A six-paper programme on the Orientation-Attention-Interpretation system, the bounded-interpretation account of strategic cognition, and the empirical methods for recovering decision culture and friction from the artefacts organisations already produce.
A recursive affect-cognition architecture linking orientations, valence, attentional selection, and the formation of valenced mental models that govern closure and override.
Strategic judgment is bounded not only by computation but by interpretation. An affective-motivational architecture determines what can register as evidence before analysis begins, and interpretations stabilise not when they achieve informational sufficiency but when they achieve affective coherence sufficient to act.
This reframes classical bounded rationality. Simon's satisficing explains computation under limits. Cohering explains closure under ambiguity. The two mechanisms are distinct, and most strategic decisions of consequence operate on the second rather than the first.
Orientation. Enduring affective-motivational architectures that carry valence toward particular categories of strategic evidence.
Attention. The valence-weighted selection mechanism that determines which environmental cues achieve sufficient salience to capture cognitive processing.
Interpretation. The formation of valenced mental models that integrate attended cues into coherent accounts of the strategic situation and that stabilise through affective fit rather than through logical adequacy.
"Bounded interpretation explains why two equally capable strategists, given identical information, can perceive different realities — and why the resulting interpretive divergence produces persistent differences in where firms search, when they commit, and how readily they adapt."
How do stable orientations shape what strategists notice, how they read it, and when a reading feels settled enough to act? Three studies put that question to experienced professionals through the Strategic Judgment Study. A fourth puts it to AI models.
The 60-item Strategic Orientation Inventory and three implicit association tests map orientations toward uncertainty, risk, exploration, autonomy and time, at the explicit and the implicit level.
Six strategic scenarios and 36 pieces of case information test whether orientations steer which information a person treats as important, urgent and relevant, and how people respond when shown how colleagues with contrasting orientations rated the same information.
Randomised decision scenarios test how choices, confidence and timing change when a decision fits or cuts against a person's orientation, and whether the effect is strongest when the options share no common measure.
The same kind of choice, put to large language models. See the AI research below.
About the Strategic Judgment StudyExecutives increasingly consult AI models on strategic questions. Many of those questions have no right answer: the options are on a par, and choosing between them is a matter of judgement. The AI Parity Benchmark asks how models close such choices, and whether they do so by an orientation of their own.
Leading open and closed models are given strategic choices built to sit at parity, alongside matched choices with a clear better option. The benchmark measures which way each model leans, how consistently, and how far that lean moves when the model is told about the person it is advising.
Each phase is preregistered on the Open Science Framework before it runs, with a held-out set of items kept back from publication.
If a model resolves parity choices by a consistent orientation, its advice is never neutral on exactly the decisions where neutrality matters most. The work connects directly to the human studies: the same domains, the same kinds of choice, and the same question about how a reading settles.
In preparation: "Closure by orientation: how language models resolve choices with no right answer."
Six interconnected papers developing the framework empirically, theoretically, and methodologically.
Cognitive signatures, process burden, and the architecture of organisational judgment. The empirical core of the programme.
A processual construct of organisational judgment distinct from climate, governance, and leadership style.
When decision burden serves judgment and when it undermines it. An empirical account of proportionality.
Configurations of decision culture and their consequences for organisational adaptation. Empirical test of the seven archetypes.
Development, validation, and psychometric properties of the multi-method Decision Culture and Friction Instrument.
Decision culture and friction at the board-management interface. The governance application of the framework.
The research programme has emerged from twenty years of senior strategy consulting practice combined with doctoral research in strategic cognition. The combination of practitioner experience and academic training is deliberate: the framework is intended to be rigorous enough to satisfy the standards of peer review and legible enough to be actionable in the room where the decision is actually being made.
The empirical programme is built around the Strategic Judgment Masterclass, which serves simultaneously as an executive development offering and as the data collection vehicle for the academic research. Every masterclass cohort contributes to the anonymised research corpus under independently approved ethics arrangements.
The academic infrastructure includes the Gorilla experiment platform for the five-session online study (surveys, implicit association tests, decision scenarios and randomised assignment); Firebase for the dyadic exercise audio capture; JASP and R for statistical analysis; and a multi-platform data pipeline that preserves the separation between commercial analytical outputs and academic research findings.
The programme maintains a disciplined separation between commercial client engagement and academic publication: commercial clients receive diagnostic value under one contractual arrangement, and anonymised aggregated findings flow to the academic programme under a separable research participation agreement.
Prior research on strategic cognition has relied almost exclusively on self-report. Decision culture, in particular, is exactly the kind of phenomenon that self-report cannot reliably measure, because the people whose deliberation would be measured are the same people whose orientations constitute the deliberation.
The programme's methodological contribution is a multi-method protocol for recovering cognitive and process signatures directly from the decision artefacts organisations already produce: board packs, strategy memos, investment committee papers, minutes, and post-investment reviews. The protocol combines LLM-assisted content analysis calibrated against human-coded ground truth, structured survey instruments, and episode-level decision audits. Together they produce evidence that is legible to both practitioners and reviewers, and that does not depend on the self-report that existing culture instruments cannot escape.