Application of Forecasting Methods in Business
Application of Forecasting Methods in Business
To cite this article: Laima Šečkute & Arnoldina Pabedinskaite (2003) Application of forecasting
methods in business, Journal of Business Economics and Management, 4:2, 144-157
--
B B
.
:
6
F7
-
E!
H
.
DS9
GE,J%(% 6
!" R !%%6
$ &&
&
$
1 < & +
1 &
1&
$ +
-$" 1
1 & & $ & &
$ & & +5 $ $ 3 &
& ,
& +
& &&
4& 3 '4) &&
*
& +
< 3
$ 11
*
& 1
&
1 1 & < 2 +
" ( 3 " ""% /
$ 2 1 4, 1
3 2 &
&
$ &
& 2
+
887
5 5Q Q547
>4
F5# G4
& 2&
& 3
1 &
+
8
& & $
&
&&&3& &
+
& 2 2
& &
+
&3&1 & &
1
& 1
<
< &
&
, +
&
+
& &&
&
&
$
+
$
& +&
8 & &
&
&
& & &
&
&
$
&
1&
2 +&
&3 +8
&
$ 1 &
<
< &+
& 2
&
&
1 & + F1
&
& $ ? +
& 2 1 & &
& $
$
& &+ & & 1 & &
8
& $ <
1 & +
7
& &3
$ 1
G
- XT9S
7
7
9
S
1
1 , + &
1
&
+
#
<
+
$
& &
1
& <
&
<
+
&
< & + •
&
$
D$ <
1E
+ 7
+
& ,&
$
&
1 & 2
9
1
&& "6K & &&
, &
+
&+
&
& &
<
$ < & $ 1
,&
J
& &3 <
& < & MKC & &3
,&
-.K/+ I &
$
+> C%& &$
1 &&
& 1
&
& &
887
5 5Q Q547
>4
F5# G4
$
& + &9
1
1
1
+ $
&
&
10 !) &9
& && &9 1
$ &3
G
- XT9S
7
7
9
S
&
&9 +
&
.+ E7 )0
7 -!
@ -!7 $!
6 &
$
&
-!7 -.
@ .@ )0
$ &
-!
@&
1 &+
$
&
& & 1
$ &
&&
,&
+ &
&
,&
&
&
+
1 &
1 + &
+ F1 $
$
<
&& $
,&
, & & +
&
$ ,&
&
11
&
2
<
2
$ &
$ &3
+
& <
< 3 +
+
.+
&
$ $
$
'& . )
&
$ + Q
'.6 )
'6.% ,& &
!+
$9 &
,&
& & $ 1 &
3
6+ 1
&&
$ J* &&+
$
2
6+ D -0
' )
2 ' )
8 1 &
887
5 5Q Q547
>4
F5# G4
&3 && 1<$
&&
& 2 +
&
< &
,&
1
B+ L
)
. 0
$ <
+ 7&
&$ $ ,& + F1
&& & 3
,& $ J 9
<
$ $
&$+
&
+
$ &&
1
& $
& $
$ &
&
+
< + 5
1 & 1
$
$
2<
1
$ $ + >
&
&$ & 9
$
$ $9
,
$
& +
+ & & &
& & $ &
$ ,& 1 < &
&
$ &
$
+
$
+ 2 $ &&
C+ ?- .
$ ,&
$
+
$
&
,&
$ $ $ &
$
2
&
+ I & 2
&$ $ &
&
&
3
$
+
&
,&
,&
+F1
$
& & &
&&$ &&
+
0 1 &
+ I &&
9 ,& 3
2 < &$ + 8 1 &
G
- XT9S
7
7
9
S
&
I & 2 $
+ &&
& <
$ $ +
&
+ F1
& < &
$
& V
& +
2
I &$
1
& & ,& $
&$ $
1 J 9
11$
V
$
<
1 & & &
& 1
+
1 ,& 3&
&+
$
,&
$ .%& +
1
+
1 ,& 3 & + 5
$
+
D
E 1
$ $ +
,& J $ 2
$ &$ < + + .% I & &
$
1 & &
,&+
1 2 $9
&
-6 B .% .. .6 .B .C .H!%/
1
1 < ∑ Ft − Yt
MAD = ; '!)
2 +
&&
n
> Q 4
&
$ !+
66B & 1
MFE =
∑ (Ft − Yt ) ; '6)
&&
1
n
1 2 &&
1
1 -.B/ & &&
1
,& 1 Ft − Yt
MAPE = ∑ . 'B)
n Yt
9 , & 1 & &
$ &
< & 2
+
Ft
887
5 5Q Q547
>4
F5# G4
$" ,&
Qual i t at i ve met hods Such evaluations may be obtained quickly and The chief executives of a company or the staff of
Juries of executive at a low cost. They may either be the basis for it s s a le s o ffic e c a n o nly d e t e r mine ge ne r a l
opinion a final forecast or may be compared with the tendencies, because there is a danger to arrive at
forecasts made with the help of other methods. a fo re c a st no t re fle c ting a c tua l situa tio n. This
The technique is an effective one when applied me t h o d r e lie s t o o h e a v ily o n a s u b j e c t iv e
as a complementary method. judgement. The results of a forecast are influenced
by the opinion of a top- ranking executive (- s),
w ho s e j ud ge me nt fo r ms t he b a s is fo r o t he r
evaluatio ns. Therefo re, there is a risk that the
opinion of a single person will be dominant.
