Implementation, Monitoring, Evaluation, and Adjustment

4. The Massively Parallel Strategy for Dealing with Scale and Complexity

This introductory article was written by ChatGPT at the direction of Heidi Burgess, who reviewed, edited, and approved the final content.
For more information

June 19, 2026

Once a theory of change has been put into practice, the work is not finished. In many ways, it has just entered its most important phase. Conflict interventions are based on assumptions about how people, groups, and institutions will respond. But in complex social systems, those assumptions often turn out to be incomplete or complete wrong. People may react defensively. Opponents may exploit the intervention. Outside events may change the political climate. This is why implementation must be accompanied by monitoring, evaluation, and adjustment. These practices help practitioners ask whether the effort is actually moving the conflict in a constructive direction, rather than simply following the original plan.

This is especially important in societal-level intractable conflicts, which are complex adaptive systems. The results of any intervention are shaped by many interacting forces, only some of which are under the practitioner’s control. Cedric de Coning’s work on adaptive peacebuilding emphasizes the need for iterative learning and adjustment under conditions of uncertainty. That is a useful model for conflict work more generally. A program may be well designed and competently implemented, yet still fail because the environment changes. Another program may appear successful for reasons that have little to do with the intervention itself. Honest monitoring and evaluation must leave room for both possibilities.

Practitioners are often uneasy about evaluation because it can feel like a threat. If a promised strategy did not produce the hoped-for results, they may fear that funders will see the project as a failure and withdraw support. This is understandable, but it creates a serious problem. If organizations are punished for learning, they will be tempted to report only success, hide uncertainty, and keep doing what they said they would do, even when the evidence suggests that they should change course. Better practice requires a different culture. Funders and practitioners both need to treat evaluation as a learning tool, not merely as an accountability weapon. CDA Collaborative’s work on theories of change in monitoring and evaluation points in this direction by encouraging programs to examine their assumptions as implementation unfolds.

Another danger is that evaluation can focus too heavily on what is easiest to count. Numbers matter, but they can be misleading when they are treated as proof of impact. Counting participants in a dialogue does not show whether trust increased. Counting workshops does not show whether people gained the courage or skill to act differently. The deeper questions are often harder to measure: Did relationships change? Did people become less afraid? Did institutions become more responsive? Did the conflict become less destructive? Did these changes last? The OECD evaluation criteria offer one widely used framework for asking broader questions about relevance, effectiveness, impact, and sustainability. But even good criteria have to be applied thoughtfully, with attention to context and to the limits of what can be known.

The purpose of monitoring and evaluation should be adaptive learning. Monitoring provides ongoing feedback while the work is still underway. Evaluation looks more carefully at what happened and why. Adjustment is the step that turns learning into improved action. In complex conflicts, this may mean strengthening a promising approach, redesigning a weak one, or abandoning a strategy that is making things worse. Approaches such as adaptive management and formative evalation are helpful because they assume that learning must occur while the system is changing. The central lesson is simple: peacebuilding and democracy-strengthening efforts should not push harder on a failing plan. They should pay close attention, learn quickly, and be willing to change course when reality shows that a different path is needed.

—————————

This page was created by ChatGPT in response to this prompt. It was then reviewed, edited, supplemented and approved by Heidi Burgess. More information about how and why we are using AI in this way, and about the growing number of ways in which Beyond Intractability is using ChatGPT, Claude and other AI systems to generate content and build out the BI system, is available on our BI/AI Overview Page.