Anticipating market-defining peaks
Coöperatie Hoogstraten, a leading growers’ cooperative, faced a familiar yet critical challenge. During the strawberry high season, they handle substantial peak volumes. Knowing exactly when these peaks will occur is essential to align commercial strategies with retailers and secure optimal prices for growers. With accurate foresight, promotional campaigns, planning, and distribution can be further optimised, increasing performance and resilience.
Our approach: AI as a strategic enabler
At Möbius, we believe in challenging the status quo to deliver breakthrough solutions. In close collaboration with Coöperatie Hoogstraten and Moet Lukken, we designed a proof-of-concept to estimate the forecastability of the peak volumes and estimate the ROI of the project. Starting small and gradually increasing the complexity is our preferred way of working. No big commitments are needed and the client is fully in charge.
Our solution combines advanced statistical techniques with machine learning algorithms. This way we were able to create a robust ensemble model that draws from more than 200 external and internal variables. These include weather conditions, crop characteristics, and historical acreage and yield data. We started with a statistical decomposition of the historic yield numbers and identified together with the strawberry experts of Coöperatie Hoogstraten relevant factors that could explain certain patterns and potentially the remaining noise.
In every step of the development and implementation, our client and partner were highly involved. The combination of business knowledge, strawberry expertise and data science resulted in a highly accurate forecasting engine that is integrated into a user-friendly web application and embedded seamlessly into day-to-day operations at every iteration.
From data to measurable impact
The AI-powered forecasting solution transformed Coöperatie Hoogstraten’s strategic positioning, with results that directly strengthen its market performance:
-
More optimal pricing – with an accuracy of ~90% in predicting the crop’s yield, the forecast enables us to take a more strategic and proactive approach to our commercial activities. The forecast itself does not automatically result in a better price — commercial action is still required for that. This includes planning and timing campaigns, programmes and promotions. Whether a major campaign starts one week earlier or later can make a significant difference if it falls just before or after a peak week. By anticipating this wisely, we can get more value out of the available volume and minimise the risk of price pressure.
-
More accurate coordination of volumes – A reliable forecast also enables better coordination with customers on desired volumes, packaging and logistics. This helps prevent both surpluses and shortages.
-
Reliable supply & stronger market trust – Unexpected shortfalls decreased, reinforcing the cooperative’s reputation for reliability and strengthening long-term retailer relationships.
Contrary to common misconceptions, advanced forecasting does not require expensive hardware or controlled environments. Our solution leverages existing data sources and is fully applicable in open-field agriculture. The result is a cost-effective, pragmatic innovation that makes cutting-edge AI accessible across the agricultural sector.
