Supply Chain Management Chapter 2

Forecasting and Demand Planning

the critical building blocks from which all supply chain planning activities are derived and are crucial components of customer satisfaction

Demand Planning

combines statistical forecasting techniques and judgment to construct demand estimates for products or services

Demand

the need for a particular product or component. Has two types (independent and dependent)

Independent Demand

demand for an item unrelated to the demand for other items, such as a finished product, a spare part, or a service part. [Demand for these items is Forecasted]

Dependent Demand

demand for an item directly related to other items or finished products, such as a component or material used in making a finished product. [Demand for these items is Calculated]

Forecasting

the business function that estimates future demand for products so they can be purchased or manufactured in appropriate quantities in advance of need

Short-term forecasting

< 3 months

Medium-term forecasting

3 months - 2 years

Long-term forecasting

2+ years

Forecasts will be

inaccurate but still useful

The forecast is the

basis for most downstream supply chain planning decisions

The 2 basic forecasting techniques

Qualitative and Quantitative forecasting

Qualitative Forecasting

Based on opinion and intuition. Used when data is limited or non-existent.

Quantitative Forecasting

Based on mathematical models and historical data

The 5 kinds of qualitative forecasting

Personal insight, jury of executive opinion, Delphi method, historical analogy, customer survey

Personal Insight

Qualitative. The forecast is based on the insight of the most experienced, knowledgeable, or senior person. Cheap but unreliable.

Jury of Executive Opinion

Qualitative. People who know the most about the product and the marketplace would likely form a jury (i.e., management panel) to discuss and determine the forecast. Cheap but may include bias.

Delphi Method

Qualitative. The same as the Jury of Executive Opinion except that each participant’s input is collected separately so that people are not influenced by one another. Less biased but more time-consuming.

Historical Analogy

Qualitative. A judgmental forecasting technique based on identifying a sales history comparable to a present situation, such as the sales history of a similar product. Can provide a significant amount of valuable data but there could also be no data to look at.

Customer Survey

Qualitative. Customers are directly approached and asked for their opinions about a particular product. Simple but time-consuming and costly.

The 2 main kinds of quantitative forecasting

Time-based and cause-and-effect

Time-series

Quantitative. Based on the assumption that the future is an extension of the past. Historical data is used to predict future demand. Most used of any forecasting technique.

Cause-and-effect

Quantitative. Assumes that one or more factors (independent variables) predict future demand .(e.g., seasonality in retail markets).

Variations in Quantitative Forecasting

Trend variations, random variations, seasonal variations, and cyclical variations

Trend Variations

A movement of a variable over time.

Random Variations

Instability in data caused by random occurrences. Short-term and often caused by unexpected accidents (weather emergencies, natural disasters, etc.). Ex: Wood demand to help rebuild after a hurricane. Considered “abnormal demand” and may be removed from data sets used for forecasts.

Seasonal Variations

Repeating patterns of demand. Can be repeated yearly, with some periods of considerably higher demand than others. Ex: Holiday shopping, swimsuits in the summer, etc.

Cyclical Variations

Wavelike patterns that last longer than one year and can extend over multiple years. Not easily predicted. Ex: Business cycles, Bull markets, etc.

Types of Time-Series Forecasting

Naive, Moving Average, Weighted Moving Average, Exponential Smoothing, Linear Trend

Naive Forecasting

Sets the demand for the next period to be precisely the same as the demand in the last period. Works well with mature products with stable demand but any variations in demand can cause inventory issues.

Simple Moving Average

Uses a calculated average of historical demand during a specified number of the most recent periods to generate the forecast. Best for short-term forecasting with stable demand and no seasonal variation. Bad with trends and seasonal spikes.
Formula: (M1 + M2 + M3 + M4) /4

Weighted Moving Average

Similar to a simple moving average except that not all periods are valued/weighted equally. More accurate than simple moving average but still lags behind actual demand.
Formula: (M1 x W1) + (M2 x W2) + (M3 x W3) + (M4 x W4)

Exponential Smoothing

Requires three elements: the last period’s actual demand, the previous period’s forecast, and a smoothing factor (a number greater than 0 and less than 1. Used as a weighting percentage). More responsive to trends, but can still lag behind trends, especially upward trends).

