Teacher development in South Africa: are some of the pieces finally coming together?

Reflections from the 2026 DBE/VVOB Teacher Development Colloquium

I spend much of my working life evaluating teacher development initiatives, often funded by foundations and other external donors. So I went into the recent DBE/VVOB Teacher Development Colloquium listening partly with that hat on.

I came away more encouraged than I expected. Not because South Africa has suddenly solved teacher professional development. We very clearly haven't. But because I saw signs that some of the pieces we have been talking about for a very long time are beginning to fit together.

Oh, and the people. Some brilliant, deeply committed people are working in government in this space. It is easy, from outside the system, to talk about “the Department” a
s though it is one big faceless institution. Spending three days with some of the people actually trying to make this work was a useful reminder that it isn't.

This wasn't just a talk shop

One thing that struck me was the level and range of people in the room.

This wasn't a conference consisting mainly of researchers and NGOs talking to one another about teacher development. It deliberately brought together people who hold different parts of the teacher education and development system, with participation at senior levels of government, including Deputy Minister and Director-General level, alongside DBE and provincial officials, DHET, SACE, unions, universities, development partners and others. That matters because fragmentation was one of the strongest themes emerging from the discussions.

We don't necessarily have a shortage of activity in teacher development. We have many programmes, institutions, providers, datasets and structures. The problem is partly that they don't always connect particularly well.

And this is happening at an interesting moment. The Integrated Strategic Planning Framework for Teacher Education and Development (ISPFTED) 2011–2025 has reached the end of its original planning period. That framework envisaged a much more integrated system in which teachers identify their development needs, learn through PLCs, access quality-assured development opportunities and connect their learning to SACE's CPTD system.

The next iteration of the teacher-development strategy therefore has an opportunity to build on fifteen years of implementation experience. (Oh, and there is still a small window for comment on it).

We have known quite a lot about the problem for some time

After the Colloquium I went back to the DBE's Teacher Professional Development Master Plan 2017–2022. It makes interesting reading now. Many of the problems we were discussing at the Colloquium aren't new discoveries.

The Master Plan already recognised, for example, that some of the barriers to teacher development weren't about the content of programmes at all. It specifically identified time for teacher training and funding for programmes as enabling constraints.

It also envisaged a system in which teacher development needs could be diagnosed, teachers could be directed towards relevant programmes, programmes could be quality assured, and learning could be linked to professional-development points. So, in some respects, the theory wasn't missing. Implementation capability was. 

And that leads to a few things I took away from the Colloquium.

1. There is a resource problem — but it is bigger than the training budget

Government simply does not have unlimited resources for teacher development. For those of us working in the externally funded part of the education system, this deserves some reflection.

Substantial philanthropic, corporate, and development-partner funding goes into teacher development in South Africa. But historically, much of that money has effectively funded a parallel teacher-development system: identify some schools, bring in a programme, employ coaches or facilitators, train teachers, evaluate it, and eventually exit.

Mary Metcalfe and others have been making versions of this argument for years: the question is not only how to fund good interventions, but how external resources strengthen the education system that ultimately has to do this work at scale.

But money isn't the only scarce resource. A teacher-development model that depends on subject advisors assumes sufficient subject-advisor capacity to provide meaningful support to teachers. A model based largely on workshops plus occasional monitoring visits assumes that this is enough contact to sustain teacher learning and changes in classroom practice. And neither assumption will hold equally well everywhere.

The old Master Plan envisaged quite substantial public infrastructure around teacher development — Provincial Teacher Development Institutes, District Teacher Development Centres, curriculum and subject advisors, schools and PLCs. But even then implementation was uneven across provinces.

So perhaps one question for external funders should increasingly be:

Not only: "Which teacher-development programme should we fund?"

But: "Which constraint in the teacher-development system could our funding help remove?"

That could lead to quite different investments.

2. There is a serious attempt to address fragmentation

This was probably my strongest impression from the Colloquium.

The system knows that fragmentation is a problem.

That includes fragmentation between actors, programmes and institutions, but also fragmentation of information. We collect considerable amounts of data about teachers, participation, professional development and learner performance. But those data don't necessarily flow easily between institutions or translate into decisions about what development a particular teacher needs next.

The fact that people responsible for different pieces of this system were actually in the same room trying to work through these problems is significant.

