Easy-to-use water system boosts revenue in Tanzania, engagement doesn’t matter

When employees at Tanzanian public water utilities find the national water information system easy to operate, the financial and operational performance of their organisations improves measurably, according to a new study published in the journal Discover Analytics. The research, led by Peter Raymond Ngunula of the Open University of Tanzania with colleagues Joseph Magali and Juma Matonya, examined the MAJI Information System, a digital platform rolled out across Tanzania’s public water utilities to strengthen billing, revenue collection and accountability. The study found that employees’ perceptions of how effortless the system is to learn and use — a construct known in technology-acceptance research as effort expectancy — was a powerful predictor of revenue collection, service-level effectiveness and efficiency in managing non-revenue water. Strikingly, employee engagement, which the researchers had expected to amplify these effects, turned out to have no significant moderating influence at all.

The study is notable because it pushes a well-established theoretical framework, the Unified Theory of Acceptance and Use of Technology, or UTAUT, beyond its usual territory. UTAUT, developed by Venkatesh and colleagues, identifies four determinants of technology use: performance expectancy, effort expectancy, social influence and facilitating conditions. Most research built on this framework, particularly in developing economies, has stopped at behavioural intention and system adoption, asking whether people will use a technology and how much. Whether that usage actually translates into organisational-level outcomes such as revenue, service quality and loss reduction has rarely been tested, especially in public utilities in low-resource settings. This study set out to fill that gap by linking a core UTAUT construct directly to financial and operational performance indicators.

The context makes the question urgent. The authors note that in Tanzania more than 40 percent of water is lost through leakages, theft or faulty metering, meaning the country’s non-revenue water level is roughly twice the recommended 20 percent threshold. Rural areas additionally suffer from poor digital infrastructure and inefficient billing systems. MAJI IS was implemented as a key reform tool to address exactly these problems, yet many utilities continued to experience low revenue recovery, inaccurate reporting and persistent water losses, suggesting a disconnect between system utilisation and financial performance. Understanding what drives that disconnect was the central motivation for the research.

Methodologically, the study adopted a positivist philosophy and an explanatory cross-sectional survey design. Data were collected between July 10 and November 30, 2025, using structured questionnaires administered to 235 staff involved in water service delivery at public water providers in the Dodoma and Katavi regions. The researchers used a multistage sampling technique: Dodoma and Katavi were purposively selected for their differing operational characteristics, ten water utilities were then randomly chosen — six from Dodoma and four from Katavi — and respondents were selected through stratified random sampling based on professional roles. The sample size was justified both by Yamane’s formula and by a prospective power analysis in G*Power 3.1, which indicated a minimum of roughly 77 observations for a three-predictor regression model at medium effect size, so the achieved sample comfortably exceeded the requirements for structural equation modelling.

Analysis was performed with Partial Least Squares Structural Equation Modelling using SmartPLS 4.1.1.8, a technique the authors chose for its suitability in prediction-oriented research and its flexibility regarding distributional assumptions. Measurement quality was assessed carefully: indicator reliability, internal consistency, convergent validity through average variance extracted, and discriminant validity through heterotrait-monotrait ratios all met recommended thresholds, with effort expectancy’s Cronbach’s alpha of 0.680 judged acceptable given its composite reliability of 0.806. Because all variables came from self-administered questionnaires at a single point in time, the team also tested for common method bias using Harman’s single-factor test and full collinearity variance inflation factors, both of which indicated the problem was not serious.

The structural model results were emphatic. Effort expectancy positively and significantly predicted revenue collection performance, with a path coefficient of 0.719, a p-value below 0.001 and an effect size of 1.084 — well above the conventional benchmark of 0.35 for a large effect, and unusually high even for PLS-SEM studies, though the authors note that comparable values have been reported when a dominant predictor explains most of the variance in an outcome. The mechanism, they argue, is straightforward: when employees find MAJI IS easy to use, they can issue bills, process payments and manage customer accounts efficiently, which improves the accuracy and timeliness of billing information and accelerates revenue collection. Customers, in turn, can pay bills more promptly than with manual systems that may require travelling to a payment centre.

