How to actually develop a leadership competency

You've probably tried to develop a leadership competency the way almost everyone does: you read the book, took the course, sat through the two-day offsite with the breakout groups and the laminated model. You came out energized, full of language for the thing you wanted to be better at — and three weeks later you were behaving exactly as before. Not because you're slow, and not because the content was wrong. Because that isn't how competencies develop. A workshop transfers knowledge about a competency. It does almost nothing to change the behavior that the competency actually is.

This is the most expensive misunderstanding in professional development, and it's baked into how companies spend their training budgets. We treat leadership competencies as if they were information to be downloaded — sit still, absorb the framework, leave improved. But a competency, as the companion piece in this pillar argues, is just a cluster of behaviors. And behaviors don't change because you understood a slide. They change through a specific, slightly unglamorous loop that you run on yourself, in the actual work, over weeks.

This piece is that loop.

Why knowing doesn't become doing

Start with the gap that workshops never close: the one between knowing what good looks like and doing it under pressure. Almost every manager who interrupts too much knows they should let people finish. Almost everyone who avoids hard conversations knows avoidance costs more later. The competency isn't missing from their knowledge — it's missing from their behavior in the three seconds where it counts, when the old pattern fires faster than the intention.

Closing that gap is a practice problem, not a comprehension problem. The research on how people actually reach high performance is unusually clear about this. Anders Ericsson, who spent his career studying expertise, found that what separates top performers isn't time logged or talent but deliberate practice — focused repetition of a specific sub-skill, at the edge of your current ability, with immediate feedback on whether you got it right. Note the two ingredients a workshop lacks entirely: repetition in the real situation, and feedback on each rep. You can't develop a behavior in a room where the behavior never has to happen.

The loop, in four moves

Developing a competency is a cycle you run deliberately. Four moves, repeated.

1. See how you actually behave. This is the move everyone skips, and skipping it dooms the rest. You cannot work on a behavior you can't see, and the behaviors that make up your competencies are largely invisible from the inside — you experience your intentions while everyone else experiences your conduct. That's why some skills are impossible to develop alone: the input has to come from outside your own head. Before you change anything, get an honest read on how you currently behave — your own assessment first, then the people who actually see you work, and the gap between the two as your real starting line.

2. Pick one behavior — not the competency. "Get better at communication" is not a target you can practice. "Stop opening meetings with my own opinion" is. The single most common reason competency development stalls is that people try to improve the whole abstraction at once, which gives them nothing concrete to do on Monday. Decompose the competency into its behaviors, find the one with the highest return, and work on only that until it's automatic. One behavior, deliberately changed, beats five behaviors vaguely intended.

3. Practice it in the real situation, with a cue. Deliberate practice for a leader doesn't happen in a simulator; it happens in your actual next meeting. The trick is to attach the new behavior to a specific, recurring trigger so it fires before the old pattern does — "when someone brings me a problem, I ask one question before I offer a solution." The cue is what turns an intention into a rep. Each real situation where you hit the cue is one repetition; each one where you miss it and notice is feedback. This is slow and a little awkward, and that awkwardness is the feeling of a behavior actually changing rather than a model being admired.

4. Re-measure. Here's the move that separates development from good intentions: after some weeks of practice, ask the same people the same questions again, and look at whether the behavior actually moved in their eyes — not yours. Improvement you can only feel is indistinguishable from improvement you imagined. The delta between your first read and your second is the only evidence that the loop worked, and it tells you whether to keep going on this behavior or move to the next one.

Why you can't do this with a mirror alone

The honest objection at this point is: can't I just watch myself? Mostly, no — and not for lack of trying. The reason is the same one that makes the first move non-negotiable: the parts of your behavior that matter most are the parts you're least able to observe in yourself, because you're inside them.

The surgeon and writer Atul Gawande made this point about people at the absolute top of their field. In Personal Best, he describes bringing in a coach to watch him operate after his results had plateaued — and the coach, just by watching, saw small habits costing him performance that Gawande, after years of practice, could no longer detect in himself. If a world-class surgeon needs an outside observer to keep developing, the manager who thinks they can self-assess their way to better leadership is kidding themselves. The outside eyes aren't a nicety in this loop. They're the sensor that makes the loop a loop instead of a guess.

A worked example

Say the competency is delegation, and your starting read — your own answers next to your team's — shows the gap is real: you think you delegate well, they experience you taking work back the moment it gets hard.

You don't work on "delegation." You decompose it and find the highest-return behavior: you re-absorb tasks at the first sign of trouble instead of coaching through them. So the one behavior becomes — when a delegated task starts to wobble, ask "what have you tried?" before touching it myself. The cue is the wobble; the rep is the question. You practice it for six weeks, badly at first, catching yourself reaching for the keyboard. Then you re-measure: same people, same question about how you delegate. If their read has shifted, the competency moved — not because you learned about delegation, but because you changed one behavior enough times that it stuck. If it hasn't, you've learned something equally useful: the behavior you picked wasn't the one that mattered, and you choose again.

That's the whole method, and it generalizes to any competency on any framework. Decompose, pick one, attach a cue, practice in the real moment, re-measure. Repeat.

How long this actually takes

Slower than a workshop promises, faster than you fear — but only if you respect the constraint that makes the loop work. A behavior that's fired the same way for years doesn't rewire in a week; give a single behavior a real run, which usually means a quarter of catching yourself at the cue, missing, and trying again. That's also why you work on one at a time. The instinct after an honest read is to fix everything at once — but five behaviors practiced vaguely is how nothing changes, while one behavior practiced until it's automatic is how the next one gets easier.

It's also why re-measuring belongs on a considered cadence, not a constant one. Asking the same people the same questions every few weeks trains them to rate you, not to notice real change; asking once a quarter or twice a year gives the behavior time to actually move before you check. Development has a tempo. Run the loop faster than the behavior can change and you just generate noise.

Where to start

The loop needs an instrument — something to take the first read, hold the behaviors you're working on, and let you re-measure later against the same baseline. That's what Mirorly is for. You run a core leadership behaviors round on yourself first, send the same behaviorally-anchored questions to the people you work with, and read your self-view next to theirs — which surfaces the gap that tells you which behavior to work on. Then, months later, you run the same round again, and the change is right there in the delta, in their answers, not just in how you feel. The framework tells you what competencies exist. This is the part that actually develops one.

The loop, in one place

• Workshops transfer knowledge about a competency; they don't change the behavior the competency is made of. Stop expecting them to.
• Development is deliberate practice — focused reps of one behavior, in the real situation, with feedback — not passive learning.
• See how you actually behave first. You can't practice a behavior you can't observe, and you can't observe your own from the inside.
• Pick one behavior, not the whole competency. Attach it to a cue so it fires before the old pattern.
• Re-measure with the same people and questions. The delta in their answers is the only proof the loop worked.
• Then pick the next behavior, and run it again.