ISP Satisfaction

1
ISP Satisfaction

• Winnie Chung (MCom Thesis)
• Presented by John Paynter

2
Agenda

• Background Information
• Research Objectives and Questions
• Scope of the Study
• Research Methodology
• Results of Data Analysis
• Conclusion

3
Background Information

• High Internet Penetration
• More than 50 of New Zealanders have access to
  the Internet and 34 use it on a regular basis
• About 50 national ISPs and 38 regional ISPs
  (http//www.netguide.co.nz)
• Different services offered by ISPs
• Different Prices, Plans, Speeds and technologies
• More potential ISPs are trying to enter the
  industry
• Only 5 of broadband adoption in NZ

4
Research Objectives and Questions
5
Scope of the Study

• Focus on retail/residential/individual ISP users
• Household ISP registration
• Exclude business ISP users
• E.g. business solutions for accessing the
  Internet
• Exclude public Internet access facilities
• E.g. Internet cafes and libraries
• Focus on urban areas

6
Literature and Background Research

• Publications (26)
• Preliminary interviews (8)
• ?Essential Constituents for the Provision of
  Internet Access Service

7
(No Transcript)
8
(No Transcript)
9
(No Transcript)
10
Theoretical Framework
11
(No Transcript)
12
(No Transcript)
13
Hypotheses
3.2 Development of Theoretical Framework and
Hypotheses
3.2.1 Illustration of the ISP User
3.2.2 Theoretical Framework
3.2.3 Hypotheses

• Hypothesis 1O Users of large ISPs have the same
  Overall Satisfaction as users of small ISPs
• Hypothesis 1A Users of large ISPs have higher
  Overall Satisfaction than users of small ISPs
• Hypothesis 2O Narrowband users have the same
  Overall Satisfaction as broadband users
• Hypothesis 2A Broadband users have higher
  Overall Satisfaction than narrowband users
• Hypothesis 3O Free service users have the same
  Overall satisfaction as charged service users
• Hypothesis 3A Free service users have higher
  Overall satisfaction than charged service users
• Hypothesis 4O Narrowband factors are not related
  to the Overall Satisfaction of narrowband users
• Hypothesis 4A Narrowband factors are positively
  related to the Overall Satisfaction of narrowband
  users
• Hypothesis 4aO Narrowband Customer Support and
  Service does not differ between large and small
  ISPs
• Hypothesis 4aA Large ISPs provide better
  narrowband Customer Support and Service than
  small ISPs
• Hypothesis 5O Broadband factors are not related
  to the Overall Satisfaction of broadband users
• Hypothesis 5A Broadband factors are positively
  related the Overall Satisfaction of broadband
  users
• Hypothesis 5aO Broadband Customer Support and
  Service does not differ between large and small
  ISPs

14
Hypotheses

• Hypothesis 9O Broadband adoption barriers are
  not related to the likelihood of future broadband
  usage
• Hypothesis 9A Broadband adoption barriers are
  negatively related to the likelihood of future
  broadband usage
• Hypothesis 10O Broadband Usage is not related to
  Gender
• Hypothesis 10A Males are more likely to use
  broadband than females
• Hypothesis 11O Broadband Usage is not related to
  Income
• Hypothesis 11A High Income people are more
  likely to use broadband than low-income people
• Hypothesis 12O Broadband Usage is not related to
  Size of ISP
• Hypothesis 12A Broadband users are more likely
  to use broadband provided by large ISPs than
  small ISPs
• Hypothesis 13O Broadband Usage is not related to
  Education Level
• Hypothesis 13A Highly educated people are more
  likely to use broadband than lowly educated
  people
• Hypothesis 14O Broadband Usage is not related to
  Internet Usage Preference
• Hypothesis 14A Broadband Usage is related to
  Internet Usage Preference
• Hypothesis 15O Broadband Usage is not related to
  Age

