Demand and Forecast Planning with Machine Learning and AI

Demand and Forecast Planning with Machine Learning and AI

Day 1: Foundations of Demand Forecasting

Module 1: Foundational Concepts of Demand Forecasting

Why do we forecast demand?

The five steps to demand planning excellence

Setting forecasting objectives

Required data and data quality

Evaluation metrics

Creating a baseline model

Forecast review processes

Module 2: Understanding Demand

Demand vs. sales vs. supply

Sales targets, supply plans, financial budgets

Capturing true (unconstrained) demand

Order management and shortage censoring

Dealing with substitution and cannibalization

Module 3: Collaboration and Hierarchies in Forecasting

How supply chains distort demand: bullwhip effect

Internal and external collaboration models

Forecasting hierarchies and aggregation levels

Granularity vs. usefulness of forecasts

Zooming in/out on different dimensions

Module 4: Forecasting Horizons and Supply Chain Alignment

Inventory optimization, lead times, and planning frequency

Lost sales vs. backorders vs. hybrid approaches

Balancing short-term accuracy with long-term visibility

One-number forecast vs. tailored forecast models

Hierarchy reconciliation and alignment challenges

Day 2: Forecasting Performance and Evaluation

Module 5: Forecasting Metrics and Interpretation

Accuracy vs. bias

Forecast error types: MAE, MAPE, RMSE

Practical exercises with metrics

Interpreting forecast performance across different situations

Module 6: Choosing and Applying the Right KPI

Dealing with extreme and intermittent demand

Case study: selecting KPIs for different business contexts

Trade-offs in metric selection

When and how to use value-weighted KPIs

Module 7: Understanding Forecast Error

Benchmarking with naïve and moving average forecasts

Seasonal benchmarks

Limitations of the Coefficient of Variation (COV)

Module 8: Forecast Accuracy Across Portfolios

Forecasting metrics for large product portfolios

How to apply segmentation to evaluate overall performance

Value-at-risk perspective in demand planning

Day 3: Forecasting with Machine Learning and AI

Module 9: Forecast Value Added and Process Efficiency

What is Forecast Value Added (FVA)?

Internal vs. external benchmarks

Evaluating process efficiency and efficacy

Best practices to optimize FVA

Module 10: Product Segmentation for Forecasting

ABC/XYZ segmentation methodology

Applications in forecasting models and review strategies

Going beyond: multi-criteria smart classification

Module 11: Forecasting Methods Overview

Time series models: level, trend, seasonality

Setting up and tuning statistical models

Predictive analytics using demand drivers

Comparing time series vs. driver-based approaches

Module 12: Machine Learning in Demand Forecasting

What is ML and how does it learn?

Overview of algorithms: tree-based, neural networks

Expectations vs. reality: where ML works best

Launching ML-driven forecasting projects

Module 13: Judgmental Forecasting and Human Insight

When to use expert judgment

Understanding cognitive and strategic biases

Using group forecasting effectively

Combining statistical, ML, and human inputs