The Software Development Model–User Satisfaction Gap in Academic Digital Systems: Empirical Evidence from Higher Education Institutions in Goma, DRC
TL;DR
The findings argue for the design of a contextual software development framework specifically adapted to the operational, financial, and organisational constraints of HEIs in resource-constrained Central Afri-can cities.
Abstract
The digitalisation of higher education institutions (HEIs) in sub-Saharan African cities presents a para-dox rarely documented in the literature: developers report high satisfaction with their development pro-cesses whilst end-users simultaneously experience low satisfaction with the resulting software products. This article introduces and empirically quantifies the Developer–Stakeholder Understanding Gap (DSU-Gap) — a structural asymmetry between developers’ self-assessed process quality and end-users’ actual product experience — through a tri-group, mixed-methods exploratory study conducted across HEIs in Goma, Democratic Republic of Congo. A total of 175 respondents participated across three groups: 17 software developers, 22 administrative and teaching staff (agent users), and 136 students. Data were collected via a structured questionnaire administered through KoBoToolbox (November 2024–November 2025) and analysed using descriptive statistics, non-parametric inferential tests (Mann-Whitney U, binomial tests), and inductive thematic analysis. Five empirically validated tensions define the DSU-Gap: (T1) Agile is the dominant development model (65%) yet 82% of developers re-port project failures (binomial test: p = .006); (T2) 88% of developers never adapt standard models de-spite 24% acknowledging their inadequacy; (T3) 73% of end-users were never consulted during software conception (p = .026), violating core requirements engineering principles; (T4) 38% of students lack ac-cess to their institution's digital services — an internal digital divide significantly exceeding a 20% threshold (p < .001); and (T5) developers’ process satisfaction is significantly higher than users’ product satisfaction (Mann-Whitney U = 246, p = .007, rank-biserial |r| = 0.45, medium-to-large effect). These findings argue for the design of a contextual software development framework specifically adapted to the operational, financial, and organisational constraints of HEIs in resource-constrained Central Afri-can cities.