China Net Market Size and Demographics Briefing
Version 2.0, Part One
March 1, 1999
The following briefing uses data from three consecutive surveys of the China Internet market conducted by the China Network Information Center between November 1997 and January 1999 (for background on the CNNIC reports, see Appendix A: Introduction to CNNIC surveys.) The analysis is based on Matrix East's own data, market experience, and consultations with industry contacts within China. The briefing was written by Matrix East president Kenneth Farrall.
|
How Many Internet Users are there?
CNNIC'S latest report provides the following data for January 1999:
total computers connected to the net: 747,000
total direct connections: 117,000
total dial-up connections: 630,000
total estimated dial-up users: 1,490,000
total estimated leased line users: 400,000
total estimated users with both leased line and dial-up: 210,000
total unique Internet users: 2.1 million
Based on Matrix East analysis the picture looks more like this:
total registered accounts: 747,000
total estimated dial-up users: 3,080,700
total estimated leased line users: 954,954
total unique Internet users: 3.8 million
Matrix East estimates are based on the incorporation of terminal sharing, unauthorized proxy networks, and account sharing habits as they appear in seven access sub-sectors (such as business leased line and consumer dial-up) into estimates of the number users per registered accounts (see table below.)
For a thorough explanation of the multiplier issue and Matrix East's estimate see Appendix B: Exploring sub-sector multipliers.
| Sub-sector | Registered Accounts | Users per Account |
Estimated Net Users |
| Leased Line Business | 67,392 | 673,920 | |
| Leased Line Government | 15,912 | 111,384 | |
| Leased Line Academic/Research | 33,930 | 169,650 | |
| Dial-up Consumer | 220,500 | 661,500 | |
| Dial-up Business | 201,600 | 1,209,600 | |
| Dial-up Government | 81,900 | 327,600 | |
| Dial-up Academic/Research | 126,000 | 882,000 | |
| Total Unique Internet Users | 3,833,872 | ||
|
 |
The Chinese Internet user today is predominantly urban, with nearly 45% concentrated in Beijing and Guangdong alone.
Access Points
44 per cent of respondents report accessing the Internet from home, while 3 per cent use increasingly popular Internet cafes.
Technical Platform
Operating SystemInternet Users: Who are they?
98 per cent of China's Internet users are using windows-based operating systems. Roughly 1 per cent report using the upstart operating system Linux.Browser Choice
Microsoft's Internet Explorer browser has a much greater market share in China than it does in Western markets. Though many analysts point towards Microsoft's aggressive marketing tactics and deals with high level government departments, the reason most net users shy away from Netscape is much more fundamental: a persistent bug (since version 3.0) that limits Netscape's ability to properly display Chinese characters. Several Chinese sites, including CNNIC, have their text replaced by empty boxes when viewed in Netscape.Modem Speed
86 per cent of Internet users go online with 33K modems or better.
According to the CNNIC reports, the typical Internet user is young, male, urban, college-educated, technically oriented, and accesses the Internet from a Windows PC. A very high percentage come form China's major cities. According to the latest report, 37 per cent of users have incomes between 400 and 1000 RMB a month.
Category by category demographic breakdown for the latest report follows:
Age and Sex
Chinese Internet users are young and male. 79.7% are between 21 and 35; 86% are male. Overall age breakdown is below:
Income
Thirty seven per cent of China's Internet users earn between 400 and 1000 RMB a month. 25% earn more than 2000 RMB a month.
Education
89% of all respondents have had college preparatory work or better.
 |
Academic - Research 29% Student 16.4% Professor 6.2% Science & Research 6.4% Business 57.4% Computer 17.4% Manufacturing and Mining 11.9% Post and Telecom related 8.7% Financial and Insurance 6% News Media 2.1% Cultural and Entertainment .5% Other Business 10.8%% Government 13.6% Govt & Party 8.1% Social Services 3.6% Health and Welfare 1.9% |
The CNNIC report breaks down the occupations of Internet users into the following categories: student, professor, science and technology, computer-related, manufacturing and mining, finance and insurance, news media, cultural and entertainment, other businesses, post and telecom-related, social services, and health and welfare. Unfortunately, this convoluted categorical scheme precludes any straightforward determination of the ratios between academic, business, and government users. Matrix East has grouped these sub-categories into larger categories -- academic & research (29%), business (57.4%), and government (13.6%) to yield a more general picture of this aspect of user demographics.
