Our learnings from working on evidence use in education indicate that the embedded policy labs approach is gaining traction not only as a way of pursuing policy impact, but also in institutionalising evidence use within decision-making structures. Yet, there are trade-offs associated with this approach, such as maintaining independence and credibility, navigating bureaucratic hurdles, and being accountable. Some argue that an insider approach (quiet diplomacy), while potentially effective, might inadvertently align with or enable authoritarian tendencies, or simply lack broader public accountability. To explore these dynamics, OTT organised a parallel session on inside, outside and in-between, how think tanks influence policy during the OTT Conference in Johannesburg from 17-18th June 2025. The aim was to:
- understand diverse models of policy influence, including embedded, external, and hybrid think tanks and how they operate in different political and institutional contexts,
- assess the trade-offs such as access vs. independence, legitimacy vs. influence associated with different models of evidence-informed policymaking,
- examine contextual factors that shape the effectiveness of each model, including the political landscape, institutional culture, and stakeholder dynamics, and
- draw lessons from real-world examples to better understand how think tanks can strategically position themselves to achieve policy impact.
The session was facilitated by Tatiana Garcia (Manager of the Colombia Evidencia Potencial en Educación initiative, Fundacion Exe), Carien Vorster (Regional Representative, Roger Federer Foundation, South Africa), Nompumelelo Nyathi (Deputy Director: Research, Monitoring & Evaluation, Department of Basic Education, South Africa), Pamla GoPaul (Senior Program Manager, AU-NEPAD), Rico Bergemann (Associate Director, IDInsight, South Africa), and Nancy Lozano Gracia (Program Manager, Institute for Economic Development, World Bank).
It examined the diverse models through which think tanks pursue policy impact, from embedded “inside track” approaches within governments to independent, externally positioned organisations engaging broader publics. Each model presents unique advantages, such as access or legitimacy, and trade-offs related to autonomy, influence, and effectiveness.
Drawing on their experience working with embedded labs, Tatiana Garcia and Carien Vorster discussed various models of think tanks and their ideal positioning for policy influence: inside (embedded within government), outside (independent), or hybrid. The discussions highlighted that:
- External think tanks often have a broader reach, influencing a wider array of stakeholders from Members of Parliament to technical bureaucrats and the media. Their independence is a key asset, allowing them to choose their focus and maintain credibility. Credibility is further bolstered by rigorous research, expert opinion, academic partnerships, and a strong reputation. Strategies for reaching stakeholders include radio shows, podcasts, and various public policy instruments. However, the effectiveness of these approaches is heavily context-dependent.
- The experience of embedded think tanks, such as the one in the Department of Basic Education within the Ministry of Education in South Africa, reveals a more complex reality. Despite being funded by the government and staffed by government officials, embedded think tanks often face the challenge of “selling” their evidence back to funders and implementers. Building relationships of trust with other government departments by assisting with their ongoing work (e.g., monitoring and reporting) becomes crucial for their evidence to be taken up in policy frameworks.
“The notion of a direct highway from evidence to policy, when embedded, is often far from the truth; internal dances and tricks are frequently required.”
—Carien Vorster
- Interestingly, there was less familiarity with hybrid positioning, suggesting an area for further exploration and discussion within the think tank community.
Whether inside, outside or in-between, Rico Bergemann and Nompumelelo Nyathi emphasised that effective evidence use, and influence go beyond simply presenting data. They reiterated that,
- Policymaking is not a singular event. Instead, it is a continuous, unfolding process that demands patience, persistence, and an understanding of its inherent temporal evolution.
- Evidence, in isolation, holds limited power. Its true impact is felt when it is carefully aligned with a clear, overarching purpose and a well-defined objective, which provides direction and meaning to the data.
- Successful influence requires an understanding of various factors, including the opportune timing for interventions, the incentives (both explicit and implicit) that drive different actors and stakeholders, and the broader political economy of the circumstance. This includes understanding power dynamics, resource flows, and institutional norms.
- Perhaps one of the most compelling insights was the critical role of narrative and storytelling. While empirical evidence is crucial, it can inevitably become outdated. A strong, consistent narrative provides a stable anchor. This narrative remains unchanged, enabling the continuous integration and updating of the latest evidence while maintaining a coherent and memorable message. This highlights the emotional and persuasive dimensions of policy advocacy.
- Reiterating the top tips by leading southern think tanks, Rico and Nompumelelo emphasised that building robust relationships and strategically identifying key champions at appropriate levels, whether political appointees with decision-making power or influential bureaucrats within the administrative apparatus, is fundamental. This also extends to understanding internal government dynamics and the often-unspoken rules of engagement within departments.
- Think tanks must evaluate their influencing strategies, recognising the potential consequences. For instance, while certain outside tactics, such as litigation, can be effective in forcing policy shifts, they can also irrevocably damage relationships and burn bridges with government entities. This makes future collaborative engagement extremely challenging. The discussions underscored the need to carefully consider where an organisation wishes to position itself within the spectrum of inside, in between, or outside, and to understand the long-term implications of these choices.
It is important to have a clear purpose and values to build trust
Pamla Gopaul and Nancy Lozano discussed how to adapt models of influence to local realities and learnings across all sectors. From these discussions, the vital role of purpose and values for any organisation seeking to impact policy emerged. It is not enough to simply have data or evidence; there must be a clear objective and a compelling reason for its use. This foundational clarity directly relates to the concept of trust, which emerged as the currency of change. Trust is not merely a desirable quality; it is essential for achieving clarity of purpose and fostering meaningful policy shifts. Whether building trust from an outside-in or inside-out perspective, or somewhere in between, its cultivation is vital in bridging the gap between evidence and action.
Integrating learning and adaptation in policy influence
The discussions also stressed that think tanks should prioritise local knowledge and contextual research for effective policy work. Pamla and Nancy noted that while a strategic long-term view is important, it is equally crucial to pay attention to the immediate questions and demands arising from the context in which the work is being done. Learning and adaptation must be grounded in local knowledge, which should be participatory and informed by contextual research.
To genuinely learn and adapt, organisations need to find ways to measure their impact and establish robust feedback loops. These loops serve as a mechanism for understanding what works and what doesn’t, allowing for continuous refinement of strategies.
“Adaptation and learning begin with co-creation.”
—Pamla Gopaul
This collaborative approach fosters shared ownership, ensuring that policies are more likely to be relevant and effective.
Finding ways of communicating and repackaging information is key. Think tanks should always tailor their messages to different audiences, taking advantage of moments in policy cycles by being on time and relevant. Pamla and Nancy ended by urging the participants to challenge their assumptions.
The discussions underscored that there is no single best position for a think tank. The distinction between inside and outside is often artificial. Instead, the approach should depend on what is most useful in a given context, rather than being rigid. The most effective approach depends on the specific context, highlighting the constant need for adaptation, strategic thinking, and a deep understanding of the intricate balance between evidence, trust, and influence.
For more resources on the embedded labs approach, see: