CA5104: Operations

Management and Total

Quality Management

© McGraw-Hill Education
MODULE
Forecasting
2

© McGraw-Hill Education
LEARNING OUTCOMES
LO2.1 Compare and contrast the features of forecasting methods.
LO2.2 Explain the elements of a good forecast.
LO2.3 Outline the steps in the forecasting process.
LO2.4 Summarize forecast errors and use summaries to make decisions.
LO2.5 Explain qualitative forecasting techniques.
LO2.6 Execute different quantitative methods to make a forecast.
LO2.7 Construct control charts and use them to monitor forecast errors.

© McGraw-Hill Education
FORECAST

Forecast – a statement about the

future value of a variable of interest
• We make forecasts about such things
as weather, demand, and resource
availability
• Forecasts are important to making
informed decisions

© McGraw-Hill Education
TWO IMPORTANT ASPECTS OF FORECASTS

• Expected level of demand

• The level of demand may be a
function of some structural
variation such as trend or
seasonal variation
• Accuracy
• Related to the potential size of
forecast error

© McGraw-Hill Education
USES OF FORECASTS
ACCOUNTING • Cost/profit estimates

FINANCE • Cash flow and funding

HUMAN RESOURCES • Hiring/recruiting/training

MARKETING • Pricing, promotion, strategy

MIS • IT/IS systems, services

OPERATIONS • Schedules, MRP, workloads

PRODUCT/SERVICE • New products and services
DESIGN

© McGraw-Hill Education
YIELD MANAGEMENT

Yield Management – relates to the

percentage of capacity being used.
Accurate forecasts can help managers
plan tactics to match capacity with
demand, thereby achieving high yield
levels

© McGraw-Hill Education
LONG-RANGE vs SHORT-RANGE

Long-range - needed for strategic

changes such as developing new products,
new market

Short-range - basis to predict

requirements for labor and other resources
needed to respond to changes in demand

© McGraw-Hill Education
USES FOR FORECAST

• Plan the system

• Generally involves long-range plans
related to:
• Types of products and services to offer
• Facility and equipment levels
• Facility location

© McGraw-Hill Education
USES FOR FORECAST
• Plan the use of the system
• Generally involves short- and
medium-range plans related to:
• Inventory management
• Workforce levels
• Purchasing
• Production
• Budgeting
• Scheduling

© McGraw-Hill Education
FEATURES COMMON TO ALL FORECASTS
1. Techniques assume some underlying
causal system that existed in the past
will persist into the future
2. Forecasts are not perfect
3. Forecasts for groups of items are more
accurate than those for individual
items
4. Forecast accuracy decreases as the
forecasting horizon increases

© McGraw-Hill Education
ELEMENTS OF A GOOD FORECAST

Techniqu
Should
es should
be
Should Should Should be simple Should
Should expresse
be be be in to be cost-
be timely d in
accurate reliable writing understa effective
meaningf
nd and
ul units
use

© McGraw-Hill Education
STEPS IN THE FORECASTING
PROCESS
1. Determine the purpose of the forecast
2. Establish a time horizon
3. Obtain, clean, and analyze appropriate data
4. Select a forecasting technique
5. Make the forecast
6. Monitor the forecast errors

© McGraw-Hill Education
APPROACHES TO FORECASTING

• Qualitative forecasting
• Qualitative techniques permit the inclusion
of soft information such as:
• Human factors
• Personal opinions
• Hunches
• These factors are difficult, or impossible, to
quantify

© McGraw-Hill Education
APPROACHES TO FORECASTING

• Quantitative forecasting
• These techniques rely on hard data
• Quantitative techniques involve either the
projection of historical data or the
development of associative methods that
attempt to use causal variables to make a
forecast

© McGraw-Hill Education
QUALITATIVE FORECASTS: JUDGMENTAL

• Forecasts that use subjective inputs such

as opinions from consumer surveys, sales
staff, managers, executives, and experts
• Executive opinions
• A small group of upper-level managers may meet
and collectively develop a forecast
• Sales force opinions
• Members of the sales or customer service staff can
be good sources of information due to their direct
contact with customers and may be aware of plans
customers may be considering for the future

