Posted by 
Why the next frontier of management is institutional judgement
You have almost certainly sat in the meeting.
A strategy is being reviewed. The numbers are assembled, the slides are clean, the language is confident. Around the table sit fifteen people, most of whom privately suspect the plan stopped describing reality months ago. Nobody says so. Not because they are cowards or lack the analysis, but because the room has already told them what contribution it wants.
The meeting ends. The strategy continues. A group of demonstrably intelligent people has just produced a decision none of them would individually defend.
Institutions full of capable people routinely produce collective judgement worse than the people inside them. We call this an anomaly – a failure of leadership, a lapse in culture, a bad quarter. I think it is the default.
We have optimised the wrong variable
For more than a century, management has been an exercise in raising institutional capability: better people, more data, sophisticated systems, redesigned structures, transformation after transformation. The modern organisation is, by any measure, more capable than its predecessors.
And still: strategies persist long after their assumptions have failed. Weak signals die on the way up through hierarchy. Expertise is overridden by incentive or status. People learn – quickly, and without being told – when not to challenge. Institutions grow more certain as reality moves away from them.
Capability is not the binding constraint. We have spent a century making organisations cleverer without asking whether cleverness was ever in short supply.
Drucker got here first
None of this would have surprised Peter Drucker: the argument that institutions must protect disagreement is his, not mine.
In The Effective Executive he set the rule that no decision should be made in the absence of disagreement, and told the story of Alfred Sloan adjourning a General Motors committee that had arrived at unanimity – postponing it, in Sloan’s phrase, to “develop disagreement” and work out what the decision was actually about. Drucker’s point was not that consensus is impolite. It was that agreement arriving early is evidence of an incomplete decision.
This is continuity, not discovery. My addition is that Sloan’s remedy no longer scales.
Sloan could manufacture disagreement because the decision and the knowledge sat in the same room, and he controlled it. Today the decisive observation is rarely at the table. It is four levels down, held by someone with the expertise to see it and no standing to be heard, and it must survive several rounds of summarisation to reach anyone with authority. The harder problem is no longer manufacturing dissent at the top. It is engineering the passage of inconvenient information through a system built for compression.
Naming the thing, carefully
The phenomenon is not unstudied. Janis named groupthink. Argyris showed how organisations learn to defend themselves against learning. Morrison and Milliken described organisational silence. Edmondson demonstrated that psychological safety determines whether people speak. Woolley and colleagues showed that groups have a measurable collective intelligence.
My claim is narrower than novelty. Most of this work concerns a single link in the chain – whether someone speaks, or why they don’t – rather than whether what they said survives four layers of summarisation and reaches a decision. What we lack is a treatment of the whole chain as one system property: observation, transmission, interpretation, decision, revision. Psychological safety governs whether someone speaks. It says nothing about whether the observation arrives intact, or changes anything when it lands. Separate links – and an institution can be strong at one and useless at the next.
I call the chain institutional judgement capability: an institution’s emergent ability to turn distributed observation, expertise, intelligence and disagreement into coherent judgement and coordinated action. The proposition I would put to this Forum is that it obeys a limit:
No institution can consistently produce collective judgements that exceed the quality of the conditions under which those judgements are formed.
Recently I worked with a technology subsidiary: ten years of growth, then two years of falling revenue in a market reordered by AI. Asked for a strategic direction, the executive team kept producing targets – half of revenue from products, growth in a named vertical. Every person in that room could describe, privately and accurately, why their proposition had stopped landing with customers. None had said so collectively. The parent’s targets had arrived first, and against a target, a contrary diagnosis sounds like an excuse.
Why ingenuity makes this urgent
This Forum has chosen ingenuity, and ingenuity is where the constraint bites hardest. Ingenuity is not scarce. In most organisations I have led or supported, the good idea already exists somewhere in the building. The scarce thing is an institution capable of receiving it.
AI sharpens this, and not in the direction we expect. We are about to make extraordinary intelligence available at negligible cost – including, for the first time, dissent on demand. A leader who wants to appear to have tested a plan can generate the strongest counter-argument in seconds, read it, and remain unmoved. Drucker’s test – was there disagreement? – becomes trivially easy to pass and much harder to fail honestly.
Intelligence has never been judgement. Poured into an institution that cannot metabolise dissent, it yields better-argued versions of the decisions it was always going to make – greater certainty, arrived at faster.
What follows
The instinct is to reach for a framework. Resist it. The honest first move is subtraction: most institutions do not need a judgement process bolted on top. They need to remove the conditions degrading the judgement they already have.
Three questions make a start. They are diagnostic, not rhetorical. Of your last significant strategic decision: what did you learn that was uncomfortable? Whose mind actually changed? And how
far did the most inconvenient evidence have to travel before someone with authority saw it – if it arrived at all?
If those questions are hard to answer, that is the finding.
The next frontier of management may not be making institutions more capable. It may be making them more honest.
About the author:
Martin Shelford is an experienced business leader, founder, and advisor, and the author of works including The Shortboard Business. His current work centres on institutional judgement capability – an institution’s ability to turn distributed intelligence and disagreement into decisions it would defend.