DYNAMICS OF INFORMATION TECHNOLOGY (IT) SUCCESSFUL IMPLEMENTATION IN DEVELOPMENT COUNTRIES: A NIGERIAN CASE STUDY

INTRODUCTION

An important question facing managers and designers of computer-based information systems (CBISs) is why technically elegant systems are not adequately and gainfully utilized by their intended end-users. Anticipated Information Systems (ISs) benefits in terms of reduction in total-cost-of-ownership (TCO), increased user productivity output and value-added enhancements to products and services do not seem to have been fully realized. This is the phenomenon aptly described as "systems being technical successes, but organizational failures" in IS circles and among IT professionals. While investments in IT are rising, the return on investment (ROI) as well as whitecollar and knowledge worker productivity gains are, as yet disappointing. This situation is truer for developing countries, particularly for those in Africa (9, 26). Literature in Management Information Systems (MIS) has identified broad categories of factors affecting the implementation of computerized information systems. Research in information systems failure has suggested that far too much emphasis has been placed on the technical aspects and that the primary cause of failure is the lack of consideration ascribed to the social and behavioral dimensions of the implementation process itself (12, 19). A further review of the literature (1, 25, 34) suggests, that while certain social and behavioral factors may influence IT acceptance and institutionalization, there is little research to date on the relative importance of similar factors connected with mission-critical information systems, especially in the context of developing countries.

BACKGROUND

The introduction of computing and specifically Information and Communications Technologies (ICTs) into the workplace is one of the most significant and dramatic changes in the realm of office work to be witnessed in recent years (23). Nigeria is not an exception in terms of this "wind of change" that is blowing across workplaces throughout the world. For example, a pre-test carried out in Nigeria prior to this study revealed that among organizations sampled, 72% engaged in the utilization of IT in one form or the other, 66% operated in a client-server computing environment, 11% had minicomputers and/or specialized workstations, while 9% still used mainframe computers in a hybrid setup and configuration with other equipment. These figures suggest among other things, that computer connectivity, networks, etc., and their deployment are becoming relatively widespread in Nigeria, and by extension, in most of Africa. However, despite the proliferation of microcomputers and workstations in the office environment, organizational embrace of IT can still be described as an innovation. An innovation is defined as "an idea, practice, or object that is perceived as new by an individual or its unit of adoption" (30). Inasmuch as the implementation and application of IT as presently practiced in Nigeria have been described as haphazard (1), it is worthwhile to study the inherent variables that if properly analyzed and harnessed, may begin to deliver IT's limitless potentials and capabilities.

STATEMENT OF THE PROBLEM

Although it is widely perceived that IT innovations, particularly in developing countries, are technical successes but organizational failures (12, 25), only a few studies have been conducted to investigate empirically why this is so. As a practical matter, and to break the vicious cycle of unsuccessful implementation outcomes, organizations have reportedly begun to take giant strides in this direction. Despite the good and bold intentions, the age-long conclusions regarding problems facing computerization in developing countries and especially in Africa as articulated by a pioneer researcher (34) remain substantially true:

...while the goal of the use of advanced technology in developing countries is to accelerate development, these countries usually lack the skills and procedures that underpin the technology ...

Apart from the issue of missing skills, other situational factors tend to bring about additional problems. Developing countries constantly encounter problems unknown to their developed counterparts (4). Some authors speculate that the main reason for the unfavorable outcomes might be that in developing nations, the problems associated with computing have been aggravated by socioeconomic problems induced by the development process itself (1).

Often, developing countries focus only on the technology, and fail to realize that organizational factors and the process followed in the actual implementation of technology are related in a direct way, to successful outcomes. The problem investigated in this study relates to the analysis of such implementation-related factors and more specifically, to determine empirically, whether organizations with varying implementation strategies ended up with different degrees of successful outcomes.

PURPOSE AND OBJECTIVE OF THE STUDY

The main purpose of this study is to assess the relationship between successful implementation and the technological, contextual and implementation process factors. Once the dominant variables are isolated and their dynamics sufficiently explained, a better-informed implementation strategy can be developed for future installation, configuration, management, and utilization of IT. This should have great value for IT practice in Nigeria and in comparable developing countries, and particularly in the African workplace.

The objective of the study is to test the implementation meta-model as depicted in Figure 1. In this connection, two main research questions are investigated.

The research questions and the corresponding hypotheses are as follows:

Research question 1 is: What are the relationships among technological, contextual and implementation process factors and implementation success?

Hypothesis 1 is: Technological validity, contextual validity, and implementation process validity are all positively related to successful implementation.

Research question 2 is: What contribution does implementation process validity make to the explanatory power of the set of independent variables?

Hypothesis 2 is: When compared with techno-logical and contextual variables, implementation process validity will account for most of the variance in explaining implementation success.