How fast is the market growing really?
It has been widely reported that China's Internet market grew a healthy 240% last year, more than tripling from 620,000 in November of 1997 to 2.1 million at the end of last year. Although it is true that official government estimates have increased to such a degree, most reports seem to have missed the fact that the number of registered Internet accounts have grown only 150% over the same time period. Over the past six months, the official estimate for total Internet users grew 75% while the number of actual registered accounts grew at half this rate. Why the discrepancy? CNNIC has been adjusting its multiplier upward to better reflect the number of users per online computer, resulting in higher total user estimates for the same number of Internet accounts. This has created a distorted picture of actual growth, which we believe to be much more in line with the 150% growth in registered accounts.
What are their thoughts on the Internet?
 |
 |
 |
Only seven per cent of respondents reported that they often click on web advertisements, while 39% never click on them at all.
Feelings about e-commerce
There is strong support for e-commerce among China Internet users, although most agree that logistical hurdles remain before it is practical. 87% of all respondents hope to engage in e-commerce in the future once these logistical problems have been solved.
- 65 % believe legal protection is necessary for purchases and that this protection is not yet in place.
- 50% feel there is currently no reliable means for transferring cash such as credit card support.
- 57% are concerned about the potential for low product and service quality online.
- 40% say they can make small purchases such as books online and that the net is useful for researching large purchases such as electronics equipment.
- 34% say it is very easy to do product research and price comparison online.
- 7% say no matter what the product or service they would still purchase at a traditional storefront and not online.
Common Complaints
Asked to list what they felt were the most serious problems with the Internet, 92% agreed access speed is too slow, while 74% complained of high fees. 49% felt there is not enough Chinese content and information available, 7% felt that using the net required technical knowhow and expertise that made it too difficult for the average person, 2% complained about their ISP, and 2% felt the Internet was just plain useless.
How do they feel about their ISPs?
China Internet users rate access speed above all else when choosing their Internet service provider; price and service remain at a distant second. Most users are generally satisfied with their ISPs; 31% say service is good or very good, while 9% say service is poor or very poor.
 |
 |
 |
 |
Time Spent Online
37% of users spend 10 hours or more a week on line. 4% spend an hour or less.
Types of Information Sought on the Net
China Internet users are seeking information of the following types, in order of popularity: science and technology information (including Internet and computer software) 76%, news 66%, sports and entertainment information 65%, business information 51%, educational information 40%, financial information (including stock reports) 34%, employment 30%, and advertising 24%.
Top Ten Most Popular Web Sites
- Yahoo -- Yahoo received votes for both its main site and Chinese Yahoo. Chinese Yahoo is based in California.
- Netease -- owned by Guangzhou-based WangYi Computer Company, Netease promotes itself as China's first true portal, a claim disputed by number three web site Sohu. Netease boasts a number of popular community services, including free web pages and very popular online bulletin boards.
- Sohu -- A web directory modeled after Yahoo, run by master Internet promoter Charles Zhang, Sohu is owned by Internet Technologies China, a Beijing based company with investment from MIT Internet guru Nicholas Negropronte.
- 163 Net -- a popular free e-mail site owned by Guangzhou Feihua Telecom Engineering Company.
- Sina Net -- started by three Stanford University students in early 1995, Sina Net is one of the oldest and most internationally known chinese language sites on the Internet. The site was purchased by Beijing-based Stone Rich Sight Information Technology Company this year.
- Shanghai Online -- Shanghai Online,run by the Shanghai bureau of the Ministry of Information Industry, is a hodgepodge of content and services collected under the Shanghai Online brand.
