E-commerce and Internet Users in Taiwan:
Research Series, 1996-2000
Wu, Tung-Xiung
Dept. of Information Management, Shih Hsin University
#1, Lane 17, Sec. 1, Mu-Cha Rd., Taipei 116, Taiwan, ROC
TEL/FAX: 886-2-22368225 ext.376
Email: txwu@cc.shu.edu.tw
ABSTRACT
This empirical research series investigated:
1. Exploring internet users profile and their behaviors as world cybercitizens. Forecasting the internet activities in Taiwan and its implication for globalization.
2. Analyzing the growth and tend of E-commerce. Proposing E-commerce strategies on market planning and public policies for private or governmental sectors.
3. Initiating suggestions for methodological issues on internet users survey.
There was 50.6% who had accessed internet after a significant change in last couple years. The qualified user who had email accounts was 23.1%. According to the S-type diffusion of innovation model, internet is passing the taking off point of the runway.
Though internet is somehow still belongs to the elite, it is not too far away to reach the oasis behind the internet mirage.
The features of internet users in Taiwan looked very similar to users in USA. The world 『cybercitizen』 is quite a real thing more than a concept.
The general profile of an internet user is young people under 30 with higher education, student or who has above average income. Their first purpose for using internet is communication.
E-commerce is still relative small and will wait for the next stage. Four strategies are under investigation. There are: market scale and trend, sales volume and trend, popular and potential products, and barricades for E-commerce.
This research team is the first and the only one to conduct internet users research series on random sampling base in Taiwan, 1996-2000.
Keywords: Internet, E-commerce, User Behavior, Diffusion of Innovation, Taiwan
SPECIFIC OBJECTIVE
Internet is a global bridge of being easy to access, friendly to interact, and free to across international boundary. It is becoming a part of human life and forms a new people of world 『cybercitizen』. It brings a key issue on marketing strategy for E-commerce, cross-cultural business and multinational cooperation. The users' behavior is the fundamental for marketing planning and decisions, it is obviously an emergent need for user studies as well.
However, there is little related research on internet users in Taiwan while Taiwan is playing such a distinguished role on information industry and international business.
This empirical research series attempted to response the scarcity to investigate the following:
1. Exploring internet users profile and their behaviors as world cybercitizens. Forecasting the internet activities in Taiwan and its implication for globalization.
2. Analyzing the growth and tend of E-commerce. Proposing E-commerce strategies on market planning and public policies for private or governmental sectors.
3. Initiating suggestions for methodological issues on internet users survey.
There are rich empirical findings on E-commerce and user behaviors in Northern America. The findings were also compared with the relevant results in USA to provide cross-cultural perspectives.
Furthermore, the researcher found there were various results in different studies. The statistics were even not convergent in different release versions of same survey conductor. The different survey methods and interpretation might produce the conflict.
Four important methodological issues were selected and discussed in order to initiate some integrated points of view. The issues were:
1. Measuring who is internet user.
2. Weighting procedure of estimating uses numbers.
3. Predicting procedure of future growth.
4. Measuring sampling precision rates.
This research is the first and single series survey on internet users in Taiwan.
LITERATURE REVIEW
Taiwan Experiences
The first internet users report, based on empirical methods in Taiwan, was the first wave of this research series that collected user data at Taipei metropolitan, 1996. Then the survey has been conducted annually, and the sample frame has extended to Taiwan regional wide.
Meanwhile, there were some discussions with indirect methods. The most popular way was to estimate by personal perspective. For instance, Liou[12] estimated the user numbers by intuition and observation of maintaining his internet server. This method would also cause a very large estimation range. The Education Ministry [6] provided an analysis of communication quantity on TaNet that is the dedicated network for academic or research institutes. The Economic Ministry examined the numbers of host domain names and counted the subscribers that were announced by ISPs. But those methods were very difficult to identify the individual users』 favor service. It was also impossible to precisely estimate the user behavior and their contribution to E-commerce.
Some ran a non-random sampling survey by putting a questionnaire on internet. An example was Jang[9]』s work. The problem was that respondents were not randomly selected. It could not draw a whole picture including both users and non-users. The method could not call an active participation, either. The Commonwealth Magazine[4] conducted another random sampling survey after 1997. However, they kept an aggressive point view to encourage E-commerce participation that was different from a scientific purpose.
Selected USA Studies
There were very useful empirical reports in USA and executed different methodological procedures. Two groups were generic. CommerceNet/Nielsen was base on general user definition and aggressive point of view. Cyber Dialogue and Find/SVP took qualified user definition and careful attitude to interpret data.
CommerceNet/Nielsen
CommerceNet/Nielsen's Internet Demographic Survey[3] was based on a probability sample of persons 16 years or older among telephone households in the US and Canada. They made re-interview skill to compare to each other study. The sample was selected from an unrestricted random digit frame of phone numbers from exchanges operating in the US and Canada. The frame was stratified by geography and approximately 280,000 phone numbers were selected. Up to 8 attempts was made to reach a household and a respondent was randomly selected. Variables were employed a 3-way weighting procedure (Age / Gender / Region) or a 4-way weighting procedure (Age / Gender / Region / Education).
They classified current internet users into long-time internet users and newcomers. Current Internet users are those respondents who have used the internet in the 6 months prior to the interview, and their classification as long-time users or newcomers are a function of whether or not they had used the internet in the 3 months prior to be interviewed.
Cyber Dialogue and Find/SVP
Find/SVP fielded American Internet User Survey[7] by telephone with randomly dialed method in 1995. They took 'qualified (heavy) user' definition and suggested that the real number of uses were more than who responded they were users.
They formulated the following calculations:
Adult users: respondent household users = 1.43: 1
Total users (including children): respondent household users = 1.61: 1
Cyber Dialogue[5] carried on American Internet User Survey after they purchased Find/SVP. Using random digit dialed survey of internet users and non-users, they fielded in the second and fourth quarter of every year. They conducted approximately 13,000 screener interviews to identify 1,000 current online users and 1,000 non-users. Their data was weighted for response bias by age, gender, income and geographic region.
In 1999, CommerceNet/Nielsen claimed the bestseller product on internet was cars. While Cyber dialogue reported that the transaction of buying a car on internet was relatively small. The real story was that there were the most people to look for information about cars on internet. But it was limited that people to purchase a new car on internet without touching it before.
These contrary evidences enlightened that we need more discreet considerations on research methodology and interpretation to conduct such surveys.
Others
Network Wizard provided Internet domain Survey to detect the numbers of valid IP addresses and host domain names. However, the recent DNS did not allow these kind of autodetect programs to get in. The variance of estimation could be large.
MIDS (Matrix Information and Directory Services) [16] produced a secondary analysis based on Network Wizard's primary data. Early on, their Internet Demographic Survey was a representative of early indirect method by surveying organizations connected to the internet.
Media Matrix ran a panel study to monitor and record users』 browsing history. It is a reference for advertising industry as an alternative rating report on internet audience.
Lycos』s Web User Study[13] was an example that their sampling frame was focused on active internet users only. Respondents were randomly selected from the Cyber Dialogue panel of users. Answers were weighted to represent national ratios for gender, age, households with children, the year they first went online, and frequency of internet usage.
Other discrete researches abstracts could be found at Cyberatlas.
The findings in USA were summarized in Table 1.
The above studies stated that different survey methods produced various results and points. They also provided grounded perspectives, and reference maps to conduct user survey in Taiwan.
The following design of study would adopt the former studies』 heuristic experiences and attempt to denote a more precise research paradigm to include more comparative respondents' opinions.
Table 1. Selected surveys summary
Org.
Year
Area
Users
Core findings & E-commerce
C.N.
1999
US &
Canada
41%
Launching E-commerce age
Female buyer significant increased
Bestseller was cars
C.N.
1998
US &
Canada
35%
First time, young group over half was user
User in aged group was also growing
8.9% buying on-line while 21.3% checking
C.N.
1997
US &
Canada
23%
Internet commerce growing of more than 50%
15% buying on-line while 73% checking
C.N.
1996
US &
Canada
24%
Non-computer professionals increasing
13% used WWW
C.N.
1995
US &
Canada
16%
New comer increasing
8% used WWW
C.D.
1999
US
30%
Most asked product: cars
Most purchased products: computers, consuming electronics, flowers, and foods.
Cars sales was under 5%
C.D.
1998
US
21%
Users emphasized WWW, Daily info and Productivity info
Users attrition
Find/
SVP
1995
US
6.4%
Personal activities
Business
1.1 million children under 18 users
Org.
Area
Other survey methods
Network Wizard
Global
Domain name survey
MIDS
Global
Secondary analysis on data from Network Wizard
Media Matrix
US & Europe
Ratings on internet audiences
Lycos
US
Survey on users only
METHODOLOGY
Table 1 would reveal different studies』 contradiction and their coherence as well. The inconsistency might be derived from their research methods. Four methodological issues could be deduced and discussed as the following.
Measuring who is internet user
Who is qualified as an internet user? This definition will affect the method and result of study.
A general user's definition had never been formally defined. It would be considered as who ever had the experience to access internet, while many studies used such similar statement in their questionnaires.
Find/SVP[7] defined a rigid concept of two requirements of qualified internet users or heavy users:
1. Who had email account.
2. Who had to be current users of at least one internet application.
Though most of other studies did not exactly check these definitions, it indicated two conceptual ideas of users whom we are going to think about.
It also explained why CommerceNet had higher percentage of users, because they implicitly took the 'general user' definition.
Generally, people would not apply an internet account unless they need a certain application. It seemed that Find/SVP』s 2 rules could be combined as 1.
Therefore, the researcher suggested the definition of 『qualified user』 was who had his/her internet account (usually it was the same as email account).
A general user was who had a chance to have a glance or borrowed an internet account to have a poke occasionally. It is obviously that their use behavior does not have same value between the two users groups.
Weighting procedure of estimating uses numbers
All former studies interviewed only one respondent who answered the phone in the sampled family. It was highly possible that there were more than 1 user in a family. In the other hand, if the respondent answered that he/she was not using internet, the non-users record would be added 1. On the contrary, the truth might be there were other users in the family.
Most of studies reported that they employed 2 possible ways to compensate such flaws. They would use weighting skill by demographic or other independent variables. They always made a further step to formulate their own ratio to estimate user number. The basis of formulation was heavily upon researchers』 experience and intuition. It would reveal why there were such large gaps between different organizations』 survey. Even more, we could find there were not totally convergent in different versions of same organization.
The researcher suggested a 3-step procedure in this research series:
1. In the opening section of interview: Asking the respondent if there was anyone who was using internet in his/her family. If there was, the interviewer must ask the user. If the user was not home, the interviewer must make an appointment and use callback skill till locate the user.
2. In the closing section of interview: Asking the respondent how many other households were internet users in the family and how many of them obtained their internet accounts.
3. In the analysis: Producing 2 tables from the collected data of step 1 and 2, formulating a quantitative ratio of family users and respondent users.
This procedure would provide a more precise estimation based on clear data.
If it was inevitable that human』s experience and intuition have to be involved in estimation. Cyber Dialogue[5] reminded a valuable methodological philosophy that it was better to keep a conservative attitude. They criticized that there were a lots inflated statistics of internet activities that were from too positive point of view.
Predicting procedure of future growth
There were also two ways that most of former studies predict internet in the future. They used historic data and their experience and intuition. The basis of prediction was mostly linear.
Although it was a quite simple approach to forecast the future, it was considerably doubtful that human would follow a linear behavioral pattern.
The researcher suggested that a classic theory of the diffusion of innovation or the S-type adoption model would draw a better picture.
This fruitful idea could be traced back to Ryan et al.』s[20] early original work in 1940』s, Katz』s[10] and Bohlen』s[1] deduced frameworks in 1960』s. Rogers et al.[18][19] defined the final concept as 『innovation model』 in 1970s. Extended applications in information technology field could be including: Gibson[8] and Nolan』s[17] four stages of EDP growth in 1970』s and Nolan et al』s IT adoption theory in 1990』s.
The basic assumption of this theory was that users of an innovation would be passing through a 4-stage S-type curve from awareness, trail, evaluation to adoption.
The researcher suggested that the S-type curve could be explained as the cumulative curve of the normal distribution for the time frame of new users. Then the certain period of each stage of innovation could be computed.
Thus, the researcher suggested a 3-step procedure for predicting the future innovation:
1. Testifying and adjusting the historic data how it fitted to the awareness stage of S-type.
2. Finding out and determining where is the critical point from awareness to adoption.
3. Predicting the time period of next stage by the last stage.
This forecast might be matching human』s behavior closer.
Measuring sampling precision rates
Sampling precision was directly related to the completion of interviewing.
Nevertheless, there was not a solid and unified definition of the rate of completion yet. A common equation was early suggested by Kuiz[11]:
Response rate = response samples/ (drawn samples- unconnected samples)
= response samples/ (response samples + refusal samples)
It was a somewhat confusing and incomplete definition.
First, ruling out unconnected samples from the denominator of designed samples would cause significant distortion. Heavy research evidences showed that people whom were difficult to be located always-retained different characteristics from whom were easy to be found.
Second, defining refusal as the incomplete part was a bias of interpretation. Refusal meant the selected respondent did pick up the phone, then hung up after understanding the survey purpose and gave no clear answers. They, eventually, responded the call and expressed their attitude for the survey: negative or non-opinion. They were meaningful in the analysis and could not excluded from the designed samples.
Third, it did not consider the 『valid』 samples and replacement of samples. An easy to be forgotten principle was: 『Do not replace drawn samples, even they were difficult to connect to.』 (Many current surveys violated this rule.) Only was one exception under the method of random digits sampling. This method would unavoidably draw a certain portion of non-working phone numbers that were not 『valid』 within the designed sampling frame. The invalid samples may or may not be replaced according to the research resource. Valid sample did not affect sampling precision, but it determined the real sample size and the denominator of sampling precision rate.
Therefore, the researcher suggested 2 rates to evaluate sampling:
1. Valid rate = samples with working phone numbers/ designed samples that drawn from sample database (+ samples of replacement)
2. Access rate = connected samples (someone picks up the phone)/ valid samples
Access rate was very useful to project the sampling precision while valid rate could also reflect the cost-efficiency of research design.
This research series based on above methodological improvements made the following research designs, as Table 2, from 1996-2000.
Sampling
Table 2. Researches time and sampling designs
Year
Sampling frame
Designed sampling size
Research period
2000
Taiwan region
3200
Apr. 25-May 31
1999
Taiwan region
1600
May 5-25
1998
Taiwan region
2000
1997 Dec. 10-1998 Jan. 10
1997
Taiwan region
3000
Apr. 25- May 10
1996
Taipei Metropolitan
1500
May 5-25
Taiwan region included Taiwan Province and Taipei Metropolitan.
Sampling method: Systematic random sampling by telephone directories.
Household selection: Random selection procedure employed.
