From bounded rationality to ecological rationality
Gigerenzer has spent four decades dismantling the assumption that rational decision-making means maximising expected utility, and this paper reads as a late summation of that project — a direct engagement with Simon's legacy and a critique of how economists domesticated bounded rationality into a tame addendum to optimisation models.
The ecological turn is the key move: rationality is not a property of a mind in isolation but of a mind fitted to its environment, which means changing the environment can be as rational as changing the decision rule.
For product direction, the implication is structural: the question is not how to nudge individuals toward better choices but how to design the decision ecology — the organisation, the interface, the information environment — so that simple heuristics produce good outcomes.
The paper fills a genuine gap in the library's decision-making coverage, where Simon is present but his heirs (Gigerenzer, Kahneman's critics) are underrepresented.
Published in Industrial and Corporate Change, it sits at the economics-cognition intersection that the library has identified as underdeveloped.
Read alongside Simon's Administrative Behavior and Kahneman's Thinking, Fast and Slow for the full triangulation of this debate.
Central argument
Gigerenzer argues that bounded rationality, as Simon originally conceived it, was subverted by economists into a deficit model — a story of cognitive shortcomings that still aspires to the optimisation ideal — and proposes ecological rationality as the corrective: a mind is rational not when it maximises expected utility but when its heuristics are well-matched to the structure of its environment. The key thesis is that simple decision rules are not second-best approximations of full rationality but can outperform complex optimisation models precisely because they are robust to noise and overfitting in uncertain, real-world conditions. Rationality, on this account, is a relational property between a decision strategy and an environment, which shifts the normative question from 'how should minds calculate better?' to 'what environments make simple minds perform well?'
Critique
The ecological framing, while powerful, risks circularity: a heuristic is deemed rational when it fits its environment, but the environment is partly constituted by the decisions made within it, making it difficult to specify ex ante which heuristics are ecologically appropriate without already knowing the outcomes. Gigerenzer's case studies tend to draw on domains — medicine, finance, natural foraging — where the environment's statistical structure is relatively stable or recoverable; it is less clear how the framework applies to environments that are rapidly designed and redesigned, such as digital product ecosystems, where the ecology itself is the artifact under construction. This creates a tension the paper may not fully resolve between ecological rationality as a descriptive theory and as a normative guide for designers of those very environments.
Why it matters for product
For a CPO, the paper reframes the design problem entirely: rather than instrumenting individual user behaviour and nudging decisions at the margin, the strategic lever is the decision ecology itself — the information architecture, the team structure, the metrics dashboard — which determines whether the heuristics people naturally use produce good or poor outcomes at scale. Concretely, this applies to how discovery processes are structured: if product teams operate in environments saturated with ambiguous data and misaligned incentives, no amount of decision-quality training will compensate, whereas redesigning the environment — clearer prioritisation criteria, smaller option sets, faster feedback loops — can make simple judgment rules reliable. It also challenges the common product org instinct to add analytical complexity (scoring models, OKR hierarchies) when performance degrades, suggesting the diagnosis should first ask whether the decision environment has become too noisy for any heuristic to work, not whether the heuristics themselves are deficient.