You told a colleague the analysis would be ready for Thursday’s steering group. It wasn’t. They went in without it.
That’s one let-down. What it does to their confidence in you depends on what they think caused it, and they’ll have a view whether or not you give them one.
Is every let-down the same kind of break?
No. Two very different stories can sit behind the same missed Thursday.
In the first, you tried and couldn’t. The data was messier than you expected, you underestimated the work, and you ran out of road. The question now open for them is Do they believe I can handle what’s needed? (Capability & Realism).
In the second, you knew by Monday it wasn’t going to happen and said “all on track” anyway. Or you blamed the data team when it wasn’t them. Now the open question is a different one: Do they believe what I say? (Clarity & Honesty).
Trust is someone’s confidence in you with something they care about. The first story dents their confidence in what you can do. The second dents their confidence in what you tell them, and that one sits underneath everything else, because it’s how they judge every update you give from now on.
The distinction isn’t ours. Researchers separate competence-based breaches from integrity-based ones, and they’ve found that what helps with one doesn’t reliably help with the other.
Why does the kind of break change the repair?
Because an apology means something different in each case.
Owning a capability failure says: I misjudged this, I can learn, here’s what I’ll do differently. Most people accept that. Everyone misjudges something eventually, and one misjudgement doesn’t mean you’ll always misjudge.
Owning an integrity failure says: yes, I did shade the truth. That confirms the very thing they were worried about. It’s still the right thing to do. It just doesn’t carry the same reassurance, because it’s harder to believe that someone who misled you once won’t do it again.
The research points the same way. In experiments by Peter Kim and colleagues, apologising repaired trust more successfully after an alleged competence breach than after an alleged integrity one. In a later study of how apologies are built, Roy Lewicki and colleagues found people rated apologies after competence breaches as more effective than apologies after integrity breaches. Both used written scenarios about people the participants didn’t know, so read them as a direction rather than a forecast about your colleague.
There’s a second finding in the Kim studies that’s easy to skip. Where later evidence showed the person really had done it, apologising had served them better than denying it. If you did it, say so. A denial that later unravels leaves you with two breaches instead of one.
Staying quiet doesn’t get you out of it either. A follow-up study by Donald Ferrin and colleagues tested “I’d rather not comment” responses and found them a weak choice for both kinds of breach. Silence doesn’t own the failure, and it doesn’t show you’ll do better.
Can trust come back after you weren’t straight with someone?
Partly, and slowly. Sometimes fully. You don’t get to decide which.
The most uncomfortable finding comes from Maurice Schweitzer and colleagues. In a repeated laboratory game, trust damaged by untrustworthy behaviour came back when people then saw a consistent run of trustworthy behaviour. Trust damaged by the same behaviour plus deception never fully recovered, even when the deceived participants received a promise, an apology and a consistent run of trustworthy behaviour afterwards. That was 262 participants, nearly all aged 19 to 22, playing for money rather than working together. It still matches something most people recognise. Being let down and being lied to aren’t the same injury.
So be honest with yourself about which one happened. “I was optimistic” and “I knew and didn’t say” are different things, and the other person will often know which it was before you admit it.
What should you do differently for each kind of break?
The order of repair doesn’t change. Our repair page sets it out once: what the break meant to them, what can be put right now, and what evidence they’ll need next. What changes is where the weight sits.
After a capability let-down, the evidence carries most of the weight. They need to see you handle the next one. That points at the four behaviours under Capability & Realism:
- Only agree to what you can realistically deliver and control.
- Prepare before key interactions and decisions.
- Ask for support when needed.
- Say clearly when something cannot be delivered.
In practice, the next commitment you make to them is one you’re sure of. If you’re stretched, say so early and ask for help rather than hoping. The repair is the next piece of work arriving as described.
After an integrity let-down, the ownership changes. It has to be complete and unprompted: what you said, what was actually true, and when you knew. No “there may have been some confusion”. Then two behaviours under Clarity & Honesty become the evidence:
- Say what is true without using vague or confusing language.
- Share relevant information openly, not selectively.
This kind of repair is slower because the evidence is harder to see. A kept deadline is visible. Being straight is only visible when it costs you something, so the moment that counts is the next time you have bad news and tell them before they find out. Raising problems early is where that shows.
What if they think it was integrity and you think it was capability?
Then their view is the one you’re repairing, not yours.
You might be certain you were simply over-optimistic. If they experienced it as being misled, because you said “on track” twice when it wasn’t, the question they’re asking is about your word. Arguing over the label won’t help. Ask what it was like from their side and own the part that’s true. “I said it was on track when I didn’t actually know. I should have told you I wasn’t sure.” That covers both.
Repair is about their confidence, and they decide when it’s back. You can’t ask for it, schedule it or point to how well you’ve done since. You can only keep giving them reasons.
When this isn’t the whole picture
Some let-downs aren’t one person’s failure. The work was undeliverable in the time, the priorities moved, or nobody decided who owned it. Say that honestly too, rather than taking all the blame to look gracious or none of it to look blameless.
If anything here involves a formal process, an investigation or a concern about someone’s safety, that comes first and nothing on this page overrides it.
The split between competence and integrity breaches comes from published research. Mapping it to the five Trust Questions and the twenty behaviours is Trust Leader’s own work, not a tested method, and none of the studies below test the Trust Leader approach. Reviewed 4 October 2026 by Scott Hunter.