Expected errors: desired confidence limits and confidence interval based on the research budget and the estimation of Wu[25].
Special design: No replacement of valid samples.
Measurement
The main questionnaire structures of surveys were divided into 3 parts to measure different problems.
1. To identify if the respondent is an internet user.
2. Users: To measure their understanding about internet and information application; to ask their using place, time, and their behavior on E-commerce including their favor resources, services, procuct categories, their purchase history and comments, and their email address.
3. Non-users: To measure if they know anything about internet and information application; to measure if they have related equipment and potential motivation to become users and why; to ask their reasons why they are not users yet.
Some question items and item values were modified every year with concerns of information technology developing and market changing.
Demographic data will be collected at the end of interview.
Interviewing
Method: Telephone interviewing.
Verification procedure: 5% complete interviews were repeatedly interviewed to verify and measure the reliability of interviewer.
Using continuous call back skill to access selected respondents who are always not home. (Wu[25])
SELECTED FINDINGS
Sampling Analysis
The description of 3 surveys』 sampling representative was as Table 3:
Table 3. Sample representative
Designed sample size
Replacement
Valid samples
Valid rate
Access samples
Access rate
2000
3200
800
3225
80.6%
2896
89.8%
1999
1600
556
1600
74.2%
1434
89.6%
1998
2000
224
2000
89.9%
1668
83.4%
1997
3000
0
2766
92.2%
2261
81.7%
1996
1500
0
1202
80.1%
905
60.3%
Most of access rates were above average except survey 1996.
Two more procedures were employed to testify sampling precision. The first was to compare the gender distribution of sample and census data. The second was to compare the using time distribution of users』 activities reported by respondents and the record of server.
The gender distribution differences were all within and/or lower than expected errors of sampling design.
On the time distribution of internet using, this study also found that was matching the report by the Education Ministry.[6]
The sampling precisions were all acceptable according to the above analyses.
Total and Trend Analysis
The historical structure of users, and non-users were summarized as well as their future trends were predicted as follows.
Table 4. Users and non-users (%)
Users
Qualified Users
Non-Users
Family
Population
Family
Population
Family
Population
2000 Taiwan
36.6
50.6
19.1
23.1
63.4
49.4
1999 Taiwan
24.3
30.6
14.5
16.0
75.7
69.4
1998 Taiwan
24.6
31.4
9.4
12.0
75.4
68.6
1997 Taiwan
13.5
17.2
7.9
9.7
86.5
82.8
2000 Taipei
42.2
61.7
24.8
31.8
57.8
38.3
1999 Taipei
25.2
31.8
15.5
17.2
74.8
68.2
1998 Taipei
36.0
45.9
10.1
12.9
64.0
54.1
1997 Taipei
19.3
24.6
7.0
8.9
80.7
75.4
1996 Taipei
17.2
21.9
9.9
11.8
82.8
78.1
Table 5. Household users(%) and average numbers per home
Use internet
Have email account
\No
1
2
3
4
5+
Ave.
1
2
3
4
5+
Ave.
2000
80.1
8.6
6.2
3.1
2.0
1.38
88
6.9
3.5
1.4
0.7
1.21
1999
85.6
6.8
4.5
2.1
1.0
1.26
92
5.2
2.0
0.3
0.1
1.10
1998
86.0
6.1
4.6
2.4
0.9
1.26
97
2.1
0.7
0.3
0.1
1.05
1997
75.7
13.5
6.8
3.7
0.3
1.39
85
9.5
4.6
0.6
0.3
1.22
Table 6. Year began to use internet of qualified users(% of qualified users)
Year
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2000
0.5
0.5
1.8
1.9
5.2
10.7
14.9
18.0
31.1
14.2
1999
1.5
0.3
1.7
2.3
4.4
13.4
26.5
32.8
17.2
1998
0.8
1.1
1.8
6.6
13.9
31.9
37.7
6.3
1997
1.4
1.8
2.7
4.1
15.8
42.5
31.7
Table 7. Cumulative percentage of qualified users(% of total population)
Year
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2000
0.1
0.2
0.6
1.1
2.3
4.8
8.2
12.4
19.5
22.8
1999
0.2
0.3
0.6
0.9
1.6
3.8
8.0
13.2
16.0
1998
0.1
0.2
0.4
1.0
2.4
5.5
9.2
9.9
1997
0.1
0.3
0.6
1.0
2.5
6.6
9.6
It was a fast growth in term of general user. In 1996, there were 13.5% respondents in Taipei metropolitan who had at least one chance to use internet. The estimation of Taiwan regional wide could be around 10%. Since the sampling unit was family, the researcher made further questions about if other households in the family also used internet as Table 5 to weight the average user numbers per home. Calculating the result of Table 5, it could estimate the users percentage of population to be 18.8% in 1997; and it made a halfway milestone (50.6%) in 2000.
However, the growth rate of qualified user, who owned an email account, was relatively a little bit slower.
The year of users began to use internet was summarized in Table 6. The data combined with Table 4 would produce the cumulative percentage of users as Table 7. The figures matched the S-type innovation theory. It also implied that internet had passed the 2nd step of diffusion of innovation.
