🎧 Listen to this article

An AI-generated podcast companion to this article, produced with Google Gemini Notebook.

You asked a colleague on your project team to send you the budget variance numbers for Thursday's leadership meeting. They said yes. Thursday morning you check your inbox and the shared folder, and neither has anything. Nobody warned you. You present without the numbers, and leadership pushes the decision on your hiring request out a month.

Most people handle the next conversation badly, and they do it in one of two ways. Some absorb the failure, redo the work at midnight, and say nothing. Others do say something, but the words come out sharp, and the other person spends the whole exchange defending themselves. Both responses feel justified in the moment. Neither repairs anything.

There is a better option. I call it OODER, a five-step conversation for restoring trust after someone has let you down: Outcome, Open, Describe, Explain, and Repair and Recommit. The steps are easy to learn. The harder skill is knowing when this conversation is the right tool, and what you need to settle with yourself before you start it. That is where most of this article goes.

Is This Actually an OODER Moment?

OODER is built for one situation: someone made a commitment, didn't keep it, and didn't tell you in time. Three questions help you check whether that's what you're looking at. Each one can change what you say, or whether you say anything at all.

Did they actually commit?

Think back to how you made the request. If you said, "It would be great to have that by Thursday," and they said, "Sure," you had a hopeful exchange. You did not have a commitment. Their yes may have meant only that they heard you.

I lay out how to close that gap in ACCORD, particularly the Confirm and Obligate steps. What matters here is the diagnosis. If the commitment was never confirmed, part of the breakdown is yours, and your opening should say so. If it was clear and explicit, you can start from firmer ground.

The same goes for context. In Get on the Boat and Row, I describe three reasons a request stalls: Commitment, Context, and Being. Before you decide this is about the person, check the second one. If their own manager pulled them onto something else on Wednesday, you aren't looking at a character problem. You may still need the conversation, but the repair will look different.

Did they tell you before the deadline?

A missed commitment and a broken one are not the same thing. Suppose your colleague had come to you on Tuesday and said, "I can't finish this by Thursday. Here's what I can get you instead." The work still wouldn't have arrived, but integrity would have held. The right response to that is thanks and a new plan.

That is the norm I call Accountability in GRRR + BSHAARP: if you can't meet a commitment, tell the person affected before the deadline, not after. OODER is for the case where that norm wasn't followed. The silence is the break, more than the missed date.

If your team never agreed to the norm, you can still have the conversation. You are just doing an extra job inside it, because the Repair and Recommit step is where you propose the agreement you wish had existed. That matters most when you can't hand out consequences. Between peers, or with people who report to someone else, a conversation built on agreement is often the only lever you have.

What kind of lapse was it?

Not every missed commitment is the same kind of failure. In Innovation Runs on Failing Well, I describe the just culture framework developed by David Marx and Sidney Dekker, which sorts behavior into three categories. An honest error gets support. At-risk behavior, where someone has drifted into a shortcut and stopped seeing the risk, gets coaching. Reckless behavior, a conscious disregard of a known risk, gets discipline. The same sorting applies to a letdown.

A colleague who misread the deadline made an honest slip. One who has quietly stopped prioritizing anything outside their own function's work has drifted. One who knew they wouldn't deliver and chose not to say so is a different case. OODER fits all three, but the weight shifts. For a slip, the conversation can be short. For drift, the follow-up about what got in the way carries most of the load. For the third, the conversation is still worth having, and it is also the point where you begin deciding whether consequences need to be formal.

The Five Steps

If the answers point to OODER, here is the conversation itself.

O
Outcome Decide what you want to be true when the conversation ends
O
Open Ask permission and start from ownership
D
Describe State what happened, facts first, including your part
E
Explain Connect the impact to the shared goal
R
Repair and Recommit Invite a fix, then ask for a clear commitment

O — Outcome: Know What You're After

Before you speak, get clear on why you're having the conversation. Some outcomes feel good and accomplish nothing: wanting them to know they messed up, needing to vent, wanting to show them who's in charge. The outcome worth aiming for is to restore trust so the two of you can work effectively toward your shared goal.

A useful test: imagine they apologize perfectly and nothing about how they work with you changes. If that would feel like a failure, acknowledgment is not your outcome. Trust is. Acknowledgment is easier to get and worth far less.

This is the same discipline I describe as the first stage of a strategic communication cycle: name the result before you draft the message. The difference here is that your own frustration is the main thing pulling you toward the wrong result. If you can't state your outcome in one sentence you'd be comfortable saying to their face, you aren't ready to have the conversation yet. Give it a day.

Example
"I want us to leave this conversation with a plan for the numbers and a clear way to count on each other for the next deadline."

O — Open: Ask Before You Start

The purpose of Open is not courtesy. It is to raise the odds that the other person arrives in a place where the outcome you chose is reachable: calm enough to listen, secure enough not to defend, and willing to solve the problem with you. Everything in this step serves that.

Ask for a moment, and get permission. Catching someone off guard puts them on defense before you've said anything useful. Choose a live conversation in private over a message in the team thread, which strips out tone and gives them an audience.

Both of those choices are about risk, and the two compound. A message in email or chat strips out tone, so a neutral sentence reads as cold and a careful one reads as passive-aggressive. It also can't adjust. You can't see the reaction, soften a phrase halfway through, or notice that the other person is stunned. And it leaves a record that gets forwarded, screenshotted, and reread later in a worse mood. A text or a direct message feels casual, but it carries the same risks.

