Stephen Ruth
Professor of Technology Management and Director
International Center for Applied Studies in Information Technology
George Mason University, Fairfax, Virginia
Robert Schware
Senior Informatics Specialist
Industry and Energy Department, Telecommunications and Informatics Division
World Bank, Washington, D.C.
This paper provides practical, experience-based suggestions to help maximize potential success of networking in developing countries by recognizing "vital signs" that can predict success or failure early on in the life of a project. Major investments in Internet connectivity have been wasted due to insufficient assessment of the readiness of the organization. Even after implementation, there are surprisingly few systems that have been rationalized sufficiently to provide a reasonable evaluation of achieving or failing to achieve stated goals. Those success predictors that can shape the outcome of the project discussed in this paper have more to do with organizational, planning, and human behavior issues than with equipment considerations. A framework for describing the characteristics of a successful implementation is presented together with specific examples from developing countries. Much of the material included here is based on a recent World Bank internal document drafted by the authors, Internet Toolkit for Task Managers (World Bank, 1996).
There has been an explosion of Internet host sites in the developed world, but less developed countries, representing most of the world's population, are on a trajectory of becoming technologically disenfranchised from the benefits of the Internet revolution. A recent report (World Bank, 1996) found that, while the number of host sites in developing countries is increasing, the fundamental statistics are relatively unchanged. North America, Western Europe, Australia and Japan, with about 10 percent of the world's population, have about 95 to 98 percent of the Internet host sites. About 180 countries with an aggregate population close to 5 billion have about a half million of the 15 million hosts sites.
This paper is based on the authors' experiences in implementing Internet connectivity in numerous countries and in every region of the world. It offers insights and specific guidelines that can assist in predicting the success of implementation for project managers. While it would seem obvious that a major investment in connectivity would be preceded by an assessment of success factors in the same manner as for other large investments, this seldom happens with Internet-related investments by bi- and multilateral donor organizations.
The following sections briefly look at the most common success factors we have found in Internet implementations in developing countries. Practical questions are presented at the end of most sections for project managers so that problems can be anticipated and solved before they become problems, and maximum use can be made of connectivity options.
The primacy of the end user is particularly applicable in the context of Internet implementation in emerging countries. While figures are difficult to confirm exactly, it appears that in the range of 40 to 60 million persons use the Internet as of December 1995, with high double-digit annual growth rates projected (MIDS, 1996; Rutkowski, 1995) Most of these users are in developed countries; until recently more than half were in the United States. The process of bringing this technology into regions or environments unprepared for it will require considerable technical and institutional guidance.
However, our experience indicates that many planned network and computer implementations are the result of detailed justification for the hardware and software, electrical connections, and other technical issues but provide little or no information on the users. A recent request to a potential donor for major networking improvements at a university in an emerging country included two paragraphs on the potential benefits to users and an inch-thick sheaf of hardware and software documentation. The request had no information on how many users would be added, whether these users were managers, researchers or lower-level employees; what effects could be expected in terms of productivity; what unique characteristics might differentiate these users from "typical" users; or even the types and volumes of message traffic that would be expected, incoming and outgoing, after implementation.
For many years, it has been known that successful introduction of any new idea, whether technology-related or not, requires a careful elaboration of the strategy of inducing the users to participate in the idea. Although the strategies vary according to the situation, a common requirement is aligning the users' behavior patterns with the use of the intended new technology. A determined, careful focus on the end user's behavior is at the heart of a successful technology transfer application. To gain leverage from a technology investment, the ultimate benchmark is some positive change, expressed in units like output, productivity, capacity, and the like. These are aggregate measures, rooted in end users' behaviors and their minute-by-minute use of the technology.
Some of the most successful networks have been characterized by a cadre of potential users anxious to be able to connect for a serious purpose and needing to be linked in an ever-widening circle. These successes, which have been "demand driven," have a common trait: the users and potential users already have a good idea of the types of services they need. For example, ZamNet, Zambia's highly successful low-cost network, was not set up initially as a full-scale system with powerful hardware and connectivity features. Instead, it gradually evolved as a community of users with common needs. As new users were added to the system, they already knew some of the basic capabilities of ZamNet and were able to broaden its applications gradually, based on individual and group needs. Users knew what they wanted and had learned to take advantage of the system, not as a toy or out of curiosity, but as a tool that enhances productivity and understanding in their lives (Bennett, 1994).
