Tài liệu consumersatisfactionservey-casestudy pdf

1




CONSUMER SATISFACTION SURVEY
(FEBRUARY-APRIL 2001)
SUMMARY REPORT



I. OBJECTIVE

• The objectives of this study are:

¾ To determine the level of customer satisfaction on overall performance
and customer service provided by the telecommunications,
broadcasting and Internet access services;
¾ To identify which areas of the service provided have the lowest and the
highest impact on customer satisfaction;
¾ To measure and evaluate the satisfaction level of customers at the
various national level.

II. SUMMARY OF FINDINGS

A. BROADCASTING - TELEVISION
• There are 6 TV stations in Malaysia. With exceptions of Mega TV (lowest
penetration), all 5 TV stations have sufficient sample size for analysis. They
are:
Free-to-air TV

¾ TV 1, TV 2, TV3 and NTV7
Subscription TV

¾ ASTRO
2
• The following table outlines the key findings:

BROADCASTING –
TV STATIONS

CUSTOMER
SATISFACTION
INDEX (CSI) –
( 1- 10 SCALE)

HIGH PRIORITY
AREAS TO
IMPROVE
(IMPORTANCE
AND CSI)


MOST OFTEN
WATCHED TV
STATIONS BY
ADULTS IN
MALAYSIAN
HOUSEHOLDS
%

FREE-TO-AIR TV

TV 1 7.22

• Transmission
Quality

• Content quality

31
TV 2 6.91

• Content quality

33

TV 3

7.44 None 83

NTV 7

7.20 None 42
SUBSCRIPTION TV


ASTRO

7.42 None 9

FREE-TO-AIR TV STATIONS
• On national level, TV 3 is the most often watched free-to-air channel in
Malaysia; followed by NTV 7, TV 2, and TV 1. All TV stations except TV 2,
have CSI score above 7. TV 2’s CSI score is lower, at 6.91.


SUBSCRIPTION TV STATION
• Astro's CSI score is 7.42.



3
B. BROADCASTING - RADIO


• The survey covered more than 35 radio stations across Malaysia. 8 stations
with highest penetration (sample size of more than 50 for each station) are
selected for detailed analysis as follows:

BROADCASTING -
RADIO STATIONS

CUSTOMER
SATISFACTION
INDEX (CSI) –
( 1- 10 SCALE)

HIGH PRIORITY
AREAS TO
IMPROVE

(IMPORTANCE
AND CSI)


MOST OFTEN
LISTENED RADIO
STATION BY
ADULTS IN
MALAYSIAN
HOUSEHOLDS
%


ERA

7.86 None 33

RADIO 3


7.54 None 20

RADIO MUSIC

7.80
• Content
Quality
16

RADIO 1

(Bahasa Malaysia)

7.56 None 13

MY FM

7.40 None 17

HITZ FM

7.39 None 10

REDIFUSSION

7.21 None 8

RADIO 5
(Mandarin)

7.38 None 9

• ERA is the most often listened radio station in Malaysia (penetration of 33%);
followed by Radio 3, MY FM, Radio Music, Radio 1 (Bahasa Malaysia), and
4
Hitz FM (all between 10 – 20% penetration). Redifussion and Radio 5
(Mandarin) have less than 10% penetration.
• Overall, the CSI scores for these 8 key stations are above 7.

C. TELECOMMUNICATIONS - FIXED LINE

• The survey shows that fixed line telephone penetration in Malaysia is 69%.
• Of those with fixed line telephone at home, Telekom Malaysia, as a fixed line
provider, has close to 100% penetration level.
• The penetration of TT dotCom, Celcom, Maxis and DIGI is negligible.


FIXED LINE PROVIDER

CUSTOMER
SATISFACTION
INDEX (CSI) -
(1- 10 SCALE)

HIGH PRIORITY
AREAS TO
IMPROVE
(IMPORTANCE
AND CSI)


FIXED LINE
PROVIDER
PENETRATION IN
MALAYSIAN
HOUSEHOLDS
%

TELEKOM MALAYSIA

7.41

• Application
process

• Installation
process

• Repair quickly


100

• Telekom Malaysia has a CSI score of 7.41.