- Microsoft
- Yeah Net -- One of the original Chinese language search engines.
- Hotmail
- Chinese Popular Computer Week
Increase in Female Users
Although the total percentage still remains low, there was a near doubling of female Internet users, from 7.2% to 14% over the six month period between June 1998 and January 1999. We expect this trend to continue, much as it has in the West, as Internet use expands among the general populous and not just within the technically-oriented, predominantly male, early adopter sector.
| Beijing | Beijing | ||
| Shanghai | Guangdong | ||
| Guangdong | Jiangsu | ||
| Jiangsu | Zhejiang | ||
| Hubei | Shanghai | ||
Guangzhou logged impressive growth in Internet users since June of 1998, accounting for more than 20 per cent of all Internet users for January 1999. Based on CNNIC's official total users numbers, Guangzhou Internet users jumped from 90,000 in June of 1998 to more than 440,000 in January 1999, an annual rate of growth of nearly 1000 per cent! Inexplicably, Shanghai net usership appears to have SHRUNK over the past six months. Again using CNNIC official numbers, there were 142,000 Internet users in June of 1998 and just over 90,000 in January of 1999!
Wealthy Users Continue to Grow Rapidly
Total percentage of net users with incomes over 2000 RMB has grown steadily with each successive CNNIC survey. The percentage grew from 3% in November of 1997 to 18% in June of 1998 and now 25% in January 1999. This is quite understandable considering the high relative cost of Internet access in China and the increasing popularity of the net in major cities where economic growth is strongest. It should be noted, however, that part of the increase between last June and January can be attributed to the net use explosion in Guangzhou, the wealthiest province in China.
2. Behavioral and Attitudinal Trends
News sites increase in importance
Respondents indicating an interest in political and economic news on the net increased dramatically, to 66 per cent from 45 per cent in June of 1998. Western news site CNN also made it into the official most popular site list, coming in at 51, despite the fact that Chinese net users are still occasionally blocked from accessing the site. Interest in sports and entertainment content remains high (65%), after a dramatic increase between the November 1997 and June 1998 surveys.
Less Attention to Online Ads
Potentially discouraging news for online advertisers is that the impact of online advertising appears to be decreasing in China, mirroring similar trends that have been noted in the West. The percentage of respondents reporting that they pay attention to banner ads dropped from 74.5% to 61%, while the percentage reporting they are annoyed by ads because of their affect on download speed rose to 10% from 2.7%. Last fall, Motorola reported tremendous success with its Digital DNA advertising campaign in China, garnering two million impressions 64,000 click-throughs, and 23,000 completed surveys during a five week period.
Increasing Interest in E-Commerce
The January survey shows an increase in both general interest in e-commerce and concern for its current limitations. 87% of respondents indicated that they hope to engage in e-commerce once current obstacles are overcome, up from 78.1% in June 1998. At the same time, a higher percentage of users recognize key roadblocks that make e-commerce difficult at present, including the lack of necessary laws and regulations (up to 65% from 62%,) money transfer limitations (up to 50% from 44.5%,) and concern over product and service quality (up to 57% from 46.1%.)
Fewer ISP Horror Stories?
Internet users have two sections within the CNNIC survey where they can sound off about their ISP. In one section, users are asked to rate the quality of their ISP, from very poor to very good. In the other section, users are asked to choose what they believe to be the most serious problems facing the Internet in China today, with choices including slow Internet speeds, high fees, and poor ISP quality. Although there was little significant change in overall ISP ratings, the percentage of users citing poor ISP quality as one of the most serious problems in China's Internet plummeted from 15.8% in June of 1998 to just 2% for January 1999.
Appendix A: Introduction to ChinaNIC Reports
ChinaNIC, roughly the Chinese equivalent of the pre- Network Solutions InterNIC, is a non-profit governmental organization dedicated to the healthy promotion and development of the Internet in China. CNNIC is responsible for the assignation of top level domains within China and the nationwide distribution of IP addresses. As part of their central role within China's Internet bureaucracy, CNNIC has embarked on a series of semi-annual surveys of the China Internet market, the results of which are made public on the Internet and in domestic newspapers.