According to the momentum of S-type model』s changing, the development of internet is standing on a critical increasing turning point. It took 5 to 6 years to finish the first stage of innovation diffusion; we could predict that it needs another 5 years to accomplish evaluation and to reach popular adoption.
Users Behavior
Internet users』 profile and their changing could be sketched as Table 8:
Table 8 Internet users』 profile (%)
Gender
Age
Education
Income
Male
Fem.
~15
16-30
31-50
51~
-Mid
High
Aca-
low
Ave.
High
2000
49.5
50.5
9.5
44.4
36.6
9.6
21.7
33.7
44.7
45.4
40.4
14.3
1999
50.9
49.1
8.7
58.7
30.7
2.0
10.3
27.9
61.7
99.6
0.4
0
1998
51.6
48.4
2.9
52.7
40.7
3.7
7.0
35.5
57.5
37.7
36.8
25.5
1997
60.2
39.5
~
54.8
28.0
17.2
12.4
25.8
63.8
56.6
28.6
21.6
1996
68.1
31.9
~
55.7
25.6
11.6
2.3
6.6
91.1
38.3
21.3
40.4
Male was more than female of internet users before 1997. After 1998 female users were growing and got nature balance.
More than half of users were younger than 30. Mid aged user group increased while senior user group stayed small.
Use of internet was correlated with higher education. Most users were educated in academy or college or above. Number of user group under middle school was the least. Users with high school diploma were moving up.
The annual income of users could be categorized into low (under US$7,000), average (US$7,000-17,500) and high (above US$17,500). The result needed a further explanation why the percentage of the category of 『low income』 was so high. The reason was that there were many student users without stable revenue. Excluding students, user』s income was above average.
Though the job fields of users were various, their ranks seemed higher. The most popular job function was professional and technical, then administrative and managerial.
No significant difference of user behavior was found between users』 gender, age and education.
Table 9. Places using internet (%)
Place
Home
School
Office
Café etc.
Other
2000
65.4
13.0
18.7
2.5
0.4
1999
65.9
11.9
20.4
0.6
1.2
1997
52.0
24.4
16.3
2.3
5.0
1996
36.0
26.0
32.0
-
6.0
Users would spend average 3 days more in a week to surf on internet.
Home, Office, School showed their changing importance of using internet by their order. The percentage of school was somehow stable as expected. It also implied internet』s growing popularity in family nowadays.
Table 10. Most important resource or service (%)
WWW
FTP
Telnet
BBS & Chat
Search & Portal
News
Others
2000
28.0
2.7
0.7
26.2
8.7
30.6
2.8
0.3
1999
36.1
1.6
10.5
20.0
6.5
21.8
1.3
1.1
1997
29.0
2.7
5.4
21.4
17.9
18.3
1.8
3.6
1996
25.9
14.1
-
23.5
21.2
7.1
7.1
1.2
Table 11. Most important information (%)
Business
Life
Sciences
Entertainment
Social
Culture & Arts
Government
Education
C&C
Medical care
2000
12.7
22.5
3.9
29.8
1.5
4.3
3.4
8.7
10.9
2.3
1999
13.1
19.1
2.1
30.5
1.5
3.6
5.8
8.0
10.8
1.8
1997
10.2
20.9
2.8
27.9
1.4
3.3
2.3
7.4
12.1
0.9
1996
17.4
19.7
2.33
26.7
4.7
4.6
2.2
5.8
15.2
1.1
WWW with their advantages of multimedia performances became users』 favor resource. Since the fast growth of websites, the importance of search services and portal sites were getting hotter. Like the significance in United States, communication including email, BBS and other chat services took major activities of users. The chat function of BBS was widely adopted than other popular chat application. The reason is the common language in BBS is Chinese while other chats, such as IRC, are English.
『Soft information』 was the basic demand of users. Browsing entertainment and life information motivated near half of users』 accessing in different time frames.
Business activity was still relatively slow move on internet.
The dominant equipment surfing on internet was PC. The most popular platform was Windows. Considering the browsers, Netscape is slightly more than I.E., while less than 10% users were using both of them.
Table 12. Top 10 internet providers (%)
HiNet
SeedNet
TANet
Ht.net/APOL
ADSL
ERANet
EBT
GCNet
Giga
FICNet
2000
55.6
10.4
7.3
1.2
0.6
0.6
0.6
0.4
0.4
0.3
1999
60.2
9.8
9.8
1.5
-
-
-
1.8
-
0.9
1997
46.8
10.1
21.1
-
-
-
-
-
-
-
1996
36.0
13.0
11.0
-
-
-
-
-
-
-
The top 3 government funded providers: HiNet, Seednet, TaNet owned major market share; the ex-monopolized telecommunication company HiNet took the largest pie. However dozens private ISPs were playing aggressive as well as competitive role in the market to develop their territory. It would also bring a fast acquisition and reorganization in small ISPs under such competition in the near future. Lot of users was using what they were given and they did have the idea who was the provider.
