A forecast is an estimate of future demand & provides the basis for planning decisions
- The goal is to minimize deviation between actual demand and forecast
- The factors that influence demand must be considered when forecasting.
- Good forecasting provides reduced inventories, costs, & stockouts, & improved production plans & customer service
- Qualitative forecasting is based on opinion & intuition.
- Quantitative forecasting uses mathematical models & historical data to make forecasts.
- Time series models are the most frequently used among all the forecasting models.
- Qualitative Forecasting Methods
Generally used when data are limited, unavailable, or not currently relevant. Forecast depends on skill & experience of forecaster(s) & available information
Four qualitative models used are –
-
- Jury of executive opinion
- Delphi method
- Sales force composite
- Consumer survey
- Jury of executive opinion
Group of senior management executives who are knowledgeable about their markets, competitors, and the business environment collectively develop the forecast
- Delphi method
Group of internal and external experts are surveyed during several rounds in terms of future events and long-term forecasts of demand, to develop a forecast
- Sales force composite
Forecast is based on the sales force’s knowledge of the market and estimates of customer needs.
- Consumer survey
Forecasts are developed from the results surveying customers on future purchasing needs, new product ideas and opinions about existing or new products
2- Quantitative Methods
- Time series forecasting – based on the assumption that the future is an extension of the past. Historical data is used to predict future demand
- Cause & Effect forecasting – assumes that one or more factors (independent variables) predict future demand
It is generally recommended to use a combination of quantitative & qualitative techniques
- Components of Time Series
Data should be plotted to detect for the following components –
-
-
- Trend variations: increasing or decreasing over many years
- Cyclical variations: wavelike movements that are longer than a year (e.g., business cycle)
- Seasonal variations: show peaks & valleys that repeat over a consistent interval such as hours, days, weeks, months, seasons, or years
- Random variations: due to unexpected or unpredictable events
-
Time Series Forecasting Models
- Weighted Moving Average Forecast – is based on an n-period weighted moving average
- Simple Moving Average Forecast – uses historical data to generate a forecast. Works well when demand is stable over time.
- Exponential Smoothing Forecast – a type of weighted moving average where only two data points are needed
Ft+1 = Ft+a(At – Ft) or Ft+1 = aAt + (1 – a) Ft
Where Ft+1 = forecast for Period t + 1
Ft = forecast for Period t
At = actual demand for Period t
a = smoothing constant (0 ≤ a ≤1)
- Cause & Effect Models
One or several external variables are identified that are related to demand
- a) Simple regression – Only one explanatory variable is used & is similar to the previous trend model. The difference is that the x variable is no longer time but an explanatory variable.
Ŷ = b0 + b1x
Where Ŷ = forecast or dependent variable
x = explanatory or independent variable
b0 = intercept of the line
b1 = slope of the line
- b) Multiple regression several explanatory variables are used to make the forecast
Ŷ = b0 + b1x1 + b2x2 + . . . Bkxk
Where
Ŷ = forecast or dependent variable
xk = kth explanatory or independent variable
b0 = intercept of the line
bk = regression coefficient of the independent variable xk
Several measures of forecasting accuracy follow –
-
- Mean absolute deviation (MAD)- a MAD of 0 indicates the forecast exactly predicted demand
- Mean absolute percentage error (MAPE)- provides a perspective of the true magnitude of the forecast error
- Mean squared error (MSE)- analogous to variance, large forecast errors are heavily penalized
Mean absolute deviation (MAD)-
- MAD of 0 indicates the forecast exactly predicted demand.
Where et = forecast error for period t
At = actual demand for period t
n = number of periods of evaluation
Mean absolute percentage error (MAPE) – provides a perspective of the true magnitude of the forecast error.
Where et = forecast error for period t
At = actual demand for period t
n = number of periods of evaluation
Mean squared error (MSE) – analogous to variance, large forecast errors are heavily penalized
Where et = forecast error for period t
n = number of periods of evaluation