Chapter 6: Analysis and consideration of the results    
Home | Abstract | Table of contents Back <> Next

The demographic profile of Supernet users

Educational attainment and income


The level of education (figure 1) in a population is bound to affect the potential usage of the Internet.  For example, an illiterate person will not be able to follow on screen instructions to be able to identify and navigate to content that they might want to access.  It would be anticipated that knowledge about available resources on the Internet and a desire to research more subjects would increase with higher educational attainment, up to a point.  It is not surprising therefore that the research shows a fairly high level of education amongst the user population.  It might have been helpful to have researched a control sample in order to put these results into context.  Since results from a sample of the general population were not collected, and no external sources of similar data were identified, it is not possible to say if the users of the Internet café are generally better educated than the population at large.  It might be plausible to suggest that they are, if this data is considered along with the family income data in figure 12.  It is not unusual to find a positive correlation between educational attainment and income in the developed world. (Expandglobe.com, 2006, based on US Census Bureau data for 2002) Figure 12 shows that most users think their family income is average or above.  The likelihood is that the education of users is therefore also average or above average.  It is unclear if the lower representation of less affluent people within the user sample is a result of their education or of their income.


The results show that around half of users have at least an undergraduate degree. This is surprisingly high, although the city does have a university.  One of the significant issues that this raises concerns the transferability of the conclusions from this research.  In many other locations in the developing world, education levels will generally be much lower.  A similar business in another location might need access to a larger population in order to generate a sufficiently large, sufficiently well educated potential customer base.


In the USA, Eszter Hargattai (2003) shows a clear correlation between Internet use and the level of education attained. Her 2001 data shows that more than 80% of graduates were users; the comparable number for those who failed to finish High School is 15%.




The results in figure 2 show more than twice as many male users in the Internet café as female users.  One has to question why.  Since the administration of the surveys was not supervised, and all the Internet café reception staff are male, perhaps the female users were not offered the chance to fill out the surveys as frequently as male users.  There is therefore the possibility of some bias in the results.  The cultural context should also be considered.  Aceh has a long tradition as one of the main departure points in Indonesia for the Haj Islamic pilgrimage to Mecca.  Aceh is a semi-autonomous province which is under Sharia law.  There is a strong male dominance within the strongly Islamic culture of Aceh.  It is conceivable that this dominance exerts itself in a number of ways:



From the author‘s observations, a difference in education between the genders

is not significant (although this was not specifically measured) and so this would probably not be a factor in explaining the user gender imbalance.


Why does this matter? Deborah Fallows‘ research (Fallows, 2005) shows that in American society, slightly more women under the age of thirty and black women are Internet users than their male peers.  If it is true that there is no intrinsic reason why there should be a gender imbalance.  This suggests that there is a market potential among women in Aceh which Supernet is not yet exploiting.  It would be worthwhile exploring factors that might attract more female users.  This might include having female receptionists, female only rooms or female only times of day.  When the questionnaires were processed, it was also noted that female respondents more frequently made suggestions and comments relating to issues of comfort and cleanliness, (e.g. clean restrooms) but this perception was not analyzed methodically.




Over 80% of the users in the survey were under the age of 30. This compares with 67% in the Uzbekistan study (Kolko, Wei and Spyridakis, 2003). Demographic data from the Pew Internet and American Life Survey (2006) state that the age spread is much more even in the USA, but that poorer and less well educated users are less well represented among users.  However, the American data reflects usage from any location (including home) whereas the data from Aceh and that from Uzbekistan records only Internet café users. 

As with education, no data was generated or found from other sources which would have allowed a comparison between the age profile of users and those of the general population. Even if some data from prior to 2005 was available,

it would not reflect any changes which may have occurred since the tsunami (maybe more younger, fitter people survived.)


It is worth reflecting on why the proportion of older users is so low.  The current surfing habits of users may partially explain this.  This subject is covered in more depth later.  However, it appears that most of the usage by younger people relates to western sports, fashion, sex and youth culture related discussions. (See figure 21)  Maybe content that appeals to older people, in the Indonesian language, is absent or older users do not know where or how to find it. Maybe they have not had the same exposure to or experience of computers and the Internet, and they have not yet taken time to learn.


Again, as with the gender differences, Supernet might do well to research this themselves and see if there are opportunities for training and for expanding services that will appeal to older users.  The “Readiness for the Networked

World” guide (Center for International Development at Harvard University, 2000) advocates the encouragement of all parts of society to use the Internet in order for the whole economy to benefit.