ZamNet offers many insights. The only viable technology solution available was Fidonet, a venerable store-and-forward system. Fidonet operates by users sending messages to other users through phone lines and having their traffic routed cooperatively from site to site. In various regions in Zambia, systems operators, "sysops," took on the responsibility of rerouting messages over phone lines from one location to the next until they were forwarded as a group to the hub in South Africa. Hundreds of users have been able to achieve good e-mail service by opting for this cooperative approach. Also, ZamNet was able to secure good cooperation from a service provider in South Africa, Uninet, which actually paid some of the line costs during the early years of ZamNet's growth. The early success of ZamNet led other users like nongovernmental organizations, universities, and research institutions to join and pay an access fee, thereby assuring financial viability.
Another insight has to do with technology. The most significant problem in the telephone lines was not in circuits across mountains or rivers; it was the last few kilometers into the node in Lusaka, the capital city. The "last mile" is an important consideration in planning for networking success. In Peru, the connection of hundreds of university faculty and students to Red Cientifica Peruana (RCP), was a similar experience. Jose Soriano, who was primarily responsible for establishing RCP, used low-cost methods to link this group. The success of RCP is measured not by investment but by message volumes, FTP calls and trends of Internet traffic, all of which show remarkable growth. In South America, where the ratio of Internet hosts to total population is low, Soriano's achievement is all the more notable. In this case, as in Zambia, the customer demand and implementer energy drove the network. Both these cases suggest that a clear sense of the applications (data gathering, e-mail, listserve, and so on) and the needs of the community of users (faculty, students, others) are significant predictors for success.
In an organization, the support of management is usually necessary for any innovation to take place. In the context of Internet connectivity in an organization, at least four levels of management support can be distinguished: energetic, active, passive, and negative.
The most positive predictor in the area of management support is characterized by a leadership structure that is fully committed to all the possible gains inherent in electronic networking. In this case, managers are taking steps to train employees as well as other managers long before the equipment arrives. Leaders already have a rough idea of how to evaluate the progress of the network, in terms like the volume of traffic in and out, applications used or trends in user behavior. Top managers themselves plan to be among the first users and are already agitating to be connected to the rest of the organization through a local area network as soon as possible. This idealized management attitude is rarely found in any organization, whether in a developed or less developed country. Nevertheless, it is undeniable that enlightened, enthusiastic support by senior management is a powerful force to ensure the success of networking projects. Examples of these occasional successes can be found in the Internet applications literature (Danowitz et al., 1995; Kramer and Dedrick; Ruth, 1992; Ruth, 1990).
Active support is a far more likely expectation than enthusiastic support. The manager showing active support is involved in the planning, is open to possibilities for improved productivity and better communication, takes an interest in some of the implementation issues even though he or she does not have time to get involved in the details, and normally projects a positive attitude at all levels in the organization. This case is more common than enthusiastic support, and there are many examples to be found in recent years (Neil et al.; Ruth, 1995).
Except in some situations where custom, religion or prevailing societal norms operate as natural inhibitors, it can usually be expected that the management will not be adverse to receiving a donation of equipment and funding for installation, implementation and training associated with electronic communication. That level of support, characterized by acceptance and acquiescence but not strong support, could be described as "passive," and is a predictor of success just as surely as the presence or absence of an assured user community. Passive support by management means that there is courteous, occasionally animated interest in the project but little of the sense of having a stake in an important effort. Passive support is typified by potential users being primarily at lower levels of the organization, or by noticeable lack of participation by senior managers in the planning process. This can result in abdicating the leadership role to persons with more technical qualifications and often restricting user populations to administrative or tactical functions rather than widespread use. Passive support can also mean that there is little anticipation of large volumes of message traffic once the expenditure is made--because the reward structure of the organization does not favor using networking over other methods--fax, for example.
