A different evaluation question

I want to suggest a different approach to evaluation, one that is grounded in Nora Bateson’s concept of symmathesy. This question is:

What can this organisation now hold in relationship that it could not hold before?

It took years of careful evaluative work — work that was methodologically sound and often inadequate to what mattered most — before I understood why a different question was needed.

The question is doing something that conventional evaluative questions do not. It is asking about an organisation or a program as a system of relationships rather than as a collection of activities and outputs. It is asking what the system has become capable of — not what it has produced.

Before I can explain why this matters, I need to name the assumptions that make it invisible.

In the early twentieth century, Frederick Winslow Taylor developed what he called scientific management: the idea that organisations could be made more efficient by breaking work into standardised components, measuring inputs and outputs precisely, and optimising each component independently of the others. Taylor's insight was genuinely useful for organisations that were literally manufacturing things. It produced the factory system. The factory system produced the industrial economy. And the industrial economy eventually produced the vocabulary through which almost every kind of organisation now describes and evaluates itself: inputs, outputs, activities, outcomes, indicators, logframes.

The problem is that this vocabulary was never designed for organisations whose work is inherently relational, contextual, and resistant to standardisation. Program logic — the evaluative framework most commonly used in the NFP and government sectors — describes the organisation as a machine: resources go in, activities are undertaken, outputs are produced, outcomes follow. Measure the outputs. Track the outcomes. Determine whether the machine performed as specified.

Human beings in conditions of complexity and disadvantage do not move through systems the way widgets move through a factory. The relationships between parts of the system — between a caseworker and a family, between an organisation and the community it serves, between the formal service and the informal networks surrounding it — are not coordination mechanisms. They are where the work actually happens. And they are precisely what program logic cannot see, because they are not inputs, outputs, or outcomes. They are the living tissue of the organisation's practice, and the machine model has no language for living tissue.

This mismatch was always a problem. But it has become something closer to urgent.

We are living through a period of converging crises — ecological, social, technological, economic — that are not merely simultaneous but causally entangled. What is sometimes called the polycrisis does not present organisations with complicated problems that can be solved by better inputs and more efficient processes. It presents them with genuine complexity: situations where cause and effect are not separable in advance, where the most important dynamics live in the connections between things rather than in the things themselves, where the capacity to respond depends not on having the right answer but on developing the relational and learning intelligence to find it together, in real time, with the people most affected.

Organisations optimised for funder satisfaction, evaluated against predetermined outcomes, oriented toward compliance rather than genuine inquiry — these organisations are not equipped for this. They have developed sophisticated systems for appearing to learn while actually defending themselves against the kind of honest examination that learning requires. They produce evaluation reports that document what happened without illuminating why. They commission research that confirms what they already believe.

“This is is not a failure of intent. It is a failure of the underlying model — a model that was designed for a different century and a different kind of problem.”

T

he biologist and cyberneticist Gregory Bateson spent much of his life arguing that the unit of analysis for any living system is not the individual organism but the organism-in-its-environment — the relationship, not the thing. His daughter Nora Bateson has extended this into what she calls symmathesy: the mutual learning that occurs in and through the relationships between parts of a living system, where the parts and the whole co-evolve through their encounter with each other and with the world. In a symmathetic system, intelligence is not located in any individual part. It is distributed across the connections — it lives in the quality of the relationships between things.

If this is true of organisations — and I believe the evidence from ecology, from developmental psychology, and from decades of organisational learning research suggests that it is — then the most important thing to evaluate about an organisation is not what its parts have produced. It is what the system as a whole has become capable of learning.

This is not an alternative methodology. It is an epistemologically honest response to what organisations actually are.

What does it mean to evaluate relational capacity?

Some years ago, working alongside organisations that were trying to do something genuinely difficult — knowing that they needed to change but unsure what that actually looked like — I began to notice that the most important developments were consistently invisible to conventional evaluation. Not because the evaluators were not looking, but because the questions they were asking could not find what was there.

What I was observing was not output. It was something more like capacity — the organisation's developing ability to hold things it had not previously been able to hold. To sustain an honest conversation about a persistent failure. To integrate the knowledge of frontline workers into decisions that had previously been made without them. To remain present to genuine disagreement without fracturing. To sit with uncertainty long enough for something genuinely new to emerge.

None of these capacities appear in a logframe. None of them can be measured against a predetermined indicator. And yet they are, I would argue, among the most important things that can happen in an organisation that is trying to serve people well in genuinely difficult conditions.

The question that eventually emerged from that observation — slowly, through accumulated experience rather than through a single insight — was this: what can this organisation now hold in relationship that it could not hold before?

When this question is taken seriously, people begin to describe things that conventional evaluation has never asked them about. The conversation that finally happened between the clinical team and the community workers, after three years of working alongside each other without genuine exchange. The moment a board member said something that nobody had been able to say in that room before, and the organisation discovered it could hold that honesty rather than deflecting it. The shift in how the organisation responded to a crisis — not the crisis management itself but the quality of collective sensemaking, the way different forms of knowledge found each other.

These are not anecdotes illustrating predetermined outcomes. They are evidence of something developing in the organisation's relational life — evidence that becomes visible only when you are asking a question that can find it.

What would it mean to build evaluation practice around this question? Not to replace existing approaches wholesale — there are genuine contexts where output measurement and outcome tracking serve real purposes — but to hold this question alongside them, as a different kind of inquiry into what an organisation is becoming capable of?

The methods are the easier part. The harder part is cultural. It requires evaluators willing to sit with genuine uncertainty rather than resolving it into a finding. It requires commissioners who understand that the most important learning may not be expressible in a report. It requires senior leaders who can hold honest inquiry without retreating into the defensive patterns that the machine model has made so familiar.

How many organisations are ready for that? More than the orthodoxy suggests, in my experience. Not enough, given what is coming. But the question itself — what can this organisation now hold in relationship that it could not hold before? — is not only an evaluative instrument. It is an invitation. It invites organisations to attend to their own relational life with the same seriousness they bring to their strategic plans and their program logic models. To treat the development of relational capacity not as a byproduct of good work but as the ground from which everything else grows.

That feels like the right place to begin.