Table 13.Barricades to users
None
Connect
Interrupt
Slow
Operation
Privacy & Security
Info & Content
2000
8.5
14.9
17.1
54.7
1.9
2.8
2.4
1999
10.6
14.5
11.7
50.2
1.9
2.2
-
1997
6.5
11.1
4.0
39.1
2.2
0.3
-
Concerning the barricades, bandwidth and related problems including: difficult to connect, interruptions during connect, and slow data transferring were users』 primary complains. Relative few users became aware of privacy and security issues. An emerging and enlightening issue was that users began to dissatisfy to find proper information and contents.
E-commerce
How many users were who considered internet was an alternative use for their normal life? The researcher pointed out 3 alternatives: media and political forum, Learning classrooms, and shopping and E-commerce place. Their answers showed that E-commerce was just beginning.
Not until 1999, CommerceNet/Nielson claimed the beginning of E-commerce year in USA, were there barely 1 tenth internet users shopping online in Taiwan. The increasing rate was even not much in 2000.
Table 14. Internet as an alternative use(%)
Use as
Media & Political forum
Learning
Shopping & E-commerce place
2000
32.8
-
10.9
1999
33.1
2.7
9.9
1998
-
0.6
-
1997
18.2
-
5.5
If we put the data of Table 15 into the same S-type innovation model, we would find that E-commerce was still slowly moving on the flat, long stage of awareness.
The average monthly purchase records were also relatively low.
Table 15. Year began to shop online(%)
1995
1996
1997
1998
1999
2000
2000
0.1
0.2
0.5
1.3
4.9
3.5
Table 16. Purchase volume(US$)
Average per month
Lower limit
Higher limit
2000
96
3
580
1999
64
9
370
On the purchased products, there were not many categories before 1999 while there were diverse choices after 2000.
Relatively hot products were: computers software, books, leisure commodity such as music CDs, and electronic hardware. Next were tickets, toys, fashions, adult accessories, furniture, and cars.
There were 2 features for online bestsellers: low price and stable quality. Service oriented products were also promoted such as 『hours of internet use』 (that was included in the item of software) and ticketing.
Table 17. Purchased products-multiple choice(%)
Books
H/W
S/W
Leisure
Toys
Furniture
Fashion
Cars
Adult
Ticket
Other
2000
2.5
1.4
3.1
1.5
0.5
0.2
0.3
0.2
0.2
0.9
1.6
1999
2.4
0.5
-
2.6
-
0.3
-
-
-
-
3.2
Were there any concerns for experienced online consumers?
Relative few people did not worry. The barricades to them were: the security of transaction, the leakage of personal privacy, the product quality, the way to pay, the after services and refund policy, the possibility of fraud, and the time for delivery.
In the other hand, the non-buyers』 reasons were almost the same. Some did not know yet that they could shop online. However, there were certain amount users who just not used to E-commerce.
A promising part was there were 14.9% non-buyers would consider to try E-commerce in the coming 6 months.
Table 18. Barricades to consumers-multiple choice (%)
None
Quality
Security
Fraud
Payment
Delivery
Services
Privacy
2000
14.4
13.8
21.6
2.9
12.1
2.4
7.3
22.3
1999
14.8
10.2
36.1
-
-
25.0
-
-
Table 19. Reasons to users don』t shop-multiple choice (%)
Don』t Know
Quality
Security
Fraud
Payment
Delivery
Services
Privacy
Not used to
2000
3.9
12.9
17.5
12.0
6.9
2.1
3.2
17.3
24.2
1999
11.5
47.2
-
-
-
-
-
-
17.9
For the possible coming buyers, they provided their shopping list. Most of categories were as same as existing products. Though, they suggested 2 new potential items: cosmetics and travel packages.
The average and range they would like to pay for the online shopping coincidentally remained the same as existing buyers.
Table 20. Potential products-multiple choice (%)
Books
H/W
S/W
Leisure
Toys
Furniture
Fashion
Cars
Adult
Ticket
Cosmetics
Travel
2000
7.2
4.0
4.5
4.1
0.5
0.6
1.0
0.5
0.2
2.3
0.6
0.2
Table 21. Acceptable price(US$)
Average highest
Lower limit
Higher limit
2000
97
6
600
Would the potential buyers pay for subscribing information?
They would agree to subscribe financial analysis, entertainment features, and computers related information. The acceptable monthly fee was around US$16.
There was also 16.1% users agreed to post internet advertisement if it was necessary.
Table 22. Acceptable categories of info to subscribe-multiple choice (%)
Financial analysis
Financial inside info
Political news
Entertainment
Education
Computers info
2000
4.9
1.1
1.0
4.2
1.8
3.6
Table 23. Acceptable fee to subscribe info(US$)
Average per month
Lower limit
Higher limit
2000
16
3
97
Summarized all findings, it should take very careful policy to run E-commerce in the near future in Taiwan.
Non-users Behavior
Non-users were asked to report if they ever heard internet, and how much they knew about internet by the scale from 1 to 100. The mean score was 28.8 in 1997, 20.0 in 1998, 31.2 in 1999, while it was 17.6 in 2000. Those who reported their knowledge below mean would be categorized as not understood.