15
3.3 Research Methodology
3.3.1 Research Design
3.3.2 Data Collection Method
3.3.3 Ethics Consideration in Data Collection
3.3.4 Data Analysis Methods
Table 3.3 Research Design Choices for the Study
16
Research Design (continue)
17

• Preliminary Interviews
• ISP users

3.3 Research Methodology
3.3.1 Research Design
3.3.2 Data Collection Method
3.3.3 Ethics Consideration in Data Collection
3.3.4 Data Analysis Methods
3.3.2.1 Questionnaires
3.3.2.2 Interviews
ISP

• ISP users
• Pilot study

18
(No Transcript)
19
(No Transcript)
20
(No Transcript)
21
(No Transcript)
22
The Breakdown of ISPs
Main ISPs
23
The Breakdown of ISPs
Secondary ISPs
24
The Breakdown of ISPs
Former ISPs
25
Length of Time Using the Main ISP, Overall
Satisfaction and Retention
Satisfaction and retention are significantly
correlated (positive) to each other at the 0.01
level (2 tailed).
26
Length of Time Using the Main ISP, Overall
Satisfaction and Retention
The calculation of the F statistics (2.72) from
the ANOVA exceeds the critical value of ?1.96,
indicating that the null hypothesis is rejected
at the 0.05 level. It is concluded that not all
group means are equal.
Tukey method (post hoc test) suggests that users
who have used their ISP between 1 - 2 years have
significantly higher satisfaction than those who
have used their ISP between 3 - 6 months.
27
ISP Characteristics and Overall Satisfaction
Speed of service
Service Charges
Size of ISP
28
Slingshot says the disadvantages of the big
companies I see that is they are very slow to
respond and in case of Xtra for instance and
Clear, they are very hard in cutting their price
because they got a large amount of revenue coming
in, so reducing their price by 10, they are
losing a lot of money.
Slingshot says Some really small ISPs have
couple hundreds of customers or may be a few
thousands of customers, they are really set up
with a tight budget and they could fall over any
day literally. Um. a good example is Voyager,
who fell over recently and there are other ISPs
are on the edge. Even Ihug is really close to
the edge financially.
Telecom says Small ISP cant afford things like
backup systems or redundant networks or
multi-tier support staff or escalations. For
example, our email systems cost 7 million dollars
which is for a small ISP they can never do.
Telecom says broadband customers are more
technically sophisticated and also want a lot
more value than they are prepared to pay for
typically. Broadband at the moment is not a
goodtheres not parameters for customers to
understand whats good value or whats not good
value. They dont have a lot of benchmarks to
compare so their expectation are higher than
dialup where people would have a higher
satisfaction with it.
29
ISP Characteristics and Overall Satisfaction
               
 
ISP Characteristics
Size of ISP
Overall Satisfaction
Speed of Service
Service Charges
30
ISP Performance Variables (5) for providing
Narrowband Connection
31
ISP Performance Variables (19) broadband and
narrowband
32
Narrowband Factors Factor Analysis
33
Narrowband Factors vs Satisfaction Stepwise
Regression

• Four variables (Connectivity, Speed, Email
  Service and Customer Support and Service) were
  added to the regression model. Web Site
  Functionality was removed, because the unique
  variance of satisfaction explained by this
  variable is not significant at the 0.05 level.
• This model explains 36 percent of the total
  variation of Overall Satisfaction.
• The regression model with the four variables is
  significant at the 0.05 level. This means that
  the null hypothesis can be rejected. Four
  narrowband factors are related to Overall
  Satisfaction.

