Digital government is often sold on speed and convenience, but a new study from Ethiopia suggests that the experience of using an online public service depends on more than whether the system technically works. Trust appears to be one of the mechanisms that turns some service qualities into satisfaction, while other features matter more directly.
Researchers from Addis Ababa University examined users of the Addis Ababa City Trade Bureau, where permits, registration, inspections and other trade-related services have increasingly moved online. Their analysis found that efficiency and ease of use were directly associated with higher customer satisfaction, while trust itself showed the strongest direct association among the reported paths.
The distinction matters because digital public services increasingly sit between citizens and essential administrative processes. A portal can be fast, secure and technically reliable, yet users may evaluate those qualities differently. Some features reduce effort immediately. Others appear to matter because they signal that the institution and system can be trusted.
A mixed-method study of digital public services
The study used an explanatory sequential mixed-method design. The researchers first distributed 400 questionnaires across selected districts in four Addis Ababa sub-cities. Of these, 342 valid responses were obtained, producing an 85.5% response rate. The sample was drawn from users with experience of the Trade Bureau’s e-government services.
The survey measured five dimensions of e-government service quality: security and privacy, efficiency, fulfillment, ease of use and language. Trust was modelled as a mediator between these service characteristics and customer satisfaction. Responses were measured on seven-point scales and analysed using partial least squares structural equation modelling in SmartPLS 4.
The mediation tests used 5,000 bootstrap resamples with 95% confidence intervals. The researchers also followed the quantitative phase with interviews. Twenty-six participants were interviewed before the researchers judged that thematic saturation had been reached, allowing them to examine how users described security, speed, reliability, usability and language in their own experiences.
Measurement checks were broadly acceptable after four weak indicators were removed. Reported factor loadings ranged from 0.68 to 0.88, average variance extracted ranged from 0.57 to 0.72, and the model passed the authors’ discriminant-validity tests. Variance inflation factors were below three, reducing concern that the predictors were excessively overlapping.
Efficiency mattered directly, while trust carried other effects
Efficiency had a positive direct association with customer satisfaction, with a path coefficient of β = 0.24 and t = 3.56. Ease of use also had a positive direct association, although smaller, at β = 0.13 and t = 2.05.
Trust showed a stronger direct relationship with satisfaction at β = 0.38 and t = 5.57. Together, the predictors and mediator explained 49% of the variance in customer satisfaction. The service-quality predictors explained 48% of the variance in trust. Adjusted R² values were 0.47 for satisfaction and 0.46 for trust.
Security and privacy did not have a statistically significant direct association with satisfaction at conventional thresholds. Its coefficient was β = 0.13, with t = 1.88 and p = 0.06. Yet the indirect path through trust was significant at β = 0.06, t = 2.17. In the researchers’ model, this was consistent with trust fully mediating the relationship.
A similar pattern appeared for fulfillment, meaning whether the digital service reliably delivered what it promised. Its direct path to satisfaction was negligible and non-significant at β = 0.02, t = 0.23 and p = 0.82. The indirect path through trust, however, was significant at β = 0.09 and t = 2.74.
Efficiency worked through both routes. It remained directly associated with satisfaction and also had a significant indirect association through trust of β = 0.12 and t = 3.58. That suggests faster and less burdensome digital services may improve the user experience immediately while also reinforcing confidence in the institution delivering them.
Ease of use followed a different pattern. Its direct association with satisfaction was significant, but its indirect path through trust was not. Language was statistically insignificant through both the direct and trust-mediated paths in the structural model.
Users described trust in practical terms
The interviews help explain why the statistical paths differed. Participants linked security and privacy to protection of personal and business information, confidentiality, secure transactions and protection against misuse. These features were therefore not simply technical attributes. They influenced whether users felt confident enough to rely on the digital service.
Efficiency was described in more immediate terms: shorter waiting times, faster processing, fewer procedures, less travel and quicker responses. This helps explain why efficiency could have both a direct relationship with satisfaction and an indirect one through trust. Users can feel the time saving directly while also interpreting efficient delivery as evidence that the institution is competent.
Fulfillment was associated with receiving the requested service completely and accurately. When a portal delivers what it promises, the benefit may be partly relational rather than purely functional. Reliable completion can strengthen the belief that the system is dependable, which then shapes the overall evaluation of the service.
Language produced one of the more nuanced results. It did not significantly predict satisfaction or trust in the structural model, but interview participants still described language barriers, limited language options, unclear terminology and difficulty understanding English. The absence of a statistically significant path therefore does not mean language accessibility is unimportant. It may instead operate as an inclusion issue that was not well captured by satisfaction in this sample.
What digital government can learn from the pattern
The findings challenge the idea that every desirable feature of a digital public service improves satisfaction through the same psychological route. Risk-related qualities such as security and privacy may matter because they build trust. Usability features such as ease of use can improve the experience without necessarily changing trust.
For public agencies, that distinction can influence where improvements are targeted. Faster processing, simpler procedures and clearer navigation can remove friction directly. Security, transparent handling of personal data, reliable completion and visible application status can help users judge whether the service and institution are dependable.
The result is particularly relevant as African governments expand digital public infrastructure. Moving a process online is not equivalent to improving it. If digitisation reproduces slow procedures, unclear status information or uncertainty about data protection, the interface may change without resolving the reasons users distrust or dislike the service.
Important limits remain
The study should not be read as evidence that the measured service characteristics caused satisfaction. Its cross-sectional design captured perceptions at one point in time, so the direction of the relationships cannot be established with the certainty of a longitudinal or experimental design.
The geographic areas were deliberately selected and the study focused on one public organisation in Addis Ababa, limiting how confidently the results can be generalised to other agencies, cities or countries. The measures were also self-reported and may be affected by response or social-desirability bias.
Another limitation is the absence of objective performance data. The analysis did not include administrative measures such as actual processing times, transaction failures, complaint resolution, repeat usage or abandonment rates. The authors therefore recommend combining user perceptions with operational records and following users over time in future research.
Even with those constraints, the study offers a useful reminder for digital transformation programmes. Citizens and businesses do not experience an online public service as a collection of isolated technical features. They experience whether it saves time, completes the task, protects their information and gives them confidence that the institution behind the screen will deliver.
Source Information
Study Title: E-government service quality and customer satisfaction in the Addis Ababa City Trade Bureau: the mediating role of trust
Authors: Tolasa Tafa Wami, Belew Dagnew Bogale and Alemu Azmeraw Bekele
Journal: Future Business Journal
Published: 30 September 2026
Method: Explanatory sequential mixed-method study using 342 valid survey responses, partial least squares structural equation modelling, 5,000 bootstrap resamples and follow-up interviews with 26 participants
Main finding: Efficiency and ease of use were directly associated with satisfaction, while trust mediated the relationships involving several service-quality dimensions, including security and privacy and fulfillment.
DOI: 10.1186/s43093-026-01005-2