And the problem isn't confined to continuing professional development.

Initial Teacher Education sits primarily within the higher education system, while much of the responsibility for in-service teacher development sits within basic education. Yet a teacher's professional learning journey obviously doesn't respect that administrative boundary.

This isn't a new insight either. The earlier Master Plan was explicitly a DBE–DHET endeavour and was approved through their joint HEDCOM structures. The current DBE Annual Performance Plan also identifies alignment of initial teacher education with universities and national priorities, national teacher induction, access to training, and stronger coordination of teacher training as areas of work.

The challenge is making those connections real.

3. Digital infrastructure changes what is possible

One of the developments I found particularly exciting was the new Online Teacher Development Platform.

On its own, an online training platform is hardly revolutionary. What is potentially important is the integration around it.

The direction of travel is towards teachers being able to access professional development, complete learning, have this linked to SACE CPTD requirements, and ultimately have a more coherent record of their professional learning. DBE has been working towards integrating the online platform with SACE so that completed programmes can be credited and training — including face-to-face training — can be captured through a common system.

Again, the idea itself isn't new. The old ISPFTED already imagined teachers identifying their own development needs and accessing appropriate development opportunities. The Master Plan envisaged digital platforms through which teachers could access development according to their needs and interests.

What has changed is that the infrastructure is becoming much more plausible. And that creates something we haven't really had before:

a realistic pathway towards differentiated teacher development at scale.

Instead of assuming that every Grade 4 teacher in a district needs the same workshop, a teacher's professional-development pathway could increasingly respond to what that teacher needs.

That doesn't guarantee good learning. People can click through online courses. Completion isn't competence, and competence doesn't automatically translate into changed classroom practice.

But differentiation is at least technically possible in a way it wasn't before.

4. And no, this doesn't mean putting everything online

This was another encouraging aspect of the discussions. There seemed to be quite a clear recognition that digital delivery is part of the solution, not the whole solution. Blended learning remains important. So does ongoing human support.

And here I found myself thinking about Professional Learning Communities.

When the 2011–2025 ISPFTED came out, I remember looking at the ambition around PLCs and wondering how feasible it really was.

Since then, many attempts have tried to figure this out. BRIDGE's Innovation in Education work was one - Now taken forward by NASCEE. VVOB has spent years working with the system on what makes PLCs function in practice. Many NGO training implementers have incorporated PLCs into their training modalities, albeit with inconsistent implementation success. 

By the time of the 2017–2022 Master Plan, DBE and VVOB had already conducted PLC capacity building in eight provinces, an M&E system had been developed, and work on subject- and programme-based PLCs was continuing.

The Master Plan actually lays out the institutional chain required to make PLCs work: national planning, provincial implementation, district support, school-level participation, resources, monitoring and links to CPTD.

Fifteen years later, PLCs haven't magically become easy. But we know considerably more about how to support them. That matters.

5. We are finally paying more attention to the transitions in a teacher's career

There is another resource gap that I think we sometimes underestimate. For years we effectively expected SMT members to induct and support novice teachers — without necessarily preparing experienced teachers to mentor, or adequately preparing novice teachers for how to continue learning once they entered a school. That too is changing.

Teacher induction and mentoring have been on the policy agenda for some time; DBE describes them as important components of teacher professional development and has subsequently developed national induction arrangements.

The same is true further upstream. We are paying more explicit attention to Initial Teacher Education, including the connection between what universities prepare teachers to do and what the schooling system subsequently expects them to do. Current policy work also includes school-based work-integrated learning, including how student teachers are placed, mentored, supervised and assessed.

None of this eliminates the institutional complexity created by teacher education spanning DHET, DBE, universities, SACE, provinces, districts and schools. But at least the whole professional pathway is increasingly visible.

So what feels different?

Perhaps this is my main takeaway. South Africa has not suddenly discovered what good teacher development should look like. Much of the vision has been there for years.

Needs-based development. Professional learning communities. Better use of technology. Quality-assured programmes. CPTD. Stronger teacher support structures. Better induction. Connections between initial teacher education and continuing professional development.

We have been trying to build these things for a long time.

What felt different at this Colloquium was that some of the infrastructure, accumulated implementation learning and institutional connections needed to make that vision actionable are beginning to come together.