© McGraw-Hill Education
QUALITATIVE FORECASTS: JUDGMENTAL

• Consumer surveys
• Since consumers ultimately determine demand,
it makes sense to solicit input from them
• Consumer surveys typically represent a sample
of consumer opinions
• Other approaches
• Managers may solicit opinions from other
managers or staff people or outside experts to
help with developing a forecast.
• The Delphi method is an iterative process
intended to achieve a consensus

© McGraw-Hill Education
TIME-SERIES FORECASTS

• Forecasts that project patterns

identified in recent time-series
observations
• Time-series – a time-ordered sequence of
observations taken at regular time intervals
• Assume that future values of the time-
series can be estimated from past
values of the time-series

© McGraw-Hill Education
ASSOCIATIVE MODELS

Use equations that consist of one or

more explanatory variables that can be
used to predict demand

© McGraw-Hill Education
FORECASTS BASED ON TIME-SERIES DATA

• Trend
• Seasonality
• Cycles
• Irregular variations
• Random variation

© McGraw-Hill Education
TRENDS AND SEASONALITY
Trend
• A long-term upward or downward movement in data
• Population shifts
• Changing income

Seasonality
• Short-term, fairly regular variations related to the calendar or
time of day
• Restaurants, service call centers, and theaters all experience
seasonal demand
© McGraw-Hill Education
Trends and Seasonality

• Trend
• A long-term upward or downward movement in data
 Population shifts
 Changing income

• Seasonality
• Short-term, fairly regular variations related to the
calendar or time of day
• Restaurants, service call centers, and theaters all
experience seasonal demand

© McGraw-Hill Education. 3-22

CYCLES AND VARIATIONS
Cycle
• Wavelike variations lasting more than one
year
• These are often related to a variety of economic,
political, or even agricultural conditions
Irregular variation
• Due to unusual circumstances that do not
reflect typical behavior
• Labor strike
• Weather event

© McGraw-Hill Education
CYCLES AND VARIATIONS
Random Variation
• Residual variation that remains after all other behaviors
have been accounted for

© McGraw-Hill Education
FORECASTING - NAÏVE FORECAST
• Naïve forecast
• Uses a single previous value of a time series
as the basis for a forecast
• The forecast for a time period is equal to the
previous time period’s value

© McGraw-Hill Education
FORECASTING - NAÏVE FORECAST
• Stable time series data
• F(t) = A(t-1)
• Future = past
• Seasonal variations
• F(t) = A(t-n)
• Future = past season
• Data with trends
• F(t) = A(t-1) + (A(t-1) – A(t-2))
• Future = past + difference from 2 periods ago

© McGraw-Hill Education
FORECASTING - NAÏVE FORECAST
• Simple to use
• Virtually no cost
• Quick and easy to prepare
• Data analysis is nonexistent
• Easily understandable
• Cannot provide high accuracy
• Can be a standard for accuracy

© McGraw-Hill Education
TIME-SERIES FORECASTING - AVERAGING
• These techniques work best when a
series tends to vary about an average
• Averaging techniques smooth variations
in the data
• They can handle step changes or
gradual changes in the level of a series
• Techniques
1. Moving average
2. Weighted moving average
3. Exponential smoothing

© McGraw-Hill Education
MOVING AVERAGE
Technique that averages a number of the most recent actual
values in generating a forecast
n

A t i
At  n  ...  At  2  At  1
Ft MA n  i 1 
n n
where
Ft Forecast for time period t
MA n n period moving average
At  i Actual value in period t  i
n Number of periods in the moving average

© McGraw-Hill Education
MOVING AVERAGE
PERIOD DEMAND
1 42
2 40
3 43
4 40
5 41

© McGraw-Hill Education
WEIGHTED MOVING AVERAGE
• The most recent values in a time series are given more
weight in computing a forecast
• The choice of weights, w, is somewhat arbitrary and involves some
trial and error
Ft wt ( At )  wt  1 ( At  1 )  ...  wt  n ( At  n )
where
wt weight for period t , wt  1 weight for period t  1, etc.
At the actual value for period t , At  1 the actual value for period t  1, etc.

© McGraw-Hill Education
Moving Average
• As new data become available, the forecast is
updated by adding the newest value and
dropping the oldest and then re-computing
the average
• The number of data points included in the
average determines the model’s sensitivity
• Fewer data points used—more responsive
• More data points used—less responsive

© McGraw-Hill Education. 3-32

 Weighted Moving Average

• The most recent values in a time series are given more

weight in computing a forecast
• The choice of weights, w, is somewhat arbitrary and involves some
trial and error
Ft wt ( At )  wt  1 ( At  1 )  ...  wt  n ( At  n )
where
wt weight for period t , wt  1 weight for period t  1, etc.
At the actual value for period t , At  1 the actual value for period t  1, etc.