The same pattern held for the other two performance outcomes. Effort expectancy significantly predicted service-level effectiveness, with a coefficient of 0.664 and an effect size of 0.788, suggesting that ease of use translates into faster responses to customer requests, better information accessibility and more efficient service delivery. It also significantly predicted non-revenue water efficiency, with a coefficient of 0.672 and an effect size of 0.814. Here the authors point to the system’s capacity for monitoring water distribution, identifying leakages, managing meter information and coordinating maintenance, including notification and GPS-based tracing of leakage or water malpractices. Across all three outcomes, the effect sizes exceeded the large-effect benchmark, leading the researchers to characterise effort expectancy as a dominant determinant of organisational performance within the MAJI IS environment rather than merely a precursor of system adoption.

The surprise came with employee engagement. The hypothesised moderating effects of engagement on all three relationships — with revenue collection, service-level effectiveness and non-revenue water efficiency — were not statistically significant. In other words, the positive impact of effort expectancy on performance remained comparatively constant regardless of how engaged employees were with their work. The authors offer several possible explanations, all carefully flagged as interpretations rather than findings. One is that once a system is perceived as easy to use, its benefits may be strong enough to affect outcomes without additional support from engagement. Another is that MAJI IS use in Tanzanian utilities may be driven by organisational policies, standard operating procedures and mandatory usage requirements, which the UTAUT literature suggests can dampen the influence of individual differences — though mandatory use was not directly measured. A third possibility is that the utilities operate in highly regulated environments where infrastructure and institutional factors outweigh employee attitudes, and a fourth is that the system has become so institutionalised and routine that engagement simply no longer matters. None of these mechanisms was measured directly, and the authors explicitly caution that these explanations remain speculative.

The study’s limitations are acknowledged candidly. The cross-sectional design cannot capture how relationships evolve across stages of technology implementation. The focus on utilities and community-based water supply organisations under the Rural Water Supply and Sanitation Agency framework may constrain generalisability to other regions, utilities or sectors. The quantitative approach excluded qualitative insight, and the model examined only a narrow set of factors, omitting potentially relevant influences such as regulatory enforcement, system quality, interoperability and environmental conditions. Employee engagement itself was measured at a single point in time. The authors also stress that the study should not be read as proposing an extensive extension of UTAUT; it retained the core constructs and simply tested their relationship with performance outcomes without modifying the theory’s assumptions.

For practitioners, the message is unambiguous: investing in usability pays. The authors recommend that public water utility managers emphasise system usability, continued support for technology users and capacity-building creativity when implementing digital technologies, and that investments be accompanied by simplified interfaces, reduced operational complexity and improved user experience, along with robust monitoring and evaluation mechanisms. At the policy level, they call for stronger coordination among sector institutions, including the Rural Water Supply and Sanitation Agency and the Energy and Water Utilities Regulatory Authority, to align digitalisation initiatives with utility operations and promote standardisation. Future research, they suggest, should adopt longitudinal designs, compare urban and rural or large and small providers, and incorporate additional organisational and institutional variables such as regulatory support, digital governance and infrastructure readiness, ideally with mixed methods that capture the lived experience of system implementation.

Subject of Research: How perceived ease of use of a water information system affects organisational performance in Tanzanian public water utilities

Article Title: Effort expectancy influences on water information system usage in Tanzania with financial implications and the moderating role of employee engagement

Article References: Ngunula, P. R., Magali, J., & Matonya, J. (2026). Effort expectancy influences on water information system usage in Tanzania with financial implications and the moderating role of employee engagement. Discover Analytics, 4(1), Article 45. https://doi.org/10.1007/s44257-026-00093-2

Image Credits: AI Generated

DOI: 10.1007/s44257-026-00093-2

Keywords: Tanzania, water utilities, MAJI Information System, UTAUT, effort expectancy, employee engagement, non-revenue water, revenue collection, PLS-SEM, digital transformation, public sector, developing economies