Table 24. Non-users knowledge about internet(%)
Non-Users
Never heard
Heard, not understood
Understood, not using
2000 Taiwan
49.4
20.1
17.8
11.4
1999 Taiwan
69.4
19.6
32.4
17.4
1998 Taiwan
68.6
21.8
22.7
24.1
1997 Taiwan
82.8
36.5
33.8
29.7
2000 Taipei
38.3
13.0
14.4
10.9
1999 Taipei
68.2
18.5
33.8
15.9
1998 Taipei
54.1
10.8
16.0
27.3
1997 Taipei
75.4
23.8
39.0
37.2
1996 Taipei
82.8
25.3
36.9
20.6
The respondents in Taipei had more knowledge about internet in each category than who was in Taiwan regional wide. It was worthy noting that there was always a considerable part of population who had have never heard of 『internet.'
Table 25. Reasons of not on internet yet (%)
Lack of knowledge
Short of facilities
No need or not interested
Fear or language concerns
Too busy to use
Cost concerns
2000
13.8
22.2
32.9
8.1
18.3
4.7
1999
18.5
10.0
33.9
7.1
4.9
20.5
1998
19.2
60.7
7.8
1.4
2.3
6.8
1997
22.3
33.5
24.4
1.7
8.9
3.0
1996
36.1
27.9
-
8.1
-
9.3
Short of facilities, lack of knowledge and no need or not interested were the biggest obstacles for potential users. The data also implied that internet did not walk in ordinary life in Taiwan yet. Anyway, it would promise a lot of attraction to the potential users in the future, if facility was more available.
Although cost concern was not a significant factor for respondents to use internet, the researcher asked them the rate range that would attract them to apply an internet account. The result was approximately about US$1.00 per hour in 1997; jumped up to US$3.00 in 1998 and 1999, and went down US$1.00 per hour again in 2000.
The demographic analysis of non-users was almost on the contrary to the users. Summarized profile was as follows.
Female was more than male. Senior person with age over 51 was the largest group.
Most of non-users』 education background was under high school. Almost half were under primary school.
The job fields of non-users were also various except about 1 third was housekeeper.
The income of non-users was lower than average. It indicated again that the social economic status was highly related the motivation to access internet.
CONCLUSION
Chapter 2 for Internet Use
Moving through the first Critical Point
According the S-type innovation model, we could announce that the internet use in Taiwan had passed the awareness stage, been entering the trial stage, and been marching toward the evaluation and adoption stage.
Figure 1. The innovation and trend of internet use
There were about a half of households who used the internet at least once in Taiwan, 2000. The population was more than 21 million in this area according to the Census Bureau[2]. It could estimate the optimistic maximum user number would be around 10 million.
The qualified users who had email accounts were 23.1%. Taipei Metropolitan is the center of prompting new ideas, new products, and new culture. Thus, the general user percentage of Taipei was significant higher, while it of Taiwan Province was lower, than the average; but the qualified user percentages seemed close to each other.
By the historical data from our research, as Table 7, we could sketch out the Figure 1. We would find the consistency of 5-year』s data, shape and trend. The figure not only matched the S-type innovation model but also could predict the future growth.
According to the S-type diffusion of innovation theory and the empirical evidence of American studies, it seems that internet using has walked through the awareness and trail periods and is running toward evaluation and adoption periods now. The slow, long and flat journey of S』 left tail is almost finished, internet is taking off the runway in Taiwan.
A Mirage and an Oasis
Internet has been becoming a very hot issue in Taiwan since 1995. All major newspapers have a special page, or pages, everyday and a special stack on internet events of their Sunday edition. Every TV network is used to present internet coverage from time to time. This research result reflected an asymmetric phenomenon that was not proportional to user percentage.
Qualified users were less than a quarter of total population. The number indicated that internet user was still a relative minority in Taiwan.
The hot flame of internet did not seem parallel to the relatively cool access of use. Internet was not so easy for people with lower social and economic status to surf on yet. It was probably still a forbidden palace for the elite, and a somehow distant mirage for the public of the pyramid』s bottom.
In the other hand, there were lots of potential users wished to enjoy the web navigation. We could expect that there is an oasis in the front.
The critical problems for non-users were short of facilities and lack of knowledge. They would make it to the oasis, if they knew the direction, and had a camel.
We could hear the bell shows that the camel is heading forward already; and we could feel a fresh breeze is coming through. It is not too far away to meet the oasis behind the mirage.
Cybercitizen and Knowledge Gap
The features of internet users in Taiwan looked very similar to users in USA. The world 『cybercitizen』 is quite a real thing more than a concept.
Gender: Male user and female user were getting closer.
Age: Young generation was playing the leading role. Main users』 ages were under 30. Senior persons did not increase their participation yet, while CommerceNet reported this hint.
Education: The higher educated person had the higher possibility to be user. Students of college and academy were one of the major groups on internet.
Income: There was an interesting two sides that internet users』 income were either above average or not stable. More than one fifth of users had good revenue while student users had no regular wage. An explanation is that students were using the free academic network TaNet.