© McGraw-Hill Education
WEIGHTED MOVING AVERAGE
Compute a weighted average forecast using a weight of .40
for the most recent period, .30 for the next most recent, .20
for the next, and .10 for the next.
PERIOD DEMAND
1 42
2 40
3 43
4 40
5 41
© McGraw-Hill Education
MOVING AVERAGE
• As new data become available, the forecast is
updated by adding the newest value and dropping
the oldest and then re-computing the average
• The number of data points included in the average
determines the model’s sensitivity
• Fewer data points used—more responsive
• More data points used—less responsive

© McGraw-Hill Education
EXPONENTIAL SMOOTHING
Premise: The most recent observations might have the highest
predictive value.
• Therefore, we should give more weight to the more recent time
periods when forecasting.
• Weighted averaging method based on previous forecast plus a
percentage of the forecast error

Ft = Ft-1 + (At-1 - Ft-1)

© McGraw-Hill Education
EXPONENTIAL SMOOTHING
Suppose the previous forecast was 42 units, actual
demand was 40 units, and α = .10. Compute for the
new forecast.
Ft = Ft-1 + (At-1 - Ft-1)

If the actual demand turns out to be 43, compute for

the next forecast.

© McGraw-Hill Education
Techniques for Trend
• Analysis of trend involves developing an
equation that will suitably describe the
trend (assuming that trend is present in
the data). Trends may be linear or
nonlinear.

3-38
© McGraw-Hill Education
Techniques for Trend
• The linear trend equation
Ft a  bt
where
Ft Forecast for period t
a Value of Ft at t 0
b Slope of the line
t Specified number of time periods from t 0

3-39
© McGraw-Hill Education
Techniques for Trend
• The slope and the intercept can be
estimated from historical data.
n  ty 
 t y
b
n  t   t 
2
2

a
 y  b t
or y  bt
n
where
n Number of periods
y Value of the time series 3-40
© McGraw-Hill Education
TREND-ADJUSTED EXPONENTIAL
SMOOTHING
• The trend adjusted forecast
consists of two components:
• Smoothed error
• Trend factor
• Alpha and beta are smoothing
constants
• Trend-adjusted exponential
smoothing has the ability to
respond to changes in trend
© McGraw-Hill Education
Trend-Adjusted Exponential Smoothing or
double-smoothing

• Trend-adjusted
exponential
smoothing has the
ability to respond to
changes in the trend

3-42
Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or
distribution without the prior written consent of McGraw-Hill Education
Example 2:
Given the time series data in the next slide,
a. Use linear trend to forecast the demand
for periods 6 to 11.
b. Use trend-adjusted exponential
smoothing to forecast the demand for
periods 6 to 11 using α=0.40 and β=0.30.

3-43
© McGraw-Hill Education
 Answer to Example 2a:
Period, Demand, Forecasted
ty t
2
t y demand,

1 700
2 724
3 720
4 728
5 740
6 742
7 758
8 750
9 770
10 775
11
© McGraw-Hill Education
Answer to Example 2a:

© McGraw-Hill Education
Steps to solve Example 2b:
• Compute the initial estimate of the trend. Since the
problem specifies that we forecast periods 6 to 11,
• Compute the average of the net changes for periods 1 to 4.
• Compute the forecast for period 5 by taking the sum of the
actual value for period 4 and the average net change for
periods 1 to 4.
• Compute S and T for the current period using the
formula provided.
• Compute TAF for the next period using the formula
provided.
Perio Deman Chang S T Forecast
d d e
=724 - 700
1 700
Step 2
Step 1
2 724 =720 - 724

3 720 =728 - 720 =(24 + (-4) +

8)/3
4 728
5 740
6 742
7 758
8 750
9 770
10 775
11
© McGraw-Hill Education
Perio Deman Chang S T Forecast
d d e
=724 - 700
1 700
Step 2
Step 1
2 724 24 =720 - 724

3 720 -4 =728 - 720 =(24 + (-4) +

8)/3
4 728 8
5 740 9.333
6 742
7 758
8 750
9 770
10 775
11
© McGraw-Hill Education
Perio Deman S T Forecast Actual -
d d Forecast
1 700
Step 3
2 724
3 720 =728+9.333

