Executive Summary
TerraTrac Equipment Parts — demand forecasting and supply planning review. Scope: 7 planned SKUs across 3 ABC classes and 4 categories, 4 distribution centers, 5 suppliers, 104 weeks of sell-through and 220 purchase orders.
Demand is predictable to within 23.1% weighted error, but inventory is not positioned to serve it: 26 of 28 SKU x DC positions sit below reorder point and 16 cannot cover one supplier lead time. The exposure is concentrated in Undercarriage, sourced single-vendor from the slowest and least reliable supplier in the network. Releasing a $8.16M catch-up buy and re-basing safety stock on measured lead times closes the near-term gap; a secondary Undercarriage source closes the structural one.
Demand is forecastable — the current planning basis is not using it
A trend + seasonality + promotion model delivers 23.1% WMAPE per SKU x DC on a blind 8-week holdout, against 24% for the moving-average approach it replaces — and, unlike the moving average, it prices in promotions and the annual cycle instead of chasing them a quarter late.
Coverage is below policy almost everywhere
26 of 28 reviewed SKU x DC positions are below reorder point and 16 cannot cover a single replenishment lead time. Cover runs days, not weeks, against lead times of 14-33 days, so a normal supplier slip becomes a customer-facing stockout.
One supplier carries most of the risk
Trackline Manufacturing (Undercarriage) averages 33.4 days with σ 10.0 days and a 51-day P95 — the slowest and least consistent source in the network, and the cause of the six-week Asia Pacific outage in weeks 40-45 that censored demand history.
History must be corrected before it is trusted
Six stockout weeks across five Undercarriage SKUs at two DCs understate true demand by roughly 100-200%. Left uncorrected, the forecast lowers safety stock precisely on the least reliable supplier — compounding the exposure rather than covering it.
Recommended actions
| Action | Owner | Timing |
|---|---|---|
Release the catch-up replenishment plan $8.16M across 7 SKUs x 4 DCs, MOQ and pack size respected. Expedite the 16 critical positions. | Planning | This week |
Adopt the seasonal + promo forecast as the planning basis Weekly refresh, WMAPE and bias tracked per SKU; promotions entered as a forward regressor rather than absorbed into the baseline. | Demand Planning | Next cycle |
Re-parameterise safety stock from actual PO history 95% service level network-wide, 97.5% for Undercarriage at Chennai and Jakarta, planned against P95 lead times for Trackline. | Inventory | 30 days |
Qualify a secondary Undercarriage source Dual-source Track Shoe and Track Chain at a 70/30 split; put measured OTIF into the supplier QBR for all five vendors. | Procurement | 90 days |
Method in one line
Weekly demand per SKU modelled as linear trend + 3-harmonic annual Fourier seasonality + promotion dummy, fitted by least squares on censoring-corrected history and validated on a blind 8-week holdout. Reorder points combine lead-time demand with a 95% service-level safety stock that accounts for both demand and lead-time variability, using lead times measured from 220 actual purchase orders rather than supplier master data.
Open questions for leadership
- Is 95% the right service level for C-class items, or should they run leaner?
- What working-capital ceiling applies to the catch-up buy?
- Is there appetite to qualify a second Undercarriage vendor this fiscal year?