinates existing instruments and
initiatives and complements them with new actions or a more
coherent policy framework where necessary.

Operational groups are the key acting entities in the EIP.

They gather farmers, advisers, researchers, businesses, and

15 funded by
other actors (e.g. civil society including NGOs and governmental bodies). The forming of OGs takes place
at the initiative of innovation actors. No specific conditions are laid down by the EC as regards the size, the
composition and the specific undertakings of an OG. OGs have to draw up a plan, describing their specific
project and the expected results of the project. Furthermore, the OGs have to disseminate the results of their
project, in particular through the EIP network. The exact content of a project plan depends on the actors
involved and the problem, issue or opportunity that needs to be tackled. Innovation brokerage can help to find
innovative ideas, and help partners to connect and set up an OG formed around concrete projects.
For funding concrete innovative actions, the EIP-AGRI is implemented through actions that are mainly supported
by two EU policies: the Rural Development Policy and Horizon 2020. Funding, implementation and prioritisation
of actions all take place through the delivery mechanisms embedded in the respective policies.

Several measures under the Rural Development Regulation 2014-2020 can be used to stimulate innovation
and the activities of OGs. The cooperation measure (Article 35) plays a key role in the implementation of the
EIP-AGRI. Support can be given both for the establishment and operation of OGs of the EIP-AGRI, and for the
implementation of their projects. This support can also be combined with support under other measures such
as training (Art.14), advice (Art.15), investments (Art. 17), etc. The Rural Development Programme can fund
bottom-up innovation projects with a 100% support rate.

Within the societal challenge “Food security, sustainable agriculture and forestry, marine and maritime and
inland water research and the bio-economy” of Horizon 2020, two new instruments were developed that
support the EIP-AGRI: multi-actor projects and thematic networks. The key feature of multi-actor projects
is to address the needs, problems and opportunities of end users and to generate the necessary interaction
between researchers and end users such as farmers / producers, advisers and enterprises by attributing a
clear role to the different actors in the work “all along the project”. This combination of practical and scientific
knowledge should generate innovative solutions that are more likely to be applied thanks to cross-fertilisation
of ideas between actors, and the co-creation and generation of co-ownership for eventual results.

Thematic networks mobilise all concerned actors on specific thematic areas. The aim is to develop end
user material to facilitate the discussion on knowledge, and the dissemination and sharing of knowledge in
an easy and accessible way. This can serve as input for education and for a research database for end users,
making results available long-term. Next to the newly developed EIP-AGRI instruments, a range of existing
instruments will continue under Horizon 2020 (collaborative projects, ERA-NETs, JPIs and COST actions).

The concept of OGs may also be applied within various funding sources. The EIP-AGRI is not exclusively linked
to the Rural Development Policy and to Horizon 2020. There are also potential synergies with other policies like
the EU Regional Development Fund, national or regional funding schemes, private funding, etc.

The EIP-AGRI is supported by the EIP-AGRI Service Point, which connects people and initiatives. Besides
operating a website (www.eip-agri.eu) including more information on the EIP-AGRI and a database of innovative
actions and projects, the Service Point produces several publications and organises interactive seminars (like
this one on data-driven business models) and EIP-AGRI Focus Groups.

EIP-AGRI Focus Groups are temporary groups of selected experts focusing on a specific subject, sharing
knowledge and experience. Each group explores practical innovative solutions to problems or opportunities in
the field, and draws on experience derived from related useful projects.
The reports of EIP-AGRI Focus Groups on precision farming, benchmarking and short supply chains are
especially useful for readers with an interest in data-driven business models.

16 funded by
8. High potential actions to trigger data-driven
business models
The EIP-AGRI seminar on data-driven business models has helped to create awareness that ICT and sensor-
based technologies, procedures and software are being implemented in agriculture and the food chain. They
enormously increase the amount of data that is collected and available in agricultural sectors and throughout
the whole supply chain (from farm to fork). The appearance and adoption of biosensors, nanotechnology, low-
cost electronics, the Internet of Things or remote sensing devices, among others, will further boost the data
domain. The EIP-AGRI seminar confirmed that this scenario offers significant potential for new data-driven
business models. However, it also is clear that we are still in a state of infancy.

