This transcript was produced by robots and left as-is. Accuracy and elegance are not guaranteed.

You notice something with one of your employees, maybe a peer. You say nothing.

Not conflict avoidance. Just too early to be sure. So you watch. Collect data. Wait for the pattern to become undeniable.

By the time it was undeniable, you weren’t curious anymore. You were building a case.

This is a decision error, not a character flaw. You treated a calibration conversation like a corrective action, a two-way door like a one-way door.

And waiting created the stakes. The conversation that was an adjustment six months ago, it’s now a performance issue. You caused that through delay.

I had this happen to me years ago. I had a young engineer, good attitude, sharp, great work ethic, but he raced through his work. It kind of reminded me of myself in high school, trying to be the first one to finish the exam. Speed over accuracy. I saw the pattern forming early, but I thought maybe it’s just the exuberance of youth.

So I said nothing. Instead, I tried to model the behavior. Slower, more thorough, hoping it would transfer. That’s an engineer’s instinct. Solve it indirectly.

It didn’t work. A few months later, he stepped into a team lead role on a high pressure assignment. It got to the point where he was basically announcing it. Look how fast I can get to 80% done. Why can’t everyone?

It blew up. Our VP overseeing the project called both of us in. My young engineer was removed from the team lead role. He came close to being removed from the team entirely.

And by the time I finally spoke with him directly, it was no longer a calibration conversation. I was delivering a verdict.

That’s what waiting does. You don’t avoid the hard conversation. You just make it harder.

You have the framework from the last episode. Here’s what it looks like in this specific conversation. When we look at the signal before the noise, open with what was observed, not what was concluded. In this case, it could have sounded like, I’m seeing you move fast and land at about 80% accurate. There’s budget left to close that gap. Let’s talk about that. That’s observable. That’s receivable. Your work isn’t good enough is a verdict. Nobody responds well to a verdict.

Next, lower the cost of dissent. Say it out loud. I might be missing something. Help me understand what’s going on. What’s driving the pace? And you need to mean it. Invite the feedback, invite the pushback. Because if they can’t push back, you’ll get agreement in the room and nothing will change outside of it.

And then you need clear movement. One behavior, one timeframe, one check-in. Something like, take one more pass before you submit that. Then we’ll check in at your next review. Anything vaguer than that isn’t a conversation. It’s a warning shot.

The signal to act isn’t when the evidence is airtight. In leadership, you cannot wait for 100% certainty. You need to move when it still feels slightly premature. That is the calibration window. Before it becomes discipline. Before you stop being curious. Before the relationship has already absorbed the cost of your silence, and all of the emotional weight you’ve been quietly adding to it.

You have the framework now. The question isn’t whether you know how to have the conversation. It’s whether you’ll use it before you’re sure you have to. Because the conversation you’re avoiding right now is easier than the one you’re creating by waiting. That’s the whole thing. Timing. Early intervention. Not to avoid difficulty, to avoid the much harder conversation that comes from delay.

Now there’s a deeper reason these conversations get avoided. It’s not just discomfort. It’s that a lot of engineers haven’t updated their identity to include this kind of work. That’s what we’re into next episode.