How to give feedback that actually changes behavior
Have you ever left a feedback conversation feeling like you nailed it, only to watch the person do the exact same thing two weeks later? You were clear. You were kind. You gave examples. And nothing changed.
The problem usually isn't courage. Most people are willing to give feedback. The problem is architecture. Feedback that changes behavior is structured differently than feedback that checks a managerial box.
I've spent twenty years coaching leaders and sales professionals, and the pattern is consistent: the feedback that sticks follows a specific design. The feedback that evaporates follows a different one. The gap between the two is learnable.
"You seemed disengaged in the meeting" is an interpretation. "You checked your phone four times during the client presentation" is an observation. The difference matters because people can argue with your interpretation. They can't argue with what happened.
Start every piece of feedback with what you saw or heard. Not what you think it meant. Not what you assume was going on internally. Just the observable behavior. This does two things: it removes defensiveness (you're describing facts, not making accusations) and it gives the other person a specific thing to change.
"When you interrupted Sarah during the review, it made me frustrated" centers the conversation on you. "When you interrupted Sarah during the review, the client shifted their attention away from our proposal" centers it on business impact.
Both are true. One changes behavior. The other creates guilt.
That's it. No editorializing. No stacking three examples when one would do. One behavior, one impact, one conversation.
The instinct after delivering feedback is to tell the person what to do differently. Resist it. Instead, ask: "What do you think you could do differently next time?"
Nine times out of ten, they already know. They just haven't been asked. And when someone generates their own solution, they're far more likely to follow through than when they're handed a directive from above.
This is Socratic coaching in action. The question does more work than the answer. When you ask someone to diagnose their own behavior, you're building their capacity to self-correct. That's the difference between fixing a moment and building a skill.
Feedback has a shelf life. The closer to the event, the more useful it is. The further away, the more it feels like an ambush.
The sweet spot is within 24 hours for most situations. Same day if possible. If you find yourself saving feedback for a quarterly review, you've already lost most of its value. The person can't remember the specifics. You can't remember the specifics. And the conversation becomes a negotiation about what actually happened instead of what to do about it.
There's one exception: if you're angry, wait. Feedback delivered in frustration is about you, not about them. Give yourself enough time to separate the emotion from the observation. Then deliver it.
Feedback given in public activates shame. Feedback given in a hallway signals it's not important. Feedback given in a formal setting signals it's serious, even when it's not.
For routine developmental feedback, the best setting is a regular one-on-one where feedback is expected. When it shows up every week in a predictable format, it stops feeling like an event and starts feeling like part of the work. That's what you want: feedback as operating rhythm, not as incident response.
You probably have one piece of feedback you've been holding. Strip it down to the observable behavior and the specific business impact. Try delivering it in the next 48 hours with this structure: "When you [behavior], the impact was [consequence]. What would you do differently?"
Then stop talking. The silence does its own work.
Drop your email and the PDF lands in your inbox immediately.
What's next
Feedback is one of the highest-leverage skills a leader can develop, and one of the hardest to self-coach. If you want to go deeper, coaching is where this work happens.
Book a discovery call →