Formula: (A1 x S) + (F1 x (1 - S))
A = Actual, F = Forecasted, S = Smoothing Factor

Linear Trend Forecasting

Imposing a best-fit line across the demand data of an entire time series. Used as the basis for forecasting future values by extending the line past the existing data and out into the future while maintaining the slope of the line. Best suited for long-term forecasts.

The 2 types of cause-and-effect forecasting

Simple linear regression and multiple linear regression. Regression uses the historical relationship between an independent and a dependent variable to predict the future values of the dependent variable, i.e., demand

Simple Linear Regression

Attempts to model the relationship between a single independent variable and a dependent variable (demand) by fitting a linear equation to the observed data. Describes the relationship between the independent and dependent variables as a straight line. Ex: the demand might depend on how much money is spent on advertising and promotion: the more money spent, the higher the demand.

Multiple Linear Regression

Attempts to model the relationship between two or more independent variables and a dependent variable (demand) by fitting a linear equation to the observed data. Ex: the demand might be dependent on how much money is spent on advertising and promotion and also on the selling price charged for the product.

The Fundamentals of Forecasting

Your forecast is most likely wrong, but that does not mean it is not useful.

The more specific the forecast, the less accurate. Forecasts for a year is more likely to be right than a forecast for a week. A national forecast vs a regional one is similar.

Simple forecast methodologies trump complex ones. They are easier to understand and analyze.

A correct forecast does not prove that it was the correct forecast method.

Old or not regularly used data should be trusted less when forecasting.

All trends will eventually end.

It is hard to eliminate bias, so most forecasts are biased.

Technology is not the solution to better forecasting.

Social media

Analysis of social sentiment can be used to:

 Evaluate the health of a brand - An understanding of how your target market feels about your company, products, and services.

 Improve demand prediction - Companies can use the Voice of the Customer (VOC) to drive improvements in forecasting and inventory positioning.

 Address a crisis - Social sentiment analysis might reveal a spike in negative posts and provide an early warning of a potential product or service issue.

 Research the competition - Social sentiment analysis can help you understand how to position against the competition.

Forecast Errors

The difference between the actual demand and the forecast demand. The error can be quantified as an absolute value or as a percentage.

Forecast Error Value: Actual - Forecasted

Forecast Error %: ((Actual - Forecasted) / Actual) x 100

Mean Absolute Deviation (MAD)

Measures the size of the forecast error in units. It is calculated as the average of the unsigned, i.e., absolute, errors over a specified period. Whether the forecast is over or under the actual demand is irrelevant; only the magnitude of the deviation matters in the MAD calculation.
MAD = ∑ ( | Actual - Forecasted | ) / n (time)
Find each period's actual - forecasted absolute value, add them, then divide by total time.

Mean Absolute Percent Error (MAPE)

Measures the size of the error in percentage terms. It is calculated as the average of the unsigned percentage error.
MAPE = ∑ (( | Actual - Forecasted |) / Actual ) / n (time)
Find each period's actual - forecasted absolute value and divide each by their actual. Then, add them all up and divide by total time.

The Bullwhip Effect

Individual supply chain participants second-guess what is happening with ordering patterns and potentially over-react, causing a ripple that effects the entire chain. One adjusts orders, the entire chain adjusts.

Collaborative Planning, Forecasting, and Replenishment (CPFR)

A business practice that combines the intelligence of multiple trading partners who share their plans, forecasts, and delivery schedules to ensure a smooth flow of goods and services across a supply chain. Helps prevent the Bullwhip effect.

Reducing the Bullwhip Effect

means the reduction of safety stocks (and associated costs) within and across the trading partners in a supply chain.

Supply Chain Agility

an organization's ability to quickly and efficiently respond to changes in demand or supply without sacrificing quality or cost.