CONCEPTUAL FRAMEWORK

The conceptual framework underlying research on implementation draws upon a variety of major topic areas: (a) organizational theory and development, (b) management science, (c) theories of organizational change, (d) sociotechnical systems theory, and (e) technological innovation process. This research derives for the most part from processbased perspective of implementation (1). Although it has not been exhaustively tested, the "three-variable" model - the iulcrum of this study, provides a reasonably comprehensive view of the implementation phenomenon. This model's focus is on process rather than on state variables. Also, given that most implementation activities occur between an organization's decision to adopt a technological innovation and occurrence of outcomes (i.e. the process-driven phase), the model appears to be the most suitable theoretical framework available. The study's emphasis on implementation strategy is in consonance with a large body of related research in MIS (1, 8, 23,).

Three classes of independent variables and one dependent variable constitute the research model. The independent variables are: (a) technological validity; (b) contextual validity; and (c) implementation process validity. The last of the three is expected to contribute the strongest influence considering the results of related past studies. The fourth variable, which is the dependent variable, is denoted by implementation success.

As each of the conceptual variables can be quite general as expressed, it became necessary to specify "how exactly" they are defined or what operations are required to achieve them. The following are the set of dependent and independent as well as their respective factors and indicators.

DEFINITION OF VARIABLES

The set of independent variables considered, measured and treated in this study falls into three conceptual categories: technological validity, (i.e., technical and social features of implemented IT), contextual validity (i.e., features of the organizational environment surrounding the implementation) and implementation process validity (i.e., features of the implementation process itself). The variables were selected, among other reasons, because of their theoretical importance as well as their potential relevance to practice. Moreover, responses given to open-ended questions in the pre-test indicated that these variables make sense and are quite meaningful to respondents.

Independent Variables: The independent variables with corresponding indicators for each subcategory are itemized as follows:

Technological Validity - relative advantage, compatibility, divisibility, access to a workstation, interaction support and functionality of the system.

Contextual Validity - linkage with multinationals, economic sector, centralization, computing infrastructure, formalization, IS unit's size, degree of professionalism, and management support for innovation.

Implementation Process Validity - degree of specificity of objectives, support of key actors, user participation, vendor involvement, mode of installation, change orientation, sociotechnical balance, training or support for learning, mutual adaptation, and IS planning.

Dependent Variable: Implementation Success - The dependent variable denoted by implementation success is conceptualized as consisting of many indicators belonging to three variable sub-categories namely - use of system, performance effects on tasks, and user satisfaction with the system.

The Research Model

The research model for studying the dynamics of IT successful implementation outcomes, as influenced by different relevant conceptual/structural factors is presented in Figure 1.

Much of the following literature review is used in laying the foundation from which pertinent factors are drawn (15, 19, 22, 28). As the conceptual model above outlines, three classes of variables determine the successful implementation of information technology innovations in organizations. Such variables are comprised of resulting zero-order technological, contextual and implementation process factors. Moreover, implementation success, as the dependent variable, is denoted by three indicators namely: use of the system, impact on user job performance, as well as overall user satisfaction with the technology.

REVIEW OF RELATED LITERATURE

The installation and implementation of computer-based information technology by organizations the world over have been profound and pervasive since the early 1960s. Both large and small organizations have over the years invested relatively large amounts of money and other resources in computer-based information systems.

As suggested by much earlier research (12), the push toward "universal" acceptance of these systems and the need to utilize them appropriately have evidently accelerated throughout the late 1990s as well as into the early part of the 21st century. The manifestation of this phenomenon is especially true in the economically developed countries of the world. Other studies (6, 10, 11, 17, 24, 33) also have suggested that in the early part of this century, IT will be available to nearly all knowledge workers in such countries and will play a dominant role in large organizations. The experience of most developing countries (DCs) is rather different. While computer-based information systems are becoming increasingly popular among them (1, 9), the understanding of how to successfully implement such systems lags behind the understanding of the technology itself.

In recent times, organizations in developing countries (DCs) have been investing quite aggressively and extensively in information and communications technologies in appreciation of their developmental potentials, and with a view to becoming part of the global economy (2, 3). Thus, DCs' organizations are willing to exploit these potentials to achieve productivity improvement, organizational effectiveness and business competitiveness (2).

ICT applications are increasingly being applied both in public and private sector organizations in DCs. Recent studies have revealed an increasing adoption of the technology in business sectors such as banking where, in addition to automating back office functions, banks are in large measures, investing in customer-facing technologies such as ATMs and smart-card technologies (3, 4, 13, 14, 37). Other studies have also expounded on the application of IT in health institutions in a number of DCs (6, 18, 24) and increasingly in the provision of government services such as the revenue collection and financial management information systems in Uganda (38), property tax reforms in India (21) and other various e-government initiatives (11). Similarly, ICTs by way of the Internet and wireless communication technologies are being used in providing access to education and much more, for the poor and the rural communities around the globe (7, 29).

