Escolar Documentos
Profissional Documentos
Cultura Documentos
Forecasting
Itwouldbeveryhelpfultodecisionmakersiftheyhad
anideaofwhatislikelytohappeninthefuture.
Itispossibletoforecastusinghistoricalnumericaldata,
e.g.salesfigures
ThisisknownasTIMESERIESANALYSIS
BusinessStudiesOnline:Slide1
Time-Series Analysis
Analysis
Thereare4stagesinvolvedincarryingouttime
seriesanalysis:
1)Identifyingthetrend
Thisisthemostimportantaspect,andrequiresbusinesses
toidentifywhathasbeenhappening.
2)Identifyingvariations
Recessionsandboomswithintheeconomywillaffectmany
typesofbusinessdata,andthusmustbeaccountedfor.
3)Identifyingseasonal/cyclicalvariations
Somebusinessesareseasonalinnature,iesalesarehigher
insummer.Thesemustalsobeaccountedfor.
4)Acceptingrandomvariations
freakeventsdooccur,andwillaffectbusinesses.E.g.
September11th2001.
BusinessStudiesOnline:Slide2
Tohelpthemtodothisbusinesseswillsmoothoutthe
rawdatabyusingMOVINGAVERAGES
Themovingaverageisusuallyeither3periodor4period,
dependinghowoftentheaverageistaken.
BusinessStudiesOnline:Slide3
Sales
(m)
67
70
64
61
59
66
73
69
64
63
69
75
72
3Point
Total
3Point
Moving
Average
201
195
184
186
198
208
206
196
196
207
216
67.00
65.00
61.33
62.00
66.00
69.33
68.67
65.33
65.33
69.00
72.00
BusinessStudiesOnline:Slide4
A Graphical Representation
Representation
Itisthenpossibletoplotboththeoriginalandtrenddata,
toseehowthedatahasbeensmoothedout:
80.00
3Point
Moving
Average
Sales
(m)
70.00
60.00
05
20
03
20
01
20
99
19
97
19
95
19
93
50.00
19
Sales(m)
TimeSeriesAnalysis:
3PointMovingAverage
Year
BusinessStudiesOnline:Slide5
Usingthepreviousdata,wewouldcalculatethe4period
movingaverageasfollows:
BusinessStudiesOnline:Slide6
Year
1993
Sales
(m)
67
1994
70
4Point
Total
8Point
Total
4Point
Moving
Average
516
64.50
504
63.00
509
63.63
526
65.75
539
67.38
541
67.63
534
66.75
536
67.00
550
68.75
262
1995
64
254
1996
61
250
1997
59
259
1998
66
267
1999
73
272
2000
69
269
2001
64
265
2002
63
An Example
4 Period
Moving
Average
Average
Ifrequiredthisdata
couldalsobe
showngraphically
271
2003
69
279
2004
75
2005
72
BusinessStudiesOnline:Slide7
20
00
19
98
19
96
19
94
19
92
19
90
80.00
75.00
70.00
65.00
60.00
55.00
50.00
19
88
Sales(m)
TimeSeriesAnalysis:
3PointMovingAverage
Lineof
BestFit
3Point
Moving
Average
Sales
(m)
Itisthena
simplecaseof
takinga
readingfrom
thelineofbest
fittoobtaina
futureforecast
Year
BusinessStudiesOnline:Slide8
The Variations
Variations
Areadingfromthelineofbestfitisunlikelytobeaccurate
becauseitdoesnotconsidervariationsinthedata,e.g.
rawdataisrarelyplottedasastraightline!
Inordertoaccountforthese,andobtainmoreaccurate
forecastsitisnecessarytocalculatethevariation.
Thisiscalculatedusingtheformula:
Variation=ActualDataTrendData
BusinessStudiesOnline:Slide9
Sales
(m)
67
70
64
61
59
66
73
69
64
63
69
75
72
3Point
3Point Moving
Total Average Variation
201
195
184
186
198
208
206
196
196
207
216
67.00
65.00
61.33
62.00
66.00
69.33
68.67
65.33
65.33
69.00
72.00
3.00
1.00
0.33
3.00
0.00
3.67
0.33
1.33
2.33
0.00
3.00
BusinessStudiesOnline:Slide10
Inourexampletheaveragevariationwouldbe:
2.00/11=0.182Million
=182,000units
Thismeansthatanyreadingobtainedfroma
lineofbestfitwouldhave182,000unitsadded.
Intheorythisshouldyieldamoreaccurate
forecast
3.00
1.00
0.33
3.00
0.00
3.67
0.33
1.33
2.33
0.00
3.00
2.00
BusinessStudiesOnline:Slide11
BusinessStudiesOnline:Slide12
Sales
4Point
(000s)
Total
120
140
580
190
590
130
610
130
640
160
670
220
670
160
680
130
700
170
710
240
740
170
760
160
770
190
780
250
180
4Point
8Point Moving
Total Average Variation
1170
1200
1250
1310
1340
1350
1380
1410
1450
1500
1530
1550
146.25
150.00
156.25
163.75
167.50
168.75
172.50
176.25
181.25
187.50
191.25
193.75
43.75
20.00
26.25
3.75
52.50
8.75
42.50
6.25
58.75
17.50
31.25
3.75
Theaverage
seasonal
variationisthen
calculatedby
addingupthe
figuresfor
corresponding
quarters,and
dividingbythe
numberof
piecesofdata
BusinessStudiesOnline:Slide13
Quarter2
(3.75+6.25+3.75)/3=4.58
Soallforecastsforquarter2wouldhave4,580subtractedfromthem
Quarter3
(43.75+52.5+58.75)/3=51.67
Soallforecastsforquarter3wouldhave51,670addedtothem
Quarter4
(20+8.75+17.5)/3=15.42
Soallforecastsforquarter4wouldhave15,420subtractedfromthem
BusinessStudiesOnline:Slide14
BusinessStudiesOnline:Slide15
Disadvantages of Using
Time Series Analysis
Analysis
Itisalongwinded,andcomplexprocess,particularlyif4
periodmovingaveragesareused
Historicaldataisnotalwaysagoodindicationofwhat
mighthappeninthefuture
Itisnotveryusefulforlongtermforecasting
BusinessStudiesOnline:Slide16