The Confidence Trap: Why Overconfidence Costs Leaders More Than They Realise
We talk a lot about confidence as a leadership asset. But here is a problem we do not name often enough: the people who sound confident are not always competent. And for leaders in new roles, that is a real risk — because most of us assume that if someone speaks decisively, they must know what they are doing.
The data says that assumption is wrong.
I saw this play out with a new CEO I was coaching. I gave him one clear instruction for his first three months: do not change anything. Just listen. Meet people, learn how the business actually works, and offer up just one suggestion of something they could change if it felt right. Within the first month, he met the leadership team and the board and proposed 72 changes. 72. He was trying to prove his competence, but what he showed instead was confidence without context. Within weeks, the board moved him out of the CEO role. Their message was blunt: maybe get to know the business first.
The issue was not a lack of intelligence. It was overconfidence — too early, too loud, too fast.
And this is where leaders often get it wrong. We assume overconfidence disappears when people get feedback. It does not. A study published in Psychological Science looked at 3,000 tournament chess players — an environment with constant, accurate public feedback. Everyone knows exactly how good they are, or should. But the researchers found that the lowest-performing players were the most overconfident by a significant margin. The gap between confidence and competence was large enough to reliably predict losing.
This builds on the Dunning-Kruger effect. People who lack skill also lack the insight to recognise their limits, which means telling them the data says otherwise rarely works — they do not yet have the mental map to interpret it.
So what does this mean for leaders? First, do not confuse confidence with readiness — especially in new roles. Second, stop trying to correct overconfidence with blunt feedback alone. That is not how calibration happens.
Instead, watch for this signal. Real confidence sounds like: here is my initial view — what am I missing? I want to test this before landing it. Who should I be consulting here? Bravado, on the other hand, sounds like certainty.
Real confidence is not about sounding certain. It is about understanding the context you are operating in, what you know, what you do not, and where your capability actually sits inside that system. When you are confident because you understand the landscape, you do not rush to conclusions. You design a process. You consult. You test your thinking before you land it. That is not hesitation. It is not weakness. It is far more impressive than bravado.
