Forecasting Techniques for Slow and Rapidly Changing Demand

Supply shortages, pandemics, military wars, trade wars, and other disruptive events have a significant impact in both consumer behaviour and product availability. Companies are becoming aware that historical sales data sets might no longer be relevant; and that the customary forecasting methods are not the best for their new current situation.

This is the reason why, demand for skilled, critical and flexible Demand Planners with broad perspective is on the rise.

In this course you will be able to decide if previously used forecasting techniques are the right ones for today's "New Normal" business environment. You will be capable of forecasting customer demand of different offerings going through different stages in their product life cycles; using causal and judgemental techniques, market research, statistical methods, time series of past sales and most recent customer orders.

You will also be able to separate relevant from non-relevant data, and mitigate the impact of low forecast accuracy in demand planning, inventory management and profitability.

By the end of this course, that is part of the edX Professional Certificate program to become a Certified Forecaster and Demand Planner (CFDP), you will be able to choose the right forecasting method for each data pattern and understand how to improve forecast performance with Machine Learning and Lean Six Sigma principles.

CFDP certified professionals are globally preferred by recruiters for decision making positions because they are capable of forecasting both slow and rapidly changing seasonal, intermittent and new product demand.

To become an ISCEA Certified Forecaster and Demand Planner (CFDP), you must complete all three preparatory courses and successfully pass the CFDP Exam.

Section 2.1. Forecasting myths, realities and challenges

2.1.1. Forecasting myth 1 - About predicting the future

2.1.2. Forecasting myth 2 - About choosing a forecasting model

2.1.3. Forecasting myth 3 - About model fitting

2.1.4. Forecasting myth 4 - About model sophistication

2.1.5. Forecasting myth 5 - About About Artificial Intelligence

2.1.6. Forecasting myth 6 - About modeling data

Section 2.2. Forecasting based on historical data

2.2.1. The role of historical data in forecasting

2.2.2. Missing data, events and outliers

2.2.3. Naïve and Moving Average Methods

2.2.4. Exponential smoothing and linear regression

2.2.5. ARIMA models

Section 2.3. Seasonal demand forecasting

2.3.1. Decomposition method

2.3.2. Decomposition method example

2.3.3. Holt-Winters method

2.3.4. Holt-Winters method example

Section 2.4. Intermittent demand forecasting

2.4.1. Croston method

Section 2.5. Judgmental and causal forecasting models

2.5.1. Judgmental forecasting models

2.5.2. Causal forecasting models

Section 2.6. Forecasting with Machine Learning

2.6.1. Machine learning fundamentals

2.6.2. Decision trees in machine learning

2.6.3. Machine learning example

Section 2.7. New product forecasting

2.7.1. New product forecasting fundamentals

2.7.2. Assumptions-based modeling and scenario analysis

2.7.3. Quantitative analysis to forecast new product demand

Section 2.8. Forecast performance

2.8.1. Forecast error

2.8.2. Impact of aggregation on forecast accuracy

2.8.3. Forecast accuracy and forecastability

Section 2.9. Impact of randomness and disruptive events in forecasts

2.9.1. Understanding randomness and disruptive events

2.9.2. Mitigating risk

Section 2.10. Improving forecasting with Lean and Six Sigma Principles

2.10.1. Forecast Value-Added Analysis

2.10.2. Applying Six Sigma methodologies to forecasting