He could describe his pattern with clinical precision. "When I feel cornered in a meeting," he told me, "I default to control. I interrupt. I talk over people. I make the decision myself so nobody else can make it badly." He paused. "I know I do this. I have known for years. I still do it every time."
This is the most common conversation in coaching. Not the one where a person discovers something they did not know about themselves. The one where a person already knows — has known for some time — and nothing has changed.
The assumption behind most development programmes is that insight produces change. Understand the problem, and the problem begins to resolve. Identify the pattern, and the pattern weakens. This assumption is so deeply embedded in how we think about human improvement that we rarely examine it directly.
It is wrong. Or rather, it is radically incomplete.
The gap that meta-analyses measure
Psychologists call it the intention-behaviour gap. It is the measurable distance between what a person intends to do and what they actually do when the moment arrives. Meta-analyses consistently show that even strong, clearly articulated intentions explain less than one-third of actual behaviour. Two-thirds of what people do is governed by something other than what they have decided to do.
This finding should be more alarming than it is. It means that the majority of human behaviour under real conditions is not controlled by conscious decision. It is shaped by habit, by emotional response, by the social dynamics of the room, by physiological state, by the cognitive resources available at the moment of action. Insight — the conscious understanding of what one should do differently — enters a competition it frequently loses.
The mechanism is not mysterious. Behaviour change requires self-regulation, and self-regulation requires executive function — the same prefrontal cortex capacity that is consumed by pressure, emotional suppression, fatigue, and the dozen other demands of any consequential professional moment. The person who understands their pattern in the calm of a coaching conversation is drawing on cognitive resources that may not be available when the pattern activates.
This is the insight trap: the belief that understanding a behaviour is sufficient to change it. It is a trap because the understanding itself feels like progress. The person has done the difficult cognitive work of recognising the pattern. They can articulate it, analyse it, trace its origins. The insight is real. What is missing is not awareness but the capacity to act on awareness under the conditions that trigger the behaviour.
Understanding is not the same as overriding
There is a specific neurological reason why insight alone fails. The behaviours that coaching and development programmes most commonly target — defensive reactions, communication habits, decision-making patterns under stress — are not stored in the same cognitive systems that produce insight.
Insight is a product of the prefrontal cortex — the slow, deliberate, analytical system that can examine behaviour from a distance. The behaviours themselves, particularly those formed over years of professional life, are stored in subcortical systems — the basal ganglia, the amygdala, the procedural memory networks that execute familiar routines without conscious deliberation.
When a leader "defaults to control" under pressure, they are not making a decision. They are running a programme. The programme was written by years of experience in which controlling the outcome produced safety — or at least the feeling of safety. It is fast, automatic, and largely invisible to the person executing it. The insight, by contrast, is slow, deliberate, and available primarily when the person is not under the conditions that trigger the programme.
This is why the same leader can describe their pattern perfectly in a Monday morning coaching session and repeat it precisely in a Tuesday afternoon board meeting. The coaching session activates the analytical system. The board meeting activates the procedural one. The two systems are not in dialogue. They operate on different timescales, in different brain regions, in response to different cues.
The rehearsal gap
If insight does not produce change, what does?
The research on habit formation and behaviour change points to something less elegant than self-knowledge but more effective: rehearsal under conditions that approximate the trigger. Not insight, but practice. Not understanding, but repetition in the right context.
This is what the best sport psychology has understood for decades. A goalkeeper does not become better at penalty saves by understanding the psychology of penalty shootouts. They become better by facing penalties — hundreds of them, in training conditions that simulate the pressure, the crowd noise, the time distortion that accompanies high-stakes moments. The understanding may help. But the understanding without the rehearsal changes nothing in the moment that matters.
In coaching, this means something specific and often uncomfortable. The conversation about the pattern is necessary but not sufficient. The client who can articulate their defensive response is at the beginning of the work, not the end. The actual work involves creating conditions — in the coaching room, in structured practice, in graduated real-world situations — where the new behaviour can be executed repeatedly until it becomes as automatic as the old one.
