Forecasting
Forecast – An estimate of the future level of
some variable.
Why Forecast?
Assess long-term capacity needs
Develop budgets, hiring plans, etc.
Plan production or order materials
Get agreement within firm and across supply
chain partners
Forecasting and TQM
• Accurate forecasting customer demand is a key to
providing good quality service
• Continuous replenishment and JIT (Just-in-Time)
complement TQM (Total-Quality-Management)
– eliminates the need for buffer inventory, which, in turn,
reduces both waste and inventory costs, a primary goal of
TQM
– smoothes process flow with no defective items
– meets expectations about on-time delivery, which is
perceived as good-quality service
11-2
Types of Forecasts
Demand
Firm-level
Market-level
Supply
Number of current producers and suppliers
Projected aggregate supply levels
Technological and political trends
Price
Cost of supplies and services
Market price for firm’s product or service
Laws of Forecasting
Forecasts are almost always wrong by some amount
(but they are still useful).
Forecasts for the near term tend to be more
accurate.
Forecasts for groups of products or services tend to
be more accurate.
Forecasts are no substitute for calculated values.
Forecasting Process
1. Identify the 2. Collect historical 3. Plot data and identify
purpose of forecast data patterns
6. Check forecast 5. Develop/compute 4. Select a forecast
accuracy with one or forecast for period of model that seems
more measures historical data appropriate for data
7.
Is accuracy of No 8b. Select new
forecast forecast model or
acceptable? adjust parameters of
existing model
Yes
9. Adjust forecast based 10. Monitor results
8a. Forecast over
on additional qualitative and measure forecast
planning horizon
information and insight accuracy
11-5
Forecasting Methods
Qualitative forecasting techniques – Forecasting
techniques based on intuition or informed opinion.
Used when situation is vague and little data exists.
E.g. forecasting of new products, sales to new market and new
technology
Involves intuition, experience
Quantitative forecasting models – Forecasting
models that use measurable, historical data to
generate forecasts.
Time series and causal models (use of
mathematical techniques)
E.g., forecasting sales of a existing or mature product and current
technology
Selecting a Forecasting Method
Figure 1
Qualitative Forecasting Methods
Market surveys
Build-up forecasts
Life-cycle analogy method
Panel consensus forecasting
Delphi method
Quantitative Forecasting Methods
Time series forecasting models – Models that
use a series of observations in chronological
order to develop forecasts.
Causal forecasting models – Models in which
forecasts are modeled as a function of
something other than time.
Demand movement
Randomness – Unpredictable movement from one
time period to the next.
Trend – Long-term movement up or down in a time
series.
Seasonality – A repeated pattern of spikes or drops in
a time series associated with certain times of the
year.
Time series with randomness
Figure 2
Idea Behind Time Series Models
Distinguish between random fluctuations and
true changes in underlying demand patterns.
Time series with
Trend and Seasonality
Figure 3
Last Period Model
Last Period Model - The simplest time series
model that uses demand for the current
period as a forecast for the next period.
Ft+1 = Dt
where Ft+1= forecast for the next period, t+1
and Dt = demand for the current period, t
Last Period Model
Table 1 Figure 4
Averaging Methods
generates a forecast for a particular time
period by averaging the observed data
values(that is the actual values of the
dependent variable) for the most recent n
time periods.
Two versions of this method are: simple
moving averages and weighted moving
averages.
Simple Moving Average Model
Moving Average Model – A time series
forecasting model that derives a forecast by
taking an average of recent demand value.
Smoothes out random fluctuations of data
Best for short term forecasts in the absence of
seasonal or cyclical variations n
D t 1 i
Ft 1 i 1
n
Moving Average Model
Period Demand
1 12 n
2 15 Dt 1 i
3 11
4 9 Ft 1 i 1
5 10 n
6 8
7 14
3-period moving average
8 12 forecast for Period 8:
= (14 + 8 + 10) / 3
= 10.67
Weighted Moving Average Model
Weighted Moving Average Model – A form of the
moving average model that allows the actual weights
applied to past observations to differ.
Weights are used to vary the effect of past data,
based on the fact that more recent data is more
important, the weights should go up and always
Weighted Moving Average Model
Period Demand
1 12
2 15
3 11
4 9
5 10
6 8
7 14
8 12 3-period weighted moving
average forecast for Period 8=
[(0.5 14) + (0.3 8) + (0.2 10)] / 1
=
11.4
Exponential Smoothing Model
Exponential Smoothing Model – A form of the
moving average model in which the forecast for the
next period is calculated as the weighted average of
the current period’s actual value and forecast.
Exponential Smoothing Model
a = .3
Period Demand Forecast
1 50 40
2 46 .3 * 50 + (1-.3) * 40 = 43
3 52 .3 * 46 + (1-.3) * 43 = 43.9
4 48 .3 * 52 + (1-.3) * 43.9 = 46.33
5 47 .3 * 48 + (1-.3) * 46.33 = 46.83
6 .3 * 47 + (1-.3) * 46.83 = 46.88
Trends
What do you think will happen to a moving
average or exponential smoothing model
when there is a trend in the data?
Same Exponential Smoothing Model as
Before:
Exponential
Actual Smoothing
Period Demand Forecast
1 11 11.00
Since the model
2 12 11.00
3 13 11.30
is based on
4 14 11.81 historical demand,
5 15 12.47 it always lags
6 16 13.23 the obvious
7 17 14.06
upward trend
8 18 14.94
9 15.86
Adjusting Exponential Smoothing for
Trend
• Add trend factor and adjust using exponential
smoothing
• Needs only two more numbers:
Tt = Trend factor for the current period t
= Weight between 0 and 1
• Then: Tt+1 = × (Ft+1 – Ft) + (1 – ) × Tt
• And the Ft+1 adjusted for trend is = Ft+1 + Tt+1
Measuring Forecast Accuracy
How do we know:
If a forecast model is “best”?
If a forecast model is still working?
What types of errors a particular
forecasting model is prone to make?
Need measures of forecast accuracy
Measures of Forecast Accuracy
Forecast error (e)
Mean forecast error (MFE) =
Mean absolute deviation (MAD) =
Measures of Forecast Accuracy
Mean absolute percentage error (MAPE) =
Tracking Signal =
Forecast Accuracy – Example 9.7
Table 9.11
Forecast Accuracy – Example 9.7
Calculate the forecast error for each week,
the absolute deviation of the forecast error,
and absolute percent errors.
Forecast Accuracy – Example 9.7
Forecast Accuracy – Example 9.7
Model 2 has the lowest MFE so it is the least
biased.
Model 2 also has the lowest MAD and MAPE
values so it appears to be superior.
Calculate the tracking signal for the first 10
weeks.
Forecast Accuracy – Example 9.7
Forecast Accuracy – Example 9.7
The tracking signal for Model 2 gets very low
in week 5, however the model recovers.
You need to continue to update the tracking
signal in the future.