4 728
5 740 9.333
6 742
7 758 =737.33+0.4(2.66
8 750 7) =740 – 737.33
9 770
10 775
Step 5 Step 4
11
© McGraw-Hill Education
Perio Deman S T Forecast Actual -
d d Forecast
1 700
Step 3
2 724
3 720 =728+9.333

4 728
5 740 738.400 9.333 737.33 2.667
0
6 742
=737.33+0.4(2.66
7 758 7) =740 – 737.33
8 750
9 770
10 775 Step 5 Step 4

11 © McGraw-Hill Education
Period Demand S T Forecast Actual - Forecast

1 700
Step 7
2 724
3 720 =9.333+0.3(747.728 - 737.33 -
9.333)
4 728
5 740 738.400 9.333 737.33 2.667
0 9.760 747.72 -5.728
745.440
6 742 8
7 758 =742 – 747.728

8 =747.728+0.4(-5.728)
750 =738.400 + 9.333
9 770 Step 8
10 775Step 9
Step 6
11 © McGraw-Hill Education
Period Demand S T Forecast Actual - Forecast

1 700
Step 7
2 724
3 720 =9.333+0.3(747.728 - 737.33 -
9.333)
4 728
5 740 738.400 9.333 737.33 2.667
0
6 742 745.440 9.760 747.728 -5.728
7 758 =742 – 747.728

8 =747.728+0.4(-5.728)
750
=738.400 + 9.333
9 770 Step 8
10 775Step 9
Step 6
11 © McGraw-Hill Education
Period Deman Change S T Forecast Error =
d Actual -
Forecast
1 700
2 724 24
3 720 -4
4 728 8
5 740 738.40 9.333 737.333 2.667
6 742 745.44 9.653 747.733 -5.733
7 758 756.26 8.965 755.093 2.907
8 750 759.13 9.314 765.221 -15.221
9 770 769.07 7.488 768.447 1.553
10 775 775.93 7.674 776.556 -1.556
11 783.607
© McGraw-Hill Education
Techniques for Seasonality
• Seasonal variations
• Regularly repeating movements in series
values that can be tied to recurring events
A. Regular annual variations

B. Daily, weekly or monthly

3-54
© McGraw-Hill Education
TECHNIQUES FOR SEASONALITY
Models of seasonality
• Additive - Seasonality is expressed as a quantity
that gets added to or subtracted from the time-
series average in order to incorporate seasonality
• Multiplicative - Seasonality is expressed as a
percentage of the average (or trend) amount
which is then used to multiply the value of a series
in order to incorporate seasonality

© McGraw-Hill Education
TECHNIQUES FOR SEASONALITY
• Seasonal variations
• Regularly repeating movements in series values that can
be tied to recurring events
• Seasonal relative
• Percentage of average or trend
• Centered moving average
• A moving average positioned at the center of the data that
were used to compute it

© McGraw-Hill Education
SEASONAL RELATIVES
• Using seasonal relatives
• To deseasonalize data
• Done in order to get a clearer picture of the nonseasonal (e.g.,
trend) components of the data series
• Divide each data point by its seasonal relative

• To incorporate seasonality in a forecast

• Obtain trend estimates for desired periods using a trend equation
• Add seasonality by applying these trend estimates by the
corresponding seasonal relative

© McGraw-Hill Education
Seasonal Relatives
The seasonal percentage used in the multiplicative
seasonally adjusted forecasting model; seasonal indices
Ways of Using Seasonal Relatives
1. To deseasonalize data
 Done to get a clearer picture of the non seasonal
(e.g., trend) components of the data series
 Divide each data point by its seasonal relative

© McGraw-Hill Education. 3-58

Example 3: Deseasonalizing data
A certain water station has
the following sales data for its
quarterly sales:

Quarte Sales (in If the quarterly relatives are

Year r liters) 1.10 for the first quarter, 1.20
1 1 360.80 for the second quarter, 0.75 for
1 2 629.76 the third quarter, and 0.95 for
1 3 501.84 the fourth quarter,
1 4 479.86 deseasonalize the station’s
2 1 432.96 sales for quarters 1 to 8.
2 2 672.40
2 3 629.76
2 4 570.06