This is harder than insight. It is slower. It requires the client to feel the discomfort of the trigger and practise a different response while that discomfort is present. It requires the coach to move from the role of interpreter — the person who helps the client understand — to the role of trainer, the person who helps the client build a new procedural response.
Why organisations invest in the wrong intervention
Most corporate development programmes are built on the insight model. A leadership programme identifies the behaviours that distinguish effective leaders. Participants are assessed against this framework. They receive feedback. They develop "action plans." They return to their roles with a new understanding of what they should do differently.
The insight is usually accurate. The action plans are usually sensible. And meta-analyses of leadership development programmes show, consistently, that the transfer rate — the proportion of learning that actually changes on-the-job behaviour — hovers between ten and twenty per cent.
Eighty per cent of what people learn in leadership programmes does not survive contact with the conditions of actual leadership.
This is not because the programmes are poorly designed. Many are excellent. It is because they are designed around the wrong theory of change. They assume that providing the right information and creating the right insight will produce the right behaviour. The research says otherwise. Information campaigns rarely change behaviour on their own. Knowledge increases motivation but does little to move people from wanting to doing.
The interventions that do change behaviour share a common feature: they create conditions for repeated practice of the target behaviour in contexts that activate the relevant triggers. Simulation-based medical training, where surgeons practise high-stakes procedures under realistic time pressure. Military stress inoculation, where soldiers rehearse decision-making under conditions that approximate combat. The principle is the same. The behaviour must be built in the conditions where it will be needed, not understood in the conditions where it is easy.
The honest conversation
In my practice, the most important moment in coaching is not the moment of insight. It is the moment after insight, when the client realises that knowing is not enough.
This is often a difficult moment. The client has invested significant cognitive and emotional effort in understanding their pattern. They expect the understanding to do the work. When it does not — when they find themselves repeating the behaviour they can now describe with precision — the temptation is to conclude that the insight was wrong, or insufficient, or that they need a deeper level of understanding.
They do not. They need a different kind of work entirely. They need to practise the alternative behaviour under conditions that are close enough to the real trigger to activate the procedural system, but structured enough to allow a different response. They need repetition, not revelation. They need the patience to build a new automatic response alongside the old one, knowing that the old one will continue to fire for some time, and that the new one will feel awkward, effortful and unconvincing long before it feels natural.
This is unglamorous work. It does not produce the satisfaction of the breakthrough moment, the flash of self-understanding that makes a coaching conversation feel transformative. But it is the work that actually changes behaviour. And until development programmes, coaching practices and organisations accept that insight is the beginning rather than the destination, we will continue to produce leaders who understand their patterns perfectly and repeat them without interruption.
Every development programme asks one question:
Does this person understand what they need to change?
The harder question, which far fewer programmes address:
Can they execute the change under the conditions that make it most difficult?
The gap between those two questions is not a failure of self-awareness. It is a failure of method. And the method — practice, repetition, rehearsal under pressure — is available. It is simply less satisfying than the alternative of sitting in a room and having an important conversation about oneself.
ALTIUS Note
This article belongs to the ALTIUS editorial library: evidence-based essays on Human Performance, Leadership and Society.
Tags: Psychology, Human Performance, Education
Selected Research Foundations
Sheeran, P., & Webb, T. L. — "The Intention–Behavior Gap" https://doi.org/10.1111/spc3.12265
Graybiel, A. M. — "Habits, Rituals, and the Evaluative Brain" https://doi.org/10.1146/annurev.neuro.29.051605.112851
Lacerenza, C. N., et al. — "Leadership Training Design, Delivery, and Implementation: A Meta-Analysis" https://doi.org/10.1037/apl0000241
Blume, B. D., et al. — "Transfer of Training: A Meta-Analytic Review" https://doi.org/10.1177/0149206309352880
Wood, W., & Rünger, D. — "Psychology of Habit" https://doi.org/10.1146/annurev-psych-122414-033417
