FACULTY OF ARTS, COMPUTING, ENGINEERING AND SCIENCES

 

BROADBAND INTERNET ACCESS IN DEVELOPING WORLD ECONOMIES:

AN INVESTIGATION OF THE FACTORS AFFECTING VIABILITY

     
 Chapter 6: Analysis and consideration of the results (Part 2)    
     
Home | Abstract | Table of contents Back <> Next
DRASTIC Website

How is the Internet café used?

Purpose and motivation for use

Question 10 asked about the purpose for their Internet use. Although the data collected is shown in figure 14, it does not seem to correlate clearly with other data collected.  There is a relatively high level of “no answer”.  On reflection, the terms chosen for the categories (Business use, Educational use, etc) and the possible answers and (Frequently, Occasionally, Never) are at best very subjective and may be considered ambiguous.  It is not felt that meaningful or reliable conclusions can be drawn from this data.

 

A similar problem exists with Question 11 (figure 15).  However, other data collected (especially see figures 20 and 21) produced contradictory information. The problems with this question are discussed earlier.  Since the data presented in figures 20 and 21 was collected through automatic monitoring software, this data is known to be accurate.

 

What applications are used?

Figure 20 was designed to find out how the computers in the Internet café were being used, in broad terms.  One term on the figure needs further explanation. Approximately 22%  of usage time is shown as “System.”  This includes interactions with the Windows Operating System (e.g. Logging on, navigating to applications, and Windows Explorer), use of the billing software etc.  This number appears to be quite high and the reason for this would bear further investigation.  Suffice it to say that this does not represent time spent using any specific user application.  For the purposes of analyzing applications elected by the user, “System” use could be excluded.  If that is done, the data shows that 87.5% of computer usage involves the use of mainline Internet applications: browsers, instant messaging clients and media players.  All other use adds up to 12.5%.  A breakdown of which specific clients that we used has some value.

 

Browsers

Supernet set up the systems so that the default browser was Firefox due to the belief that it is generally less vulnerable to malware and other security issues. 

Nonetheless, 77% of users chose to find and use Microsoft Internet Explorer.

 

Instant Messaging Clients

The Yahoo client totally dominated usage.  MSN Messenger had a 2% share and other clients, which are common in the West (e.g. AOL‘s AIM, Skype, Google Talk and  ICQ) were hardly used at all.

 

Media Players

Real Player dominated in this area, with a 79% share

 

The significance of the applications chosen by users is that it would appear to show less diversity than might be expected among users in the developed world. For example, worldwide, Neilson/Netratings attributed both AOL and MSN with much higher user counts than Yahoo in July 2006, yet their usage in Aceh is negligible. (BigBlueBall.com, 2006)  No data was collected to explain this.  The unprompted user installations of Microsoft Office discussed earlier demonstrate that users could load other applications if they wished to.  Yet the data suggests that they elected to use familiar applications. Whilst analyzing the data behind figure 21, a similar pattern was seen in the websites that were visited. In each category of interest, a few major sites tended to dominate usage.  This leads one to question if the user community (and therefore Supernet) is making optimum use of the rich resources available on the Internet. 

Perhaps Acehnese users are not culturally as inquisitive as some other cultures.  Maybe the channels that are used in the West to disseminate news about new applications and possibilities are not available in Indonesian. What-ever the reason, perhaps Supernet has a role to play in expanding horizons for its Indonesian users.  Apple‘s retail stores have two interesting approaches that might prove adaptable for Supernet. One in a regular program of workshops, introducing users to a range of applications and demonstrating what can be achieved with them.  The other is their “Genius Bar” concept, which provides individual problem solving, advice and usage coaching. (Apple.com, 2006)  These might be used to develop customer loyalty, as well as longer and more frequent usage. Once again, it  would be worthwhile researching the range of web based applications that are available in the local language.  This may well be a limiting factor.  Clearly, the available range of resources that users are aware of will inform the value that they attach to using the Internet.  In turn, this is likely to influence intensity, duration and frequency of use.  These will impact the viability of local Internet businesses.

 

Pornography

The analysis of surfing habits shown in figure 21 has some significant implications for Supernet, and for the viability of similar businesses beyond those discussed above.

 

The data showed over 12 percent of web surfing was associated with pornographic material.  There are three main consequences of this for the viability of an Internet café :

 

Supernet is considering the introduction of  content filtering, which will be implemented at the egress point on to the internet in Atlanta, Georgia.  This would prevent filtered content from using satellite bandwidth at all (compared with filtering using a locally located server.) Although this might cause the loss of some customers, it will allow Supernet to promote their compliance with the local Islamic sensitivities and maybe attract some new users.  At the same time, they believe that the other negative impacts can be dealt with. (See notes at bottom of page)

 

Internet café usage frequency and duration

 

There is a good spread of frequency of visits to Internet cafés.  Nearly half visit three times or more a week and nearly 75% visit twice or more. (See figure 16) About a quarter of users spend more than 7 hours a week online.  To put this in context, this represents a cost of approximately 1.5 times the average daily wage. Over half spend more than 4 hours a week online. (See figure 17).   It is common practice in Indonesia for users to share a user session, with sometimes up to four friends in the same cubicle. Question 17, from which this data is generated, specifically instructs users to discount shared time.  The actual phrasing of the question can be seen in Appendix B.  The data suggests that further increases in usage per person per week might be limited by affordability.  Increasing the number of customers might be a better way to increase revenue.

