Lean Thinking in Scaling Agricultural Technology: The Importance of Right-Sizing Solutions for Different Stages of Business

When building and scaling Ukko Agro Inc. (https://ukko.ag/) from the ground up, we encountered numerous challenges. One crucial lesson we learned from this journey is that solving a problem at different scales presents a different set of challenges. Thus, it is vital to appreciate the value of properly sizing the solution to the given scale. In the agricultural technology field, this is particularly significant because right-sizing a solution does not imply releasing an incomplete product; rather, it means delivering a complete solution that meets the user’s requirements. For example, if we need to travel from point A to point B, releasing a skateboard instead of a car could be the appropriate solution for the given scale.

When creating ForeSite™, our predictive analytics platform that collects weather data from both weather stations and a Weather API, in addition to the other data sources, we had to address a fundamental question for our customers: how many weather stations are necessary to produce a high-quality prediction for the entire farm, and where should these devices be placed to capture the highest risk across various areas of the farm? In this post, we will explain how we applied lean thinking to tackle this challenge, size the solution for different stages of our business, and how each solution stage is currently being employed throughout the customer journey.

Establish a culture of lean thinking. It is the key to successful execution, particularly when strategies fall short

The culture of lean thinking emphasizes efficiency, continuous improvement, and delivering value to customers. When fostering this culture, organizations can optimize processes, minimize waste, and adapt quickly to changes in the market. In our case, we understood the challenge but not the problem well enough to even know where to start. We aimed to make fewer mistakes and only new ones. The lean culture we adopted involved a mindset of problem-solving and experimentation, working closely with customers to identify and address issues quickly and effectively. This cultural value has enabled us to achieve successful execution even in the face of obstacles or setbacks.

Solve a problem manually to validate your assumptions and see through that the results connect with the target outcomes.

Initially, we developed an algorithm that factored in variables such as field topography, sun direction, and more to determine the optimal location for installing devices that could best track disease-conducive conditions and the conditions for the plant growth stages. Through manual computation and customer feedback, we validated the results, which served as the foundation for the Device Placement Logic in our system. For the size and scale of our business, we manually configured each weather station to transfer data to our servers and link with each and every field for a given farm. For the size and scale of the business, this was good enough. However, configuring each device manually in our system was not a sustainable approach as we aimed to scale up our business. Thus, we recognized the need to automate the process to facilitate our scaling efforts.

Automate for the target scale when the manual approach becomes too expensive (both in time and opportunity cost)

We updated our computing algorithm to enable our platform to suggest the ideal location for weather station installation based on user inputs, using automated device placement logic. We also automated the weather station configurations so they could be onboarded onto our system using a bar-code scan or GPS mapping. This was the first iteration of our automated device placement logic. In the customer journey, this is still a preferred method for smaller-scale customer operations looking for maximum disease pressure coverage over a local area. The data from these weather stations automatically feed our quantitative models that are monitoring several fields. Automating these processes allowed us to scale up by reducing the need for hours of manual work.

Adjust the entire solution and the processes, not just a single feature when the scale changes for a solved problem

As our customer base expanded (both in scope and size), and they began onboarding hundreds of weather stations, we faced the challenge of balancing their weather station budget (capital expense) with the need for acceptable system-level accuracy for our quantitative models. We evolved our solutions to suggest the optimal number of weather stations to cover the maximum area within their budget. Our automated platform allowed us to offer different solutions to different clients.

With this iteration of our Device Placement logic, we also minimized the required effort (friction points) from users while still providing acceptable system-level accuracy. We introduced the ability to validate crop staging and provide region and field-specific models. We also provided our customers with reports to ensure the accuracy of the presented data.

By solving the challenge of balancing client budgets with system accuracy, we built a solution that brought weather stations from across the region into a shared (private or public) weather network, benefiting our entire customer base.

Observe customer behavior carefully by segments for the next growth opportunities and apply lean thinking for building on the existing foundation

The latest iteration of our Device Placement logic now includes a weather grid option, allowing users to set up weather networks using grids. This option has become the preferred choice for our large-scale retail clients who oversee a large number of farming operations with thousands of fields. By using the weather grid in combination with several products on our platform, our clients have experienced numerous benefits, such as providing their clients with more localized weather data, relevant to their fields, combined with agronomic insights to grow successful crops. One of the key advantages of the weather grid for our clients has been the ability to attract new customers to their business. This new iteration of our Device Placement logic is a result of years of determining optimal device placements and previous iterations.

In summary, as you build a solution and scale it, different sets of challenges will arise for the same problem. Applying lean thinking is a good starting point, but to evolve your solution effectively, it’s important to observe customer behavior and understand their needs at each scale. Before building or evolving your solution, take time to validate your assumptions and partner with your customers, as much as possible, to build a minimum viable product for each stage.