There are still enormous gaps. Capacity differs dramatically between contexts. Funding remains constrained. Digital systems can fail. PLCs can become meetings rather than learning communities. CPTD can become point collecting. Subject advisors cannot provide intensive support to unlimited numbers of teachers. And coordination structures don't automatically produce coordination.

None of the solutions is foolproof, and none of the constraints looks the same everywhere. But there is something increasingly resembling an actionable theory of action:

Identify needs → differentiate development → widen access → combine digital learning with human support → strengthen PLCs and local support → connect learning across a teacher's career → improve the flow and use of data → and coordinate the institutions responsible for making all of this happen.

For those of us evaluating — and particularly those funding — teacher-development interventions from outside government, I think that creates an interesting challenge. Perhaps our unit of thinking needs to shift a little.

Yes: does our teacher-development intervention work?

BUT ALSO: What part of the teacher-development system does it strengthen — and what will be left behind when our funding ends?

Photo by Antoinette Plessis on Unsplash

AI disclosure: ChatGPT helped me think this one through — interrogating ideas, reviewing source material, structuring the argument and editing drafts. I remain responsible for the interpretations, judgements and final post. Charlie did however write this joke for you:  Knock knock. Who’s there? CPTD. CPTD who? Exactly. Please complete the module. 

A second challenge in the Teacher Development Ecosystem: capacity for interpretation and synthesis

Making research and evaluation more visible is necessary, but it is not sufficient. If we succeed in connecting evaluation reports, research, administrative data and information about interventions, we will have access to a much richer body of evidence. But somebody still needs to make sense of it.

This requires more than technical data skills. We need enough people across the education system who can critically interpret evidence, assess the strength and limitations of different studies, recognise when findings can or cannot be compared, synthesise evidence across multiple sources, and translate that evidence into useful insights for decision-makers.

As I argued in an earlier post, the interesting work often starts after the information has been assembled. Good sensemaking requires moving between data and the real world it represents, while retaining uncertainty, context and different perspectives. (M&E Blog)

So alongside investment in better data and evidence infrastructure, there is a human-capability question:

Do we have enough people, in the right places in the system, with the skills and time to turn an increasingly rich evidence base into useful knowledge?

Possible responses: strengthening interpretation and sensemaking capacity

Improving access to research and data is only part of the solution. We also need to strengthen the people, processes and institutions that turn information into useful knowledge for decision-making. These ideas could work together as different parts of that system. Some of them are already underway and not new initiatives. 

1. Create an annual “Sensing” publication for teacher development. Develop an annual synthesis, similar in concept to the South African Child Gauge, that brings together key data trends, research and evaluation findings, policy developments, emerging initiatives and evidence gaps. The New Leaders Foundation is already doing some really interesting work to explore what is happening in the system. The purpose would be to help the sector periodically stand back and ask what we are learning collectively about teacher development.

2. Connect academic research hubs into the evidence system. Build more deliberate links with universities and research centres producing relevant education research. Rather than expecting decision-makers to find and interpret academic publications themselves, relevant findings could be translated into short, accessible evidence products and incorporated into the wider evidence base. Academic partners could also contribute to periodic synthesis and sensemaking. Dr. Cally Ardington is doing some very interesting work through ADEA, RESEP is a reliable hub of interesting research outputs, and our joint data repositories are growing for example - Data First. Our colleagues at the Research Coordination, Monitoring and Evaluation (RCME) Directorate in the DBE are also very influential in this system. 

3. Build data capability around real DBE decisions. 

Our colleagues at the RCME are also instrumental in providing high-quality analysis and reporting to inform the broader system. But, we need more people able to use the available data in SA-SAMS/DDD to inform teacher development decisions. The Subject Advisors in Districts are potential key users. Develop customised reports using SA-SAMS, DDD and other available data around the actual decisions that officials need to make, accompanied by practical data-use bootcamps. The emphasis would be less on generic data literacy and more on helping people ask the right questions, interpret particular indicators and use evidence appropriately in their work.

4. Strengthen the analysts who support decision-makers. Develop custom reports and offer more advanced bootcamps for analysts, M&E specialists and research staff in NGOs, foundations, corporates and government. These could focus on triangulation, synthesis, interpretation, visualisation and communicating uncertainty, building a distributed network of people capable of translating complex evidence for decision-makers.