F o re c a sting b a se d o n C onsumer questioning provides reliable data. In the consumption market, the method is suitable
consumer surveys The accuracy of a forecast increases when it o nly whe n a s ho rt- te rm fo re c a s t is p re p a re d .
makes use of a probability scale rather than O therwise, the representation of the forecast is
consumers' answers about their intentions to buy reduced. Sometimes consumers do not answer the
or not to buy a specific product. Consumer polls company's questions, and their actual behaviour
supply a company with information about the differs from their intentions. In many cases, the
in t e n t io n s t o b u y a p r o d u c t , le v e l o f number of consumers is too large or it is impossible
consumption, buying periods, and reasons for to identify potential customers. For a survey to be
buying. With respect to industrial goods, a poll s uc c e s s ful the me tho d r e q uir e s c o ns id e r a b le
is successful when there are several customers experience, it is necessary to properly prepare the
who are able and willing to predict their future question- naires, conduct the survey, and evaluate
behaviour. Such surveys provide the data that the results.
cannot be obtained in any other way.
The Delphi method The responses of experts are anonymous, which The p ro b le m o f c o nse rva tive te nd e nc ie s is o f
encourages them to be open. This method is a c o ns id e r a b le s ignific a nc e he r e . To p r e d ic t
convenient one for people who cannot meet in technological changes is quite complicated. If an
o ne p a r tic ula r p la c e . At e a c h s ta ge a ne w expert is not motivated to conduct a survey, he/she
questionnaire is prepared on the basis of the data may present his/her evaluation indifferently and
obtained at an earlier stage, which increases the irresponsibly. It is possible to arrive at a misleading
volume of the data on which the participants can c o nc lus io n. This me tho d is le s s r e lia b le tha n
base their opinion. Such a systematic taking into consumer surveys.
account of an expert's opinion may be applied
to solve almost any problem, no matter what
products serve as the basis for the forecast.
Sales force composites B y ma k ing us e o f t he ir ma r k e t e xp e r t is e , The re is a ris k tha t s a le s re p re s e nta tive s will
salespeople should develop quite a successful u n d e r e s t ima t e t h e d e ma n d f o r t h e f e a r o f
forecast, because their closeness to the market overestimating the sales volumes. According to Mr
helps them to take into consideration even short- F. Kotler, judgement by a member of the sales staff
term changes. The interviews of the sales staff may be influenced by his/her personality (pessimist,
o ffe r a n o p p o r tunity to ge ne r a lis e s p e c ific optimist) and his/her success or failure at work.
tendencies, to determine the advantages and Moreover, they are not always able to make a
limitations a company's products and services, judgement about the tendencies of the economic
competitors' strategy, characteristic features of development of the country they work in. Thus, the
the ma ss c o nsume r, a nd to p re p a re a sa le s sales staff may incorrectly interpret consumer's
fo r e c a s t o n t he b a s is o f d iffe r e nt p r o d uc t needs and may be unable to distinguish between
c a te go r ie s , c o ns ume r gr o up s , a nd r e gio ns . what consumers would like to do and what they
Salespeople are in direct contact with consumers ac tually will d o . F urthe rmo re , sale s re p re se n-
and most probably know their likes and dislikes. tatives may not have enough time to thoroughly
Q uite often they also know about their future prepare a forecast or may consider forecasting
plans and expectations. completely meaningless.
G
- XT9S
7
7
9
S
Quant i t at i ve met hods The method is uncomplicated in application, One presumes that the regularities that manifested
Trend extrapolation b e c a us e it d o e s no t r e q uir e a ny a d d itio na l t he m in t he p a s t w ill c o nt inue in t he fut ur e .
expenditure except for computation cost. However, such an assumption is risky, especially
when the mark et is d evelo p ing d ynamically o r
when the forecast under consideration is a long-
term one. Any further planning, alongside all the
numerical and statistical methods should also take
into account qualitative evaluations. However, the
s a le s fluc tua tio ns tha t to o k p la c e in the p a s t,
e c o no mic s itua tio n, e ve r - c ha nging c o ns ume r
t a s t e s , c o mp e t it io n, a nd d e gr e e o f ma r k e t
saturation are ignored.
