Forecasting in Finance/Accounting
Used as the basis for budgeting and cost control.
Forecasting in Marketing
Used in personnel compensation and new product planning.
Forecasting in Production
Uses forecasts to select suppliers, determine capacity requirements, and make purchasing/staffing/inventory decisions.
Strategic Forecasts
Used for decisions about overall directions
Tactical Forecasts
Used to guide day-to-day decisions
All forecasts have errors
UNDERSTANDING and MINIMIZING ERROR is the key to effective forecasting process.
Types of Forecasting
1) Qualitative
2) Time Series Analysis (primary focus)
- Based on idea that data relating to past demand can be used to predict future demand.
3) Causal Relationships
4) Simulation
Components of Demand
1) Average demand for a period of time
2) Trend
3) Seasonal Element
4) Cyclical Elements
5) Random Variation
6) Autocorrelation
Choosing appropriate Forecasting Model depends on:
1) Time horizon to be forecast
2) Data availability
3) Accuracy required
4) Size of forecasting budget
5) Availability of qualified personnel
Simple Moving Average
Useful when demand is relatively consistent and no seasonality exists.
Removes some random fluctuation from the data. Longer periods provide more smoothing, shorter periods react more quickly to data.
Weighted Moving Average
Allows unequal weighting for prior time periods.
Exponential Smoothing
Includes all past data in the forecasting calculation.
*Most used of all forecasting techniques.
Reasons for this:
1) Exponential models are surprisingly accurate.
2) Formulating an exponential model is relatively easy.
3) The user can understand how the
What is the effect of trends in Exponential Smoothing
Trends cause the exponential smoothing forecast to always lag behind the actual data. Corrected by adding trend adjustment delta.
Choosing Alpha and Delta
Relatively small values are common for both (0.1 to 0.3)
Alpha depends on how much random variation is present.
Delta depends on how steady the trend is.
Measurements of forecast error can be used to select values to minimize overall forecast error.
Forecast Error
The difference between the forecast value and what actually occurred. All forecasts contain some level of error.
Sources of Error:
Bias - A consistent mistake is made
Random
Measures of Error
MAD - Mean Absolute Deviation
MAPE - Mean Absolute Percentage Error
Tracking Signal
MAD
Ideally, zero. Higher MAD means less accurate model.
skewing
when X bar isn't in the desired range
X bar
average # of observation
Process Capability
Is my design capable of delivering to my specifications?
Reducing variability
improves process capability