A Decade of Crop Monitoring

The Genesis

When I founded Geocledian in 2013 I never thought about what it would be like in 10 years. 5 years maybe, but 10 was far too far away. We were a startup at this time and I knew that only a small number of companies would survive. To make things worse I didn't even have a clear business idea. But I was proud of the name I found after a little while of searching: Geocledian. I derived it from "Euclidean geometry" by swapping the "i" and the "e". "Cledian" sounded like a roman name, which I wondered why it actually didn't exist. But it works well in most latin influenced languages. But the best thing about Geocledian was that a Google search resulted in a blank page. Something I've never had seen before. Done!

To fill the gap of the missing business idea, I remembered an idea that I had developed some time ago in a previous start-up. At this time remote sensing analysis was a time consuming project work. It took usually a few months for experts to process gigabytes of data, analyse them and produce fancy maps or dull statistics. Why not making it much simpler? Most users were just interested in their small piece of land and didn't care about the hundreds of square kilometres around them. So I needed to think it from this side. Request the data for just these few fields and get back almost instantly all usable images with no clouds and analysis ready. It also should be possible to use this data with any standard software and not just with the specialised packages like ERDAS or ENVI, which were popular at this time. Hence, I needed to use a simple REST API and no OGC standard, which nobody outside the geo community understood and no supporting frameworks available. The idea sounded weird at this time, two years before Sentinel 2 was launched, and it was.

Bootstrapping to release Version 1.0

The first version I built myself in the evenings hours. I was using PostGIS, as it provided imaging processing capabilities inside the database, however in SQL, which turned out to be not the best idea in the long run. But I don't want to got into the technical details. To learn more about our technological Genesis please check Johannes article on the technological evolution of Geocledian.

Since I did a lot of consultancy work in parallel to earn the money - I worked on the technical design of CloudEO together with Martin - I needed to find somebody to continue my work. That was when Toni joined. He became the first employee of Geocledian, even before I was one.

To raise some money I had to sell the idea first - before being able to sell the product. Bringing an investor in was not an option after my previous experiences with this type of financing, so bootstrapping was the preferred choice. ESA BIC together with a project fund called SmartAgriFood were our first supporters. Both with different but similar strange underlying conditions. We had to move to Germering, near DLR, and to use FIWARE. Both luckily did not persist, but the money allowed us to develop the first version of AgKnowledge, how the API was now called. It served cloud free NDVI time series based on satellite images from Landsat 8 for each field requested and it was released in 2016.

First Customers

Now selling of the new product had to start. The team was enforced with Florian, Johannes and Dimo. Our first customer was ABACO, with whom we had previously built a business relation ship through various consulting projects. Almost at the same time CropIn, an ambitious startup from India joined. I never imagined to sell IT services to India, before I met Krishna in Abu Dhabi. CropIn requested us to provide AgKnowledge for the whole of India. This was an early success but presented us with some technical challenges as you might imagine.

The customer base grew. Several other startups joined, mostly from the new sector of upcoming farm management systems. But big names were not coming in. At the same time our competition grew. The big cooperates like Bayer started to build their own advisory systems from scratch following the hype fuelled by the launch of the Sentinel 2 system. Unfortunately these were not so interested in our service, as they had the power to built everything themselves. Today I know that it was too early for our understanding of the business. Placing an API like ours within the digital supply chain requires a more mature market and industry. The specialisation of the individual supply chain players was not yet present.

Now our bootstrapping approach paid off. We didn't have to match high ROI expectations of investors. We stayed independent and had the time to improve our product, while others vanished or got absorbed.

Growing up

In this constellation we had time to develop a broad expertise and a more powerful platform. We removed the technical and conceptional flaws of the early days. Also the market got more mature. The big promises got scaled down by reality. Selling data services is not as easy as selling green, red or blue horsepowers. Also the big names had to learn that. But data can be as powerful if applied in an intelligent way.

We learned our lessons as well. Just providing data was not enough. Analytics and actionable results were required. The service portfolio got much more complete. Now it ranges from the sub field level up to the regional scale. The end-user needs and not only the ones of enabling IT systems had to become our main focus. Technology must serve the solution and never the other way round. And we had to strengthen our marketing and sales activities.

Today AgKnowledge together with the products built on it form a solid offering around crop monitoring, serving farm management systems, the processing industry, input suppliers, certification bodies, insurances and governments alike.

The Road Ahead

The last 10 years have been an exiting journey. Some of our first customers we still serve. We built many personal relationships in the community, no matter if partner, competitor or customer. We had many great encounters.

New challenges are lying ahead in the coming years. New sustainability regulations like EUDR or ESRS require implementation. Climate change needs to be addressed and the sustainability of agricultural production improved. Industrial technologies caused many troubles we have now to overcome, but technology can also help to solve these. Data analytics can help us to understand what is going on, help us to change our behaviour, and help to improve lives of the millions of smallholders our global agricultural production heavily relies on. I'm excited that we can help to drive this change. And I'm happy that the best team I had to date is working with me to create this impact. Now we can source from our 10 years of experience to build the custom fit solutions our customers need.