A serious negative predictor is a management that looks upon a networking system as a threat. The perception of threat can be attributable to many factors. For example, since a significant number of the users of the Internet and WWW are males under the age of 30, some managers may resent the use of a technology that has that bias. Incidentally, the disparity of representation by age and gender on the Internet is gradually diminishing. Some managers will feel threatened by the wide variety of services available on the Internet and WWW. If an organization has been operating in an atmosphere of data deprivation and the possibility of a tidal wave of Internet sources looms ahead, it is not surprising that the reaction could be less than positive. There could be concerns about the distractions from work that occur as employees start "surfing" the networks. (Research indicates that this concern can be valid if steps are not taken to curb inappropriate use of Internet at the work site. The network does provide many distractions.) There could be genuine fears that Internet tools could make many traditional data search methodologies obsolete. There could be concern about bypassing traditional channels of authority, a problem that has surfaced in many developed countries.
Another negative predictor for networking success is the possible effect of some religious beliefs, traditions and affiliations (Goodman, et al.). Anti-pornography and religious groups in developed and developing countries have expressed concern at the availability of pornography on some Usenet groups and a more general skepticism at the unregulated nature of the Internet. If an organization's management supports this view of the Internet as unregulated, ungodly and a threat to morality, successful implementation is obviously unlikely.
While support from government agencies is needed for most technology transfer activities in developing nations, the manifestations of that support vary dramatically (Ruth, 1994). Full support for a networking project can be characterized by a willingness to plan for high usage at all levels of the organization. This means that agency heads, staff and even the most senior officials have formulated plans to use the network themselves and openly expect that staff will use it. This level of support is rare; in most cases, a relatively small group will wish to be connected to the network for specific purposes. A good success indicator would be that at least one-fifth of the senior government managers not only wish to be connected to the Internet, but anticipate sending and receiving about a dozen messages per week, half e-mail and half information gathering, from either Gopher or WWW services. It should be noted that the individuals do not themselves need to send and receive the messages, sitting at their computers and punching keys. Many successful implementations have been characterized by senior officials having their secretaries or staff send and receive Internet traffic, a perfectly acceptable approach, but they clearly understand the power and potential of networking and are extracting value from high use.
The PTT plays a major role in connectivity in most developing countries. Sometimes the PTT has monopoly power in pricing services for leased lines, Internet connectivity services, and in some cases even the price of modems. In Zimbabwe, for example, the Public Telephone Company can require a user to rent slow modems at high prices. The Zimbabwe PTT also has a statutory monopoly of the third party use of its telecommunications network, thereby making difficult the emergence of competitive Internet service providers (World Bank, 1995). It is often necessary to enter into direct negotiations with the PTT to preclude problems with service later on. Ideally, the PTT will see the emergence of Internet connectivity projects as a way to stimulate growth of the telecommunications sector, but in cases where the PTT is not cooperative, prospects for success are significantly diminished. The PTT can be responsible for the time required to obtain permits, the imposition of special charges for telecommunications equipment, and the quality and cost of leased lines. Hence, the PTT's cooperation is a major predictor for success in a networking project.
It is possible to anticipate the implementation's success by using some of the techniques already described and setting up a simple matrix that matches expectations with measured results (Ruth, 1993; Goodman et al., 1994). Table 1 describes some of the specific metrics that can be used. They are obviously measurable and available in most organizations. It is a management decision as to which of the proposed metrics can best predict and then measure success, but clearly a half dozen of them would be more than enough to get the required degree of coverage.
Table 1. Examples of Monitorable Indicators
Table 2 is an example of a real-world application of the ideas presented in this paper. It is a segment of a project proposal made to the Andrew W. Mellon Foundation in late 1989 requesting funds for specific scientific connectivity tasks in the Czech and Slovak Federated Republic (CSFR), now the Czech Republic and the Slovak Republic. At that time, there was virtually no internetworking in the country and the proposal set out some specific objectives. Several characteristics of this proposal can be helpful. First, there is an explicit description of success: users, quality of service and message volumes are all shown. Second, it is recognized that even without this specific investment there would be some network activity. The reader of the proposal is not told that this investment is the only solution, but it is clear that the investment promises far greater opportunities for CSFR than would occur without it.
Table 2. Excerpt from A.W. Mellon Foundation Project Plan for Telecommunications Upgrade Proposal for Czech and Slovak Federated Republic 1989
Phase 1 Prague Implementation (July 1990-September 1991)
Includes initial node point selection, connection of major universities and research institutions and administrative steps to manage the long-range plan.