Language and ethnicity


Figures 4 and 5 show the survey results relating to the language fluency and ethnic backgrounds of the Internet café users.  One reason this information was collected was to determine the proportion of indigenous people who are using the facilities of the Internet café.  Since the tsunami, there has been a large number of foreigners in the area working with the various aid and development organizations.  While these people may currently be using the Internet café, their presence is transitory and so they may not contribute to the  long term viability of the business.  The results showed that very few foreigners were included in the sample.  Casual observation accords with this finding: the users are almost all local people.  Even those from other parts of Indonesia are predominantly from nearby regions.  This is encouraging as the rest of the survey is validated as being highly applicable to the Supernet business.


As already mentioned, the availability of Indonesian language resources on the Internet will have an impact on the extent to which Indonesians use the Internet and the purposes for which it is used. These factors have a direct bearing on viability.


Experience with PC’s and the Internet


Question 7 in the survey asked how long users had been using computers and the Internet (See figures 6 and 7).  A high proportion said they had been using both for more than 3 years.  It would have been interesting to re-phrase the question to see how many had been using for even longer periods of 5, 7 or even 10 years.  Unfortunately, these options were not given to the users.  This uncovers a biased perception on the part of the author, who thought it unlikely that Internet use had been very long term.  Once again, a pilot study might have uncovered this and allowed these questions to be phrased more appropriately. 


The high levels of use for more than three years is especially interesting given the high proportion of users under 20 years of age. 


The length of experience might be related to the range of activities for which the Internet is used, but this was not explicitly studied. It would be interesting to know what applications first appeal to users and get them “hooked.”  It would also be interesting to see how usage develops over time.  The author‘s impression is that users seem to have quite a narrow repertoire of use (after manually reviewing 50,000 individual records relating to which Internet sites are visited.)  Again, it would require further research to establish if there is a factual basis for that impression.  This information could be useful to Supernet and others.  Maybe there is an opportunity to run training which would introduce users to new possibilities.  This may engender more customer loyalty, more intensive and frequent use.  


Alternative access to computers and the Internet


Questions 8 and 9 attempted to understand what alternatives users have to visiting Supernet.  The results are shown in figures 8 and 9. Around 75% of users said that they have access to computers for personal use at other locations (there is a slight difference in the results from the two questions).  Unless these alternate computers are poor, old, unreliable and slow, it would be puzzling to see people traveling from these other locations and paying to use a computer at Supernet. Indeed, usage data in figure 20 confirms that almost all usage of the computers at the Internet café involves using the Internet, not just the computers. The raw data showed the usage of Microsoft Office and OpenOffice applications to be almost non-existent. (although MS Office mysteriously arrived on most machines even though Supernet didn‘t install it!)  Figure 9 also shows that 27 of the 109 respondents had no Internet connection at an alternative location.  38% of users have alternative Internet access (not including other Internet cafés) that is either dial-up or “slow” broadband.  Most of these are at home, at work or at school/college, where presumably they do not have to pay for access.  Yet, they chose to visit Supernet and pay for access.  This must suggest some competitive advantage that Supernet has.  Some suggestions might be:



It is important for Supernet to fully understand that many users have alternatives, and to precisely identify what it is about Supernet that gives it a competitive edge. If any of those alternatives become better than Supernet in important respects, customers could quickly switch their custom elsewhere. Indeed, Supernet should seek to increase its competitive advantage if possible.  This would construct some degree of barrier to entry for competitors.


Travel time


Figure 10 shows the average travel time for customers from their homes and places of work to Supernet.  Surprisingly, some respondents said that they travel for more than an hour to get there, but on average travel time is about 15 minutes from home and 22 minutes from the place of work.  Since there are several other Internet cafés within 5-10 minutes from Supernet, the data suggests that users do not just select the nearest facility.  Once again, the implication is that Supernet’s users perceive that it has a competitive advantage.


Aceh, like many Asian locations has very wide ownership of small motorcycles.  This market characteristic certainly adds to the distance which can easily be traveled and provides Supernet with a geographically large feasible catchment area.  In poorer cultures, where even a small motorcycle is out of reach for most people, one suspects that it would be very unlikely that an Internet business could be viable.  Sufficient rapid personal transport is both an indicator of relative affluence and provides the mobility by which a sufficiently large target market is made accessible.




