September 2026

When the administration clerk at a rural secondary school in KwaZulu-Natal left, the principal had to take on everything herself, from leading the school to capturing its data. Emails kept arriving from the Data Driven Districts (DDD) Dashboard, which turns the data schools already capture into regular updates on learners’ attendance and results. With no one to share the load, for a long time she didn’t open them. Data felt like one more administrative task.

Then, at a training session, her school’s data was projected for everyone to see: 100% learner attendance, teachers always present, every task submitted on time. She knew it wasn’t true. Without a clerk, the school’s data hadn’t been captured properly, and nobody had been checking.

She could have looked away. Instead, she looked closer. She asked for access, learned the dashboard, involved her staff and rebuilt the school’s records. Absence alerts she once had no time for became reasons to phone parents. What changed her school was her decision to look honestly at the data.

Where the gap really is

South Africa does not have a data problem. Every term, schools capture attendance, marks and progression in the Department of Basic Education’s school administration and management system, SA-SAMS. The challenge has always been getting that information to leaders in time to act on it. One official described the old way: data went from school to circuit, circuit to district, district to province. “By the time you get it, it is too late.”

That official was one of 77 interviewed across six provinces for an independent review of the Data Driven Districts programme by the behavioural science organisation ideas42, published in 2026. The review also surveyed more than 3 000 education officials and compared the programme with seven similar initiatives around the world. Together with two years of provincial intervention reports, including where interventions didn’t work, it points to three lessons.

1. Responsiveness is built into routines

Timely data matters. In the review, 90% of dashboard users had taken an action or made a decision after their most recent visit, and 84% said they trust the data.

But timely data isn’t enough on its own. In the Eastern Cape and Limpopo, provincial leadership treats the dashboard as a core planning tool, and 95% or more of the previous term’s data is consistently in before the new term begins. Districts there can plan support in the first week of term, not the fifth. Where submissions run late, the window to act closes, however good the technology.

2. Service comes before accountability

The same data can be used in two ways. Used for accountability, it flows up the system and asks, “Who is failing?” People respond the way people always respond to being judged: they make the numbers look good. That is how a school ends up showing 100% attendance.

Used as a service, data flows to the person closest to the learner and asks, “Who needs what, and when?” Because bad news leads to help rather than punishment, people are willing to surface it early. Public education needs accountability, but the order matters. When service comes first, accountability follows, because the numbers are true. As one circuit manager put it: “The conversations are evidence-based now. It’s not about opinions anymore.”

This is where equity lives, because it changes who gets seen. Equality gives everyone the same thing. Equity asks which barriers hold particular learners back from reaching their potential, and works to remove them. Finding those barriers is what the data is for.

In 2025, working with provincial departments in schools identified for support, our teams followed about 2 000 Grade 12 learners through the year. Around 40% had been progressed to Grade 12 without meeting the previous year’s requirements. Of learners who had been promoted, 55% passed the National Senior Certificate, the Grade 12 school-leaving exam. Of those who had been progressed, 28% did.

Those learners were visible in the data a full year before their results. An accountability system sees them in December, as a pass rate. A service system sees them in February, as individual learners who need something now.

At one Gauteng primary school, a principal noticed that Grade 6 social sciences results had been low for two years. Looking closer, she found the subject had never been taught by a trained teacher, and reassigned it. By her account, the Grade 6 average rose from 49% to 78% between 2023 and 2025. The data didn’t tell her what to do. It told her where to look.

3. The difference is leadership

Does this change results? The honest answer is yes, with caveats. In 2024, 23 of 26 districts targeted by a year-long Grade 12 intervention drive met the 85% pass rate target. Among the districts receiving district-level support, six of seven reached it, up from one of seven the year before. But national results also rose that year, so not all of that improvement is ours. In 2025, when the work reached into primary schools for the first time, results were mixed, and they were strongest where district leadership was strongest.

Data-driven improvement is possible. It is not automatic. The review and our own work point to five leadership practices that make the difference:

1. Ask for the data first. Officials are far more likely to use the dashboard when their own leaders encourage it. Yet only 24% say their supervisor actively promotes it.
2. Make it safe to show bad numbers. If a true number is punished and a false one praised, you will get false numbers.
3. Build habits and peers, not one-off events. The programme has trained more than 260 Master Trainers (officials who train and support their colleagues) in two years, but habits fade when leaders stop reinforcing them.
4. Close the loop. Only 15% of officials use the dashboard to check whether an intervention actually worked.
5. Pair evidence with professional judgement. The dashboard shows where to look. Experienced educators decide what to do.

What still stands in the way

None of this scales fairly unless we are honest about the obstacles. In the review, nearly three in ten officials had no or unreliable internet, one in five said they lacked time to check the dashboard and follow up, and almost half reported only moderate or lower comfort using it. The least connected schools are often the ones that most need the insight. If that isn’t addressed, data-driven leadership becomes one more advantage for schools that already have advantages.

There is also the question of sustainability. After more than a decade, full government ownership of the DDD Dashboard is still a goal in progress. It depends on aligned funding, dedicated capacity inside the system and clear accountability, and that work continues with the Department.

Looking earlier

For years, the system focused its attention on Grade 12. That is changing, as the Department puts foundational literacy and numeracy first. In 2025, provincial interventions reached Grades 6 and 9 for the first time, and officials are now asking for simpler tools for the Foundation Phase (Grades R to 3). The review highlights two moments when learners are most likely to fall behind: the move from Grade 3 to Grade 4, when the language of instruction changes, and from Grade 7 to Grade 8, when learners move to secondary school.

The further back we look, the cheaper and kinder the intervention.

Data-driven leadership is not mainly a technology. It is a leadership practice: evidence, professional judgement and the discipline to learn from both. The principal in KwaZulu-Natal didn’t get extra staff or new resources. She got an honest picture, and she chose to act on it.

The Data Driven Districts programme is implemented by New Leaders Foundation on behalf of the Department of Basic Education, supported by the Michael & Susan Dell Foundation.