10 Forecasting
10 Forecasting
and
FORECASTING
Operations Management
Dr. Ron Tibben-Lembke
Demand Management
Coordinate sources of demand for supply chain
to run efficiently, deliver on time
Independent Demand
Things demanded by end users
Dependent Demand
Demand known, once demand for end items is
known
Affecting Demand
Increasing demand
Marketing campaigns
Sales force efforts, cut prices
Time Horizons
Different decisions require projections about
different time periods:
Short-range: who works when, what to make each
day (weeks to months)
Medium-range: when to hire, lay off (months to
years)
Long-range: where to build plants, enter new
markets, products (years to decades)
Forecast Impact
Finance & Accounting: budget planning
Human Resources: hiring, training, laying off
employees
Capacity: not enough, customers go away angry,
too much, costs are too high
Supply-Chain Management: bringing in new
vendors takes time, and rushing it can lead to
quality problems later
Qualitative Methods
Sales force composite / Grass Roots
Market Research / Consumer market surveys &
interviews
Jury of Executive Opinion / Panel Consensus
Delphi Method
Historical Analogy - DVDs like VCRs
Nave approach
Quantitative Methods
Time Series Methods
0.
All-Time Average
1. Simple Moving Average
2. Weighted Moving Average
3. Exponential Smoothing
4. Exponential smoothing with trend
5. Linear regression
Causal Methods
Linear Regression
All-Time Average
To forecast next period, take the average of all
previous periods
Advantages: Simple to use
Disadvantages: Ends up with a lot of data
Gives equal importance to very old data
Moving Average
Compute forecast using n most recent periods
Jan
Feb
Mar
Apr
May
Jun
Jul
Moving Average
Advantages:
Disadvantages:
Ja F M Ap My Jn Jl Au S O N D
M
Avg of Jan-Dec gives average of month 6.5:
Ja F
(1+2+3+4+5+6+7+8+9+10+11+12)/12=
Avg of Feb-Jan gives average of month 6.5:
6.5
(2+3+4+5+6+7+8+9+10+11+12+13)/12
=7.5
How get a July average? Average of other two averages
Old Data
Comparison of simple, moving averages clearly
shows that getting rid of old data makes forecast
respond to trends faster
Moving average still lags the trend, but it suggests
to us we give newer data more weight, older data
less weight.
Exponential Smoothing
Ft Ft 1 At 1 Ft 1
Ft 1 Ft 1 At 1
Exponential Smoothing
Smoothing Constant between 0.1-0.3
Easier to compute than moving average
Most widely used forecasting method, because of
its easy use
F1 = 1,050, = 0.05, A1 = 1,000
F2 = F1 + (A1 - F1)
= 1,050 + 0.05(1,000 1,050)
= 1,050 + 0.05(-50) = 1,047.5 units
BTW, we have to make a starting forecast to get
started. Often, use actual A1
Alpha = 0.3
Alpha = 0.5
Exponential Smoothing
Ft At 1 1 Ft 1
We take:
And substitute in
to get:
Ft 1 At 2 1 Ft 2
Ft At 1 1 At 2 1 Ft 2
2
Choosing
Low : if demand is stable, we dont want to get
thrown into a wild-goose chase, over-reacting to
trends that are really just short-term variation
High : If demand really is changing rapidly, we
want to react as quickly as possible
Averaging Methods
Simple Average
Moving Average
Weighted Moving Average
Exponentially Weighted Moving Average
(Exponential Smoothing)
They ALL take an average of the past
With a trend, all do badly
Average must be in-between
30
20
10
T t T t 1 Ft FITt 1
where and are smoothing constants
T 2 T t 1 Ft FITt 1 T 1 F2 FIT1
0.20 0.30
F3 FIT2 A2 FIT2
Selecting and
You could:
Try an initial value for each parameter.
Try lots of combinations and see what looks best.
But how do we decide what looks best?
Evaluating Forecasts
How far off is the forecast?
