By Grace Matlhape, Chief Executive Director, and Rebecca Hickman, Chief Ecosystem and Policy Officer, at SmartStart
August 5, 2025 – To achieve lasting change at scale, we’re thinking differently. SmartStart is using a systems thinking approach to build a national early learning platform in South Africa. Our goal is not just to be a bigger program, but to create a fundamentally different delivery mechanism that can reach every child.
Why Systems Thinking?
The challenge when a new concept becomes popular is that it can be co-opted in ways that dilute or distort its essence. The word ‘system’ is hardly new, but systems thinking does not mean just thinking about how to improve systems. It invites us to use a whole new framework to consider the intricate interweavings of the variables that bear on the problem we want to solve.
Systems thinking is a way of understanding complex systems that sees the whole system—rather than single elements, events, or actors—and considers relationships, interdependencies, and system behaviours. This means the focus is on underlying structures and patterns of cause and effect over time.
Used in this way, systems thinking has helped SmartStart consider questions such as: Have we named the problem correctly? Are we seeing all the upstream causes? Have we thought through all the downstream effects of our solution? And, crucially, are we intervening at the highest point of leverage?
Some of our systems work has been a form of retrofitting—bringing a new theoretical framework to bear on our theory of change and intervention design. But much of it has also helped us understand what we need to redesign, where we should shift focus and resource allocation, and what the wider implications are for policy and regulatory reform.
Understanding the Implementation Gap
We began by analyzing government systems that comprise the ECD regulatory framework—such as systems for financing, curriculum, qualifications, and health and safety—and observed two things. Firstly, these systems do not, by themselves, either unlock new supply or build quality in early learning programs. In a quasi-market of provision with endemic under-resourcing, programs do not automatically self-establish or self-improve just because the right state regulatory systems exist. Moreover, solving for delivery is very different from regulating.
Secondly, government systems do not naturally bend toward equity. In fact, where the internal imperatives of bureaucracies are disconnected from the everyday lives and environments of people, there is a high risk that those bureaucracies will reproduce—or even magnify—social injustices.
We call this divergence between intended policy outcomes and what is actually happening on the ground “the implementation gap.” The systems mapping tools we have gained from the ECDAN working group have enabled us to develop a much fuller understanding of this implementation gap and how societal systems interact with government systems.
Beyond State Systems
This process showed us that we were not wrong to focus on state systems, but that this focus provided only a partial view. Transforming the whole system requires us to have a clear line of sight into the system dynamics that create structural barriers and opportunities, including the role of deep-seated cultural, social, political, and economic factors.
For example, we see how the highly informal nature of many early learning programs in low-income communities is both an upstream cause and a downstream effect of an inappropriate regulatory regime—creating a reinforcing loop. Many providers in informal contexts struggle to engage with state systems predicated on an assumed level of formality, and so they remain outside the regulatory framework, preventing it from performing its essential purpose.
This understanding has helped SmartStart recognize that one of its core functions at the field level is to serve as a bridge between state regulatory funding systems and the real-world contexts of early learning programs.
Shifting the Government Lens
It is in the DNA of much of government to think that policy, legislative, and funding actions can solve everything. Sitting in the complexity of a comprehensive causal loop diagram blows that assumption apart. It invites humility on the part of governments—to see the wider and deeper systems that are at play: social systems, knowledge systems, and power systems, to name a few.
This insight helps governments see that there is very little they can solve alone.
In South Africa, this understanding has helped inform the new 2030 Strategy for ECD Programmes, which the government published earlier this year. The Strategy proposes a new Social Compact for ECD, bringing all players together to co-produce solutions—the ECD sector, social partners, businesses, donors, and, most importantly, families and communities.
The strategy also acknowledges that closing the access and quality gaps requires new forms of partnership. Government has a pivotal role to play in leading, planning, and driving, but it cannot in the near term build a new public service from scratch. Its priority is radical ecosystem enablement—working closely with social partners who are well-placed to expand service delivery rapidly and equitably.
