Supply Chain Management: Demand Planning: Forecasting and Demand Management (CH 12)

Qualitative Forecasting

-subjective (based on people's opinions) -can incorporate expertise that can be hard to codify -Opinions can dominate and/or bias the forecast -Delphi Method -Market Survey -Grass Roots

5 steps to help managers forecast

1. Identify the users and decision-making processes that the forecast will support 2. Identify likely sources of the best data inputs 3. Select forecasting techniques that will most effectively transform data into timely, accurate forecasts over the appropriate planning horizon 4. Document and apply the proposed technique to the data gathered for the appropriate business process 5. Monitor the performance of the forecasting process for continuous improvement

Mean Absolute Deviation (MAD)

The average size of forecast errors, irrespective of their directions (Mean Absolute Error)

Demand Planning

The combined process of forecasting and managing customer demands to create a planned pattern of demand that meets the firm's operational and financial goals

Autocorrelation

The correlation of current demand values with past demand values

Forecast Error

The difference between a forecast and the actual demand

Primary Goal in designing a forecasting process

generate forecasts that are usable, timely, and accurate

Exponential Smoothing with Trend Effects

more responsive to changes than exponential smoothing trend component makes these reactive models proactive FORMULA ON PAGE 406

Components of Demand (Quantitative)

-Average demand for the period -Trend -Seasonality -Cycles -Autocorrelation -Random Variation

Broad Categories of Forecasting Techniques

-Judgment based techniques (Qualitative) -statistical model based techniques (Quantitative) -techniques for assessing forecast error and providing feedback to the forecasting system

Causal Relationship Techniques

-Linear Regression

Time Series Techniques

-Moving Average -Weighted Moving Average -Exponential Smoothing

Quantitative Forecasting

-Objective (based on numeric data and equations) -Consistency -Use large amounts of data -Must have data -Time Series -Casual Relationship

Keep in mind while designing a forecasting process

-Time Horizon -Level of Detail -Accuracy versus Cost -Fit with existing business processes

Mean Absolute Percentage Error (MAPE)

The MAD represented as a percentage of demand

Medium Range Forecast

3 months to 3 years Sales and personal planning, budgeting

Long Range Forecast

3+ years new product planning, facility allocation

Demand Forecasting

A decision process in which managers predict demand patterns short term is more accurate than long term Aggregated forecasts are more accurate than individual forecasts

Weighted Moving Average

A forecasting model that assigns a different to each period's demand according to its importance -permits unequal weights on prior demand -when a detectable trend or pattern is present, weights can be used to place more emphasis on recent values -the assignment of weights rely on intuition -Formula on Slide 20, Page 7

Moving Average

A forecasting model that computes a forecast as the average of demands over a number of immediate past period -useful when demand is relatively stable without trend or seasonality -useful when demand is stable -requires much historical data -increasing the number periods (n) reduces the impact of random or atypical demands in isolated time periods but it also reduces the sensitivity to actual shifts in demand MA= (Sum of demands in n periods)/n

Marketing Research

A forecasting technique that bases forecasts on the purchasing patterns and attitudes of current or potential customers customer surveys, interviews, and focus groups

Historical Analogy

A forecasting technique that uses data and experience from similar products to forecast the demand for a new product

Mean Squared Error (MSE)

A more sensitive measure of forecast errors that approximated the error variance

Exponential Smoothing

A moving average approach that applies exponentially decreasing weights to each demand that occurred farther back in time -most recent data are weighted most -Requires smoothing constant (alpha) -alpha ranges from 0 to 1 which is subjectively chosen to reflect responsiveness FORMULA ON PAGE 405

Shift or Step Change

A one-time change in demand, usually due to some external influence on demand

Smoothing Coefficient

A parameter indicating the weight given to the most recent demand alpha ranges from 0 to 1 if you ave a higher alpha it rates the most recent time period more heavily

Demand Management

A proactive approach in which managers attempt to influence the pattern of demand

Naive Model

A simple forecasting approach that assumes that recent history is a good predictor of the near future sometimes effective ignores trend, seasonal, or other components so it can create highly erratic forecasts

Grassroots Forecasting

A techniques that seeks inputs from people who are in close contact with customers and products -experts may unconsciously bias their forecasts on most recent results instead of overall experiences -each salesperson projects their sales and estimates from individual salespersons are reviewed for reasonableness -tend to be overly optimistic -good for short-term

Mean Percent Error (MPE)

Average error represented as a percentage of demand

Time Series Analysis Models

Forecasting models that compute forecasts using historical data arranged in the order of occurrence -chronologically ordered data that may contain one or more components of demand -set of evenly spaced numerical data -forecasts based only on past values -assumes that factor influencing past and present will continue to influence in future -a stationary belief

Executive Forecasting

Forecasting techniques that use input from high-level experienced managers good for long-term

Delphi Method

Forecasts developed by asking a panel of experts to individually and repeatedly respond to a series of questions -Iterative group process -reduces "group think" by concealing identities -usually achieves satisfactory results in three iterations -repeated until a consensus is achieved

Root Mean Squared Error (RMSE)

Gives an approximation of the forecast error standard deviation

LOOK AT SLIDE 13

ITS ON PAGE 5

Trend

The general sloping tendency of demand, either upward or downward, in a linear or non linear fashion

Forecast Accuracy

The measure of how closely the forecast aligns with the observations over time -MSE (mean squared error) -MAD (mean absolute deviation) -MAPE (mean absolute percent error) -RSFE (running sum of forecast errors) -TS (tracking signal)

Tracking Signal

The ratio of a running total of forecast error to MAD that indicates when the pattern of forecast error is changing significantly -a measure that indicates whether the forecast average is keeping pace with any genuine upward or downward changes in demand -TS=RSFE/MAD -typical control limits are set at +-3 standard deviations 1 standard deviation = 1.25 MAD 3 standard deviation = 3.75 MAD

Causal Models

Where time series models use only past demand values as indicators of future demand, causal models use other independent, observed data to predict demand

Short Range Forecast

up to one year, usually less than three months job scheduling, work assignments