The motivation for the question about users‘ occupations attempts to gain an indication of the stability of users‘ incomes, and the extent to which Internet access is being used to underpin economic development. In hindsight, it would have been helpful to re-phrase the answers to provide separate data on those working for NGOs and those occupied in businesses.  The reason for this is that a large number of NGO jobs are available as a result of the tsunami.  One can anticipate that over the next few years, such employment will decline.  If  the proportion of users who are working for NGO‘s is high, then the loss of these jobs could impact Supernet’s business.  The author’s sense is that most of the 40%  who said they were employed by businesses or NGO‘s were in fact employed by NGO‘s.  Some evidence that supports this notion is the opinion of several colleagues that wages are said to have approximately doubled since the tsunami, as a result of NGO employment.  In the medium term, many users’ incomes are therefore likely to be unstable.


It is encouraging that about 18% of users are either self employed or employ others, but from an analysis of Internet use, there is little evidence that Supernet is used to generate economic activity.


Occupational and income stability are aspects of user demographics that should be understood when examining the potential viability of Internet businesses in other locations too. 


Mobile phone ownership


The author had a suspicion that there might have been a correlation between mobile phone ownership and Internet café use, because owning a mobile phone indicates a level of relative affluence.  In fact, as figure 13 shows, nearly 82% of users owned mobile phones.  A cursory observation of the general population demonstrated that this was probably the general level of mobile phone ownership.  Data to support that observation could not be found.  It would be interesting to research a number of indicators such as mobile phone and motor vehicle ownership to see if a basket of such indicators were a reliable predictor of the susceptibility of a community to premium Internet provision.  This thinking has already been adopted as a general economics.  So called “burgernomics” suggests that the price of a McDonald‘s Big Mac in different economies is an excellent indicator of purchasing power parity in those countries.  The Economist regularly publishes a “Big Mac Index” as a result. (The Economist, 2006)  


Another issue raised by wide mobile phone ownership is that it reduces the potential to raise the revenue of an Internet café through Internet Protocol (Voice over IP or VoIP) telephone services.  Although the business model originally developed for Supernet called for two VoIP lines, only one was implemented for Supernet’s own internal use, mainly for international calls.  Since services like Skype (www.skype.com) have become widespread, Internet cafés often rent out headphones for a little extra on the hourly rate, and users maintain their own accounts.  This is quite a change from a few years ago, where VSAT usage was dominated by telephony services (Hartshorn, 2004). In many cities, GSM phones services are so widespread and affordable that Internet cafés cannot expect to supplement income with phone services. (BBC, 2006)


BBC (2006) Digital Planet. [Podcast] BBC World Service [online] Last accessed on 6/19/2006 at URL:


CENTER FOR INTERNATIONAL DEVELOPMENT AT HARVARD UNIVERSITY (2000) Readiness for the Networked World: A Guide for Developing

Countries. [online] Last accessed on 6/28/2006 at URL: http://cyber.law.harvard.edu/readinessguide/guide.pdf

EXPANDGLOBE.COM. (2006) Education and Income [online] Last accessed on 9/17/2006 at URL: http://www.expandglobal.com/General_info/education_and_income.htm

FALLOWS, D. (2005) How Women and Men use the Internet [online] Last accessed on 9/16/2006 at URL: http://www.pewinternet.org/PPF/r/171/report_display.asp


HARGITTAI, E. (2003) The Digital Divide and What to Do About It [online] Lastaccessed on 9/16/2006 at


HARTSHORN, D. (2004) Open and closed skies... satellite access in Africa. DS Air Ltd, Horsham p.25 Also available [online] Last accessed on 7/23/2006

at URL: http://www.gvf.org/database/regulatoryDB/africaskiesdownload.cfm


KOLKO, B.E., WEI, C.Y. AND SPYRIDAKIS, J.H (2003) Internet Use in Uzbekistan: Developing a Methodology for Tracking Information Technology Im-

plementation Success Information Technology and International Development 1 (2), 1-19

PEW INTERNET & AMERICAN LIFE PROJECT (2006) Demographics of Internet Users [online] Last accessed on 8/2/2006 at URL:


THE ECONOMIST (2006) The Big Bac Index [online] Last accessed on 9/17/2006 at URL: http://www.economist.com/markets/bigmac/about.cfm

Edited by the author for the web.

© Copyright, 2006  Rob Longhurst (rlonghurst@drasticom.org)