This recognition of the power of partnerships also lies at the heart of the SmartStart model. Our platform architecture centers the network, which is understood broadly to include not only our implementing partners and practitioners, but also children, families, and communities. We see enormous potential in the organic interactions between these different players and feel a related responsibility to ensure that our platform design enables collaborative environments that amplify value.
Systems Not Programs
In many low- and middle-income countries, ECD service provision has been strongly influenced by a “development” model, with nongovernmental organizations (NGOs) playing a vital role in compensating for service delivery failures. However, out of necessity, NGOs generally tackle challenges programmatically. This means they develop interventions that engage with specific failings in the system—e.g., lack of training or resources—or address issues at the sub-population level.
Such programs have a valuable role to play, but there is also a risk that they work at low points of leverage within the system. This means they might unintentionally prop up the very system that needs transforming or contribute to destabilizing levels of fragmentation.
In practice, this has informed our understanding that to close the access and quality gaps, an end-to-end delivery platform like SmartStart is needed. “Platform” in this sense means an intervention that organizes the continuum of service delivery options into a single value chain, and thereby can engineer access and quality wherever it is needed.
Connecting the Dots
Systems thinking does not permit siloed solutions. This is because we cannot pick one part of the system and somehow seal it off from the rest. If our focus is early learning, then the solution must consider how factors such as stunting, toxic stress, and indigenous knowledge systems—factors that might sit outside the conventional parameters of an early learning program—directly influence early childhood outcomes.
The recently published Lancet series on the “The Next 1000 Days” emphasizes the importance of multisectoral partnerships across education, health, nutrition, and social protection sectors. This goes beyond these different sectors simply communicating better and challenges us to think increasingly strategically about how the intersections between different service areas can be more effectively leveraged.
For SmartStart, this presents an interesting conundrum as we plan for the future. On the one hand, ambition must be tempered with pragmatism. There is a risk that overloading the delivery channel dilutes rather than increases impact. On the other hand, reimagining ways of partnering for multicomponent interventions presents an exciting opportunity to test new approaches.
Seeing Strength
Systems thinking has also encouraged SmartStart to pay attention to the value of what is already there—in Donella Meadows’ words, to understand the wisdom of the system and watch how it behaves. For us, this is about an asset-based approach that sees South Africa’s strengths and resources, rather than its deficits.
It has propelled us to design a delivery system that actively harnesses the nascent assets in our communities—such as under-employed women with the aptitudes to provide high-quality care and learning; home and community venues where millions of children are already cared for every day; and vibrant forms of informal economic activity.
Mental Models
One key theme that has interlinked the application of systems thinking in SmartStart and government has been the power of mental models at every level of the system—societal assumptions that care work is women’s work; parental beliefs that children do not need learning and development opportunities before they start school; and a range of government beliefs, including that high-quality care and learning cannot be provided in informal settings or that the cost of funding early learning provisioning outweighs the social and economic benefits.
Each of these mental models is a powerful leverage point in the system, influencing how we allocate effort and resources.
Toward Consensus
Finally, the process of creating the childcare systems mapping with ECDAN has demonstrated the power of systems thinking as a consensus-building tool. We can debate endlessly why ‘x’ intervention is better than ‘y’ intervention or why ‘z’ policy isn’t working. Systems mapping forces us to take a step back and use evidence to describe what is actually happening.
Within ECDAN’s systems working group, while we have had many debates over what to include in the map and how to name variables, there have been fewer disagreements over how the system actually functions. This is powerful. It provides a framework for understanding why, in terms of expanding access to early learning, the dial has barely flickered over the past decade. More importantly, the shared narration of events enables us to build consensus around where and how we need to intervene, and what to prioritize, in order to transform outcomes for children.
The resulting systems map is intricate, multilayered, and even messy. It takes time and effort to interact with it and extract insights and waymarkers. But arguably, it is our tendency toward simplicity that has confounded the most well-intentioned policies and programs. In early learning, as elsewhere, if the problem is multidimensional, then the solution must be too.