?
Email Services
Connectivity
?
Speed
?
Customer Support and Service
?
Overall Satisfaction
Web site Functionalites
34
Connectivity
Slingshot says that modems (internal or external)
OS, noise on the line and whether one is running
some programs (those that use up a lot of memory)
behind are factors that may affect the connection
Telecom says And if people are having problem
maintaining a connection often there are software
settings on their PC, which we can help them
change to get them more reliable. For instance,
the OS settings, Dialup setting, TCP settings.
Telecom says Theres just diversity. Theres
hundred of different kind of modems, theres
hundred of different kind of PC and theres many
of different kind of operating systems. The
permutation or combination of that mix combined
with some variable line quality in some areas
mean that you can never guaranteeyou know.
Slingshot says We think that it is more
important to get things like the bandwidth, the
reliability the fact that we have a policy of not
allowing busy tones.
35
Price is the main thing customers focus on, but
we view that it is not the thing they actually
want. What we seen is growth of non-priced
attributes, quality of network, quality of
service, stability to connected and stay
connected. Price has to be competitive. Quality
of email service, outages, helpdesk services and
support, all the components. the whole range of
services.
Customer Support and Service
Slingshot says We know that our target market
dont use helpdesk much so we dont believe that
priority which we charge for helpdesk phone calls
but we do provide free email. It doesnt mean
that if customer has a problem, if it is a minor
problem then they will think twice about it
before calling us.
Slingshot says We do have technical support.
Its free emails and free fax. Um..for emails,
we normally respond within a couple of hours. So
its pretty fast turn around like for instance if
you use ClearNet and you send an email, they
normally within a day but we respond within a
couple of hours.
36
T-test of Narrowband Customer Support and Service
Against ISP Size

• T-test for equality of means is significant at
  the 0.05 level (the null hypothesis is rejected).
  
• This suggests that the narrowband Customer
  Support and Service differs from large ISPs to
  small ISPs.
• Small ISPs have a significantly lower narrowband
  Customer Support and Service level than large
  ISPs.

37
Broadband Factors Factor Analysis
38
Broadband Factors vs Satisfaction Stepwise
Regression
?
Customer Support and Service

• Two variables (Speed and Customer Support and
  Service) were added to the regression model.
• This model explains 39 percent of the total
  variation of Overall Satisfaction.
• Email Service, Value, Additional Service and Web
  Site Functionality were removed because the
  unique variance of Overall Satisfaction explained
  by each of these variables is not significant.
• The regression model with the two variables is
  significant at the 0.05 level (means that the
  null hypothesis can be rejected). Two broadband
  factors are related to Overall Satisfaction.

Email Services
Speed
?
Value
Additional Services
Overall Satisfaction
Web site functionalities
39
T-test of Broadband Customer Support and Service
Against ISP Size

• T-test for equality of means is not significant
  at the 0.05 level.
• The null hypothesis cannot be rejected.
• This suggests that the broadband Customer Support
  and Service do not differ from large ISPs to
  small ISPs.

40
Satisfaction of Different Broadband Technologies
Test of difference between means of DSL, Cable
and Satellite
It was found that using ANOVA to test for
difference is not a feasible in this case because
there are insufficient observations in some cells
(Satellite and Cable). The number of broadband
users is small and most broadband users use DSL.
For this reason, ANOVA was not conducted.
Although the means of Satellite, Cable and DSL
are different, we have no confidence that they
are significantly different.
41
Broadband technologies
Slingshot says The disadvantages of satellite is
that you get really long, very high latency which
mean basically the time takes for the signal to
go out and then come back in again is very long.
So even they say that they can provide broadband
for a large customer base it is a very slow
response service.. And I think for upload
trafficThey use a modemso really the only
broadband in the market place at the moment is
Telecom ADSL.
Telecom says Satellite is expensive and the
quality is no good
Slingshot says The cost of bandwidth provides
these broadband services (Cable) have been in the
past quite privative.
Telecom says Say you got ADSL, cable, Satellite,
you got an install cost, a quality and a
throughput cost or marginal cost. ADSL has a
very low install cost, cable you got to dig up
everyones drive way, thats an exceedingly high
cost, satellite you got to buy a dish, thats a
very high cost.
42
Broadband Adoption Barriers (21)
43
Broadband Adoption Barriers - Factor Analysis
44
Broadband Adoption Barriers Stepwise Regression
Cost
?
Limited Opportunity to change

• Only one variable (Cost) was added to the
  regression model. The other five variables were
  all removed because the unique variance of
  Likelihood of Future Broadband Usage explained by
  each of these five variables is not significant.
  