Seminar participants welcomed the opportunity to connect initiatives by bringing the community together.
This gives people the opportunity to discuss, test and enrich their ideas while meeting other members of the
community. In line with practices in the ICT community, participants decided to create an online platform to
stay connected and exchange information on Slack: https://agridata-eu.slack.com/

In addition to what innovators can do themselves, there is the question if others, and especially the European
Commission, could take actions that overcome the limiting factors that were identified at the EIP-AGRI seminar.
Based on the limiting factors (see chapter 6) and further reflections during the seminar, five actions are
suggested that have a high potential to promote data-driven business models:

1. Develop an EU ICT Architecture Strategy for AGRI-FOOD to provide guidance on the desired
information
architecture, for a food sector based on family farms and SMEs in food processing. For them, data-intensive
processes to guarantee quality (tracing and tracking, food integrity) are essential. Such a strategy would
provide a more secure investment environment for innovators with data-driven business models. It would
solve ongoing discussions on data ownership and data governance as identified during the seminar (second
and fourth limiting factors in chapter 6). Most likely, the information architecture would stress (and speed
up) interoperability via standardisation at a European level, to make competition and collaboration between
data platforms and apps possible and effective. This could also reduce network and lock-in effects. It would
sketch how farmers and others (e.g. with authorisations and e-recognition procedures) could execute the
“ownership” of their data, to create more trust in big data solutions. It would also sketch how farmers could
be simply connected to different cloud data platforms of agri-businesses and governments, from their own
integrated dashboard (in farm management software). Such an EU ICT Architecture Strategy for agri-food
could probably be delivered by a high level panel (supported by a ‘sherpa’-group) with representatives from
farmers’ organisations, ICT companies (big and small), standard organisations, advisory services, science,
and the food, input and machinery industry as well as NGOs (consumer and sustainability interests, NGOs
on privacy and data protection).

2. Open data and active provision of digital data by government organisations is important for the
development of data-driven business models. In many areas new services have been set up by innovative
ICT companies and others, with the use of open data of governments. In agriculture, weather data, satellite
data and soil maps are some of the examples. However, in many countries more can be done, as mentioned
already in chapter 6. Examples of good practices include paying agencies that make land use data at
individual field level (from the applications for CAP subsidies as stored in the IACS system) collected over
the last years available to the public. FADN data on regional average farm results could be made available
in a format that is easy to use in benchmark software. The same holds for inspection results from food
safety authorities or customs data (to make markets more transparent). Government agencies should also

17 funded by
 make it possible for farmers (or their advisers) to submit data in digital form from their farm management
software. Also, if they use the website of the government agency to enter data, farmers could forward
their digital data to their business partners, such as advisers and agri-businesses. If farmers have informed
the paying agency where their sugar beet fields are, it should be easy to inform their sugar company and
their service contractor who is taking care of the harvest by forwarding the data entered in the government
system. Developing best practices in this area of open data and active provision of digital data could be
stimulated by a thematic network or a multi-actor H2020 project. This does not need to wait for the EU
Strategy advocated under point 1. Actions, such as a seminar, encouraging Managing Authorities to support
Operational Groups in the data domain could also be beneficial.

3. Given the discussions on data ownership and data governance, it makes sense to start an action on
this point, with or without the EU strategy advocated under point 1. This action should not only study the
legal aspects, but also look for best practices and realise solutions. Denmark has strong leadership from
the cooperatives and farmers’ organisations in this field, which may be one of the potential solutions. In
the Netherlands a solution with authorisations and digital identification seems to work (see Box 3), but
companies are asking for international solutions. Best practices could be explored and stimulated by a
thematic network or a multi-actor H2020 project.

4. The development of data-driven business models could benefit from more in-depth analysis of what works
and what does not work. There are still clear gaps between business developers and final clients at needs,
knowledge and technical levels as mentioned in the first type of limiting factors identified in chapter 6. Short
presentations, as presented at the EIP-AGRI seminar, help to identify interesting cases but do not provide
enough detailed information to understand why a particular solution is working and whether it can be
scaled up or copied to other regions or sectors. This asks for a more in-depth analysis of data-driven
business models with formal methods like Osterwalder’s Canvas model or Harvard case studies. In the
current H2020 call for proposals there is a call to submit (by February 2017) a proposal for a multi-actor
project on business models (not necessarily data-driven) in agriculture. If this call leads to insufficient
attention for data-driven business models, a follow-up call specifically oriented at data-driven business
models might be interesting.