Forecasts
Demands
Evaluating Forecasts
n
Mean Absolute
Deviation
MAD (1/n) At Ft
i1
n
Mean Squared
Error
MSE (1/n) At Ft
Mean Absolute
Percent Error
At Ft
MAPE (1/n)
100
Di
i1
i1
Tracking Signal
To monitor, compute tracking signal
RSFE
Tracking Signal
MAD
n
RSFE At Ft
t 1
Forecast Error
Upper Limit
-10
Forecast Period
Lower Limit
Updating MAD
Simplified calculation avoids keeping
running total of all errors and demands:
1.25 MAD
yt a bt
Computing Values
xy n x y
b
x nx
y b x
a
y bx
2
S yx
(
y
Y
)
i
i
i 1
n2
Linear Regression
Three methods
Causal Forecasting
Linear regression seeks a linear relationship
between the input variable and the output
quantity.
yc a bx
920/500 = 1.84
1.84 * 50 million = 92 million videos?
Fortunately, not that dumb.
Lessons Learned
Flooded market with DVDs
Guaranteed Sales
Promised the retailer they would sell them, or else the
retailer could return them
Didnt know how many would come back
5 years ago
Typical movie 30% of sales in first week
Animated movies even lower than that
1993
1994
1995
1996
Seasonality
Seasonality is regular up or down
movements in the data
Can be hourly, daily, weekly, yearly
Nave method
N1: Assume January sales will be same as
December
N2: Assume this Fridays ticket sales will be
same as last
Seasonal Factors
Seasonal factor for May is 1.20, means May sales
are typically 20% above the average
Factor for July is 0.90, meaning July sales are
typically 10% below the average
Sales
Factor
200200/250 = 0.8
350350/250 = 1.4
300300/250 = 1.2
150150/250 = 0.6
Total
1,000
Avg 1,000/4=250
1994
1995
1996
1990
1991
1992
1993
1994
1995
1996
1989-2007
1989-2007
1998-2007
Cache
Creek
9/11
Thunder CC
Valley Expands
2003Q3 - 2007Q3
2003Q2 - 2007Q3
2003-2007
Compute Indexes
Q
For each Q:
Date
Quarter
Win
59276,371
60235,766
2004
61240,221
62259,350
63279,758
64245,811
2005
65231,608
66259,687
67297,414
68260,149
2006
69245,775
70269,670
Avg
Index
1240,562
0.9168
2265,456
1.0117
3289,187
1.1022
0.9693
Divide
Total Avg by Total Avg to get Ind
240,562/262,382
Avg.
262,382 = 0.9168
2003-2007
period
Win
59276,371
2004
2005
2006
Deseasonalize
d
250,755
60235,766
243,236
61240,221
262,010
62259,350
256,347
63279,758
253,828
64245,811
253,598
65231,608
252,616
66259,687
256,681
67297,414
269,847
68260,149
268,391
69245,775
268,069
70269,670
266,548
71294,839
267,511
Do LR on deseasonalized da
Linear
intercept
185,538.00
slope
1,119.91
251,613
252,733
253,853
254,972
256,092
257,212
258,332
259,452
260,572
261,692
262,812
263,932
265,052
266,172
rsq
0.497
Create
Linear
Forecasts
Int + slope * period
Seasonal Forecast
2004
2005
2006
58257,062
Deseasonalized Linear
Forecast
59276,371
250,755
251,613
277,317
60235,766
243,236
252,733
244,972
61240,221
262,010
253,853
232,741
62259,350
256,347
254,972
257,959
63279,758
253,828
256,092
282,254
64245,811
253,598
257,212
249,314
65231,608
252,616
258,332
236,848
66259,687
256,681
259,452
262,491
67297,414
269,847
260,572
287,191
68260,149
268,391
261,692
253,656
69245,775
268,069
262,812
240,956
70269,670
266,548
263,932
267,023
Q
1
2
3
4
Index
0.9168
1.0117
1.1022
0.9693
Multiply Linear
forecast by
indexes
251,613 * 1.1022
= 277,317
267,291 * 0.9168
= 245,063