D. PUBLIC PAYHONES
• 98% of the respondents interviewed mentioned that a key concern is the poor
functioning of the payphones.

E. TELECOMMUNCIATIONS - MOBILE


• In total, there are 7 service providers covered in this survey.
5
• Time Cel, Celcom (010), and Mobikom have the lowest penetration with small
sample size achieved. Hence, detailed analysis focus on the four major
service providers which accounted for 96% of the penetration:



MOBILE SERVICE

PROVIDER

CUSTOMER
SATISFACTION
INDEX (CSI) -
(1- 10 SCALE)

HIGH PRIORITY
AREAS TO IMPROVE
(IMPORTANCE
AND CSI)


MOBILE
SERVICE
PROVIDER
PENETRATION
AMONGST
ADULTS IN
MALAYSIAN
HOUSEHOLDS
%


MAXIS

7.39
• Coverage
36
TM TOUCH 7.13

• Coverage

• Capacity

• Network Quality

16
DIGI 6.96

• Coverage

• Network Quality

• Capacity

18
CELCOM 019 6.75

• Coverage

• Capacity

• Network Quality

26


• Maxis (012) has the highest penetration level, followed by Celcom (019), DIGI
(016), and TM Touch (013).

6
F. INTERNET

• TM Net has the highest penetration (82%) followed by Jaring (24%).
• Maxis Net, Time Net and Celcom Net are still very small; their combined
penetration is less than 5%.
• The following outlines key findings for TM Net and Jaring:

INTERNET
SERVICE
PROVIDER

CUSTOMER
SATISFACTION
INDEX (CSI) –
(1- 10 SCALE)

HIGH PRIORITY
AREAS TO
IMPROVE
(IMPORTANCE
AND CSI)

INTERNET
SERVICE
PROVIDER
PENETRATION
AMONGST
ADULTS IN

MALAYSIAN
HOUSEHOLDS
%


JARING

7.4
• Pricing
24
TM NET 6.62

• Ease to log in

• Customer
Service

• Reliable line
connection

82


III. RESEARCH METHODOLOGY

In brief, the fieldwork was conducted from 1
st
of February 2001 to 15
th
of April

2001. Total sample of 2001 was achieved where a face-to-face random interview
was conducted throughout the country. The base of the survey is a random
sampling, a method used to obtain results that are representative to the
population of the country (18 years old and above). The interview length was 30
to 40 minutes.

7
Below is the detailed explanation of this Appendix:
1. Coverage
2. Time Reference
3. Sampling
4. Data Collection
5. Response
6. Quality Control

1. COVERAGE


Geographical Coverage

Geographically, the study covered Peninsular Malaysia, Sabah and Sarawak.

Demographic Coverage

Considering population coverage, only those aged 18 years old and above were
covered. It was felt that it is only those in this age group who would be able to
understand and provide reliable answers to questions on performance of the
various services identified. Thus, the survey would reflect 65% of Malaysia’s
population who are 18 years old and above.

2. TIME REFERENCE

The fieldwork for the survey was conducted between February and Mid-April
2001. Thus the data in this report refer to this period.

3. SAMPLING

The survey was a household sample survey. A multi-stage probability sample
design was adopted.

8
The lowest area unit with identifiable boundaries is the Electoral District (ED).
The EDs, which cover all areas of the country comprehensively, are the basis of
the sampling frame. EDs have been created within States and Federal Territory
Kuala Lumpur. For the States, selected EDs were designated as “urban” based
on their location in the 129 towns which have been classified as urban in 1991
Population Census.
1
The definition of urban is as follows:

“Gazetted areas with their adjoining built-up areas which had a combined
population of 10,000 or more at the time of the 1991 Population census.”
2


The remaining EDs are treated as “rural”.

Since the EDs are of unequal size, they can be selected either with probability
proportional to size or subdivided into units of equal size and selected with equal

probability. The latter approach was adopted. Each ED was artificially divided in
blocks with a measure of size (MOS) of 100 households using the electoral
population as a guide. The blocks formed the primary sampling units for the
study.

Stratification of primary sampling units was according to state and urban/rural
areas. The allocation of the sample among the strata was primarily proportional
to the population, but also took in account the expected lower variances in the
rural areas/less developed states and related costs and difficulties of transport
and travel.

