What to say when someone repeats the same mistake

You've given the feedback. Clearly, kindly, anchored to a specific moment — you did it properly. And then it happens again. So you give it a second time, slightly more firmly. It happens a third time. Now you're in the situation every manager eventually lands in, and the instinct is almost universal: say the same thing again, louder, with a harder edge. That instinct is wrong, and understanding why changes the whole conversation. Feedback that hasn't landed after two or three honest attempts is no longer telling you something about the other person's motivation. It's telling you something about your diagnosis. Repetition means you've been treating a problem you haven't correctly identified yet — and the third round of the same message will fail exactly like the first two did.

Why repeating it doesn't work

The unspoken assumption behind repetition is that the person heard you, understood you, could have done otherwise, and chose not to — so what's needed is more pressure. Occasionally that's true. Usually it isn't, and the assumption is doing real damage: each repetition without a change in your approach costs you credibility, teaches the person that your feedback is noise they can absorb rather than signal they must act on, and gradually converts a solvable problem into a relationship problem.

There's also a subtler cost. By the third time, the conversation has usually shifted, in both of your heads, from the behavior to the person — from "this thing keeps happening" to "you're the kind of person who does this." Once it becomes an identity claim, defensiveness is guaranteed and learning stops. Knowledge at Wharton's work on behavior change at work keeps landing on the same point: persistent behavior usually has a cause that pressure alone can't reach.

The five gaps behind a repeated mistake

Before you say anything a third time, work out which of these you're actually looking at. They look identical from the outside and need completely different responses.

• The clarity gap — they don't know what "right" looks like. You told them what was wrong; you may never have described the correct version concretely. "Don't ship before design sees it" doesn't tell someone at which point, to whom, or in what form. People can't hit a target they can only describe in the negative.
• The awareness gap — they can't see themselves doing it. They agree with you completely in the abstract and genuinely don't detect it in the moment. This is the most common one and the least addressed, because more feedback about the pattern doesn't help someone who can't perceive the instances. What helps is a signal in the moment — a checklist, an agreed cue, a peer who flags it live.
• The constraint gap — the system makes the right behavior expensive. They skip the review step because the deadline you set doesn't fit including it. They don't delegate because they're measured on output this week. When the environment punishes the behavior you're asking for, the behavior loses — every time. INSEAD's research on organizational behavior is consistent that structure usually beats exhortation. If this is your gap, the fix is yours to make, not theirs.
• The capability gap — they can't yet, rather than won't. Giving clearer priorities requires being able to identify priorities; running a tight meeting is a skill. Feedback tells someone that they need to improve; it doesn't teach them how. This gap needs training, pairing, or a worked example — not another conversation about the standard.
• The buy-in gap — they disagree and haven't said so. They think your way is wrong, or that the tradeoff isn't worth it, and they've decided arguing with you isn't worth the friction. So they nod, and then keep doing what they think is right. This is the most frequently missed of the five, because from the outside compliance-then-relapse looks exactly like carelessness.

Notice that only one of these five — a genuine buy-in problem that survives discussion — is remotely close to "won't." The others are all things a manager can fix, and four of them are made worse by pressure.

Repeating vs diagnosing

The pattern conversation is a different conversation

Here's the specific move that unlocks this, and it's worth naming because most managers never make it: at some point you have to stop giving feedback about instances and have a single explicit conversation about the pattern. These are genuinely different conversations. Instance feedback says "in Thursday's review, X happened." Pattern conversation says "this is the third time X has happened, so something structural is going on and I want to figure out what."

The pattern conversation has three properties that make it work. It's explicit — you say out loud that it's recurring, which stops the polite fiction that each instance is isolated. It's collaborative — you're investigating a cause together rather than issuing a third verdict. And it assumes a reason exists, which is both more respectful and more accurate than assuming indifference. In practice that sounds like: "I've raised this three times, which usually means I'm missing something. What's making this hard?" Nine times out of ten the answer is one of the five gaps, and the person tells you which — often immediately, because nobody enjoys repeatedly disappointing their manager and they've usually been sitting on the reason.

Then, and only then, escalate

Diagnosis first doesn't mean infinite patience. If you've named the pattern, found no clarity, awareness, constraint, or capability gap, and there's no real disagreement — or there was and you've resolved it — then you're looking at a genuine performance issue, and it deserves a clear conversation about expectations and consequences. That conversation is management, not feedback, and it should be unambiguous: what needs to change, by when, and what happens if it doesn't.

The order matters enormously, though. Escalating before diagnosing is how managers end up putting someone on a formal warning for a problem the manager themselves created with an impossible deadline or a standard they never actually specified. Escalating after diagnosing is fair, defensible, and — because you've eliminated the other explanations — usually lands as reasonable even to the person receiving it. And if the whole exchange keeps stalling on "that's not what happened," the issue may be that the original feedback wasn't specific or observed closely enough to discuss as fact.

How to run this

1. 1

   Stop after the second repetition

   Treat "I've said this twice" as a hard trigger to change approach rather than volume. A third identical attempt is wasted, and it costs you credibility you'll want later.
2. 2

   Say the pattern out loud

   Name explicitly that it's recurring, and frame it as a shared puzzle rather than a verdict: "third time, so I think something structural is going on." This ends the fiction that each instance stands alone.
3. 3

   Diagnose which of the five gaps you're in

   Ask what happens in the moment. Do they know what right looks like? Can they tell when it's happening? Is something making the right choice expensive? Can they actually do it? Do they quietly disagree? Let them answer before you decide.
4. 4

   Match the fix to the gap

   A clarity gap needs a concrete description of good. An awareness gap needs an in-the-moment signal. A constraint gap needs you to change something. A capability gap needs teaching. A buy-in gap needs an honest argument you're willing to lose.
5. 5

   Escalate only once the gaps are ruled out

   If none of them applies, have the clear expectations-and-consequences conversation. Skipping straight here is unfair; arriving here after diagnosis is legitimate — and expect the defensive reflex either way.

Where Mirorly fits

Look again at the awareness gap — someone agreeing with feedback in principle while being unable to detect the behavior in the moment. That isn't a character defect; it's the ordinary condition of being human, and it's the exact reason feedback needs to be specific and repeated back to people over time. Which raises the uncomfortable question: what's your repeated pattern? The one two or three people have noticed, mentioned once in a softened way, and then dropped when it didn't land — because you're their manager and pushing a third time isn't worth it to them. Mirorly is how you find out: your team answers specific behavioral questions about how you actually work, aggregated with an anonymity threshold so honesty is safe, and asked again next quarter so a pattern shows up as a pattern rather than a one-off comment. The core leadership behaviours template is built for that, and it's the closest thing there is to seeing your own awareness gap from the outside.

Common questions

The one-line summary

When the same mistake keeps recurring after two honest attempts at feedback, the repetition is information about your diagnosis rather than their motivation — so stop saying it louder, name the pattern explicitly as a shared puzzle, and work out which of the five gaps you're in (clarity, awareness, constraint, capability, or buy-in), because four of them are yours to fix and all of them are made worse by pressure; escalate only once you've genuinely ruled them out.