ChinaNIC now engages in a nationwide survey of Internet issue two times a year. So far, three such studies have been conducted -- in October 1997, June 1998, and December 1998 -- providing information on the size, demographics, attitudes, and behavior of the Chinese Internet audience. Having three studies over the course of 18 months, with nearly identical methodologies and structure gives more credence to overall findings and provides some indication of possible demographic trends. There have been other studies (see source list), but either their methodology is not clearly laid out or the survey has only been conducted one time. The CNNIC reports provide the best picture of the current market available and will only get more valuable with each succeeding study.
3.Methodology
The surveys are conducted through the distribution and tabulation of questionnaires. The 1997 study was performed in in cooperation with computer magazine China Computer World and conducted via ISP and web site home pages. The June 1998 study was performed in cooperation with the State Council Information Office. 3098 surveys were released and 2494 were returned. The most recent study, conducted last December with the cooperation of China's top web sites and ISPs, resulted in 22,177 completed surveys, by far the largest sampling to date.
4. Reliability
How reliable can the demographic and behavioral information reported in the CNNIC surveys be expected to be? This has been a hot topic among industry analysts over the past few months. Some have questioned the integrity of the data, suggesting that only Internet "newbies" would fill out the lengthy survey questionnaire, skewing the sample. This kind of question has been raised time and time again in relation to online marketing surveys, but in fact, results of surveys conducted online tend to mirror parallel studies conducted using traditional means like phone surveys and mailed questionnaires. Changes in question format over the past three surveys and the occasional poor question wording in the CNNIC survey do have an impact on data integrity, but the large sample size (now more than 20,000) continues to put the CNNIC survey, in terms of the raw data provided, well ahead of any comparable study available today.
Appendix B: Exploring Sub-sector Multipliers
Traditional estimates of Internet user populations generally proceed in three steps. First, determine the number of computers with Internet access; second, determine the average number of individuals per computer; third, multiply the first and second numbers together. CNNIC reports there are 747,000 computers connected to the Internet in China and roughly 2.8 Internet users per connected computer, resulting in an estimate of 2.1 million real Internet users.
Since all Chinese individuals and businesses who wish to obtain an Internet account are required to register with public security bureau, CNNIC can be expected to have a fairly accurate number of registered Internet accounts, from which it appears they derive their number of 747,000 computers connected to the Internet.
There is a significant difference, however, between the number of registered accounts and the actual number of computers connected to the Internet in China, and a much greater difference between the number of registered accounts and actual Internet users than the government is presently willing to acknowledge.
A single registered Internet account can provide access to multiple users in three ways: terminal sharing, unauthorized proxy networks, and account sharing. Terminal sharing, where two or more individuals may make use of the same computer over the course of a day or week, is well-known, accepted phenomenon and is taken into account by official CNNIC user estimates. The latter two, however, are more controversial. Existing regulations on Internet use in China expressly forbid the "lending" or "transferal" of accounts so it is understandable why such practices may be ignored in CNNIC's official estimate. Proxy networks allow a company, for example, to link ten or more computers via ethernet to a single, registered host PC connected to the net via a dial-up or leased line. Companies can, and often do, use such networks to expand their business use of the Internet without the added hassle and expense of registering each new computer with its own account. The practice is against Chinese Internet law, but is increasingly wide-spread due to a lack of monitoring and enforcement. Finally, account sharing, where several individuals purchase and share use of a single Internet account, is also gaining in popularity despite being illegal.
The extent to which terminal sharing, proxy networks, and account sharing occur in the Chinese Internet market varies along with certain "sub-sector demographics." For example, account sharing occurs most commonly within the student community, where students rarely have enough disposable cash to pay for their own Internet account and where university-centered social activities provide the perfect backdrop for popularizing and organizing the practice. A large, cost conscious import-export company would have more motivation and more available expertise to use proxy networks than a local government agency that has just received a directive to go online.
