The challenge

Northwind ran forecasting on spreadsheets maintained by three regional teams. Errors compounded across the network, leading to empty trucks on some routes and missed deliveries on others.

Our approach

We started with a two-week clarity sprint to map data sources, interview planners, and size the opportunity. From there we scoped a pilot on the two busiest regions, with planners involved from day one.

• Unified shipment and order data into a single model

• Built a forecast that planners could question and override

• Set a clear success metric before writing any code

The outcome

The pilot reduced forecast error by 23 percent within its first quarter. Northwind is now rolling the approach out to every region, owned by its own planning team.