There were half users accessed internet at home. It matched the most popular information were 『soft』 materials. It also implied that internet subscribers might share their account to their folks and friends.
The hottest information that users were looking for was entertainment and life. Users need internet when they want to sit back.
Search services were getting hotter because the webs were getting more interwoven. Users would be lost without a reliable navigator.
As the moving of see-saw, services with command interfaces were falling down when WWW with multimedia interface was keep on rising up.
Communication was the most important service for users. Internet was a confidant and a comfortable way to get in touch with others.
Users』 favor of communication was similar to the users in States. It implied that the internet community shared the similar interests without international differences.
On the contrary, there was continuously more than 1 fifth of total population who had never heard of 『internet』 even under the bombing of media『s promotion raid. (An explanation might be 『recognition dissonance』 that implied media failed to leave any trace on their audience』s memory). Their profiles were different from cybercitizen and belonged to low social economic status.
This comparison pointed out the warning of knowledge gap and its possible consequences.
E-commerce Strategies
Market scale
Users in Taiwan did not recognize internet as shopping place yet, while CommerceNet informally announced that USA was entering the age of E-commerce.
Business info was considered more than 10%, but the real transaction was much pretty low.
An alternative for current marketing strategy was to promote business to business market and provide unique goods or services.
Purchase Volume
The user percentage in Taiwan was relatively closer to the Cyber Dialogue and Find/SVP『s finding. Though it might imply a cross-cultural behavioral pattern in the same time frame, its meaning was not the same because the United States』 market size was much larger than Taiwan『s. Same number did not produce the same values. Internet commerce is still under fermenting in Taiwan. Furthermore, we would like to remember Cyber Dialogue』s warning: Do not inflate the internet activities.
The reasonable pricing policy should be around US$100.
Purchase Products
More than 1% users have bought computers software including hours of internet use, books, leisure commodity such as music CDs, and electronic hardware. More than 0.1% buyers used to purchase tickets, toys, fashions, adult accessories, furniture, and cars.
Potential customers expected two promising products: cosmetics and travel package.
Barricades to E-commerce
Most of online shoppers kept cautious. Before 2000, they worried about the security of transaction, the time for delivery and the exaggerated advertisement. Now, their prior concerns were on the leakage of personal privacy, the product quality, the way to pay, and the after services and refund policy.
For the internet users who did not shop online, their reasons were almost the same except they consider fraud was a more serious problem. An interesting fact was nearly a quarter of users who did not have any concern, they just not used to E-commerce.
Looking forward to the mirage of E-commerce, we enjoy the longing bell rhyme of the moving straight camel. We need a long walk to the oasis. We could be optimistic, and should be careful.
References
[2] Census Bureau; Government Statistics; http://www.dgbasey.gov.tw/; 2000
[3] CommerceNet Consortium/Nielsen Media Research﹔The CommerceNet/ Nielsen Internet Demographics Survey; http://www.commerce.net /research; 2000
[4] Commonwealth Magazine; Internet Survey; Commonwealth; Issue: Nov.: 124-38; 1997
[5] Cyber Dialogue; American Internet User Survey; http://www.cyberdialogue.com; 2000
[6] Education Ministry﹔The Traffic Growth of Services of TaNet; http://taiwan.csie.ntu.edu.tw/inet98/traffic/,2000
[7] Find/SVP﹔The American Internet User Survey; New Survey Highlights: http;// etrg.findsvp.com 1996
[12] Liou, Da-Cuan; How to Perceive Net Media Corectly;tw.talk;1996
[13] Lycos; Web User Study; http://www.cyberdialogue.com/press.html; 2000
[14] Media Matrix; http://www.mediametrix.com/usa/index.jsp; 2000
[15] Network Wizards; http://www.nw.com/; 2000
[16] MIDS Internet Demographic Survey﹔ http://www2.mids.org, 2000
[18] Rogers, E. M.; Communication and Development: The Passing of a dominant paradigm; Communication Research; 3: 213-40; 1976
[19] Rogers, E. M. and F. Shoemaker; Communication of Innovations; NY: Free Press; 1973
[21] Wu, Tung-Xiung; Survey Research on Household Internet Users in Taiwan, 2000; Taipei: Research paper sponsored by Economics Ministry and III; 2000e
[22] Wu, Tung-Xiung; Research Series on Internet Users in Taiwan, 1996-1999；Paper at the Communications and Technology Conference, 1999; Shin Zhu: Chiao Tung Univ; 1999i
[23] Wu, Tung-Xiung; From Mirage toward Oasis: Research Series on Internet Users in Taiwan; in Despotis, Dimitris K. and Constantin Zopounidis (eds.) Integrating Technology & Human Decisions: Global Bridges into the 21st Century; Athens: New technologies Publications; Vol. 1, p: 873-7; 1999e
[24] Wu, Tung-Xiung; A Distended Mirage and a Distant Oasis: The First Empirical Study on Internet Users in Taiwan; WDSI: Proceedings of the 26th Annual Meeting for the Western Decision Sciences Institute; Hawaii: WDSI; p:811-6; 1997