The Plan Was Perfect. The System Never Read It. The limits of organizational control, and the quiet arrival of the illusion of progress

 Every output delivered. Every target met. Every report green. And yet, when you walk into the community where the programme worked, the stubborn thing it was built to shift is standing exactly where you left it.

Have you ever stood in that gap?

Organizations control activities and outputs. They influence outcomes. They do not control how systems respond.

For much of my professional life, I believed that better planning would lead to better results. The logic was clean. Clarify the objectives, strengthen the theory of change, sharpen the indicators, tighten accountability, and performance would rise, with societal outcomes following close behind. Much of the management thinking that shaped the social sector over several decades rests on this promise.

Experience taught me otherwise. I met programmes that were professionally designed, adequately funded and competently implemented, yet fell short of expectations. I met interventions that produced outcomes nobody had designed. I watched some initiatives succeed because of forces far beyond their reach, while others struggled despite disciplined delivery.

Picture a composite that many practitioners will recognise. A district health programme trains every community health worker, delivers every consignment of vaccines, and reports its coverage targets met at the quarterly review. In the villages, meanwhile, the decision to bring a child to the clinic is still being made by trust in the nurse, the counsel of grandmothers, the cost of transport and the memory of a stockout last season. The programme did what it promised. The system did what systems do.

The difficulty does not lie in the quality of the plan. It lies in the nature of the systems within which planning takes place.

Four Quiet Promises: What Results-Based Management Assumes About Change

Every management discipline rests on assumptions about how change occurs. These assumptions shape what managers notice, how they judge progress, and which actions seem reasonable. Results-Based Management (RBM) is no exception, and four of its assumptions matter most.

Linearity. RBM typically assumes that change travels along a chain: inputs support activities, activities generate outputs, outputs contribute to outcomes, and outcomes lead to impact. The chain brings clarity and helps organizations organize action. It also implies that cause and effect connect in orderly, predictable ways, an implication that feedback loops, delays and tipping points in social systems routinely disturb.

Controllability. RBM often assumes that organizations can steer results through effective planning and execution. Over activities and outputs, they can. But influence thins with every step outward from the programme boundary, until it reaches the household, the market and the ministry, where other hands are on the wheel.

Attribution. Performance systems rightly want to connect interventions to outcomes. In complex environments, however, outcomes arise from the combined influence of many actors, conditions, relationships and events. The question “what did we cause?” quietly gives way to “what did we contribute to?”

Stability. Results frameworks assume that conditions will remain steady enough for plans and targets to stay valid. Yet political priorities shift, institutions evolve, crises arrive, technologies disrupt and expectations change. A five-year target set in a calm season is a promise made to a future that has not yet been consulted.

None of these assumptions is flawed. Each is a useful simplification for planning, organizing, monitoring and accountability. Trouble begins only when a practical approximation is promoted to a complete description of reality. What starts as a tool becomes an explanation, and the explanation slowly stops being true.

Complicated or Complex? The Distinction That Changes the Method

Why does RBM perform so well in some settings and so unevenly in others? One answer lies in the difference between complicated and complex problems (Glouberman and Zimmerman, 2002; Snowden and Boone, 2007).

Complicated problems have many components, but the relationships among them can largely be understood through expertise and analysis. Building a bridge, implementing a payroll system, constructing a hospital or conducting a national census is demanding work, yet the pathway from action to outcome is largely knowable. Clear plans, milestones, indicators and reporting lines give such work the discipline it needs, and RBM serves it superbly.

Complex problems behave differently. Improving learning outcomes, strengthening child protection systems, reducing inequality, improving governance, building climate resilience and restoring public trust all involve many actors whose behaviour continually shapes one another. Outcomes emerge through interaction, adaptation, feedback, institutional incentives, social norms and changing circumstances. A community-based child protection model may thrive where social workers, traditional leaders and the police already trust one another, and stall where those relationships are thin. A reform may take root in one district and wither in the next, though the design, the budget and the training were identical.

Complex problems are not merely more difficult. They behave differently.

Management approaches built for predictable environments therefore grow less reliable as organizations engage more deeply with complex adaptive social systems (Ramalingam, 2013).

A DIAGNOSTIC PAUSE Think of the hardest result your organization is accountable for. Is the pathway to it knowable through expertise, or does it depend on how many other people choose to act?

When Good Management Produces the Illusion of Progress

A paradox sometimes appears when RBM meets complexity. Organizations become increasingly skilled at managing performance while remaining uncertain about their contribution to lasting change. Three consequences follow.

The illusion of progress. Programmes meet targets, complete activities, deliver outputs and file positive reports while deeper system conditions stay largely as they were. The organization looks successful because it delivered what it promised, yet the challenge that motivated the work remains substantially intact. Pritchett, Woolcock and Andrews (2013) describe a close cousin of this pattern, in which institutions adopt the forms of success without acquiring its function.

