Medium

Nicolas Vandeput

Featured Book

Demand Forecasting Best Practices

Demand Forecasting Best Practices

Demand Forecasting Best Practices teaches you to optimize demand planning to deliver a more effective supply chain.

2023

Nicolas Vandeput

Nicolas Vandeput

Stories

Pinned

Nicolas Vandeput

·

Jul 22, 2024

Demand Planners Rulebook

What should demand planners do and not do? First, we emphasize critical practices for demand planning (viewing forecasting as an…

Engraving of a crowded academy: scholars in robes study globes, maps, geometry drawings and books, and debate in small groups among columns and steps.

Pinned

Nicolas Vandeput

·

Jul 5, 2019

Forecast KPI: RMSE, MAE, MAPE & Bias

The article below is an extract from my book Data Science for Supply Chain Forecast, available here. You can find my other articles here:

An archer draws a bow and aims at a row of round targets in an indoor range, a picture of aiming a forecast at the real demand.

Nicolas Vandeput

·

Aug 19

How Demand Planners Should (Not) Collaborate with Finance

Every company runs on several numbers at once: a budget, sales targets, a supply plan, and a demand forecast. Sooner or later, someone…

Hand-coloured satirical print: gentlemen in top hats peer closely at paintings of pigs and country scenes that cover a gallery wall.

Nicolas Vandeput

·

Jul 15

Should You Forecast Demand per Customer? (Usually Not)

Sooner or later, someone asks, “Can we forecast demand per customer?” It usually comes from sales or finance, and it sounds perfectly…

Old painting of a blue and yellow macaw perched by a tree trunk, with a small brown songbird singing on a branch behind it.

Nicolas Vandeput

·

May 11

Outgrowing the Safety Stock Formula

The classical safety stock formula, Ss = zσ√(L + R), which most software vendors and consultants use, is based on multiple assumptions and…

Safety stock formula in five levels, from worst to best: zσ√L, zσ√(L+R), z·RMSE·√(L+R), z·RMSE, then k·RMSE or k·MAE, and level 4: forecast-driven policies and probabilistic forecasts.

Nicolas Vandeput

·

Mar 30

Cumulative and Lag-1 Forecasts are the Most Important

Most supply chain leaders analyze forecasting accuracy over a single (or just a few) lags. This article explains why cumulative forecasting…

Summary box beside an old engraving of a print workshop. The article explains why cumulative accuracy beats period accuracy, why lag 1 matters most, and how to measure cumulative accuracy.

Nicolas Vandeput

·

Mar 23

Forecasting Variability: Causes, Solutions, and Why It (doesn’t) Matter

Earlier in 2025, we started measuring forecasting variability (or stability) in our projects at SupChains. As we discussed the subject…

Table of monthly demand and forecasts from three cycles, June to August 2024. The customer's own forecasts appear, vanish and jump between cycles, while SupChains forecasts change in small steps.

Nicolas Vandeput

·

Mar 11

Integrating Machine Learning in Demand Planning: Where Humans Still Matter

Machine learning (ML) is transforming the role of demand planners. Moving them away from mechanical tasks focused on adjusting forecasts to…

Summary: which demand planning practices are outdated, how planners should enrich an ML baseline, and how ML models differ from statistical ones. Beside it, an old engraving of printers at work.

Nicolas Vandeput

·

Jan 26

My Learning Points from VN2, the First Inventory Competition

Taking place from 29th September to 10th of November 2025, VN2 is the first global inventory planning competition.[1] Participants had to…

Tier list of VN2 inventory techniques. A: optimization, inventory projection, mean and distribution forecasts. B: net inventory ordering, reinforcement learning. C: k × RMSE, quantiles. D: DDMRP, ABC, z·σ·√(L+R).

Nicolas Vandeput

·

Jan 7

Forecasting Demand for a Chemical Manufacturer using Confirmed Future Orders

SupChains partnered with Vantage, a US chemical manufacturer, to improve its demand planning capabilities. Together, we redefined their…

Line chart of the share of orders two customers confirm in advance. 30 days before delivery, about 63% and 42% of orders are known. At 90 days, about 10%. Near none after 180 days.

Nicolas Vandeput

Nicolas Vandeput

Book Author

Founder of SupChains. Our demand forecasting and inventory models cut forecast error by 30% 📈 and inventory by 20% 📦. Author of 3 books. supchains.com

Following

Help

Status

About

Careers

Press

Blog

Store

Privacy

Rules

Terms

Text to speech