Optimisation of production planning
How a monthly sales forecast with an 18 month rolling horizon let a manufacturer plan production against expected demand, cutting delivery bottlenecks by 73 percent while average inventory came down 2 percent.
Published results
73% fewer bottlenecks
Delivery bottlenecks after production planning was rebuilt around the forecast.
2% less inventory
Average inventory fell at the same time. Stock buys lead time at the cost of liquidity, so lower inventory alongside fewer bottlenecks is the harder half of the result.
18 months ahead
The forecast horizon, rolled forward monthly from global sales history and external market data.
The business challenge
A manufacturer was experiencing high production volatilities and long lead times due to challenges with the co-ordination of sales forecasting and production planning.
The solution
A new sales forecast model was built using DetectX® Business Monitoring and the underlying predictive analytics capability of the DetectX® platform. Prospero were able to create predictive models that forecast sales volumes based on sales history data and external market data. Based on global sales data, this model is run on a monthly basis, providing an 18 month rolling sales forecast.
The results
Based on the new forecasting model the manufacturer was able to optimise production planning. This reduced delivery bottlenecks by 73 percent while at the same time reducing average inventory by 2 percent. That balance is the point: high stock levels can improve lead times but they tie up liquidity, so the challenge is to keep inventory low while still improving delivery lead times to customers.
Where to look next
- Beyond financial services
Where DetectX® is applied outside banking and insurance.
- DetectX® platform
The predictive analytics layer these models are built on.
- Early detection of delivery delays
The same module, finding the factors behind late deliveries rather than forecasting demand.
- Process optimisation in aluminium company
An early warning model on input data measured before the production run.
Feed a sales forecast into your production planning
No scripted demo. A working session on your own data and your own questions.