We run a multi-plant network with 20-26 week components, and most trainings I’ve taken stop at EOQ/ABC - fine for firefighting, not for forecasting seasonal demand and sizing safety stock across echelons. If you’ve done CPIM Part 2, IBF’s forecasting cert, MITx MicroMasters, or DDMRP, did it materially improve your long-range planning and S&OP decisions, or is there another program that truly moved the needle?
EOQ/ABC wasn’t cutting it for our 20–26 week components either — ugh — CPIM2 helped the S&OP cadence, but the MITx MicroMasters analytics moved the needle. We layered seasonal decomposition with a basic multi‑echelon safety stock calc and set forward coverage to lead time plus P80 forecast error; expedites dropped about 25% — do you have any MEIO tool or just ERP fields?
Co-sign @nicole29 — what really helped our 26‑week components was switching from more inventory to safety time: set seasonal indices at the family level, then push a 1–2 week time buffer into MRP during peak months to smooth releases and cut expediting; have you tried time buffers yet?
But we deal with half‑year supplier lead times too, and the IBF forecasting cert was the first thing that improved our long-horizon calls — EOQ/ABC just drives me nuts there. The big win was running Forecast Value Add reviews and moving to probabilistic plans — “treat lead time as a distribution, not a number” — so we pre-build only when the percentile risk justifies it. DDMRP helped execution buffers, but not the 12–18 month view; if you try IBF, their FVA primer is a solid start: https://ibf.org/knowledge/glossary/forecast-value-added.
For our ‘multi-plant network’ with 6‑month parts, MITx SCM (SC2x/SC3x) was the first thing that forced us to model multi‑echelon safety stock and test seasonal profiles — it changed our long-horizon calls. CPIM Part 2 gave structure but not the math; DDMRP helped execution buffers, not sizing months out. Quick win meanwhile: run Forecast Value Added and use family‑level seasonality with variance‑to‑lead‑time scaling; IBF’s explainer is decent: https://ibf.org/knowledge/fva.
Biggest lift for our long-horizon calls came from setting service time targets by node and using the square‑root rule to decide where to hold buffer; MITx SC3x gave me the tooling to test that and we shifted about 30% of stock upstream — like moving sandbags uphill before the storm. DDMRP smoothed execution for us, but we still needed a seasonal family profile feeding pre‑S&OP — “plan variability, not averages”; @thomill, have you tried service time instead of a flat service level?
Quick example: DDMRP plus a simple prebuild calendar did more for our 26‑week parts than CPIM — setting decoupling points and using “variable buffer profiles” by season let us size stock by node without overcooking FG. We paired it with IBF’s focus on WAPE to pick winners among models, but caveat: you’ll still need a multi‑echelon calc (I liked this MITx SC3x module: Supply Chain Management MicroMasters® Program) to set targets. Felt like switching from a flashlight to a headlamp.