Optimization around what is measurable. Indicators attract managerial attention, and attention shapes behaviour. Once a measure becomes a target, it stops being a reliable measure (Strathern, 1997). Activities become targets, outputs become proxies for change, and reporting begins to overshadow learning.

Weakened learning. When plans are treated as commitments rather than hypotheses, a deviation reads as an implementation failure rather than as information. Feedback becomes something to explain away. Organizations grow better at executing their existing assumptions and less able to ask whether those assumptions still hold.

None of this arises from poor management. It arises from managing performance successfully inside systems that require something more.

Before You Diagnose the System, Diagnose Your Own Assumptions

In Adaptive Results Management (ARM), the first discipline, Define Purpose, orients everything that follows. The second discipline, Diagnose, begins somewhere less comfortable: with the assumptions we carry into the work. Before an organization can diagnose a system, it must diagnose its own assumptions about control.

The stakes are practical. Assume linearity, and we design linear solutions for nonlinear problems. Assume controllability, and we measure what we can control rather than what matters. Assume attribution, and we claim credit for what many hands produced. Assume stability, and we write plans that break when the weather changes.

The diagnostic question, then, is not whether these assumptions are wrong. They are not. It is where they have become the whole of our management thinking, rather than useful simplifications within it.

The Four Assumptions, Side by Side

The diagram below places the four assumptions in four quadrants. Each quadrant pairs a useful simplification with the point at which it stops being enough. The arrow along the base marks the shift from useful simplification to problematic, when a simplification is mistaken for a complete description of reality.

Alt text: A four-quadrant diagram showing the four assumptions beneath Results-Based Management: linearity, controllability, attribution and stability. Each quadrant contains the assumption and the reason it becomes limiting in complex adaptive social systems. An arrow along the bottom indicates the shift from useful simplification to problematic when treated as a complete description of reality.

What Practitioners Already Know

Many practitioners recognise these limits intuitively. Those who work closest to communities, institutions and frontline delivery meet the evidence daily: adaptation, unintended consequences, shifting incentives, informal networks and unexpected pathways of change.

Yet formal management systems often demand more certainty than the environment can give. Targets imply predictability. Reporting systems imply attribution. Planning frameworks imply stability. Practitioners therefore live between two realities: the complex one they experience and the simplified one their systems require them to describe.

That gap is becoming harder to sustain. It limits learning, discourages honest reflection, and weakens an organization’s capacity to respond intelligently to change. The answer is not to abandon performance discipline. It is to extend it, so that rigour and adaptation travel together.

The Assumption Audit: Forty-Five Minutes That Could Change Your Next Review

Set aside 45 to 60 minutes, alone or with your team, and work through one live strategy, programme or results framework. 

1	Choose one live plan. Pick a strategy, programme or results framework that is being implemented now, not one resting on a shelf.
2	Put the results chain on one page. Draw the theory of change with every arrow visible.
3	Mark linearity. Circle each arrow that runs from an output to an outcome. Ask which of these depends on feedback, delay, or a decision made by someone outside your control.
4	Mark controllability. Underline every result your team can deliver alone. Then list the results that need ministries, communities, partners or markets to move.
5	Mark attribution. Find each claim that reads “because of our intervention”. Ask who else is working on the same outcome, and what they would say caused it.
6	Mark stability. List the conditions the plan needs to hold for its full duration, such as funding, leadership, policy, security, climate and prices. Note which have already shifted.
7	Ask for evidence. For each of the four assumptions, write down what evidence you have that it holds in this place, at this time.
8	Choose one hypothesis. Select one assumption to treat as a hypothesis to test. State it as a sentence, name who will test it, what they will look for, and by when.

 Output: a completed one-page audit with one hypothesis to test. Debrief: what changed when you treated an assumption as a hypothesis? Use it in: strategy reviews, programme design and results framework development.

A QUESTION TO HOLD Which assumption about control most shapes your current plan, and what would change if you treated it as a hypothesis rather than a certainty?

A plan is a lantern, not a landscape. It lights the next few steps with great fidelity and says nothing about how the ground will move. The work ahead is to carry the lantern with discipline, and to keep our eyes on the terrain.

The next piece in this sequence turns to what organizations actually manage when they cannot control outcomes, and why the strategic portfolio becomes the primary object of management.

References

Glouberman, S. and Zimmerman, B. (2002) Complicated and complex systems: what would successful reform of Medicare look like? Discussion Paper No. 8. Ottawa: Commission on the Future of Health Care in Canada.

Pritchett, L., Woolcock, M. and Andrews, M. (2013) ‘Looking like a state: techniques of persistent failure in state capability for implementation’, The Journal of Development Studies, 49(1), pp. 1-18.

Ramalingam, B. (2013) Aid on the edge of chaos: rethinking international cooperation in a complex world. Oxford: Oxford University Press.

Snowden, D.J. and Boone, M.E. (2007) ‘A leader’s framework for decision making’, Harvard Business Review, 85(11), pp. 68-76.

Strathern, M. (1997) ‘“Improving ratings”: audit in the British university system’, European Review, 5(3), pp. 305-321.