Knowledge Management in a Cooperative Farm

Mediterranean dinner at a farm – source

Team: Marius Ursu, Paul (Herenboeren)

Client: Herenboeren Wilhelminapark

Description of the challenges and solution

Poster summarizing the project

In a HerenBoeren farm, the owners – 150 families – are also the consumers, therefore eliminating the middle men and revolutionizing the supply chain. Members have a word over what and how it is grown. In 2018 the “herenboeren” had access to seasonal vegetables, meat (beef, pork) and eggs (chicken). The farm is located in Boxtel and more are in development in Weert, Rotterdam, Ede, etc.

 Storytelling video about Iteration 1

The project facilitated the learning that takes place between the members of the community. It consisted of a digital platform which allows the users to share knowledge of tips and recipes. The sharing leads to a stronger community and a more meaningful experience. Users are presented with a personalized homepage which recommends relevant tips and recipes. The content is seasonal and based on the patterns in the family history.

This is valuable because the members take the food directly from the farm, which is not something the modern human is used to. It requires a set of skills to clean, store and cook the vegetables and meat. 

Process overview

Starting from the Co-founder’s description of the concept and it’s challenges, I conducted wide need finding research among the users through: safaris, observations, and interviews.

The data was analyzed into insights and placed on an Journey Map, later validated with the users. This was followed by framing a challenge (not previously articulated by the founder) and testing it’s perceived usefulness with members of the community through a rapid click-through prototype.

The attempt to complement the physical user journey with a digital one was validated, resulting in a later iteration involving members of the community as designers, engineers and content creators (by the community for the community). The solution enabled knowledge sharing of recipes and tips through a live, high-fidelity prototype.

The resulting work is in production, serving real users, with the usual administration, maintenance and continuous tinkering challenges that come with the real world.

Project outcome

A high fidelity prototype – implemented with real users

Homepage – Live and implemented with real users in the community – Software:  WP Ultimate Recipe with custom CSS

Starting point

The initial challenges we first discussed with the Co-founder were wide and multiple, among which: problems with planning the crops and harvest; educating other farmers on other farms of how to run a farm. After the initial user research using contextual interviews and shadowing, a set of frustrations became clear. The members had an unmet need of avoiding waste and improvising better with what nature provided in terms of seasonal vegetables. They also wanted to have a stronger feeling of a community.

User Journey starting every week on Saturday

It is important to note that the founders or the management board were not fully aware of all these issues (in red), which were only discovered during my initial research observations and proposing an initial prototype.

These issues would later become the defining characteristics of a new design, which had impact over the Value Proposition of the farm, and over the Strategy / Market in which it is active – starting the move

>> from a farm which sells products on a subscription basis with a community membership

>>>> to a service which offers a communal cooking and eating experience, a learning and confidence-building experience, and a feeling of being part of a (food) tribe.

User Research

The Research consisted of observation at the farm (tens of people), shadowing 1 user at home, informal discussions with many members who are either designers or operational excellence professionals and about 5 contextual interviews.

Picking up the products at the farm on Saturday

Fridge storage after the products are cleaned

Room temperature storage after vegetables are cleaned


After the initial discovery, the direction was chosen to be:

To facilitate the learning that takes place between the members of the community by building a platform which allows the users to share knowledge of tips and recipes. 

This is valuable because the members take the food directly from the farm, which is not something the modern man is used to. It requires a set of skills to properly clean, store and cook the vegetables and meat. These skills are very contextual and procedural (learned by doing, not reading or seeing), and so it matters to have a touchpoint at home where dynamic information is easily accessible, as opposed to having educational discussions only on Saturday at the farm.

The members wish to avoid waste and improvise with what nature provides in terms of seasonal vegetables. They also wish to have a stronger community, and the knowledge sharing helps in accomplishing those goals.

The challenges that could only be enumerated after the first Co-Reflections

As is often the case, the challenges above became clear only after the first iteration was presented for user evaluation in the form of a rapid prototype (Adobe XD mockup); paired with an experience flow in order to stimulate co-reflection.

Rapid Prototype – Iteration 1

Homepage – Iteration 1 using Hellofresh recipes and Adobe XD

The role of the first iteration was to test the MVP described in the User Research

It came to be through ideation with the use of tools like: scenarios, storytelling, which slowly led to functionality and wireframes.

Sometimes a bit of acting out can do wonders for the process and getting unstuck. Below you can see a set of tangible figurines which were useful in acting out and storytelling for ideation purposes; as well as for presentation and the concept video.

The latest iteration of the platform can be found at the top of this page. It is different in that it is of higher fidelity and more precise in the features it offers.

A newer version is currently running live in the real world with real people contributing and using it.

In depth User Testing is not described on this page, but suffice to say, following these prototypes, I discovered a group of people within the community (led by Paul) who were aiming to build the same platform, and we joined forces. Therefore, this work was indeed desired, and realized in a collaborative manner.

Machine Learning

The Machine Learning algorithm was prototyped in order to make possible the “recommender” behind the assistant. One of the limitations for the content moderator of the platform was that they had to manually propose a selection of content. Using an algorithm enabled us to offload that responsibility, and opened up the possibility of a customized experience – which was expected by the community members

Visitors during Demo Day had the opportunity to tangibly act out the algorithm and pretend to be in the mind of the computer.

Acting out the machine learning algorithm in a tangible way


In retrospect, My role as a designer became to express what the members wanted in the first place when they joined the community, but found hard to say by themselves. I would later take into account how their experience changed over the time of being a member: my design would later serve as a connector between those who had a good experience to those who were still frustrated. I was aiming to help them organize and problem solve in this decentralized organization. They have been doing it before I arrived, and would continue to do so after I left.