© McGraw-Hill Education. 3-59

Example 3: Deseasonalizing data
Solution:
Quarte Sales (in Quarter Deseasonalized
Year r liters) Relatives Sales
1 1 360.80 1.10 328.00
1 2 629.76 1.20 524.80
1 3 501.84 0.75 669.12
1 4 479.86 0.95 505.12
2 1 432.96 1.10 393.60
2 2 672.40 1.20 560.33
2 3 629.76 0.75 839.68
2 4 570.06 0.95 600.06

© McGraw-Hill Education. 3-60

Seasonal Relatives
Ways of Using seasonal relatives
2. To incorporate seasonality in a forecast
 Obtain trend estimates for desired periods using a
trend equation
 Add seasonality by applying these trend estimates to
the corresponding seasonal relative

Incorporating seasonality in a forecast is useful when the

demand has both trend (or average) and seasonal components.

© McGraw-Hill Education. 3-61

Example 4 (Incorporating seasonality)
A certain farm ships boxes of fruits to key cities in NCR.
Using the following information, a manager wants to forecast
shipments for the first four months of the following year.

The monthly forecast being used is: Ft = 402 + 3t, where

t = 0 = January of this year and Ft = Forecast of shipments
for month t. Incorporate seasonality into the trend estimates.

© McGraw-Hill Education. 3-62

Example 4 Solution
Step 1: Determine the trend amounts for the first four months
of the following year using the given monthly forecast equation
(You may also use Excel to get this equation in case it is not
given in the problem.)

Seasonal
Month t Ft=402 + 3t
Relative
Jan. 12 438
Feb. 13 441
March 14 444
April 15 447

© McGraw-Hill Education. 3-63

Example 4 Solution
Step 1: Determine the trend amounts for the first four months
of the following year using the given monthly forecast equation
(You may also use Excel to get this equation in case it is not
given in the problem.)

Seasonal
Month t Ft
Relative
Jan. 1.2 12 402 + 3(12) = 438
Feb. 1.3 13 402 + 3(13) = 441
March 1.3 14 402 + 3(14) = 444
April 1.1 15 402 + 3(15) = 447

© McGraw-Hill Education. 3-64

Example 4 Solution
Step 2: Multiply each monthly trend by the corresponding
seasonal relative for that month.
Seasonal
Month Relative t Ft Forecast
402 + 3(12) =
Jan. 1.2 12 525.60
438
402 + 3(13) =
Feb. 1.3 13 573.30
441
402 + 3(14) =
March 1.3 14 577.20
444
402 + 3(15) =
April 1.1 15 491.70
447

© McGraw-Hill Education. 3-65

Computing Seasonal
Relatives
Seasonal Relatives can be computed using two methods:
1. The Centered Moving Average is the most widely used method
for computing seasonal relatives since it effectively accounts
for any trend (linear or curvilinear) that might be present in
the data. The use of the software is recommended since the
manual computation is a bit cumbersome.
2. The Simple Average (SA) Method is an alternative way. Each
seasonal relative is the average for that season divided by the
average for all seasons. It can be used to obtain fairly good
values of seasonal relatives as long as the ratio of the
intercept to the slope is large.

3-66
© McGraw-Hill Education
Example 5 (Computing seasonal relatives)
Given the seasons (in days) for every quarter of Years 1 to 3.
Compute the seasonal relatives for each quarter using the SA
method. Step 1: Step 3:
Quarter
Quarter Year 1 Year 2 Year 3 Quarter Average
Relative
=Mean(20, 23,
1 20 23 17
17)
2 10 12 8
3 25 19 22
4 28 26 30
Mean of Quarter
Step 2: Get the average
Averages.
© McGraw-Hill Education. 3-67
Example 5: Answer
Given the seasons (in days) for every quarter of Years 1 to 3.
Compute the seasonal relatives for each quarter using the SA
method. Step 1: Step 3:
Quarter
Quarter Year 1 Year 2 Year 3 Quarter Average
Relative
1 20 23 17 20 20/20 = 1.00
2 10 12 8 10 10/20 = 0.50
3 25 19 22 22 22/20 = 1.10
4 28 26 30 28 28/20 = 1.40
Mean 80/4 = 20
Step 2: Get the average of
Quarter Averages.
© McGraw-Hill Education. 3-68
TECHNIQUES FOR CYCLES
• Cycles are similar to seasonal variations but are of
longer duration
• Explanatory approach
• Search for another variable that relates to, and leads, the variable
of interest
• Housing starts precede demand for products and services
directly related to construction of new homes
• If a high correlation can be established with a leading variable,
an equation can be developed that describes the relationship,
enabling forecasts to be made