Results without proposed services (Estimate): 400-500 users and monthly message volumes of 4,000 totalling about 250 megabytes per month. 20 sites on line with plans for Phase 2 inclusion of 40 more. Relatively sparse representation of arts, economics, literature, law and other faculties frequently omitted in such plans.
Results with proposed services (Estimate): Over 1,000 users and monthly message volumes of 10,000 totalling about 500 megabytes per month. 20 sites online with plans for Phase 2 inclusion of 80 more. Relatively high representation of arts, economics, literature, law and other faculties frequently omitted in such plans.
Source: The Andrew W. Mellon Foundation.
Third, and perhaps most important, this approach gives the donor a "scorecard," a basis for assessing whether the investment really paid off. Incidentally, in the case of this proposal, all the goals were met well ahead of schedule (Ruth, 1993).
The potential of the Internet is still largely untapped in developing countries and is likely to remain so unless a comprehensive approach is taken to investments and project implementation. The authors' experiences in a number of projects in developing countries have led to what may be termed "best practices," which appears mostly as a framework within which to implement a project. Their work is by no means exhaustive, but problem areas get raised over and over. Earlier sections pointed out that the assessment of potential benefits beforehand as well as the evaluation of results afterward require criteria that make sense in the context of a particular investment program. These criteria must be useful, stable over time, quantitative or qualitative (but capable of showing measurable improvement or decline), and understandable to both lenders and institutions receiving the technology. Establishing these criteria can serve as a measuring rod for assessment; without such monitorable indicators it is difficult to assess progress objectively.
American Association for the Advancement of Science, Electronic Networking in West African Universities, 1994.
Benzine and P. Gerland, Assessing and Using the Internet, Revised Edition, August 1995. United Nations.
M. Bennett, postings on ZamNet in 1994 and 1995 and information on Zambia's WWW site.
A. Danowitz, Y. Nassef, and S. Goodman, "Cyberspace Across the Sahara: Computing in North Africa," Communications of the ACM, 38(12), December 1995, pp. 23-28.
S. Goodman, L. Press, S. Ruth and A. Rutkowski, "The Global Diffusion of Internet: Patterns and Problems," Communications of the ACM, 17(8), August 1994, pp. 27-31.
S.R. Hiltz and K. Johnson, "User Satisfaction with Computer Mediated Communication Systems," Management Science, 36(6), June 1990, pp. 739-764.
Internet Society, Internet Host Connections by Region, September 1995.
Project Proposal for Internet Connection in Ethiopia, May 1995. Copies available from Thomas Musa (thomas@padis.gn.apc.org).
P. Renaud, Guide de l'Internet en Afrique, 1994.
S. Ruth, "Measuring and Comparing the Investment on Computer-mediated Empowerment: A Developing Nations Perspective," Proceedings of the INET Fifth International Conference, Honolulu, Hawaii, June 29, 1995.
S. Ruth, "Determining Predictor Variables for Successful Entry into Electronic Networking in Developing Nations," Higher Education Policy, 6(4), Winter 1994, pp. 39-42.
S. Ruth and R. Gouet, "Must Invisible Colleges be Invisible? An Approach to Examining Large Communities of Network Users," Electronic Networking Research and Practice, 3(1), Spring 1993, pp. 34-60.
S. Ruth, "Using Powerful, Low Cost High Yield Academic Networks in Developing Nations: A Value Added Services Perspective," International Information Systems, 1(3), Summer 1992, pp. 87-99.
S. Ruth, F Utreras and R. Brecsia, "The South American Network Within One Year: An Attainable, Low Cost, High Yield Reality," Interciencia, Volume 5, September-October 1990, pp. 259-264.
A. Rutkowski, "Global Growth of the Internet," 1995, available from the Internet Society (http://info.isoc.org/).
World Bank, "Increasing Internet Connectivity in Sub-Saharan Africa, Issues and Options, and World Bank Group Role," March 29, 1995.
World Bank, "Internet Toolkit for Task Managers," Africa Region and Telecommunications and Informatics Division (IENTI), February 1996.