Forecast
Seasonal regression, and Croston for intermittent demand, chosen by a rolling backtest. Accuracy and bias are reported against a naive forecast, which is often harder to beat than expected.
The Supply & Planning module covers forecast to stock, order to cash and the inputs to S&OP. Forecasts are chosen by backtest against a naive baseline, inventory policy is set per SKU, and excess stock is valued as capital tied up, never as profit.
Seasonal regression, and Croston for intermittent demand, chosen by a rolling backtest. Accuracy and bias are reported against a naive forecast, which is often harder to beat than expected.
Days of inventory against policy, the share of SKUs above policy, stock-outs, stock stuck in the channel, and how much of the excess comes from forecast bias.
Safety stock and reorder points derived from a target service level, with coverage policies by ABC-XYZ class.
Receivables aging next to inventory, so the working capital picture is complete.
Which forecast should planning trust?
Rolling backtest against a naive baseline · WMAPE and bias
How much should each SKU hold?
Safety stock from service level · ABC-XYZ and newsvendor coverage
Where is capital tied up, and why?
Days of inventory against policy · channel stock · bias to stock
What does fixing it free up?
Valued as one-off capital, never as recurring profit
Where do demand, target and capacity disagree?
Actual against target against capacity
In development: constrained S&OP planning and forecast challengers. Delivery service (OTIF and fill rate) is reported only when delivery data is measured.
A Value Tree in Supply & Planning sizes the capital and the service at stake in 14 days.