Womack Report

Quiz next week, which will be open from 6:00 PM to 10:00 PM Sunday. Very weird system. During class next week we’ll have a test, which will be conducted online during the class time. Turned in our first case this Sunday.

Chapter 5 — Forecasting

Forecasting is useful for making informed decisions about things which will take place in the future.

There are eight steps in forecasting

1. Determine the purpose of the forecast
2. Select the items or quantities to be forecasted
3. Determine the time horizon of the forecast
4. Select the forecasting model or models
5. Gather the data needed to make the forecast
6. Validate the forecasting model
7. Make the forecast
8. Implement the results

There are three major families of forecasts

1. Qualitative Models – attempt to include subjective factors
1. Delphi Methods – Independently question many knowledgeable people in the relevant field and have mediator form consensus
2. Jury of Executive Opinion – Executives put their heads together and make a forecast collectively
3. Sales Force Composite – Have each seller estimate his or her upcoming sales, and add all the results together.
4. Customer Market Survey – Survey customer opinions
2. Time-Series Models – include historical data over a time interval
1. Moving Average
2. Exponential Smoothing
3. Trend Projections
4. Decomposition
3. Causal Methods – Include a variety of factors
1. Regression Analysis
2. Multiple Regression

Time Series can be decomposed into:

• Trend – Gradual up or down movement over time
• Seasonality – Pattern of fluctuations above or below trend that occur every year
• Cycles – Patterns in data that occur over the course of multiple years
• Random Variations – Variations in data caused by chance, unpredictable, or unusual situations.

Moving Average methods consist of computing an average of the most recent n data values for the time series and using this average for the forecast of the next period.

A simple moving average simply averages the data from the last n periods.  This will often be inaccurate, particularly when there is a trend upward or downward.

The weighted moving average works the same, but applies greater value to some data than other data.  Typically the most recent data is assigned greater weight.

Exponential Smoothing is a type of moving average technique that involves little record keeping for previous data.  The new forecast is equal to the previous forecast + alpha * the error in the previous forecast.  Alpha is a measure of confidence in the past data.  Also called the smoothing constant.  The higher the alpha, the more importance you are placing in the current data.  This is because the most current data is reflected in the error of the previous forecast, not the current forecast.