Why Behaviour is so Difficult to Evaluate

Executive Summary

Behavioural performance evaluation South Africa challenges stem not from poor system design but from the inherent nature of behaviour itself — this is the central argument of Part 2 in Camille Rabier’s five-part series on rethinking performance. Having established in Part 1 that performance systems are structurally incomplete when they assess outcomes without assessing how those outcomes are achieved, this second article explains why that gap has proved so persistent and so difficult to close.

Rabier identifies five properties of behaviour that together make it genuinely resistant to consistent evaluation: it is context-dependent, meaning the same behaviour can be effective in one situation and counterproductive in another; it is interpretive, meaning observation does not produce a single objective reading but a range of judgements shaped by experience and cognitive bias; it is multi-dimensional, made up of multiple underlying elements that cannot be cleanly isolated; it is continuous, unfolding over time rather than at measurable points; and it is entangled with outcomes, meaning the two cannot be separated even though they are not interchangeable. Together these properties explain why performance systems default to outcomes — not because organisations do not value behaviour, but because behaviour is genuinely harder to specify, observe and compare than outcome data. This insight page extends that analysis into what it means for South African organisations designing or reforming their performance frameworks.

Key Strategic Takeaways

Reframed for South African HR and Remuneration Leaders:

• The difficulty of evaluating behaviour is inherent, not incidental. Behaviour is context-dependent, interpretive, multi-dimensional, continuous, and entangled with outcomes — each of these properties individually makes consistent evaluation harder; together they explain why most performance systems simply stop at outcomes.
• Cognitive bias is a structural feature of behavioural evaluation, not an individual failure. Because behaviour is interpreted rather than simply observed, individual managers bring different lenses — different experience, different priorities, different tolerance for ambiguity — which is why consistency requires deliberate design, not good intentions.
• The same behaviour can represent good performance in one context and poor performance in another. Context-dependence means there is no universal behavioural standard that can be applied across roles, seniority levels, or team dynamics — any evaluation framework must account for this rather than pretend it away.
• Outcomes cannot be used as proxies for behaviour. Strong outcomes can be achieved through behaviours that are unsustainable or damaging; effective behaviours may not produce immediate visible results. The same outcome can reflect very different underlying approaches — and these distinctions matter for long-term performance.
• The goal is not to make behaviour objective but to make its evaluation more disciplined. Accepting that behaviour involves interpretation is not a concession — it is the starting point for designing evaluation systems that structure and anchor that interpretation rather than leaving it entirely to individual managerial discretion.

21st Century Commentary – Insights

What the Properties of Behaviour Mean for South African Performance System Design

Rabier’s analysis gives precise language to a set of difficulties that 21C’s performance management and climate survey work consistently surfaces: the gap between what organisations say they value and what they actually differentiate in practice is not explained by a lack of intent, but by the genuine design challenge of making behaviour evaluable. This distinction matters because it shifts the intervention from culture change — telling people to behave better or value behaviour more — to system design: how do you build a performance framework that can actually capture and compare behavioural contribution?

Inconsistency as a systemic rather than personal failure

In 21C’s engagement and performance calibration work, inconsistent evaluation is among the most common sources of perceived unfairness reported by employees. Individuals delivering similar results are rated differently; similar behaviours are assessed differently by different managers. Rabier’s analysis explains why this happens even when managers are trying to be fair: without shared criteria and a structured approach to interpreting context, judgement diverges. The implication for South African organisations is that consistency cannot be achieved through manager training alone — it requires the kind of structural operationalisation that Parts 3 and 4 of this series describe.

The entanglement of behaviour and outcomes in the South African pay equity context

South Africa’s Employment Equity Act and the evolving pay transparency environment make the entanglement of behaviour and outcomes a governance issue, not just a management challenge. Where organisations cannot explain how performance ratings were reached — because behavioural criteria were absent or inconsistently applied — differentiated pay decisions become vulnerable to challenge on grounds of subjectivity or unfairness. 21C’s RemCom and HR advisory work increasingly involves supporting organisations to articulate and document the evidence base for performance differentiation, and outcome-only systems routinely fail to produce the documentation that makes those decisions defensible.

South African Business Implications

The cognitive bias dimension Rabier identifies has a specific South African resonance: in organisations managing transformation-related diversity, the risk that managerial interpretation of behaviour is unconsciously influenced by factors outside of performance — familiarity, communication style, cultural norms around directness or authority — is material and well-documented. Performance systems that leave behavioural evaluation entirely to individual managerial discretion, without shared criteria or calibration mechanisms, are more, not less, exposed to this risk in the South African context. Structured behavioural evaluation frameworks reduce this exposure, not by removing judgement but by anchoring it.

King IV’s governance expectations around ethical leadership and organisational culture connect directly to the continuous and context-dependent nature of behaviour Rabier describes: culture is not an event but a pattern of behaviour expressed over time, and governance oversight of culture requires the ability to assess that pattern rather than rely on point-in-time observations. South African boards overseeing organisational culture as a governance responsibility should be asking whether their performance systems can actually capture behavioural contribution over time — or whether they are, by design, measuring only the most recent and visible outcomes.

Related Insights

Related Services

Attribution