Despite the increase in investment, the use of ICTs in DCs has, over the years, been suggested as not living up to the expectations of their implementation and the return on ICT investment in these regions is said to be generally poor (20). It is apt to reference Heeks (2002) who identified four types of implementation failure: (a) total failure of a system that never works; (b) partial failure where major implementation goals were never realized; (c) sustainability failure where the implementation succeeded initially, but eventually became a failure, and (d) replication failure of a pilot project that could not be reproduced (11). All types of implementation failures in varying degrees and combinations, are known to have presented themselves in developing nations of the world. Other authors have argued that an effective exploitation of ICT potentials lies not just in putting an expensive infrastructure and access in place, but more importantly in paying close attention to local factors in the context of implementation (5). Of relevance also is the development of ICT-related human capabilities and incentives to use information effectively (20, 27).

Similarly, Sahay and Avgerou (2002) and Thatcher, et. al. (2003) suggested that an understanding of local improvisations and work practices are important for information technologies to be effectively integrated into organizational practices in DCs (31, 33).

These suggestions imply the view that solutions and management techniques developed from the industrialized countries of the 'west' cannot be mechanistically implemented in DCs with the expectation of achieving similar and comparable results as in the 'west.' The foregoing arguments have hence called for an understanding of local contextual factors and how these can lead to a successful implementation of IT in DCs. It is therefore reassuring that deliberate efforts are being made to adopt context-sensitive strategies in bringing rural communities around the world to join the global online community (29). Such time-honored and well-informed implementation strategy cannot but lead to more successful implementation outcomes.

RESEARCH METHODOLOGY

Research Design

The overall research plan consisted of two main steps: step one, which was a sample survey, emphasized statistical inference, and step two, which was personal interview, emphasized qualitative data.

Sample Design

The initial sample of participating organizations was drawn from the directories of institutional members of:

1. Nigerian Computer Society (N.C.S.), and

2. The Federal Ministry of Trade and Industry, Computer Users Group (C.U.G) register.

These two sources represent the most current lists of computer installations (in both the public and private sectors) within the country. They constituted the primary sample frame. Since the N.C.S and the C.U.G registers are established by law, one would expect the entries to be fairly exhaustive and up-to-date. The two sources of data were used to create a unified list of sufficiently large IT-dependent organizations.

DATA COLLECTION

The data gathering method employed was a cross-sectional field survey using a variety of approaches as well as datagathering sources. "Multiple operationalism" - the use of alternative measures that attempt to tap the same phenomenon (35) - was used to serve as a validity check on the operational measures for different variables. In this connection therefore, data gathering involved a variety of instruments and data sources, in compliance with the principle of triangulation (32). These were surveys administered to relevant users within the various units of the organizations (an omnibus questionnaire, designed to explore several issues related to technological, contextual, and implementation process factors, and success variables). Personal interviews augmented the questionnaire. In general, the variables were operationalized through selfadministered questionnaires that were hand-delivered or mailed to each of the participating organizations. The scales used in the measurement process included indexes of implementation success, technological, contextual and implementation process variables. By and large, all the variables were measured with a variety of fill-ins, check-offs and scaled responses. Questionnaires were sent to all industrial sectors, in roughly the same proportion as the sector's share of the total Nigerian economy. One hundred and fifty two organizations were contacted; 137 agreed to participate in the study. One hundred and eight organizations returned 242 usable questionnaires out of 361, representing 64% return rate.

Table 1 shows a breakdown by economic sector of the 108 organizations that fully participated in the study (sample size, n = 108).

The industry-category breakdown of the sample is extremely varied, covering five of the six identified areas. Although more than half of the organizations belong in the financial and insurance industries, the conclusions derivable from their varying experiences should be instructive. The organizations surveyed are predominantly privately owned businesses with 63% belonging to this group. The remaining 37% is in the public sector. The majority of them conform to the organizational-size requirement as set forth before the commencement of the research. Nevertheless, the parity in the number of years of IT acquisition and the operational experience between right-sized and other smaller organizations made it possible for the smaller organizations to be included in the study. While larger organizations (having 75 to 100 employees) are in the majority (about 81% of the total), their working experience with IT seems to be surprisingly marginal. Mean years of IT experience stood at 11.2 years. In contrast, smaller organizations generally seem to have had a relatively richer and longer experience with IT. The reported average years of IT experience for the group was 15.7 years.

The distribution of computer processing power is diverse. It is interesting to note that nearly a quarter of the organizations employ some kind of mini/micro-computer mix to carry out necessary work. Two-thirds use microcomputers exclusively, while some negligible proportion uses a combination of all grades of computers. The data revealed that most IS groups (73%) use systems from two or more manufacturers. IBM machines and IBM clones are most popular. Connectivity, exemplified by client/server computing, is a phenomenon that is becoming increasingly popular among the Nigerian business community. Nearly 60% of the sample reported having their systems networked or are in the process of doing so.