Below, we break down the China Internet population into seven sub-sectors -- three for leased-line access and four for dial-up -- and then examine typical account useage patterns for each to come up with an estimated number of real Internet users per registered account.
It is important to note here that assumptions for sub sector multipliers is largely anecdotal and difficult to pin down. Companies that connect multiple computers to a single registered Internet account are reluctant to report such behavior because it is illegal under current Internet law. We have chosen to be conservative in choosing our multipliers so that we err on the low side instead of making an overestimate. As our data improves, expect multipliers and our corresponding estimates to increase. In addition, we use these sub sector multipliers for the purpose of estimating different patterns of account sharing and not as indicators of possible online behavior. It is quite common for Internet users in a work environment to use the Internet in "consumer mode" visiting entertainment and news sites and engaging informal online chat.
Leased-line: 117,000 registered accounts
We break down leased line accounts according to the occupational breakdowns determined in section B.2: business 57.4, government 13.6%, and academic/research use 29%.
Leased-line business: 10
After going through the paperwork and guanxi-stroking necessities of obtaining leased line connections to their local ISPs, large private and state-owned industries find themselves free and clear to wire more and more corporate computers for Internet access without the additional hassle of registering these accounts. Although the practice is illegal there is very little enforcement against unauthorized networking. Most company managers are reluctant to discuss the practice for fear of making themselves easy targets, but it is very widespread. If one combines unauthorized networking with terminal sharing, some companies are supporting 40 or more Internet users with a single registered account. From anecdotal research and personal experience with a wide variety of leased-line access set ups, Matrix East estimates there are at least 10 real Internet users for every registered leased line business account.
Leased-line government: 7
Proxy networking and account sharing on individual computers also appears common in government settings. Based on informal surveys of government IS managers at a recent "government online" conference as well as additional anecdotal evidence, we estimate 10 Internet users per government dial-up connection.
Leased-line academic/research: 5
In academic and research settings, where account management is more strict, unauthorized networking of PCs or account sharing appears to be less common.
Dial-up Accounts:
For the purposes of this analysis, we break down dial-up accounts into the following sub sector percentages: consumer (the typical home user) 35%, business 32%, government 13%, and academic/research 20%. These percentages are based on a combination of data sources, including occupational and access data from the CNNIC report and information from China ISPs
Dial-up Consumers: 3
Consumer Internet dial-up accounts are unlikely to have more than one computer in the house and thus rarely, if ever, will have more than a single computer connected to the Internet at one time. While there is some sharing of accounts among consumer families it does not appear to be widespread.
Dial-up Business: 7
As commercial access to the net became available on a large scale in early 1997, a very large number of new accounts were for businesses. The vast majority of these companies lacked the facilities to support a leased line connection. Those that did have corporate networks found obtaining a leased line connection to be a time consuming and red tape-laden process. Most began with dial-up accounts. As companies expanded their Internet activities, these single dial-up accounts, with the help of proxy-connection software like Wingate, would be connected to ethernets of as many as twenty PCs. At a conference of national web site companies in Xiamen in June 1998, most attested to similar setups, with tens of individuals accessing the Net via a single dial-up.
Dial-up government: 4
As China prepares to enter "Government Online Year", with more and more government departments both getting access and developing their own online presences, the technological sophistication and "wired" nature of the average dial-up government employee is increasing. With their respective government agency footing the bill, state workers can afford to stay on line longer and share their experiences with family and friends much as is being reported in academic circles.
Dial-up Academic/Research: 7
Apparently classrooms and other social settings in colleges and universities are providing a major breeding ground for the sharing of account log-in names and passwords. In Beijing, it is not uncommon for 12 students to pitch in together to buy and share a single Internet account, which they then dial-up individually from their dorm rooms or family apartments. Professors are known to share accounts as well though not as extensively as there students. Though not all dial-up users in this category practice account sharing, the practice is widespread enough to push the academic/research multiplier to 7.
Based on these new assumptions, Matrix East estimates there were 3.8 million Internet users in China as of January 1999.