Time series smoothing The technique is not complicated in application The variables that determine the values of time
(the methods of moving and helps to prepare a forecast quickly and series remain unid entified . P ast trend s will no t
averages and e a s ily. T he me t ho d und e r d is c us s io n is a necessarily co ntinue in the future. The metho d
exponential smoothing) forecasting technique that is well understood by should not be applied when corporate conditions
consumers. It is possible to adapt to changing are unstable and frequently change. The technique
conditions by means of changing the value of a is most suitable when applied to relatively stable
smoothing constant and the length of the period, processes or stationary time series.
which is the basis for averaging.
Correlation and Forecasting on the basis of these methods is Ind e p e nd e nt va ria b le s ma y b e inte rre la te d . A
regression analysis done at a relatively low cost. The techniques forecast may be misleading if important factors
take into consideration the factors that determine change d uring the fo recasted p erio d o r if new
the forecasted value as well as the degree of their fa c to rs a p p e a r. An a c c ura te fo re c a st re q uire s
impact. extensive historical data. The forecasted values are
fr e q ue nt ly influe nc e d b y t he fa c t o r s w ho s e
quantitative expression is rather complicated.
$
$ sa0 = sa1 ⋅
∑ xi2 , 'H)
&+ n
7
'
& 1 2
&
$ ) & + 5 2 $ 1
Yˆ = Y + a1 (x − x ), 'K)
Y = a0 + a1 x + ε, 'M)
1 S > 2 #
<
1 ∑
1 %
1 ,
<
%
.
1 &
>
# + +
,
1
sres
sa1 = ,
'")
∑ (xi − x )2 Yˆ − t ⋅ s ; Yˆk + t α ⋅sˆ .
k α ;n −2 Yˆk ;n − 2 Yk '..)
2 2
887
5 5Q Q547
>4
F5# G4
(Yˆ − 2,069 ⋅ s
k Yˆk
; Yˆk + 2,069 ⋅ sYˆ .
k
) '.!) $ $
1
S $ , 2 ,& 1
S
2 $ &
&9
sYind = sres 1 +
1
+
(xk − x ) . 2 2
$
2 1
+
2
&
1 < &
$9 + 4,& 1
$
+ Q
+
$ &
& $
1
+
+
1
&
2
&
$
&
&&
$ 2
&
&
+
&&, 2
$ &
#
&
1
m 9
$ ,
& &
#
&
1 &
1 $
9
$ ,
F −t ⋅ s ; F + t ⋅ s ,& + Q !6 $
int α ;n −2 int int α int . '.M)
;n − 2 $ &
2 2
+ 1 $
+4
1 <
G
- XT9S
7
7
9
S
,& 1 1 )
$
0 2 &
+ & &
&
$ <
$ $
+
&
& $
1 2
&
<
,& & +
& + Q && &3 &
$ && # &
+
+
<
$
&3 &&
&
1<
<
&;&&
< + $ & & & 3
& 1 + &
&
, $
)$ & &
"# B"% #2
1 $
#
& $1 1$
&
1
&3
&
&$ && +
&&
<
<
&
Q+.+
1 1< &
&
$
&
1 &
&
& &3 0
+ •
$
0
,&
&
< 1 &
1
&
+
$
& &3
0
$ * < 1
1
) 2
&
1 &
1 &
& , < 3 1
3 +
$ & & <
&
'1
887
5 5Q Q547
>4
F5# G4
%&#"'% #&)-%. !.0"'/) !$")
!"#$%&#"'% # /#&0#"'%) /#&0#"'%) /#&0#"'%)
'()') +))'% -%$% 0." -%$% 0." -%$% 0."
)(%%"*'+ ##&,)') (# (# (#
1
2
34
5
G
- XT9S
7
7
9
S
&
$
1 & ,&
<1
$ <
+ &+
<
$ 2
2
3
&
1 1 &
&3 & &
&
+
& < +
1
& +
$
+&&
$
F
+
&
&& 1
4 && 2
&
2 1
&
1
$ $9 2
1 $
1 $
&3 &&
1 ,& 3 & 1
&
+
& 3
$ ,&
!+
&
$
'& . )
&
$
&&
,
2
< 1
&
+ 2
2 +F1
,& 9 ,
& ,
&
&
,&
,& 3 & & & &&
$ &
&$
$
& &
887
5 5Q Q547
>4
F5# G4
+
K+ >i + '!%%%)+ 4< 9 &*+
< & + + .!" &+
..+ 5 O + F + R + ' . K K C ) + 4 _
$ <
+
G $ +
1
O C6M +
"3""#%"
.MM &+
.6+ I+R+'.KK%)+8
J5& >
.+ >+ '.KK6)+ F
$< >
+ G F 1
1+ K!K &+
2 + O 8
+
$ + BBM + .M+ ] h<\ + '.KK6)+ l< <& &
*+
><
$+ 6 9-
9 +
B+ G <1* 4+ + O + +
I+ '.KHK)+
<+
>< + 1+ "C! &+