Naming Impact: Counting What Counts in Social Enterprise

Efficiency, perception, and the value of what we fail to measure

In a lecture hosted by the Adam Smith Institute, Rory Sutherland offers a deceptively simple insight: progress often comes not from doing things better, but from seeing things differently. For social enterprise, this observation is more than rhetorical. It challenges the very foundations of how impact is defined, measured, and governed.

Drawing implicitly on Adam Smith and more explicitly aligned with Iain McGilchrist’s distinction between analytical and contextual modes of thinking, we are confronted with a central tension: social systems are increasingly managed through left-brain logics of efficiency and measurement, while the outcomes they seek—dignity, belonging, agency—are inherently right-brain phenomena.

This tension becomes particularly acute in an era shaped by AI, automation, and data-driven governance. If we are not careful, we risk building systems that measure activity perfectly while missing impact entirely.

The Measurement Problem: When Outputs Replace Outcomes

Across social enterprise sectors—disability services, sustainability ventures, community programmes—performance frameworks tend to privilege:

• outputs (units delivered, hours billed, services completed)
• efficiency (cost per beneficiary, throughput)
• standardisation (replicable, scalable models)

These are not inherently problematic. However, they become dangerous when they are mistaken for impact itself.

Consider disability services within the NDIS context. A provider may report:

• hours of support delivered
• punctual attendance
• tasks completed

Yet none of these metrics capture whether a participant:

• feels more confident
• leaves the house independently
• forms social connections
• regains a sense of identity

A participant who, months earlier, could not leave their home but is now baking cakes for others has undergone a profound transformation. The system, however, may still see only “hours delivered”.

This is the central failure: we measure what is visible, not what is valuable.

The Efficiency Trap: From Useful Metric to Systemic Distortion

This failure follows a recognisable trajectory:

• Toy: a useful metric is introduced (e.g. hours of service)
• Target: the metric becomes a goal
• Tradition: it becomes institutionalised
• Trap: it replaces the original purpose

At this point, organisations optimise the proxy rather than the outcome.

In international development, counting wells built led to thousands of non-functional systems.
In sustainability, measuring tonnes recycled can obscure whether consumption behaviour has actually changed.
In disability services, billing precision can displace relational continuity.

The pattern is consistent: what is easy to count becomes what counts.

Sub-Optimising the Parts to Optimise the Whole

This leads to a counterintuitive but essential principle:

To optimise the whole, you must sub-optimise the parts.

In mechanical systems, optimisation is additive. In human systems, it is not.

Disability Pathways Example

A support worker who spends an additional ten minutes in unstructured conversation may appear inefficient within a tightly optimised schedule. Yet that moment may:

• build trust
• surface hidden concerns
• prevent escalation
• increase long-term independence

From a narrow lens, this is inefficiency.
From a systems perspective, it is high-leverage intervention.

Sustainability Enterprise Example

A circular fashion enterprise that:

• slows production
• invests in storytelling
• prioritises design quality

may produce fewer units. Yet it may:

• shift consumer identity
• increase willingness to pay
• normalise sustainable consumption

Again, the part is sub-optimised; the whole is strengthened.

The AI Problem: Mispricing Human Work

The rise of AI intensifies this issue. Automation systems:

• excel at codified tasks
• optimise repeatable processes
• scale standardised outputs

But they struggle with:

• tacit knowledge
• relational judgement
• contextual adaptation
• micro-innovation

Many roles in social enterprise—support workers, community facilitators, programme leads—derive much of their value from these invisible contributions.

This is where the logic of kaizen becomes instructive. In the Toyota system, continuous improvement emerges not from top-down design, but from frontline workers identifying small inefficiencies and making incremental changes. These micro-inventions rarely appear in formal metrics, yet they accumulate into systemic advantage.

If organisations value labour only by its documented outputs, they risk:

• overestimating what AI can replace
• underestimating what humans actually contribute

The danger is not simply technological displacement. It is conceptual misunderstanding of work itself.

Observation Beyond Automation: Recognising the Unexpected

A deeper issue sits beneath measurement and automation: the nature of observation.

In many systems, observation is treated as:

• data capture
• process monitoring
• compliance verification

Yet true observation involves:

• recognising emerging value
• noticing unexpected change
• interpreting subtle shifts in behaviour

Returning to the earlier example: the participant baking cakes is not merely completing a task. They are:

• expressing agency
• contributing socially
• reshaping identity
• experiencing joy

These are not incidental outcomes. They are the impact.

Yet they often remain unmeasured because they are:

• qualitative
• emergent
• difficult to standardise

Reframing Impact Through Social Enterprise Sherpa

The Social Enterprise Sherpa offers a useful way to structure this challenge.

Rather than asking:

• What can we measure?

It asks:

• What value is emerging that we have not yet named?
• What would disappear if this role were removed?

• What small changes signal deeper transformation?

This shifts the focus from:

• measurement as control
  to
• measurement as sense-making

A Multi-Dimensional View of Impact

To “count what counts”, social enterprises must expand their lens:

Most organisations operate comfortably in the first dimension.
The real work—and the real value—lies in the latter three.

Linking to the Sustainable Development Goals (SDGs)

The United Nations Sustainable Development Goals already anticipate this complexity through their structure of goals, targets, and indicators.

Relevant alignments include:

• Goal 3: Health and Wellbeing (lived capability, not just service access)
• Goal 8: Decent Work (quality and meaning of work, not just employment rates)
• Goal 10: Reduced Inequalities (participation and inclusion)
• Goal 12: Responsible Consumption (behavioural change, not just recycling volume)
• Goal 17: Partnerships (system-level coordination)

The SDG framework implicitly recognises that:

• outputs are insufficient
• outcomes must be interpreted
• systems must be understood holistically

A Practical Checklist: Counting What Counts

Naming

• Does our language reflect aspiration or deficit?

Measurement

• Are we measuring outputs, outcomes, or identity shifts?

Labour Value

• What tacit knowledge exists in this role?

AI Exposure

• What tasks are automatable?
• What human contributions are not?

Relational Impact

• Are trust and dignity visible in our metrics?

Trap Detection

• Has a useful metric become the goal?

Conclusion: The Moral Choice of Measurement

If Adam Smith showed that markets depend on human behaviour, and Rory Sutherland reminds us that behaviour depends on perception, then social enterprise must recognise a deeper truth:

Measurement is not neutral. It encodes a theory of what matters.

To count what counts is therefore not simply a technical task. It is:

• a strategic decision
• a narrative act
• a design challenge
• and ultimately, a moral commitment

The question is no longer whether we can measure impact.

It is whether we are willing to see it properly in the first place.

Final Reflective Question

• If your organisation measures service delivery perfectly, but cannot detect confidence, dignity, participation, or joy, are you measuring impact—or merely administration?

References

• Sutherland, R. (2024). Behavioural Science and Adam Smith Lecture. Adam Smith Institute. https://youtu.be/md9UAorQ_B0
• Smith, A. (1776). The Wealth of Nations.
• McGilchrist, I. (2009). The Master and His Emissary. Yale University Press.
• Madhavan, G. (2021). Wicked Problems. W.W. Norton & Company.
• United Nations. Sustainable Development Goals. https://sdgs.un.org/goals