© McGraw-Hill Education
FORECAST ACCURACY AND
CONTROL
Forecast errors should be monitored
• Error = Actual – Forecast
• If errors fall beyond acceptable bounds,
corrective action may be necessary
• Positive error- forecast is too low
• Negative error-forecast is too high

© McGraw-Hill Education
FORECAST ACCURACY METRICS
MAD 
 Actual t  Forecast t MAD weights all errors
evenly
n

 Actual t  Forecast t 
2
MSE weights errors according to
MSE  their squared values
n 1

Actual t  Forecast t
 Actual t
100 MAPE weights errors according
MAPE  to relative error
n

© McGraw-Hill Education
FORECAST ACCURACY METRICS
FORECAST 1 FORECAST 2
QUARTERS
ACTUAL FORECAST ACTUAL FORECAST
First Quarter 10 10 10 9
Second Quarter 12 11 12 12
Third Quarter 11 9 11 9
Fourth Quarter 14 15 14 12

Actualt  Forecast t
   2

MAD 
Actualt  Forecast t
MSE 
Actual t  Forecast t  Actualt
100
n n 1 MAPE 
n

© McGraw-Hill Education
Monitoring the Forecast
• Tracking and analyzing forecast errors can provide useful
insight into whether forecasts perform satisfactorily.
• Possible Sources of forecast errors:
1. The model may be inadequate due to the omission of an
important variable, a change or shift in the variable the model
cannot handle, or the appearance of a new variable
2. Irregular variations may occur due to severe weather or
other natural phenomena, temporary shortages or
breakdowns, catastrophes, or similar events.
3. Randomness or the inherent variation that remains in the
data after all causes or variations have been accounted for.

© McGraw-Hill Education. 3-73

Monitoring the Forecast
• A forecast is generally deemed to perform
adequately when the errors exhibit only
random variations.
• If the forecast errors are not random, it is
necessary to investigate to determine which
of the other sources is present and how to
correct the problem.

© McGraw-Hill Education. 3-74

Tools for Monitoring the Forecast
1. Control Charts – a very useful tool for detecting
non-randomness in errors.
2. Tracking Signal – Relates the cumulative
forecast to the average absolute error (MAE) to
detect any bias in errors over time.
Bias – the persistent tendency for forecasts to be
greater or less than the actual values of a time
series.
The control chart is generally superior to the
tracking signal approach since the latter uses
cumulative errors. © McGraw-Hill Education. 3-75
Control Charts

Forecast errors are “in control” or random if:

1. All the errors are within the control limits.
2. No patterns (trends, cycles, non-centered data) are
present.
© McGraw-Hill Education. 3-76
Control Charts with non-random errors

© McGraw-Hill Education. 3-77

Control Chart Construction
1. Compute the MSE.
2. Estimate of standard deviation of the distributi on of errors
s  MSE
3. UCL : 0  z MSE
4. LCL : 0  z MSE
where z  Number of standard deviations from the mean

© McGraw-Hill Education. 3-78

• If non-randomness is found, corrective action is
needed.
• The removal of a pattern (evidence of non-
randomness) usually results in less variability in
forecast errors and narrower control limits

© McGraw-Hill Education. 3-79

Comment
• Control chart approach is generally superior
• Major weakness of tracking signal approach is the
use of cumulative errors (large positive and
negative values cancel each other
• With control charts, every error is judged
individually

