Shipments Forecast Disaggregation Considering Weighted Time Buckets

Summary:

Our company is very seasonal. We sell agricultural products to small customers. Often, customers will place one or two orders for the year and show shipment history in 2 out of 12 months of the year. When shipments forecast is calculated at the aggregate level, the forecast follows some of the aggregate seasonal trends. When the shipments forecast is then disaggregated to the customers' for their items, we see the forecast smoothed out for the year. The total year volume looks appropriate compared to historical trends, but the forecast in the individual months are way off. Ideally, we'd like to