5. Provide ongoing support through an imagined Ask Ngolwazi. Training does not solve the problem when someone encounters an unfamiliar indicator or report six months later. Ask Ngolwazi, an AI-enabled data coach, could provide just-in-time support: helping users understand indicators, interrogate reports, formulate useful questions and recognise what the available data can—and cannot—tell them. It would complement human analytical capacity rather than replace it.

6. Create a mechanism for partners to contribute capacity and resources. Drawing inspiration from the Bana Pele collaborative framework, develop a mechanism through which DBE can identify priority evidence, analytical and capacity needs and invite donors, foundations, corporates, universities and other partners to contribute against them. This could help shift external investment from fragmented individual initiatives towards shared system priorities and capabilities.

Taken together, these are not simply six training or knowledge-management activities. They begin to describe an evidence-use system:

Generate evidence → make it visible → translate and synthesise it → build people's capacity to interpret it → support them when they use it → mobilise additional capacity where the system needs it.

7. Use futures scenarios to stay adaptive

We need to recognise that some of the ideas we develop today may become irrelevant surprisingly quickly because we are in a period of transition. Borrowing from futures research, we could develop a small set of plausible scenarios for the future of teacher development and use them to test our assumptions, notice emerging signals, and keep adapting our investments and approaches. 

 

Interesting Initiatives That May Help Us Make Research and Evaluation More Visible

A recurring challenge in education is that we do not have a good view of what research and evaluation is happening, where it is happening, and what has already been learned. Yet we are not starting from scratch. Several existing initiatives already hold different pieces of this puzzle. The opportunity may be less about creating another repository and more about exploring how these pieces could connect.

1. NED Connect – linking interventions to the research and evaluation around them

What exists: NED Connect, developed and stewarded by NASCEE, is building a national picture of who is doing what, where, and with whom in the education ecosystem. It includes organisation profiles, geographic information and interactive maps, including school-level mapping of reported support.

The possibility: Could NED Connect include an additional research and evaluation layer? An intervention could be tagged to indicate that research or evaluation is planned, underway or completed, with basic metadata and links to reports or other evidence. We could then begin to see not only where interventions are happening, but also where evidence is being generated about them.

2. AfrED – building on an existing repository of African evaluation evidence

What exists: CREST's African Evaluation Database (AfrED) already captures bibliographic information about evaluation reports, journal articles and doctoral theses. It currently contains around 7,800 documents, including more than 4,000 evaluation reports, and is specifically intended to make African evaluation production more visible.

The possibility: Rather than creating another repository for completed evaluation reports, could we strengthen the connection between AfrED and other education platforms? NED Connect might show that an intervention has been evaluated, while AfrED could provide the searchable evidence record. Over time, this could create a much clearer pathway from intervention → evaluation → evidence.

3. DDD – connecting evaluation with longitudinal system data

What exists: Data Driven Districts (DDD), a partnership involving the DBE and New Leaders Foundation, turns SA-SAMS data into accessible information for education officials. It currently incorporates data on around 12 million learners across approximately 24,000 schools, including information on attendance and performance.

The possibility: DDD raises a different opportunity. Could evaluation information be linked to the rich longitudinal information already available about schools and learners? Knowing which interventions and evaluations have taken place in a school, alongside its administrative and performance trajectory, could open new possibilities for secondary analysis, longer-term follow-up and new evaluation designs. It may also allow us to learn more from data already being collected rather than repeatedly creating parallel datasets. And here is a novel idea - Why don't we set up the system so that the schools can also use the evaluation and research output? 

4. A more connected ethics and research approval process – making planned research visible earlier

What exists: Research and evaluation currently passes through multiple processes: funders and researchers commission work; ethics bodies review studies; and provincial education departments manage research approvals. Each process has a legitimate purpose, but they are largely separate. No single actor therefore has a complete view of the research and evaluation landscape.

The possibility: Could a central or coordinated education ethics review function become another point of connection? It might draw basic metadata from provincial approval processes and ethics applications, identify possible overlaps, and trigger follow-up where appropriate: Another evaluation is already taking place in these schools. Similar data have recently been collected. Have the two teams spoken? Is there existing evidence that should be considered? The aim would not be another layer of bureaucracy, but a light-touch coordinating function that helps the system notice connections before data collection begins. This builds on the idea of pre-ethics review, a research and evaluation registry, and a small enabling coordination function already explored in the Weave concept.