STATISTICAL AND DATA ANALYSIS

The original intention was to test the entire model using canonical correlational analysis; as the objective was to derive a summary statement of the relationship between some combination of the predictor set of variables and a combination of variables in the criterion set, i.e. many-to-many patterns of association. The preliminary factor analysis, leading to the reduction of the dependent variables to one, rendered this approach unnecessary. The strategy then became one of estimating least squares regression for each of the variables remaining in the model. Multiple regression therefore, became a method of interest for data analysis.

The merit of this technique is that it provides a unique partitioning of the variance associated with different sets of independent variables. Moreover, it allows tests of explicit hypotheses about variance attributable to certain variables after others assumed to be prior have been accounted for. Having determined the minimum number of independent variables necessary to reproduce the variation in the original data sets by transforming the data through a factor analytic procedure, a regression model consisting of three diagnostic variables and one dependent variable was hypothesized. The independent variables were introduced in three blocks. Model I F-tests of significance were used to assess the change in R^sup 2^ resulting from the addition of each new set of predictors.

RESULTS

The summary statistics describing the levels of association between the resulting indices are presented and discussed in this section. Table 2 highlights the zero-order correlations between the resulting indices. From the table, it can be gleaned that two of the three indices (CONT. and TECH.) are relatively moderately related to the level of implementation success observed and that significant relationships exist only between contextual and success variables. The factor with the greatest impact is contextual validity, while technological validity ranks next to it. It is interesting to note the negligible impact of the implementation process validity. Overall, the predictor variables are positively associated with success. Hypothesis 1 states that technological validity, contextual validity, and the implementation process validity are positively related to successful implementation.

As Table 2 shows, all three validities are positively correlated to implementation success (although technological validity and implementation process variables are not significant at p = 0.05), and all parameter estimates of the independent variables are positive (i.e., the direction of the relationships indicated by the beta coefficients in Table 3 is consistent with the research expectations). Hence, all validities are necessary for successful implementation of Information Technology. On the basis of this analysis therefore, Hypothesis 1 partially is supported.

As stated above, multiple regression was used to test the second hypothesis as well. Table 3 shows the results of the regression of implementation outcomes on the predictor variables. It shows the standardized regression coefficients (ß), adjusted coefficients of determination (R) and squared semipartial correlations ([black triangle up]R^sup 2^) after the inclusion of each block of variables. After all the blocks had been incorporated in the model, F*(3,104)-1.904.

FACTORS PROMOTING SUCCESSFUL IMPLEMENTATION OF INFORMATION TECHNOLOGY

The results of data analysis for hypotheses 1 and 2 suggest that indeed, only a few predictor variables (particularly the contextual variables) contribute to successful implementation to some degree. Regression analysis for the model yielded a value of F of 1.904. Since F* = 1.904 is less than 2.68 (the test statistic); df (3,104)), the null hypothesis (that the beta coefficients are 0) is not rejected, and the sample result is deemed statistically insignificant at the 0.05 level. The set of contextual validity characteristics explained the highest amount of variation in successful implementation followed by the impacts of technological and process factors respectively, although the last two were left almost out of account.

In addition, all three successful implementation outcomes were readily predicted; the strongest finding was the industrial sector being a critical predictor of the use of the system, satisfaction with the system and productivity gains resulting from the use of the system. The emergence of the economic sector as a good predictor of favorable outcomes is consistent with existing theory that organizations respond to their preordained environmental conditions (36). After all, in Nigeria, the private sector is expected to be the pacesetter and technology champion in innovative and technological advancements. Qualitative analysis of interviews conducted as a follow-up to the study corroborated, in part, these findings.

SUMMARY AND CONCLUSIONS

The results, as with similar implementation studies, are interpretable in multiple ways. First, they are a test of the research model, a possible clue to implementing information technologies in developing countries with particular reference to Nigeria. The results demonstrate at least some model support for a tangible portion of the research model. That is, there is partial support for the hypothesis that some antecedents of implementation in terms of the technology, the context and the strategies contribute positively to success.

The overall conclusion of this research is that contextual factors (such as industrial sector, level of leadership support for technology, degree of professionalism as well as the level of professionalism within an IS unit) contribute considerably to the complex reality of technology implementation in Nigeria. To the extent that inferences could be drawn about the population from the sample, these conclusions seem to portray a picture of an industrial sector precepts and examples, the consequence of which is the need to focus on, and learn more from what "best practices" firms are doing right in institutionalizing their information technologies. Also, it is conceivable that the results of this study should be of interest for what they reveal about the nature and issues concerning implementation of technology in a developing country as well as their general implications for implementation research.