Seasonal forecasting for fertilizer demand
A global fertilizer manufacturer
Legacy demand forecasting process for the clients global operation was excel-based and heavily dependent on personal judgement. This resulted in low accuracy of 20-30%. Inventory and place replenishment orders from point of sales suffered as a result.
We followed a ‘Prepare-Explore-Model’ process.
Built a time series model to analyze the past data and forecast 6 months ahead
External factors impact was introduced in the process
New system suggests a point forecast and four range forecast for each month in the horizon.
Forecast are generated in a form that integrated with their Inventory Management Systems (IMS)
Overall forecast accuracy improved by 50%
Huge time savings, with optimized inventory planning that freed up working capital.