5. The Zenex–CREST meta-review – showing the value of bringing evaluation evidence back together

What exists: The Zenex Foundation has recently supported CREST to undertake a meta-review of selected evaluations commissioned by the Foundation since 2015, across Foundation, Intermediate and Senior Phase work. Rather than looking at another individual programme, the exercise brings together a body of evaluation evidence accumulated over roughly a decade and examines it systematically.

The possibility: This illustrates why visibility matters. Individual evaluations answer important questions about individual programmes. But when we can find, organise and synthesise evaluations across time, another level of learning becomes possible: What findings recur? Which approaches appear promising across contexts? Where does evidence disagree? What have we repeatedly tried? And what important questions remain unanswered?

---

The bigger opportunity

These initiatives do different things, and they should not necessarily be combined into one large system. But together they suggest the beginnings of an evidence infrastructure:

NED Connect helps us see what is happening where.
Provincial approvals and ethics processes could help us see what research is about to happen.
AfrED can help us find what research and evaluation has already been completed.
DDD provides longitudinal system data that can potentially deepen what we can learn from evaluations.
Meta-analysis and synthesis allow us to step back periodically and ask what the accumulated evidence is telling us.

The opportunity may therefore not be to build one more database. It may be to make the connections between what already exists — so that each new research or evaluation investment contributes not only to an individual programme, but progressively to what the education system knows.

AI disclosure: I used ChatGPT to help structure, refine and edit this post, based on my own ideas, experience and source material. I reviewed and take responsibility for the final content. 

We have a visibility problem in our education research ecosystem

 One of the challenges in education research and evaluation is surprisingly basic: we don't really know what is happening.

At any given time, universities, government departments, foundations, NGOs and evaluators are conducting research and evaluations across the education system. Schools are participating in studies. Learners are being assessed. Teachers are completing surveys and being observed. Programmes are generating monitoring data. Reports are being produced.

But it is remarkably difficult to get a picture of what research and evaluation is happening where, what has already been completed, what data have already been collected, and what we have learned from it all.

This matters for very practical reasons. A funder may commission an evaluation without knowing that somebody else has investigated a similar question. A researcher may collect data that already exist elsewhere. A school may participate in several studies while another comparable school participates in none. And an evaluation may produce useful findings that live on a funder's website—or in someone's Google Drive—without becoming part of the evidence available to the wider system.

The problem, then, is not simply that research isn't disseminated well enough. We have a visibility problem across the research and evaluation lifecycle. We struggle to see what is planned, what is underway, what has been completed, what evidence and data already exist, and how individual pieces of work relate to one another.

That makes it much harder to move from lots of individual evaluations to cumulative learning about the education system.

The interesting thing is that we may not need to build a giant new repository to solve this. Several pieces of the infrastructure already exist, or are beginning to emerge. The question I've been thinking about is whether we could connect them differently. In the next post I share some ideas about how we can make research and evaluation more visibile. 

AI disclosure: I used ChatGPT to help structure, refine and edit this post, based on my own ideas, experience and source material. I reviewed and take responsibility for the final content.

A potentially important shift in teacher development

Some exciting news from the DBE/VVOB 2026 Teacher Development Colloquium today: the DBE launched its new Online Teacher Development Platform.

Teachers can access targeted courses and microcredentials, build their own learning pathways, and have completed learning linked directly to SACE CPTD points. External providers can also link their training to the platform.

Very exciting.

Of course, we should not expect miracles. Internet access and technology will remain barriers. Some people may game the system for points. Completion does not automatically mean learning, and learning does not automatically translate into classroom practice.

But I think it would be hard to argue that this is not a step up. In many places, subject advisors simply cannot reach every teacher. And even where teachers attend workshops, we often know surprisingly little about how much of the training they actually engaged with. Sometimes even tracking who completed most of the programme is difficult.

So this is potentially more than another online platform. It could change some of the infrastructure of teacher professional development.

The next question is how we support it well. Districts often do not have the ICT support capacity that a system like this will need. Could funders and partners help strengthen that public infrastructure, rather than build yet another parallel programme?

There are so many possibilities. I’m excited to see where the next two days lead us.