• The simple regression model with one variable is
  significant at the 0.05 level. This means that
  the null hypothesis can be rejected.
• This model only explains 11 percent of the total
  variation of the Likelihood of Future Broadband
  Usage.

Lack of broadband information and knowledge
Uncertainty about supports and services
Likelihood to adoption broadband in the near
future
Reluctance to change
Limited subscription plan available
45
Potential of Broadband in NZ
Slingshot says It will be at a point where the
majority of connections are broadband there will
still be a big chunk of dial-up connections
because of the very price sensitive people. But
it will take Telecom to severely change their
price structure for broadband before mainstream.

Slingshot says It is in my opinion anyway
Telecoms entire fault as to why broadband hasnt
really taken off to the extent that it should has
because nowadays bandwidth is not a cost factor,
well it is but it is not a significant cost
factor. The infrastructure out there in the
market place but unless Telecom changes their
price structure to make it at least break-even
for other competitors, umno one really wants to
take it up and really heavily promote it.
Slingshot says The price structure is just
really bad for JetStream. They just cost too
much like um. so thats why a lot of customers
are on JetStart because they perceive it is a
flat right.
Telecom says the southern cross cable, you know
how much it cost, 1.2 billion dollars.Thats
12,000 million dollars. how many people are
there in NZ? 4 millions Ok, 4 millions, how
much each do we have to pay?..Divided them up300
dollars. thats for everybody was on the
Internet but only 1 million people is on the
Internet. So the marginal cost is very high, the
fixed cost is very high.
46
Characteristics of Broadband Users
47
Characteristics of Broadband Users
48
(No Transcript)
49
ISP Viewpoint of Satisfaction Factors
50
Advantages and Disadvantages of Large and Small
ISPs
51
ISP appraisal of Broadband Internet in NZ
52
Conclusion

• Results indicate that user satisfaction directly
  reflects retention
• Narrowband and broadband users have slightly
  different satisfaction requirements.
• Connectivity, Speed, Email Service and Customer
  Support and Service (in order) are significantly
  important for narrowband users.
• Broadband users however perceived that Speed and
  Customer Support and Service (in order) are the
  two significant importance factors.
• ISPs should focus attention based on the relative
  importance of these factors.
• In addition to these factors, ISPs perceived that
  Value and Web Site Functionality should not be
  ignored.
• There was no evidence to suggest that Speed of
  Service (narrowband or broadband), Service
  Charges (free or charged) and Size of ISP (large
  or small) effect overall user satisfaction
• However, narrowband users perceived that large
  ISPs provide significantly better Customer
  Support and Service than small ISPs.

53
Conclusion

• ISPs should spend special effort on new
  subscribers (those that who stay with an ISP just
  between one to three months), as they are
  significantly less satisfied than those who stay
  with an ISP between one to two years. If ISPs
  can ensure the satisfaction of these users, they
  are less likely to switch.
• Broadband users are more likely to subscribe to
  broadband with large ISPs than those smaller
  ones. Male, Preference for Heavy Entertainment
  Traffic, high General Computer Skills and
  preference to use longer hours on the Internet
  per day, are the characteristics of potential
  broadband users.
• If ISPs want to increase their broadband customer
  base, they should concentrate targeting users who
  fit these characteristics.
• The broadband market in NZ is still in an infancy
  period with too little proper choice available
  for users. The penetration is low and Cost is
  the barrier that significantly prevents the
  adoption of broadband. The future of broadband
  in NZ looks good, as the infrastructure is ready.
  However, this will not takeoff unless if price
  is reduced, either initiated by ISPs, or by
  government subsidy.