 

Usage Influence Factors

 

Earlier in this chapter, the need for a clean competitive advantage was discussed.  Figure 18 shows which factors are considered by users to be the most important.  This is a vital reading of where customers place value. The red line on the figure isolates the four most important factors to the right of the line.  These fall into two categories: Access performance and user comfort. 

The two performance factors:

The two user comfort factors:

 

Surprisingly, cost of access was not seen as very important, being the eighth most important factor out of fifteen factors.  This is validated by the speed of expansion of demand at Supernet despite being 20% more expensive than average.  Although comparative data from the developed world was not found, it accords with the author‘s own preferences and those that might be expected of many user communities.

 

This valuable data throws doubt on some other models for Internet cafés in the developing world, which concentrate on delivering some connectivity capability, however poor, at the lowest possible cost.  As C.K. Prahalad puts it, poorer people in the developing world have the same aspirations as affluent people in the developed world.  The challenge is therefore around delivering excellent performance at an affordable cost (high value) rather than delivering cheap second rate products and services. (Prahald, 2005).  Many Western based companies still do not appear to understand this. For example, Microsoft were at one point  planning a “Starter” edition of their forthcoming Vista operating system aimed at “developing markets,” with many of the advanced features removed or disabled. (Wikipedia, 2006)  One has to question why anyone would buy a “hobbled” version of Visa from Microsoft, when they can buy an illegal copy of just about any software, including Windows XP and Server editions , complete with a range of access codes, for little more than the cost of a  CD (usually about $1.50) or a free copy of fully functioning Linux distributions. It is noted that the Starter edition does not appear in the current comparison of planned editions of Vista on the Microsoft Website. (Microsoft, 2006)

 

The business model developed and implemented at Supernet delivers the value that users want, evidenced by the rapid expansion that has been possible.  However, it should not be assumed that value will be identically perceived by customers in every location. 

 

Utilization percentages for computers

In the Internet cafés, when computers are being used, revenue is being generated. When they are not being used, revenue is not being generated. Utilization levels therefore critically affect viability.  Figure 19 shows the utilization percentage achieved each hour over a monitored period.  It can be seen that there was a large variation in the utilization during this time.  Since the author was not present during this period, the reasons for spells of very low utilization have not been established.  By the time the data was received and analyzed, Supernet staff were unable to remember what events occurred on specific days.  This business has suffered periods of unreliability, due to a number of problems such as:

On the best days, when everything is operating as designed, utilization levels close to 60% are achieved.  In the months preceding this research, Supernet reported that an average revenue of $215 per computer had been achieved, indicating an average utilization of around 45% (Assuming $1=9000Rp).  At an average of around 36%, seen in the results data, this would produce only about $173 per computer per month.  If the 60% levels were maintainable, monthly revenue would be about $288 per computer.  With around 30 computers between their two sites, the difference in revenue between the monitored average and the best days equates to about $3500 per month. That size difference in revenues could easily make the difference between a very unsustainable business and a comfortably profitable one.  Utilization improvements can only happen if the management of the business is able to address the unreliability issues.   It has already been demonstrated that reliability is also a key factor in determining which Internet café users will patronize.  This is another incentive to focus on achieving it.

 

Bandwidth Utilization

 

Just as the utilization of the computers in the Internet café affects viability, the utilization of the available bandwidth is also important.  Bandwidth represents a significant fixed cost in the business, from which revenue must be optimized to achieve viability.

 

Figure 22 shows the way the consumption of bandwidth varies with time on one sample day.  Ideally, it would be possible to maintain 100% utilization of available bandwidth day and night and to generate good revenues from all usage.  In practice, as figure 23 shows, different types of users have a different consumption pattern.  The assumption that a proportion of bandwidth could be resold to other businesses (whose peak usage is offset by several hours from the peak usage at the Internet cafés) was validated at Supernet.  However, there are still periods of the day where bandwidth utilization could be increased with-out compromising the performance for existing users.

 

Figure 24 shows that download throughput averages are always above the contractual minimum and Upload averages are close to or slightly below.  This would indicate that generally good level of utilization is being achieved.


References:

APPLE.COM (2006) Genius Bar [online] Last accessed on 9/16/2006 at URL: http://www.apple.com/retail/geniusbar/

BIGBLUEBALL.COM (2006) IM Market Share [online]  Last accessed on

9/16/2006 at URL:http://www.bigblueball.com/forums/general-im-news/34413-im-market-share.html

MICROSOFT (2006) The Editions [online] Last accessed on 9/16/2006 at URL: http://www.microsoft.com/windowsvista/getready/editions/d

PRAHALAD, C.K. (2005) The Fortune at The Bottom of The Pyramid: Eradicating Poverty Through Profits. Wharton School Publishing, Upper Saddle River, NJ

WIKIPEDIA, (2006) High Speed Downlink Packet Access [online] Last accessed on 9/13/2006 at URL: http://en.wikipedia.org/wiki/High-Speed_Downlink_Packet_Access

Notes (not part of the original dissertation):

Since this research was completed, content filtering has been implemented.  The early results were approximately 25% reduction in bandwidth consumption, and an estimated loss of about 10% of the Internet café customer.  The long term effects and the  impact on malware management have not yet been measured.



Edited by the author for the web.

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

 

 

 


Ó