A total of 200 blocks were selected for the survey. The distribution according to
strata is shown in Table 1 below.

9




Table 1: Distribution of Sample Blocks and Persons

State Blocks Persons Targeted
Total Urban Rural Total Urban Rural
Kedah 14 6 8 140 60 80
Perlis 2 1 1 20 10 10
Penang 13 11 2 130 110 20
Perak 21 14 7 210 140 70
NORTH 50 32 18 500 320 180
Selangor 31 27 4 310 270 40
K. Lumpur 16 16 - 160 160 -

N. Sembilan 8 4 4 80 40 40
CENTRAL 55 47 8 550 470 80
Malacca 6 3 3 60 30 30
Johor 25 14 11 250 140 110
SOUTH 31 17 14 310 170 140
Pahang 12 5 7 120 50 70
Terengganu 9 4 5 90 40 50
Kelantan 13 6 7 130 60 70
EAST 34 15 19 340 150 190
PENINSULAR
MALAYSIA
170 111 59 1700 1110 590
Sabah &
Labuan
18 10 8 180 100 80
Sarawak 12 7 5 120 70 50
SABAH &
SARAWAK
30 17 13 300 170 130
MALAYSIA 200 128 72 2000 1280 720

The penultimate sampling unit was the household defined as “…a group of one
or more persons living together who make common provisions for food or other
essentials of living.”
3
For each sample block households were listed and 10
households were selected based on a systematic sampling. Thus, the sample
consisted of 2,000 households. Some of the results are presented at the
household level.

The ultimate or elementary sampling units were persons aged 18 years old and
above. One person was selected per sample household using the Kish selection
10
grid. The distribution of the 2,000 persons targeted is given in Table 1 above.
Most of the results in this report are presented for this sample. The results are
presented in proportions and averages (for CSI only). Variances and standard
errors were estimated for the CSI and are presented in Table 2 below.

Due to the fact that a sample size of 2001 was used to study five service
categories, the sample size for many of the services that had lower incidence
levels ended up being very small. A decision was made to present data only for
services which had a sample size of 50 or more.
4
Thus

the results are presented
only for 31 services in the five product categories. Also the presentation of the
data by urban/rural areas followed the requirement of a minimum sample size of
50.

The relatively small sample size for many of the services resulted in many
instances of single cases in the sample blocks. While each case is valuable for
the overall result, this fact posed problems of calculating within-cluster variances.
It was thus decided to treat the sample as only stratified sampling. Each of the 31
services was treated as a domain of study. Mean CSI scores and associated
variances were estimated within domains using standard formulae.
5
Strata
weights were based on the 1999 mid-year population estimates of the
Department of Statistics since the required data from the 2000 Population

Census was not yet available.
6


4. DATA COLLECTION


The data was collected by face-to-face interviews, which lasted for 30 – 40
minutes using a structured questionnaire.




11
5. RESPONSE

Non-response is an endemic problem in most surveys. Overall there was a 45%
non-response, 48% in the urban areas and 38% in the rural areas.

To overcome this problem, random substitutes were done to achieve the sample
size of 10 households per block and consequently a random sample of 2001
households.

6. QUALITY CONTROL

The following procedures under the Quality Control System for Interviewing
(QCSI) were adopted for the survey:

• All interviewers were briefed by researchers in charge of the survey
• No interviewer was allowed to do more than 20% of the sample in order to

prevent bias
• 20% of each interviewer’s work was validated randomly by Supervisory staff
and the Quality Assurance Unit

All questionnaires were check-edited and coded before data capture.

12

1
Department of Statistics, Malaysia. (1992). Preliminary Count Report for the
Urban and Rural Areas. Kuala Lumpur. (p.58)
2
Ibid. (p. 12)
3
. Department of Statistics, Malaysia. (2000). Preliminary Count Summary. Kuala
Lumpur. (The definition here is the general one adopted; see p. 14.)
4
Vavra, T. G. (1997). Improving Your Measurement of Customer Satisfaction.
Wheeler Publishing, New Delhi. (As suggested in p. 79)
5
Cochran, W. G. (1977). Sampling Techniques. Wiley. New York. (pp. 142 –
144)
6
Department of Statistics, Malaysia. (2000). Op. cit .