5. Data is not information. The deluge of data coming out of all sensors and other Internet of Things sources
will not automatically lead to better decision making by men or machines. Often agronomic models are
needed that interpret the real-time data. Many scientific models are from an era without such abundant
real-time information and are based on measurements in specific research sites. This implies that science
also has to learn to cope with big data and the needs from practice for improved models based on real-time
data. Programming the most urgent needs for updated agronomic models in a research agenda for real-
time agronomic models is a useful action. This can for instance be done by an EIP-AGRI Focus Group or
a working group in SCAR (the EU’s Standing Committee on Agricultural Research), to support data-driven
business models. This agenda should also identify how to incentivise scientists to make sure they do not
only make their publication available as an open source, but also their research data, their source code
and - in a multi-actor approach - set up big data advisory services that apps can tap into. Researchers,
who work in a heavily competitive market, often have incentives for not making their data and source code
available, and funders might consider changing this. At the same time, this knowledge has to be properly
delivered to and interpreted by the ICT experts. A close collaboration between content experts and ICT
experts is needed, and actions promoting Operational Groups in the agriculture data domain will facilitate
such collaboration.

18 funded by
 References
EU SCAR [2015], Agricultural knowledge and innovation systems towards the future – a foresight paper,
Brussels.

Ge, L. and M.J. Bogaardt [2015], Bites into the Bits. Governance of Data Harvesting Initiatives in Agrifood
Chains. Paper prepared for presentation at the 148th seminar of the EAAE, “Does Europe need a Food
Policy?”.

Osterwalder, Alexander, Yves Pigneur, Alan Smith, and 470 practitioners from 45 countries [2010]. Business
Model Generation, Wiley

Perez, C. [2002] Technological Revolutions and Financial Capital: The Dynamics of Bubbles and Golden
Ages. Edward Elgar. Cheltenham. United Kingdom.

Poppe, Krijn J., Sjaak Wolfert, Cor Verdouw and Tim Verwaart [2013]: Information and Communication
Technology as a Driver for Change in Agri-food Chains in: EuroChoices vol 12. Nr. 1, 2013 pages 60–65

Poppe, Krijn, Sjaak Wolfert, Cor Verdouw and Alan Renwick [2015a]: A European perspective on the
economics of big data in: Farm Policy Journal, Vol. 12, no. 1, autumn quarter 2015 p 11-19.

Poppe, Krijn and Elke Saggau: ICT as a driver of change in the agri and food sector [2015b] In: EU SCAR
(2015).

Poppe, Krijn J., Marc-Jeroen Bogaardt and Tamme van der Wal [2016]: The economics and governance of
digitalisation and precision agriculture. Paper for European Parliament’s Science and Technology Options
Assessment Panel, 2016.

Van’t Spijker, A. [2014], The New Oil. Using innovative business models to turn data into profit. Technics
Publications.

19 funded by
 The European Innovation Partnership ‘Agricultural Productivity and
Sustainability’ (EIP-AGRI) is one of five EIPs launched by the European
Commission in a bid to promote rapid modernisation by stepping up innovation
efforts.

The EIP-AGRI aims to catalyse the innovation process in the agricultural

and forestry sectors by bringing research and practice closer together
– in research and innovation projects as well as through the EIP-AGRI network.

EIPs aim to streamline, simplify and better coordinate existing instruments

and initiatives and complement them with actions where necessary. Two
specific funding sources are particularly important for the EIP-AGRI:

• the EU Research and Innovation framework, Horizon 2020

• the EU Rural Development Policy

funded by

Join the EIP-AGRI Network & Register via www.eip-agri.eu

www.eip-agri.eu +32 2 543 73 48 [email protected] Avenue de la Toison d’Or 72 1060 Brussels Belgium