Try this the next time someone on the team needs to hear something difficult.
Plan the exact sentence in advance, then notice what actually comes out of your mouth once you’re in the room. Chances are it arrives with three qualifiers, a compliment stacked on each side, and a closing line that quietly cancels the first half of what was said. The other person leaves the conversation feeling fine and knowing nothing new.
That’s the usual trade with a colleague who takes feedback personally, and it happens almost automatically, without anyone deciding to make it.
The instinct behind it makes sense. Watching someone’s face fall, or hearing their voice tighten, is uncomfortable enough that softening the message can feel like the kind thing to do. A message softened past the point of usefulness stops functioning as feedback, though. It turns into a performance of care that leaves the actual problem untouched, and the other person is left to guess at what was really meant, sometimes walking away convinced everything is fine. Left unaddressed long enough, the same issue tends to resurface later in a much less forgiving form, usually during a performance review or a moment when there’s no more room for a gentle version.
The steps below are about keeping the substance of the message intact while still accounting for how someone this sensitive is likely to receive it.
1. Anchor it to one specific moment
Vague feedback gives a defensive listener room to argue about the wrong thing entirely. A comment like “you’ve been dismissive in meetings lately” invites a debate about whether that’s even true, because there’s nothing concrete to point to.
Naming one meeting, one exchange, or one specific email gives the conversation a fixed reference point that’s much harder to relitigate than a general impression. The Center for Creative Leadership built an entire feedback framework, known as situation-behavior-impact, around this idea. It starts every piece of feedback with the exact situation before naming anything else, because a floating generalization is usually where these conversations go wrong first.
2. Describe the specific behavior
The next line matters just as much as the first. Naming a specific action, like noting that a colleague interrupted a teammate twice on the same call, describes something that happened. A phrase like “you don’t respect other people’s ideas” reaches for something bigger: a verdict about character, which is a far harder thing to sit with than a description of one moment.
Sheila Heen and Douglas Stone, who teach negotiation at Harvard Law School, describe this as an identity trigger: the point where feedback stops sounding like information about an action and starts sounding like a verdict on who someone is. Once that trigger fires, the actual content of the feedback barely registers, because the listener is busy defending a sense of self instead of processing a message. Staying with observable behavior, and leaving out words like “always” or “never,” is one of the few things that keeps that switch from flipping.
3. Attach a concrete impact
Naming a behavior without saying what it caused can come across as nitpicking, so the situation-behavior-impact model pairs the two anyway. Together they give the other person something specific to work with: what happened, and what it actually caused, whether that’s a missed deadline, a frustrated client, or a colleague who quietly stopped bringing ideas to the table.
This is also the point where most people start over-explaining, adding extra examples and hedging qualifiers meant to pre-empt objections nobody has raised yet. One clear impact statement, said once, does more work than five hedged ones stacked on top of each other, and it leaves less room for the conversation to drift into reassurance instead of information. “This meant the client presentation went out with the wrong numbers” lands, and stays landed, in a way that a paragraph of caveats never quite manages.
4. Say the actual point first
Kim Scott managed teams at Google and Apple before writing about a framework she calls radical candor. She describes the more common failure as ruinous empathy: caring about someone’s feelings so much that the real message never actually gets delivered. Scott has said, “We’re conditioned from an early age to avoid hurting people’s feelings.”
That conditioning is exactly what produces the compliment sandwich and the long wind-up before the actual point ever arrives. With someone who takes feedback personally, the instinct is often to add even more padding, and that tends to backfire twice: it buries the message the listener actually needs, and it signals, correctly, that the message is being treated as too fragile to say plainly. None of this requires a harsh tone. It just means the first sentence should already contain the point, rather than working up to it three paragraphs later.
5. Let the reaction happen without managing it
A person who takes feedback personally is likely to look upset, go quiet, or push back harder than the moment seems to call for. That reaction is coming from the same identity trigger described above, not from anything specifically wrong with the feedback itself, so it isn’t a sign that the message needs to be walked back on the spot. It’s possible to sit through someone else’s discomfort without rescuing them from it, and without turning the rest of the conversation into damage control for their reaction instead of the original point.
A short acknowledgment that the moment is hard, something as plain as “I know this isn’t easy to hear,” followed by returning to the same specific behavior and impact already named, tends to work better than a fresh round of softening. The next follow-up conversation should pick up that same thread, not reopen a negotiation over how the last one felt.
Final words
The conversation may still be uncomfortable, but discomfort is not the same as harm. What matters is whether the other person leaves with something they can actually act on, rather than a vague sense that something was wrong and a much clearer sense that the conversation is finally over.
The goal, then, is not to make difficult feedback feel painless. It is to make it clear enough to be useful and respectful enough to be heard. That means naming the moment, describing the behavior, explaining the impact, and resisting the urge to dilute the message when the reaction is difficult.
Someone may still take it personally. They may need time before they can respond well. Neither reaction means the conversation failed. The more meaningful measure is whether they leave knowing exactly what happened, why it mattered, and what needs to change.
Clarity can feel unkind in the moment, especially when reassurance would make the room more comfortable. But reassurance that obscures the truth is not kindness. The kinder choice is to trust the other person with a message they can actually use.
