Sonke Gender Justice is one of South Africa's most significant civil society organisations working on gender equity, men and boys' engagement, and the reduction of gender-based violence. Their programmes operate across multiple communities and geographies, involving a large field workforce, complex beneficiary relationships, and ongoing obligations to report meaningful impact data to funders and partners.
Organisations like Sonke live and die by the quality of their data. Funders want rigorous evidence. Programme managers need to understand what is working and what is not. Leadership needs to allocate resources to where they are creating the most impact. All of this depends on having data that is accurate, complete, accessible, and reportable — and a system that makes it easy to keep it that way.
When Sonke approached GEMIS, they had data. What they lacked was a coherent system to manage it.
The challenge: data spread across systems that could not speak to each other
Sonke's monitoring and evaluation function is central to the organisation's credibility and effectiveness. They track beneficiary engagement across programmes, record activity data from field teams, monitor progress against log frame indicators, and produce regular reports for funders and management. The data volumes involved are substantial, and the data relationships are complex — a single beneficiary may be touched by multiple programmes, across multiple sites, over several years.
Legacy data that couldn't be abandoned. Years of programme records existed in legacy formats — spreadsheets, older databases, and manual records — representing a significant organisational asset that could not simply be left behind. Any new system had to absorb this history, not start from zero.
A data model complex enough to reflect real programme logic. The relationships between beneficiaries, programme activities, staff, sites, and reporting periods in Sonke's work are genuinely complex. A generic off-the-shelf database would not capture these relationships accurately, leading to the same reporting limitations that already existed.
Decision-makers who needed answers without needing to be data analysts. The goal was not just a better database — it was a system that could surface meaningful insights to programme managers and leadership without requiring them to write queries or export data to Excel to make sense of it.
How GEMIS approached it
The engagement followed a structured process that began with deep engagement with Sonke's M&E team and programme managers before a line of code was written. Understanding the data model required understanding the programmes themselves — how beneficiaries moved through them, how activities were recorded in the field, how reporting obligations were structured, and how the organisation actually made decisions using data.
Data model design
Working with Sonke's M&E team to map the full entity model — beneficiaries, programmes, activities, sites, staff, and the relationships between them — before committing to any database structure. This phase determined what questions the system would be able to answer, which is the only meaningful way to evaluate a data model.
Legacy data audit and migration strategy
Assessing the state of existing data — its formats, its completeness, its consistency — and designing a migration approach that would bring it into the new system accurately. Data migration is frequently underestimated in systems projects. For Sonke, the integrity of their historical programme data was non-negotiable, and the migration received the same rigour as the build itself.
System build
Building the database management system on Azure SQL, with a structured data capture interface that reflected Sonke's actual data collection workflows. The system was designed for the people who would use it every day — field coordinators, data capturers, and M&E officers — not just for the administrators who would manage it.
Performance dashboards
Deploying Power BI dashboards that connected directly to the live database, giving programme managers and leadership real-time visibility of key performance indicators without any manual report compilation. Dashboards were designed around the specific decisions Sonke's management team needed to make — not generic charts.
Transition support and ongoing assistance
Providing active support through the transition period — working with Sonke's team to ensure adoption was genuine, that the system was being used as intended, and that issues surfaced during initial operation were resolved quickly. The system's value is only realised when people trust it and use it; that trust is built in the transition period, not the build.
GemIS exhibited a deep understanding of our organisation's unique needs — their tailored approach to designing and implementing the Database Management System exceeded our expectations.
— Jacob Segale, Research, Monitoring, Evaluation & Learning Unit Manager, Sonke Gender JusticeWhat the system delivers for Sonke
The most significant outcome is one that is difficult to quantify but immediately felt by anyone who has managed M&E data in a large civil society organisation: the shift from spending time on data management to spending time on data use.
Before the system, Sonke's M&E team spent a significant proportion of their time on the mechanics of data collection, consolidation, and reporting — time that was not available for the analysis and sense-making that is the actual purpose of an M&E function. After, the mechanics are handled by the system, and the team's time is freed for interpretation, for conversations with programme managers about what the data means, and for the kind of evidence-based decision-making that makes programmes more effective.
Unified data architecture
All programme data — historical and current — lives in a single, coherent system with consistent entity relationships. The fragmentation that previously made cross-programme reporting almost impossible is gone.
Live performance dashboards
Management and programme leads have real-time visibility of progress against targets, beneficiary reach, and activity completion — without waiting for a monthly report to be compiled manually.
Faster, more reliable reporting
Funder reports and management packs that previously required days of data consolidation can now be generated in a fraction of the time, with greater confidence in their accuracy.
Institutional knowledge preserved
Years of programme history that existed in ageing and inconsistent formats has been brought into the system accurately — preserving an organisational asset that would otherwise have continued to degrade.
Why this kind of project is different
NGO and civil society technology projects have specific characteristics that make them different from commercial SME work, and that require a developer who understands the environment.
The first is the nature of M&E data itself. Monitoring and evaluation in the development sector involves complex logical frameworks, indicator hierarchies, disaggregated reporting requirements, and multi-year tracking of programme participants across sites. This is not standard database complexity — it reflects the genuine intricacy of the programmes being tracked, and it demands a data model that is built with that intricacy in mind from the start.
The second is the accountability context. Civil society organisations working with institutional funders operate under significant reporting obligations. Data that cannot be demonstrated to be accurate and complete is a reputational and compliance risk, not just an operational inconvenience. The standard of rigour required in the data management system is correspondingly high.
The third is the resource context. NGOs typically operate with constrained technology budgets relative to the complexity of what they are trying to manage. Getting the scoping right — building what is actually needed, rather than everything that might be nice to have — is particularly important in this environment. It requires a developer who will have that conversation honestly rather than scope-creeping towards a larger project.
GEMIS has worked with civil society organisations — including Sonke Gender Justice and Tore's Foundation — long enough to understand these dynamics. The systems we build for NGOs reflect that understanding.
We understand your data environment
M&E systems, donor reporting platforms, and beneficiary databases for the development sector require specific expertise. Start with a Technology Fit Assessment — a structured half-day session that maps your needs and identifies realistic options, without commitment to a build.
Start the conversation →