The Blame Trap
Walk into many development organisations in South Asia and ask what happens when a programme fails to meet its targets. The answer, more often than not, involves some version of blame. The field team blames the community for low participation. The programme team blames the field team for poor implementation. The management blames the programme team for unrealistic planning. And the donors blame the management for poor results. Everyone is pointing fingers, and no one is learning.
This blame culture is not unique to the development sector, but it is particularly damaging here. Development work operates in conditions of profound uncertainty — interventions interact with complex social systems in ways that are inherently unpredictable. Failure is not a sign of incompetence; it is an inevitable feature of working at the frontier of social change. Yet most organisations treat failure as something to be hidden, explained away, or punished rather than examined and learned from. The cross-sector synthesis behind ALNAP's "Lessons of lessons" review makes the same point about humanitarian work: the same lessons resurface evaluation after evaluation precisely because organisations rarely act on what their own reviews already tell them.
The cost of this culture is enormous. When staff fear punishment for honest reporting, data quality deteriorates. When programme managers hide negative findings, the organisation loses the feedback loops needed to improve. When donors penalise failure rather than rewarding learning, organisations optimise for favourable reports rather than genuine impact. The entire system conspires against the learning that development work desperately needs. As we explore in stories of evidence-based pivots, it takes real courage to follow uncomfortable evidence.
What Psychological Safety Looks Like
The concept of psychological safety — originally introduced by Edgar Schein and Warren Bennis in the 1960s and later developed by Harvard professor Amy Edmondson — describes an environment where people feel safe to take interpersonal risks: admitting mistakes, asking questions, offering dissenting opinions, and flagging problems without fear of punishment or humiliation. Edmondson's study of hospital teams produced a now-famous counterintuitive finding: the better-performing units reported more errors, not fewer — not because they made more mistakes, but because they felt safe enough to surface them and learn. In development organisations, that same safety is the foundation on which a learning culture must be built.
Psychological safety does not mean lowering standards or avoiding accountability. It means separating the judgment of the work from the judgment of the person. A programme that fails to achieve its outcomes is a learning opportunity, not an indictment of the team that implemented it. A field officer who reports that beneficiaries are not engaging with the programme is providing valuable intelligence, not confessing to failure.
In practice, psychological safety manifests in specific behaviours. Leaders who model vulnerability by sharing their own mistakes and uncertainties. Team meetings where dissent is actively invited rather than merely tolerated. Performance reviews that assess learning and adaptation alongside target achievement. Reporting systems that ask "what did we learn?" alongside "what did we achieve?"—framed around indicators that actually matter, not just vanity metrics.
"In a learning organisation, the question is never 'whose fault was it?' The question is 'what can we do differently?' That single shift changes everything." — A veteran NGO leader in Mumbai
After-Action Reviews: A Practical Tool
One of the most effective tools for building a learning culture is the after-action review (AAR) — a structured reflection process developed by the US Army in the 1970s and increasingly adopted by humanitarian and development organisations. ALNAP's after-action review guidance adapts the method for aid agencies around four simple questions: What was supposed to happen? What actually happened? Why was there a difference? What will we do differently next time? AARs are conducted after significant activities — programme launches, training events, emergency responses, evaluation exercises — while the experience is still fresh.
The power of AARs lies in their structure and regularity. They are not occasional exercises but routine practices embedded in the organisation's workflow. A livelihoods programme might conduct an AAR after every market linkage event. A health programme might review every community health camp. An education programme might debrief after every teacher training cycle. The cumulative effect of these regular reflections is a steady accumulation of practical knowledge that improves programme quality over time.
Critical success factors for AARs include: timeliness (conducting them within days of the event, while memories are fresh), inclusivity (involving all team members regardless of hierarchy), documentation (recording insights in a shared, searchable format), and follow-through (actually implementing the changes identified). An AAR that generates recommendations but no action is worse than no AAR at all — it teaches staff that reflection is performative rather than genuine.
Safe Spaces for Failure
The boldest precedent comes from Engineers Without Borders Canada, which from 2008 published an annual Failure Report — candidly documenting projects that went wrong, from a farmers'-organisation programme in Burkina Faso to a water-monitoring system in Malawi the government could not afford to sustain — and later spun the idea into a wider AdmittingFailure platform. EWB's experience punctured a common fear: rather than losing donors, the organisation found that transparency about failure built trust. Some development organisations in South Asia have begun creating similar intentional spaces for sharing and learning from failure, and participatory MEL approaches can reinforce this by giving communities a voice in defining what success and failure look like. A handful of NGO networks now run quarterly "Fail Fest"-style sessions where programme managers present their biggest failures from the preceding quarter, what they learned, and how they adapted — deliberately framed as celebratory rather than sombre, with recognition for the "most instructive failure" and the "bravest pivot."
Other teams borrow "red team" exercises for major programme designs, where a designated group's role is to challenge assumptions, identify weaknesses, and argue for alternative approaches. These exercises surface potential failure points before implementation rather than after, and they normalise critical questioning as a constructive rather than adversarial practice.
A natural extension is the internal "failure library" — a documented collection of programme components that did not work, with analysis of why and what was learned. Where organisations require new staff to read it during onboarding, and ask programme designers to check fresh proposals against documented failures, the library becomes a defence against repeating known mistakes — turning what ALNAP describes as the costly habit of relearning the same lessons into institutional memory.
The Leadership Imperative
Culture change starts at the top. No amount of AAR facilitation guides or learning frameworks will transform an organisation whose leaders punish honesty and reward spin. The single most powerful thing a development leader can do to build a learning culture is to publicly share their own failures and what they learned from them. When a CEO tells the team, "I made a strategic error last quarter, here is what happened and what I learned," it sends an unmistakable signal that learning from failure is valued at the highest levels.
Donors have a crucial role too. Funding frameworks that require only success stories and upward-trending indicators actively undermine learning cultures. The clearest institutional attempt to flip this is USAID's Collaborating, Learning and Adapting (CLA) framework, which treats organisational culture and adaptive management as enabling conditions for effectiveness — and which evidence suggests is associated with better organisational and development outcomes. The deeper principle is one Michael Quinn Patton has argued for decades through utilisation-focused evaluation: an evaluation should be judged not by its methodological elegance but by whether intended users actually use it to make decisions. Yet, as ALNAP's work on evaluating humanitarian action repeatedly shows, the uptake of evaluation findings remains the sector's weakest link — making the culture this article describes the precondition for evidence to be used at all. Tools like the MEL Plan Lab can help teams design learning-oriented frameworks from the outset. This shift, while still minority practice, represents the kind of systemic change that the development sector needs to truly become a learning enterprise.