Professional Documents
Culture Documents
Explatory research
Marketing Research Systematic gathering
:
¥ Preliminary examination
of problem Situation
of
marketing related data producer
→ -
research
Descriptive
wno ? what ? when ?
qualitative vs .
quantitative
causa research
Delermines wether one ( or more )
FEATURE QUALITATIVE QUANTITATIVE variables causes an other
Typf Exploatory
"" "
Ptivesual
question maire
SAMPLE SIZE Sma " large
① Introduction
QUESTIONS unstrut und Structure d
② Screening Question
'
ANALYSIS subjective
Objective ③ Survey Question
Statistical
④ Demographie information
GENERALHABIHTY limited high ⑤ ( losing
- - -
- - - - - - - - - - - - - - -
or " , →
multiple
-
choice
⑥ or ④ questions .
mumm
D-
Ordinul Ranking manageable amount of answers
:
a
D -
( Winner )
occasionally
ranking
DVD
or Situation Always
-
-
. . .
Intervall :
Resutt or answer
Ying in a rixed
range → scaled response
-
soooo
-
00000
Ratio : answer or fait of the time
One right ,
most - oo o oo
_ 00000
E.t.a.
: Game amount
Scale bad answers which can be identitied without
missunderstanding .
° "
- /
,
,
i
' '
.
. : : -
'
.
-
.
:
-
Cases section one :
① done ? 0
"
-
→ Excersise
÷;;
- i
② done ?
.
. .
→ Excersise Q
→ Excersise ③ done ? OV
notreliable rentable reliable
→ Excersise ④ done ? Or
not valid notvalid valid
→ Excersise ⑤ done ? 0
HANK HHHNHIIIYIK MIR IINNINIDN
assigning a 1 =
last ROW .
placinga mark .
This
way the relative im Porta nee can be Show M
Endsrepresent
.
> 0 , a I 100
the extremes .
Likert Scale :
widdy used rating Scale that Semantic differential :
Rating scale with
congideraticenf in
oceatincf marketing Halef
:
① Number of
categories
:
No pre fixed-
amount ,
between fire and nine are kommended .
④ Force d vs .
Unforced :
If the responderts have rio opinion , data maybe improved by using a non
Force d Scale .
⑤ Verbal description :
⑥ Physica form : A humber of Option should be fried and the best one Select d .
In
Systematic error
accuracy in Observation or
/ DE Ability
vaeidity
of a measurement scale to
=
.
→ ordinal ,
nominal . . .
,
→
→ How to ask the
questions ! Companion → which one fits best
→ Oder of rank
;
.
→
Giving point s
Non Comparative Scales
Reliability Validity
.
vs
-
i
the
opposite a- R the some Output
→
Always
- - -
-
→ :
→ v :
Chasing the
right method
Multi item Scale
Single
vs.
→
only measuring one Concept ( Single ) Considerations in marketing Scales
→ measure more
Lagers of a concept ( Multi) →
numberofcategories balanced , vs .
Unbalanced ,
MOVING AVERAGE
Hanning
base
caecueating of
average
→ on
→
good if data is Stable
2020
2021
for
january
:
Moving average
191-21 20
2- Month : =
3- month :
191-211-19 = 19,5
3
1Gt 211-19+22
4- month : =
20,25
4
of 2
Most recent month
wright
:
,
1. 19 t 1 -
21 t 2 -
19 =
19,5
4
Haarwäsche actyalvalveperiodt
EXPONENT IAL SMOOTH ING Et t d .
Htt -
Et )
:
Etta
data
weights recent past more
strongly
per:{¥ aufginge tforecastnext
→
→ × dose to 1 =
greater reaction to most recent demand period
→ 2=0
,
forecast some as last period
zozo
Dez 14,71-0,1-(21-14,7)=15,3
Jan 15,31-0,1-(19-15,3)=15,67
It I
:
at b. xt E
I habe
'
( time )
Period Slope
55
F =
:
5,5
so
873
g- 87,3
=
=
10
|
5342 10 5,5 87,3 540,5
455
-
Jt
- -
6,552
=
" = =
Linear trend line : :
51,27 t .
×
82.5
385-10 .
g. sz
6,55 -
5,5=87,3 -
36,04=51,27 Yt :
51,27T 6,55 *
11=123,33
2020 :
(521-511-51)=51,33
3
220
HANK IIHNHHNYK MIR 1114MDM
Hanning
ERROR S Forecasting error
=
actual value -
forecast valve :
At Ft
-
Ft *
Er:c:p!!! "
Shooting .
\
0
MAD
:
402 =
Summe At Ft
vnon
-
MSE :
'
Summe von ( At Ft )-
%
:
MAPE :
Summe At Ft
vonn
-
.
100
( 0,089/1 )
- i -
.
100 =
HAHA HIINHIINYK MIR IINNNIIDN
linear
negereggion
dependent
§
of
Regression is a
way of predicting the value of ßestmäted
variable
value
independent variable
Variable from
>
One an other r
→ linear relationship =
equations of a
straight line § = a t b Xt .
Eins erregern
in ept
with Slope of line ,
[ AL ( UL AT I ON OF y -
axis strengt of
relationship
REGRESSION LINE
34
:
7
779 900-12 .
249,67 '
254,75
D=
:
1,048 8
763910-12 .
249,672 q
10
a
:
254,75 -
1048 '
249,67=-6904 '
11
§ :
-6,90 t 1,05 ×
12
;
↳ Mittelwert
Assessment of regression
coefficients
eine
Regression
:
effects
of demand Variation →
model fit ( in
predidingy )
→ t tests
-
significance
( deviding coefficients by
it's Standard error )
Ö
" "
0,011<0105
→ The
average selling price
of a house inverses by
✓ for
1047,79€ every Square
\ meter increase in size
HAHA IIHNHIINYK MIR IHN )
CONJOINT ANALYSIS
Relative importance consumers attacke to saeient attributes and the
Utilities they attack to the levels of attributes .
numberoflevelsof attribute i
→
decompositional approach humber of
f sit the jth level
attributes
] k, 9%.tt?buiethispresent
Relative importance wrights UCX ) : [ [ % Xiyoiiothewise .
{ i)Jt f
of an
utility part North Contribution
-
alternative
or
utility associated
with the jth level
④ Overall
Utility for nonstop
flight with air
cologne
and Price 409€
↳ 2 t 0,50T 0
=
2,5
attribute :
span
- be used
Conjoint analysis can
① (1-1-1,50)) t (0,67 C- 0,67 ))
- t
(2 -
C -
1,33 ) )
=
7,17 for :
→
Measuring Price elasticity
2,5 1,34 3,33 7,17
=
t
ideas for new
t
→
generali ng
Products
② 2 'S =
0,3486 134,86% ) Relative importance
→
determining relative importance
7,17 Airline of attributes in Consumer
choice process
STEPS
1,34 =
0,1869 ( 18,69% ) Relative importance
7,17 Price Formula le the problem
]
Conduct the Stimuli
{ Decide of the form of
3,33 Relative Importeure
0,4644 ( 46,44% ) input data
z
'
Man
-
7,17 fügt
Select a
procedure
[ Interpret the resutts
z
Acessreliabiüty .
and vaeidity
HAHA HHNHIINYK MIR 1114MDM
PART -
WORTH CALCULATION
Reference level :
O
Sweet 1 > 0 ,
he prefers Sweet
RB :
-3,0 & E :
-2,0 L O .
he prerers Booster -
Part worths
-
prices :
I Lzs -
das =
b2 = > 221=4 t das lasert 223
= -
2,35
I daz -
das =
b} Lzz :
3T das das =
4 -
2,35 =
1,65
2,35 =
0,65
I & II in III :
223
= -
2,35
4T Lz } t 3 t 223 t 223 =
0
1,663×23
Price 0,99 f
=
7+3223 =
0 , -
0,66×23
Price 1,49 E :
= -
7 / :3
Price 1,99€ = -
2,35
:
-2,35
called Clusters
geneovs groups Object in each Cluster tend
. to be similar to
DISTANCE MEASURE
Güven Resets :
-2
4
§
b
|
-
(
-
4-
f- 2 4 J2 3- Jb
City -
Block Matrix :
Example calculation
:
13-5 It I 2- 41 t 11 -
71=21-21-6
= 10
Euclidean Distance :
Example calculation :
( 2- 6ft ( 3- 7ft ( 3- 6) =
42
?
4
'
t t 3
=
41
( hebychev Distance :
Example calculation :
1- 5 2-4 1-7
4 2 6
✓
max .
= 6
HAHA HHNHHNYK MIR 1114MDM
DENDROGRAM based on
single linkage
WARDS PROCEDURE
Varian ( e method in which the Square d Euclidean distance to the cluster mean
is minimized .
The Word value tells ,
which
Object should be cwslered E- Iowest
} =
1,5
' I
(1-1,5)
( z-1.si WWDAß
=
11-1,512+(2-1,5)<+12-2,53
, ?
(3- 2. 5) 2+(2-2,5)
?
(0,25 4) tritt
. = t ( 3- 2,5 ) t
?
I t ( 5- 4. 5)
'
t ( 4-4,55+(6-6,5)
2¥ =
,
t ( t -
6,5)
? =
2,5
! Ward
B. ( =
( 2- 1,5ft (1- 1,5ft
' ' ?
( 3- 3) t
-
3) t (2-1,5) t
"
:
=
HNHKN.YHHHH.lt/IMKllYlkIlNnNlDN
Example
distance
%}
|
di wasitcwskj.ec#ouchedbefore? YYYI.gg?Ig;y ?
¥:{ an ?
stasi :
"
-
Person >
lobjectslbrands
Example
Step -
first
end e O
d
Cluster 1 =
[ 1. 5. 2,7 ]
Cluster 3
=
[ 3,6 , 4,10 ,
6. 8,9 ]