© McGraw-Hill Education. 3-80

 Example 5 Forecast demand using exponential smoothing.
Determine if the forecast is “in control” using a control chart.
Exponential
Smoothing
Period,
t
Demand
(α =
Error Error^2 Step 1: Forecast
1 47
0.10)
using the indicated
2 45 47 technique.
3 48 46.80 Step 2: Compute the
4 45 46.92
forecast errors.
5 46 46.73
6 54 46.66 Step 3: Compute the
7 51 47.39 average error.
8 49 47.75 Step 4: Compute the
9 50 47.88
10 43 48.09
standard deviation
11 45 47.58 using
12 47.32 s=
© McGraw-Hill Education
Example 5 Forecast demand using exponential smoothing.
Determine if the forecast is “in control” using a control chart.
Exponential
Period, Smoothing
Demand Error Error^2
t (α =
0.10) Step 5: Determine
1 47 the control limits
2 45 47
using .
3 48 46.80
Step 6: Check if all
4 45 46.92
5 46 46.73 errors are within the
6 54 46.66 limits.
7 51 47.39 Step 7: Plot the data
8 49 47.75 to check for non-
9 50 47.88 random patterns.
10 43 48.09
11 45 47.58
© McGraw-Hill Education
Example 5 solution
Exponential
Period, Deman Smoothing
Error Error^2
t d (α =
0.10)
MSE
1 47
s
2 45 47.00
3 48 46.80
4 45 46.92 Control limit
5 46 46.73
UCL
6 54 46.66 LCL
7 51 47.39
8 49 47.75
9 50 47.88
10 43 48.09
11 45 47.58
12 47.32 © McGraw-Hill Education
 Example 5 solution
Exponential
Period, Deman Smoothing
Error Error^2
t d (α =
0.10) MSE 12.8057
1 47
2 45 47 -2 4.0000 s 3.5785
3 48 46.8 1.2 1.4400
4 45 46.92 -1.92 3.6864 0±2(3.5785
5 46 46.73 -0.728 0.5300 Control limit )
6 54 46.66 7.3448 53.9461
7 51 47.39 3.6103 13.0344 UCL 0+7.157 7.1570
8 49 47.75 1.2493 1.5607 LCL 0-7.157 -7.1570
9 50 47.88 2.1244 4.5129
10 43 48.09 -5.0881 25.8885
11 45 47.58 -2.5793 6.6526
12 47.32

© McGraw-Hill Education
 Plot the errors on the chart.
An error above
the UCL.

e forecast error, 7.3448, is beyond the control limits. Hence, the forecast was not “in con
forecast error exceeded 2 standard deviations above the control limit.
© McGraw-Hill Education
Tracking Signals
This method relates the cumulative forecast error to the
average absolute error (MAE or MAD). The intent is to
detect any bias in errors over time. The tracking signal is
computed period by period using the formula:

Values can be positive or negative.

• A value of zero would be ideal.
• Limits of ±4 or ±5 are often used for a range of
acceptable values of the tracking signal.
• If a value outside the range occurs, that would be taken
as a signal that there is bias in the forecast.
© McGraw-Hill Education. 3-86
Tracking Signals
After an initial value of MAD has been determined,
MAD can be updated and smoothed (SMAD) using
exponential smoothing:

© McGraw-Hill Education. 3-87

Example 6 Tracking Signals
Monthly attendance at financial planning seminars
for the past 24 months, forecasts, and errors for those
months are shown in the given table (please refer to the
Excel file Tracking signal Example).

Determine if the forecast is working using a tracking

signal, beginning with month 10, updating MAD with
exponential smoothing. Use limits of ±4 and α = 0.20.

© McGraw-Hill Education. 3-88

Example 6 Tracking Signals

© McGraw-Hill Education. 3-89

Example 6 Tracking Signals

© McGraw-Hill Education. 3-90

Choosing a Forecasting Technique
• Factors to consider
• Cost
• Accuracy
• Availability of historical data
• Availability of forecasting software
• Time needed to gather and analyze data and
prepare a forecast
• Forecast horizon

© McGraw-Hill Education. 3-91

Guide to selecting an appropriate forecasting method

© McGraw-Hill Education. 3-92

Forecast factors, by range of forecast

© McGraw-Hill Education. 3-93

Operations Strategy
• The better forecasts are, the more able
organizations will be to take advantage of future
opportunities and reduce potential risks
• A worthwhile strategy is to work to improve
short-term forecasts
 Accurate, up-to-date information can have
a significant effect on forecast accuracy:
o Prices
o Demand
o Other important variables
© McGraw-Hill Education. 3-94
Operations Strategy
• Reduce the time horizon forecasts have
to cover – shortening lead time needed
to respond to a forecast
• Sharing forecasts or demand data
through the supply chain can improve
forecast quality in the supply chain,
resulting in lower costs and shorter
lead times. HP & IBM require resellers
to include such information in their
contracts
© McGraw-Hill Education. 3-95
Assignment:

© McGraw-Hill Education
End of presentation