An audience changes the question the other person is answering. In private, they decide how to respond to you. In front of others, whether in a meeting, on a call, or in a team thread, they decide how to protect their standing, and that usually means defending, deflecting, or pushing back. Even a fair, factual description lands as a public correction. The people watching also become participants. Some take sides, and others quietly decide what it would cost them to be next. That can damage trust in you more than the missed deliverable did.

Written notes still have a place. Use one to ask for the private conversation, or to confirm the repair plan afterward. It is a poor place for the conversation itself. If the lapse already surfaced in front of others, acknowledge it briefly in the moment and take the real conversation offline.

Asking permission can feel like deference when you're the one who was let down. It isn't. It gives the other person a measure of control, which raises the odds they arrive ready to solve the problem. They can say "later today." They can't say "never," not credibly, and you've lost nothing.

Example
"I'd like to talk with you about the variance numbers I was counting on for Thursday's meeting. Do you have a few minutes now, or would later today work better?"

D — Describe: What Happened, Including Your Part

State the breakdown as a sequence of facts. Skip labels like "irresponsible" or "sloppy." A label invites an argument about the label. A fact invites a response.

Speak from your own experience, with "I" statements. "I understood the numbers would arrive Wednesday" is a description. "You blew the deadline" is a verdict, and a "you" statement invites the defensiveness you're trying to avoid. Staying in a response-able frame, a term I take from Fred Kofman's work on unconditional responsibility, means you describe what you saw and what you did or didn't do, and leave motives and character out of it.

This is also where you account for your own role, and there is a fair objection to that: why should I take responsibility for something they did? You shouldn't, unless part of it is yours. If you never confirmed the deadline, say so. If you did everything right, don't invent a share of the blame. Performed humility is easy to spot, and it makes the rest of what you say harder to trust.

Example
"I understood the numbers would be in the shared folder by Wednesday evening. When I checked Thursday morning, they weren't there, and I hadn't heard from you. I also never confirmed that deadline with you in writing, and I should have checked in on Wednesday."

E — Explain: Connect the Impact to the Goal

Now tie the breakdown to what it cost, and keep it about the work. This is not a guilt trip. It is the reason the conversation matters. Name the specific consequences other people felt, because specifics are harder to dismiss than general disappointment.

Then stop talking. Explain is the step where most people keep going, and the extra words turn information into accusation.

Example
"Without them, I presented the request unsupported, and leadership pushed the decision to next month. Two teams were waiting on that approval before they could post the open roles."

R — Repair and Recommit: Build the Path Back

Invite them to propose how to make it right. The person who broke the commitment should own the fix, and people keep plans they designed more reliably than plans they were handed.

Example
"I know the leadership team was really counting on this. What do you think you can do to make this right?"

Sometimes the ship has sailed. The meeting is over, and nothing they do now can change the outcome. Then making it right is not about recovering the work. It means acknowledging what happened and what it cost, and recommitting on the next one. You can invite that without demanding it: "I can't get Thursday back. What do you want to do from here?" An apology offered freely carries weight. One extracted under pressure does not.

Once there's a plan, ask for a clear recommitment. It is the same ask as the Obligate step in ACCORD, with one difference: you now have evidence about what gets in the way, so you can ask a better version of the question.

Example
"That works. Can I count on you to have the numbers to me by Monday at noon?"

Then schedule a short follow-up. Use it to ask what got in the way last time, what would help prevent it next time, and what they need from you or the team. Integrity isn't restored when the work arrives. It is restored when trust returns, and trust returns after the next commitment is kept.

One distinction is worth holding onto through this whole step. Keeping your word means doing what you said you would do. Honoring your word, a distinction I draw from the integrity work of Werner Erhard and Michael Jensen, which Jensen lays out in a 2009 interview in Rotman Magazine, is what you do when you can't: you tell the people counting on you as soon as you know, and you clean up what the lapse left behind. Your colleague didn't keep their word. This conversation is their chance to honor it. If they decline, or can't commit to a repair they can deliver, that matters too. A plain "I can't take this on right now" is worth more than a yes you can't rely on.

When It Happens Again

One lapse followed by a real repair is a repaired relationship. A second lapse deserves a second conversation and a closer look at your diagnosis. By the third, the pattern is itself information.

Before you conclude the person is the problem, recheck the ground you've already covered. Was the commitment explicit? Was the context workable? Did you ask what they need to be able to deliver? If all three hold and the pattern continues, you are looking at someone who is unwilling or unable to honor their commitments. That person may not belong on the team, or at least not in a role others depend on.

Peers often can't make that call. What you can do is raise it with whoever can, restructure the work so you stop depending on that person, or both. It can feel harsh, and it's tempting to keep extending chances. But a team that keeps absorbing the same failure pays for it in everyone else's trust.

The quieter alternative is to take the work back and do it yourself. It's fast, it avoids the conversation, and it protects the deadline. Once, to save Thursday, that is a reasonable decision. As a habit, it teaches the team that dropping something has no consequence, and it teaches you that you're the one who catches it. I've written about where that road leads in The Indispensability Trap.

The Question

Next time someone lets you down, ask one thing before you plan a single word: did I have a commitment, or did I have a hope? The answer tells you which conversation you need.

Christensen, K. (2009). Integrity is a matter of a person's word – nothing more and nothing less [Interview with M. C. Jensen]. Rotman Magazine, Fall, 16–20.

Dekker, S. (2016). Just culture: Balancing safety and accountability (3rd ed.). CRC Press.

Erhard, W., & Jensen, M. C. (2017). Putting integrity into finance: A purely positive approach. Capitalism and Society, 12(1).

Kofman, F. (2006). Conscious business: How to build value through values. Sounds True.

Marx, D. (2001). Patient safety and the "just culture": A primer for health care executives. Columbia University.